svn commit: r212616 - head/sys/dev/mps

[Kenneth D. Merry]

Author: ken
Date: Tue Sep 14 17:22:06 2010
New Revision: 212616
URL: http://svn.freebsd.org/changeset/base/212616

Log:
  MFp4: (//depot/projects/mps/...)
  
  Report data overruns properly.
  
  Submitted by: scottl

Modified:
  head/sys/dev/mps/mps_sas.c

Modified: head/sys/dev/mps/mps_sas.c
==
--- head/sys/dev/mps/mps_sas.c  Tue Sep 14 16:22:22 2010(r212615)
+++ head/sys/dev/mps/mps_sas.c  Tue Sep 14 17:22:06 2010(r212616)
@@ -1219,11 +1219,9 @@ mpssas_scsiio_complete(struct mps_softc 
ccb->ccb_h.status = CAM_REQ_CMP;
break;
case MPI2_IOCSTATUS_SCSI_DATA_OVERRUN:
-   /*
-* XXX any way to report this?
-*/
+   /* resid is ignored for this condition */
ccb->csio.resid = 0;
-   ccb->ccb_h.status = CAM_REQ_CMP;
+   ccb->ccb_h.status = CAM_DATA_RUN_ERR;
break;
case MPI2_IOCSTATUS_SCSI_INVALID_DEVHANDLE:
case MPI2_IOCSTATUS_SCSI_DEVICE_NOT_THERE:
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r212624 - head/share/man/man4

[Kenneth D. Merry]

Author: ken
Date: Tue Sep 14 21:40:29 2010
New Revision: 212624
URL: http://svn.freebsd.org/changeset/base/212624

Log:
  MFp4 (//depot/projects/mps/...)
  
  Add a man page for the mps(4) driver, and reference it in the
  mpt(4) driver man page.
  
  Sponsored by: Spectra Logic Corporation

Added:
  head/share/man/man4/mps.4   (contents, props changed)
Modified:
  head/share/man/man4/Makefile
  head/share/man/man4/mpt.4

Modified: head/share/man/man4/Makefile
==
--- head/share/man/man4/MakefileTue Sep 14 20:55:25 2010
(r212623)
+++ head/share/man/man4/MakefileTue Sep 14 21:40:29 2010
(r212624)
@@ -219,6 +219,7 @@ MAN=aac.4 \
mmcsd.4 \
mn.4 \
mouse.4 \
+   mps.4 \
mpt.4 \
msk.4 \
mtio.4 \

Added: head/share/man/man4/mps.4
==
--- /dev/null   00:00:00 1970   (empty, because file is newly added)
+++ head/share/man/man4/mps.4   Tue Sep 14 21:40:29 2010(r212624)
@@ -0,0 +1,159 @@
+.\" 
+.\" Copyright (c) 2010 Spectra Logic Corporation
+.\" All rights reserved.
+.\" 
+.\" Redistribution and use in source and binary forms, with or without
+.\" modification, are permitted provided that the following conditions
+.\" are met:
+.\" 1. Redistributions of source code must retain the above copyright
+.\"notice, this list of conditions, and the following disclaimer,
+.\"without modification.
+.\" 2. Redistributions in binary form must reproduce at minimum a disclaimer
+.\"substantially similar to the "NO WARRANTY" disclaimer below
+.\"("Disclaimer") and any redistribution must be conditioned upon
+.\"including a substantially similar Disclaimer requirement for further
+.\"binary redistribution.
+.\" 
+.\" NO WARRANTY
+.\" THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+.\" "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+.\" LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR
+.\" A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+.\" HOLDERS OR CONTRIBUTORS BE LIABLE FOR SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+.\" STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
+.\" IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+.\" POSSIBILITY OF SUCH DAMAGES.
+.\" 
+.\" mps driver man page.
+.\"
+.\" Author: Ken Merry 
+.\"
+.\" $Id: //depot/SpectraBSD/head/share/man/man4/mps.4#1 $
+.\" $FreeBSD$
+.\"
+.Dd September 13, 2010
+.Dt MPS 4
+.Os
+.Sh NAME
+.Nm mps
+.Nd LSI Fusion-MPT 2 Serial Attached SCSI driver
+.Sh SYNOPSIS
+To compile this driver into your kernel,
+place the following lines in your kernel configuration file:
+.Bd -ragged -offset indent
+.Cd "device scbus"
+.Cd "device mps"
+.Ed
+.Pp
+Or, to load the driver as a module at boot, place the following line in
+.Xr loader.conf 5 :
+.Bd -literal -offset indent
+mpt_load="YES"
+.Ed
+.Sh DESCRIPTION
+The
+.Nm
+driver provides support for LSI Logic Fusion-MPT 2
+.Tn SAS
+controllers.
+.Sh HARDWARE
+The following controllers are supported by the
+.Nm
+driver
+.Pp
+.Bl -bullet -compact
+.It
+LSI Logic SAS2004 (4 Port
+.Tn SAS )
+.It
+LSI Logic SAS2008 (8 Port
+.Tn SAS )
+.It
+LSI Logic SAS2108 (8 Port
+.Tn SAS )
+.It
+LSI Logic SAS2116 (16 Port
+.Tn SAS )
+.It
+LSI Logic SAS2208 (8 Port
+.Tn SAS )
+.El
+.Sh CONFIGURATION
+To disable MSI interrupts for all
+.Nm
+driver instances, set the following tunable value in
+.Xr loader.conf 5 :
+.Bd -literal -offset indent
+hw.mps.disable_msi=1
+.Ed
+.Pp
+To disable MSI-X interrupts for all
+.Nm
+driver instances, set the following tunable value in
+.Xr loader.conf 5 :
+.Bd -literal -offset indent
+hw.mps.disable_msix=1
+.Ed
+.Sh DEBUGGING
+To enable debugging prints from the
+.Nm
+driver, set the
+.Bd -literal -offset indent
+hw.mps.X.debug_level
+.Ed
+.Pp
+variable, where X is the adapter number, either in
+.Xr loader.conf 5
+or via
+.Xr sysctl 8 .
+The following bits have the described effects:
+.Bl -tag -offset indent
+.It 0x01
+Enable informational prints.
+.It 0x02
+Enable tracing prints.
+.It 0x04
+Enable prints for driver faults.
+.It 0x08
+Enable prints for controller events.
+.El
+.Sh SEE ALSO
+.Xr cd 4 ,
+.Xr ch 4 ,
+.Xr da 4 ,
+.Xr mpt 4 ,
+.Xr pci 4 ,
+.Xr sa 4 ,
+.Xr scsi 4 ,
+.Xr targ 4
+.Sh BUGS
+This driver is still in development, it has only been tested on the amd64
+architecture and has some known shortcomings:
+.Bl -bullet -compact
+.It
+No IR (Integrated RAID) support.
+.It 
+No userland utility support (e.g.
+.Xr mptutil 8).
+.It
+Sometimes the driver gets into a state where devices arrive and depart multiple
+ti

svn commit: r212669 - head/share/man/man4

[Kenneth D. Merry]

Author: ken
Date: Wed Sep 15 16:05:51 2010
New Revision: 212669
URL: http://svn.freebsd.org/changeset/base/212669

Log:
  MFp4 (//depot/projects/mps/...)
  
  Fix a typo.
  
  Submitted by: pluknet at gmail dot com

Modified:
  head/share/man/man4/mps.4

Modified: head/share/man/man4/mps.4
==
--- head/share/man/man4/mps.4   Wed Sep 15 16:05:40 2010(r212668)
+++ head/share/man/man4/mps.4   Wed Sep 15 16:05:51 2010(r212669)
@@ -51,7 +51,7 @@ place the following lines in your kernel
 Or, to load the driver as a module at boot, place the following line in
 .Xr loader.conf 5 :
 .Bd -literal -offset indent
-mpt_load="YES"
+mps_load="YES"
 .Ed
 .Sh DESCRIPTION
 The
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r212772 - head/sys/dev/mps

[Kenneth D. Merry]

Author: ken
Date: Thu Sep 16 22:54:56 2010
New Revision: 212772
URL: http://svn.freebsd.org/changeset/base/212772

Log:
  MFp4 (//depot/projects/mps/...):
  
  According to the MPT2 spec, task management commands are
  serialized, and so no I/O should start while task management
  commands are active.
  
  So, to comply with that, freeze the SIM queue before we send any
  task management commands (abort, target reset, etc.) down to the
  IOC.  We unfreeze the queue once the task management command
  completes.
  
  It isn't clear from the spec whether multiple simultaneous task
  management commands are supported.  Right now it is possible to
  have multiple outstanding task management commands, especially in
  the abort case.  Multiple outstanding aborts do complete
  successfully, so it may be supported.
  
  We also don't yet have any recovery mechanism (e.g. reset the IOC)
  if the task management command fails.

Modified:
  head/sys/dev/mps/mps_sas.c

Modified: head/sys/dev/mps/mps_sas.c
==
--- head/sys/dev/mps/mps_sas.c  Thu Sep 16 22:38:27 2010(r212771)
+++ head/sys/dev/mps/mps_sas.c  Thu Sep 16 22:54:56 2010(r212772)
@@ -438,6 +438,7 @@ mpssas_prepare_remove(struct mpssas_soft
cm->cm_desc.Default.RequestFlags = MPI2_REQ_DESCRIPT_FLAGS_DEFAULT_TYPE;
cm->cm_complete = mpssas_remove_device;
cm->cm_targ = targ;
+   xpt_freeze_simq(sc->sassc->sim, 1);
mps_map_command(sc, cm);
 }
 
@@ -453,6 +454,7 @@ mpssas_remove_device(struct mps_softc *s
 
reply = (MPI2_SCSI_TASK_MANAGE_REPLY *)cm->cm_reply;
handle = cm->cm_targ->handle;
+   xpt_release_simq(sc->sassc->sim, 1);
if (reply->IOCStatus != MPI2_IOCSTATUS_SUCCESS) {
mps_printf(sc, "Failure 0x%x reseting device 0x%04x\n", 
   reply->IOCStatus, handle);
@@ -983,6 +985,11 @@ mpssas_abort_complete(struct mps_softc *
mps_printf(sc, "%s: abort request on handle %#04x SMID %d "
   "complete\n", __func__, req->DevHandle, req->TaskMID);
 
+   /*
+* Release the SIM queue, we froze it when we sent the abort.
+*/
+   xpt_release_simq(sc->sassc->sim, 1);
+
mps_free_command(sc, cm);
 }
 
@@ -1013,10 +1020,19 @@ mpssas_recovery(struct mps_softc *sc, st
cm->cm_data = NULL;
cm->cm_desc.Default.RequestFlags = MPI2_REQ_DESCRIPT_FLAGS_DEFAULT_TYPE;
 
+   /*
+* Freeze the SIM queue while we issue the abort.  According to the
+* Fusion-MPT 2.0 spec, task management requests are serialized,
+* and so the host should not send any I/O requests while task
+* management requests are pending.
+*/
+   xpt_freeze_simq(sc->sassc->sim, 1);
+
error = mps_map_command(sc, cm);
 
if (error != 0) {
mps_printf(sc, "%s: error mapping abort request!\n", __func__);
+   xpt_release_simq(sc->sassc->sim, 1);
}
 #if 0
error = mpssas_reset(sc, targ, &resetcm);
@@ -1361,7 +1377,13 @@ mpssas_resetdev(struct mpssas_softc *sas
cm->cm_data = NULL;
cm->cm_desc.Default.RequestFlags = MPI2_REQ_DESCRIPT_FLAGS_DEFAULT_TYPE;
 
+   xpt_freeze_simq(sassc->sim, 1);
+
error = mps_map_command(sassc->sc, cm);
+
+   if (error != 0)
+   xpt_release_simq(sassc->sim, 1);
+
return (error);
 }
 
@@ -1385,6 +1407,9 @@ mpssas_resetdev_complete(struct mps_soft
ccb->ccb_h.status = CAM_REQ_CMP_ERR;
 
mps_free_command(sc, cm);
+
+   xpt_release_simq(sc->sassc->sim, 1);
+
xpt_done(ccb);
 }
 
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r212802 - head/sys/dev/mps

[Kenneth D. Merry]

Author: ken
Date: Fri Sep 17 21:53:56 2010
New Revision: 212802
URL: http://svn.freebsd.org/changeset/base/212802

Log:
  Fix a couple of mps problems.
  
  When the driver is completely saturated with commands (1024 in the
  case of the SAS2008 in my test system), I/O stops.  If we tell CAM
  that we have one less command slot than we have actually allocated,
  everything works fine.  We also need a few extra command slots to
  allow for aborts and other task management commands to be sent down.
  
  This needs more investigation to determine the root cause, but for
  now this fixes things in my testing.
  
  mps.c:Change a printf() to mps_printf().
  
  mps_sas.c:Subtract 5 command slots when we tell CAM how many
commands we can handle.
  
Add some commented-out logic to print the contents
the CDBs for timed-out commands.  This can help
in debugging devices that are timing out.  This
will be uncommented once I bring some CAM changes in.
  
  Reported by:  Andrew Boyer 

Modified:
  head/sys/dev/mps/mps.c
  head/sys/dev/mps/mps_sas.c

Modified: head/sys/dev/mps/mps.c
==
--- head/sys/dev/mps/mps.c  Fri Sep 17 19:20:39 2010(r212801)
+++ head/sys/dev/mps/mps.c  Fri Sep 17 21:53:56 2010(r212802)
@@ -1416,7 +1416,7 @@ mps_data_cb(void *arg, bus_dma_segment_t
chain = mps_alloc_chain(sc);
if (chain == NULL) {
/* Resource shortage, roll back! */
-   printf("out of chain frames\n");
+   mps_printf(sc, "out of chain frames\n");
return;
}
 

Modified: head/sys/dev/mps/mps_sas.c
==
--- head/sys/dev/mps/mps_sas.c  Fri Sep 17 19:20:39 2010(r212801)
+++ head/sys/dev/mps/mps_sas.c  Fri Sep 17 21:53:56 2010(r212802)
@@ -596,6 +596,7 @@ mps_attach_sas(struct mps_softc *sc)
 {
struct mpssas_softc *sassc;
int error = 0;
+   int num_sim_reqs;
 
mps_dprint(sc, MPS_TRACE, "%s\n", __func__);
 
@@ -605,15 +606,30 @@ mps_attach_sas(struct mps_softc *sc)
sc->sassc = sassc;
sassc->sc = sc;
 
-   if ((sassc->devq = cam_simq_alloc(sc->num_reqs)) == NULL) {
+   /*
+* Tell CAM that we can handle 5 fewer requests than we have
+* allocated.  If we allow the full number of requests, all I/O
+* will halt when we run out of resources.  Things work fine with
+* just 1 less request slot given to CAM than we have allocated.
+* We also need a couple of extra commands so that we can send down
+* abort, reset, etc. requests when commands time out.  Otherwise
+* we could wind up in a situation with sc->num_reqs requests down
+* on the card and no way to send an abort.
+*
+* XXX KDM need to figure out why I/O locks up if all commands are
+* used.
+*/
+   num_sim_reqs = sc->num_reqs - 5;
+
+   if ((sassc->devq = cam_simq_alloc(num_sim_reqs)) == NULL) {
mps_dprint(sc, MPS_FAULT, "Cannot allocate SIMQ\n");
error = ENOMEM;
goto out;
}
 
sassc->sim = cam_sim_alloc(mpssas_action, mpssas_poll, "mps", sassc,
-   device_get_unit(sc->mps_dev), &sc->mps_mtx, sc->num_reqs, 
sc->num_reqs,
-   sassc->devq);
+   device_get_unit(sc->mps_dev), &sc->mps_mtx, num_sim_reqs,
+   num_sim_reqs, sassc->devq);
if (sassc->sim == NULL) {
mps_dprint(sc, MPS_FAULT, "Cannot allocate SIM\n");
error = EINVAL;
@@ -930,6 +946,9 @@ mpssas_scsiio_timeout(void *data)
struct mps_softc *sc;
struct mps_command *cm;
struct mpssas_target *targ;
+#if 0
+   char cdb_str[(SCSI_MAX_CDBLEN * 3) + 1];
+#endif
 
cm = (struct mps_command *)data;
sc = cm->cm_sc;
@@ -954,6 +973,22 @@ mpssas_scsiio_timeout(void *data)
 
xpt_print(ccb->ccb_h.path, "SCSI command timeout on device handle "
  "0x%04x SMID %d\n", targ->handle, cm->cm_desc.Default.SMID);
+   /*
+* XXX KDM this is useful for debugging purposes, but the existing
+* scsi_op_desc() implementation can't handle a NULL value for
+* inq_data.  So this will remain commented out until I bring in
+* those changes as well.
+*/
+#if 0
+   xpt_print(ccb->ccb_h.path, "Timed out command: %s. CDB %s\n",
+ scsi_op_desc((ccb->ccb_h.flags & CAM_CDB_POINTER) ?
+   ccb->csio.cdb_io.cdb_ptr[0] :
+   ccb->csio.cdb_io.cdb_bytes[0], NULL),
+ scsi_cdb_string((ccb->ccb_h.flags & CAM_CDB_POINTER) ?
+

svn commit: r213535 - in head: share/man/man4 sys/dev/mps

[Kenneth D. Merry]

Author: ken
Date: Thu Oct  7 21:56:10 2010
New Revision: 213535
URL: http://svn.freebsd.org/changeset/base/213535

Log:
  Turn on serialization of task management commands going down to the
  controller, but make it optional.
  
  After a problem report from Andrew Boyer, it looks like the LSI
  chip may have issues (the watchdog timer fired) if too many aborts
  are sent down to the chip at the same time.  We know that task
  management commands are serialized, and although the manual doesn't
  say it, it may be a good idea to just send one at a time.
  
  But, since I'm not certain that this is necessary, add a tunable
  and sysctl variable (hw.mps.%d.allow_multiple_tm_cmds) to control
  the driver's behavior.
  
  mps.c:Add support for the sysctl and tunable, and add a
comment about the possible return values to
mps_map_command().
  
  mps_sas.c:Run all task management commands through two new
routines, mpssas_issue_tm_request() and
mpssas_complete_tm_request().
  
This allows us to optionally serialize task
management commands.  Also, change things so that
the response to a task management command always
comes back through the callback.  (Before it could
come via the callback or the return value.)
  
  mpsvar.h: Add softc variables for the list of active task
management commands, the number of active commands,
and whether we should allow multiple active task
management commands.  Add an active command flag.
  
  mps.4:Describe the new sysctl/loader tunable variable.
  
  Sponsored by: Spectra Logic Corporation

Modified:
  head/share/man/man4/mps.4
  head/sys/dev/mps/mps.c
  head/sys/dev/mps/mps_sas.c
  head/sys/dev/mps/mpsvar.h

Modified: head/share/man/man4/mps.4
==
--- head/share/man/man4/mps.4   Thu Oct  7 20:31:07 2010(r213534)
+++ head/share/man/man4/mps.4   Thu Oct  7 21:56:10 2010(r213535)
@@ -31,7 +31,7 @@
 .\"
 .\" Author: Ken Merry 
 .\"
-.\" $Id: //depot/SpectraBSD/head/share/man/man4/mps.4#1 $
+.\" $Id: //depot/SpectraBSD/head/share/man/man4/mps.4#4 $
 .\" $FreeBSD$
 .\"
 .Dd September 13, 2010
@@ -97,6 +97,20 @@ driver instances, set the following tuna
 .Bd -literal -offset indent
 hw.mps.disable_msix=1
 .Ed
+.Pp
+To allow the driver to send multiple task management commands (like abort,
+LUN reset, etc.), set the following variable:
+.Bd -literal -offset indent
+hw.mps.X.allow_multiple_tm_cmds=1
+.Ed
+.Pp
+via
+.Xr loader.conf 5
+or
+.Xr sysctl 8 ,
+where X is the adapter number.
+By default the driver only sends one task management command at a time, to
+avoid causing a potential controller lock-up.
 .Sh DEBUGGING
 To enable debugging prints from the
 .Nm

Modified: head/sys/dev/mps/mps.c
==
--- head/sys/dev/mps/mps.c  Thu Oct  7 20:31:07 2010(r213534)
+++ head/sys/dev/mps/mps.c  Thu Oct  7 21:56:10 2010(r213535)
@@ -805,6 +805,9 @@ mps_attach(struct mps_softc *sc)
snprintf(tmpstr, sizeof(tmpstr), "hw.mps.%d.debug_level",
device_get_unit(sc->mps_dev));
TUNABLE_INT_FETCH(tmpstr, &sc->mps_debug);
+   snprintf(tmpstr, sizeof(tmpstr), "hw.mps.%d.allow_multiple_tm_cmds",
+   device_get_unit(sc->mps_dev));
+   TUNABLE_INT_FETCH(tmpstr, &sc->allow_multiple_tm_cmds);
 
mps_dprint(sc, MPS_TRACE, "%s\n", __func__);
 
@@ -831,6 +834,11 @@ mps_attach(struct mps_softc *sc)
OID_AUTO, "debug_level", CTLFLAG_RW, &sc->mps_debug, 0,
"mps debug level");
 
+   SYSCTL_ADD_INT(&sc->sysctl_ctx, SYSCTL_CHILDREN(sc->sysctl_tree),
+   OID_AUTO, "allow_multiple_tm_cmds", CTLFLAG_RW,
+   &sc->allow_multiple_tm_cmds, 0,
+   "allow multiple simultaneous task management cmds");
+
if ((error = mps_transition_ready(sc)) != 0)
return (error);
 
@@ -873,6 +881,7 @@ mps_attach(struct mps_softc *sc)
sc->facts->MaxReplyDescriptorPostQueueDepth) - 1;
TAILQ_INIT(&sc->req_list);
TAILQ_INIT(&sc->chain_list);
+   TAILQ_INIT(&sc->tm_list);
 
if (((error = mps_alloc_queues(sc)) != 0) ||
((error = mps_alloc_replies(sc)) != 0) ||
@@ -1470,6 +1479,10 @@ mps_data_cb(void *arg, bus_dma_segment_t
return;
 }
 
+/*
+ * Note that the only error path here is from bus_dmamap_load(), which can
+ * return EINPROGRESS if it is waiting for resources.
+ */
 int
 mps_map_command(struct mps_softc *sc, struct mps_command *cm)
 {

Modified: head/sys/dev/mps/mps_sas.c
==
--- head/sys/dev/mps/mps_sas.c  Thu Oct  7 20:31:07 2010(r213534)
+++ head/sys/dev/mps/mps_sas.c  T

svn commit: r216088 - in head: lib/libcam sbin/camcontrol sys/cam sys/cam/scsi sys/conf sys/dev/mps sys/sys

[Kenneth D. Merry]

Author: ken
Date: Tue Nov 30 22:39:46 2010
New Revision: 216088
URL: http://svn.freebsd.org/changeset/base/216088

Log:
  Add Serial Management Protocol (SMP) passthrough support to CAM.
  
  This includes support in the kernel, camcontrol(8), libcam and the mps(4)
  driver for SMP passthrough.
  
  The CAM SCSI probe code has been modified to fetch Inquiry VPD page 0x00
  to determine supported pages, and will now fetch page 0x83 in addition to
  page 0x80 if supported.
  
  Add two new CAM CCBs, XPT_SMP_IO, and XPT_GDEV_ADVINFO.  The SMP CCB is
  intended for SMP requests and responses.  The ADVINFO is currently used to
  fetch cached VPD page 0x83 data from the transport layer, but is intended
  to be extensible to fetch other types of device-specific data.
  
  SMP-only devices are not currently represented in the CAM topology, and so
  the current semantics are that the SIM will route SMP CCBs to either the
  addressed device, if it contains an SMP target, or its parent, if it
  contains an SMP target.  (This is noted in cam_ccb.h, since it will change
  later once we have the ability to have SMP-only devices in CAM's topology.)
  
  smp_all.c,
  smp_all.h:New helper routines for SMP.  This includes
SMP request building routines, response parsing
routines, error decoding routines, and structure
definitions for a number of SMP commands.
  
  libcam/Makefile:  Add smp_all.c to libcam, so that SMP functionality
is available to userland applications.
  
  camcontrol.8,
  camcontrol.c: Add smp passthrough support to camcontrol.  Several
new subcommands are now available:
  
'smpcmd' functions much like 'cmd', except that it
allows the user to send generic SMP commands.
  
'smprg' sends the SMP report general command, and
displays the decoded output.  It will automatically
fetch extended output if it is available.
  
'smppc' sends the SMP phy control command, with any
number of potential options.  Among other things,
this allows the user to reset a phy on a SAS
expander, or disable a phy on an expander.
  
'smpmaninfo' sends the SMP report manufacturer
information and displays the decoded output.
  
'smpphylist' displays a list of phys on an
expander, and the CAM devices attached to those
phys, if any.
  
  cam.h,
  cam.c:Add a status value for SMP errors
(CAM_SMP_STATUS_ERROR).
  
Add a missing description for CAM_SCSI_IT_NEXUS_LOST.
  
Add support for SMP commands to cam_error_string().
  
  cam_ccb.h:Rename the CAM_DIR_RESV flag to CAM_DIR_BOTH.  SMP
commands are by nature bi-directional, and we may
need to support bi-directional SCSI commands later.
  
Add the XPT_SMP_IO CCB.  Since SMP commands are
bi-directional, there are pointers for both the
request and response.
  
Add a fill routine for SMP CCBs.
  
Add the XPT_GDEV_ADVINFO CCB.  This is currently
used to fetch cached page 0x83 data from the
transport later, but is extensible to fetch many
other types of data.
  
  cam_periph.c: Add support in cam_periph_mapmem() for XPT_SMP_IO
and XPT_GDEV_ADVINFO CCBs.
  
  cam_xpt.c:Add support for executing XPT_SMP_IO CCBs.
  
  cam_xpt_internal.h:   Add fields for VPD pages 0x00 and 0x83 in struct
cam_ed.
  
  scsi_all.c:   Add scsi_get_sas_addr(), a function that parses
VPD page 0x83 data and pulls out a SAS address.
  
  scsi_all.h:   Add VPD page 0x00 and 0x83 structures, and a
prototype for scsi_get_sas_addr().
  
  scsi_pass.c:  Add support for mapping buffers in XPT_SMP_IO and
XPT_GDEV_ADVINFO CCBs.
  
  scsi_xpt.c:   In the SCSI probe code, first ask the device for
VPD page 0x00.  If any VPD pages are supported,
that page is required to be implemented.  Based on
the response, we may probe for the serial number
(page 0x80) or device id (page 0x83).
  
Add support for the XPT_GDEV_ADVINFO CCB.
  
  sys/conf/files:   Add smp_all.c.
  
  mps.c:Add support for passing in a

svn commit: r216093 - head/sys/modules/cam

[Kenneth D. Merry]

Author: ken
Date: Wed Dec  1 05:39:02 2010
New Revision: 216093
URL: http://svn.freebsd.org/changeset/base/216093

Log:
  Add smp_all.c to the cam module build to un-break it.
  
  Reported by:  jkim

Modified:
  head/sys/modules/cam/Makefile

Modified: head/sys/modules/cam/Makefile
==
--- head/sys/modules/cam/Makefile   Wed Dec  1 05:24:29 2010
(r216092)
+++ head/sys/modules/cam/Makefile   Wed Dec  1 05:39:02 2010
(r216093)
@@ -25,6 +25,7 @@ SRCS+=scsi_ses.c
 SRCS+= scsi_sg.c
 SRCS+= scsi_targ_bh.c scsi_target.c
 SRCS+= scsi_xpt.c
+SRCS+= smp_all.c
 SRCS+= ata_all.c
 SRCS+= ata_xpt.c
 SRCS+= ata_da.c
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r216361 - in head: sbin/camcontrol sys/cam sys/cam/scsi

[Kenneth D. Merry]

Author: ken
Date: Fri Dec 10 21:38:51 2010
New Revision: 216361
URL: http://svn.freebsd.org/changeset/base/216361

Log:
  Fix a few issues related to the XPT_GDEV_ADVINFO CCB.
  
  camcontrol.c: In buildbusdevlist(), don't attempt to get call
getdevid() for an unconfigured device, even when the
verbose flag is set.  The cam_open_btl() call will almost
certainly fail.
  
Probe for the buffer size when issuing the XPT_GDEV_ADVINFO
CCB.  Probing for the buffer size first helps us avoid
allocating the maximum buffer size when it really may not
be necessary.  This also helps avoid errors from
cam_periph_mapmem() if we attempt to map more than MAXPHYS.
  
  cam_periph.c: In cam_periph_mapmem(), if the XPT_GDEV_ADVINFO CCB
shows a bufsiz of 0, we don't have anything to map,
so just return.
  
Also, set the maximum mapping size to MAXPHYS
instead of DFLTPHYS for XPT_GDEV_ADVINFO CCBs,
since they don't actually go down to the hardware.
  
  scsi_pass.c:  Don't bother mapping the buffer in XPT_GDEV_ADVINFO
CCBs if bufsiz is 0.

Modified:
  head/sbin/camcontrol/camcontrol.c
  head/sys/cam/cam_periph.c
  head/sys/cam/scsi/scsi_pass.c

Modified: head/sbin/camcontrol/camcontrol.c
==
--- head/sbin/camcontrol/camcontrol.c   Fri Dec 10 20:27:50 2010
(r216360)
+++ head/sbin/camcontrol/camcontrol.c   Fri Dec 10 21:38:51 2010
(r216361)
@@ -5026,14 +5026,7 @@ getdevid(struct cam_devitem *item)
goto bailout;
}
 
-   item->device_id_len = CAM_SCSI_DEVID_MAXLEN;
-   item->device_id = malloc(item->device_id_len);
-   if (item->device_id == NULL) {
-   warn("%s: unable to allocate %d bytes", __func__,
-item->device_id_len);
-   retval = 1;
-   goto bailout;
-   }
+   item->device_id_len = 0;
 
ccb = cam_getccb(dev);
if (ccb == NULL) {
@@ -5044,12 +5037,19 @@ getdevid(struct cam_devitem *item)
 
bzero(&(&ccb->ccb_h)[1],
  sizeof(union ccb) - sizeof(struct ccb_hdr));
+
+   /*
+* On the first try, we just probe for the size of the data, and
+* then allocate that much memory and try again.
+*/
+retry:
ccb->ccb_h.func_code = XPT_GDEV_ADVINFO;
ccb->ccb_h.flags = CAM_DIR_IN;
ccb->cgdai.flags = CGDAI_FLAG_PROTO;
ccb->cgdai.buftype = CGDAI_TYPE_SCSI_DEVID;
ccb->cgdai.bufsiz = item->device_id_len;
-   ccb->cgdai.buf = (uint8_t *)item->device_id;
+   if (item->device_id_len != 0)
+   ccb->cgdai.buf = (uint8_t *)item->device_id;
 
if (cam_send_ccb(dev, ccb) < 0) {
warn("%s: error sending XPT_GDEV_ADVINFO CCB", __func__);
@@ -5063,6 +5063,29 @@ getdevid(struct cam_devitem *item)
goto bailout;
}
 
+   if (item->device_id_len == 0) {
+   /*
+* This is our first time through.  Allocate the buffer,
+* and then go back to get the data.
+*/
+   if (ccb->cgdai.provsiz == 0) {
+   warnx("%s: invalid .provsiz field returned with "
+"XPT_GDEV_ADVINFO CCB", __func__);
+   retval = 1;
+   goto bailout;
+   }
+   item->device_id_len = ccb->cgdai.provsiz;
+   item->device_id = malloc(item->device_id_len);
+   if (item->device_id == NULL) {
+   warn("%s: unable to allocate %d bytes", __func__,
+item->device_id_len);
+   retval = 1;
+   goto bailout;
+   }
+   ccb->ccb_h.status = CAM_REQ_INPROG;
+   goto retry;
+   }
+
 bailout:
if (dev != NULL)
cam_close_device(dev);
@@ -5158,9 +5181,8 @@ buildbusdevlist(struct cam_devlist *devl
dev_result = 
 &ccb.cdm.matches[i].result.device_result;
 
-   if ((dev_result->flags
-& DEV_RESULT_UNCONFIGURED)
-&& ((arglist & CAM_ARG_VERBOSE) == 0)) {
+   if (dev_result->flags &
+   DEV_RESULT_UNCONFIGURED) {
skip_device = 1;
break;
} else

Modified: head/sys/cam/cam_periph.c
==
--- head/sys/cam/cam_periph.c   Fri Dec 10 20:27:50 2010(r216360)
+++ head/sys/cam/cam_periph.c   Fri Dec 

svn commit: r216363 - head/sys/dev/mps

[Kenneth D. Merry]

Author: ken
Date: Fri Dec 10 21:45:10 2010
New Revision: 216363
URL: http://svn.freebsd.org/changeset/base/216363

Log:
  Fix an event handling bug with the mps(4) driver.
  
  This bug manifested itself after repeated device arrivals and
  departures.  The root of the problem was that the last entry in the
  reply array wasn't initialized/allocated.  So every time we got
  around to that event, we had a bogus address.
  
  There were a couple more problems with the code that are also fixed:
  
   - The reply mechanism was being treated as sequential (indexed by
 sc->replycurindex) even though the spec says that the driver
 should use the ReplyFrameAddress field of the post queue
 descriptor to figure out where the reply is.  There is no
 guarantee that the reply descriptors will be used in sequential
 order.
  
   - The second word of the reply post queue descriptor wasn't being
 checked in mps_intr_locked() to make sure that it wasn't
 0x.  So the driver could potentially come across a
 partially DMAed descriptor.
  
   - The number of replies allocated was one less than the actual
 size of the queue.  Instead, it was the size of the number of
 replies that can be used at one time.  (Which is one less than
 the size of the queue.)
  
  mps.c:When initializing the entries in the reply free
queue, make sure we initialize the full number that
we tell the chip we have (sc->fqdepth), not the
number that can be used at any one time (sc->num_replies).
  
When allocating replies, make sure we allocate the
number of replies that we've told the chip exist,
not just the number that can be used simultaneously.
  
Use the ReplyFrameAddress field of the post queue
descriptor to figure out which reply is being
referenced.  This is what the spec says to do, and
the spec doesn't guarantee that the replies will be
used in order.
  
Put a check in to verify that the reply address passed
back from the card is valid.  (Panic if it isn't, we'll
panic when we try to deference the reply pointer in any
case.)
  
In mps_intr_locked(), verify that the second word of the
post queue descriptor is not 0x in addition to
verifying that the unused flag is not set, so we can
make sure we didn't get a partially DMAed descriptor.
  
Remove references to sc->replycurindex, it isn't needed
now.
  
  mpsvar.h: Remove replycurindex from the softc, it isn't needed now.
  
  Reviewed by:  scottl

Modified:
  head/sys/dev/mps/mps.c
  head/sys/dev/mps/mpsvar.h

Modified: head/sys/dev/mps/mps.c
==
--- head/sys/dev/mps/mps.c  Fri Dec 10 21:43:20 2010(r216362)
+++ head/sys/dev/mps/mps.c  Fri Dec 10 21:45:10 2010(r216363)
@@ -43,6 +43,7 @@ __FBSDID("$FreeBSD$");
 #include 
 #include 
 #include 
+#include 
 
 #include 
 #include 
@@ -607,9 +608,16 @@ mps_alloc_queues(struct mps_softc *sc)
 static int
 mps_alloc_replies(struct mps_softc *sc)
 {
-   int rsize;
+   int rsize, num_replies;
 
-   rsize = sc->facts->ReplyFrameSize * sc->num_replies * 4; 
+   /*
+* sc->num_replies should be one less than sc->fqdepth.  We need to
+* allocate space for sc->fqdepth replies, but only sc->num_replies
+* replies can be used at once.
+*/
+   num_replies = max(sc->fqdepth, sc->num_replies);
+
+   rsize = sc->facts->ReplyFrameSize * num_replies * 4; 
 if (bus_dma_tag_create( sc->mps_parent_dmat,/* parent */
4, 0,   /* algnmnt, boundary */
BUS_SPACE_MAXADDR_32BIT,/* lowaddr */
@@ -782,11 +790,19 @@ mps_init_queues(struct mps_softc *sc)
 
memset((uint8_t *)sc->post_queue, 0xff, sc->pqdepth * 8);
 
+   /*
+* According to the spec, we need to use one less reply than we
+* have space for on the queue.  So sc->num_replies (the number we
+* use) should be less than sc->fqdepth (allocated size).
+*/
if (sc->num_replies >= sc->fqdepth)
return (EINVAL);
 
-   for (i = 0; i < sc->num_replies; i++)
-   sc->free_queue[i] = sc->reply_busaddr + i * 
sc->facts->ReplyFrameSize * 4;
+   /*
+* Initialize all of the free queue entries.
+*/
+   for (i = 0; i < sc->fqdepth; i++)
+   sc->free_queue[i] = sc->reply_busaddr + (i * 
sc->facts->ReplyFrameSize * 4);
sc->replyfreeindex = sc->num_replies;
 
return (0);
@@ -907,7 +923,6 @@ mps_attach(struct mps_softc *sc)
 * replies.
 */
sc-

svn commit: r216368 - head/sys/dev/mps

[Kenneth D. Merry]

Author: ken
Date: Sat Dec 11 00:36:35 2010
New Revision: 216368
URL: http://svn.freebsd.org/changeset/base/216368

Log:
  Fix setting LUN numbers in the mps(4) driver.
  
  Prior to this change, the addressing method wasn't getting set, and
  so the LUN field could be set incorrectly in some instances.
  
  This fix should allow for LUN numbers up to 16777215 (and return an error
  for anything larger, which wouldn't fit into the flat addressing model).
  
  Submitted by: scottl (in part)

Modified:
  head/sys/dev/mps/mps_sas.c

Modified: head/sys/dev/mps/mps_sas.c
==
--- head/sys/dev/mps/mps_sas.c  Fri Dec 10 23:57:55 2010(r216367)
+++ head/sys/dev/mps/mps_sas.c  Sat Dec 11 00:36:35 2010(r216368)
@@ -121,6 +121,7 @@ struct mpssas_devprobe {
 
 MALLOC_DEFINE(M_MPSSAS, "MPSSAS", "MPS SAS memory");
 
+static __inline int mpssas_set_lun(uint8_t *lun, u_int ccblun);
 static struct mpssas_target * mpssas_alloc_target(struct mpssas_softc *,
 struct mpssas_target *);
 static struct mpssas_target * mpssas_find_target(struct mpssas_softc *, int,
@@ -163,6 +164,43 @@ static void mpssas_resetdev_complete(str
 static void mpssas_freeze_device(struct mpssas_softc *, struct mpssas_target 
*);
 static void mpssas_unfreeze_device(struct mpssas_softc *, struct mpssas_target 
*) __unused;
 
+/*
+ * Abstracted so that the driver can be backwards and forwards compatible
+ * with future versions of CAM that will provide this functionality.
+ */
+#define MPS_SET_LUN(lun, ccblun)   \
+   mpssas_set_lun(lun, ccblun)
+
+static __inline int
+mpssas_set_lun(uint8_t *lun, u_int ccblun)
+{
+   uint64_t *newlun;
+
+   newlun = (uint64_t *)lun;
+   *newlun = 0;
+   if (ccblun <= 0xff) {
+   /* Peripheral device address method, LUN is 0 to 255 */
+   lun[1] = ccblun;
+   } else if (ccblun <= 0x3fff) {
+   /* Flat space address method, LUN is <= 16383 */
+   scsi_ulto2b(ccblun, lun);
+   lun[0] |= 0x40;
+   } else if (ccblun <= 0xff) {
+   /* Extended flat space address method, LUN is <= 16777215 */
+   scsi_ulto3b(ccblun, &lun[1]);
+   /* Extended Flat space address method */
+   lun[0] = 0xc0;
+   /* Length = 1, i.e. LUN is 3 bytes long */
+   lun[0] |= 0x10;
+   /* Extended Address Method */
+   lun[0] |= 0x02;
+   } else {
+   return (EINVAL);
+   }
+
+   return (0);
+}
+
 static struct mpssas_target *
 mpssas_alloc_target(struct mpssas_softc *sassc, struct mpssas_target *probe)
 {
@@ -1366,14 +1404,12 @@ mpssas_action_scsiio(struct mpssas_softc
break;
}
 
-   /* XXX Need to handle multi-level LUNs */
-   if (csio->ccb_h.target_lun > 255) {
+   if (MPS_SET_LUN(req->LUN, csio->ccb_h.target_lun) != 0) {
mps_free_command(sc, cm);
ccb->ccb_h.status = CAM_LUN_INVALID;
xpt_done(ccb);
return;
}
-   req->LUN[1] = csio->ccb_h.target_lun;
 
if (csio->ccb_h.flags & CAM_CDB_POINTER)
bcopy(csio->cdb_io.cdb_ptr, &req->CDB.CDB32[0], csio->cdb_len);
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r197208 - head/sys/cam

[Kenneth D. Merry]

Author: ken
Date: Tue Sep 15 00:15:24 2009
New Revision: 197208
URL: http://svn.freebsd.org/changeset/base/197208

Log:
  Fix some instances where CAM rescans get hung up or take a long time to
  complete.
  
