svn commit: r232249 - head
Author: pluknet
Date: Tue Feb 28 11:06:52 2012
New Revision: 232249
URL: http://svn.freebsd.org/changeset/base/232249
Log:
Add lib32 part after libarchive 3.0.3 update.
Modified:
head/ObsoleteFiles.inc
Modified: head/ObsoleteFiles.inc
==
--- head/ObsoleteFiles.inc Tue Feb 28 08:36:38 2012(r232248)
+++ head/ObsoleteFiles.inc Tue Feb 28 11:06:52 2012(r232249)
@@ -56,6 +56,9 @@ OLD_FILES+=man/man3/archive_read_data_in
man/man3/archive_write_set_compression_none.3.gz \
man/man3/archive_write_set_compression_program.3.gz
OLD_LIBS+=usr/lib/libarchive.so.5
+.if ${TARGET_ARCH} == "amd64" || ${TARGET_ARCH} == "powerpc64"
+OLD_LIBS+=usr/lib32/libarchive.so.5
+.endif
# 20120113: removal of wtmpcvt(1)
OLD_FILES+=usr/bin/wtmpcvt
OLD_FILES+=usr/share/man/man1/wtmpcvt.1.gz
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r232250 - in head: sbin/ifconfig sbin/ipfw share/man/man4 sys/conf sys/dev/ath sys/dev/siba sys/ia64/ia64 sys/mips/rt305x
Author: gavin Date: Tue Feb 28 13:19:34 2012 New Revision: 232250 URL: http://svn.freebsd.org/changeset/base/232250 Log: Correct capitalization of "Hz" in user-visible text (manpages, printf(), etc). MFC after:3 days Modified: head/sbin/ifconfig/ifconfig.8 head/sbin/ipfw/ipfw.8 head/share/man/man4/ath.4 head/share/man/man4/net80211.4 head/sys/conf/NOTES head/sys/dev/ath/if_ath_sysctl.c head/sys/dev/siba/siba_core.c head/sys/ia64/ia64/machdep.c head/sys/mips/rt305x/rt305x_sysctl.c Modified: head/sbin/ifconfig/ifconfig.8 == --- head/sbin/ifconfig/ifconfig.8 Tue Feb 28 11:06:52 2012 (r232249) +++ head/sbin/ifconfig/ifconfig.8 Tue Feb 28 13:19:34 2012 (r232250) @@ -1036,7 +1036,7 @@ Enable Dynamic Frequency Selection (DFS) DFS embodies several facilities including detection of overlapping radar signals, dynamic transmit power control, and channel selection according to a least-congested criteria. -DFS support is mandatory for some 5Ghz frequencies in certain +DFS support is mandatory for some 5GHz frequencies in certain locales (e.g. ETSI). By default DFS is enabled according to the regulatory definitions specified in /etc/regdomain.xml and the current country code, regdomain, @@ -1091,7 +1091,7 @@ By default DTIM is 1 (i.e., DTIM occurs .It Cm quiet Enable the use of quiet IE. Hostap will use this to silent other stations to reduce interference for radar detection when -operating on 5Ghz frequency and doth support is enabled. +operating on 5GHz frequency and doth support is enabled. Use .Fl quiet to disable this functionality. Modified: head/sbin/ipfw/ipfw.8 == --- head/sbin/ipfw/ipfw.8 Tue Feb 28 11:06:52 2012(r232249) +++ head/sbin/ipfw/ipfw.8 Tue Feb 28 13:19:34 2012(r232250) @@ -2235,7 +2235,7 @@ specifies the scheduling algorithm to us is just a FIFO scheduler (which means that all packets are stored in the same queue as they arrive to the scheduler). FIFO has O(1) per-packet time complexity, with very low -constants (estimate 60-80ns on a 2Ghz desktop machine) +constants (estimate 60-80ns on a 2GHz desktop machine) but gives no service guarantees. .It Cm wf2qp implements the WF2Q+ algorithm, which is a Weighted Fair Queueing Modified: head/share/man/man4/ath.4 == --- head/share/man/man4/ath.4 Tue Feb 28 11:06:52 2012(r232249) +++ head/share/man/man4/ath.4 Tue Feb 28 13:19:34 2012(r232250) @@ -95,8 +95,8 @@ with transmit speeds appropriate to each AR5416-class devices are capable of 802.11n operation but are supported only in legacy modes (802.11a, 11b, 11g). Most chips also support an Atheros Turbo Mode (TM) that