  Also, allow xpt_rescan() to rescan a LUN instead of a full bus.
  
  MFC after:3 days
  Sponsored by: Copan Systems, Inc.

Modified:
  head/sys/cam/cam_xpt.c

Modified: head/sys/cam/cam_xpt.c
==
--- head/sys/cam/cam_xpt.c  Mon Sep 14 23:46:40 2009(r197207)
+++ head/sys/cam/cam_xpt.c  Tue Sep 15 00:15:24 2009(r197208)
@@ -794,8 +794,9 @@ xpt_scanner_thread(void *dummy)
 * processed.
 */
xpt_lock_buses();
-   msleep(&xsoftc.ccb_scanq, &xsoftc.xpt_topo_lock, PRIBIO,
-   "ccb_scanq", 0);
+   if (TAILQ_EMPTY(&xsoftc.ccb_scanq))
+   msleep(&xsoftc.ccb_scanq, &xsoftc.xpt_topo_lock, PRIBIO,
+  "ccb_scanq", 0);
TAILQ_INIT(&queue);
TAILQ_CONCAT(&queue, &xsoftc.ccb_scanq, sim_links.tqe);
xpt_unlock_buses();
@@ -806,9 +807,12 @@ xpt_scanner_thread(void *dummy)
sim = ccb->ccb_h.path->bus->sim;
CAM_SIM_LOCK(sim);
 
-   ccb->ccb_h.func_code = XPT_SCAN_BUS;
+   if( ccb->ccb_h.path->target->target_id == 
CAM_TARGET_WILDCARD )
+   ccb->ccb_h.func_code = XPT_SCAN_BUS;
+   else
+   ccb->ccb_h.func_code = XPT_SCAN_LUN;
ccb->ccb_h.cbfcnp = xptdone;
-   xpt_setup_ccb(&ccb->ccb_h, ccb->ccb_h.path, 5);
+   xpt_setup_ccb(&ccb->ccb_h, ccb->ccb_h.path, 1);
cam_periph_runccb(ccb, NULL, 0, 0, NULL);
xpt_free_path(ccb->ccb_h.path);
xpt_free_ccb(ccb);
@@ -828,6 +832,7 @@ xpt_rescan(union ccb *ccb)
xpt_lock_buses();
TAILQ_FOREACH(hdr, &xsoftc.ccb_scanq, sim_links.tqe) {
if (xpt_path_comp(hdr->path, ccb->ccb_h.path) == 0) {
+   wakeup(&xsoftc.ccb_scanq);
xpt_unlock_buses();
xpt_print(ccb->ccb_h.path, "rescan already queued\n");
xpt_free_path(ccb->ccb_h.path);
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r197318 - in stable/8/sys: . amd64/include/xen cam cddl/contrib/opensolaris contrib/dev/acpica contrib/pf dev/ciss dev/xen/xenpci

[Kenneth D. Merry]

Author: ken
Date: Fri Sep 18 20:35:05 2009
New Revision: 197318
URL: http://svn.freebsd.org/changeset/base/197318

Log:
  Merge change r197208 from head to stable/8:
  
  Fix some instances where CAM rescans get hung up or take a long time to
  complete.
  
  Also, allow xpt_rescan() to rescan a LUN instead of a full bus.
  
  Sponsored by: Copan Systems, Inc.
  Approved by:  re (kib)

Modified:
  stable/8/sys/   (props changed)
  stable/8/sys/amd64/include/xen/   (props changed)
  stable/8/sys/cam/cam_xpt.c
  stable/8/sys/cddl/contrib/opensolaris/   (props changed)
  stable/8/sys/contrib/dev/acpica/   (props changed)
  stable/8/sys/contrib/pf/   (props changed)
  stable/8/sys/dev/ciss/   (props changed)
  stable/8/sys/dev/xen/xenpci/   (props changed)

Modified: stable/8/sys/cam/cam_xpt.c
==
--- stable/8/sys/cam/cam_xpt.c  Fri Sep 18 17:09:33 2009(r197317)
+++ stable/8/sys/cam/cam_xpt.c  Fri Sep 18 20:35:05 2009(r197318)
@@ -794,8 +794,9 @@ xpt_scanner_thread(void *dummy)
 * processed.
 */
xpt_lock_buses();
-   msleep(&xsoftc.ccb_scanq, &xsoftc.xpt_topo_lock, PRIBIO,
-   "ccb_scanq", 0);
+   if (TAILQ_EMPTY(&xsoftc.ccb_scanq))
+   msleep(&xsoftc.ccb_scanq, &xsoftc.xpt_topo_lock, PRIBIO,
+  "ccb_scanq", 0);
TAILQ_INIT(&queue);
TAILQ_CONCAT(&queue, &xsoftc.ccb_scanq, sim_links.tqe);
xpt_unlock_buses();
@@ -806,9 +807,12 @@ xpt_scanner_thread(void *dummy)
sim = ccb->ccb_h.path->bus->sim;
CAM_SIM_LOCK(sim);
 
-   ccb->ccb_h.func_code = XPT_SCAN_BUS;
+   if( ccb->ccb_h.path->target->target_id == 
CAM_TARGET_WILDCARD )
+   ccb->ccb_h.func_code = XPT_SCAN_BUS;
+   else
+   ccb->ccb_h.func_code = XPT_SCAN_LUN;
ccb->ccb_h.cbfcnp = xptdone;
-   xpt_setup_ccb(&ccb->ccb_h, ccb->ccb_h.path, 5);
+   xpt_setup_ccb(&ccb->ccb_h, ccb->ccb_h.path, 1);
cam_periph_runccb(ccb, NULL, 0, 0, NULL);
xpt_free_path(ccb->ccb_h.path);
xpt_free_ccb(ccb);
@@ -828,6 +832,7 @@ xpt_rescan(union ccb *ccb)
xpt_lock_buses();
TAILQ_FOREACH(hdr, &xsoftc.ccb_scanq, sim_links.tqe) {
if (xpt_path_comp(hdr->path, ccb->ccb_h.path) == 0) {
+   wakeup(&xsoftc.ccb_scanq);
xpt_unlock_buses();
xpt_print(ccb->ccb_h.path, "rescan already queued\n");
xpt_free_path(ccb->ccb_h.path);
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r224973 - head/sys/dev/mps

[Kenneth D. Merry]

Author: ken
Date: Thu Aug 18 16:07:41 2011
New Revision: 224973
URL: http://svn.freebsd.org/changeset/base/224973

Log:
  Set the max_lun field in the path inquiry CCB for the mps(4) driver to 8.
  
  This allows LUNs greater than 0 to be probed.  The value can be increased
  later if need be.
  
  Approved by:  re (kib)

Modified:
  head/sys/dev/mps/mps_sas.c

Modified: head/sys/dev/mps/mps_sas.c
==
--- head/sys/dev/mps/mps_sas.c  Thu Aug 18 16:00:32 2011(r224972)
+++ head/sys/dev/mps/mps_sas.c  Thu Aug 18 16:07:41 2011(r224973)
@@ -925,7 +925,7 @@ mpssas_action(struct cam_sim *sim, union
cpi->hba_misc = PIM_NOBUSRESET;
cpi->hba_eng_cnt = 0;
cpi->max_target = sassc->sc->facts->MaxTargets - 1;
-   cpi->max_lun = 0;
+   cpi->max_lun = 8;
cpi->initiator_id = 255;
strncpy(cpi->sim_vid, "FreeBSD", SIM_IDLEN);
strncpy(cpi->hba_vid, "LSILogic", HBA_IDLEN);
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r225950 - in head: sbin/camcontrol share/examples/scsi_target share/misc sys/cam sys/cam/scsi sys/dev/ciss sys/dev/firewire sys/dev/iir sys/dev/iscsi/initiator sys/dev/isp sys/dev/mly s...

[Kenneth D. Merry]

Author: ken
Date: Mon Oct  3 20:32:55 2011
New Revision: 225950
URL: http://svn.freebsd.org/changeset/base/225950

Log:
  Add descriptor sense support to CAM, and honor sense residuals properly in
  CAM.
  
  Desriptor sense is a new sense data format that originated in SPC-3.  Among
  other things, it allows for an 8-byte info field, which is necessary to
  pass back block numbers larger than 4 bytes.
  
  This change adds a number of new functions to scsi_all.c (and therefore
  libcam) that abstract out most access to sense data.
  
  This includes a bump of CAM_VERSION, because the CCB ABI has changed.
  Userland programs that use the CAM pass(4) driver will need to be
  recompiled.
  
  camcontrol.c: Change uses of scsi_extract_sense() to use
scsi_extract_sense_len().
  
Use scsi_get_sks() instead of accessing sense key specific
data directly.
  
  scsi_modes:   Update the control mode page to the latest version (SPC-4).
  
  scsi_cmds.c,
  scsi_target.c:Change references to struct scsi_sense_data to struct
scsi_sense_data_fixed.  This should be changed to allow the
user to specify fixed or descriptor sense, and then use
scsi_set_sense_data() to build the sense data.
  
  ps3cdrom.c:   Use scsi_set_sense_data() instead of setting sense data
manually.
  
  cam_periph.c: Use scsi_extract_sense_len() instead of using
scsi_extract_sense() or accessing sense data directly.
  
  cam_ccb.h:Bump the CAM_VERSION from 0x15 to 0x16.  The change of
struct scsi_sense_data from 32 to 252 bytes changes the
size of struct ccb_scsiio, but not the size of union ccb.
So the version must be bumped to prevent structure
mis-matches.
  
  scsi_all.h:   Lots of updated SCSI sense data and other structures.
  
Add function prototypes for the new sense data functions.
  
Take out the inline implementation of scsi_extract_sense().
It is now too large to put in a header file.
  
Add macros to calculate whether fields are present and
filled in fixed and descriptor sense data
  
  scsi_all.c:   In scsi_op_desc(), allow the user to pass in NULL inquiry
data, and we'll assume a direct access device in that case.
  
Changed the SCSI RESERVED sense key name and description
to COMPLETED, as it is now defined in the spec.
  
Change the error recovery action for a number of read errors
to prevent lots of retries when the drive has said that the
block isn't accessible.  This speeds up reconstruction of
the block by any RAID software running on top of the drive
(e.g. ZFS).
  
In scsi_sense_desc(), allow for invalid sense key numbers.
This allows calling this routine without checking the input
values first.
  
Change scsi_error_action() to use scsi_extract_sense_len(),
and handle things when invalid asc/ascq values are
encountered.
  
Add a new routine, scsi_desc_iterate(), that will call the
supplied function for every descriptor in descriptor format
sense data.
  
Add scsi_set_sense_data(), and scsi_set_sense_data_va(),
which build descriptor and fixed format sense data.  They
currently default to fixed format sense data.
  
Add a number of scsi_get_*() functions, which get different
types of sense data fields from either fixed or descriptor
format sense data, if the data is present.
  
Add a number of scsi_*_sbuf() functions, which print
formatted versions of various sense data fields.  These
functions work for either fixed or descriptor sense.
  
Add a number of scsi_sense_*_sbuf() functions, which have a
standard calling interface and print the indicated field.
These functions take descriptors only.
  
Add scsi_sense_desc_sbuf(), which will print a formatted
version of the given sense descriptor.
  
Pull out a majority of the scsi_sense_sbuf() function and
put it into scsi_sense_only_sbuf().  This allows callers
that don't use struct ccb_scsiio to easily utilize the
printing routines.  Revamp that function to handle
descriptor sense and use the new sense fetching and
printing routines.
  
Move scsi_extract_sense() into scsi_all.c, and implement it
in terms of the new function, scsi_extract_sense_len().
The _len() version takes a l

svn commit: r226067 - in stable/9: sbin/camcontrol share/examples/scsi_target share/misc sys/cam sys/cam/scsi sys/dev/ciss sys/dev/firewire sys/dev/iir sys/dev/iscsi/initiator sys/dev/isp sys/dev/m...

[Kenneth D. Merry]

Author: ken
Date: Thu Oct  6 19:15:51 2011
New Revision: 226067
URL: http://svn.freebsd.org/changeset/base/226067

Log:
  MFC r225950:
  
  Add descriptor sense support to CAM, and honor sense residuals properly in
  CAM.
  
  Desriptor sense is a new sense data format that originated in SPC-3.  Among
  other things, it allows for an 8-byte info field, which is necessary to
  pass back block numbers larger than 4 bytes.
  
  This change adds a number of new functions to scsi_all.c (and therefore
  libcam) that abstract out most access to sense data.
  
  This includes a bump of CAM_VERSION, because the CCB ABI has changed.
  Userland programs that use the CAM pass(4) driver will need to be
  recompiled.
  
  camcontrol.c: Change uses of scsi_extract_sense() to use
scsi_extract_sense_len().
  
Use scsi_get_sks() instead of accessing sense key specific
data directly.
  
  scsi_modes:   Update the control mode page to the latest version (SPC-4).
  
  scsi_cmds.c,
  scsi_target.c:Change references to struct scsi_sense_data to struct
scsi_sense_data_fixed.  This should be changed to allow the
user to specify fixed or descriptor sense, and then use
scsi_set_sense_data() to build the sense data.
  
  ps3cdrom.c:   Use scsi_set_sense_data() instead of setting sense data
manually.
  
  cam_periph.c: Use scsi_extract_sense_len() instead of using
scsi_extract_sense() or accessing sense data directly.
  
  cam_ccb.h:Bump the CAM_VERSION from 0x15 to 0x16.  The change of
struct scsi_sense_data from 32 to 252 bytes changes the
size of struct ccb_scsiio, but not the size of union ccb.
So the version must be bumped to prevent structure
mis-matches.
  
  scsi_all.h:   Lots of updated SCSI sense data and other structures.
  
Add function prototypes for the new sense data functions.
  
Take out the inline implementation of scsi_extract_sense().
It is now too large to put in a header file.
  
Add macros to calculate whether fields are present and
filled in fixed and descriptor sense data
  
  scsi_all.c:   In scsi_op_desc(), allow the user to pass in NULL inquiry
data, and we'll assume a direct access device in that case.
  
Changed the SCSI RESERVED sense key name and description
to COMPLETED, as it is now defined in the spec.
  
Change the error recovery action for a number of read errors
to prevent lots of retries when the drive has said that the
block isn't accessible.  This speeds up reconstruction of
the block by any RAID software running on top of the drive
(e.g. ZFS).
  
In scsi_sense_desc(), allow for invalid sense key numbers.
This allows calling this routine without checking the input
values first.
  
Change scsi_error_action() to use scsi_extract_sense_len(),
and handle things when invalid asc/ascq values are
encountered.
  
Add a new routine, scsi_desc_iterate(), that will call the
supplied function for every descriptor in descriptor format
sense data.
  
Add scsi_set_sense_data(), and scsi_set_sense_data_va(),
which build descriptor and fixed format sense data.  They
currently default to fixed format sense data.
  
Add a number of scsi_get_*() functions, which get different
types of sense data fields from either fixed or descriptor
format sense data, if the data is present.
  
Add a number of scsi_*_sbuf() functions, which print
formatted versions of various sense data fields.  These
functions work for either fixed or descriptor sense.
  
Add a number of scsi_sense_*_sbuf() functions, which have a
standard calling interface and print the indicated field.
These functions take descriptors only.
  
Add scsi_sense_desc_sbuf(), which will print a formatted
version of the given sense descriptor.
  
Pull out a majority of the scsi_sense_sbuf() function and
put it into scsi_sense_only_sbuf().  This allows callers
that don't use struct ccb_scsiio to easily utilize the
printing routines.  Revamp that function to handle
descriptor sense and use the new sense fetching and
printing routines.
  
Move scsi_extract_sense() into scsi_all.c, and implement it
in terms of the new function, scsi_extract_sense_len().
The _len()

svn commit: r217016 - head/sbin/camcontrol

[Kenneth D. Merry]

Author: ken
Date: Wed Jan  5 18:27:38 2011
New Revision: 217016
URL: http://svn.freebsd.org/changeset/base/217016

Log:
  The camcontrol smppc option to clear a PHY error log is 'clearerrlog',
  not 'clearerrorlog'.

Modified:
  head/sbin/camcontrol/camcontrol.8

Modified: head/sbin/camcontrol/camcontrol.8
==
--- head/sbin/camcontrol/camcontrol.8   Wed Jan  5 17:10:10 2011
(r217015)
+++ head/sbin/camcontrol/camcontrol.8   Wed Jan  5 18:27:38 2011
(r217016)
@@ -677,7 +677,7 @@ Send the HARD RESET command to the phy.
 .It disable
 Send the DISABLE command to the phy.
 Note that the LINK RESET or HARD RESET commands should re-enable the phy.
-.It clearerrorlog
+.It clearerrlog
 Send the CLEAR ERROR LOG command.
 This clears the error log counters for the specified phy.
 .It clearaffiliation
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r218810 - in stable/8/sys: amd64/conf conf dev/bwn dev/mps dev/siba dev/sis mips/mips modules modules/mps

[Kenneth D. Merry]

Author: ken
Date: Fri Feb 18 16:29:38 2011
New Revision: 218810
URL: http://svn.freebsd.org/changeset/base/218810

Log:
  MFC: 212420, 212616, 212772, 212802, 213535, 213702, 213704, 213707, 213708,
   213743, 213839, 213840, 213882, 213898, 216088, 216227, 216363, 216368:
  
  Merge the mps(4) driver into stable/8.  This is currently only included in
  GENERIC on amd64, since that is the only architecture it has been tested
  on.

Added:
  stable/8/sys/dev/mps/
 - copied from r212420, head/sys/dev/mps/
  stable/8/sys/modules/mps/
 - copied from r212420, head/sys/modules/mps/
Modified:
  stable/8/sys/amd64/conf/GENERIC
  stable/8/sys/conf/files
  stable/8/sys/dev/bwn/if_bwn.c
  stable/8/sys/dev/mps/mps.c
  stable/8/sys/dev/mps/mps_ioctl.h
  stable/8/sys/dev/mps/mps_pci.c
  stable/8/sys/dev/mps/mps_sas.c
  stable/8/sys/dev/mps/mps_user.c
  stable/8/sys/dev/mps/mpsvar.h
  stable/8/sys/dev/siba/siba_bwn.c
  stable/8/sys/dev/sis/if_sisreg.h
  stable/8/sys/mips/mips/mp_machdep.c
  stable/8/sys/modules/Makefile
  stable/8/sys/modules/mps/Makefile
Directory Properties:
  stable/8/sys/   (props changed)
  stable/8/sys/amd64/include/xen/   (props changed)
  stable/8/sys/cddl/contrib/opensolaris/   (props changed)
  stable/8/sys/contrib/dev/acpica/   (props changed)
  stable/8/sys/contrib/pf/   (props changed)

Modified: stable/8/sys/amd64/conf/GENERIC
==
--- stable/8/sys/amd64/conf/GENERIC Fri Feb 18 16:21:09 2011
(r218809)
+++ stable/8/sys/amd64/conf/GENERIC Fri Feb 18 16:29:38 2011
(r218810)
@@ -114,6 +114,7 @@ device  hptiop  # Highpoint RocketRaid 3
 device isp # Qlogic family
 #deviceispfw   # Firmware for QLogic HBAs- normally a 
module
 device mpt # LSI-Logic MPT-Fusion
+device mps # LSI-Logic MPT-Fusion 2
 #devicencr # NCR/Symbios Logic
 device sym # NCR/Symbios Logic (newer chipsets + those of 
`ncr')
 device trm # Tekram DC395U/UW/F DC315U adapters

Modified: stable/8/sys/conf/files
==
--- stable/8/sys/conf/files Fri Feb 18 16:21:09 2011(r218809)
+++ stable/8/sys/conf/files Fri Feb 18 16:29:38 2011(r218810)
@@ -1303,6 +1303,11 @@ dev/mmc/mmcbr_if.m   standard
 dev/mmc/mmcbus_if.mstandard
 dev/mmc/mmcsd.coptional mmcsd
 dev/mn/if_mn.c optional mn pci
+dev/mps/mps.c  optional mps
+dev/mps/mps_pci.c  optional mps pci
+dev/mps/mps_sas.c  optional mps
+dev/mps/mps_table.coptional mps
+dev/mps/mps_user.c optional mps
 dev/mpt/mpt.c  optional mpt
 dev/mpt/mpt_cam.c  optional mpt
 dev/mpt/mpt_debug.coptional mpt

Modified: stable/8/sys/dev/bwn/if_bwn.c
==
--- stable/8/sys/dev/bwn/if_bwn.c   Fri Feb 18 16:21:09 2011
(r218809)
+++ stable/8/sys/dev/bwn/if_bwn.c   Fri Feb 18 16:29:38 2011
(r218810)
@@ -2882,7 +2882,7 @@ bwn_set_channel(struct ieee80211com *ic)
 
error = bwn_switch_band(sc, ic->ic_curchan);
if (error)
-   goto fail;;
+   goto fail;
bwn_mac_suspend(mac);
bwn_set_txretry(mac, BWN_RETRY_SHORT, BWN_RETRY_LONG);
chan = ieee80211_chan2ieee(ic, ic->ic_curchan);
@@ -8260,7 +8260,7 @@ bwn_switch_band(struct bwn_softc *sc, st
device_printf(sc->sc_dev, "switching to %s-GHz band\n",
IEEE80211_IS_CHAN_2GHZ(chan) ? "2" : "5");
 
-   down_dev = sc->sc_curmac;;
+   down_dev = sc->sc_curmac;
status = down_dev->mac_status;
if (status >= BWN_MAC_STATUS_STARTED)
bwn_core_stop(down_dev);

Modified: stable/8/sys/dev/mps/mps.c
==
--- head/sys/dev/mps/mps.c  Fri Sep 10 15:03:56 2010(r212420)
+++ stable/8/sys/dev/mps/mps.c  Fri Feb 18 16:29:38 2011(r218810)
@@ -43,6 +43,7 @@ __FBSDID("$FreeBSD$");
 #include 
 #include 
 #include 
+#include 
 
 #include 
 #include 
@@ -380,7 +381,7 @@ mps_request_sync(struct mps_softc *sc, v
return (0);
 }
 
-static void
+void
 mps_enqueue_request(struct mps_softc *sc, struct mps_command *cm)
 {
 
@@ -607,9 +608,16 @@ mps_alloc_queues(struct mps_softc *sc)
 static int
 mps_alloc_replies(struct mps_softc *sc)
 {
-   int rsize;
+   int rsize, num_replies;
+
+   /*
+* sc->num_replies should be one less than sc->fqdepth.  We need to
+* allocate space for sc->fqdepth replies, but only sc->num_replies
+* replies can be used at once.
+*/
+   num_replies = max(sc->fqdepth, sc->num_replies);

svn commit: r218811 - head/sys/dev/mps

[Kenneth D. Merry]

Author: ken
Date: Fri Feb 18 17:01:57 2011
New Revision: 218811
URL: http://svn.freebsd.org/changeset/base/218811

Log:
  In the MPS driver, during device removal processing, don't assume that
  the controller firmware will return all of our commands.  Instead, keep
  track of outstanding I/Os and return them to CAM once device removal
  processing completes.
  
  mpsvar.h: Declare the new "io_list" in the mps_softc.
  
  mps.c:Initialize the new "io_list" in the mps softc.
  
  mps_sas.c:o Track SCSI I/O requests on the io_list from the
  time of mpssas_action() through mpssas_scsiio_complete().
o Zero out the request structures used for device
  removal commands prior to filling them out.
o Once the target reset task management function completes
  during device removal processing, assume any SCSI I/O
  commands that are still oustanding will never return
  from the controller, and process them manually.
  
  Submitted by: gibbs
  MFC after:3 days

Modified:
  head/sys/dev/mps/mps.c
  head/sys/dev/mps/mps_sas.c
  head/sys/dev/mps/mpsvar.h

Modified: head/sys/dev/mps/mps.c
==
--- head/sys/dev/mps/mps.c  Fri Feb 18 16:29:38 2011(r218810)
+++ head/sys/dev/mps/mps.c  Fri Feb 18 17:01:57 2011(r218811)
@@ -898,6 +898,7 @@ mps_attach(struct mps_softc *sc)
TAILQ_INIT(&sc->req_list);
TAILQ_INIT(&sc->chain_list);
TAILQ_INIT(&sc->tm_list);
+   TAILQ_INIT(&sc->io_list);
 
if (((error = mps_alloc_queues(sc)) != 0) ||
((error = mps_alloc_replies(sc)) != 0) ||

Modified: head/sys/dev/mps/mps_sas.c
==
--- head/sys/dev/mps/mps_sas.c  Fri Feb 18 16:29:38 2011(r218810)
+++ head/sys/dev/mps/mps_sas.c  Fri Feb 18 17:01:57 2011(r218811)
@@ -486,7 +486,10 @@ mpssas_prepare_remove(struct mpssas_soft
return;
}
 
+   mps_dprint(sc, MPS_INFO, "Preparing to remove target %d\n", targ->tid);
+
req = (MPI2_SCSI_TASK_MANAGE_REQUEST *)cm->cm_req;
+   memset(req, 0, sizeof(*req));
req->DevHandle = targ->handle;
req->Function = MPI2_FUNCTION_SCSI_TASK_MGMT;
req->TaskType = MPI2_SCSITASKMGMT_TASKTYPE_TARGET_RESET;
@@ -507,6 +510,7 @@ mpssas_remove_device(struct mps_softc *s
MPI2_SCSI_TASK_MANAGE_REPLY *reply;
MPI2_SAS_IOUNIT_CONTROL_REQUEST *req;
struct mpssas_target *targ;
+   struct mps_command *next_cm;
uint16_t handle;
 
mps_dprint(sc, MPS_TRACE, "%s\n", __func__);
@@ -523,11 +527,13 @@ mpssas_remove_device(struct mps_softc *s
return;
}
 
-   mps_printf(sc, "Reset aborted %d commands\n", reply->TerminationCount);
+   mps_dprint(sc, MPS_INFO, "Reset aborted %u commands\n",
+   reply->TerminationCount);
mps_free_reply(sc, cm->cm_reply_data);
 
/* Reuse the existing command */
req = (MPI2_SAS_IOUNIT_CONTROL_REQUEST *)cm->cm_req;
+   memset(req, 0, sizeof(*req));
req->Function = MPI2_FUNCTION_SAS_IO_UNIT_CONTROL;
req->Operation = MPI2_SAS_OP_REMOVE_DEVICE;
req->DevHandle = handle;
@@ -539,6 +545,17 @@ mpssas_remove_device(struct mps_softc *s
mps_map_command(sc, cm);
 
mps_dprint(sc, MPS_INFO, "clearing target handle 0x%04x\n", handle);
+   TAILQ_FOREACH_SAFE(cm, &sc->io_list, cm_link, next_cm) {
+   union ccb *ccb;
+
+   if (cm->cm_targ->handle != handle)
+   continue;
+
+   mps_dprint(sc, MPS_INFO, "Completing missed command %p\n", cm);
+   ccb = cm->cm_complete_data;
+   ccb->ccb_h.status = CAM_DEV_NOT_THERE;
+   mpssas_scsiio_complete(sc, cm);
+   }
targ = mpssas_find_target(sc->sassc, 0, handle);
if (targ != NULL) {
targ->handle = 0x0;
@@ -1430,6 +1447,7 @@ mpssas_action_scsiio(struct mpssas_softc
cm->cm_complete_data = ccb;
cm->cm_targ = targ;
 
+   TAILQ_INSERT_TAIL(&sc->io_list, cm, cm_link);
callout_reset(&cm->cm_callout, (ccb->ccb_h.timeout * hz) / 1000,
   mpssas_scsiio_timeout, cm);
 
@@ -1449,6 +1467,7 @@ mpssas_scsiio_complete(struct mps_softc 
mps_dprint(sc, MPS_TRACE, "%s\n", __func__);
 
callout_stop(&cm->cm_callout);
+   TAILQ_REMOVE(&sc->io_list, cm, cm_link);
 
sassc = sc->sassc;
ccb = cm->cm_complete_data;
@@ -1470,8 +1489,10 @@ mpssas_scsiio_complete(struct mps_softc 
 
/* Take the fast path to completion */
if (cm->cm_reply == NULL) {
-   ccb->ccb_h.status = CAM_REQ_CMP;
-   ccb->csio.scsi_status = SCSI_STATUS_OK;
+   if ((ccb->ccb_h.status & CAM_STATUS_MASK) == CAM_REQ_INPROG) {
+ 

svn commit: r218812 - head/sys/dev/mps

[Kenneth D. Merry]

Author: ken
Date: Fri Feb 18 17:06:06 2011
New Revision: 218812
URL: http://svn.freebsd.org/changeset/base/218812

Log:
  Fix several issues with the mps(4) driver.
  
  When the driver ran out of DMA chaining buffers, it kept the timeout for
  the I/O, and I/O would stall.
  
  The driver was not freezing the device queue on errors.
  
  mps.c:Pull command completion logic into a separate
function, and call the callback/wakeup for commands
that are never sent due to lack of chain buffers.
  
Add a number of extra diagnostic sysctl variables.
  
Handle pre-hardware errors for configuration I/O.
This doesn't panic the system, but it will fail the
configuration I/O and there is no retry mechanism.
So the device probe will not succeed.  This should
be a very uncommon situation, however.
  
  mps_sas.c:Freeze the SIM queue when we run out of chain
buffers, and unfreeze it when more commands
complete.
  
Freeze the device queue when errors occur, so that
CAM can insure proper command ordering.
  
Report pre-hardware errors for task management
commands.  In general, that shouldn't be possible
because task management commands don't have S/G
lists, and that is currently the only error path
before we get to the hardware.
  
Handle pre-hardware errors (like out of chain
elements) for SMP requests.  That shouldn't happen
either, since we should have enough space for two
S/G elements in the standard request.
  
For commands that end with
MPI2_IOCSTATUS_SCSI_IOC_TERMINATED and
MPI2_IOCSTATUS_SCSI_EXT_TERMINATED, return them
with CAM_REQUEUE_REQ to retry them unconditionally.
These seem to be related to back end, transport
related problems that are hopefully transient.  We
don't want to go through the retry count for
something that is not a permanent error.
  
Keep track of the number of outstanding I/Os.
  
  mpsvar.h: Track the number of free chain elements.
  
Add variables for the number of outstanding I/Os,
and I/O high water mark.
  
Add variables to track the number of free chain
buffers and the chain low water mark, as well as
the number of chain allocation failures.
  
Add I/O state flags and an attach done flag.
  
  MFC after:3 days

Modified:
  head/sys/dev/mps/mps.c
  head/sys/dev/mps/mps_sas.c
  head/sys/dev/mps/mpsvar.h

Modified: head/sys/dev/mps/mps.c
==
--- head/sys/dev/mps/mps.c  Fri Feb 18 17:01:57 2011(r218811)
+++ head/sys/dev/mps/mps.c  Fri Feb 18 17:06:06 2011(r218812)
@@ -62,6 +62,7 @@ static void mps_startup(void *arg);
 static void mps_startup_complete(struct mps_softc *sc, struct mps_command *cm);
 static int mps_send_iocinit(struct mps_softc *sc);
 static int mps_attach_log(struct mps_softc *sc);
+static __inline void mps_complete_command(struct mps_command *cm);
 static void mps_dispatch_event(struct mps_softc *sc, uintptr_t data, 
MPI2_EVENT_NOTIFICATION_REPLY *reply);
 static void mps_config_complete(struct mps_softc *sc, struct mps_command *cm);
 static void mps_periodic(void *);
@@ -387,6 +388,15 @@ mps_enqueue_request(struct mps_softc *sc
 
mps_dprint(sc, MPS_TRACE, "%s\n", __func__);
 
+   if (sc->mps_flags & MPS_FLAGS_ATTACH_DONE)
+   mtx_assert(&sc->mps_mtx, MA_OWNED);
+
+   if ((cm->cm_desc.Default.SMID < 1)
+|| (cm->cm_desc.Default.SMID >= sc->num_reqs)) {
+   mps_printf(sc, "%s: invalid SMID %d, desc %#x %#x\n",
+  __func__, cm->cm_desc.Default.SMID,
+  cm->cm_desc.Words.High, cm->cm_desc.Words.Low);
+   }
mps_regwrite(sc, MPI2_REQUEST_DESCRIPTOR_POST_LOW_OFFSET,
cm->cm_desc.Words.Low);
mps_regwrite(sc, MPI2_REQUEST_DESCRIPTOR_POST_HIGH_OFFSET,
@@ -732,6 +742,7 @@ mps_alloc_requests(struct mps_softc *sc)
chain->chain_busaddr = sc->chain_busaddr +
i * sc->facts->IOCRequestFrameSize * 4;
mps_free_chain(sc, chain);
+   sc->chain_free_lowwater++;
}
 
/* XXX Need to pick a more precise value */
@@ -855,6 +866,26 @@ mps_attach(struct mps_softc *sc)
&sc->allow_multiple_tm_cmds, 0,
"allow multiple simultaneous task management cmds");
 
+   SYSCTL_ADD_INT(&sc->sysctl_ctx, SYSCTL_CHILDREN(sc->sysctl_tree),
+   OID_AUTO, "io_cmds_active", CTLFLAG_RD,
+ 

svn commit: r218921 - stable/8/sys/dev/mps

[Kenneth D. Merry]

Author: ken
Date: Mon Feb 21 16:55:53 2011
New Revision: 218921
URL: http://svn.freebsd.org/changeset/base/218921

Log:
  MFC: r218811:
  
  In the MPS driver, during device removal processing, don't assume that
  the controller firmware will return all of our commands.  Instead, keep
  track of outstanding I/Os and return them to CAM once device removal
  processing completes.
  
  mpsvar.h: Declare the new "io_list" in the mps_softc.
  
  mps.c:Initialize the new "io_list" in the mps softc.
  
  mps_sas.c:o Track SCSI I/O requests on the io_list from the
  time of mpssas_action() through mpssas_scsiio_complete().
o Zero out the request structures used for device
  removal commands prior to filling them out.
o Once the target reset task management function completes
  during device removal processing, assume any SCSI I/O
  commands that are still oustanding will never return
  from the controller, and process them manually.
  
  Submitted by: gibbs

Modified:
  stable/8/sys/dev/mps/mps.c
  stable/8/sys/dev/mps/mps_sas.c
  stable/8/sys/dev/mps/mpsvar.h
Directory Properties:
  stable/8/sys/   (props changed)
  stable/8/sys/amd64/include/xen/   (props changed)
  stable/8/sys/cddl/contrib/opensolaris/   (props changed)
  stable/8/sys/contrib/dev/acpica/   (props changed)
  stable/8/sys/contrib/pf/   (props changed)

Modified: stable/8/sys/dev/mps/mps.c
==
--- stable/8/sys/dev/mps/mps.c  Mon Feb 21 16:33:01 2011(r218920)
+++ stable/8/sys/dev/mps/mps.c  Mon Feb 21 16:55:53 2011(r218921)
@@ -898,6 +898,7 @@ mps_attach(struct mps_softc *sc)
TAILQ_INIT(&sc->req_list);
TAILQ_INIT(&sc->chain_list);
TAILQ_INIT(&sc->tm_list);
+   TAILQ_INIT(&sc->io_list);
 
if (((error = mps_alloc_queues(sc)) != 0) ||
((error = mps_alloc_replies(sc)) != 0) ||

Modified: stable/8/sys/dev/mps/mps_sas.c
==
--- stable/8/sys/dev/mps/mps_sas.c  Mon Feb 21 16:33:01 2011
(r218920)
+++ stable/8/sys/dev/mps/mps_sas.c  Mon Feb 21 16:55:53 2011
(r218921)
@@ -486,7 +486,10 @@ mpssas_prepare_remove(struct mpssas_soft
return;
}
 
+   mps_dprint(sc, MPS_INFO, "Preparing to remove target %d\n", targ->tid);
+
req = (MPI2_SCSI_TASK_MANAGE_REQUEST *)cm->cm_req;
+   memset(req, 0, sizeof(*req));
req->DevHandle = targ->handle;
req->Function = MPI2_FUNCTION_SCSI_TASK_MGMT;
req->TaskType = MPI2_SCSITASKMGMT_TASKTYPE_TARGET_RESET;
@@ -507,6 +510,7 @@ mpssas_remove_device(struct mps_softc *s
MPI2_SCSI_TASK_MANAGE_REPLY *reply;
MPI2_SAS_IOUNIT_CONTROL_REQUEST *req;
struct mpssas_target *targ;
+   struct mps_command *next_cm;
uint16_t handle;
 
mps_dprint(sc, MPS_TRACE, "%s\n", __func__);
@@ -523,11 +527,13 @@ mpssas_remove_device(struct mps_softc *s
return;
}
 
-   mps_printf(sc, "Reset aborted %d commands\n", reply->TerminationCount);
+   mps_dprint(sc, MPS_INFO, "Reset aborted %u commands\n",
+   reply->TerminationCount);
mps_free_reply(sc, cm->cm_reply_data);
 
/* Reuse the existing command */
req = (MPI2_SAS_IOUNIT_CONTROL_REQUEST *)cm->cm_req;
+   memset(req, 0, sizeof(*req));
req->Function = MPI2_FUNCTION_SAS_IO_UNIT_CONTROL;
req->Operation = MPI2_SAS_OP_REMOVE_DEVICE;
req->DevHandle = handle;
@@ -539,6 +545,17 @@ mpssas_remove_device(struct mps_softc *s
mps_map_command(sc, cm);
 
mps_dprint(sc, MPS_INFO, "clearing target handle 0x%04x\n", handle);
+   TAILQ_FOREACH_SAFE(cm, &sc->io_list, cm_link, next_cm) {
+   union ccb *ccb;
+
+   if (cm->cm_targ->handle != handle)
+   continue;
+
+   mps_dprint(sc, MPS_INFO, "Completing missed command %p\n", cm);
+   ccb = cm->cm_complete_data;
+   ccb->ccb_h.status = CAM_DEV_NOT_THERE;
+   mpssas_scsiio_complete(sc, cm);
+   }
targ = mpssas_find_target(sc->sassc, 0, handle);
if (targ != NULL) {
targ->handle = 0x0;
@@ -1430,6 +1447,7 @@ mpssas_action_scsiio(struct mpssas_softc
cm->cm_complete_data = ccb;
cm->cm_targ = targ;
 
+   TAILQ_INSERT_TAIL(&sc->io_list, cm, cm_link);
callout_reset(&cm->cm_callout, (ccb->ccb_h.timeout * hz) / 1000,
   mpssas_scsiio_timeout, cm);
 
@@ -1449,6 +1467,7 @@ mpssas_scsiio_complete(struct mps_softc 
mps_dprint(sc, MPS_TRACE, "%s\n", __func__);
 
callout_stop(&cm->cm_callout);
+   TAILQ_REMOVE(&sc->io_list, cm, cm_link);
 
sassc = sc->sassc;
ccb = cm->cm_complete_data;
@@ -1470,8 +1489,10 @@ mpssas_scsiio

svn commit: r218922 - stable/8/sys/dev/mps

[Kenneth D. Merry]

Author: ken
Date: Mon Feb 21 18:11:56 2011
New Revision: 218922
URL: http://svn.freebsd.org/changeset/base/218922

Log:
  MFC: r218812
  
  Fix several issues with the mps(4) driver.
  
  When the driver ran out of DMA chaining buffers, it kept the timeout for
  the I/O, and I/O would stall.
  
  The driver was not freezing the device queue on errors.
  
  mps.c:Pull command completion logic into a separate
function, and call the callback/wakeup for commands
that are never sent due to lack of chain buffers.
  
Add a number of extra diagnostic sysctl variables.
  
Handle pre-hardware errors for configuration I/O.
This doesn't panic the system, but it will fail the
configuration I/O and there is no retry mechanism.
So the device probe will not succeed.  This should
be a very uncommon situation, however.
  
  mps_sas.c:Freeze the SIM queue when we run out of chain
buffers, and unfreeze it when more commands
complete.
  
Freeze the device queue when errors occur, so that
CAM can insure proper command ordering.
  
Report pre-hardware errors for task management
commands.  In general, that shouldn't be possible
because task management commands don't have S/G
lists, and that is currently the only error path
before we get to the hardware.
  
Handle pre-hardware errors (like out of chain
elements) for SMP requests.  That shouldn't happen
either, since we should have enough space for two
S/G elements in the standard request.
  
For commands that end with
MPI2_IOCSTATUS_SCSI_IOC_TERMINATED and
MPI2_IOCSTATUS_SCSI_EXT_TERMINATED, return them
with CAM_REQUEUE_REQ to retry them unconditionally.
These seem to be related to back end, transport
related problems that are hopefully transient.  We
don't want to go through the retry count for
something that is not a permanent error.
  
Keep track of the number of outstanding I/Os.
  
  mpsvar.h: Track the number of free chain elements.
  
Add variables for the number of outstanding I/Os,
and I/O high water mark.
  
Add variables to track the number of free chain
buffers and the chain low water mark, as well as
the number of chain allocation failures.
  
Add I/O state flags and an attach done flag.

Modified:
  stable/8/sys/dev/mps/mps.c
  stable/8/sys/dev/mps/mps_sas.c
  stable/8/sys/dev/mps/mpsvar.h
Directory Properties:
  stable/8/sys/   (props changed)
  stable/8/sys/amd64/include/xen/   (props changed)
  stable/8/sys/cddl/contrib/opensolaris/   (props changed)
  stable/8/sys/contrib/dev/acpica/   (props changed)
  stable/8/sys/contrib/pf/   (props changed)

Modified: stable/8/sys/dev/mps/mps.c
==
--- stable/8/sys/dev/mps/mps.c  Mon Feb 21 16:55:53 2011(r218921)
+++ stable/8/sys/dev/mps/mps.c  Mon Feb 21 18:11:56 2011(r218922)
@@ -62,6 +62,7 @@ static void mps_startup(void *arg);
 static void mps_startup_complete(struct mps_softc *sc, struct mps_command *cm);
 static int mps_send_iocinit(struct mps_softc *sc);
 static int mps_attach_log(struct mps_softc *sc);
+static __inline void mps_complete_command(struct mps_command *cm);
 static void mps_dispatch_event(struct mps_softc *sc, uintptr_t data, 
MPI2_EVENT_NOTIFICATION_REPLY *reply);
 static void mps_config_complete(struct mps_softc *sc, struct mps_command *cm);
 static void mps_periodic(void *);
@@ -387,6 +388,15 @@ mps_enqueue_request(struct mps_softc *sc
 
mps_dprint(sc, MPS_TRACE, "%s\n", __func__);
 
+   if (sc->mps_flags & MPS_FLAGS_ATTACH_DONE)
+   mtx_assert(&sc->mps_mtx, MA_OWNED);
+
+   if ((cm->cm_desc.Default.SMID < 1)
+|| (cm->cm_desc.Default.SMID >= sc->num_reqs)) {
+   mps_printf(sc, "%s: invalid SMID %d, desc %#x %#x\n",
+  __func__, cm->cm_desc.Default.SMID,
+  cm->cm_desc.Words.High, cm->cm_desc.Words.Low);
+   }
mps_regwrite(sc, MPI2_REQUEST_DESCRIPTOR_POST_LOW_OFFSET,
cm->cm_desc.Words.Low);
mps_regwrite(sc, MPI2_REQUEST_DESCRIPTOR_POST_HIGH_OFFSET,
@@ -732,6 +742,7 @@ mps_alloc_requests(struct mps_softc *sc)
chain->chain_busaddr = sc->chain_busaddr +
i * sc->facts->IOCRequestFrameSize * 4;
mps_free_chain(sc, chain);
+   sc->chain_free_lowwater++;
}
 
/* XXX Need to pick a more precise value */
@@ -855,6 +866,26 @@ mps_att

svn commit: r219036 - head/sys/dev/mps

[Kenneth D. Merry]

Author: ken
Date: Fri Feb 25 17:30:25 2011
New Revision: 219036
URL: http://svn.freebsd.org/changeset/base/219036

Log:
  Silence 'out of chain frames' warnings and bump the number of frames.
  
  mps.c:Hide the 'out of chain frames' warning behind MPS_INFO.
  
  mps_sas.c:Hide the SIM queue freeze/unfreeze messages behind MPS_INFO.
  
  mpsvar.h: Bump the number of chain frames from 1024 to 2048.  From
testing, it looks like this makes it less likely that we'll
run out of chain frames, and it doesn't cost much memory
(32K).
  
  MFC after:3 days

Modified:
  head/sys/dev/mps/mps.c
  head/sys/dev/mps/mps_sas.c
  head/sys/dev/mps/mpsvar.h

Modified: head/sys/dev/mps/mps.c
==
--- head/sys/dev/mps/mps.c  Fri Feb 25 16:08:31 2011(r219035)
+++ head/sys/dev/mps/mps.c  Fri Feb 25 17:30:25 2011(r219036)
@@ -1705,7 +1705,7 @@ mps_data_cb(void *arg, bus_dma_segment_t
sflags, nsegs - i);
if (error != 0) {
/* Resource shortage, roll back! */
-   mps_printf(sc, "out of chain frames\n");
+   mps_dprint(sc, MPS_INFO, "out of chain frames\n");
cm->cm_flags |= MPS_CM_FLAGS_CHAIN_FAILED;
mps_complete_command(cm);
return;

Modified: head/sys/dev/mps/mps_sas.c
==
--- head/sys/dev/mps/mps_sas.c  Fri Feb 25 16:08:31 2011(r219035)
+++ head/sys/dev/mps/mps_sas.c  Fri Feb 25 17:30:25 2011(r219036)
@@ -1542,8 +1542,8 @@ mpssas_scsiio_complete(struct mps_softc 
if ((sassc->flags & MPSSAS_QUEUE_FROZEN) == 0) {
xpt_freeze_simq(sassc->sim, 1);
sassc->flags |= MPSSAS_QUEUE_FROZEN;
-   mps_printf(sc, "Error sending command, freezing "
-  "SIM queue\n");
+   mps_dprint(sc, MPS_INFO, "Error sending command, "
+  "freezing SIM queue\n");
}
}
 
@@ -1556,7 +1556,8 @@ mpssas_scsiio_complete(struct mps_softc 
if (sassc->flags & MPSSAS_QUEUE_FROZEN) {
ccb->ccb_h.status |= CAM_RELEASE_SIMQ;
sassc->flags &= ~MPSSAS_QUEUE_FROZEN;
-   mps_printf(sc, "Unfreezing SIM queue\n");
+   mps_dprint(sc, MPS_INFO,
+  "Unfreezing SIM queue\n");
}
} else {
ccb->ccb_h.status |= CAM_DEV_QFRZN;

Modified: head/sys/dev/mps/mpsvar.h
==
--- head/sys/dev/mps/mpsvar.h   Fri Feb 25 16:08:31 2011(r219035)
+++ head/sys/dev/mps/mpsvar.h   Fri Feb 25 17:30:25 2011(r219036)
@@ -34,7 +34,7 @@
 #define MPS_REQ_FRAMES 1024
 #define MPS_EVT_REPLY_FRAMES   32
 #define MPS_REPLY_FRAMES   MPS_REQ_FRAMES
-#define MPS_CHAIN_FRAMES   1024
+#define MPS_CHAIN_FRAMES   2048
 #define MPS_SENSE_LEN  SSD_FULL_SIZE
 #define MPS_MSI_COUNT  1
 #define MPS_SGE64_SIZE 12
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r223062 - head/sys/kern

[Kenneth D. Merry]

Author: ken
Date: Mon Jun 13 22:08:24 2011
New Revision: 223062
URL: http://svn.freebsd.org/changeset/base/223062

Log:
  Instead of using an atomic operation to determine whether the devstat(9)
  device node has been created, pass MAKEDEV_CHECKNAME in so that the devfs
  code will do the check.
  
  Use a regular static variable as before, that's good enough to keep us from
  calling into devfs most of the time.
  
  Suggested by: kib
  MFC after:1 week
  Sponsored by: Spectra Logic Corporation

Modified:
  head/sys/kern/subr_devstat.c

Modified: head/sys/kern/subr_devstat.c
==
--- head/sys/kern/subr_devstat.cMon Jun 13 21:21:02 2011
(r223061)
+++ head/sys/kern/subr_devstat.cMon Jun 13 22:08:24 2011
(r223062)
@@ -470,9 +470,11 @@ devstat_alloc(void)
static int once;
 
mtx_assert(&devstat_mutex, MA_NOTOWNED);
-   if (!once && atomic_cmpset_int(&once, 0, 1)) {
-   make_dev_credf(MAKEDEV_ETERNAL, &devstat_cdevsw, 0, NULL,
-   UID_ROOT, GID_WHEEL, 0400, DEVSTAT_DEVICE_NAME);
+   if (!once) {
+   make_dev_credf(MAKEDEV_ETERNAL | MAKEDEV_CHECKNAME,
+   &devstat_cdevsw, 0, NULL, UID_ROOT, GID_WHEEL, 0400,
+   DEVSTAT_DEVICE_NAME);
+   once = 1;
}
spp2 = NULL;
mtx_lock(&devstat_mutex);
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

Re: svn commit: r223061 - head/sys/kern

[Kenneth D. Merry]

On Tue, Jun 14, 2011 at 00:31:50 +0300, Kostik Belousov wrote:
> On Mon, Jun 13, 2011 at 09:21:02PM +, Justin T. Gibbs wrote:
> > Author: gibbs
> > Date: Mon Jun 13 21:21:02 2011
> > New Revision: 223061
> > URL: http://svn.freebsd.org/changeset/base/223061
> > 
> > Log:
> >   Fix a couple of race conditions in devstat(9) initialization.
> >   
> >   In devstat_new_entry(), there is no need to initialize the queue
> >   and the mutex in this function.  There are ways to do static
> >   initialization on both, so use STAILQ_HEAD_INITIALIZER and
> >   MTX_SYSINIT to initialize the queue and the mutex.
> >   
> >   In devstat_alloc(), use an atomic test and set routine to guard
> >   making our entry in /dev.  Using just a plain static variable
> >   creates a race condition on multiprocessor machines.  If you
> >   attempt to create a second entry in devfs, the kernel will panic.
> Devfs returns an error if MAKEDEV_CHECKNAME flag is supplied and
> attempt is made to create the existing node. The static guard is
> still useful, since make_dev() call is costly, but you can remove
> the atomic, since the race should be of limited scope.

Done, thanks for the suggestion!

Ken
-- 
Kenneth Merry
k...@freebsd.org
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r219099 - stable/8/sys/dev/mps

[Kenneth D. Merry]

Author: ken
Date: Mon Feb 28 16:39:15 2011
New Revision: 219099
URL: http://svn.freebsd.org/changeset/base/219099

Log:
  MFC: r219036
  
  Silence 'out of chain frames' warnings and bump the number of frames.
  
  mps.c:Hide the 'out of chain frames' warning behind MPS_INFO.
  
  mps_sas.c:Hide the SIM queue freeze/unfreeze messages behind MPS_INFO.
  
  mpsvar.h: Bump the number of chain frames from 1024 to 2048.  From
testing, it looks like this makes it less likely that we'll
run out of chain frames, and it doesn't cost much memory
(32K).

Modified:
  stable/8/sys/dev/mps/mps.c
  stable/8/sys/dev/mps/mps_sas.c
  stable/8/sys/dev/mps/mpsvar.h
Directory Properties:
  stable/8/sys/   (props changed)
  stable/8/sys/amd64/include/xen/   (props changed)
  stable/8/sys/cddl/contrib/opensolaris/   (props changed)
  stable/8/sys/contrib/dev/acpica/   (props changed)
  stable/8/sys/contrib/pf/   (props changed)

Modified: stable/8/sys/dev/mps/mps.c
==
--- stable/8/sys/dev/mps/mps.c  Mon Feb 28 15:58:30 2011(r219098)
+++ stable/8/sys/dev/mps/mps.c  Mon Feb 28 16:39:15 2011(r219099)
@@ -1705,7 +1705,7 @@ mps_data_cb(void *arg, bus_dma_segment_t
sflags, nsegs - i);
if (error != 0) {
/* Resource shortage, roll back! */
-   mps_printf(sc, "out of chain frames\n");
+   mps_dprint(sc, MPS_INFO, "out of chain frames\n");
cm->cm_flags |= MPS_CM_FLAGS_CHAIN_FAILED;
mps_complete_command(cm);
return;

Modified: stable/8/sys/dev/mps/mps_sas.c
==
--- stable/8/sys/dev/mps/mps_sas.c  Mon Feb 28 15:58:30 2011
(r219098)
+++ stable/8/sys/dev/mps/mps_sas.c  Mon Feb 28 16:39:15 2011
(r219099)
@@ -1542,8 +1542,8 @@ mpssas_scsiio_complete(struct mps_softc 
if ((sassc->flags & MPSSAS_QUEUE_FROZEN) == 0) {
xpt_freeze_simq(sassc->sim, 1);
sassc->flags |= MPSSAS_QUEUE_FROZEN;
-   mps_printf(sc, "Error sending command, freezing "
-  "SIM queue\n");
+   mps_dprint(sc, MPS_INFO, "Error sending command, "
+  "freezing SIM queue\n");
}
}
 
@@ -1556,7 +1556,8 @@ mpssas_scsiio_complete(struct mps_softc 
if (sassc->flags & MPSSAS_QUEUE_FROZEN) {
ccb->ccb_h.status |= CAM_RELEASE_SIMQ;
sassc->flags &= ~MPSSAS_QUEUE_FROZEN;
-   mps_printf(sc, "Unfreezing SIM queue\n");
+   mps_dprint(sc, MPS_INFO,
+  "Unfreezing SIM queue\n");
}
} else {
ccb->ccb_h.status |= CAM_DEV_QFRZN;

Modified: stable/8/sys/dev/mps/mpsvar.h
==
--- stable/8/sys/dev/mps/mpsvar.h   Mon Feb 28 15:58:30 2011
(r219098)
+++ stable/8/sys/dev/mps/mpsvar.h   Mon Feb 28 16:39:15 2011
(r219099)
@@ -34,7 +34,7 @@
 #define MPS_REQ_FRAMES 1024
 #define MPS_EVT_REPLY_FRAMES   32
 #define MPS_REPLY_FRAMES   MPS_REQ_FRAMES
-#define MPS_CHAIN_FRAMES   1024
+#define MPS_CHAIN_FRAMES   2048
 #define MPS_SENSE_LEN  SSD_FULL_SIZE
 #define MPS_MSI_COUNT  1
 #define MPS_SGE64_SIZE 12
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r219678 - stable/8/share/man/man4

[Kenneth D. Merry]

Author: ken
Date: Tue Mar 15 21:34:10 2011
New Revision: 219678
URL: http://svn.freebsd.org/changeset/base/219678

Log:
  MFC: 212624, 212669, 213535
  
  Merge the mps(4) man page.
  
r212624 | ken | 2010-09-14 15:40:29 -0600 (Tue, 14 Sep 2010) | 7 lines
  
MFp4 (//depot/projects/mps/...)
  
Add a man page for the mps(4) driver, and reference it in the
mpt(4) driver man page.
  
Sponsored by:   Spectra Logic Corporation
  
r212669 | ken | 2010-09-15 10:05:51 -0600 (Wed, 15 Sep 2010) | 6 lines
  
MFp4 (//depot/projects/mps/...)
  
Fix a typo.
  
Submitted by:   pluknet at gmail dot com
  
r213535 | ken | 2010-10-07 15:56:10 -0600 (Thu, 07 Oct 2010) | 36 lines
  
Turn on serialization of task management commands going down to the
controller, but make it optional.
  
After a problem report from Andrew Boyer, it looks like the LSI
chip may have issues (the watchdog timer fired) if too many aborts
are sent down to the chip at the same time.  We know that task
management commands are serialized, and although the manual doesn't
say it, it may be a good idea to just send one at a time.
  
But, since I'm not certain that this is necessary, add a tunable
and sysctl variable (hw.mps.%d.allow_multiple_tm_cmds) to control
the driver's behavior.
  
mps.c:  Add support for the sysctl and tunable, and add a
comment about the possible return values to
mps_map_command().
  
mps_sas.c:  Run all task management commands through two new
routines, mpssas_issue_tm_request() and
mpssas_complete_tm_request().
  
This allows us to optionally serialize task
management commands.  Also, change things so that
the response to a task management command always
comes back through the callback.  (Before it could
come via the callback or the return value.)
  
mpsvar.h:   Add softc variables for the list of active task
management commands, the number of active commands,
and whether we should allow multiple active task
management commands.  Add an active command flag.
  
mps.4:  Describe the new sysctl/loader tunable variable.
  
Sponsored by:   Spectra Logic Corporation
  
  Reminded by:  Hubert Tournier 

Added:
  stable/8/share/man/man4/mps.4
 - copied, changed from r212624, head/share/man/man4/mps.4
Modified:
  stable/8/share/man/man4/Makefile
  stable/8/share/man/man4/mpt.4
Directory Properties:
  stable/8/share/man/man4/   (props changed)

Modified: stable/8/share/man/man4/Makefile
==
--- stable/8/share/man/man4/MakefileTue Mar 15 20:24:21 2011
(r219677)
+++ stable/8/share/man/man4/MakefileTue Mar 15 21:34:10 2011
(r219678)
@@ -218,6 +218,7 @@ MAN=aac.4 \
mmcsd.4 \
mn.4 \
mouse.4 \
+   mps.4 \
mpt.4 \
msk.4 \
mtio.4 \

Copied and modified: stable/8/share/man/man4/mps.4 (from r212624, 
head/share/man/man4/mps.4)
==
--- head/share/man/man4/mps.4   Tue Sep 14 21:40:29 2010(r212624, copy 
source)
+++ stable/8/share/man/man4/mps.4   Tue Mar 15 21:34:10 2011
(r219678)
@@ -31,7 +31,7 @@
 .\"
 .\" Author: Ken Merry 
 .\"
-.\" $Id: //depot/SpectraBSD/head/share/man/man4/mps.4#1 $
+.\" $Id: //depot/SpectraBSD/head/share/man/man4/mps.4#4 $
 .\" $FreeBSD$
 .\"
 .Dd September 13, 2010
@@ -51,7 +51,7 @@ place the following lines in your kernel
 Or, to load the driver as a module at boot, place the following line in
 .Xr loader.conf 5 :
 .Bd -literal -offset indent
-mpt_load="YES"
+mps_load="YES"
 .Ed
 .Sh DESCRIPTION
 The
@@ -97,6 +97,20 @@ driver instances, set the following tuna
 .Bd -literal -offset indent
 hw.mps.disable_msix=1
 .Ed
+.Pp
+To allow the driver to send multiple task management commands (like abort,
+LUN reset, etc.), set the following variable:
+.Bd -literal -offset indent
+hw.mps.X.allow_multiple_tm_cmds=1
+.Ed
+.Pp
+via
+.Xr loader.conf 5
+or
+.Xr sysctl 8 ,
+where X is the adapter number.
+By default the driver only sends one task management command at a time, to
+avoid causing a potential controller lock-up.
 .Sh DEBUGGING
 To enable debugging prints from the
 .Nm

Modified: stable/8/share/man/man4/mpt.4
==
--- stable/8/share/man/man4/mpt.4   Tue Mar 15 20:24:21 2011
(r219677)
+++ stable/8/share/man/man4/mpt.4   Tue Mar 15 21:34:10 2011
(r219678)
@@ -152,6 +152,7 @@ can take on - no separate compilation is
 .Xr cd 4 ,
 .Xr ch 4 ,
 .Xr da 4 ,
+.Xr mps 4 ,
 .Xr pci 4 ,
 .Xr sa 4 ,
 .Xr scsi 4 ,
___
svn-src-all@freebsd.org ma

svn commit: r222537 - in head/sys: kern sys

[Kenneth D. Merry]

Author: ken
Date: Tue May 31 17:29:58 2011
New Revision: 222537
URL: http://svn.freebsd.org/changeset/base/222537

Log:
  Fix apparent garbage in the message buffer.
  
  While we have had a fix in place (options PRINTF_BUFR_SIZE=128) to fix
  scrambled console output, the message buffer and syslog were still getting
  log messages one character at a time.  While all of the characters still
  made it into the log (courtesy of atomic operations), they were often
  interleaved when there were multiple threads writing to the buffer at the
  same time.
  
  This fixes message buffer accesses to use buffering logic as well, so that
  strings that are less than PRINTF_BUFR_SIZE will be put into the message
  buffer atomically.  So now dmesg output should look the same as console
  output.
  
  subr_msgbuf.c:Convert most message buffer calls to use a new 
spin
lock instead of atomic variables in some places.
  
Add a new routine, msgbuf_addstr(), that adds a
NUL-terminated string to a message buffer.  This
takes a priority argument, which allows us to
eliminate some races (at least in the the string
at a time case) that are present in the
implementation of msglogchar().  (dangling and
lastpri are static variables, and are subject to
races when multiple callers are present.)
  
msgbuf_addstr() also allows the caller to request
that carriage returns be stripped out of the
string.  This matches the behavior of msglogchar(),
but in testing so far it doesn't appear that any
newlines are being stripped out.  So the carriage
return removal functionality may be a candidate
for removal later on if further analysis shows
that it isn't necessary.
  
  subr_prf.c:   Add a new msglogstr() routine that calls
msgbuf_logstr().
  
Rename putcons() to putbuf().  This now handles
buffered output to the message log as well as
the console.  Also, remove the logic in putcons()
(now putbuf()) that added a carriage return before
a newline.  The console path was the only path that
needed it, and cnputc() (called by cnputs())
already adds a carriage return.  So this
duplication resulted in kernel-generated console
output lines ending in '\r''\r''\n'.
  
Refactor putchar() to handle the new buffering
scheme.
  
Add buffering to log().
  
Change log_console() to use msglogstr() instead of
msglogchar().  Don't add extra newlines by default
in log_console().  Hide that behavior behind a
tunable/sysctl (kern.log_console_add_linefeed) for
those who would like the old behavior.  The old
behavior led to the insertion of extra newlines
for log output for programs that print out a
string, and then a trailing newline on a separate
write.  (This is visible with dmesg -a.)
  
  msgbuf.h: Add a prototype for msgbuf_addstr().
  
Add three new fields to struct msgbuf, msg_needsnl,
msg_lastpri and msg_lock.  The first two are needed
for log message functionality previously handled
by msglogchar().  (Which is still active if
buffering isn't enabled.)
  
Include sys/lock.h and sys/mutex.h for the new
mutex.
  
  Reviewed by:  gibbs

Modified:
  head/sys/kern/subr_msgbuf.c
  head/sys/kern/subr_prf.c
  head/sys/sys/msgbuf.h

Modified: head/sys/kern/subr_msgbuf.c
==
--- head/sys/kern/subr_msgbuf.c Tue May 31 17:24:18 2011(r222536)
+++ head/sys/kern/subr_msgbuf.c Tue May 31 17:29:58 2011(r222537)
@@ -31,8 +31,16 @@
 
 #include 
 #include 
+#include 
+#include 
 #include 
 
+/*
+ * Maximum number conversion buffer length: uintmax_t in base 2, plus <>
+ * around the priority, and a terminating NUL.
+ */
+#defineMAXPRIBUF   (sizeof(intmax_t) * NBBY + 3)
+
 /* Read/write sequence numbers are modulo a multiple of the buffer size. */
 #define SEQMOD(size) ((size) * 16)
 
@@ -51,6 +59,9 @@ msgbuf_init(struct msgbuf *mbp, void *pt
mbp->msg_s

Re: svn commit: r222537 - in head/sys: kern sys

[Kenneth D. Merry]

On Tue, May 31, 2011 at 14:00:18 -0700, m...@freebsd.org wrote:
> On Tue, May 31, 2011 at 10:29 AM, Kenneth D. Merry  wrote:
> > Author: ken
> > Date: Tue May 31 17:29:58 2011
> > New Revision: 222537
> > URL: http://svn.freebsd.org/changeset/base/222537
> >
> > Log:
> > ?Fix apparent garbage in the message buffer.
> >
> > ?While we have had a fix in place (options PRINTF_BUFR_SIZE=128) to fix
> > ?scrambled console output, the message buffer and syslog were still getting
> > ?log messages one character at a time. ?While all of the characters still
> > ?made it into the log (courtesy of atomic operations), they were often
> > ?interleaved when there were multiple threads writing to the buffer at the
> > ?same time.
> 
> This seems to panic my box with "lock "msgbuf" 0xfe0127e0
> already initialized".
> 
> Unfortunately, though I booted with a fresh CURRENT this morning
> successfully, both /boot/kernel and /boot/kernel.old give this panic.
> To add insult to injury, when the kernel drops into the debugger, my
> keyboard input no longer works so I can't get a stack, etc.

Uh-oh!

> So:
> 
> 1) Is there anything else I can do to help debug this?
> 2) how can I resurrect this box without a reinstall?
> 
> I will try to repro on a virtual machine so I have a snapshot to come back to.

My guess is that this is an issue with the message buffer reinitialization
path.  lock_init() (called by mtx_init()) has an assert to make sure that
the lock is initialized, and that is just a flag check.

Since the spin lock is part of the message buffer structure, if it is held
over from a previous boot, the LO_INITIALIZED flag may still be set.

Try power cycling the machine.  If it is an issue with re-initialization,
that should clear the memory and allow you to boot.

My testing has been with VMs (under Xen), so the reinit path has probably
not been tested as fully as it should have been.  Sorry about that!

As for the debugger, that's another issue altogether.  It does work for me,
but then again if the spin lock initialization is broken for the message
buffer that may affect things.

Try a cold boot and see if that helps.  If so, I think we can probably just
bzero the mutex in msgbuf_reinit() and that will fix things.

Ken
-- 
Kenneth Merry
k...@freebsd.org
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

Re: svn commit: r222537 - in head/sys: kern sys

[Kenneth D. Merry]

On Tue, May 31, 2011 at 15:02:37 -0700, m...@freebsd.org wrote:
> On Tue, May 31, 2011 at 2:46 PM, Kenneth D. Merry  wrote:
> > On Tue, May 31, 2011 at 14:00:18 -0700, m...@freebsd.org wrote:
> >> On Tue, May 31, 2011 at 10:29 AM, Kenneth D. Merry  
> >> wrote:
> >> > Author: ken
> >> > Date: Tue May 31 17:29:58 2011
> >> > New Revision: 222537
> >> > URL: http://svn.freebsd.org/changeset/base/222537
> >> >
> >> > Log:
> >> > ?Fix apparent garbage in the message buffer.
> >> >
> >> > ?While we have had a fix in place (options PRINTF_BUFR_SIZE=128) to fix
> >> > ?scrambled console output, the message buffer and syslog were still 
> >> > getting
> >> > ?log messages one character at a time. ?While all of the characters still
> >> > ?made it into the log (courtesy of atomic operations), they were often
> >> > ?interleaved when there were multiple threads writing to the buffer at 
> >> > the
> >> > ?same time.
> >>
> >> This seems to panic my box with "lock "msgbuf" 0xfe0127e0
> >> already initialized".
> >>
> >> Unfortunately, though I booted with a fresh CURRENT this morning
> >> successfully, both /boot/kernel and /boot/kernel.old give this panic.
> >> To add insult to injury, when the kernel drops into the debugger, my
> >> keyboard input no longer works so I can't get a stack, etc.
> >
> > Uh-oh!
> >
> >> So:
> >>
> >> 1) Is there anything else I can do to help debug this?
> >> 2) how can I resurrect this box without a reinstall?
> >>
> >> I will try to repro on a virtual machine so I have a snapshot to come back 
> >> to.
> >
> > My guess is that this is an issue with the message buffer reinitialization
> > path. ?lock_init() (called by mtx_init()) has an assert to make sure that
> > the lock is initialized, and that is just a flag check.
> >
> > Since the spin lock is part of the message buffer structure, if it is held
> > over from a previous boot, the LO_INITIALIZED flag may still be set.
> >
> > Try power cycling the machine. ?If it is an issue with re-initialization,
> > that should clear the memory and allow you to boot.
> 
> Hmm, apparently my previous presses of the power button weren't long
> enough.  I let it sit off for 20 seconds and it boots okay now.

Okay, so it probably is the re-initialization code.  Can you try this patch
and see if it survives a warm boot?  I also changed the initialization
path, so we don't get tripped up by garbage left in memory.

Also, does the debugger work now that it has booted successfully?

Thanks,

Ken
-- 
Kenneth Merry
k...@freebsd.org
Index: subr_msgbuf.c
===
--- subr_msgbuf.c   (revision 222537)
+++ subr_msgbuf.c   (working copy)
@@ -61,6 +61,7 @@
mbp->msg_magic = MSG_MAGIC;
mbp->msg_lastpri = -1;
mbp->msg_needsnl = 0;
+   bzero(&mbp->msg_lock, sizeof(mbp->msg_lock));
mtx_init(&mbp->msg_lock, "msgbuf", NULL, MTX_SPIN);
 }
 
@@ -95,6 +96,7 @@
mbp->msg_lastpri = -1;
/* Assume that the old message buffer didn't end in a newline. */
mbp->msg_needsnl = 1;
+   bzero(&mbp->msg_lock, sizeof(mbp->msg_lock));
mtx_init(&mbp->msg_lock, "msgbuf", NULL, MTX_SPIN);
 }
 
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r222550 - head/sys/kern

[Kenneth D. Merry]

Author: ken
Date: Tue May 31 22:39:32 2011
New Revision: 222550
URL: http://svn.freebsd.org/changeset/base/222550

Log:
  Fix a bug introduced in revision 222537.
  
  In msgbuf_reinit() and msgbuf_init(), we weren't initializing the mutex.
  Depending on the contents of memory, the LO_INITIALIZED flag might be
  set on the mutex (either due to a warm reboot, and the message buffer
  remaining in place, or due to garbage in memory) and in that case, with
  INVARIANTS turned on, we would trigger an assertion that the mutex had
  already been initialized.
  
  Fix this by bzeroing the message buffer mutex for the _init() and _reinit()
  paths.
  
  Reported by:  mdf

Modified:
  head/sys/kern/subr_msgbuf.c

Modified: head/sys/kern/subr_msgbuf.c
==
--- head/sys/kern/subr_msgbuf.c Tue May 31 21:42:34 2011(r222549)
+++ head/sys/kern/subr_msgbuf.c Tue May 31 22:39:32 2011(r222550)
@@ -61,6 +61,7 @@ msgbuf_init(struct msgbuf *mbp, void *pt
mbp->msg_magic = MSG_MAGIC;
mbp->msg_lastpri = -1;
mbp->msg_needsnl = 0;
+   bzero(&mbp->msg_lock, sizeof(mbp->msg_lock));
mtx_init(&mbp->msg_lock, "msgbuf", NULL, MTX_SPIN);
 }
 
@@ -95,6 +96,7 @@ msgbuf_reinit(struct msgbuf *mbp, void *
mbp->msg_lastpri = -1;
/* Assume that the old message buffer didn't end in a newline. */
mbp->msg_needsnl = 1;
+   bzero(&mbp->msg_lock, sizeof(mbp->msg_lock));
mtx_init(&mbp->msg_lock, "msgbuf", NULL, MTX_SPIN);
 }
 
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

Re: svn commit: r222537 - in head/sys: kern sys

[Kenneth D. Merry]

On Tue, May 31, 2011 at 15:16:05 -0700, m...@freebsd.org wrote:
> On Tue, May 31, 2011 at 3:11 PM, Kenneth D. Merry  wrote:
> > On Tue, May 31, 2011 at 15:02:37 -0700, m...@freebsd.org wrote:
> >> On Tue, May 31, 2011 at 2:46 PM, Kenneth D. Merry  wrote:
> >> > On Tue, May 31, 2011 at 14:00:18 -0700, m...@freebsd.org wrote:
> >> >> On Tue, May 31, 2011 at 10:29 AM, Kenneth D. Merry  
> >> >> wrote:
> >> >> > Author: ken
> >> >> > Date: Tue May 31 17:29:58 2011
> >> >> > New Revision: 222537
> >> >> > URL: http://svn.freebsd.org/changeset/base/222537
> >> >> >
> >> >> > Log:
> >> >> > ?Fix apparent garbage in the message buffer.
> >> >> >
> >> >> > ?While we have had a fix in place (options PRINTF_BUFR_SIZE=128) to 
> >> >> > fix
> >> >> > ?scrambled console output, the message buffer and syslog were still 
> >> >> > getting
> >> >> > ?log messages one character at a time. ?While all of the characters 
> >> >> > still
> >> >> > ?made it into the log (courtesy of atomic operations), they were often
> >> >> > ?interleaved when there were multiple threads writing to the buffer 
> >> >> > at the
> >> >> > ?same time.
> >> >>
> >> >> This seems to panic my box with "lock "msgbuf" 0xfe0127e0
> >> >> already initialized".
> >> >>
> >> >> Unfortunately, though I booted with a fresh CURRENT this morning
> >> >> successfully, both /boot/kernel and /boot/kernel.old give this panic.
> >> >> To add insult to injury, when the kernel drops into the debugger, my
> >> >> keyboard input no longer works so I can't get a stack, etc.
> >> >
> >> > Uh-oh!
> >> >
> >> >> So:
> >> >>
> >> >> 1) Is there anything else I can do to help debug this?
> >> >> 2) how can I resurrect this box without a reinstall?
> >> >>
> >> >> I will try to repro on a virtual machine so I have a snapshot to come 
> >> >> back to.
> >> >
> >> > My guess is that this is an issue with the message buffer 
> >> > reinitialization
> >> > path. ?lock_init() (called by mtx_init()) has an assert to make sure that
> >> > the lock is initialized, and that is just a flag check.
> >> >
> >> > Since the spin lock is part of the message buffer structure, if it is 
> >> > held
> >> > over from a previous boot, the LO_INITIALIZED flag may still be set.
> >> >
> >> > Try power cycling the machine. ?If it is an issue with re-initialization,
> >> > that should clear the memory and allow you to boot.
> >>
> >> Hmm, apparently my previous presses of the power button weren't long
> >> enough. ?I let it sit off for 20 seconds and it boots okay now.
> >
> > Okay, so it probably is the re-initialization code. ?Can you try this patch
> > and see if it survives a warm boot? ?I also changed the initialization
> > path, so we don't get tripped up by garbage left in memory.
> 
> This patch survived a warm reboot (shutdown -r now).

Cool!  I just checked in the patch.

> > Also, does the debugger work now that it has booted successfully?
> 
> The debugger (or rather, my keyboard in the debugger) works on a
> successful boot.  I used sysctl debug.kdb.enter=1 to test it.

Ahh, glad things are working there too.

Let me know if you (or anyone else) run into any more issues.

Thanks for testing this!

Ken
-- 
Kenneth Merry
k...@freebsd.org
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

Re: svn commit: r222537 - in head/sys: kern sys

[Kenneth D. Merry]

On Wed, Jun 01, 2011 at 20:07:31 +1000, Bruce Evans wrote:
> On Wed, 1 Jun 2011, Andriy Gapon wrote:
> 
> >on 31/05/2011 20:29 Kenneth D. Merry said the following:
> >>+   mtx_init(&mbp->msg_lock, "msgbuf", NULL, MTX_SPIN);
> >
> >Sorry that I didn't gather myself together for a review before this change 
> >got
> >actually committed.
> >Do you see any reason not to make this spinlock recursive?
> 
> I see reasons why it must not exist.  The message buffer code cannot use
> any normal locking, and was carefully written to avoid doing so.
> 
> >I am a little bit worried about "exotic" situations like receiving an NMI 
> >in the
> >middle of printing and wanting to print in the NMI context, or similar 
> >things
> >that penetrate contexts with disabled interrupts - e.g. Machine Check 
> >Exception.
> >Also it's not clear to me if there won't any bigger damage in the 
> >situations
> >like those described above.
> >
> >P.S. I have been thinking about fixing the problem in a different fashion, 
> >via
> >reserving portions of dmesg buffer for a whole message using CAS:
> >http://lists.freebsd.org/pipermail/freebsd-hackers/2010-April/031535.html
> 
> Something like this might work.
> 
> PRINTF_BUFR size should not exist either, especially now that it is
> much more complicated and broken.  Here is some of my old mail about
> adding (necessarily not normal) locking to to printf.  No one replied,
> so I never finished this :-(.

That particular solution doesn't lock the normal kernel printf path, and so
won't fix what we're trying to fix.  (We've got lots of disks in each
system, and when there are various SES events and disks arriving and
departing, there are multiple kernel contexts doing printfs simultaneously,
and that results in scrambled dmesg output.)

I think the goals should be:

 - console output, syslog output, and dmesg output are not scrambled.
   (i.e. individual kernel printfs make it out atomically, or at least
   with a certain granularity, like PRINTF_BUFR_SIZE.)

 - Can be called by any kernel routine (i.e. doesn't sleep)

 - Offers a string at a time interface.

 - Does the right thing for log messages (priority, etc.)

It looks like we should add "does not use normal locking" to the list as
well.

As Justin mentioned, we started off down the path of using atomic
operations, and then figured out that wouldn't fully work.  Then we
discussed doing a per-CPU message buffer, with each message tagged with a
sequence number, and having the reader re-serialize the messages in the
right order.  The complexity started to get large enough that we decided
that using a spin lock would be a much easier approach.

cnputs() already uses a spinlock, so we're no worse off than before from a
locking standpoint.

We could perhaps make the message buffer mutex recursive and set the
MTX_NOWITNESS flag as well, that might make things a little better from a
side effect standpoint.

If we can come up with a solution that meets the above goals, Justin and I
would be happy to help implement it.

Ken
-- 
Kenneth Merry
k...@freebsd.org
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

Re: svn commit: r222537 - in head/sys: kern sys

[Kenneth D. Merry]

On Fri, Jun 03, 2011 at 20:47:00 +1000, Bruce Evans wrote:
> On Thu, 2 Jun 2011, Kenneth D. Merry wrote:
> 
> >On Wed, Jun 01, 2011 at 20:07:31 +1000, Bruce Evans wrote:
> >>On Wed, 1 Jun 2011, Andriy Gapon wrote:
> >>
> >>>on 31/05/2011 20:29 Kenneth D. Merry said the following:
> >>>>+ mtx_init(&mbp->msg_lock, "msgbuf", NULL, MTX_SPIN);
> >>>
> >>>Sorry that I didn't gather myself together for a review before this 
> >>>change
> >>>got
> >>>actually committed.
> >>>Do you see any reason not to make this spinlock recursive?
> 
> It does need to be recursive to avoid deadlock (as I described), but this
> prevents it actually working as a lock.  I think the new code needs it to
> work as a lock.

Yes, it does.

> >>I see reasons why it must not exist.  The message buffer code cannot use
> >>any normal locking, and was carefully written to avoid doing so.
> >>
> >>>I am a little bit worried about "exotic" situations like receiving an NMI
> >>>in the
> >>>middle of printing and wanting to print in the NMI context, or similar
> >>>things
> >>>that penetrate contexts with disabled interrupts - e.g. Machine Check
> >>>Exception.
> >>>Also it's not clear to me if there won't any bigger damage in the
> >>>situations
> >>>like those described above.
> >>>
> >>>P.S. I have been thinking about fixing the problem in a different 
> >>>fashion,
> >>>via
> >>>reserving portions of dmesg buffer for a whole message using CAS:
> >>>http://lists.freebsd.org/pipermail/freebsd-hackers/2010-April/031535.html
> >>
> >>Something like this might work.
> >>
> >>PRINTF_BUFR size should not exist either, especially now that it is
> >>much more complicated and broken.  Here is some of my old mail about
> >>adding (necessarily not normal) locking to to printf.  No one replied,
> >>so I never finished this :-(.
> >
> >That particular solution doesn't lock the normal kernel printf path, and so
> >won't fix what we're trying to fix.  (We've got lots of disks in each
> >system, and when there are various SES events and disks arriving and
> >departing, there are multiple kernel contexts doing printfs simultaneously,
> >and that results in scrambled dmesg output.)
> 
> But it does.  PRINTF_BUFR_SIZE (at least used to) only be applied to
> vprintf(), but my printf_lock() is applied to all calls to kvprintf()
> that set TOLOG.  (If I missed some, this takes 2 lines each to fix.)
> The calls in tprintf() and log() have comments saying that this is
> intentional "to serialize msgbuf", though I fear this might not be safe
> for at least tprintf(), since it may block unboundedly.  But even when
> it blocks unboundedly, nothing really bad happens: the output just
> becomes very slow due to everything waiting for the timeout until the
> blockage clears, and may be interleaved.  Recursion gives similar
> behaviour, except it takes unrecursing to clear the blockage.  I
> consider the very slow output to be a feature.  Unbounded blockages
> and recursion are bugs and the slow output makes them more obvious.