operates in -the 5Ghz frequency range with 2x the transmit speeds. -Some chips also support Turbo mode in the 2.4Ghz range with 802.11g +the 5GHz frequency range with 2x the transmit speeds. +Some chips also support Turbo mode in the 2.4GHz range with 802.11g though this support is not presently available due to regulatory requirements. (Note that Turbo modes are, however, only interoperable with other Atheros-based devices.) Modified: head/share/man/man4/net80211.4 == --- head/share/man/man4/net80211.4 Tue Feb 28 11:06:52 2012 (r232249) +++ head/share/man/man4/net80211.4 Tue Feb 28 13:19:34 2012 (r232250) @@ -280,7 +280,7 @@ Return whether or not Dynamic Frequency DFS embodies several facilities including detection of overlapping radar signals, dynamic transmit power control, and channel selection according to a least-congested criteria. -DFS support is mandatory for some 5Ghz frequencies in certain +DFS support is mandatory for some 5GHz frequencies in certain locales (e.g. ETSI). By default DFS is enabled according to the regulatory definitions and the current country code, regdomain, and channel. Modified: head/sys/conf/NOTES == --- head/sys/conf/NOTES Tue Feb 28 11:06:52 2012(r232249) +++ head/sys/conf/NOTES Tue Feb 28 13:19:34 2012(r232250) @@ -2417,11 +2417,11 @@ device cmx # or # options BROOKTREE_SYSTEM_DEFAULT=BROOKTREE_NTSC # Specifies the default video capture mode. -# This is required for Dual Crystal (28&35Mhz) boards where PAL is used +# This is required for Dual Crystal (28&35MHz) boards where PAL is used # to prevent hangs during initialization, e.g. VideoLogic Captivator PCI. # # options BKTR_USE_PLL -# This is required for PAL or SECAM boards with a 28Mhz crystal and no 35Mhz +# This is required for PAL or SECAM boards with a 28MHz crystal and no 35MHz # crystal, e.g. some new Bt878 cards. #
svn commit: r232253 - head
Author: pluknet
Date: Tue Feb 28 14:35:36 2012
New Revision: 232253
URL: http://svn.freebsd.org/changeset/base/232253
Log:
Correct pathname for manpages after libarchive 3.0.3 update.
Modified:
head/ObsoleteFiles.inc
Modified: head/ObsoleteFiles.inc
==
--- head/ObsoleteFiles.inc Tue Feb 28 14:22:54 2012(r232252)
+++ head/ObsoleteFiles.inc Tue Feb 28 14:35:36 2012(r232253)
@@ -39,22 +39,22 @@
# done
# 20120225: libarchive 3.0.3
-OLD_FILES+=man/man3/archive_read_data_into_buffer.3.gz \
- man/man3/archive_read_support_compression_all.3.gz \
- man/man3/archive_read_support_compression_bzip2.3.gz \
- man/man3/archive_read_support_compression_compress.3.gz \
- man/man3/archive_read_support_compression_gzip.3.gz \
- man/man3/archive_read_support_compression_lzma.3.gz \
- man/man3/archive_read_support_compression_none.3.gz \
- man/man3/archive_read_support_compression_program.3.gz \
- man/man3/archive_read_support_compression_program_signature.3.gz \
- man/man3/archive_read_support_compression_xz.3.gz \
- man/man3/archive_write_set_callbacks.3.gz \
- man/man3/archive_write_set_compression_bzip2.3.gz \
- man/man3/archive_write_set_compression_compress.3.gz \
- man/man3/archive_write_set_compression_gzip.3.gz \
- man/man3/archive_write_set_compression_none.3.gz \
- man/man3/archive_write_set_compression_program.3.gz
+OLD_FILES+=usr/share/man/man3/archive_read_data_into_buffer.3.gz \
+ usr/share/man/man3/archive_read_support_compression_all.3.gz \
+ usr/share/man/man3/archive_read_support_compression_bzip2.3.gz \
+ usr/share/man/man3/archive_read_support_compression_compress.3.gz \
+ usr/share/man/man3/archive_read_support_compression_gzip.3.gz \