Ahh.  It wasn't 100% clear from looking at the patch which calls had the
lock applied.  In this case at least, it's skipping the lock when the
priority is -1.  Is this tprintf()?

%  /*
%   * Warn that a system table is full.
% @@ -198,5 +221,9 @@
%   pca.flags = flags;
%   va_start(ap, fmt);
% + if (pri != -1)
% + printf_lock();  /* To serialize msgbuf.  XXX: long-held
lock? */
%   kvprintf(fmt, putchar, &pca, 10, ap);
% + if (pri != -1)
% + printf_unlock();
%   va_end(ap);
%   if (sess != NULL)
%

> >I think the goals should be:
> >
> >- console output, syslog output, and dmesg output are not scrambled.
> >  (i.e. individual kernel printfs make it out atomically, or at least
> >  with a certain granularity, like PRINTF_BUFR_SIZE.)
> 
> Console output includes debugger output.  This must be able to break into
> any otherwise-atomic kernel printf.  Possibly similarly for NMIs (if they
> print.  They really shouldn't print, but they at least used to get this
> horribly wrong from the beginning on i386).  Similarly for certain (non
> maskable) traps.  These really shouldn't happen while we're printing, but
> if they do then they may need to panic and then print for things like
> machine check exceptions.
> 
> Debugger output is the easiest way of showing why prin

svn commit: r222804 - head/sys/kern

[Kenneth D. Merry]

Author: ken
Date: Tue Jun  7 05:04:37 2011
New Revision: 222804
URL: http://svn.freebsd.org/changeset/base/222804

Log:
  Set pca.p_bufr to NULL when we haven't allocated a buffer.
  
  Otherwise, p_bufr is set to garbage on the stack, and if that garbage
  happens to be non-NULL, and the TOLOG or TOCONS flag is set, putbuf()
  will get called and attempt to fill the non-existent buffer.
  
  This is really only relevant for tprintf() (and only when the priority is
  not -1), but set it in uprintf() and ttyprintf() for completeness.
  
  The next step, to avoid log buffer scrambling, would be to add the
  PRINTF_BUFR_SIZE code to tprintf(), but this should prevent panics.
  
  Submitted by: rmacklem
  Found by: pho

Modified:
  head/sys/kern/subr_prf.c

Modified: head/sys/kern/subr_prf.c
==
--- head/sys/kern/subr_prf.cTue Jun  7 04:38:33 2011(r222803)
+++ head/sys/kern/subr_prf.cTue Jun  7 05:04:37 2011(r222804)
@@ -163,6 +163,7 @@ uprintf(const char *fmt, ...)
goto out;
}
pca.flags = TOTTY;
+   pca.p_bufr = NULL;
va_start(ap, fmt);
tty_lock(pca.tty);
retval = kvprintf(fmt, putchar, &pca, 10, ap);
@@ -206,6 +207,7 @@ tprintf(struct proc *p, int pri, const c
pca.pri = pri;
pca.tty = tp;
pca.flags = flags;
+   pca.p_bufr = NULL;
va_start(ap, fmt);
if (pca.tty != NULL)
tty_lock(pca.tty);
@@ -234,6 +236,7 @@ ttyprintf(struct tty *tp, const char *fm
va_start(ap, fmt);
pca.tty = tp;
pca.flags = TOTTY;
+   pca.p_bufr = NULL;
retval = kvprintf(fmt, putchar, &pca, 10, ap);
va_end(ap);
return (retval);
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r222883 - head/sbin/rtsol

[Kenneth D. Merry]

Author: ken
Date: Wed Jun  8 21:59:07 2011
New Revision: 222883
URL: http://svn.freebsd.org/changeset/base/222883

Log:
  Add dump.c to the rtsol build.  It is needed now that sec2str is non-static
  and used in rtsold.c.

Modified:
  head/sbin/rtsol/Makefile

Modified: head/sbin/rtsol/Makefile
==
--- head/sbin/rtsol/MakefileWed Jun  8 21:58:19 2011(r222882)
+++ head/sbin/rtsol/MakefileWed Jun  8 21:59:07 2011(r222883)
@@ -19,7 +19,7 @@ SRCDIR=   ${.CURDIR}/../../usr.sbin/rtsold
 .PATH: ${SRCDIR}
 
 PROG=  rtsol
-SRCS=  rtsold.c rtsol.c if.c probe.c rtsock.c
+SRCS=  rtsold.c rtsol.c if.c probe.c dump.c rtsock.c
 NO_MAN=
 
 WARNS?=3
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r265236 - in head: share/man/man4 sys/amd64/conf sys/conf sys/dev/mpr sys/dev/mpr/mpi sys/i386/conf sys/ia64/conf sys/mips/conf sys/modules sys/modules/mpr sys/sparc64/conf

[Kenneth D. Merry]

Author: ken
Date: Fri May  2 20:25:09 2014
New Revision: 265236
URL: http://svnweb.freebsd.org/changeset/base/265236

Log:
  Bring in the mpr(4) driver for LSI's MPT3 12Gb SAS controllers.
  
  This is derived from the mps(4) driver, but it supports only the 12Gb
  IT and IR hardware including the SAS 3004, SAS 3008 and SAS 3108.
  
  Some notes about this driver:
   o The 12Gb hardware can do "FastPath" I/O, and that capability is included in
 this driver.
  
   o WarpDrive functionality has been removed, since it isn't supported in
 the 12Gb driver interface.
  
   o The Scatter/Gather list handling code is significantly different between
 the 6Gb and 12Gb hardware.  The 12Gb boards support IEEE Scatter/Gather
 lists.
  
  Thanks to LSI for developing and testing this driver for FreeBSD.
  
  share/man/man4/mpr.4:
mpr(4) man page.
  
  sys/dev/mpr/*:
mpr(4) driver files.
  
  sys/modules/Makefile,
  sys/modules/mpr/Makefile:
Add a module Makefile for the mpr(4) driver.
  
  sys/conf/files:
Add the mpr(4) driver.
  
  sys/amd64/conf/GENERIC,
  sys/i386/conf/GENERIC,
  sys/mips/conf/OCTEON1,
  sys/sparc64/conf/GENERIC:
Add the mpr(4) driver to all config files that currently
have the mps(4) driver.
  
  sys/ia64/conf/GENERIC:
Add the mps(4) and mpr(4) drivers to the ia64 GENERIC
config file.
  
  sys/i386/conf/XEN:
Exclude the mpr module from building here.
  
  Submitted by: Steve McConnell 
  MFC after:3 days
  Tested by:Chris Reeves 
  Sponsored by: LSI, Spectra Logic
  Relnotes: LSI 12Gb SAS driver mpr(4) added

Added:
  head/share/man/man4/mpr.4   (contents, props changed)
  head/sys/dev/mpr/
  head/sys/dev/mpr/mpi/
  head/sys/dev/mpr/mpi/mpi2.h   (contents, props changed)
  head/sys/dev/mpr/mpi/mpi2_cnfg.h   (contents, props changed)
  head/sys/dev/mpr/mpi/mpi2_hbd.h   (contents, props changed)
  head/sys/dev/mpr/mpi/mpi2_history.txt   (contents, props changed)
  head/sys/dev/mpr/mpi/mpi2_init.h   (contents, props changed)
  head/sys/dev/mpr/mpi/mpi2_ioc.h   (contents, props changed)
  head/sys/dev/mpr/mpi/mpi2_ra.h   (contents, props changed)
  head/sys/dev/mpr/mpi/mpi2_raid.h   (contents, props changed)
  head/sys/dev/mpr/mpi/mpi2_sas.h   (contents, props changed)
  head/sys/dev/mpr/mpi/mpi2_targ.h   (contents, props changed)
  head/sys/dev/mpr/mpi/mpi2_tool.h   (contents, props changed)
  head/sys/dev/mpr/mpi/mpi2_type.h   (contents, props changed)
  head/sys/dev/mpr/mpr.c   (contents, props changed)
  head/sys/dev/mpr/mpr_config.c   (contents, props changed)
  head/sys/dev/mpr/mpr_ioctl.h   (contents, props changed)
  head/sys/dev/mpr/mpr_mapping.c   (contents, props changed)
  head/sys/dev/mpr/mpr_mapping.h   (contents, props changed)
  head/sys/dev/mpr/mpr_pci.c   (contents, props changed)
  head/sys/dev/mpr/mpr_sas.c   (contents, props changed)
  head/sys/dev/mpr/mpr_sas.h   (contents, props changed)
  head/sys/dev/mpr/mpr_sas_lsi.c   (contents, props changed)
  head/sys/dev/mpr/mpr_table.c   (contents, props changed)
  head/sys/dev/mpr/mpr_table.h   (contents, props changed)
  head/sys/dev/mpr/mpr_user.c   (contents, props changed)
  head/sys/dev/mpr/mprvar.h   (contents, props changed)
  head/sys/modules/mpr/
  head/sys/modules/mpr/Makefile   (contents, props changed)
Modified:
  head/sys/amd64/conf/GENERIC
  head/sys/conf/files
  head/sys/i386/conf/GENERIC
  head/sys/i386/conf/XEN
  head/sys/ia64/conf/GENERIC
  head/sys/mips/conf/OCTEON1
  head/sys/modules/Makefile
  head/sys/sparc64/conf/GENERIC

Added: head/share/man/man4/mpr.4
==
--- /dev/null   00:00:00 1970   (empty, because file is newly added)
+++ head/share/man/man4/mpr.4   Fri May  2 20:25:09 2014(r265236)
@@ -0,0 +1,226 @@
+.\"
+.\" Copyright (c) 2010 Spectra Logic Corporation
+.\" Copyright (c) 2014 LSI Corp
+.\" All rights reserved.
+.\"
+.\" Redistribution and use in source and binary forms, with or without
+.\" modification, are permitted provided that the following conditions
+.\" are met:
+.\" 1. Redistributions of source code must retain the above copyright
+.\"notice, this list of conditions, and the following disclaimer,
+.\"without modification.
+.\" 2. Redistributions in binary form must reproduce at minimum a disclaimer
+.\"substantially similar to the "NO WARRANTY" disclaimer below
+.\"("Disclaimer") and any redistribution must be conditioned upon
+.\"including a substantially similar Disclaimer requirement for further
+.\"binary redistribution.
+.\"
+.\" NO WARRANTY
+.\" THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+.\" "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+.\" LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR
+.\" A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+.\" HOLDERS OR CONTRIBUTORS BE LIABLE FOR SPECIAL, EXEMPLAR

svn commit: r265237 - head/share/man/man4

[Kenneth D. Merry]

Author: ken
Date: Fri May  2 20:36:20 2014
New Revision: 265237
URL: http://svnweb.freebsd.org/changeset/base/265237

Log:
  Add the mpr(4) man page to the man4 Makefile.
  
  This should have been included in r265236.
  
  Submitted by: Steve McConnell 
  MFC after:3 days
  Sponsored by: LSI, Spectra Logic

Modified:
  head/share/man/man4/Makefile

Modified: head/share/man/man4/Makefile
==
--- head/share/man/man4/MakefileFri May  2 20:25:09 2014
(r265236)
+++ head/share/man/man4/MakefileFri May  2 20:36:20 2014
(r265237)
@@ -262,6 +262,7 @@ MAN=aac.4 \
mod_cc.4 \
mos.4 \
mouse.4 \
+   mpr.4 \
mps.4 \
mpt.4 \
msk.4 \
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r265386 - head/sys/dev/mpr

[Kenneth D. Merry]

Author: ken
Date: Mon May  5 19:53:03 2014
New Revision: 265386
URL: http://svnweb.freebsd.org/changeset/base/265386

Log:
  Adjust #if statements inside mprsas_send_smpcmd() to more accurately
  reflect when unmapped I/O support was added.
  
  For FreeBSD 10, it arrived just prior to __FreeBSD_version 128.
  For FreeBSD 9, it arrived just prior to __FreeBSD_version 902001.
  
  Also, fix compiler warnings in mprsas_send_smpcmd() that happen in the
  i386 PAE build for non-unmapped I/O builds.  These were fixed in mps(4)
  in revision 241145, but didn't make it into the mpr(4) driver.  This
  change should only affect FreeBSD versions outside the above revisions,
  and thus doesn't affect head.
  
  MFC after:3 days
  Sponsored by: Spectra Logic Corporation

Modified:
  head/sys/dev/mpr/mpr_sas.c

Modified: head/sys/dev/mpr/mpr_sas.c
==
--- head/sys/dev/mpr/mpr_sas.c  Mon May  5 19:38:29 2014(r265385)
+++ head/sys/dev/mpr/mpr_sas.c  Mon May  5 19:53:03 2014(r265386)
@@ -2501,7 +2501,8 @@ mprsas_send_smpcmd(struct mprsas_softc *
sg = NULL;
error = 0;
 
-#if __FreeBSD_version >= 129
+#if (__FreeBSD_version >= 128) || \
+((__FreeBSD_version >= 902001) && (__FreeBSD_version < 100))
switch (ccb->ccb_h.flags & CAM_DATA_MASK) {
case CAM_DATA_PADDR:
case CAM_DATA_SG_PADDR:
@@ -2561,7 +2562,7 @@ mprsas_send_smpcmd(struct mprsas_softc *
xpt_done(ccb);
return;
}
-#else //__FreeBSD_version < 129
+#else /* __FreeBSD_version < 128 */
/*
 * XXX We don't yet support physical addresses here.
 */
@@ -2604,7 +2605,7 @@ mprsas_send_smpcmd(struct mprsas_softc *
bus_dma_segment_t *req_sg;
 
req_sg = (bus_dma_segment_t *)ccb->smpio.smp_request;
-   request = (uint8_t *)req_sg[0].ds_addr;
+   request = (uint8_t *)(uintptr_t)req_sg[0].ds_addr;
} else
request = ccb->smpio.smp_request;
 
@@ -2612,14 +2613,14 @@ mprsas_send_smpcmd(struct mprsas_softc *
bus_dma_segment_t *rsp_sg;
 
rsp_sg = (bus_dma_segment_t *)ccb->smpio.smp_response;
-   response = (uint8_t *)rsp_sg[0].ds_addr;
+   response = (uint8_t *)(uintptr_t)rsp_sg[0].ds_addr;
} else
response = ccb->smpio.smp_response;
} else {
request = ccb->smpio.smp_request;
response = ccb->smpio.smp_response;
}
-#endif //__FreeBSD_version >= 129
+#endif /* __FreeBSD_version < 128 */
 
cm = mpr_alloc_command(sc);
if (cm == NULL) {
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r265388 - in stable/10: share/man/man4 sys/amd64/conf sys/conf sys/dev/mpr sys/i386/conf sys/ia64/conf sys/mips/conf sys/modules sys/modules/mpr sys/sparc64/conf

[Kenneth D. Merry]

Author: ken
Date: Mon May  5 20:35:35 2014
New Revision: 265388
URL: http://svnweb.freebsd.org/changeset/base/265388

Log:
  MFC the mpr(4) driver for LSI's 12Gb SAS cards.
  
  This includes r265236, r265237, r265241 and r265261:
  

r265236 | ken | 2014-05-02 14:25:09 -0600 (Fri, 02 May 2014) | 51 lines
  
Bring in the mpr(4) driver for LSI's MPT3 12Gb SAS controllers.
  
This is derived from the mps(4) driver, but it supports only the 12Gb
IT and IR hardware including the SAS 3004, SAS 3008 and SAS 3108.
  
Some notes about this driver:
 o The 12Gb hardware can do "FastPath" I/O, and that capability is included 
in
   this driver.
  
 o WarpDrive functionality has been removed, since it isn't supported in
   the 12Gb driver interface.
  
 o The Scatter/Gather list handling code is significantly different between
   the 6Gb and 12Gb hardware.  The 12Gb boards support IEEE Scatter/Gather
   lists.
  
Thanks to LSI for developing and testing this driver for FreeBSD.
  
share/man/man4/mpr.4:
mpr(4) man page.
  
sys/dev/mpr/*:
mpr(4) driver files.
  
sys/modules/Makefile,
sys/modules/mpr/Makefile:
Add a module Makefile for the mpr(4) driver.
  
sys/conf/files:
Add the mpr(4) driver.
  
sys/amd64/conf/GENERIC,
sys/i386/conf/GENERIC,
sys/mips/conf/OCTEON1,
sys/sparc64/conf/GENERIC:
Add the mpr(4) driver to all config files that currently
have the mps(4) driver.
  
sys/ia64/conf/GENERIC:
Add the mps(4) and mpr(4) drivers to the ia64 GENERIC
config file.
  
sys/i386/conf/XEN:
Exclude the mpr module from building here.
  
Submitted by:   Steve McConnell 
Tested by:  Chris Reeves 
Sponsored by:   LSI, Spectra Logic
Relnotes:   LSI 12Gb SAS driver mpr(4) added
  


r265237 | ken | 2014-05-02 14:36:20 -0600 (Fri, 02 May 2014) | 8 lines
  
Add the mpr(4) man page to the man4 Makefile.
  
This should have been included in r265236.
  
Submitted by:   Steve McConnell 
MFC after:  3 days
Sponsored by:   LSI, Spectra Logic
  


r265241 | brueffer | 2014-05-02 15:14:28 -0600 (Fri, 02 May 2014) | 2 lines
  
Use our standard SYNOPSIS wording; perform some cleanup while here.
  


r265261 | brueffer | 2014-05-03 05:15:28 -0600 (Sat, 03 May 2014) | 2 lines
  
Add a missing colon.
  

  
  Submitted by: Steve McConnell 
  Tested by:Chris Reeves 
  Sponsored by: LSI, Spectra Logic
  Relnotes: LSI 12Gb SAS driver mpr(4) added

Added:
  stable/10/share/man/man4/mpr.4
 - copied, changed from r265236, head/share/man/man4/mpr.4
  stable/10/sys/dev/mpr/
 - copied from r265236, head/sys/dev/mpr/
  stable/10/sys/modules/mpr/
 - copied from r265236, head/sys/modules/mpr/
Modified:
  stable/10/share/man/man4/Makefile
  stable/10/sys/amd64/conf/GENERIC
  stable/10/sys/conf/files
  stable/10/sys/i386/conf/GENERIC
  stable/10/sys/i386/conf/XEN
  stable/10/sys/ia64/conf/GENERIC
  stable/10/sys/mips/conf/OCTEON1
  stable/10/sys/modules/Makefile
  stable/10/sys/sparc64/conf/GENERIC
Directory Properties:
  stable/10/   (props changed)

Modified: stable/10/share/man/man4/Makefile
==
--- stable/10/share/man/man4/Makefile   Mon May  5 20:13:25 2014
(r265387)
+++ stable/10/share/man/man4/Makefile   Mon May  5 20:35:35 2014
(r265388)
@@ -259,6 +259,7 @@ MAN=aac.4 \
mod_cc.4 \
mos.4 \
mouse.4 \
+   mpr.4 \
mps.4 \
mpt.4 \
msk.4 \

Copied and modified: stable/10/share/man/man4/mpr.4 (from r265236, 
head/share/man/man4/mpr.4)
==
--- head/share/man/man4/mpr.4   Fri May  2 20:25:09 2014(r265236, copy 
source)
+++ stable/10/share/man/man4/mpr.4  Mon May  5 20:35:35 2014
(r265388)
@@ -36,7 +36,7 @@
 .\" $Id$
 .\" $FreeBSD$
 .\"
-.Dd Apr 28, 2014
+.Dd May 2, 2014
 .Dt MPR 4
 .Os
 .Sh NAME
@@ -51,7 +51,8 @@ kernel configuration file:
 .Cd "device mpr"
 .Ed
 .Pp
-Or, to load the driver as a module at boot, place the following line in
+Alternatively, to load the driver as a
+module at boot time, place the following line in
 .Xr loader.conf 5 :
 .Bd -literal -offset i

svn commit: r265424 - head/sys/dev/mpr

[Kenneth D. Merry]

Author: ken
Date: Tue May  6 06:18:43 2014
New Revision: 265424
URL: http://svnweb.freebsd.org/changeset/base/265424

Log:
  Fix a problem with async notifications in the mpr(4) driver.
  
  This problem only occurs on versions of FreeBSD prior to the recent CAM
  locking changes.  (i.e. stable/9 and older versions of stable/10)  This
  change should be a no-op for head and stable/10.
  
  If a path isn't specified, xpt_register_async() will create a fully
  wildcarded path and acquire a lock (the XPT lock in older versions,
  and via xpt_path_lock() in newer versions) to call xpt_action() for the
  XPT_SASYNC_CB CCB.  It will then drop the lock and if the requested event
  includes AC_FOUND_DEVICE or AC_PATH_REGISTERED, it will get the caller up
  to date with any device arrivals or path registrations.
  
  The issue is that before the locking changes, each SIM lock would get
  acquired in turn during the EDT tree traversal process.  If a path is
  specified for xpt_register_async(), it won't acquire and drop its own lock,
  but instead expects the caller to hold its own SIM lock.  That works for
  the first part of xpt_register_async(), but causes a recursive lock
  acquisition once the EDT traversal happens and it comes to the SIM in
  question.  And it isn't possible to call xpt_action() without holding a SIM
  lock.
  
  The locking changes fix this by using the XPT topology lock for EDT
  traversal, so it is no longer an issue to hold the SIM lock while calling
  xpt_register_async().
  
  The solution for FreeBSD versions before the locking changes is to request
  notification of all device arrivals (so we pass a NULL path into
  xpt_register_async()) and then filter out the arrivals that are not ours.
  
  MFC After:3 days
  Sponsored by: Spectra Logic Corporation

Modified:
  head/sys/dev/mpr/mpr_sas.c

Modified: head/sys/dev/mpr/mpr_sas.c
==
--- head/sys/dev/mpr/mpr_sas.c  Tue May  6 04:22:37 2014(r265423)
+++ head/sys/dev/mpr/mpr_sas.c  Tue May  6 06:18:43 2014(r265424)
@@ -813,8 +813,49 @@ mpr_attach_sas(struct mpr_softc *sc)
 #else
event = AC_FOUND_DEVICE;
 #endif
+
+   /*
+* Prior to the CAM locking improvements, we can't call
+* xpt_register_async() with a particular path specified.
+*
+* If a path isn't specified, xpt_register_async() will
+* generate a wildcard path and acquire the XPT lock while
+* it calls xpt_action() to execute the XPT_SASYNC_CB CCB.
+* It will then drop the XPT lock once that is done.
+* 
+* If a path is specified for xpt_register_async(), it will
+* not acquire and drop the XPT lock around the call to
+* xpt_action().  xpt_action() asserts that the caller
+* holds the SIM lock, so the SIM lock has to be held when
+* calling xpt_register_async() when the path is specified.
+* 
+* But xpt_register_async calls xpt_for_all_devices(),
+* which calls xptbustraverse(), which will acquire each
+* SIM lock.  When it traverses our particular bus, it will
+* necessarily acquire the SIM lock, which will lead to a
+* recursive lock acquisition.
+* 
+* The CAM locking changes fix this problem by acquiring
+* the XPT topology lock around bus traversal in
+* xptbustraverse(), so the caller can hold the SIM lock
+* and it does not cause a recursive lock acquisition.
+*
+* These __FreeBSD_version values are approximate, especially
+* for stable/10, which is two months later than the actual
+* change.
+*/
+
+#if (__FreeBSD_version < 1000703) || \
+((__FreeBSD_version >= 110) && (__FreeBSD_version < 112))
+   mpr_unlock(sc);
+   status = xpt_register_async(event, mprsas_async, sc,
+   NULL);
+   mpr_lock(sc);
+#else
status = xpt_register_async(event, mprsas_async, sc,
sassc->path);
+#endif
+
if (status != CAM_REQ_CMP) {
mpr_dprint(sc, MPR_ERROR,
"Error %#x registering async handler for "
@@ -2988,6 +3029,18 @@ mprsas_async(void *callback_arg, uint32_
break;
 
/*
+* See the comment in mpr_attach_sas() for a detailed
+* explanation.  In these versions of FreeBSD we register
+* for all events and filter out the events that don't
+* apply to us.
+*/
+#if (__FreeBSD_version < 1000703) 

svn commit: r265473 - head/sys/dev/mpr

[Kenneth D. Merry]

Author: ken
Date: Tue May  6 22:13:38 2014
New Revision: 265473
URL: http://svnweb.freebsd.org/changeset/base/265473

Log:
  Change the device name for mpr(4) from /dev/mpr_N to /dev/mprN.
  
  This is more consistent with the existing mps(4) behavior.
  
  Reviewed by:  Steve McConnell 
  MFC after:3 days

Modified:
  head/sys/dev/mpr/mpr_user.c

Modified: head/sys/dev/mpr/mpr_user.c
==
--- head/sys/dev/mpr/mpr_user.c Tue May  6 22:06:39 2014(r265472)
+++ head/sys/dev/mpr/mpr_user.c Tue May  6 22:13:38 2014(r265473)
@@ -212,7 +212,7 @@ mpr_attach_user(struct mpr_softc *sc)
 
unit = device_get_unit(sc->mpr_dev);
sc->mpr_cdev = make_dev(&mpr_cdevsw, unit, UID_ROOT, GID_OPERATOR,
-   0640, "mpr_%d", unit);
+   0640, "mpr%d", unit);
if (sc->mpr_cdev == NULL) {
return (ENOMEM);
}
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r265484 - head/sys/dev/mpr

[Kenneth D. Merry]

Author: ken
Date: Wed May  7 05:11:16 2014
New Revision: 265484
URL: http://svnweb.freebsd.org/changeset/base/265484

Log:
  Remove some debugging code.
  
  Submitted by: Steve McConnell 
  MFC after:3 days

Modified:
  head/sys/dev/mpr/mpr.c

Modified: head/sys/dev/mpr/mpr.c
==
--- head/sys/dev/mpr/mpr.c  Wed May  7 05:06:55 2014(r265483)
+++ head/sys/dev/mpr/mpr.c  Wed May  7 05:11:16 2014(r265484)
@@ -2182,8 +2182,6 @@ mpr_add_chain(struct mpr_command *cm, in
 * code other than 0.
 */
if (cm->cm_flags & MPR_CM_FLAGS_SGE_SIMPLE) {
-//SLM-test
-printf("Trying to add a chain element to an MPI SGL\n");
mpr_dprint(sc, MPR_ERROR, "A chain element cannot be added to "
"an MPI SGL.\n");
return(ENOBUFS);
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r265485 - head/sys/dev/mpr

[Kenneth D. Merry]

Author: ken
Date: Wed May  7 05:14:48 2014
New Revision: 265485
URL: http://svnweb.freebsd.org/changeset/base/265485

Log:
  Hold the SIM lock when calling xpt_create_path() and xpt_action() in
  mprsas_SSU_to_SATA_devices().
  
  This fixes an assertion on shutdown with INVARIANTS enabled with SATA
  drives present on an IR firmware controller.
  
  Reviewed by:  Steve McConnell .
  MFC after:3 days

Modified:
  head/sys/dev/mpr/mpr_sas_lsi.c

Modified: head/sys/dev/mpr/mpr_sas_lsi.c
==
--- head/sys/dev/mpr/mpr_sas_lsi.c  Wed May  7 05:11:16 2014
(r265484)
+++ head/sys/dev/mpr/mpr_sas_lsi.c  Wed May  7 05:14:48 2014
(r265485)
@@ -1026,6 +1026,8 @@ mprsas_SSU_to_SATA_devices(struct mpr_so
char path_str[64];
struct timeval cur_time, start_time;
 
+   mpr_lock(sc);
+
/*
 * For each LUN of each target, issue a StartStopUnit command to stop
 * the device.
@@ -1041,6 +1043,7 @@ mprsas_SSU_to_SATA_devices(struct mpr_so
SLIST_FOREACH(lun, &target->luns, lun_link) {
ccb = xpt_alloc_ccb_nowait();
if (ccb == NULL) {
+   mpr_unlock(sc);
mpr_dprint(sc, MPR_FAULT, "Unable to alloc "
"CCB to stop unit.\n");
return;
@@ -1057,6 +1060,7 @@ mprsas_SSU_to_SATA_devices(struct mpr_so
mpr_dprint(sc, MPR_FAULT, "Unable to "
"create LUN path to stop unit.\n");
xpt_free_ccb(ccb);
+   mpr_unlock(sc);
return;
}
xpt_path_string(ccb->ccb_h.path, path_str,
@@ -1092,6 +1096,8 @@ mprsas_SSU_to_SATA_devices(struct mpr_so
}
}
 
+   mpr_unlock(sc);
+
/*
 * Wait until all of the SSU commands have completed or time has
 * expired (60 seconds).  pause for 100ms each time through.  If any
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r265584 - svnadmin/conf

[Kenneth D. Merry]

Author: ken
Date: Wed May  7 18:22:30 2014
New Revision: 265584
URL: http://svnweb.freebsd.org/changeset/base/265584

Log:
  Add Steve McConnell (slm) as a src committer.
  
  Steve works for LSI (which is now Avago) and will be maintaing drivers for
  LSI's SAS and RAID products (mps(4), mpr(4), mrsas(4)).
  
  scottl, ambrisko and ken will be his mentors.
  
  Approved by:  core

Modified:
  svnadmin/conf/access
  svnadmin/conf/mentors

Modified: svnadmin/conf/access
==
--- svnadmin/conf/accessWed May  7 18:21:38 2014(r265583)
+++ svnadmin/conf/accessWed May  7 18:22:30 2014(r265584)
@@ -197,6 +197,7 @@ se
 silby
 simon
 sjg
+slm
 smh
 sobomax
 stas

Modified: svnadmin/conf/mentors
==
--- svnadmin/conf/mentors   Wed May  7 18:21:38 2014(r265583)
+++ svnadmin/conf/mentors   Wed May  7 18:22:30 2014(r265584)
@@ -28,6 +28,7 @@ monthadar adrian
 odeds  jhb Co-mentor: alfred
 peterj jhb Co-mentor: grog
 royger gibbs
+slmken Co-mentor: scottl, ambrisko
 snbdwmalone
 tychon neelCo-mentor: grehan
 versus gavin   Co-mentor: fjoe
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r252214 - in stable/9: bin/chio sys/cam/scsi sys/sys

[Kenneth D. Merry]

Author: ken
Date: Tue Jun 25 21:43:49 2013
New Revision: 252214
URL: http://svnweb.freebsd.org/changeset/base/252214

Log:
  MFC: 249658, 249701
  
  Update chio(1) and ch(4) to support reporting element designators.
  
  This allows mapping a tape drive in a changer (as reported by
  'chio status') to a sa(4) driver instance by comparing the
  serial numbers.
  
  The designators can be ASCII (which is printed out directly), binary
  (which is printed in hex format) or UTF-8, which is printed in either
  native UTF-8 format if the terminal can support it, or in %XX notation
  for non-ASCII characters.  Thanks to Hiroki Sato  for the
  explanation and example UTF-8 printing code.
  
  chio.h:   Modify the changer_element_status structure to add new
fields and definitions from the SMC3r16 spec.
  
Rename the original CHIOGSTATUS ioctl to OCHIOGTATUS and
define a new CHIOGSTATUS ioctl.
  
Clean up some tab/space issues.
  
  chio.c:   For the 'status' subcommand, print the designator field
if it is supplied by a device.
  
  scsi_ch.h:Add new flags for DVCID and CURDATA to the READ
ELEMENT STATUS command structure.
  
Add a read_element_status_device_id structure
for the data fields in the new standard. Add new
unions, dt_or_obsolete and voltage_devid, to hold
and address data from either SCSI-2 or newer devices.
  
  scsi_ch.c:Implement support for fetching device IDs with READ
ELEMENT STATUS data.
  
Add new arguments to scsi_read_element_status() to
allow the user to request the DVCID and CURDATA bits.
This isn't compiled into libcam (it's only an internal
kernel interface), so we don't need any special
handling for the API change.
  
If the user issues the new CHIOGSTATUS ioctl, copy all of
the available element status data out.  If he issues the
OCHIOGSTATUS ioctl, we don't copy the new fields in the
structure.
  
Fix a bug in chopen() that would result in the peripheral
never getting unheld if chgetparams() failed.
  
  Sponsored by: Spectra Logic
  Submitted by: Po-Li Soong

Modified:
  stable/9/bin/chio/chio.c
  stable/9/sys/cam/scsi/scsi_ch.c
  stable/9/sys/cam/scsi/scsi_ch.h
  stable/9/sys/sys/chio.h
Directory Properties:
  stable/9/   (props changed)
  stable/9/bin/   (props changed)
  stable/9/sys/   (props changed)
  stable/9/sys/sys/   (props changed)

Modified: stable/9/bin/chio/chio.c
==
--- stable/9/bin/chio/chio.cTue Jun 25 21:34:21 2013(r252213)
+++ stable/9/bin/chio/chio.cTue Jun 25 21:43:49 2013(r252214)
@@ -54,6 +54,8 @@ __FBSDID("$FreeBSD$");
 #include 
 #include 
 #include 
+#include 
+#include 
 
 #include "defs.h"
 #include "pathnames.h"
@@ -81,6 +83,7 @@ staticint do_status(const char *, int, 
 static int do_ielem(const char *, int, char **);
 static int do_return(const char *, int, char **);
 static int do_voltag(const char *, int, char **);
+static void print_designator(const char *, u_int8_t, u_int8_t);
 
 #ifndef CHET_VT
 #defineCHET_VT 10  /* Completely Arbitrary 
*/
@@ -723,6 +726,10 @@ do_status(const char *cname, int argc, c
putchar('?');
putchar('>');
}
+   if (ces->ces_designator_length > 0)
+   print_designator(ces->ces_designator,
+ces->ces_code_set,
+ces->ces_designator_length);
putchar('\n');
}
 
@@ -1177,3 +1184,66 @@ usage(void)
"arg1 arg2 [arg3 [...]]\n", getprogname());
exit(1);
 }
+
+#defineUTF8CODESET "UTF-8"
+
+static void
+print_designator(const char *designator, u_int8_t code_set,
+u_int8_t designator_length)
+{
+   printf(" serial number: <");
+   switch (code_set) {
+   case CES_CODE_SET_ASCII: {
+   /*
+* The driver insures that the string is always NUL terminated.
+*/
+   printf("%s", designator);
+   break;
+   }
+   case CES_CODE_SET_UTF_8: {
+   char *cs_native;
+
+   setlocale(LC_ALL, "");
+   cs_native = nl_langinfo(CODESET);
+
+   /* See if we can natively print UTF-8 */
+   if (strcmp(cs_native, UTF8CODESET) == 0)
+   cs_native = NULL;
+
+   if (cs_native == NULL) {
+   /* We can natively print UTF-8, so use printf. */
+

Re: svn commit: r252214 - in stable/9: bin/chio sys/cam/scsi sys/sys

[Kenneth D. Merry]

On Tue, Jul 02, 2013 at 07:53:33 +0200, Andre Albsmeier wrote:
> On Tue, 25-Jun-2013 at 23:43:50 +0200, Kenneth D. Merry wrote:
> > Author: ken
> > Date: Tue Jun 25 21:43:49 2013
> > New Revision: 252214
> > URL: http://svnweb.freebsd.org/changeset/base/252214
> 
> Hi Ken,
> 
> > Log:
> >   MFC: 249658, 249701
> > 
> >   Update chio(1) and ch(4) to support reporting element designators.
> > 
> >   This allows mapping a tape drive in a changer (as reported by
> >   'chio status') to a sa(4) driver instance by comparing the
> >   serial numbers.
> > 
> >   The designators can be ASCII (which is printed out directly), binary
> >   (which is printed in hex format) or UTF-8, which is printed in either
> >   native UTF-8 format if the terminal can support it, or in %XX notation
> >   for non-ASCII characters.  Thanks to Hiroki Sato  for the
> >   explanation and example UTF-8 printing code.
> > 
> >   chio.h:   Modify the changer_element_status structure to add 
> > new
> > fields and definitions from the SMC3r16 spec.
> > 
> > Rename the original CHIOGSTATUS ioctl to OCHIOGTATUS and
> > define a new CHIOGSTATUS ioctl.
> > 
> > Clean up some tab/space issues.
> > 
> >   chio.c:   For the 'status' subcommand, print the designator field
> > if it is supplied by a device.
> > 
> >   scsi_ch.h:Add new flags for DVCID and CURDATA to the READ
> > ELEMENT STATUS command structure.
> > 
> > Add a read_element_status_device_id structure
> > for the data fields in the new standard. Add new
> > unions, dt_or_obsolete and voltage_devid, to hold
> > and address data from either SCSI-2 or newer devices.
> > 
> >   scsi_ch.c:Implement support for fetching device IDs with READ
> > ELEMENT STATUS data.
> > 
> > Add new arguments to scsi_read_element_status() to
> > allow the user to request the DVCID and CURDATA bits.
> 
> This broke "chio status" when talking to my QUALSTAR TLS-8211 library:
> 
> Jul  2 07:08:22  server kernel: (ch0:ahd3:0:1:0): READ ELEMENT 
> STATUS. CDB: b8 10 fd e8 00 01 03 00 04 00 00 00
> Jul  2 07:08:22  server kernel: (ch0:ahd3:0:1:0): CAM status: SCSI 
> Status Error
> Jul  2 07:08:22  server kernel: (ch0:ahd3:0:1:0): SCSI status: 
> Check Condition
> Jul  2 07:08:22  server kernel: (ch0:ahd3:0:1:0): SCSI sense: 
> ILLEGAL REQUEST asc:24,0 (Invalid field in CDB)
> Jul  2 07:08:22  server kernel: (ch0:ahd3:0:1:0): Command byte 6 
> is invalid
> Jul  2 07:08:22  server kernel: (ch0:ahd3:0:1:0): Error 22, 
> Unretryable error
> 
> scsi_cmd->flags (Byte 6) must be zero for this library so I had to use
> this workaround to make it work again until a better solution is available:
> 
> --- sys/cam/scsi/scsi_ch.c.ORI2013-06-26 13:38:54.0 +0200
> +++ sys/cam/scsi/scsi_ch.c2013-07-02 07:42:24.0 +0200
> @@ -1245,8 +1245,8 @@
>/* tag_action */ MSG_SIMPLE_Q_TAG,
>/* voltag */ want_voltags,
>/* sea */ softc->sc_firsts[chet],
> -  /* dvcid */ 1,
> -  /* curdata */ 1,
> +  /* dvcid */ 0,
> +  /* curdata */ 0,
>/* count */ 1,
>/* data_ptr */ data,
>/* dxfer_len */ 1024,
> @@ -1284,8 +1284,8 @@
>/* voltag */ want_voltags,
>/* sea */ softc->sc_firsts[chet]
>+ cesr->cesr_element_base,
> -  /* dvcid */ 1,
> -  /* curdata */ 1,
> +  /* dvcid */ 0,
> +  /* curdata */ 0,
>/* count */ cesr->cesr_element_count,
>/* data_ptr */ data,
>/* dxfer_len */ size,
> 
>   -Andre

Oops, sorry.

We need to check the SCSI version to see whether to set those bits, and
also fall back in case it doesn't work.

I am on vacation and have very spotty net access.  I can't do anything
about it until I get back next week.

Justin and Alan (CCed) can work on the fix, though.