+ usr/share/man/man3/archive_read_support_compression_lzma.3.gz \
+ usr/share/man/man3/archive_read_support_compression_none.3.gz \
+ usr/share/man/man3/archive_read_support_compression_program.3.gz \
+
usr/share/man/man3/archive_read_support_compression_program_signature.3.gz \
+ usr/share/man/man3/archive_read_support_compression_xz.3.gz \
+ usr/share/man/man3/archive_write_set_callbacks.3.gz \
+ usr/share/man/man3/archive_write_set_compression_bzip2.3.gz \
+ usr/share/man/man3/archive_write_set_compression_compress.3.gz \
+ usr/share/man/man3/archive_write_set_compression_gzip.3.gz \
+ usr/share/man/man3/archive_write_set_compression_none.3.gz \
+ usr/share/man/man3/archive_write_set_compression_program.3.gz
OLD_LIBS+=usr/lib/libarchive.so.5
.if ${TARGET_ARCH} == "amd64" || ${TARGET_ARCH} == "powerpc64"
OLD_LIBS+=usr/lib32/libarchive.so.5
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r232254 - in head: share/man/man5 share/man/man9 sys/netatalk
Author: kevlo
Date: Tue Feb 28 15:07:05 2012
New Revision: 232254
URL: http://svn.freebsd.org/changeset/base/232254
Log:
Fix typos
Modified:
head/share/man/man5/passwd.5
head/share/man/man9/byteorder.9
head/sys/netatalk/aarp.c
Modified: head/share/man/man5/passwd.5
==
--- head/share/man/man5/passwd.5Tue Feb 28 14:35:36 2012
(r232253)
+++ head/share/man/man5/passwd.5Tue Feb 28 15:07:05 2012
(r232254)
@@ -232,7 +232,7 @@ user's home phone number
.Pp
The full
.Ar name
-may contain a ampersand
+may contain an ampersand
.Pq Ql &
which will be replaced by
the capitalized login
Modified: head/share/man/man9/byteorder.9
==
--- head/share/man/man9/byteorder.9 Tue Feb 28 14:35:36 2012
(r232253)
+++ head/share/man/man9/byteorder.9 Tue Feb 28 15:07:05 2012
(r232254)
@@ -123,7 +123,7 @@ The
.Fn htobe32 ,
and
.Fn htobe64
-functions return a integer in the system's native
+functions return an integer in the system's native
byte order converted to big endian byte order.
The return value will be the same as the argument on big endian systems.
.Pp
Modified: head/sys/netatalk/aarp.c
==
--- head/sys/netatalk/aarp.cTue Feb 28 14:35:36 2012(r232253)
+++ head/sys/netatalk/aarp.cTue Feb 28 15:07:05 2012(r232254)
@@ -467,7 +467,7 @@ at_aarpinput(struct ifnet *ifp, struct m
if (aat != NULL) {
if (op == AARPOP_PROBE) {
/*
-* Someone's probing for spa, dealocate the one we've
+* Someone's probing for spa, deallocate the one we've
* got, so that if the prober keeps the address,
* we'll be able to arp for him.
*/
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r232255 - head/sys/dev/amr
Author: kevlo Date: Tue Feb 28 15:09:56 2012 New Revision: 232255 URL: http://svn.freebsd.org/changeset/base/232255 Log: Fix checks for error return from amr_sglist_map() and amr_ccb_map() Modified: head/sys/dev/amr/amr_pci.c Modified: head/sys/dev/amr/amr_pci.c == --- head/sys/dev/amr/amr_pci.c Tue Feb 28 15:07:05 2012(r232254) +++ head/sys/dev/amr/amr_pci.c Tue Feb 28 15:09:56 2012(r232255) @@ -339,11 +339,11 @@ amr_pci_attach(device_t dev) /* * Build the scatter/gather buffers. */ -if (amr_sglist_map(sc)) +if ((error = amr_sglist_map(sc)) != 0) goto out; debug(2, "s/g list mapped"); -if (amr_ccb_map(sc)) +if ((error = amr_ccb_map(sc)) != 0) goto out; debug(2, "ccb mapped"); ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r232256 - head/sys/dev/acpi_support
Author: kevlo
Date: Tue Feb 28 15:12:26 2012
New Revision: 232256
URL: http://svn.freebsd.org/changeset/base/232256
Log:
Add missing MODULE_DEPEND() so that acpi.ko and aibs.ko can be
loaded dynamically.
Modified:
head/sys/dev/acpi_support/atk0110.c
Modified: head/sys/dev/acpi_support/atk0110.c
==
--- head/sys/dev/acpi_support/atk0110.c Tue Feb 28 15:09:56 2012
(r232255)
+++ head/sys/dev/acpi_support/atk0110.c Tue Feb 28 15:12:26 2012
(r232256)
@@ -97,7 +97,7 @@ static driver_t aibs_driver = {
static devclass_t aibs_devclass;
DRIVER_MODULE(aibs, acpi, aibs_driver, aibs_devclass, NULL, NULL);
-
+MODULE_DEPEND(aibs, acpi, 1, 1, 1);
static char* aibs_hids[] = {
"ATK0110",
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r232181 - in head/sys: kern sys
On Tuesday, February 28, 2012 1:34:25 am Julian Elischer wrote: > On 2/27/12 1:29 AM, Konstantin Belousov wrote: > > On Mon, Feb 27, 2012 at 10:49:59AM +0200, Mikolaj Golub wrote: > >> On Mon, 27 Feb 2012 09:28:11 +0100 Pawel Jakub Dawidek wrote: > >> > >> PJD> On Sun, Feb 26, 2012 at 02:25:48PM +, Mikolaj Golub wrote: > >> >> Author: trociny > >> >> Date: Sun Feb 26 14:25:48 2012 > >> >> New Revision: 232181 > >> >> URL: http://svn.freebsd.org/changeset/base/232181 > >> >> > >> >> Log: > >> >>Add sysctl to retrieve or set umask of another process. > >> > >> PJD> "set umask of another process"? This seems... weird. What's the > >> purpose > >> PJD> of this change? > >> > >> When we were discussing this with Kostik and Robert, and I asked if it > >> could > >> be useful to have the sysctl rw, Kostik described a real situation when he > >> had > >> had to change umask of another process: umask had not been set properly on > >> an > >> aplication start but it could not be restarted until the end of the day. > >> Kostik was able to fix it using gdb but having an easier way looked useful. > > kgdb, not gdb. > > > > It is indeed possible to write a ptrace-based utility that inject a code > > payload that would change umask. Since this is very risky but indeed > > possible, > > having the straighforward kernel facility is justified. > Why not have a sysctl to change a process' uid, cwd, memory limits, > etc. etc. uid and cwd would be rediculous to change. However, we recently added sysctls to allow a sysadmin to read and write the limits of other processes (and that is a very useful feature indeed since it is not unusual for a long-running process to require more resources than it was initially allocated, as is the ability to easily query the limits that a given process is subject to). > I don't think this belongs in the kernel by default. It's not exactl a > call for backout but It's teh next thing short of that. a call for "do > you REALLY think we need this particular specific case catered for?" That said, the umask bit does strike me as a bit more odd than the limits case. I would have more need of a tool to let me adjust the listen queue length of a socket than to adjust umask (I've had to do that multiple times via kgdb). -- John Baldwin ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r232257 - in head: share/man/man4 sys/dev/usb sys/dev/usb/net
Author: kevlo
Date: Tue Feb 28 15:45:42 2012
New Revision: 232257
URL: http://svn.freebsd.org/changeset/base/232257
Log:
Add support for the MCS7832
Obtained from:OpenBSD
Modified:
head/share/man/man4/mos.4
head/sys/dev/usb/net/if_mos.c
head/sys/dev/usb/net/if_mosreg.h
head/sys/dev/usb/usbdevs
Modified: head/share/man/man4/mos.4
==
--- head/share/man/man4/mos.4 Tue Feb 28 15:12:26 2012(r232256)
+++ head/share/man/man4/mos.4 Tue Feb 28 15:45:42 2012(r232257)
@@ -20,7 +20,7 @@
.Os
.Sh NAME
.Nm mos
-.Nd Moschip MCS7730/MCS7840 USB Ethernet driver
+.Nd Moschip MCS7730/MCS7830/MCS7832 USB Ethernet driver
.Sh SYNOPSIS
To compile this driver into the kernel,
place the following lines in your
@@ -44,9 +44,9 @@ if_mos_load="YES"
The
.Nm
driver provides support for USB Ethernet adapters based on the
-Moschip MCS7730/MCS7830 chipset.