Ken
-- 
Kenneth Merry
k...@freebsd.org
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

Re: svn commit: r252214 - in stable/9: bin/chio sys/cam/scsi sys/sys

[Kenneth D. Merry]

On Wed, Jul 03, 2013 at 07:43:49 +0200, Andre Albsmeier wrote:
> On Wed, 03-Jul-2013 at 07:22:36 +0200, Kenneth D. Merry wrote:
> > On Tue, Jul 02, 2013 at 07:53:33 +0200, Andre Albsmeier wrote:
> > > On Tue, 25-Jun-2013 at 23:43:50 +0200, Kenneth D. Merry wrote:
> > > > Author: ken
> > > > Date: Tue Jun 25 21:43:49 2013
> > > > New Revision: 252214
> > > > URL: http://svnweb.freebsd.org/changeset/base/252214
> > > 
> > > Hi Ken,
> > > 
> > > > Log:
> > > >   MFC: 249658, 249701
> > > > 
> > > >   Update chio(1) and ch(4) to support reporting element designators.
> > > > 
> > > >   This allows mapping a tape drive in a changer (as reported by
> > > >   'chio status') to a sa(4) driver instance by comparing the
> > > >   serial numbers.
> > > > 
> > > >   The designators can be ASCII (which is printed out directly), binary
> > > >   (which is printed in hex format) or UTF-8, which is printed in either
> > > >   native UTF-8 format if the terminal can support it, or in %XX notation
> > > >   for non-ASCII characters.  Thanks to Hiroki Sato  for the
> > > >   explanation and example UTF-8 printing code.
> > > > 
> > > >   chio.h:   Modify the changer_element_status structure to 
> > > > add new
> > > > fields and definitions from the SMC3r16 spec.
> > > > 
> > > > Rename the original CHIOGSTATUS ioctl to OCHIOGTATUS and
> > > > define a new CHIOGSTATUS ioctl.
> > > > 
> > > > Clean up some tab/space issues.
> > > > 
> > > >   chio.c:   For the 'status' subcommand, print the designator field
> > > > if it is supplied by a device.
> > > > 
> > > >   scsi_ch.h:Add new flags for DVCID and CURDATA to the READ
> > > > ELEMENT STATUS command structure.
> > > > 
> > > > Add a read_element_status_device_id structure
> > > > for the data fields in the new standard. Add new
> > > > unions, dt_or_obsolete and voltage_devid, to hold
> > > > and address data from either SCSI-2 or newer devices.
> > > > 
> > > >   scsi_ch.c:Implement support for fetching device IDs with READ
> > > > ELEMENT STATUS data.
> > > > 
> > > > Add new arguments to scsi_read_element_status() to
> > > > allow the user to request the DVCID and CURDATA bits.
> > > 
> > > This broke "chio status" when talking to my QUALSTAR TLS-8211 library:
> > > 
> > > Jul  2 07:08:22  server kernel: (ch0:ahd3:0:1:0): READ ELEMENT 
> > > STATUS. CDB: b8 10 fd e8 00 01 03 00 04 00 00 00
> > > Jul  2 07:08:22  server kernel: (ch0:ahd3:0:1:0): CAM status: 
> > > SCSI Status Error
> > > Jul  2 07:08:22  server kernel: (ch0:ahd3:0:1:0): SCSI status: 
> > > Check Condition
> > > Jul  2 07:08:22  server kernel: (ch0:ahd3:0:1:0): SCSI sense: 
> > > ILLEGAL REQUEST asc:24,0 (Invalid field in CDB)
> > > Jul  2 07:08:22  server kernel: (ch0:ahd3:0:1:0): Command byte 
> > > 6 is invalid
> > > Jul  2 07:08:22  server kernel: (ch0:ahd3:0:1:0): Error 22, 
> > > Unretryable error
> > > 
> > > scsi_cmd->flags (Byte 6) must be zero for this library so I had to use
> > > this workaround to make it work again until a better solution is 
> > > available:
> > > 
> > > --- sys/cam/scsi/scsi_ch.c.ORI2013-06-26 13:38:54.0 +0200
> > > +++ sys/cam/scsi/scsi_ch.c2013-07-02 07:42:24.0 +0200
> > > @@ -1245,8 +1245,8 @@
> > >/* tag_action */ MSG_SIMPLE_Q_TAG,
> > >/* voltag */ want_voltags,
> > >/* sea */ softc->sc_firsts[chet],
> > > -  /* dvcid */ 1,
> > > -  /* curdata */ 1,
> > > +  /* dvcid */ 0,
> > > +  /* curdata */ 0,
> > >/* count */ 1,
> > >/* data_ptr */ data,
> > >/* dxfer_len */ 1024,
> > > @@ -1284,8 +1284,8 @@
> > >/* voltag */ want

Re: svn commit: r252214 - in stable/9: bin/chio sys/cam/scsi sys/sys

[Kenneth D. Merry]

On Wed, Jul 10, 2013 at 20:58:33 +0200, Andre Albsmeier wrote:
> On Wed, 10-Jul-2013 at 20:49:15 +0200, Kenneth D. Merry wrote:
> > On Wed, Jul 03, 2013 at 07:43:49 +0200, Andre Albsmeier wrote:
> > > On Wed, 03-Jul-2013 at 07:22:36 +0200, Kenneth D. Merry wrote:
> > > > On Tue, Jul 02, 2013 at 07:53:33 +0200, Andre Albsmeier wrote:
> > > > > On Tue, 25-Jun-2013 at 23:43:50 +0200, Kenneth D. Merry wrote:
> > > > > > Author: ken
> > > > > > Date: Tue Jun 25 21:43:49 2013
> > > > > > New Revision: 252214
> > > > > > URL: http://svnweb.freebsd.org/changeset/base/252214
> > > > > 
> > > > > Hi Ken,
> > > > > 
> > > > > > Log:
> > > > > >   MFC: 249658, 249701
> > > > > > 
> > > > > >   Update chio(1) and ch(4) to support reporting element designators.
> > > > > > 
> > > > > >   This allows mapping a tape drive in a changer (as reported by
> > > > > >   'chio status') to a sa(4) driver instance by comparing the
> > > > > >   serial numbers.
> > > > > > 
> > > > > >   The designators can be ASCII (which is printed out directly), 
> > > > > > binary
> > > > > >   (which is printed in hex format) or UTF-8, which is printed in 
> > > > > > either
> > > > > >   native UTF-8 format if the terminal can support it, or in %XX 
> > > > > > notation
> > > > > >   for non-ASCII characters.  Thanks to Hiroki Sato  for the
> > > > > >   explanation and example UTF-8 printing code.
> > > > > > 
> > > > > >   chio.h:   Modify the changer_element_status structure 
> > > > > > to add new
> > > > > > fields and definitions from the SMC3r16 spec.
> > > > > > 
> > > > > > Rename the original CHIOGSTATUS ioctl to 
> > > > > > OCHIOGTATUS and
> > > > > > define a new CHIOGSTATUS ioctl.
> > > > > > 
> > > > > > Clean up some tab/space issues.
> > > > > > 
> > > > > >   chio.c:   For the 'status' subcommand, print the designator 
> > > > > > field
> > > > > > if it is supplied by a device.
> > > > > > 
> > > > > >   scsi_ch.h:Add new flags for DVCID and CURDATA to the READ
> > > > > > ELEMENT STATUS command structure.
> > > > > > 
> > > > > > Add a read_element_status_device_id structure
> > > > > > for the data fields in the new standard. Add new
> > > > > > unions, dt_or_obsolete and voltage_devid, to hold
> > > > > > and address data from either SCSI-2 or newer 
> > > > > > devices.
> > > > > > 
> > > > > >   scsi_ch.c:Implement support for fetching device IDs with READ
> > > > > > ELEMENT STATUS data.
> > > > > > 
> > > > > > Add new arguments to scsi_read_element_status() to
> > > > > > allow the user to request the DVCID and CURDATA 
> > > > > > bits.
> > > > > 
> > > > > This broke "chio status" when talking to my QUALSTAR TLS-8211 library:
> > > > > 
> > > > > Jul  2 07:08:22  server kernel: (ch0:ahd3:0:1:0): READ 
> > > > > ELEMENT STATUS. CDB: b8 10 fd e8 00 01 03 00 04 00 00 00
> > > > > Jul  2 07:08:22  server kernel: (ch0:ahd3:0:1:0): CAM 
> > > > > status: SCSI Status Error
> > > > > Jul  2 07:08:22  server kernel: (ch0:ahd3:0:1:0): SCSI 
> > > > > status: Check Condition
> > > > > Jul  2 07:08:22  server kernel: (ch0:ahd3:0:1:0): SCSI 
> > > > > sense: ILLEGAL REQUEST asc:24,0 (Invalid field in CDB)
> > > > > Jul  2 07:08:22  server kernel: (ch0:ahd3:0:1:0): Command 
> > > > > byte 6 is invalid
> > > > > Jul  2 07:08:22  server kernel: (ch0:ahd3:0:1:0): Error 
> > > > > 22, Unretryable error
> > > > > 
> > > > > scsi_cmd->flags (Byte 6) must be zero for this library so I had to u

Re: svn commit: r252214 - in stable/9: bin/chio sys/cam/scsi sys/sys

[Kenneth D. Merry]

On Thu, Jul 11, 2013 at 14:53:32 -0600, Kenneth D. Merry wrote:
> On Wed, Jul 10, 2013 at 20:58:33 +0200, Andre Albsmeier wrote:
> > On Wed, 10-Jul-2013 at 20:49:15 +0200, Kenneth D. Merry wrote:
> > > On Wed, Jul 03, 2013 at 07:43:49 +0200, Andre Albsmeier wrote:
> > > > On Wed, 03-Jul-2013 at 07:22:36 +0200, Kenneth D. Merry wrote:
> > > > > On Tue, Jul 02, 2013 at 07:53:33 +0200, Andre Albsmeier wrote:
> > > > > > On Tue, 25-Jun-2013 at 23:43:50 +0200, Kenneth D. Merry wrote:
> > > > > > > Author: ken
> > > > > > > Date: Tue Jun 25 21:43:49 2013
> > > > > > > New Revision: 252214
> > > > > > > URL: http://svnweb.freebsd.org/changeset/base/252214
> > > > > > 
> > > > > > Hi Ken,
> > > > > > 
> > > > > > > Log:
> > > > > > >   MFC: 249658, 249701
> > > > > > > 
> > > > > > >   Update chio(1) and ch(4) to support reporting element 
> > > > > > > designators.
> > > > > > > 
> > > > > > >   This allows mapping a tape drive in a changer (as reported by
> > > > > > >   'chio status') to a sa(4) driver instance by comparing the
> > > > > > >   serial numbers.
> > > > > > > 
> > > > > > >   The designators can be ASCII (which is printed out directly), 
> > > > > > > binary
> > > > > > >   (which is printed in hex format) or UTF-8, which is printed in 
> > > > > > > either
> > > > > > >   native UTF-8 format if the terminal can support it, or in %XX 
> > > > > > > notation
> > > > > > >   for non-ASCII characters.  Thanks to Hiroki Sato  for the
> > > > > > >   explanation and example UTF-8 printing code.
> > > > > > > 
> > > > > > >   chio.h:   Modify the changer_element_status 
> > > > > > > structure to add new
> > > > > > > fields and definitions from the SMC3r16 spec.
> > > > > > > 
> > > > > > > Rename the original CHIOGSTATUS ioctl to 
> > > > > > > OCHIOGTATUS and
> > > > > > > define a new CHIOGSTATUS ioctl.
> > > > > > > 
> > > > > > > Clean up some tab/space issues.
> > > > > > > 
> > > > > > >   chio.c:   For the 'status' subcommand, print the designator 
> > > > > > > field
> > > > > > > if it is supplied by a device.
> > > > > > > 
> > > > > > >   scsi_ch.h:Add new flags for DVCID and CURDATA to the READ
> > > > > > > ELEMENT STATUS command structure.
> > > > > > > 
> > > > > > > Add a read_element_status_device_id structure
> > > > > > > for the data fields in the new standard. Add new
> > > > > > > unions, dt_or_obsolete and voltage_devid, to hold
> > > > > > > and address data from either SCSI-2 or newer 
> > > > > > > devices.
> > > > > > > 
> > > > > > >   scsi_ch.c:Implement support for fetching device IDs with 
> > > > > > > READ
> > > > > > > ELEMENT STATUS data.
> > > > > > > 
> > > > > > > Add new arguments to scsi_read_element_status() to
> > > > > > > allow the user to request the DVCID and CURDATA 
> > > > > > > bits.
> > > > > > 
> > > > > > This broke "chio status" when talking to my QUALSTAR TLS-8211 
> > > > > > library:
> > > > > > 
> > > > > > Jul  2 07:08:22  server kernel: (ch0:ahd3:0:1:0): READ 
> > > > > > ELEMENT STATUS. CDB: b8 10 fd e8 00 01 03 00 04 00 00 00
> > > > > > Jul  2 07:08:22  server kernel: (ch0:ahd3:0:1:0): CAM 
> > > > > > status: SCSI Status Error
> > > > > > Jul  2 07:08:22  server kernel: (ch0:ahd3:0:1:0): SCSI 
> > > > > > status: Check Condition
> > > > > > Jul  2 07:08:22  server ker

svn commit: r253274 - head/sys/cam/scsi

[Kenneth D. Merry]

Author: ken
Date: Fri Jul 12 17:09:50 2013
New Revision: 253274
URL: http://svnweb.freebsd.org/changeset/base/253274

Log:
  Fix a problem with READ ELEMENT STATUS that occurs on some
  changers that don't support the DVCID and CURDATA bits that were
  introduced in the SMC spec.
  
  These changers will return an Illegal Request type error if the
  bits are set.  This causes "chio status" to fail.
  
  The fix is two-fold.  First, for changers that claim to be SCSI-2
  or older, don't set the DVCID and CURDATA bits for READ ELEMENT
  STATUS.  For newer changers (SCSI-3 and newer), we default to
  setting the new bits, but back off and try the READ ELEMENT STATUS
  without the bits if we get an Illegal Request type error.
  
  This has been tested on a Qualstar TLS-8211, which is a SCSI-2
  changer that does not support the new bits, and a Spectra T-380,
  which is a SCSI-3 changer that does support the new bits.  In the
  absence of a SCSI-3 changer that does not support the bits, I
  tested that with some error injection code.  (The SMC spec says
  that support for CURDATA is mandatory, and DVCID is optional.)
  
  scsi_ch.c:Add a new quirk, CH_Q_NO_DVCID that gets set for
SCSI-2 and older libraries, or newer libraries that
report errors when the DVCID/CURDATA bits are set.
  
In chgetelemstatus(), use the new quirk to
determine whether or not to set DVCID and CURDATA.
If we get an error with the bits set, back off and
try without the bits.  Set the quirk flag if the
read element status succeeds without the bits set.
  
Increase the READ ELEMENT STATUS timeout to 60
seconds after testing with a Spectra T-380.  The
previous value was 10 seconds, and too short for
the T-380.  This may be decreased later after
some additional testing and investigation.
  
  Tested by:Andre Albsmeier 
  Sponsored by: Spectra Logic
  MFC after:3 days

Modified:
  head/sys/cam/scsi/scsi_ch.c

Modified: head/sys/cam/scsi/scsi_ch.c
==
--- head/sys/cam/scsi/scsi_ch.c Fri Jul 12 16:41:58 2013(r253273)
+++ head/sys/cam/scsi/scsi_ch.c Fri Jul 12 17:09:50 2013(r253274)
@@ -102,7 +102,7 @@ static const u_int32_t  CH_TIMEOUT_MODE_S
 static const u_int32_t CH_TIMEOUT_MOVE_MEDIUM   = 10;
 static const u_int32_t CH_TIMEOUT_EXCHANGE_MEDIUM   = 10;
 static const u_int32_t CH_TIMEOUT_POSITION_TO_ELEMENT   = 10;
-static const u_int32_t CH_TIMEOUT_READ_ELEMENT_STATUS   = 1;
+static const u_int32_t CH_TIMEOUT_READ_ELEMENT_STATUS   = 6;
 static const u_int32_t CH_TIMEOUT_SEND_VOLTAG   = 1;
 static const u_int32_t CH_TIMEOUT_INITIALIZE_ELEMENT_STATUS = 50;
 
@@ -122,12 +122,14 @@ typedef enum {
 
 typedef enum {
CH_Q_NONE   = 0x00,
-   CH_Q_NO_DBD = 0x01
+   CH_Q_NO_DBD = 0x01,
+   CH_Q_NO_DVCID   = 0x02
 } ch_quirks;
 
 #define CH_Q_BIT_STRING\
"\020"  \
-   "\001NO_DBD"
+   "\001NO_DBD"\
+   "\002NO_DVCID"
 
 #define ccb_state  ppriv_field0
 #define ccb_bp ppriv_ptr1
@@ -396,6 +398,14 @@ chregister(struct cam_periph *periph, vo
periph->softc = softc;
softc->quirks = CH_Q_NONE;
 
+   /*
+* The DVCID and CURDATA bits were not introduced until the SMC
+* spec.  If this device claims SCSI-2 or earlier support, then it
+* very likely does not support these bits.
+*/
+   if (cgd->inq_data.version <= SCSI_REV_2)
+   softc->quirks |= CH_Q_NO_DVCID;
+
bzero(&cpi, sizeof(cpi));
xpt_setup_ccb(&cpi.ccb_h, periph->path, CAM_PRIORITY_NORMAL);
cpi.ccb_h.func_code = XPT_PATH_INQ;
@@ -1208,6 +1218,8 @@ chgetelemstatus(struct cam_periph *perip
caddr_t data = NULL;
size_t size, desclen;
int avail, i, error = 0;
+   int curdata, dvcid, sense_flags;
+   int try_no_dvcid = 0;
struct changer_element_status *user_data = NULL;
struct ch_softc *softc;
union ccb *ccb;
@@ -1239,14 +1251,31 @@ chgetelemstatus(struct cam_periph *perip
cam_periph_lock(periph);
ccb = cam_periph_getccb(periph, CAM_PRIORITY_NORMAL);
 
+   sense_flags = SF_RETRY_UA;
+   if (softc->quirks & CH_Q_NO_DVCID) {
+   dvcid = 0;
+   curdata = 0;
+   } else {
+   dvcid = 1;
+   curdata = 1;
+   /*
+* Don't print anything for an Illegal Request, because
+* these flags can cause some changers to complain.  We'll
+* retry without them if we get an error.
+*/
+   sense_flags |= SF_QUIET_IR;
+   }
+
+retry_einval:
+
scsi_read_element_status(&cc

Re: svn commit: r252214 - in stable/9: bin/chio sys/cam/scsi sys/sys

[Kenneth D. Merry]

On Fri, Jul 12, 2013 at 07:39:50 +0200, Andre Albsmeier wrote:
> On Fri, 12-Jul-2013 at 00:02:11 +0200, Kenneth D. Merry wrote:
> > On Thu, Jul 11, 2013 at 14:53:32 -0600, Kenneth D. Merry wrote:
> > > On Wed, Jul 10, 2013 at 20:58:33 +0200, Andre Albsmeier wrote:
> > > > On Wed, 10-Jul-2013 at 20:49:15 +0200, Kenneth D. Merry wrote:
> > > > > On Wed, Jul 03, 2013 at 07:43:49 +0200, Andre Albsmeier wrote:
> > > > > > On Wed, 03-Jul-2013 at 07:22:36 +0200, Kenneth D. Merry wrote:
> > > > > > > On Tue, Jul 02, 2013 at 07:53:33 +0200, Andre Albsmeier wrote:
> > > > > > > > On Tue, 25-Jun-2013 at 23:43:50 +0200, Kenneth D. Merry wrote:
> > > > > > > > > Author: ken
> > > > > > > > > Date: Tue Jun 25 21:43:49 2013
> > > > > > > > > New Revision: 252214
> > > > > > > > > URL: http://svnweb.freebsd.org/changeset/base/252214
> > > > > > > > 
> > > > > > > > Hi Ken,
> > > > > > > > 
> > > > > > > > > Log:
> > > > > > > > >   MFC: 249658, 249701
> > > > > > > > > 
> > > > > > > > >   Update chio(1) and ch(4) to support reporting element 
> > > > > > > > > designators.
> > > > > > > > > 
> > > > > > > > >   This allows mapping a tape drive in a changer (as reported 
> > > > > > > > > by
> > > > > > > > >   'chio status') to a sa(4) driver instance by comparing the
> > > > > > > > >   serial numbers.
> > > > > > > > > 
> > > > > > > > >   The designators can be ASCII (which is printed out 
> > > > > > > > > directly), binary
> > > > > > > > >   (which is printed in hex format) or UTF-8, which is printed 
> > > > > > > > > in either
> > > > > > > > >   native UTF-8 format if the terminal can support it, or in 
> > > > > > > > > %XX notation
> > > > > > > > >   for non-ASCII characters.  Thanks to Hiroki Sato  for 
> > > > > > > > > the
> > > > > > > > >   explanation and example UTF-8 printing code.
> > > > > > > > > 
> > > > > > > > >   chio.h:   Modify the changer_element_status 
> > > > > > > > > structure to add new
> > > > > > > > > fields and definitions from the SMC3r16 spec.
> > > > > > > > > 
> > > > > > > > > Rename the original CHIOGSTATUS ioctl to 
> > > > > > > > > OCHIOGTATUS and
> > > > > > > > > define a new CHIOGSTATUS ioctl.
> > > > > > > > > 
> > > > > > > > > Clean up some tab/space issues.
> > > > > > > > > 
> > > > > > > > >   chio.c:   For the 'status' subcommand, print the 
> > > > > > > > > designator field
> > > > > > > > > if it is supplied by a device.
> > > > > > > > > 
> > > > > > > > >   scsi_ch.h:Add new flags for DVCID and CURDATA to the 
> > > > > > > > > READ
> > > > > > > > > ELEMENT STATUS command structure.
> > > > > > > > > 
> > > > > > > > > Add a read_element_status_device_id structure
> > > > > > > > > for the data fields in the new standard. Add 
> > > > > > > > > new
> > > > > > > > > unions, dt_or_obsolete and voltage_devid, to 
> > > > > > > > > hold
> > > > > > > > > and address data from either SCSI-2 or newer 
> > > > > > > > > devices.
> > > > > > > > > 
> > > > > > > > >   scsi_ch.c:Implement support for fetching device IDs 
> > > > > > > > > with READ
> > > > > > > > > ELEMENT STATUS data.
> > > > > >

svn commit: r253368 - head/sys/cam/scsi

[Kenneth D. Merry]

Author: ken
Date: Mon Jul 15 16:38:48 2013
New Revision: 253368
URL: http://svnweb.freebsd.org/changeset/base/253368

Log:
  Fix an argument reversal in calls to scsi_read_element_status().
  
  Reported by:  Ulrich Spoerlein 
  MFC after:3 days

Modified:
  head/sys/cam/scsi/scsi_ch.c

Modified: head/sys/cam/scsi/scsi_ch.c
==
--- head/sys/cam/scsi/scsi_ch.c Mon Jul 15 15:40:57 2013(r253367)
+++ head/sys/cam/scsi/scsi_ch.c Mon Jul 15 16:38:48 2013(r253368)
@@ -1274,8 +1274,8 @@ retry_einval:
 /* tag_action */ MSG_SIMPLE_Q_TAG,
 /* voltag */ want_voltags,
 /* sea */ softc->sc_firsts[chet],
-/* dvcid */ dvcid,
 /* curdata */ curdata,
+/* dvcid */ dvcid,
 /* count */ 1,
 /* data_ptr */ data,
 /* dxfer_len */ 1024,
@@ -1342,8 +1342,8 @@ retry_einval:
 /* voltag */ want_voltags,
 /* sea */ softc->sc_firsts[chet]
 + cesr->cesr_element_base,
-/* dvcid */ dvcid,
 /* curdata */ curdata,
+/* dvcid */ dvcid,
 /* count */ cesr->cesr_element_count,
 /* data_ptr */ data,
 /* dxfer_len */ size,
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

Re: svn commit: r253274 - head/sys/cam/scsi

[Kenneth D. Merry]

On Mon, Jul 15, 2013 at 17:14:00 +0200, Ulrich Sp??rlein wrote:
> On Fri, 2013-07-12 at 17:09:50 +0000, Kenneth D. Merry wrote:
> > Author: ken
> > Date: Fri Jul 12 17:09:50 2013
> > New Revision: 253274
> > URL: http://svnweb.freebsd.org/changeset/base/253274
> > 
> > Log:
> >   Fix a problem with READ ELEMENT STATUS that occurs on some
> >   changers that don't support the DVCID and CURDATA bits that were
> >   introduced in the SMC spec.
> >   
> >   These changers will return an Illegal Request type error if the
> >   bits are set.  This causes "chio status" to fail.
> >   
> >   The fix is two-fold.  First, for changers that claim to be SCSI-2
> >   or older, don't set the DVCID and CURDATA bits for READ ELEMENT
> >   STATUS.  For newer changers (SCSI-3 and newer), we default to
> >   setting the new bits, but back off and try the READ ELEMENT STATUS
> >   without the bits if we get an Illegal Request type error.
> >   
> >   This has been tested on a Qualstar TLS-8211, which is a SCSI-2
> >   changer that does not support the new bits, and a Spectra T-380,
> >   which is a SCSI-3 changer that does support the new bits.  In the
> >   absence of a SCSI-3 changer that does not support the bits, I
> >   tested that with some error injection code.  (The SMC spec says
> >   that support for CURDATA is mandatory, and DVCID is optional.)
> >   
> >   scsi_ch.c:Add a new quirk, CH_Q_NO_DVCID that gets set for
> > SCSI-2 and older libraries, or newer libraries that
> > report errors when the DVCID/CURDATA bits are set.
> >   
> > In chgetelemstatus(), use the new quirk to
> > determine whether or not to set DVCID and CURDATA.
> > If we get an error with the bits set, back off and
> > try without the bits.  Set the quirk flag if the
> > read element status succeeds without the bits set.
> >   
> > Increase the READ ELEMENT STATUS timeout to 60
> > seconds after testing with a Spectra T-380.  The
> > previous value was 10 seconds, and too short for
> > the T-380.  This may be decreased later after
> > some additional testing and investigation.
> >   
> >   Tested by:Andre Albsmeier 
> >   Sponsored by: Spectra Logic
> >   MFC after:3 days
> > 
> > Modified:
> >   head/sys/cam/scsi/scsi_ch.c
> > 
> > Modified: head/sys/cam/scsi/scsi_ch.c
> > ==
> > --- head/sys/cam/scsi/scsi_ch.c Fri Jul 12 16:41:58 2013
> > (r253273)
> > +++ head/sys/cam/scsi/scsi_ch.c Fri Jul 12 17:09:50 2013
> > (r253274)
> > @@ -1284,8 +1342,8 @@ chgetelemstatus(struct cam_periph *perip
> >  /* voltag */ want_voltags,
> >  /* sea */ softc->sc_firsts[chet]
> >  + cesr->cesr_element_base,
> > -/* dvcid */ 1,
> > -/* curdata */ 1,
> > +/* dvcid */ dvcid,
> > +/* curdata */ curdata,
> >  /* count */ cesr->cesr_element_count,
> >  /* data_ptr */ data,
> >  /* dxfer_len */ size,
> 
> Are you sure? Coverity flags this as being in the wrong argument order
> (there's no CID for this yet).
> 
> CID null (#2 of 2): Arguments in wrong order (SWAPPED_ARGUMENTS)
> swapped_arguments: The positions of arguments curdata and dvcid are 
> inconsistent with the positions of the corresponding parameters for 
> "scsi_read_element_status(struct ccb_scsiio *, u_int32_t, void (*)(struct 
> cam_periph *, union ccb *), u_int8_t, int, u_int32_t, int, int, u_int32_t, 
> u_int8_t *, u_int32_t, u_int8_t, u_int32_t)". [show details]
> 1338scsi_read_element_status(&ccb->csio,
> 1339 /* retries */ 1,
> 1340 /* cbfcnp */ chdone,
> 1341 /* tag_action */ MSG_SIMPLE_Q_TAG,
> 1342 /* voltag */ want_voltags,
> 1343 /* sea */ softc->sc_firsts[chet]
> 1344 + cesr->cesr_element_base,
> 1345 /* dvcid */ dvcid,
> 1346 /* curdata */ curdata,
> 1347 

svn commit: r253369 - stable/9/sys/cam/scsi

[Kenneth D. Merry]

Author: ken
Date: Mon Jul 15 18:02:26 2013
New Revision: 253369
URL: http://svnweb.freebsd.org/changeset/base/253369

Log:
  MFC r253274 and r253368:
  
r253274 | ken | 2013-07-12 11:09:50 -0600 (Fri, 12 Jul 2013) | 40 lines
  
Fix a problem with READ ELEMENT STATUS that occurs on some
changers that don't support the DVCID and CURDATA bits that were
introduced in the SMC spec.
  
These changers will return an Illegal Request type error if the
bits are set.  This causes "chio status" to fail.
  
The fix is two-fold.  First, for changers that claim to be SCSI-2
or older, don't set the DVCID and CURDATA bits for READ ELEMENT
STATUS.  For newer changers (SCSI-3 and newer), we default to
setting the new bits, but back off and try the READ ELEMENT STATUS
without the bits if we get an Illegal Request type error.
  
This has been tested on a Qualstar TLS-8211, which is a SCSI-2
changer that does not support the new bits, and a Spectra T-380,
which is a SCSI-3 changer that does support the new bits.  In the
absence of a SCSI-3 changer that does not support the bits, I
tested that with some error injection code.  (The SMC spec says
that support for CURDATA is mandatory, and DVCID is optional.)
  
scsi_ch.c:  Add a new quirk, CH_Q_NO_DVCID that gets set for
SCSI-2 and older libraries, or newer libraries that
report errors when the DVCID/CURDATA bits are set.
  
In chgetelemstatus(), use the new quirk to
determine whether or not to set DVCID and CURDATA.
If we get an error with the bits set, back off and
try without the bits.  Set the quirk flag if the
read element status succeeds without the bits set.
  
Increase the READ ELEMENT STATUS timeout to 60
seconds after testing with a Spectra T-380.  The
previous value was 10 seconds, and too short for
the T-380.  This may be decreased later after
some additional testing and investigation.
  
Tested by:  Andre Albsmeier 
Sponsored by:   Spectra Logic
  

r253368 | ken | 2013-07-15 10:38:48 -0600 (Mon, 15 Jul 2013) | 5 lines
  
Fix an argument reversal in calls to scsi_read_element_status().
  
Reported by:Ulrich Spoerlein 
  
  Approved by:  re (kib)

Modified:
  stable/9/sys/cam/scsi/scsi_ch.c
Directory Properties:
  stable/9/sys/   (props changed)

Modified: stable/9/sys/cam/scsi/scsi_ch.c
==
--- stable/9/sys/cam/scsi/scsi_ch.c Mon Jul 15 16:38:48 2013
(r253368)
+++ stable/9/sys/cam/scsi/scsi_ch.c Mon Jul 15 18:02:26 2013
(r253369)
@@ -102,7 +102,7 @@ static const u_int32_t  CH_TIMEOUT_MODE_S
 static const u_int32_t CH_TIMEOUT_MOVE_MEDIUM   = 10;
 static const u_int32_t CH_TIMEOUT_EXCHANGE_MEDIUM   = 10;
 static const u_int32_t CH_TIMEOUT_POSITION_TO_ELEMENT   = 10;
-static const u_int32_t CH_TIMEOUT_READ_ELEMENT_STATUS   = 1;
+static const u_int32_t CH_TIMEOUT_READ_ELEMENT_STATUS   = 6;
 static const u_int32_t CH_TIMEOUT_SEND_VOLTAG   = 1;
 static const u_int32_t CH_TIMEOUT_INITIALIZE_ELEMENT_STATUS = 50;
 
@@ -122,12 +122,14 @@ typedef enum {
 
 typedef enum {
CH_Q_NONE   = 0x00,
-   CH_Q_NO_DBD = 0x01
+   CH_Q_NO_DBD = 0x01,
+   CH_Q_NO_DVCID   = 0x02
 } ch_quirks;
 
 #define CH_Q_BIT_STRING\
"\020"  \
-   "\001NO_DBD"
+   "\001NO_DBD"\
+   "\002NO_DVCID"
 
 #define ccb_state  ppriv_field0
 #define ccb_bp ppriv_ptr1
@@ -396,6 +398,14 @@ chregister(struct cam_periph *periph, vo
periph->softc = softc;
softc->quirks = CH_Q_NONE;
 
+   /*
+* The DVCID and CURDATA bits were not introduced until the SMC
+* spec.  If this device claims SCSI-2 or earlier support, then it
+* very likely does not support these bits.
+*/
+   if (cgd->inq_data.version <= SCSI_REV_2)
+   softc->quirks |= CH_Q_NO_DVCID;
+
bzero(&cpi, sizeof(cpi));
xpt_setup_ccb(&cpi.ccb_h, periph->path, CAM_PRIORITY_NORMAL);
cpi.ccb_h.func_code = XPT_PATH_INQ;
@@ -1208,6 +1218,8 @@ chgetelemstatus(struct cam_periph *perip
caddr_t data = NULL;
size_t size, desclen;
int avail, i, error = 0;
+   int curdata, dvcid, sense_flags;
+   int try_no_dvcid = 0;
struct changer_element_status *user_data = NULL;
struct ch_softc *softc;
union ccb *ccb;
@@ -1239,14 +1251,31 @@ chgetelemstatus(struct cam_periph *perip
cam_periph_lock(periph);
ccb = cam_periph_getccb(periph, CAM_PRIORITY_NORMAL);
 
+   sense_flags = SF_RETRY_UA;
+   if (softc->

svn commit: r253550 - head/sys/dev/mps

[Kenneth D. Merry]

Author: ken
Date: Mon Jul 22 18:41:53 2013
New Revision: 253550
URL: http://svnweb.freebsd.org/changeset/base/253550

Log:
  Merge in phase 14+ -> 16 mps driver fixes from LSI:
  
  ---
  System panics during a Port reset with ouststanding I/O
  ---
  It is possible to call mps_mapping_free_memory after this
  memory is already freed, causing a panic. Removed this extra
  call to mps_mappiing_free_memory and call mps_mapping_exit
  in place of the mps_mapping_free_memory call so that any
  outstanding mapping items can be flushed before memory is
  freed.
  
  ---
  Correct memory leak during a Port reset with ouststanding I/O
  ---
  In mps_reinit function, the mapping memory was not being
  freed before being re-allocated. Added line to call the
  memory free function for mapping memory.
  
  ---
  Use CAM_SIM_QUEUED flag in Driver IO path.
  ---
  This flag informs the XPT that successful abort of a CCB
  requires an abort ccb to be issued to the SIM.  While
  processing SCSI IO's, set the CAM_SIM_QUEUED flag in the
  status for the IO. When the command completes, clear this
  flag.
  
  ---
  Check for CAM_REQ_INPROG in I/O path.
  ---
  Added a check in mpssas_action_scsiio for the In Progress
  status for the IO. If this flag is set, the IO has already
  been aborted by the upper layer (before CAM_SIM_QUEUED was
  set) and there is no need to send the IO. The request will
  be completed without error.
  
  ---
  Improve "doorbell handshake method" for mps_get_iocfacts
  ---
  Removed call to get Port Facts since this information is
  not used currently.
  
  Added mps_iocfacts_allocate function to allocate memory
  that is based on IOC Facts data.  Added mps_iocfacts_free
  function to free memory that is based on IOC Facts data.
  Both of the functions are used when a Diag Reset is performed
  or when the driver is attached/detached. This is needed in
  case IOC Facts changes after a Diag Reset, which could
  happen if FW is upgraded.
  
  Moved call of mps_bases_static_config_pages from the attach
  routine to after the IOC is ready to process accesses based
  on the new memory allocations (instead of polling through
  the Doorbell).
  
  ---
  Set TimeStamp in INIT message in millisecond format Set the IOC
  ---
  
  ---
  Prefer mps_wait_command to mps_request_polled
  ---
  Instead of using mps_request_polled, call mps_wait_command
  whenever possible. Change the mps_wait_command function to
  check the current context and either use interrupt context
  or poll if required by using the pause or DELAY function.
  Added a check after waiting 50mSecs to see if the command
  has timed out. This is only done if polliing, the msleep
  command will automatically timeout if the command has taken
  too long to complete.
  
  ---
  Integrated RAID: Volume Activation Failed error message is
  displayed though the volume has been activated.
  ---
  Instead of failing an IOCTL request that does not have a
  large enough buffer to hold the complete reply, copy as
  much data from the reply as possible into the user's buffer
  and log a message saying that the user's buffer was smaller
  than the returned data.
  
  ---
  mapping_add_new_device failure due to persistent table FULL
  ---
  When a new device is added, if it is determined that the
  device persistent table is being used and is full, instead
  of displaying a message for this condition every time, only
  log a message if the MPS_INFO bit is set in the debug_flags.
  
  Submitted by: LSI
  MFC after:1 week

Modified:
  head/sys/dev/mps/mps.c
  head/sys/dev/mps/mps_config.c
  head/sys/dev/mps/mps_mapping.c
  head/sys/dev/mps/mps_sas.c
  head/sys/dev/mps/mps_sas_lsi.c
  head/sys/dev/mps/mps_user.c
  head/sys/dev/mps/mpsvar.h

Modified: head/sys/dev/mps/mps.c
==
--- head/sys/dev/mps/mps.c  Mon Jul 22 18:37:07 

svn commit: r253549 - in head/sys: cam dev/mps sys

[Kenneth D. Merry]

Author: ken
Date: Mon Jul 22 18:37:07 2013
New Revision: 253549
URL: http://svnweb.freebsd.org/changeset/base/253549

Log:
  CAM and mps(4) driver scanning changes.
  
  Add a PIM_NOSCAN flag to the CAM path inquiry CCB.  This tells CAM
  not to perform a rescan on a bus when it is registered.
  
  We now use this flag in the mps(4) driver.  Since it knows what
  devices it has attached, it is more efficient for it to just issue
  a target rescan on the targets that are attached.
  
  Also, remove the private rescan thread from the mps(4) driver in
  favor of the rescan thread already built into CAM.  Without this
  change, but with the change above, the MPS scanner could run before
  or during CAM's initial setup, which would cause duplicate device
  reprobes and announcements.
  
  sys/param.h:
Bump __FreeBSD_version to 139 for the inclusion of the
PIM_RESCAN CAM path inquiry flag.
  
  sys/cam/cam_ccb.h:
  sys/cam/cam_xpt.c:
Added a PIM_NOSCAN flag.  If a SIM sets this in the path
inquiry ccb, then CAM won't rescan the bus in
xpt_bus_regsister.
  
  sys/dev/mps/mps_sas.c
For versions of FreeBSD that have the PIM_NOSCAN path
inquiry flag, don't freeze the sim queue during scanning,
because CAM won't be scanning this bus.  Instead, hold
up the boot.  Don't call mpssas_rescan_target in
mpssas_startup_decrement; it's redundant and I don't
know why it was in there.
  
Set PIM_NOSCAN in path inquiry CCBs.
  
Remove methods related to the internal rescan daemon.
  
Always use async events to trigger a probe for EEDP support.
In older versions of FreeBSD where AC_ADVINFO_CHANGED is
not available, use AC_FOUND_DEVICE and issue the
necessary READ CAPACITY manually.
  
Provide a path to xpt_register_async() so that we only
receive events for our own SCSI domain.
  
Improve error reporting in cases where setup for EEDP
detection fails.
  
  sys/dev/mps/mps_sas.h:
Remove softc flags and data related to the scanner thread.
  
  sys/dev/mps/mps_sas_lsi.c:
Unconditionally rescan the target whenever a device is added.
  