+Moschip MCS7730/MCS7830/MCS7832 chipset.
.Pp
-The adapters that contain the Moschip MCS7730/MCS7830 chipset
+The adapters that contain the Moschip MCS7730/MCS7830/MCS7832 chipset
will operate at 100Base-TX and full-duplex.
.Pp
The Moschip contains a 10/100
Modified: head/sys/dev/usb/net/if_mos.c
==
--- head/sys/dev/usb/net/if_mos.c Tue Feb 28 15:12:26 2012
(r232256)
+++ head/sys/dev/usb/net/if_mos.c Tue Feb 28 15:45:42 2012
(r232257)
@@ -82,7 +82,7 @@
__FBSDID("$FreeBSD$");
/*
- * Moschip MCS7730/MCS7830 USB to Ethernet controller
+ * Moschip MCS7730/MCS7830/MCS7832 USB to Ethernet controller
* The datasheet is available at the following URL:
* http://www.moschip.com/data/products/MCS7830/Data%20Sheet_7830.pdf
*/
@@ -149,6 +149,7 @@ SYSCTL_INT(_hw_usb_mos, OID_AUTO, debug,
static const STRUCT_USB_HOST_ID mos_devs[] = {
{USB_VPI(USB_VENDOR_MOSCHIP, USB_PRODUCT_MOSCHIP_MCS7730, MCS7730)},
{USB_VPI(USB_VENDOR_MOSCHIP, USB_PRODUCT_MOSCHIP_MCS7830, MCS7830)},
+ {USB_VPI(USB_VENDOR_MOSCHIP, USB_PRODUCT_MOSCHIP_MCS7832, MCS7832)},
{USB_VPI(USB_VENDOR_SITECOMEU, USB_PRODUCT_SITECOMEU_LN030, MCS7830)},
};
@@ -715,6 +716,8 @@ mos_attach(device_t dev)
MOS_DPRINTFN("model: MCS7730");
} else if (sc->mos_flags & MCS7830) {
MOS_DPRINTFN("model: MCS7830");
+ } else if (sc->mos_flags & MCS7832) {
+ MOS_DPRINTFN("model: MCS7832");
}
error = uether_ifattach(ue);
if (error) {
Modified: head/sys/dev/usb/net/if_mosreg.h
==
--- head/sys/dev/usb/net/if_mosreg.hTue Feb 28 15:12:26 2012
(r232256)
+++ head/sys/dev/usb/net/if_mosreg.hTue Feb 28 15:45:42 2012
(r232257)
@@ -152,6 +152,7 @@
#defineMCS7730 0x0001
#defineMCS7830 0x0002
+#defineMCS7832 0x0004
#defineMOS_INC(x, y) (x) = (x + 1) % y
Modified: head/sys/dev/usb/usbdevs
==
--- head/sys/dev/usb/usbdevsTue Feb 28 15:12:26 2012(r232256)
+++ head/sys/dev/usb/usbdevsTue Feb 28 15:45:42 2012(r232257)
@@ -2306,6 +2306,7 @@ product MOSCHIP MCS7703 0x7703 MCS7703
product MOSCHIP MCS77300x7730 MCS7730 Ethernet
product MOSCHIP MCS78200x7820 MCS7820 Serial Port Adapter
product MOSCHIP MCS78300x7830 MCS7830 Ethernet
+product MOSCHIP MCS78320x7832 MCS7832 Ethernet
product MOSCHIP MCS78400x7840 MCS7840 Serial Port Adapter
/* Motorola products */
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r232258 - head/share/man/man4
Author: kevlo Date: Tue Feb 28 15:47:39 2012 New Revision: 232258 URL: http://svn.freebsd.org/changeset/base/232258 Log: Mention mos(4) Modified: head/share/man/man4/usb.4 Modified: head/share/man/man4/usb.4 == --- head/share/man/man4/usb.4 Tue Feb 28 15:45:42 2012(r232257) +++ head/share/man/man4/usb.4 Tue Feb 28 15:47:39 2012(r232258) @@ -150,6 +150,7 @@ specifications can be found at: .Xr cue 4 , .Xr ehci 4 , .Xr kue 4 , +.Xr mos 4 , .Xr ohci 4 , .Xr pci 4 , .Xr rue 4 , ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r232260 - head/sys/dev/aac