  Sponsored by: Spectra Logic
  MFC after:1 week

Modified:
  head/sys/cam/cam_ccb.h
  head/sys/cam/cam_xpt.c
  head/sys/dev/mps/mps_sas.c
  head/sys/dev/mps/mps_sas.h
  head/sys/dev/mps/mps_sas_lsi.c
  head/sys/sys/param.h

Modified: head/sys/cam/cam_ccb.h
==
--- head/sys/cam/cam_ccb.h  Mon Jul 22 18:18:21 2013(r253548)
+++ head/sys/cam/cam_ccb.h  Mon Jul 22 18:37:07 2013(r253549)
@@ -571,6 +571,7 @@ typedef enum {
PIM_NO_6_BYTE   = 0x08, /* Do not send 6-byte commands */
PIM_SEQSCAN = 0x04, /* Do bus scans sequentially, not in parallel */
PIM_UNMAPPED= 0x02,
+   PIM_NOSCAN  = 0x01  /* SIM does its own scanning */
 } pi_miscflag;
 
 /* Path Inquiry CCB */

Modified: head/sys/cam/cam_xpt.c
==
--- head/sys/cam/cam_xpt.c  Mon Jul 22 18:18:21 2013(r253548)
+++ head/sys/cam/cam_xpt.c  Mon Jul 22 18:37:07 2013(r253549)
@@ -3895,18 +3895,23 @@ xpt_bus_register(struct cam_sim *sim, de
 
/* Notify interested parties */
if (sim->path_id != CAM_XPT_PATH_ID) {
-   union   ccb *scan_ccb;
 
xpt_async(AC_PATH_REGISTERED, path, &cpi);
-   /* Initiate bus rescan. */
-   scan_ccb = xpt_alloc_ccb_nowait();
-   if (scan_ccb != NULL) {
-   scan_ccb->ccb_h.path = path;
-   scan_ccb->ccb_h.func_code = XPT_SCAN_BUS;
-   scan_ccb->crcn.flags = 0;
-   xpt_rescan(scan_ccb);
+   if ((cpi.hba_misc & PIM_NOSCAN) == 0) {
+   union   ccb *scan_ccb;
+
+   /* Initiate bus rescan. */
+   scan_ccb = xpt_alloc_ccb_nowait();
+   if (scan_ccb != NULL) {
+   scan_ccb->ccb_h.path = path;
+   scan_ccb->ccb_h.func_code = XPT_SCAN_BUS;
+   scan_ccb->crcn.flags = 0;
+   xpt_rescan(scan_ccb);
+   } else
+   xpt_print(path,
+ "Can't allocate CCB to scan bus\n");
} else
-   xpt_print(path, "Can't allocate CCB to scan bus\n");
+   xpt_free_path(path);
} else
xpt_free_path(path);
return (CAM_SUCCESS);

Modified: head/sys/dev/mps/mps_sas.c
==
--- head/sys/dev/mps/mps_sas.c  Mon Jul 22 18:18:21 2013(r253548)
+++ head/

Re: svn commit: r247814 - in head: . sys/amd64/conf sys/cam/ctl sys/conf sys/i386/conf

[Kenneth D. Merry]

On Wed, Mar 13, 2013 at 22:09:51 +0100, Pawel Jakub Dawidek wrote:
> On Wed, Mar 13, 2013 at 11:08:26AM -0400, John Baldwin wrote:
> > On Tuesday, March 12, 2013 5:09:21 pm Pawel Jakub Dawidek wrote:
> > > On Mon, Mar 04, 2013 at 09:18:45PM +, Kenneth D. Merry wrote:
> > > > Author: ken
> > > > Date: Mon Mar  4 21:18:45 2013
> > > > New Revision: 247814
> > > > URL: http://svnweb.freebsd.org/changeset/base/247814
> > > > 
> > > > Log:
> > > >   Re-enable CTL in GENERIC on i386 and amd64, but turn on the CTL 
> > > > disable
> > > >   tunable by default.
> > > >   
> > > >   This will allow GENERIC configurations to boot on small memory boxes, 
> > > > but
> > > >   not require end users who want to use CTL to recompile their kernel.  
> > > > They
> > > >   can simply set kern.cam.ctl.disable=0 in loader.conf.
> > > 
> > > Could you rename it to kern.cam.ctl.enable(d)? There was discussion at
> > > some point about sysctl/tunable names and the consensus was, AFAIR, to
> > > use positive(?) names as they are more obvious.
> > 
> > Except that all the hints we use for devices are hint.foo.X.disable=1 :)
> 
> Well we have many sysctls too, but would be nice not to introduce new ones.

This particular sysctl/tunable is not new.  It was in 9.1.

My concern with changing it is that it will lead to surprises for users who
already have it set.

Ken
-- 
Kenneth Merry
k...@freebsd.org
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r248916 - svnadmin/conf

[Kenneth D. Merry]

Author: ken
Date: Fri Mar 29 21:09:00 2013
New Revision: 248916
URL: http://svnweb.freebsd.org/changeset/base/248916

Log:
  Release Will into the wild.

Modified:
  svnadmin/conf/mentors

Modified: svnadmin/conf/mentors
==
--- svnadmin/conf/mentors   Fri Mar 29 21:08:20 2013(r248915)
+++ svnadmin/conf/mentors   Fri Mar 29 21:09:00 2013(r248916)
@@ -33,4 +33,3 @@ peterjjhb Co-mentor: grog
 smhpjd Co-mentor: avg
 snbdwmalone
 versus gavin   Co-mentor: fjoe
-will   ken
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r249025 - head/sys/kern

[Kenneth D. Merry]

Author: ken
Date: Tue Apr  2 16:49:49 2013
New Revision: 249025
URL: http://svnweb.freebsd.org/changeset/base/249025

Log:
  Add support for XPT_CONT_TARGET_IO CCBs in _bus_dmamap_load_ccb().
  
  Declare CCB types in their respective switch blocks.
  
  Sponsored by: Spectra Logic

Modified:
  head/sys/kern/subr_bus_dma.c

Modified: head/sys/kern/subr_bus_dma.c
==
--- head/sys/kern/subr_bus_dma.cTue Apr  2 14:27:44 2013
(r249024)
+++ head/sys/kern/subr_bus_dma.cTue Apr  2 16:49:49 2013
(r249025)
@@ -158,8 +158,6 @@ static int
 _bus_dmamap_load_ccb(bus_dma_tag_t dmat, bus_dmamap_t map, union ccb *ccb,
int *nsegs, int flags)
 {
-   struct ccb_ataio *ataio;
-   struct ccb_scsiio *csio;
struct ccb_hdr *ccb_h;
void *data_ptr;
int error;
@@ -169,18 +167,33 @@ _bus_dmamap_load_ccb(bus_dma_tag_t dmat,
error = 0;
ccb_h = &ccb->ccb_h;
switch (ccb_h->func_code) {
-   case XPT_SCSI_IO:
+   case XPT_SCSI_IO: {
+   struct ccb_scsiio *csio;
+
csio = &ccb->csio;
data_ptr = csio->data_ptr;
dxfer_len = csio->dxfer_len;
sglist_cnt = csio->sglist_cnt;
break;
-   case XPT_ATA_IO:
+   }
+   case XPT_CONT_TARGET_IO: {
+   struct ccb_scsiio *ctio;
+
+   ctio = &ccb->ctio;
+   data_ptr = ctio->data_ptr;
+   dxfer_len = ctio->dxfer_len;
+   sglist_cnt = ctio->sglist_cnt;
+   break;
+   }
+   case XPT_ATA_IO: {
+   struct ccb_ataio *ataio;
+
ataio = &ccb->ataio;
data_ptr = ataio->data_ptr;
dxfer_len = ataio->dxfer_len;
sglist_cnt = 0;
break;
+   }
default:
panic("_bus_dmamap_load_ccb: Unsupported func code %d",
ccb_h->func_code);
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r249028 - head/sys/cam/ctl

[Kenneth D. Merry]

Author: ken
Date: Tue Apr  2 17:29:17 2013
New Revision: 249028
URL: http://svnweb.freebsd.org/changeset/base/249028

Log:
  Fix sending virtual scatter/gather lists from the CTL CAM frontend
  peripheral.
  
  Sponsored by: Spectra Logic

Modified:
  head/sys/cam/ctl/scsi_ctl.c

Modified: head/sys/cam/ctl/scsi_ctl.c
==
--- head/sys/cam/ctl/scsi_ctl.c Tue Apr  2 17:11:07 2013(r249027)
+++ head/sys/cam/ctl/scsi_ctl.c Tue Apr  2 17:29:17 2013(r249028)
@@ -916,7 +916,7 @@ ctlfestart(struct cam_periph *periph, un
if (io->io_hdr.flags & CTL_FLAG_BUS_ADDR)
flags |= CAM_DATA_SG_PADDR;
else
-   flags &= ~CAM_DATA_SG;
+   flags |= CAM_DATA_SG;
data_ptr = (uint8_t *)cam_sglist;
dxfer_len = io->scsiio.kern_data_len;
} else {
@@ -938,6 +938,10 @@ ctlfestart(struct cam_periph *periph, un
data_ptr = sglist[*ti].addr;
dxfer_len = sglist[*ti].len;
csio->sglist_cnt = 0;
+   if (io->io_hdr.flags & CTL_FLAG_BUS_ADDR)
+   flags |= CAM_DATA_PADDR;
+   else
+   flags |= CAM_DATA_VADDR;
cmd_info->flags |= CTLFE_CMD_PIECEWISE;
(*ti)++;
}
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r249167 - stable/9/sys/cam/ctl

[Kenneth D. Merry]

Author: ken
Date: Fri Apr  5 19:33:31 2013
New Revision: 249167
URL: http://svnweb.freebsd.org/changeset/base/249167

Log:
  MFC r244052 and r245288:
  
r244052 | ken | 2012-12-09 12:53:21 -0700 (Sun, 09 Dec 2012) | 20 lines
  
Fix a couple of CTL locking issues and clean up some duplicated code.
  
ctl_frontend_cam_sim.c: Coalesce cfcs_online() and cfcs_offline()
into a single function since these were
identical except for one line.
  
Make sure we hold the SIM lock around path
creation, and calling xpt_rescan().
  
scsi_ctl.c: In ctlfe_onoffline(), make sure we hold the
SIM lock around path creation and free
calls, as well as xpt_action().
  
In ctlfe_lun_enable(), hold the SIM lock
around path and peripheral operations that
require it.
  
Sponsored by:   Spectra Logic Corporation
MFC after:  1 week
  

r245228 | ken | 2013-01-09 10:02:08 -0700 (Wed, 09 Jan 2013) | 43 lines
  
Make CTL work a little better with loading and unloading drivers.
  
Previously CTL would leave individual LUNs enabled in the target
driver, whether or not the port as a whole was enabled.  It would
also leave the wildcard LUN enabled indefinitely.
  
This change means that CTL will enable and disable any active LUNs,
as well as the wildcard LUN, when enabling and disabling a port.
  
Also, fix a bug that could crop up due to an uninitialized CCB
type.
  
ctl.c:  Before calling ctl_frontend_online(), run through
the LUN list and enable all active LUNs.
  
After calling ctl_frontend_offline(), run through
the LUN list and disble all active LUNs.
  
scsi_ctl.c: Before bringing a port online, allocate the
wildcard peripheral for that bus.  And after taking
a port offline, invalidate the wildcard peripheral
for that bus.
  
Make sure that we hold the SIM lock around all
calls to xpt_action() and other transport layer
interfaces that require it.
  
Use CAM_SIM_{LOCK|UNLOCK} consistently to acquire
and release the SIM lock.
  
Update a number of outdated comments.  Some of
these should have been fixed long ago.
  
Actually do LUN disbables now.  The newer drivers
in the tree work correctly for this as far as I
know.
  
Initialize the CCB type to CTLFE_CCB_DEFAULT to
avoid a panic due to uninitialized memory.
  
Submitted by:   Chuck Tuffli (partially)
MFC after:  1 week

Modified:
  stable/9/sys/cam/ctl/ctl.c
  stable/9/sys/cam/ctl/ctl_frontend_cam_sim.c
  stable/9/sys/cam/ctl/scsi_ctl.c
Directory Properties:
  stable/9/sys/   (props changed)

Modified: stable/9/sys/cam/ctl/ctl.c
==
--- stable/9/sys/cam/ctl/ctl.c  Fri Apr  5 18:09:43 2013(r249166)
+++ stable/9/sys/cam/ctl/ctl.c  Fri Apr  5 19:33:31 2013(r249167)
@@ -2263,11 +2263,31 @@ ctl_ioctl(struct cdev *dev, u_long cmd, 
 */
mtx_unlock(&softc->ctl_lock);
 
-   if (cmd == CTL_ENABLE_PORT)
+   if (cmd == CTL_ENABLE_PORT) {
+   struct ctl_lun *lun;
+
+   STAILQ_FOREACH(lun, &softc->lun_list,
+  links) {
+   fe->lun_enable(fe->targ_lun_arg,
+   lun->target,
+   lun->lun);
+   }
+
ctl_frontend_online(fe);
-   else if (cmd == CTL_DISABLE_PORT)
+   } else if (cmd == CTL_DISABLE_PORT) {
+   struct ctl_lun *lun;
+
ctl_frontend_offline(fe);
 
+   STAILQ_FOREACH(lun, &softc->lun_list,
+  links) {
+   fe->lun_disable(
+   fe->targ_lun_arg,
+   lun->target,
+   lun->lun);
+   }
+   }
+
mtx_lock(&

svn commit: r249256 - head/sys/cam/ctl

[Kenneth D. Merry]

Author: ken
Date: Mon Apr  8 15:36:26 2013
New Revision: 249256
URL: http://svnweb.freebsd.org/changeset/base/249256

Log:
  Fix a memory leak that showed up when we delete LUNs.  The memory used for
  the LUN was never freed.
  
  ctl.c:Adjust ctl_alloc_lun() to make sure we don't clear the
CTL_LUN_MALLOCED flag.
  
  Reported by:  Sreenivasa Honnur 
  Sponsored by: Spectra Logic
  MFC after:3 days

Modified:
  head/sys/cam/ctl/ctl.c

Modified: head/sys/cam/ctl/ctl.c
==
--- head/sys/cam/ctl/ctl.c  Mon Apr  8 15:13:15 2013(r249255)
+++ head/sys/cam/ctl/ctl.c  Mon Apr  8 15:36:26 2013(r249256)
@@ -4223,7 +4223,7 @@ ctl_alloc_lun(struct ctl_softc *ctl_soft
 {
struct ctl_lun *nlun, *lun;
struct ctl_frontend *fe;
-   int lun_number, i;
+   int lun_number, i, lun_malloced;
 
if (be_lun == NULL)
return (EINVAL);
@@ -4245,11 +4245,15 @@ ctl_alloc_lun(struct ctl_softc *ctl_soft
}
if (ctl_lun == NULL) {
lun = malloc(sizeof(*lun), M_CTL, M_WAITOK);
-   lun->flags = CTL_LUN_MALLOCED;
-   } else
+   lun_malloced = 1;
+   } else {
+   lun_malloced = 0;
lun = ctl_lun;
+   }
 
memset(lun, 0, sizeof(*lun));
+   if (lun_malloced)
+   lun->flags = CTL_LUN_MALLOCED;
 
mtx_lock(&ctl_softc->ctl_lock);
/*
@@ -4301,7 +4305,7 @@ ctl_alloc_lun(struct ctl_softc *ctl_soft
 * The processor LUN is always enabled.  Disk LUNs come on line
 * disabled, and must be enabled by the backend.
 */
-   lun->flags = CTL_LUN_DISABLED;
+   lun->flags |= CTL_LUN_DISABLED;
lun->backend = be_lun->be;
be_lun->ctl_lun = lun;
be_lun->lun_id = lun_number;
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r249334 - head/usr.bin/ctlstat

[Kenneth D. Merry]

Author: ken
Date: Wed Apr 10 16:01:45 2013
New Revision: 249334
URL: http://svnweb.freebsd.org/changeset/base/249334

Log:
  Fix a time calculation error in ctlstat_standard().
  
  ctlstat.c:When converting a timeval to a floating point
number in ctlstat_standard(), cast the nanoseconds
calculation to a long double, so we don't lose
precision.  Without the cast, we wind up with a
time in whole seconds only.
  
  Sponsored by: Spectra Logic
  MFC after:3 days

Modified:
  head/usr.bin/ctlstat/ctlstat.c

Modified: head/usr.bin/ctlstat/ctlstat.c
==
--- head/usr.bin/ctlstat/ctlstat.c  Wed Apr 10 11:26:30 2013
(r249333)
+++ head/usr.bin/ctlstat/ctlstat.c  Wed Apr 10 16:01:45 2013
(r249334)
@@ -416,9 +416,10 @@ ctlstat_standard(struct ctlstat_context 
if (F_CPU(ctx) && (getcpu(&ctx->cur_cpu) != 0))
errx(1, "error returned from getcpu()");
 
-   cur_secs = ctx->cur_time.tv_sec + (ctx->cur_time.tv_nsec / 10);
+   cur_secs = ctx->cur_time.tv_sec +
+   ((long double)ctx->cur_time.tv_nsec / 10);
prev_secs = ctx->prev_time.tv_sec +
-(ctx->prev_time.tv_nsec / 10);
+((long double)ctx->prev_time.tv_nsec / 10);
 
etime = cur_secs - prev_secs;
 
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r249347 - head/sys/cam/ata

[Kenneth D. Merry]

Author: ken
Date: Wed Apr 10 22:12:21 2013
New Revision: 249347
URL: http://svnweb.freebsd.org/changeset/base/249347

Log:
  Add a callback to the ada(4) driver so that it knows when GEOM has released
  references to it.
  
  This is the functional equivalent to change r237518, which added this
  functionality to the cd(4) and da(4) drivers.
  
  This fix prevents a panic caused by GEOM calling adaopen() while the device
  is going away.  We now keep the device around until GEOM has finished
  cleaning up its state.
  
  ata_da.c: In adaregister(), add a d_gone callback to the GEOM disk
structure registered for the ada driver.  Increment the
peripheral reference count for GEOM.
  
Add a new callback, adadiskgonecb(), that GEOM calls when
it is done with its resources.  This callback releases the
reference acquired in adaregister().
  
  Submitted by: Po-Li Soong
  Sponsored by: Spectra Logic
  MFC After:5 days

Modified:
  head/sys/cam/ata/ata_da.c

Modified: head/sys/cam/ata/ata_da.c
==
--- head/sys/cam/ata/ata_da.c   Wed Apr 10 20:38:15 2013(r249346)
+++ head/sys/cam/ata/ata_da.c   Wed Apr 10 22:12:21 2013(r249347)
@@ -801,6 +801,20 @@ adainit(void)
}
 }
 
+/*
+ * Callback from GEOM, called when it has finished cleaning up its
+ * resources.
+ */
+static void
+adadiskgonecb(struct disk *dp)
+{
+   struct cam_periph *periph;
+
+   periph = (struct cam_periph *)dp->d_drv1;
+
+   cam_periph_release(periph);
+}
+
 static void
 adaoninvalidate(struct cam_periph *periph)
 {
@@ -1157,6 +1171,7 @@ adaregister(struct cam_periph *periph, v
softc->disk->d_strategy = adastrategy;
softc->disk->d_getattr = adagetattr;
softc->disk->d_dump = adadump;
+   softc->disk->d_gone = adadiskgonecb;
softc->disk->d_name = "ada";
softc->disk->d_drv1 = periph;
maxio = cpi.maxio;  /* Honor max I/O size of SIM */
@@ -1222,6 +1237,17 @@ adaregister(struct cam_periph *periph, v
}
} else
legacy_id = -1;
+   /*
+* Acquire a reference to the periph before we register with GEOM.
+* We'll release this reference once GEOM calls us back (via
+* adadiskgonecb()) telling us that our provider has been freed.
+*/
+   if (cam_periph_acquire(periph) != CAM_REQ_CMP) {
+   xpt_print(periph->path, "%s: lost periph during "
+ "registration!\n", __func__);
+   cam_periph_lock(periph);
+   return (CAM_REQ_CMP_ERR);
+   }
disk_create(softc->disk, DISK_VERSION);
cam_periph_lock(periph);
cam_periph_unhold(periph);
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r249384 - head/usr.bin/ctlstat

[Kenneth D. Merry]

Author: ken
Date: Thu Apr 11 21:18:04 2013
New Revision: 249384
URL: http://svnweb.freebsd.org/changeset/base/249384

Log:
  Fix bugs in the elapsed time calculation in ctlstat_standard()
  pointed out by bde:
  
  - Casting to long double isn't needed.
  - The division isn't needed, multiplication can be used.
"When 1 nanosecond is in a floating point literal, the whole
expression is automatically promoted correctly."
  - non-KNF indentation (1 tab) for the newly split line
  - different non-KNF indentation (5 spaces) for the previously split
line
  - exessive parentheses around the division operation
  - bogus blank line which splits up the etime initialization
  - general verboseness from the above.
  
  Submitted by: bde
  MFC after:3 days

Modified:
  head/usr.bin/ctlstat/ctlstat.c

Modified: head/usr.bin/ctlstat/ctlstat.c
==
--- head/usr.bin/ctlstat/ctlstat.c  Thu Apr 11 21:17:49 2013
(r249383)
+++ head/usr.bin/ctlstat/ctlstat.c  Thu Apr 11 21:18:04 2013
(r249384)
@@ -404,7 +404,7 @@ ctlstat_json(struct ctlstat_context *ctx
 
 static void
 ctlstat_standard(struct ctlstat_context *ctx) {
-   long double cur_secs, prev_secs, etime;
+   long double etime;
uint64_t delta_jiffies, delta_idle;
uint32_t port;
long double cpu_percentage;
@@ -416,12 +416,8 @@ ctlstat_standard(struct ctlstat_context 
if (F_CPU(ctx) && (getcpu(&ctx->cur_cpu) != 0))
errx(1, "error returned from getcpu()");
 
-   cur_secs = ctx->cur_time.tv_sec +
-   ((long double)ctx->cur_time.tv_nsec / 10);
-   prev_secs = ctx->prev_time.tv_sec +
-((long double)ctx->prev_time.tv_nsec / 10);
-
-   etime = cur_secs - prev_secs;
+   etime = ctx->cur_time.tv_sec - ctx->prev_time.tv_sec +  
+   (ctx->prev_time.tv_nsec - ctx->cur_time.tv_nsec) * 1e-9; 
 
if (F_CPU(ctx)) {
ctx->prev_total_jiffies = ctx->cur_total_jiffies;
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

Re: svn commit: r249334 - head/usr.bin/ctlstat

[Kenneth D. Merry]

On Thu, Apr 11, 2013 at 17:06:44 +1000, Bruce Evans wrote:
> On Wed, 10 Apr 2013, Kenneth D. Merry wrote:
> 
> >Log:
> > Fix a time calculation error in ctlstat_standard().
> >
> > ctlstat.c:  When converting a timeval to a floating point
> > number in ctlstat_standard(), cast the nanoseconds
> > calculation to a long double, so we don't lose
> > precision.  Without the cast, we wind up with a
> > time in whole seconds only.
> >...
> >Modified: head/usr.bin/ctlstat/ctlstat.c
> >==
> >--- head/usr.bin/ctlstat/ctlstat.c   Wed Apr 10 11:26:30 2013 (r249333)
> >+++ head/usr.bin/ctlstat/ctlstat.c   Wed Apr 10 16:01:45 2013 (r249334)
> >@@ -416,9 +416,10 @@ ctlstat_standard(struct ctlstat_context
> > if (F_CPU(ctx) && (getcpu(&ctx->cur_cpu) != 0))
> > errx(1, "error returned from getcpu()");
> >
> >-cur_secs = ctx->cur_time.tv_sec + (ctx->cur_time.tv_nsec / 
> >10);
> >+cur_secs = ctx->cur_time.tv_sec +
> >+((long double)ctx->cur_time.tv_nsec / 10);
> > prev_secs = ctx->prev_time.tv_sec +
> >- (ctx->prev_time.tv_nsec / 10);
> >+ ((long double)ctx->prev_time.tv_nsec / 10);
> >
> > etime = cur_secs - prev_secs;
> 
> long double is rarely necessary.  It mainly asks for slowness (10-50%)
> on i386 and extreme slowness (hundreds of times slower) on space64.
> Double precision is plenty.  Many arches in FreeBSD have long double ==
> double, so you can't depend on long double being more precise than double,
> and statistics utilities are especially not in need of much precision.
> (Float precision would be enough here.  It would be accurate to 1/16
> of a microsecond.  Not to 1 nanosecond, but you don't need that.  The
> integer division was only accurate to 1/4 second, so ist error was
> noticeable.)
> 
> There is no need for any casts.  There is no need for any divisions.
> Simply multiply by 1 nanosecond.  This must be in floating point, since
> integers can't represent 1 nanosecond (neither can floating, but the
> error of ~2**-53 nanosecnds for double precision is neglogible).  When
> 1 nanosecond is in a floating point literal, the whole expression is
> automatically promoted correctly.
> 
> Other style bugs in the above:
> - non-KNF indentation (1 tab) for the newly split line
> - different non-KNF indentation (5 spaces) for the previously split line
> - exessive parentheses around the division operation
> - bogus blank line which splits up the etime initialization
> - general verboseness from the above.
> 
> Fixing these gives:
> 
>   cur_secs = ctx->cur_time.tv_sec + ctx->cur_time.tv_nsec * 1e-9;
>   prev_secs = ctx->prev_time.tv_sec + ctx->prev_time.tv_nsec * 1e-9
>   etime = cur_secs - prev_secs;
> 
> It is now clear that this is still too verbose, since cur_secs and prev_secs
> are not used except to initialize etime.  Simplifying this gives:
> 
>   etime = ctx->cur_time.tv_sec - ctx->prev_time.tv_sec +
>   (ctx->prev_time.tv_nsec - ctx->cur_time.tv_nsec) * 1e-9;
> 
> This might need casting to double of ctx->cur_time.tv_sec, in case time_t
> is unsigned and the time went backwards.  Otherwise, this should be the
> usual expression for subtracting timespecs.

The time can't go backwards in this case, because it is the system uptime.
Using wall clock time causes problems measuring performance when NTP
decides that the system time needs to change.  I have a local patch to dd
to fix instances of bogus performance numbers due to its using wall clock
time to measure elapsed time.

This should be fixed now, thanks for the review!

Ken
-- 
Kenneth Merry
k...@freebsd.org
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r249592 - in head/sys: conf fs/nfs fs/nfsclient fs/nfsserver modules/nfsd modules/nfsserver nfsserver

[Kenneth D. Merry]

Author: ken
Date: Wed Apr 17 21:00:22 2013
New Revision: 249592
URL: http://svnweb.freebsd.org/changeset/base/249592

Log:
  Revamp the old NFS server's File Handle Affinity (FHA) code so that
  it will work with either the old or new server.
  
  The FHA code keeps a cache of currently active file handles for
  NFSv2 and v3 requests, so that read and write requests for the same
  file are directed to the same group of threads (reads) or thread
  (writes).  It does not currently work for NFSv4 requests.  They are
  more complex, and will take more work to support.
  
  This improves read-ahead performance, especially with ZFS, if the
  FHA tuning parameters are configured appropriately.  Without the
  FHA code, concurrent reads that are part of a sequential read from
  a file will be directed to separate NFS threads.  This has the
  effect of confusing the ZFS zfetch (prefetch) code and makes
  sequential reads significantly slower with clients like Linux that
  do a lot of prefetching.
  
  The FHA code has also been updated to direct write requests to nearby
  file offsets to the same thread in the same way it batches reads,
  and the FHA code will now also send writes to multiple threads when
  needed.
  
  This improves sequential write performance in ZFS, because writes
  to a file are now more ordered.  Since NFS writes (generally
  less than 64K) are smaller than the typical ZFS record size
  (usually 128K), out of order NFS writes to the same block can
  trigger a read in ZFS.  Sending them down the same thread increases
  the odds of their being in order.
  
  In order for multiple write threads per file in the FHA code to be
  useful, writes in the NFS server have been changed to use a LK_SHARED
  vnode lock, and upgrade that to LK_EXCLUSIVE if the filesystem
  doesn't allow multiple writers to a file at once.  ZFS is currently
  the only filesystem that allows multiple writers to a file, because
  it has internal file range locking.  This change does not affect the
  NFSv4 code.
  
  This improves random write performance to a single file in ZFS, since
  we can now have multiple writers inside ZFS at one time.
  
  I have changed the default tuning parameters to a 22 bit (4MB)
  window size (from 256K) and unlimited commands per thread as a
  result of my benchmarking with ZFS.
  
  The FHA code has been updated to allow configuring the tuning
  parameters from loader tunable variables in addition to sysctl
  variables.  The read offset window calculation has been slightly
  modified as well.  Instead of having separate bins, each file
  handle has a rolling window of bin_shift size.  This minimizes
  glitches in throughput when shifting from one bin to another.
  
  sys/conf/files:
Add nfs_fha_new.c and nfs_fha_old.c.  Compile nfs_fha.c
when either the old or the new NFS server is built.
  
  sys/fs/nfs/nfsport.h,
  sys/fs/nfs/nfs_commonport.c:
Bring in changes from Rick Macklem to newnfs_realign that
allow it to operate in blocking (M_WAITOK) or non-blocking
(M_NOWAIT) mode.
  
  sys/fs/nfs/nfs_commonsubs.c,
  sys/fs/nfs/nfs_var.h:
Bring in a change from Rick Macklem to allow telling
nfsm_dissect() whether or not to wait for mallocs.
  
  sys/fs/nfs/nfsm_subs.h:
Bring in changes from Rick Macklem to create a new
nfsm_dissect_nonblock() inline function and
NFSM_DISSECT_NONBLOCK() macro.
  
  sys/fs/nfs/nfs_commonkrpc.c,
  sys/fs/nfsclient/nfs_clkrpc.c:
Add the malloc wait flag to a newnfs_realign() call.
  
  sys/fs/nfsserver/nfs_nfsdkrpc.c:
Setup the new NFS server's RPC thread pool so that it will
call the FHA code.
  
Add the malloc flag argument to newnfs_realign().
  
Unstaticize newnfs_nfsv3_procid[] so that we can use it in
the FHA code.
  
  sys/fs/nfsserver/nfs_nfsdsocket.c:
In nfsrvd_dorpc(), add NFSPROC_WRITE to the list of RPC types
that use the LK_SHARED lock type.
  
  sys/fs/nfsserver/nfs_nfsdport.c:
In nfsd_fhtovp(), if we're starting a write, check to see
whether the underlying filesystem supports shared writes.
If not, upgrade the lock type from LK_SHARED to LK_EXCLUSIVE.
  
  sys/nfsserver/nfs_fha.c:
Remove all code that is specific to the NFS server
implementation.  Anything that is server-specific is now
accessed through a callback supplied by that server's FHA
shim in the new softc.
  
There are now separate sysctls and tunables for the FHA
implementations for the old and new NFS servers.  The new
NFS server has its tunables under vfs.nfsd.fha, the old
NFS server's tunables are under vfs.nfsrv.fha as before.
  
In fha_extract_info(), use callouts for all server-specific
code.  Getting file handles and offsets is now done in the
individual server's shim module.
  
In fha_hash_entry_choose_thread(), change the way w

svn commit: r249596 - in head/sys: conf fs/nfsserver modules/nfsd modules/nfsserver nfs nfsserver

[Kenneth D. Merry]

Author: ken
Date: Wed Apr 17 22:42:43 2013
New Revision: 249596
URL: http://svnweb.freebsd.org/changeset/base/249596

Log:
  Move the NFS FHA (File Handle Affinity) code from sys/nfsserver to
  sys/nfs, since it is now shared by the two NFS servers.
  
  Suggested by: rmacklem
  Sponsored by: Spectra Logic
  MFC after:2 weeks

Added:
  head/sys/nfs/nfs_fha.c
 - copied, changed from r249595, head/sys/nfsserver/nfs_fha.c
  head/sys/nfs/nfs_fha.h
 - copied unchanged from r249595, head/sys/nfsserver/nfs_fha.h
Deleted:
  head/sys/nfsserver/nfs_fha.c
  head/sys/nfsserver/nfs_fha.h
Modified:
  head/sys/conf/files
  head/sys/fs/nfsserver/nfs_fha_new.c
  head/sys/fs/nfsserver/nfs_nfsdkrpc.c
  head/sys/modules/nfsd/Makefile
  head/sys/modules/nfsserver/Makefile
  head/sys/nfsserver/nfs_fha_old.c
  head/sys/nfsserver/nfs_srvkrpc.c

Modified: head/sys/conf/files
==
--- head/sys/conf/files Wed Apr 17 21:21:27 2013(r249595)
+++ head/sys/conf/files Wed Apr 17 22:42:43 2013(r249596)
@@ -3203,6 +3203,7 @@ nfs/bootp_subr.c  optional bootp nfsclie
 nfs/krpc_subr.coptional bootp nfsclient | bootp nfscl
 nfs/nfs_common.c   optional nfsclient | nfsserver
 nfs/nfs_diskless.c optional nfsclient nfs_root | nfscl nfs_root
+nfs/nfs_fha.c  optional nfsserver | nfsd
 nfs/nfs_lock.c optional nfsclient | nfscl | nfslockd | nfsd
 nfsclient/nfs_bio.coptional nfsclient
 nfsclient/nfs_node.c   optional nfsclient
@@ -3211,7 +3212,6 @@ nfsclient/nfs_subs.c  optional nfsclient
 nfsclient/nfs_nfsiod.c optional nfsclient
 nfsclient/nfs_vfsops.c optional nfsclient
 nfsclient/nfs_vnops.c  optional nfsclient
-nfsserver/nfs_fha.coptional nfsserver | nfsd
 nfsserver/nfs_fha_old.coptional nfsserver
 nfsserver/nfs_serv.c   optional nfsserver
 nfsserver/nfs_srvkrpc.coptional nfsserver

Modified: head/sys/fs/nfsserver/nfs_fha_new.c
==
--- head/sys/fs/nfsserver/nfs_fha_new.c Wed Apr 17 21:21:27 2013
(r249595)
+++ head/sys/fs/nfsserver/nfs_fha_new.c Wed Apr 17 22:42:43 2013
(r249596)
@@ -30,11 +30,11 @@ __FBSDID("$FreeBSD$");
 #include 
 
 #include 
+#include 
 #include 
 #include 
 #include 
 #include 
-#include 
 #include 
 
 static void fhanew_init(void *foo);

Modified: head/sys/fs/nfsserver/nfs_nfsdkrpc.c
==
--- head/sys/fs/nfsserver/nfs_nfsdkrpc.cWed Apr 17 21:21:27 2013
(r249595)
+++ head/sys/fs/nfsserver/nfs_nfsdkrpc.cWed Apr 17 22:42:43 2013
(r249596)
@@ -42,7 +42,7 @@ __FBSDID("$FreeBSD$");
 #include 
 #include 
 
-#include 
+#include 
 #include 
 
 #include 

Modified: head/sys/modules/nfsd/Makefile
==
--- head/sys/modules/nfsd/Makefile  Wed Apr 17 21:21:27 2013
(r249595)
+++ head/sys/modules/nfsd/Makefile  Wed Apr 17 22:42:43 2013
(r249596)
@@ -1,6 +1,6 @@
 # $FreeBSD$
 
-.PATH: ${.CURDIR}/../../fs/nfsserver ${.CURDIR}/../../nfsserver
+.PATH: ${.CURDIR}/../../fs/nfsserver ${.CURDIR}/../../nfs
 KMOD=  nfsd
 SRCS=  vnode_if.h \
nfs_fha.c \

Modified: head/sys/modules/nfsserver/Makefile
==
--- head/sys/modules/nfsserver/Makefile Wed Apr 17 21:21:27 2013
(r249595)
+++ head/sys/modules/nfsserver/Makefile Wed Apr 17 22:42:43 2013
(r249596)
@@ -1,6 +1,6 @@
 # $FreeBSD$
 
-.PATH: ${.CURDIR}/../../nfsserver
+.PATH: ${.CURDIR}/../../nfsserver ${.CURDIR}/../../nfs
 KMOD=  nfsserver
 SRCS=  vnode_if.h \
nfs_fha.c nfs_fha_old.c nfs_serv.c nfs_srvkrpc.c nfs_srvsubs.c \

Copied and modified: head/sys/nfs/nfs_fha.c (from r249595, 
head/sys/nfsserver/nfs_fha.c)
==
--- head/sys/nfsserver/nfs_fha.cWed Apr 17 21:21:27 2013
(r249595, copy source)
+++ head/sys/nfs/nfs_fha.c  Wed Apr 17 22:42:43 2013(r249596)
@@ -38,7 +38,7 @@ __FBSDID("$FreeBSD$");
 #include 
 
 #include 
-#include 
+#include 
 
 static MALLOC_DEFINE(M_NFS_FHA, "NFS FHA", "NFS FHA");
 

Copied: head/sys/nfs/nfs_fha.h (from r249595, head/sys/nfsserver/nfs_fha.h)
==
--- /dev/null   00:00:00 1970   (empty, because file is newly added)
+++ head/sys/nfs/nfs_fha.h  Wed Apr 17 22:42:43 2013(r249596, copy 
of r249595, head/sys/nfsserver/nfs_fha.h)
@@ -0,0 +1,112 @@
+/*-
+ * Copyright (c) 2008 Isilon Inc http://www.isilon.com/
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are per

svn commit: r249632 - stable/9/usr.bin/ctlstat

[Kenneth D. Merry]

Author: ken
Date: Fri Apr 19 03:55:54 2013
New Revision: 249632
URL: http://svnweb.freebsd.org/changeset/base/249632

Log:
  MFC r249334 and r249384:
  

r249334 | ken | 2013-04-10 10:01:45 -0600 (Wed, 10 Apr 2013) | 11 lines
  
Fix a time calculation error in ctlstat_standard().
  
ctlstat.c:  When converting a timeval to a floating point
number in ctlstat_standard(), cast the nanoseconds
calculation to a long double, so we don't lose
precision.  Without the cast, we wind up with a
time in whole seconds only.
  
Sponsored by:   Spectra Logic

r249384 | ken | 2013-04-11 15:18:04 -0600 (Thu, 11 Apr 2013) | 17 lines
  
Fix bugs in the elapsed time calculation in ctlstat_standard()
pointed out by bde:
  
- Casting to long double isn't needed.
- The division isn't needed, multiplication can be used.
  "When 1 nanosecond is in a floating point literal, the whole
  expression is automatically promoted correctly."
- non-KNF indentation (1 tab) for the newly split line
- different non-KNF indentation (5 spaces) for the previously split
  line
- exessive parentheses around the division operation
- bogus blank line which splits up the etime initialization
- general verboseness from the above.
  
Submitted by:   bde
  
  Sponsored by: Spectra Logic

Modified:
  stable/9/usr.bin/ctlstat/ctlstat.c
Directory Properties:
  stable/9/usr.bin/ctlstat/   (props changed)

Modified: stable/9/usr.bin/ctlstat/ctlstat.c
==
--- stable/9/usr.bin/ctlstat/ctlstat.c  Fri Apr 19 00:30:52 2013
(r249631)
+++ stable/9/usr.bin/ctlstat/ctlstat.c  Fri Apr 19 03:55:54 2013
(r249632)
@@ -404,7 +404,7 @@ ctlstat_json(struct ctlstat_context *ctx
 
 static void
 ctlstat_standard(struct ctlstat_context *ctx) {
-   long double cur_secs, prev_secs, etime;
+   long double etime;
uint64_t delta_jiffies, delta_idle;
uint32_t port;
long double cpu_percentage;
@@ -416,11 +416,8 @@ ctlstat_standard(struct ctlstat_context 
if (F_CPU(ctx) && (getcpu(&ctx->cur_cpu) != 0))
errx(1, "error returned from getcpu()");
 
-   cur_secs = ctx->cur_time.tv_sec + (ctx->cur_time.tv_nsec / 10);
-   prev_secs = ctx->prev_time.tv_sec +
-(ctx->prev_time.tv_nsec / 10);
-
-   etime = cur_secs - prev_secs;
+   etime = ctx->cur_time.tv_sec - ctx->prev_time.tv_sec +  
+   (ctx->prev_time.tv_nsec - ctx->cur_time.tv_nsec) * 1e-9; 
 
if (F_CPU(ctx)) {
ctx->prev_total_jiffies = ctx->cur_total_jiffies;
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r249658 - in head: bin/chio sys/cam/scsi sys/sys

[Kenneth D. Merry]

Author: ken
Date: Fri Apr 19 20:03:51 2013
New Revision: 249658
URL: http://svnweb.freebsd.org/changeset/base/249658

Log:
  Update chio(1) and ch(4) to support reporting element designators.
  
  This allows mapping a tape drive in a changer (as reported by
  'chio status') to a sa(4) driver instance by comparing the
  serial numbers.
  
  The designators can be ASCII (which is printed out directly), binary
  (which is printed in hex format) or UTF-8, which is printed in either
  native UTF-8 format if the terminal can support it, or in %XX notation
  for non-ASCII characters.  Thanks to Hiroki Sato  for the
  explaining UTF-8 printing and example UTF-8 printing code.
  
  chio.h:   Modify the changer_element_status structure to add new
fields and definitions from the SMC3r16 spec.
  
Rename the original CHIOGSTATUS ioctl to OCHIOGTATUS and
define a new CHIOGSTATUS ioctl.
  
Clean up some tab/space issues.
  
  chio.c:   For the 'status' subcommand, print the designator field
if it is supplied by a device.
  
  scsi_ch.h:Add new flags for DVCID and CURDATA to the READ
ELEMENT STATUS command structure.
  
Add a read_element_status_device_id structure
for the data fields in the new standard. Add new
unions, dt_or_obsolete and voltage_devid, to hold
and address data from either SCSI-2 or newer devices.
  
  scsi_ch.c:Implement support for fetching device IDs with READ
ELEMENT STATUS data.
  
Add new arguments to scsi_read_element_status() to
allow the user to request the DVCID and CURDATA bits.
This isn't compiled into libcam (it's only an internal
kernel interface), so we don't need any special
handling for the API change.
  
If the user issues the new CHIOGSTATUS ioctl, copy all of
the available element status data out.  If he issues the
OCHIOGSTATUS ioctl, we don't copy the new fields in the
structure.
  
Fix a bug in chopen() that would result in the peripheral
never getting unheld if chgetparams() failed.
  
  Sponsored by: Spectra Logic
  Submitted by: Po-Li Soong
  MFC After:1 week

Modified:
  head/bin/chio/chio.c
  head/sys/cam/scsi/scsi_ch.c
  head/sys/cam/scsi/scsi_ch.h
  head/sys/sys/chio.h

Modified: head/bin/chio/chio.c
==
--- head/bin/chio/chio.cFri Apr 19 19:45:00 2013(r249657)
+++ head/bin/chio/chio.cFri Apr 19 20:03:51 2013(r249658)
@@ -54,6 +54,8 @@ __FBSDID("$FreeBSD$");
 #include 
 #include 
 #include 
+#include 
+#include 
 
 #include "defs.h"
 #include "pathnames.h"
@@ -81,6 +83,7 @@ staticint do_status(const char *, int, 
 static int do_ielem(const char *, int, char **);
 static int do_return(const char *, int, char **);
 static int do_voltag(const char *, int, char **);
+static void print_designator(const char *, u_int8_t, u_int8_t);
 
 #ifndef CHET_VT
 #defineCHET_VT 10  /* Completely Arbitrary 
*/
@@ -723,6 +726,10 @@ do_status(const char *cname, int argc, c
putchar('?');
putchar('>');
}
+   if (ces->ces_designator_length > 0)
+   print_designator(ces->ces_designator,
+ces->ces_code_set,
+ces->ces_designator_length);
putchar('\n');
}
 
@@ -1177,3 +1184,66 @@ usage(void)
"arg1 arg2 [arg3 [...]]\n", getprogname());
exit(1);
 }
+
+#defineUTF8CODESET "UTF-8"
+
+static void
+print_designator(const char *designator, u_int8_t code_set,
+u_int8_t designator_length)
+{
+   printf(" serial number: <");
+   switch (code_set) {
+   case CES_CODE_SET_ASCII: {
+   /*
+* The driver insures that the string is always NUL terminated.
+*/
+   printf("%s", designator);
+   break;
+   }
+   case CES_CODE_SET_UTF_8: {
+   char *cs_native;
+
+   setlocale(LC_ALL, "");
+   cs_native = nl_langinfo(CODESET);
+
+   /* See if we can natively print UTF-8 */
+   if (strcmp(cs_native, UTF8CODESET) == 0)
+   cs_native = NULL;
+
+   if (cs_native == NULL) {
+   /* We can natively print UTF-8, so use printf. */
+   printf("%s", designator);
+   } else {
+   int i;
+
+   /*
+* We can't nati

svn commit: r249661 - stable/9/sys/cam/ata

[Kenneth D. Merry]

Author: ken
Date: Fri Apr 19 21:09:27 2013
New Revision: 249661
URL: http://svnweb.freebsd.org/changeset/base/249661

Log:
  MFC r249347:

r249347 | ken | 2013-04-10 16:12:21 -0600 (Wed, 10 Apr 2013) | 22 lines
  
Add a callback to the ada(4) driver so that it knows when GEOM has released
references to it.
  
This is the functional equivalent to change r237518, which added this
functionality to the cd(4) and da(4) drivers.
  
This fix prevents a panic caused by GEOM calling adaopen() while the device
is going away.  We now keep the device around until GEOM has finished
cleaning up its state.
  
ata_da.c:   In adaregister(), add a d_gone callback to the GEOM disk
structure registered for the ada driver.  Increment the
peripheral reference count for GEOM.
  
Add a new callback, adadiskgonecb(), that GEOM calls when
it is done with its resources.  This callback releases the
reference acquired in adaregister().
  
  Submitted by: Po-Li Soong
  Sponsored by: Spectra Logic

Modified:
  stable/9/sys/cam/ata/ata_da.c
Directory Properties:
  stable/9/sys/   (props changed)

Modified: stable/9/sys/cam/ata/ata_da.c
==
--- stable/9/sys/cam/ata/ata_da.c   Fri Apr 19 21:08:56 2013
(r249660)
+++ stable/9/sys/cam/ata/ata_da.c   Fri Apr 19 21:09:27 2013
(r249661)
@@ -792,6 +792,20 @@ adainit(void)
}
 }
 
+/*
+ * Callback from GEOM, called when it has finished cleaning up its
+ * resources.
+ */
+static void
+adadiskgonecb(struct disk *dp)
+{
+   struct cam_periph *periph;
+
+   periph = (struct cam_periph *)dp->d_drv1;
+
+   cam_periph_release(periph);
+}
+
 static void
 adaoninvalidate(struct cam_periph *periph)
 {
@@ -1131,6 +1145,7 @@ adaregister(struct cam_periph *periph, v
softc->disk->d_strategy = adastrategy;
softc->disk->d_getattr = adagetattr;
softc->disk->d_dump = adadump;
+   softc->disk->d_gone = adadiskgonecb;
softc->disk->d_name = "ada";
softc->disk->d_drv1 = periph;
maxio = cpi.maxio;  /* Honor max I/O size of SIM */
@@ -1194,6 +1209,17 @@ adaregister(struct cam_periph *periph, v
}
} else
legacy_id = -1;
+   /*
+* Acquire a reference to the periph before we register with GEOM.
+* We'll release this reference once GEOM calls us back (via
+* adadiskgonecb()) telling us that our provider has been freed.
+*/
+   if (cam_periph_acquire(periph) != CAM_REQ_CMP) {
+   xpt_print(periph->path, "%s: lost periph during "
+ "registration!\n", __func__);
+   cam_periph_lock(periph);
+   return (CAM_REQ_CMP_ERR);
+   }
disk_create(softc->disk, DISK_VERSION);
cam_periph_lock(periph);
cam_periph_unhold(periph);
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r249701 - head/sys/cam/scsi

[Kenneth D. Merry]

Author: ken
Date: Sat Apr 20 14:33:55 2013
New Revision: 249701
URL: http://svnweb.freebsd.org/changeset/base/249701

Log:
  Fix compilation.
  
  Pointy hat to:ken

Modified:
  head/sys/cam/scsi/scsi_ch.c

Modified: head/sys/cam/scsi/scsi_ch.c
==
--- head/sys/cam/scsi/scsi_ch.c Sat Apr 20 14:01:22 2013(r249700)
+++ head/sys/cam/scsi/scsi_ch.c Sat Apr 20 14:33:55 2013(r249701)
@@ -1147,15 +1147,11 @@ copy_element_status(struct ch_softc *sof
ces->ces_designator_length = devid->designator_length;
/*
 * Make sure we are always NUL terminated.  The
-* buffer should be sized for the maximum
-* designator length plus 1, but this will make sure
-* there is always a NUL at the end.  This won't
-* matter for the binary code set, since the user
-* will only pay attention to the length field.
+* This won't matter for the binary code set,
+* since the user will only pay attention to the
+* length field.
 */
-   ces->ces_designator[
-   MIN(sizeof(ces->ces_designator) - 1,
-   devid->designator_length)]= '\0';
+   ces->ces_designator[devid->designator_length]= '\0';
}
if (devid->piv_assoc_designator_type &
READ_ELEMENT_STATUS_PIV_SET) {
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

Re: svn commit: r249658 - in head: bin/chio sys/cam/scsi sys/sys

[Kenneth D. Merry]

On Sat, Apr 20, 2013 at 05:45:43 -0600, Warner Losh wrote:
> Hey Ken,
> 
> this change doesn't compile for me.  See below
> 
> Warner
> 
> On Apr 19, 2013, at 2:03 PM, Kenneth D. Merry wrote:
[ ...]
> > +   ces->ces_designator_length = devid->designator_length;
> > +   /*
> > +* Make sure we are always NUL terminated.  The
> > +* buffer should be sized for the maximum
> > +* designator length plus 1, but this will make sure
> > +* there is always a NUL at the end.  This won't
> > +* matter for the binary code set, since the user
> > +* will only pay attention to the length field.
> > +*/
> > +   ces->ces_designator[
> > +   MIN(sizeof(ces->ces_designator) - 1,
> > +   devid->designator_length)]= '\0';
> 
> compiler complains here that this comparison is always false due to data 
> ranges. I hacked it in my copy by always using devid->designator_length, but 
> I know that's a lame fix. Can you look into it?
> 

Not sure what happened there.  Different warns levels on different
architectures?

In any case, I put in the same fix you did until I can figure out a better
way to do it.  (Hard to type with a toddler on your lap. :)

Ken
-- 
Kenneth Merry
k...@freebsd.org
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r249861 - svnadmin/conf

[Kenneth D. Merry]

Author: ken
Date: Wed Apr 24 22:48:21 2013
New Revision: 249861
URL: http://svnweb.freebsd.org/changeset/base/249861

Log:
  Add Alan Somers (asomers) as a new src committer.
  
  Alan works for Spectra Logic and will be working on ZFS, zfsd, and lots of
  other bug fixes.
  
  gibbs, will and I will be his mentors.
  
  Approved by:  core

Modified:
  svnadmin/conf/access
  svnadmin/conf/mentors

Modified: svnadmin/conf/access
==
--- svnadmin/conf/accessWed Apr 24 21:41:24 2013(r249860)
+++ svnadmin/conf/accessWed Apr 24 22:48:21 2013(r249861)
@@ -32,6 +32,7 @@ andrew
 antoine
 ariff
 art
+asomers
 attilio
 avatar
 avg

Modified: svnadmin/conf/mentors
==
--- svnadmin/conf/mentors   Wed Apr 24 21:41:24 2013(r249860)
+++ svnadmin/conf/mentors   Wed Apr 24 22:48:21 2013(r249861)
@@ -13,6 +13,7 @@
 achim  scottl  Co-mentor: emaste
 anchie bz
 artavg Co-mentor: marcel
+asomersken Co-mentor: gibbs, will
 benl   philip  Co-mentor: simon
 bgray  cognet
 carl   jimharris
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r249862 - head/share/misc

[Kenneth D. Merry]

Author: ken
Date: Wed Apr 24 23:32:47 2013
New Revision: 249862
URL: http://svnweb.freebsd.org/changeset/base/249862

Log:
  Add myself to the committers list.

Modified:
  head/share/misc/committers-src.dot

Modified: head/share/misc/committers-src.dot
==
--- head/share/misc/committers-src.dot  Wed Apr 24 22:48:21 2013
(r249861)
+++ head/share/misc/committers-src.dot  Wed Apr 24 23:32:47 2013
(r249862)
@@ -194,6 +194,7 @@ jwd [label="John De Boskey\njwd@FreeBSD.
 kaiw [label="Kai Wang\nk...@freebsd.org\n2007/09/26"]
 kan [label="Alexander Kabaev\n...@freebsd.org\n2002/07/21"]
 kargl [label="Steven G. Kargl\nka...@freebsd.org\n2011/01/17"]
+ken [label="Ken Merry\n...@freebsd.org\n1998/09/08"]
 kensmith [label="Ken Smith\nkensm...@freebsd.org\n2004/01/23"]
 kevlo [label="Kevin Lo\nke...@freebsd.org\n2006/07/23"]
 kib [label="Konstantin Belousov\n...@freebsd.org\n2006/06/03"]
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

Re: svn commit: r262575 - head/sys/dev/mps

[Kenneth D. Merry]

On Thu, Feb 27, 2014 at 21:48:34 +, Alexander Motin wrote:
> Author: mav
> Date: Thu Feb 27 21:48:34 2014
> New Revision: 262575
> URL: http://svnweb.freebsd.org/changeset/base/262575
> 
> Log:
>   Restore SIM freeze/release match, broken at r253549.
>   
>   This fixes problem with SIM left in frozen state after reinit, for example,
>   after firmware update.

Oops!  Thanks for fixing it!

Ken
-- 
Kenneth Merry
k...@freebsd.org
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

Re: svn commit: r253550 - head/sys/dev/mps

[Kenneth D. Merry]

On Sat, Aug 10, 2013 at 22:14:54 +0400, Slawa Olhovchenkov wrote:
> On Mon, Jul 22, 2013 at 06:41:54PM +0000, Kenneth D. Merry wrote:
> 
> > Author: ken
> > Date: Mon Jul 22 18:41:53 2013
> > New Revision: 253550
> > URL: http://svnweb.freebsd.org/changeset/base/253550
> > 
> > Log:
> >   Merge in phase 14+ -> 16 mps driver fixes from LSI:
> >   Submitted by: LSI
> >   MFC after:1 week
> 
> Not done?

I was waiting for the stable/9 freeze to end.

Now that it's done, I hope to merge that change to stable/9 this week.

Ken
-- 
Kenneth Merry
k...@freebsd.org
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r254372 - head/sys/dev/isp

[Kenneth D. Merry]

Author: ken
Date: Thu Aug 15 16:41:27 2013
New Revision: 254372
URL: http://svnweb.freebsd.org/changeset/base/254372

Log:
  Export the maxio field in the CAM XPT_PATH_INQ CCB in the isp(4)
  driver.
  
  This tells consumers up the stack the maximum I/O size that the
  controller can handle.
  
  The I/O size is bounded by the number of scatter/gather segments
  the controller can handle and the page size.  For an amd64 system,
  it works out to around 5MB.
  
  Reviewed by:  mjacob
  MFC after:3 days
  Sponsored by: Spectra Logic

Modified:
  head/sys/dev/isp/isp_freebsd.c

Modified: head/sys/dev/isp/isp_freebsd.c
==
--- head/sys/dev/isp/isp_freebsd.c  Thu Aug 15 16:03:09 2013
(r254371)
+++ head/sys/dev/isp/isp_freebsd.c  Thu Aug 15 16:41:27 2013
(r254372)
@@ -5445,6 +5445,11 @@ isp_action(struct cam_sim *sim, union cc
cpi->max_target = ISP_MAX_TARGETS(isp) - 1;
cpi->max_lun = ISP_MAX_LUNS(isp) - 1;
cpi->bus_id = cam_sim_bus(sim);
+   if (isp->isp_osinfo.sixtyfourbit)
+   cpi->maxio = (ISP_NSEG64_MAX - 1) * PAGE_SIZE;
+   else
+   cpi->maxio = (ISP_NSEG_MAX - 1) * PAGE_SIZE;
+
bus = cam_sim_bus(xpt_path_sim(cpi->ccb_h.path));
if (IS_FC(isp)) {
fcparam *fcp = FCPARAM(isp, bus);
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r254389 - in head/sys: dev/nvme geom kern sys

[Kenneth D. Merry]

Author: ken
Date: Thu Aug 15 22:52:39 2013
New Revision: 254389
URL: http://svnweb.freebsd.org/changeset/base/254389

Log:
  Change the way that unmapped I/O capability is advertised.
  
  The previous method was to set the D_UNMAPPED_IO flag in the cdevsw
  for the driver.  The problem with this is that in many cases (e.g.
  sa(4)) there may be some instances of the driver that can handle
  unmapped I/O and some that can't.  The isp(4) driver can handle
  unmapped I/O, but the esp(4) driver currently cannot.  The cdevsw
  is shared among all driver instances.
  
  So instead of setting a flag on the cdevsw, set a flag on the cdev.
  This allows drivers to indicate support for unmapped I/O on a
  per-instance basis.
  
  sys/conf.h:   Remove the D_UNMAPPED_IO cdevsw flag and replace it
with an SI_UNMAPPED cdev flag.
  
  kern_physio.c:Look at the cdev SI_UNMAPPED flag to determine
whether or not a particular driver can handle
unmapped I/O.
  
  geom_dev.c:   Set the SI_UNMAPPED flag for all GEOM cdevs.
Since GEOM will create a temporary mapping when
needed, setting SI_UNMAPPED unconditionally will
work.
  
Remove the D_UNMAPPED_IO flag.
  
  nvme_ns.c:Set the SI_UNMAPPED flag on cdevs created here
if NVME_UNMAPPED_BIO_SUPPORT is enabled.
  
  vfs_aio.c:In aio_qphysio(), check the SI_UNMAPPED flag on a
cdev instead of the D_UNMAPPED_IO flag on the cdevsw.
  
  sys/param.h:  Bump __FreeBSD_version to 145 for the switch from
setting the D_UNMAPPED_IO flag in the cdevsw to setting
SI_UNMAPPED in the cdev.
  
  Reviewed by:  kib, jimharris
  MFC after:1 week
  Sponsored by: Spectra Logic

Modified:
  head/sys/dev/nvme/nvme_ns.c
  head/sys/geom/geom_dev.c
  head/sys/kern/kern_physio.c
  head/sys/kern/vfs_aio.c
  head/sys/sys/conf.h
  head/sys/sys/param.h

Modified: head/sys/dev/nvme/nvme_ns.c
==
--- head/sys/dev/nvme/nvme_ns.c Thu Aug 15 22:33:27 2013(r254388)
+++ head/sys/dev/nvme/nvme_ns.c Thu Aug 15 22:52:39 2013(r254389)
@@ -133,11 +133,7 @@ nvme_ns_strategy(struct bio *bp)
 
 static struct cdevsw nvme_ns_cdevsw = {
.d_version =D_VERSION,
-#ifdef NVME_UNMAPPED_BIO_SUPPORT
-   .d_flags =  D_DISK | D_UNMAPPED_IO,
-#else
.d_flags =  D_DISK,
-#endif
.d_read =   physread,
.d_write =  physwrite,
.d_open =   nvme_ns_open,
@@ -348,6 +344,9 @@ nvme_ns_construct(struct nvme_namespace 
NULL, UID_ROOT, GID_WHEEL, 0600, "nvme%dns%d",
device_get_unit(ctrlr->dev), ns->id);
 #endif
+#ifdef NVME_UNMAPPED_BIO_SUPPORT
+   ns->cdev->si_flags |= SI_UNMAPPED;
+#endif
 
if (ns->cdev != NULL)
ns->cdev->si_drv1 = ns;

Modified: head/sys/geom/geom_dev.c
==
--- head/sys/geom/geom_dev.cThu Aug 15 22:33:27 2013(r254388)
+++ head/sys/geom/geom_dev.cThu Aug 15 22:52:39 2013(r254389)
@@ -79,7 +79,7 @@ static struct cdevsw g_dev_cdevsw = {
.d_ioctl =  g_dev_ioctl,
.d_strategy =   g_dev_strategy,
.d_name =   "g_dev",
-   .d_flags =  D_DISK | D_TRACKCLOSE | D_UNMAPPED_IO,
+   .d_flags =  D_DISK | D_TRACKCLOSE,
 };
 
 static g_taste_t g_dev_taste;
@@ -237,6 +237,7 @@ g_dev_taste(struct g_class *mp, struct g
g_free(sc);
return (NULL);
}
+   dev->si_flags |= SI_UNMAPPED;
sc->sc_dev = dev;
 
/* Search for device alias name and create it if found. */
@@ -251,6 +252,7 @@ g_dev_taste(struct g_class *mp, struct g
freeenv(val);
make_dev_alias_p(MAKEDEV_CHECKNAME | MAKEDEV_WAITOK,
&adev, dev, "%s", buf);
+   adev->si_flags |= SI_UNMAPPED;
break;
}
}

Modified: head/sys/kern/kern_physio.c
==
--- head/sys/kern/kern_physio.c Thu Aug 15 22:33:27 2013(r254388)
+++ head/sys/kern/kern_physio.c Thu Aug 15 22:52:39 2013(r254389)
@@ -93,8 +93,7 @@ physio(struct cdev *dev, struct uio *uio
 
csw = dev->si_devsw;
if (uio->uio_segflg == UIO_USERSPACE) {
-   if (csw != NULL &&
-(csw->d_flags & D_UNMAPPED_IO) != 0)
+   if (dev->si_flags & SI_UNMAPPED)
mapped = 0;
else
mapped = 1;

Modified: head/sys/kern/vfs_aio.c

svn commit: r254416 - head/sys/cam/scsi

[Kenneth D. Merry]

Author: ken
Date: Fri Aug 16 16:14:32 2013
New Revision: 254416
URL: http://svnweb.freebsd.org/changeset/base/254416

Log:
  Add unmapped I/O and larger I/O support to the sa(4) driver.
  
  We now pay attention to the maxio field in the XPT_PATH_INQ CCB,
  and if it is set, propagate it up to physio via the si_iosize_max
  field in the cdev structure.
  
  We also now pay attention to the PIM_UNMAPPED capability bit in the
  XPT_PATH_INQ CCB, and set the new SI_UNMAPPED cdev flag when the
  underlying SIM supports unmapped I/O.
  
  scsi_sa.c:Add unmapped I/O support and propagate the SIM's
maximum I/O size up.
  
Adjust scsi_tape_read_write() in the same way that
scsi_read_write() was changed to support unmapped
I/O.  We overload the readop parameter with bits
that tell us whether it's an unmapped I/O, and we
need to set the CAM_DATA_BIO CCB flag.  This change
should be backwards compatible in source and
binary forms.
  
  MFC after:1 week
  Sponsored by: Spectra Logic

Modified:
  head/sys/cam/scsi/scsi_sa.c

Modified: head/sys/cam/scsi/scsi_sa.c
==
--- head/sys/cam/scsi/scsi_sa.c Fri Aug 16 14:22:20 2013(r254415)
+++ head/sys/cam/scsi/scsi_sa.c Fri Aug 16 16:14:32 2013(r254416)
@@ -212,6 +212,7 @@ struct sa_softc {
sa_statestate;
sa_flagsflags;
sa_quirks   quirks;
+   u_int   si_flags;
struct  bio_queue_head bio_queue;
int queue_count;
struct  devstat *device_stats;
@@ -221,6 +222,7 @@ struct sa_softc {
int blk_shift;
u_int32_t   max_blk;
u_int32_t   min_blk;
+   u_int32_t   maxio;
u_int32_t   comp_algorithm;
u_int32_t   saved_comp_algorithm;
u_int32_t   media_blksize;
@@ -447,7 +449,7 @@ static struct cdevsw sa_cdevsw = {
.d_ioctl =  saioctl,
.d_strategy =   sastrategy,
.d_name =   "sa",
-   .d_flags =  D_TAPE | D_NEEDGIANT,
+   .d_flags =  D_TAPE,
 };
 
 static int
@@ -1506,10 +1508,29 @@ saregister(struct cam_periph *periph, vo
DEVSTAT_BS_UNAVAILABLE, SID_TYPE(&cgd->inq_data) |
XPORT_DEVSTAT_TYPE(cpi.transport), DEVSTAT_PRIORITY_TAPE);
 
+   /*
+* If maxio isn't set, we fall back to DFLTPHYS.  If it is set, we
+* take it whether or not it's larger than MAXPHYS.  physio will
+* break it down into pieces small enough to fit in a buffer.
+*/
+   if (cpi.maxio == 0)
+   softc->maxio = DFLTPHYS;
+   else
+   softc->maxio = cpi.maxio;
+
+   /*
+* If the SIM supports unmapped I/O, let physio know that we can
+* handle unmapped buffers.
+*/
+   if (cpi.hba_misc & PIM_UNMAPPED)
+   softc->si_flags = SI_UNMAPPED;
+
softc->devs.ctl_dev = make_dev(&sa_cdevsw, SAMINOR(SA_CTLDEV,
0, SA_ATYPE_R), UID_ROOT, GID_OPERATOR,
0660, "%s%d.ctl", periph->periph_name, periph->unit_number);
softc->devs.ctl_dev->si_drv1 = periph;
+   softc->devs.ctl_dev->si_iosize_max = softc->maxio;
+   softc->devs.ctl_dev->si_flags |= softc->si_flags;
 
for (i = 0; i < SA_NUM_MODES; i++) {
 
@@ -1518,18 +1539,24 @@ saregister(struct cam_periph *periph, vo
UID_ROOT, GID_OPERATOR, 0660, "%s%d.%d",
periph->periph_name, periph->unit_number, i);
softc->devs.mode_devs[i].r_dev->si_drv1 = periph;
+   softc->devs.mode_devs[i].r_dev->si_iosize_max = softc->maxio;
+   softc->devs.mode_devs[i].r_dev->si_flags |= softc->si_flags;
 
softc->devs.mode_devs[i].nr_dev = make_dev(&sa_cdevsw,
SAMINOR(SA_NOT_CTLDEV, i, SA_ATYPE_NR),
UID_ROOT, GID_OPERATOR, 0660, "n%s%d.%d",
periph->periph_name, periph->unit_number, i);
softc->devs.mode_devs[i].nr_dev->si_drv1 = periph;
+   softc->devs.mode_devs[i].nr_dev->si_iosize_max = softc->maxio;
+   softc->devs.mode_devs[i].nr_dev->si_flags |= softc->si_flags;
 
softc->devs.mode_devs[i].er_dev = make_dev(&sa_cdevsw,
SAMINOR(SA_NOT_CTLDEV, i, SA_ATYPE_ER),
UID_ROOT, GID_OPERATOR, 0660, "e%s%d.%d",
periph->periph_name, periph->unit_number, i);
softc->devs.mode_devs[i].er_dev->si_drv1 = periph;
+   softc->devs.mode_devs[i].er_dev->si_iosize_max = softc->maxio;
+   softc->devs.mode_devs[i].er_dev->si_flags |= softc->si_flags;
 
/*
 * Make the (well known) aliases for the first mode.
@@ -1540,12 +1567,20 @@ saregister(struct cam_periph *

svn commit: r254615 - head/sys/dev/mps

[Kenneth D. Merry]

Author: ken
Date: Wed Aug 21 21:30:56 2013
New Revision: 254615
URL: http://svnweb.freebsd.org/changeset/base/254615

Log:
  Fix mps(4) driver breakage that came in in change 253550 that
  manifested itself in out of chain frame conditions.
  
  When the driver ran out of chain frames, the request in question
  would get completed early, and go through mpssas_scsiio_complete().
  
  In mpssas_scsiio_complete(), the negation of the CAM status values
  (CAM_STATUS_MASK | CAM_SIM_QUEUED) was ORed in instead of being
  ANDed in.  This resulted in a bogus CAM CCB status value.  This
  didn't show up in the non-error case, because the status was reset
  to something valid (e.g. CAM_REQ_CMP) later on in the function.
  
  But in the error case, such as when the driver ran out of chain
  frames, the CAM_REQUEUE_REQ status was ORed in to the bogus status
  value.  This led to the CAM transport layer repeatedly releasing
  the SIM queue, because it though that the CAM_RELEASE_SIMQ flag had
  been set.  The symptom was messages like this on the console when
  INVARIANTS were enabled:
  
  xpt_release_simq: requested 1 > present 0
  xpt_release_simq: requested 1 > present 0
  xpt_release_simq: requested 1 > present 0
  
  mps_sas.c:In mpssas_scsiio_complete(), use &= to take status
bits out.  |= adds them in.
  
In the error case in mpssas_scsiio_complete(), set
the status to CAM_REQUEUE_REQ, don't OR it in.
  
  MFC after:3 days
  Sponsored by: Spectra Logic

Modified:
  head/sys/dev/mps/mps_sas.c

Modified: head/sys/dev/mps/mps_sas.c
==
--- head/sys/dev/mps/mps_sas.c  Wed Aug 21 21:30:06 2013(r254614)
+++ head/sys/dev/mps/mps_sas.c  Wed Aug 21 21:30:56 2013(r254615)
@@ -2103,7 +2103,7 @@ mpssas_scsiio_complete(struct mps_softc 
cm->cm_targ->completed++;
cm->cm_targ->outstanding--;
TAILQ_REMOVE(&cm->cm_targ->commands, cm, cm_link);
-   ccb->ccb_h.status |= ~(CAM_STATUS_MASK | CAM_SIM_QUEUED);
+   ccb->ccb_h.status &= ~(CAM_STATUS_MASK | CAM_SIM_QUEUED);
 
if (cm->cm_state == MPS_CM_STATE_TIMEDOUT) {
TAILQ_REMOVE(&cm->cm_targ->timedout_commands, cm, cm_recovery);
@@ -2145,7 +2145,7 @@ mpssas_scsiio_complete(struct mps_softc 
 * because there can be no reply when we haven't actually
 * gone out to the hardware.
 */
-   ccb->ccb_h.status |= CAM_REQUEUE_REQ;
+   ccb->ccb_h.status = CAM_REQUEUE_REQ;
 
/*
 * Currently the only error included in the mask is
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r254627 - in head: bin/chflags bin/ls lib/libc/gen lib/libc/sys sys/cddl/contrib/opensolaris/uts/common/fs/zfs sys/fs/msdosfs sys/fs/smbfs sys/sys sys/ufs/ufs

[Kenneth D. Merry]

Author: ken
Date: Wed Aug 21 23:04:48 2013
New Revision: 254627
URL: http://svnweb.freebsd.org/changeset/base/254627

Log:
  Expand the use of stat(2) flags to allow storing some Windows/DOS
  and CIFS file attributes as BSD stat(2) flags.
  
  This work is intended to be compatible with ZFS, the Solaris CIFS
  server's interaction with ZFS, somewhat compatible with MacOS X,
  and of course compatible with Windows.
  
  The Windows attributes that are implemented were chosen based on
  the attributes that ZFS already supports.
  
  The summary of the flags is as follows:
  
  UF_SYSTEM:Command line name: "system" or "usystem"
ZFS name: XAT_SYSTEM, ZFS_SYSTEM
Windows: FILE_ATTRIBUTE_SYSTEM
  
This flag means that the file is used by the
operating system.  FreeBSD does not enforce any
special handling when this flag is set.
  
  UF_SPARSE:Command line name: "sparse" or "usparse"
ZFS name: XAT_SPARSE, ZFS_SPARSE
Windows: FILE_ATTRIBUTE_SPARSE_FILE
  
This flag means that the file is sparse.  Although
ZFS may modify this in some situations, there is
not generally any special handling for this flag.
  
  UF_OFFLINE:   Command line name: "offline" or "uoffline"
ZFS name: XAT_OFFLINE, ZFS_OFFLINE
Windows: FILE_ATTRIBUTE_OFFLINE
  
This flag means that the file has been moved to
offline storage.  FreeBSD does not have any special
handling for this flag.
  
  UF_REPARSE:   Command line name: "reparse" or "ureparse"
ZFS name: XAT_REPARSE, ZFS_REPARSE
Windows: FILE_ATTRIBUTE_REPARSE_POINT
  
This flag means that the file is a Windows reparse
point.  ZFS has special handling code for reparse
points, but we don't currently have the other
supporting infrastructure for them.
  
  UF_HIDDEN:Command line name: "hidden" or "uhidden"
ZFS name: XAT_HIDDEN, ZFS_HIDDEN
Windows: FILE_ATTRIBUTE_HIDDEN
  
This flag means that the file may be excluded from
a directory listing if the application honors it.
FreeBSD has no special handling for this flag.
  
The name and bit definition for UF_HIDDEN are
identical to the definition in MacOS X.
  
  UF_READONLY:  Command line name: "urdonly", "rdonly", "readonly"
ZFS name: XAT_READONLY, ZFS_READONLY
Windows: FILE_ATTRIBUTE_READONLY
  
This flag means that the file may not written or
appended, but its attributes may be changed.
  
ZFS currently enforces this flag, but Illumos
developers have discussed disabling enforcement.
  
The behavior of this flag is different than MacOS X.
MacOS X uses UF_IMMUTABLE to represent the DOS
readonly permission, but that flag has a stronger
meaning than the semantics of DOS readonly permissions.
  
  UF_ARCHIVE:   Command line name: "uarch", "uarchive"
ZFS_NAME: XAT_ARCHIVE, ZFS_ARCHIVE
Windows name: FILE_ATTRIBUTE_ARCHIVE
  
The UF_ARCHIVED flag means that the file has changed and
needs to be archived.  The meaning is same as
the Windows FILE_ATTRIBUTE_ARCHIVE attribute, and
the ZFS XAT_ARCHIVE and ZFS_ARCHIVE attribute.
  
msdosfs and ZFS have special handling for this flag.
i.e. they will set it when the file changes.
  
  sys/param.h:  Bump __FreeBSD_version to 147 for the
addition of new stat(2) flags.
  
  chflags.1:Document the new command line flag names
(e.g. "system", "hidden") available to the
user.
  
  ls.1: Reference chflags(1) for a list of file flags
and their meanings.
  
  strtofflags.c:Implement the mapping between the new
command line flag names and new stat(2)
flags.
  
  chflags.2:Document all of the new stat(2) flags, and
explain the intended behavior in a little
more detail.  Explain how they map to
Windows file attributes.
  
Different filesystems behave differently
with respect to flags, so warn the
application developer to take care when
using them.
  
  zfs_vnops.c:  Add support for getting and setting the
UF_ARCHIVE, UF_READONLY, UF_SYSTEM, UF_HIDDEN,
UF_REPARSE, UF

Re: svn commit: r254615 - head/sys/dev/mps

[Kenneth D. Merry]

On Thu, Aug 22, 2013 at 16:42:41 +0400, Dmitry Morozovsky wrote:
> Ken,
> 
> On Wed, 21 Aug 2013, Kenneth D. Merry wrote:
> 
> > Author: ken
> > Date: Wed Aug 21 21:30:56 2013
> > New Revision: 254615
> > URL: http://svnweb.freebsd.org/changeset/base/254615
> > 
> > Log:
> >   Fix mps(4) driver breakage that came in in change 253550 that
> >   manifested itself in out of chain frame conditions.
> >   
> >   When the driver ran out of chain frames, the request in question
> >   would get completed early, and go through mpssas_scsiio_complete().
> >   
> >   In mpssas_scsiio_complete(), the negation of the CAM status values
> >   (CAM_STATUS_MASK | CAM_SIM_QUEUED) was ORed in instead of being
> >   ANDed in.  This resulted in a bogus CAM CCB status value.  This
> >   didn't show up in the non-error case, because the status was reset
> >   to something valid (e.g. CAM_REQ_CMP) later on in the function.
> >   
> >   But in the error case, such as when the driver ran out of chain
> >   frames, the CAM_REQUEUE_REQ status was ORed in to the bogus status
> >   value.  This led to the CAM transport layer repeatedly releasing
> >   the SIM queue, because it though that the CAM_RELEASE_SIMQ flag had
> >   been set.  The symptom was messages like this on the console when
> >   INVARIANTS were enabled:
> >   
> >   xpt_release_simq: requested 1 > present 0
> >   xpt_release_simq: requested 1 > present 0
> >   xpt_release_simq: requested 1 > present 0
> 
> what is real impact of the bug?

Your system will essentially hang, certainly as far as anything connected
to the controller in question.

> >   
> >   mps_sas.c:In mpssas_scsiio_complete(), use &= to take status
> > bits out.  |= adds them in.
> >   
> > In the error case in mpssas_scsiio_complete(), set
> > the status to CAM_REQUEUE_REQ, don't OR it in.
> >   
> >   MFC after:3 days
> 
> This patch does not apply cleanly as r253550 had not been merged, and the 
> first 
> masking does not occur on contemporary stable/9. Comments?

As far as I know, this is not a problem on the version of the driver in
stable/9.  But then again, I have not tested the out of chain frames code
since early 2011 when I last fixed it.

If you want to verify the behavior is correct in stable/9, do this:

1. enable INVARIANTS

2. In /boot/loader.conf:
hw.mps.max_chains=32

3. Use up most of your memory.  If you're using ZFS, just do a sequential
write to a file so that the ARC starts filling up with cached data.  Look
at the free memory in top to see how much you've used.  This will cause
enough fragmentation to lead to more scatter/gather segments getting used
in the driver.

4. Do something like this:

((i=0)); while [ $i -lt 60 ]; do dd if=/dev/da0 of=/dev/null bs=1m &
((i++)); done

5.  Look for an out of chain frames message on the console.  To see how far
you are towards using the chain frames, run 'sysctl dev.mps'.  You can see
how many chain frames you have free, and how many requests have failed.

This change just needs to be merged along with the other changes to avoid
having the regression in stable.

Ken
-- 
Kenneth Merry
k...@freebsd.org
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r254760 - in head: share/man/man4 sys/cam/scsi sys/kern sys/sys

[Kenneth D. Merry]

Author: ken
Date: Sat Aug 24 04:52:22 2013
New Revision: 254760
URL: http://svnweb.freebsd.org/changeset/base/254760

Log:
  Add support to physio(9) for devices that don't want I/O split and
  configure sa(4) to request no I/O splitting by default.
  
  For tape devices, the user needs to be able to clearly understand
  what blocksize is actually being used when writing to a tape
  device.  The previous behavior of physio(9) was that it would split
  up any I/O that was too large for the device, or too large to fit
  into MAXPHYS.  This means that if, for instance, the user wrote a
  1MB block to a tape device, and MAXPHYS was 128KB, the 1MB write
  would be split into 8 128K chunks.  This would be done without
  informing the user.
  
  This has suboptimal effects, especially when trying to communicate
  status to the user.  In the event of an error writing to a tape
  (e.g. physical end of tape) in the middle of a 1MB block that has
  been split into 8 pieces, the user could have the first two 128K
  pieces written successfully, the third returned with an error, and
  the last 5 returned with 0 bytes written.  If the user is using
  a standard write(2) system call, all he will see is the ENOSPC
  error.  He won't have a clue how much actually got written.  (With
  a writev(2) system call, he should be able to determine how much
  got written in addition to the error.)
  
  The solution is to prevent physio(9) from splitting the I/O.  The
  new cdev flag, SI_NOSPLIT, tells physio that the driver does not
  want I/O to be split beforehand.
  
  Although the sa(4) driver now enables SI_NOSPLIT by default,
  that can be disabled by two loader tunables for now.  It will not
  be configurable starting in FreeBSD 11.0.  kern.cam.sa.allow_io_split
  allows the user to configure I/O splitting for all sa(4) driver
  instances.  kern.cam.sa.%d.allow_io_split allows the user to
  configure I/O splitting for a specific sa(4) instance.
  
  There are also now three sa(4) driver sysctl variables that let the
  users see some sa(4) driver values.  kern.cam.sa.%d.allow_io_split
  shows whether I/O splitting is turned on.  kern.cam.sa.%d.maxio shows
  the maximum I/O size allowed by kernel configuration parameters
  (e.g. MAXPHYS, DFLTPHYS) and the capabilities of the controller.
  kern.cam.sa.%d.cpi_maxio shows the maximum I/O size supported by
  the controller.
  
  Note that a better long term solution would be to implement support
  for chaining buffers, so that that MAXPHYS is no longer a limiting
  factor for I/O size to tape and disk devices.  At that point, the
  controller and the tape drive would become the limiting factors.
  
  sys/conf.h:   Add a new cdev flag, SI_NOSPLIT, that allows a
driver to tell physio not to split up I/O.
  
  sys/param.h:  Bump __FreeBSD_version to 149 for the addition
of the SI_NOSPLIT cdev flag.
  
  kern_physio.c:If the SI_NOSPLIT flag is set on the cdev, return
any I/O that is larger than si_iosize_max or
MAXPHYS, has more than one segment, or would have
to be split because of misalignment with EFBIG.
(File too large).
  
In the event of an error, print a console message to
give the user a clue about what happened.
  
  scsi_sa.c:Set the SI_NOSPLIT cdev flag on the devices created
for the sa(4) driver by default.
  
Add tunables to control whether we allow I/O splitting
in physio(9).
  
Explain in the comments that allowing I/O splitting
will be deprecated for the sa(4) driver in FreeBSD
11.0.
  
Add sysctl variables to display the maximum I/O
size we can do (which could be further limited by
read block limits) and the maximum I/O size that
the controller can do.
  
Limit our maximum I/O size (recorded in the cdev's
si_iosize_max) by MAXPHYS.  This isn't strictly
necessary, because physio(9) will limit it to
MAXPHYS, but it will provide some clarity for the
application.
  
Record the controller's maximum I/O size reported
in the Path Inquiry CCB.
  
  sa.4: Document the block size behavior, and explain that
the option of allowing physio(9) to split the I/O
will disappear in FreeBSD 11.0.
  
  Sponsored by: Spectra Logic

Modified:
  head/share/man/man4/sa.4
  head/sys/cam/scsi/scsi_sa.c
  head/sys/kern/kern_physio.c
  head/sys/sys/conf.h
  head/sys/sys/param.h

Modified: head/share/man/man4/sa.4
==
--- head/share/man/man4/sa.4Sat Aug 24 01:50:31 2013(r254759)
+++ head/share/man/man4/sa.4Sat Aug 24 04:52:22 2013(r254760)
@@ -25,7 +25,7

svn commit: r254802 - head/sys/kern

[Kenneth D. Merry]

Author: ken
Date: Sat Aug 24 19:02:36 2013
New Revision: 254802
URL: http://svnweb.freebsd.org/changeset/base/254802

Log:
  Fix a printf format warning on 32-bit mips and powerpc.
  
  Reported by:  bde, gjb
  Pointy hat to:ken

Modified:
  head/sys/kern/kern_physio.c

Modified: head/sys/kern/kern_physio.c
==
--- head/sys/kern/kern_physio.c Sat Aug 24 17:45:42 2013(r254801)
+++ head/sys/kern/kern_physio.c Sat Aug 24 19:02:36 2013(r254802)
@@ -117,10 +117,10 @@ physio(struct cdev *dev, struct uio *uio
 * This device does not want I/O to be split.
 */
if (dev->si_flags & SI_NOSPLIT) {
-   printf("%s: request ptr %#jx is not "
+   printf("%s: request ptr %p is not "
"on a page boundary, cannot split "
"request\n", devtoname(dev),
-   (uintmax_t)bp->b_data);
+   bp->b_data);
error = EFBIG;
goto doerror;
}
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

Re: svn commit: r254760 - in head: share/man/man4 sys/cam/scsi sys/kern sys/sys

[Kenneth D. Merry]

On Sat, Aug 24, 2013 at 12:44:46 -0400, Glen Barber wrote:
> On Sat, Aug 24, 2013 at 04:52:22AM +0000, Kenneth D. Merry wrote:
> > Author: ken
> > Date: Sat Aug 24 04:52:22 2013
> > New Revision: 254760
> > URL: http://svnweb.freebsd.org/changeset/base/254760
> > 
> 
> > [...]
> 
> > @@ -83,6 +113,17 @@ physio(struct cdev *dev, struct uio *uio
> >  */
> > iolen = ((vm_offset_t) bp->b_data) & PAGE_MASK;
> > if ((bp->b_bcount + iolen) > bp->b_kvasize) {
> > +   /*
> > +* This device does not want I/O to be split.
> > +*/
> > +   if (dev->si_flags & SI_NOSPLIT) {
> > +   printf("%s: request ptr %#jx is not "
> > +   "on a page boundary, cannot split "
> > +   "request\n", devtoname(dev),
> > +   (uintmax_t)bp->b_data);
> 
> This breaks LINT for at least powerpc.
> 
> cc1: warnings being treated as errors
> /src/sys/kern/kern_physio.c: In function 'physio':
> /src/sys/kern/kern_physio.c:123: warning: cast from pointer to integer of 
> different size [-Wpointer-to-int-cast]
> *** Error code 1
> 
> 
> http://tinderbox.freebsd.org/tinderbox-head-build-HEAD-powerpc-powerpc.full

Oops, sorry about that!

Fixed.

Ken
-- 
Kenneth Merry
k...@freebsd.org
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r254938 - in stable/9/sys: cam dev/mps sys

[Kenneth D. Merry]

Author: ken
Date: Mon Aug 26 21:34:43 2013
New Revision: 254938
URL: http://svnweb.freebsd.org/changeset/base/254938

Log:
  MFC mps(4) driver changes 253460, 253549, 253550 and 254615.
  
  There are some slight changes here from the version in head.
  
  __FreeBSD_version has been bumped to 902502 for the inclusion of the
  PIM_RESCAN CAM path inquiry flag.  The ifdefs in the mps(4) driver
  have been changed accordingly.
  
  In head, the TDP_NOSLEEPING thread flag has been removed but it
  still exists in stable/9.
  

r253460 | scottl | 2013-07-18 18:12:41 -0600 (Thu, 18 Jul 2013) | 5 lines
  
Overhaul error, information, and debug logging.
  
Obtained from:  Netflix
  

r253549 | ken | 2013-07-22 12:37:07 -0600 (Mon, 22 Jul 2013) | 57 lines
  
CAM and mps(4) driver scanning changes.
  
Add a PIM_NOSCAN flag to the CAM path inquiry CCB.  This tells CAM
not to perform a rescan on a bus when it is registered.
  
We now use this flag in the mps(4) driver.  Since it knows what
devices it has attached, it is more efficient for it to just issue
a target rescan on the targets that are attached.
  
Also, remove the private rescan thread from the mps(4) driver in
favor of the rescan thread already built into CAM.  Without this
change, but with the change above, the MPS scanner could run before
or during CAM's initial setup, which would cause duplicate device
reprobes and announcements.
  
sys/param.h:
Bump __FreeBSD_version to 139 for the inclusion of the
PIM_RESCAN CAM path inquiry flag.
  
sys/cam/cam_ccb.h:
sys/cam/cam_xpt.c:
Added a PIM_NOSCAN flag.  If a SIM sets this in the path
inquiry ccb, then CAM won't rescan the bus in
xpt_bus_regsister.
  
sys/dev/mps/mps_sas.c
For versions of FreeBSD that have the PIM_NOSCAN path
inquiry flag, don't freeze the sim queue during scanning,
because CAM won't be scanning this bus.  Instead, hold
up the boot.  Don't call mpssas_rescan_target in
mpssas_startup_decrement; it's redundant and I don't
know why it was in there.
  
Set PIM_NOSCAN in path inquiry CCBs.
  
Remove methods related to the internal rescan daemon.
  
Always use async events to trigger a probe for EEDP support.
In older versions of FreeBSD where AC_ADVINFO_CHANGED is
not available, use AC_FOUND_DEVICE and issue the
necessary READ CAPACITY manually.
  
Provide a path to xpt_register_async() so that we only
receive events for our own SCSI domain.
  
Improve error reporting in cases where setup for EEDP
detection fails.
  
sys/dev/mps/mps_sas.h:
Remove softc flags and data related to the scanner thread.
  
sys/dev/mps/mps_sas_lsi.c:
Unconditionally rescan the target whenever a device is added.
  
Sponsored by:   Spectra Logic
  

r253550 | ken | 2013-07-22 12:41:53 -0600 (Mon, 22 Jul 2013) | 93 lines
  
Merge in phase 14+ -> 16 mps driver fixes from LSI:
  
---
System panics during a Port reset with ouststanding I/O
---
It is possible to call mps_mapping_free_memory after this
memory is already freed, causing a panic. Removed this extra
call to mps_mappiing_free_memory and call mps_mapping_exit
in place of the mps_mapping_free_memory call so that any
outstanding mapping items can be flushed before memory is
freed.
  
---
Correct memory leak during a Port reset with ouststanding I/O
---
In mps_reinit function, the mapping memory was not being
freed before being re-allocated. Added line to call the
memory free function for mapping memory.
  
---
Use CAM_SIM_QUEUED flag in Driver IO path.
---
This flag informs the XPT that successful abort of a CCB
requires an abort ccb to be issued to the SIM.  While
processing SCSI IO's, set the CAM_SIM_QUEUED flag in the
status for the IO. When the command completes, clear this
flag.
  
---
Check for CAM_REQ_INPROG in I/O path.
---
Added a check in mpssas_action_scsiio for the In Progress
status for the IO. If this flag is set, the IO has already
been aborted by the upper laye

Re: svn commit: r253550 - head/sys/dev/mps

[Kenneth D. Merry]

On Tue, Aug 13, 2013 at 19:30:29 +0400, Slawa Olhovchenkov wrote:
> On Mon, Aug 12, 2013 at 01:02:29PM -0600, Kenneth D. Merry wrote:
> 
> > If you really want one now, I've attached a patch from stable/9 on June
> > 27th.  It may or may not apply now.
> 
> Its apply witch litle edit.
> 
> > > > Now that it's done, I hope to merge that change to stable/9 this week.
> > > 
> > > Also, I see strange behaviour of LSI 9211-8i -- after some activity
> > > got timeout and HBA put in looped reset.
> > 
> > That's not good.  Try with the newer driver and see whether if affects the
> > behavior.
> 
> Witch this driver I got same behavior: after intesive i/o got timeout,
> driver try re-init controller without success. On all HDD activity LED
> ON w/o blinking at this moment.

Do you have any logs or dmesg output showing the problem?

I have merged the latest driver changes to stable/9, including a bug fix
for out of chain frame handling.

Ken
-- 
Kenneth Merry
k...@freebsd.org
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

Re: svn commit: r253550 - head/sys/dev/mps

[Kenneth D. Merry]

On Tue, Aug 27, 2013 at 02:25:26 +0400, Slawa Olhovchenkov wrote:
> On Mon, Aug 26, 2013 at 03:40:27PM -0600, Kenneth D. Merry wrote:
> 
> > On Tue, Aug 13, 2013 at 19:30:29 +0400, Slawa Olhovchenkov wrote:
> > > On Mon, Aug 12, 2013 at 01:02:29PM -0600, Kenneth D. Merry wrote:
> > > 
> > > > If you really want one now, I've attached a patch from stable/9 on June
> > > > 27th.  It may or may not apply now.
> > > 
> > > Its apply witch litle edit.
> > > 
> > > > > > Now that it's done, I hope to merge that change to stable/9 this 
> > > > > > week.
> > > > > 
> > > > > Also, I see strange behaviour of LSI 9211-8i -- after some activity
> > > > > got timeout and HBA put in looped reset.
> > > > 
> > > > That's not good.  Try with the newer driver and see whether if affects 
> > > > the
> > > > behavior.
> > > 
> > > Witch this driver I got same behavior: after intesive i/o got timeout,
> > > driver try re-init controller without success. On all HDD activity LED
> > > ON w/o blinking at this moment.
> > 
> > Do you have any logs or dmesg output showing the problem?
> 
> Currenly not awailable, but simmilar as here
> http://lists.freebsd.org/pipermail/freebsd-current/2013-May/041943.html

That one is relatively recent, but all it shows is timeouts.  There are
lots of potential causes.

Without dmesg output, we don't know what firmware version he's running, and
that could be an issue.  As can SATA drives behind an expander.

> or here http://forums.freebsd.org/showthread.php?t=28252.

This one shows lots of IOC terminated errors, but it looks like his problem
was solved by upgrading his SSD firmware.

> At the time of problem -- no any messages in the log or concole, just
> stopping all operations. After some time (aprox 30-60 seconds) --
> circular messages about timeout, reset and massive 'SCSI command
> timeout'.

What was your topology?  i.e. mps(4) controller connected directly to the
drives, or via an expander?  SAS or SATA drives?

> > I have merged the latest driver changes to stable/9, including a bug fix
> > for out of chain frame handling.
> 
> Now I can't test anymore -- controller replaced to asr.

An asr(4) controller?

Ken
-- 
Kenneth Merry
k...@freebsd.org
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r254970 - head/sys/cam/scsi

[Kenneth D. Merry]

Author: ken
Date: Tue Aug 27 19:47:03 2013
New Revision: 254970
URL: http://svnweb.freebsd.org/changeset/base/254970

Log:
  If a drive returns ASC/ASCQ 0x04,0x11 "Logical unit not ready,
  notify (enable spinup) required", instead of doing the normal
  retries, poll for a change in status.
  
  We will poll every half second for a minute for the status to
  change.
  
  Hitachi drives (and likely other SAS drives) return that ASC/ASCQ
  when they are waiting to spin up.  What it means is that they are
  waiting for the SAS expander to send them the SAS
  NOTIFY (ENABLE SPINUP) primitive.
  
  That primitive is the mechanism expanders/enclosures use to
  sequence drive spinup to avoid overloading power supplies.
  
  Sponsored by: Spectra Logic
  MFC after:3 days

Modified:
  head/sys/cam/scsi/scsi_all.c

Modified: head/sys/cam/scsi/scsi_all.c
==
--- head/sys/cam/scsi/scsi_all.cTue Aug 27 19:46:56 2013
(r254969)
+++ head/sys/cam/scsi/scsi_all.cTue Aug 27 19:47:03 2013
(r254970)
@@ -1118,7 +1118,7 @@ static struct asc_table_entry asc_table[
{ SST(0x04, 0x10, SS_RDEF,  /* XXX TBD */
"Logical unit not ready, auxiliary memory not accessible") },
/* DT  WRO AEB VF */
-   { SST(0x04, 0x11, SS_RDEF,  /* XXX TBD */
+   { SST(0x04, 0x11, SS_TUR | SSQ_MANY | SSQ_DECREMENT_COUNT | EBUSY,
"Logical unit not ready, notify (enable spinup) required") },
/*MV  */
{ SST(0x04, 0x12, SS_RDEF,  /* XXX TBD */
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r254971 - stable/9/sys/dev/isp

[Kenneth D. Merry]

Author: ken
Date: Tue Aug 27 20:34:11 2013
New Revision: 254971
URL: http://svnweb.freebsd.org/changeset/base/254971

Log:
  MFC 254372:
  
  Export the maxio field in the CAM XPT_PATH_INQ CCB in the isp(4)
  driver.
  
  This tells consumers up the stack the maximum I/O size that the
  controller can handle.
  
  The I/O size is bounded by the number of scatter/gather segments
  the controller can handle and the page size.  For an amd64 system,
  it works out to around 5MB.
  
  Reviewed by:  mjacob
  Sponsored by: Spectra Logic

Modified:
  stable/9/sys/dev/isp/isp_freebsd.c

Modified: stable/9/sys/dev/isp/isp_freebsd.c
==
--- stable/9/sys/dev/isp/isp_freebsd.c  Tue Aug 27 19:47:03 2013
(r254970)
+++ stable/9/sys/dev/isp/isp_freebsd.c  Tue Aug 27 20:34:11 2013
(r254971)
@@ -5447,6 +5447,11 @@ isp_action(struct cam_sim *sim, union cc
cpi->max_target = ISP_MAX_TARGETS(isp) - 1;
cpi->max_lun = ISP_MAX_LUNS(isp) - 1;
cpi->bus_id = cam_sim_bus(sim);
+   if (isp->isp_osinfo.sixtyfourbit)
+   cpi->maxio = (ISP_NSEG64_MAX - 1) * PAGE_SIZE;
+   else
+   cpi->maxio = (ISP_NSEG_MAX - 1) * PAGE_SIZE;
+
bus = cam_sim_bus(xpt_path_sim(cpi->ccb_h.path));
if (IS_FC(isp)) {
fcparam *fcp = FCPARAM(isp, bus);
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r254972 - in stable/9/sys: . dev dev/isp

[Kenneth D. Merry]

Author: ken
Date: Tue Aug 27 20:43:27 2013
New Revision: 254972
URL: http://svnweb.freebsd.org/changeset/base/254972

Log:
  Properly record mergeinfo for the merge of revision 254372 from head into
  stable/9.
  
  This should have been done in change 254971.
  
  Pointy hat to:ken

Modified:
Directory Properties:
  stable/9/sys/   (props changed)
  stable/9/sys/dev/   (props changed)
  stable/9/sys/dev/isp/   (props changed)
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r254979 - in stable/9/sys: dev/nvme geom kern sys

[Kenneth D. Merry]

Author: ken
Date: Tue Aug 27 23:02:20 2013
New Revision: 254979
URL: http://svnweb.freebsd.org/changeset/base/254979

Log:
  MFC change 254389:
  
  There are some slight differences from the version in FreeBSD/head.
  __FreeBSD_version has been bumped to 902503 for the availability
  of the SI_UNMAPPED cdev flag, and the D_UNMAPPED_IO cdevsw flag
  remains in place.  D_UNMAPPED_IO no longer does anything.  Drivers that
  use that flag will just wind up having mapped I/O by default.  The
  impact will only be on performance, not functionality.
  
Change the way that unmapped I/O capability is advertised.
  
The previous method was to set the D_UNMAPPED_IO flag in the cdevsw
for the driver.  The problem with this is that in many cases (e.g.
sa(4)) there may be some instances of the driver that can handle
unmapped I/O and some that can't.  The isp(4) driver can handle
unmapped I/O, but the esp(4) driver currently cannot.  The cdevsw
is shared among all driver instances.
  
So instead of setting a flag on the cdevsw, set a flag on the cdev.
This allows drivers to indicate support for unmapped I/O on a
per-instance basis.
  
sys/conf.h: Remove the D_UNMAPPED_IO cdevsw flag and replace it
with an SI_UNMAPPED cdev flag.
  
kern_physio.c:  Look at the cdev SI_UNMAPPED flag to determine
whether or not a particular driver can handle
unmapped I/O.
  
geom_dev.c: Set the SI_UNMAPPED flag for all GEOM cdevs.
Since GEOM will create a temporary mapping when
needed, setting SI_UNMAPPED unconditionally will
work.
  
Remove the D_UNMAPPED_IO flag.
  
nvme_ns.c:  Set the SI_UNMAPPED flag on cdevs created here
if NVME_UNMAPPED_BIO_SUPPORT is enabled.
  
vfs_aio.c:  In aio_qphysio(), check the SI_UNMAPPED flag on a
cdev instead of the D_UNMAPPED_IO flag on the cdevsw.
  
sys/param.h:Bump __FreeBSD_version to 145 for the switch from
setting the D_UNMAPPED_IO flag in the cdevsw to setting
SI_UNMAPPED in the cdev.
  
Reviewed by:kib, jimharris
Sponsored by:   Spectra Logic

Modified:
  stable/9/sys/dev/nvme/nvme_ns.c
  stable/9/sys/geom/geom_dev.c
  stable/9/sys/kern/kern_physio.c
  stable/9/sys/kern/vfs_aio.c
  stable/9/sys/sys/conf.h
  stable/9/sys/sys/param.h
Directory Properties:
  stable/9/sys/   (props changed)
  stable/9/sys/dev/   (props changed)
  stable/9/sys/sys/   (props changed)

Modified: stable/9/sys/dev/nvme/nvme_ns.c
==
--- stable/9/sys/dev/nvme/nvme_ns.c Tue Aug 27 22:37:29 2013
(r254978)
+++ stable/9/sys/dev/nvme/nvme_ns.c Tue Aug 27 23:02:20 2013
(r254979)
@@ -133,11 +133,7 @@ nvme_ns_strategy(struct bio *bp)
 
 static struct cdevsw nvme_ns_cdevsw = {
.d_version =D_VERSION,
-#ifdef NVME_UNMAPPED_BIO_SUPPORT
-   .d_flags =  D_DISK | D_UNMAPPED_IO,
-#else
.d_flags =  D_DISK,
-#endif
.d_read =   physread,
.d_write =  physwrite,
.d_open =   nvme_ns_open,
@@ -348,6 +344,9 @@ nvme_ns_construct(struct nvme_namespace 
NULL, UID_ROOT, GID_WHEEL, 0600, "nvme%dns%d",
device_get_unit(ctrlr->dev), ns->id);
 #endif
+#ifdef NVME_UNMAPPED_BIO_SUPPORT
+   ns->cdev->si_flags |= SI_UNMAPPED;
+#endif
 
if (ns->cdev != NULL)
ns->cdev->si_drv1 = ns;

Modified: stable/9/sys/geom/geom_dev.c
==
--- stable/9/sys/geom/geom_dev.cTue Aug 27 22:37:29 2013
(r254978)
+++ stable/9/sys/geom/geom_dev.cTue Aug 27 23:02:20 2013
(r254979)
@@ -79,7 +79,7 @@ static struct cdevsw g_dev_cdevsw = {
.d_ioctl =  g_dev_ioctl,
.d_strategy =   g_dev_strategy,
.d_name =   "g_dev",
-   .d_flags =  D_DISK | D_TRACKCLOSE | D_UNMAPPED_IO,
+   .d_flags =  D_DISK | D_TRACKCLOSE,
 };
 
 static g_taste_t g_dev_taste;
@@ -237,6 +237,7 @@ g_dev_taste(struct g_class *mp, struct g
g_free(sc);
return (NULL);
}
+   dev->si_flags |= SI_UNMAPPED;
sc->sc_dev = dev;
 
/* Search for device alias name and create it if found. */
@@ -251,6 +252,7 @@ g_dev_taste(struct g_class *mp, struct g
freeenv(val);
make_dev_alias_p(MAKEDEV_CHECKNAME | MAKEDEV_WAITOK,
&adev, dev, "%s", buf);
+   adev->si_flags |= SI_UNMAPPED;
break;
}
}

Modified: stable/9/sys/kern/kern_physio.c
==
--- stable/9/sys/kern/kern_physio.c Tue Aug 27 22:37:29 2013
(r254978)
+++ stable/9/s

svn commit: r255000 - stable/9/sys/cam/scsi

[Kenneth D. Merry]

Author: ken
Date: Wed Aug 28 19:22:09 2013
New Revision: 255000
URL: http://svnweb.freebsd.org/changeset/base/255000

Log:
  MFC 254416:
  

r254416 | ken | 2013-08-16 10:14:32 -0600 (Fri, 16 Aug 2013) | 24 lines
  
Add unmapped I/O and larger I/O support to the sa(4) driver.
  
We now pay attention to the maxio field in the XPT_PATH_INQ CCB,
and if it is set, propagate it up to physio via the si_iosize_max
field in the cdev structure.
  
We also now pay attention to the PIM_UNMAPPED capability bit in the
XPT_PATH_INQ CCB, and set the new SI_UNMAPPED cdev flag when the
underlying SIM supports unmapped I/O.
  
scsi_sa.c:  Add unmapped I/O support and propagate the SIM's
maximum I/O size up.
  
Adjust scsi_tape_read_write() in the same way that
scsi_read_write() was changed to support unmapped
I/O.  We overload the readop parameter with bits
that tell us whether it's an unmapped I/O, and we
need to set the CAM_DATA_BIO CCB flag.  This change
should be backwards compatible in source and
binary forms.
  
Sponsored by:   Spectra Logic

Modified:
  stable/9/sys/cam/scsi/scsi_sa.c
Directory Properties:
  stable/9/sys/   (props changed)

Modified: stable/9/sys/cam/scsi/scsi_sa.c
==
--- stable/9/sys/cam/scsi/scsi_sa.c Wed Aug 28 19:06:22 2013
(r254999)
+++ stable/9/sys/cam/scsi/scsi_sa.c Wed Aug 28 19:22:09 2013
(r255000)
@@ -212,6 +212,7 @@ struct sa_softc {
sa_statestate;
sa_flagsflags;
sa_quirks   quirks;
+   u_int   si_flags;
struct  bio_queue_head bio_queue;
int queue_count;
struct  devstat *device_stats;
@@ -221,6 +222,7 @@ struct sa_softc {
int blk_shift;
u_int32_t   max_blk;
u_int32_t   min_blk;
+   u_int32_t   maxio;
u_int32_t   comp_algorithm;
u_int32_t   saved_comp_algorithm;
u_int32_t   media_blksize;
@@ -447,7 +449,7 @@ static struct cdevsw sa_cdevsw = {
.d_ioctl =  saioctl,
.d_strategy =   sastrategy,
.d_name =   "sa",
-   .d_flags =  D_TAPE | D_NEEDGIANT,
+   .d_flags =  D_TAPE,
 };
 
 static int
@@ -1506,10 +1508,29 @@ saregister(struct cam_periph *periph, vo
DEVSTAT_BS_UNAVAILABLE, SID_TYPE(&cgd->inq_data) |
XPORT_DEVSTAT_TYPE(cpi.transport), DEVSTAT_PRIORITY_TAPE);
 
+   /*
+* If maxio isn't set, we fall back to DFLTPHYS.  If it is set, we
+* take it whether or not it's larger than MAXPHYS.  physio will
+* break it down into pieces small enough to fit in a buffer.
+*/
+   if (cpi.maxio == 0)
+   softc->maxio = DFLTPHYS;
+   else
+   softc->maxio = cpi.maxio;
+
+   /*
+* If the SIM supports unmapped I/O, let physio know that we can
+* handle unmapped buffers.
+*/
+   if (cpi.hba_misc & PIM_UNMAPPED)
+   softc->si_flags = SI_UNMAPPED;
+
softc->devs.ctl_dev = make_dev(&sa_cdevsw, SAMINOR(SA_CTLDEV,
0, SA_ATYPE_R), UID_ROOT, GID_OPERATOR,
0660, "%s%d.ctl", periph->periph_name, periph->unit_number);
softc->devs.ctl_dev->si_drv1 = periph;
+   softc->devs.ctl_dev->si_iosize_max = softc->maxio;
+   softc->devs.ctl_dev->si_flags |= softc->si_flags;
 
for (i = 0; i < SA_NUM_MODES; i++) {
 
@@ -1518,18 +1539,24 @@ saregister(struct cam_periph *periph, vo
UID_ROOT, GID_OPERATOR, 0660, "%s%d.%d",
periph->periph_name, periph->unit_number, i);
softc->devs.mode_devs[i].r_dev->si_drv1 = periph;
+   softc->devs.mode_devs[i].r_dev->si_iosize_max = softc->maxio;
+   softc->devs.mode_devs[i].r_dev->si_flags |= softc->si_flags;
 
softc->devs.mode_devs[i].nr_dev = make_dev(&sa_cdevsw,
SAMINOR(SA_NOT_CTLDEV, i, SA_ATYPE_NR),
UID_ROOT, GID_OPERATOR, 0660, "n%s%d.%d",
periph->periph_name, periph->unit_number, i);
softc->devs.mode_devs[i].nr_dev->si_drv1 = periph;
+   softc->devs.mode_devs[i].nr_dev->si_iosize_max = softc->maxio;
+   softc->devs.mode_devs[i].nr_dev->si_flags |= softc->si_flags;
 
softc->devs.mode_devs[i].er_dev = make_dev(&sa_cdevsw,
SAMINOR(SA_NOT_CTLDEV, i, SA_ATYPE_ER),
UID_ROOT, GID_OPERATOR, 0660, "e%s%d.%d",
periph->periph_name, periph->unit_number, i);
softc->devs.mode_devs[i].er_dev->si_drv1 = periph;
+   softc->devs.mode_devs[i].er_dev

Re: svn commit: r254627 - in head: bin/chflags bin/ls lib/libc/gen lib/libc/sys sys/cddl/contrib/opensolaris/uts/common/fs/zfs sys/fs/msdosfs sys/fs/smbfs sys/sys sys/ufs/ufs

[Kenneth D. Merry]

On Wed, Aug 28, 2013 at 15:54:18 -0500, Bryan Drewery wrote:
> On 2013-08-21 18:04, Kenneth D. Merry wrote:
> >Author: ken
> >Date: Wed Aug 21 23:04:48 2013
> >New Revision: 254627
> >URL: http://svnweb.freebsd.org/changeset/base/254627
> >
> >Log:
> >  Expand the use of stat(2) flags to allow storing some Windows/DOS
> >  and CIFS file attributes as BSD stat(2) flags.
> >
> >  This work is intended to be compatible with ZFS, the Solaris CIFS
> >  server's interaction with ZFS, somewhat compatible with MacOS X,
> >  and of course compatible with Windows.
> >
> >  The Windows attributes that are implemented were chosen based on
> >  the attributes that ZFS already supports.
> >
> >  The summary of the flags is as follows:
> [...]
> >
> >  UF_ARCHIVE:Command line name: "uarch", "uarchive"
> > ZFS_NAME: XAT_ARCHIVE, ZFS_ARCHIVE
> > Windows name: FILE_ATTRIBUTE_ARCHIVE
> >
> > The UF_ARCHIVED flag means that the file has changed and
> > needs to be archived.  The meaning is same as
> > the Windows FILE_ATTRIBUTE_ARCHIVE attribute, and
> > the ZFS XAT_ARCHIVE and ZFS_ARCHIVE attribute.
> >
> > msdosfs and ZFS have special handling for this flag.
> > i.e. they will set it when the file changes.
> 
> Is it intended that this flag is automatically added to all new and 
> existing ZFS files?

Yes, that is intentional.  ZFS already has the flag internally, this just
exposes it to FreeBSD.  With ZFS at least, it is set any time a file
changes.  If an application clears it on files it has archived, it will
know when the file has changed and it needs to archive the file again.
This is the inverse (more or less) of the SF_ARCHIVED flag.

With UFS, the flag is just passed through and stored.

One application for this is to support CIFS servers that need to store
DOS/CIFS/Windows attributes on a FreeBSD server.

> # touch test
> # ls -alo test
> -rw-r--r--- 1 root wheel  uarch  0 Aug 28 15:46 test
> 
> This breaks 'cp -p' to tmpfs as tmpfs does not allow this flag.
> 
> # mkdir /tmp/tmpfs
> # mount -t tmpfs tmpfs /tmp/tmpfs
> # cp -f test /tmp/tmpfs
> cp: test: Operation not supported

Right.  For some filesystems, like UFS and probably tmpfs, the right answer
may be to just pass through most all of the file flags, and reject specific
flags that it doesn't support.

For some filesystems, like smbfs, the right answer seems to be to just
ignore flags that are set that aren't supported.

Other filesystems, like msdosfs, support a small number of flags and reject
any that aren't supported.

In other words, this isn't a new problem, and it would have cropped up if
you tried to copy a file with the SF_IMMUTABLE flag set from UFS to msdosfs
with cp -p.

> [...]
> >Modified: head/sys/ufs/ufs/ufs_vnops.c
> >==
> >--- head/sys/ufs/ufs/ufs_vnops.c Wed Aug 21 22:57:29 2013 (r254626)
> >+++ head/sys/ufs/ufs/ufs_vnops.c Wed Aug 21 23:04:48 2013 (r254627)
> >@@ -528,9 +528,11 @@ ufs_setattr(ap)
> > return (EINVAL);
> > }
> > if (vap->va_flags != VNOVAL) {
> >-if ((vap->va_flags & ~(UF_NODUMP | UF_IMMUTABLE | UF_APPEND |
> >-UF_OPAQUE | UF_NOUNLINK | SF_ARCHIVED | SF_IMMUTABLE |
> >-SF_APPEND | SF_NOUNLINK | SF_SNAPSHOT)) != 0)
> >+if ((vap->va_flags & ~(SF_APPEND | SF_ARCHIVED | 
> >SF_IMMUTABLE |
> >+SF_NOUNLINK | SF_SNAPSHOT | UF_APPEND | UF_ARCHIVE |
> >+UF_HIDDEN | UF_IMMUTABLE | UF_NODUMP | UF_NOUNLINK |
> >+UF_OFFLINE | UF_OPAQUE | UF_READONLY | UF_REPARSE |
> >+UF_SPARSE | UF_SYSTEM)) != 0)
> > return (EOPNOTSUPP);
> > if (vp->v_mount->mnt_flag & MNT_RDONLY)
> > return (EROFS);
> 
> Seems a similar change is needed in tmpfs_subr.c:tmpfs_chflags() 
> (antoine pointed this out)

Sure, I can fix tmpfs.

Ken
-- 
Kenneth Merry
k...@freebsd.org
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r255008 - head/sys/fs/tmpfs

[Kenneth D. Merry]

Author: ken
Date: Wed Aug 28 22:12:56 2013
New Revision: 255008
URL: http://svnweb.freebsd.org/changeset/base/255008

Log:
  Support storing 7 additional file flags in tmpfs:
  
  UF_SYSTEM, UF_SPARSE, UF_OFFLINE, UF_REPARSE, UF_ARCHIVE, UF_READONLY,
  and UF_HIDDEN.
  
  Sort the file flags tmpfs supports alphabetically.  tmpfs now
  supports the same flags as UFS, with the exception of SF_SNAPSHOT.
  
  Reported by:  bdrewery, antoine
  Sponsored by: Spectra Logic

Modified:
  head/sys/fs/tmpfs/tmpfs_subr.c

Modified: head/sys/fs/tmpfs/tmpfs_subr.c
==
--- head/sys/fs/tmpfs/tmpfs_subr.c  Wed Aug 28 21:10:37 2013
(r255007)
+++ head/sys/fs/tmpfs/tmpfs_subr.c  Wed Aug 28 22:12:56 2013
(r255008)
@@ -1433,9 +1433,10 @@ tmpfs_chflags(struct vnode *vp, u_long f
 
node = VP_TO_TMPFS_NODE(vp);
 
-   if ((flags & ~(UF_NODUMP | UF_IMMUTABLE | UF_APPEND | UF_OPAQUE |
-   UF_NOUNLINK | SF_ARCHIVED | SF_IMMUTABLE | SF_APPEND |
-   SF_NOUNLINK)) != 0)
+   if ((flags & ~(SF_APPEND | SF_ARCHIVED | SF_IMMUTABLE | SF_NOUNLINK |
+   UF_APPEND | UF_ARCHIVE | UF_HIDDEN | UF_IMMUTABLE | UF_NODUMP |
+   UF_NOUNLINK | UF_OFFLINE | UF_OPAQUE | UF_READONLY | UF_REPARSE |
+   UF_SPARSE | UF_SYSTEM)) != 0)
return (EOPNOTSUPP);
 
/* Disallow this operation if the file system is mounted read-only. */
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r255032 - head/sys/kern

[Kenneth D. Merry]

Author: ken
Date: Thu Aug 29 16:41:40 2013
New Revision: 255032
URL: http://svnweb.freebsd.org/changeset/base/255032

Log:
  Fix some issues in change 254760 pointed out by Bruce Evans:
  
   - Remove excessive parenthesis
   - Use KNF continuation indentation
   - Cut down on excessive continuation lines
   - More consistent style in messages
   - Use uprintf() instead of printf()
  
  Submitted by: bde

Modified:
  head/sys/kern/kern_physio.c

Modified: head/sys/kern/kern_physio.c
==
--- head/sys/kern/kern_physio.c Thu Aug 29 16:26:04 2013(r255031)
+++ head/sys/kern/kern_physio.c Thu Aug 29 16:41:40 2013(r255032)
@@ -58,25 +58,22 @@ physio(struct cdev *dev, struct uio *uio
 * If the driver does not want I/O to be split, that means that we
 * need to reject any requests that will not fit into one buffer.
 */
-   if ((dev->si_flags & SI_NOSPLIT) &&
-   ((uio->uio_resid > dev->si_iosize_max) ||
-(uio->uio_resid > MAXPHYS) ||
-(uio->uio_iovcnt > 1))) {
+   if (dev->si_flags & SI_NOSPLIT &&
+   (uio->uio_resid > dev->si_iosize_max || uio->uio_resid > MAXPHYS ||
+   uio->uio_iovcnt > 1)) {
/*
 * Tell the user why his I/O was rejected.
 */
if (uio->uio_resid > dev->si_iosize_max)
-   printf("%s: request size %zd > si_iosize_max=%d, "
+   uprintf("%s: request size=%zd > si_iosize_max=%d; "
"cannot split request\n", devtoname(dev),
uio->uio_resid, dev->si_iosize_max);
-
if (uio->uio_resid > MAXPHYS)
-   printf("%s: request size %zd > MAXPHYS=%d, "
+   uprintf("%s: request size=%zd > MAXPHYS=%d; "
"cannot split request\n", devtoname(dev),
uio->uio_resid, MAXPHYS);
-
if (uio->uio_iovcnt > 1)
-   printf("%s: request vectors=%d > 1, "
+   uprintf("%s: request vectors=%d > 1; "
"cannot split request\n", devtoname(dev),
uio->uio_iovcnt);
 
@@ -117,8 +114,8 @@ physio(struct cdev *dev, struct uio *uio
 * This device does not want I/O to be split.
 */
if (dev->si_flags & SI_NOSPLIT) {
-   printf("%s: request ptr %p is not "
-   "on a page boundary, cannot split "
+   uprintf("%s: request ptr %p is not "
+   "on a page boundary; cannot split "
"request\n", devtoname(dev),
bp->b_data);
error = EFBIG;
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r255043 - head/sys/cam/scsi

[Kenneth D. Merry]

Author: ken
Date: Thu Aug 29 21:25:27 2013
New Revision: 255043
URL: http://svnweb.freebsd.org/changeset/base/255043

Log:
  Bump up the default timeouts for move commands in the ch(4) driver
  to 15 minutes, and 5 minutes for things like READ ELEMENT STATUS.
  
  This is needed to account for the worst case scenarios on at least
  some Spectra Logic tape libraries.
  
  Sponsored by: Spectra Logic
  MFC after:3 days

Modified:
  head/sys/cam/scsi/scsi_ch.c

Modified: head/sys/cam/scsi/scsi_ch.c
==
--- head/sys/cam/scsi/scsi_ch.c Thu Aug 29 20:51:12 2013(r255042)
+++ head/sys/cam/scsi/scsi_ch.c Thu Aug 29 21:25:27 2013(r255043)
@@ -99,10 +99,10 @@ __FBSDID("$FreeBSD$");
  */
 
 static const u_int32_t CH_TIMEOUT_MODE_SENSE= 6000;
-static const u_int32_t CH_TIMEOUT_MOVE_MEDIUM   = 10;
-static const u_int32_t CH_TIMEOUT_EXCHANGE_MEDIUM   = 10;
-static const u_int32_t CH_TIMEOUT_POSITION_TO_ELEMENT   = 10;
-static const u_int32_t CH_TIMEOUT_READ_ELEMENT_STATUS   = 6;
+static const u_int32_t CH_TIMEOUT_MOVE_MEDIUM   = 15 * 60 * 1000;
+static const u_int32_t CH_TIMEOUT_EXCHANGE_MEDIUM   = 15 * 60 * 1000;
+static const u_int32_t CH_TIMEOUT_POSITION_TO_ELEMENT   = 15 * 60 * 1000;
+static const u_int32_t CH_TIMEOUT_READ_ELEMENT_STATUS   = 5 * 60 * 1000;
 static const u_int32_t CH_TIMEOUT_SEND_VOLTAG   = 1;
 static const u_int32_t CH_TIMEOUT_INITIALIZE_ELEMENT_STATUS = 50;
 
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

svn commit: r255501 - head/sys/dev/mps

[Kenneth D. Merry]

Author: ken
Date: Thu Sep 12 22:06:12 2013
New Revision: 255501
URL: http://svnweb.freebsd.org/changeset/base/255501

Log:
  Fix an issue that caused Integrated RAID volumes on LSI mps(4) controllers
  to not get scanned on boot.
  
  The problem originated in change 253549.  With the change to the mps(4)
  driver to scan only targets that it knows it has (as opposed to scanning
  the entire bus), scanning RAID volumes on boot was omitted.
  
  So, for versions of FreeBSD that have the scanning changes
  (__FreeBSD_version 139 and higher), scan RAID volumes that are added
  whether or not we're booting.
  
  PR:   kern/181784
  Reported by:  Xiguang Wang 
  Tested by:Dennis Glatting 
  Sponsored by: Spectra Logic
  Approved by:  re (delphij)
  MFC After:3 days

Modified:
  head/sys/dev/mps/mps_sas_lsi.c

Modified: head/sys/dev/mps/mps_sas_lsi.c
==
--- head/sys/dev/mps/mps_sas_lsi.c  Thu Sep 12 21:24:59 2013
(r255500)
+++ head/sys/dev/mps/mps_sas_lsi.c  Thu Sep 12 22:06:12 2013
(r255501)
@@ -898,7 +898,9 @@ mpssas_volume_add(struct mps_softc *sc, 
free(lun, M_MPT2);
}
SLIST_INIT(&targ->luns);
+#if __FreeBSD_version < 139
if ((sassc->flags & MPSSAS_IN_STARTUP) == 0)
+#endif
mpssas_rescan_target(sc, targ);
mps_dprint(sc, MPS_MAPPING, "RAID target id %d added (WWID = 0x%jx)\n",
targ->tid, wwid);
___
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"