Author: emaste Date: Tue Feb 28 17:29:31 2012 New Revision: 232260 URL: http://svn.freebsd.org/changeset/base/232260 Log: Avoid transfers crossing a 4GB boundary, which can lead to data corruption. Thanks to scottl@ for the suggestion. This change will likely be revised after consideration of a general method to address this type of issue for other drivers. Sponsored by: Sandvine Incorporated MFC after: 3 days Modified: head/sys/dev/aac/aac_pci.c Modified: head/sys/dev/aac/aac_pci.c == --- head/sys/dev/aac/aac_pci.c Tue Feb 28 15:52:01 2012(r232259) +++ head/sys/dev/aac/aac_pci.c Tue Feb 28 17:29:31 2012(r232260) @@ -403,7 +403,8 @@ aac_pci_attach(device_t dev) * Note that some of these controllers are 64-bit capable. */ if (bus_dma_tag_create(NULL,/* parent */ - PAGE_SIZE, 0,/* algnmnt, boundary */ + PAGE_SIZE, /* alignment */ + ((bus_size_t)((uint64_t)1 << 32)), /* boundary*/ BUS_SPACE_MAXADDR, /* lowaddr */ BUS_SPACE_MAXADDR, /* highaddr */ NULL, NULL, /* filter, filterarg */ ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r232261 - in head/sys: amd64/include i386/include pc98/include x86/include
Author: tijl Date: Tue Feb 28 18:15:28 2012 New Revision: 232261 URL: http://svn.freebsd.org/changeset/base/232261 Log: Copy amd64 _types.h to x86 and merge with i386 _types.h. Replace existing amd64/i386/pc98 _types.h with stubs. Added: head/sys/x86/include/_types.h - copied, changed from r232259, head/sys/amd64/include/_types.h Modified: head/sys/amd64/include/_types.h head/sys/i386/include/_types.h head/sys/pc98/include/_types.h Modified: head/sys/amd64/include/_types.h == --- head/sys/amd64/include/_types.h Tue Feb 28 17:29:31 2012 (r232260) +++ head/sys/amd64/include/_types.h Tue Feb 28 18:15:28 2012 (r232261) @@ -1,116 +1,6 @@ /*- - * Copyright (c) 2002 Mike Barcroft - * Copyright (c) 1990, 1993 - * The Regents of the University of California. 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. - * 2. Redistributions in binary form must reproduce the above copyright - *notice, this list of conditions and the following disclaimer in the - *documentation and/or other materials provided with the distribution. - * 3. All advertising materials mentioning features or use of this software - *must display the following acknowledgement: - * This product includes software developed by the University of - * California, Berkeley and its contributors. - * 4. Neither the name of the University nor the names of its contributors - *may be used to endorse or promote products derived from this software - *without specific prior written permission. - * - * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND - * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE - * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE - * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE - * FOR ANY DIRECT, INDIRECT, INCIDENTAL, 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 DAMAGE. - * - * From: @(#)ansi.h8.2 (Berkeley) 1/4/94 - * From: @(#)types.h 8.3 (Berkeley) 1/5/94 - * $FreeBSD$ + * This file is in the public domain. */ +/* $FreeBSD$ */ -#ifndef _MACHINE__TYPES_H_ -#define_MACHINE__TYPES_H_ - -#ifndef _SYS_CDEFS_H_ -#error this file needs sys/cdefs.h as a prerequisite -#endif - -#define __NO_STRICT_ALIGNMENT - -/* - * Basic types upon which most other types are built. - */ -typedefsigned char __int8_t; -typedefunsigned char __uint8_t; -typedefshort __int16_t; -typedefunsigned short __uint16_t; -typedefint __int32_t; -typedefunsigned int__uint32_t; -typedeflong__int64_t; -typedefunsigned long __uint64_t; - -/* - * Standard type definitions. - */ -typedef__int32_t __clock_t; /* clock()... */ -typedef__int64_t __critical_t; -typedefdouble __double_t; -typedeffloat __float_t; -typedef__int64_t __intfptr_t; -typedef__int64_t __intmax_t; -typedef__int64_t __intptr_t; -typedef__int32_t __int_fast8_t; -typedef__int32_t __int_fast16_t; -typedef__int32_t __int_fast32_t; -typedef__int64_t __int_fast64_t; -typedef__int8_t__int_least8_t; -typedef__int16_t __int_least16_t; -typedef__int32_t __int_least32_t; -typedef__int64_t __int_least64_t; -typedef__int64_t __ptrdiff_t;/* ptr1 - ptr2 */ -typedef__int64_t __register_t; -typedef__int64_t __segsz_t; /* segment size (in pages) */ -typedef__uint64_t __size_t; /* sizeof() */ -typedef__int64_t __ssize_t; /* byte count or error */ -typedef__int64_t __time_t; /* time()... */ -typedef__uint64_t __uintfptr_t; -typedef__uint64_t __uintmax_t; -typedef__uint64_t __uintptr_t; -typedef__uint32_t __uint_fast8_t; -typedef__uint32_t __uint_fast16_t; -typedef__uint32_t __uint_fast32_t;
svn commit: r232262 - in head/sys: amd64/include i386/include pc98/include x86/include
Author: tijl Date: Tue Feb 28 18:24:28 2012 New Revision: 232262 URL: http://svn.freebsd.org/changeset/base/232262 Log: Copy amd64 _limits.h to x86 and merge with i386 _limits.h. Replace amd64/i386/pc98 _limits.h with stubs. Added: head/sys/x86/include/_limits.h - copied, changed from r232259, head/sys/amd64/include/_limits.h Modified: head/sys/amd64/include/_limits.h head/sys/i386/include/_limits.h head/sys/pc98/include/_limits.h Modified: head/sys/amd64/include/_limits.h == --- head/sys/amd64/include/_limits.hTue Feb 28 18:15:28 2012 (r232261) +++ head/sys/amd64/include/_limits.hTue Feb 28 18:24:28 2012 (r232262) @@ -1,87 +1,6 @@ /*- - * Copyright (c) 1988, 1993 - * The Regents of the University of California. 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. - * 2. Redistributions in binary form must reproduce the above copyright - *notice, this list of conditions and the following disclaimer in the - *documentation and/or other materials provided with the distribution. - * 4. Neither the name of the University nor the names of its contributors - *may be used to endorse or promote products derived from this software - *without specific prior written permission. - * - * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND - * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE - * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE - * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE - * FOR ANY DIRECT, INDIRECT, INCIDENTAL, 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 DAMAGE. - * - * @(#)limits.h8.3 (Berkeley) 1/4/94 - * $FreeBSD$ + * This file is in the public domain. */ +/* $FreeBSD$ */ -#ifndef_MACHINE__LIMITS_H_ -#define_MACHINE__LIMITS_H_ - -/* - * According to ANSI (section 2.2.4.2), the values below must be usable by - * #if preprocessing directives. Additionally, the expression must have the - * same type as would an expression that is an object of the corresponding - * type converted according to the integral promotions. The subtraction for - * INT_MIN, etc., is so the value is not unsigned; e.g., 0x8000 is an - * unsigned int for 32-bit two's complement ANSI compilers (section 3.1.3.2). - */ - -#define__CHAR_BIT 8 /* number of bits in a char */ - -#define__SCHAR_MAX 0x7f/* max value for a signed char */ -#define__SCHAR_MIN (-0x7f - 1) /* min value for a signed char */ - -#define__UCHAR_MAX 0xff/* max value for an unsigned char */ - -#define__USHRT_MAX 0x /* max value for an unsigned short */ -#define__SHRT_MAX 0x7fff /* max value for a short */ -#define__SHRT_MIN (-0x7fff - 1) /* min value for a short */ - -#define__UINT_MAX 0x /* max value for an unsigned int */ -#define__INT_MAX 0x7fff /* max value for an int */ -#define__INT_MIN (-0x7fff - 1) /* min value for an int */ - -#define__ULONG_MAX 0x /* max for an unsigned long */ -#define__LONG_MAX 0x7fff /* max for a long */ -#define__LONG_MIN (-0x7fff - 1) /* min for a long */ - - /* max value for an unsigned long long */ -#define__ULLONG_MAX0xULL -#define__LLONG_MAX 0x7fffLL/* max value for a long long */ -#define__LLONG_MIN (-0x7fffLL - 1) /* min for a long long */ - -#define__SSIZE_MAX __LONG_MAX /* max value for a ssize_t */ - -#define__SIZE_T_MAX__ULONG_MAX /* max value for a size_t */ - -#define__OFF_MAX __LONG_MAX /* max value for an off_t */ -#define__OFF_MIN __LONG_MIN /* min value for an off_t */ - -/* Quads and longs are the same on the amd64. Ensure they stay in sync. */ -#define__UQUAD_MAX __ULONG_MAX /* max value for a uquad_t */ -#define__QUAD_MAX __LONG_MAX /* max value for a quad_t */ -#define__QUAD_MIN
svn commit: r232263 - in head: lib/libsm lib/libthread_db lib/libz libexec/mail.local libexec/smrsh sbin/fsdb share/mk sys/boot/i386/boot0 sys/boot/i386/boot2 sys/boot/i386/btx/btx sys/boot/i386/bt...
Author: dim
Date: Tue Feb 28 18:30:18 2012
New Revision: 232263
URL: http://svn.freebsd.org/changeset/base/232263
Log:
Define several extra macros in bsd.sys.mk and sys/conf/kern.pre.mk, to
get rid of testing explicitly for clang (using ${CC:T:Mclang}) in
individual Makefiles.
Instead, use the following extra macros, for use with clang:
- NO_WERROR.clang (disables -Werror)
- NO_WCAST_ALIGN.clang (disables -Wcast-align)
- NO_WFORMAT.clang(disables -Wformat and friends)
- CLANG_NO_IAS(disables integrated assembler)
- CLANG_OPT_SMALL (adds flags for extra small size optimizations)
As a side effect, this enables setting CC/CXX/CPP in src.conf instead of
make.conf! For clang, use the following:
CC=clang
CXX=clang++
CPP=clang-cpp
MFC after:2 weeks
Modified:
head/lib/libsm/Makefile
head/lib/libthread_db/Makefile
head/lib/libz/Makefile
head/libexec/mail.local/Makefile
head/libexec/smrsh/Makefile
head/sbin/fsdb/Makefile
head/share/mk/bsd.sys.mk
head/sys/boot/i386/boot0/Makefile
head/sys/boot/i386/boot2/Makefile
head/sys/boot/i386/btx/btx/Makefile
head/sys/boot/i386/btx/btxldr/Makefile
head/sys/boot/i386/gptboot/Makefile
head/sys/boot/i386/gptzfsboot/Makefile
head/sys/boot/i386/libi386/Makefile
head/sys/boot/i386/pxeldr/Makefile
head/sys/boot/i386/zfsboot/Makefile
head/sys/boot/pc98/btx/btx/Makefile
head/sys/boot/pc98/btx/btxldr/Makefile
head/sys/conf/Makefile.amd64
head/sys/conf/Makefile.i386
head/sys/conf/kern.pre.mk
head/sys/modules/bios/smapi/Makefile
head/sys/modules/linux/Makefile
head/usr.bin/netstat/Makefile
head/usr.bin/vacation/Makefile
head/usr.sbin/sendmail/Makefile
Modified: head/lib/libsm/Makefile
==
--- head/lib/libsm/Makefile Tue Feb 28 18:24:28 2012(r232262)
+++ head/lib/libsm/Makefile Tue Feb 28 18:30:18 2012(r232263)
@@ -18,12 +18,10 @@ CFLAGS+=${SENDMAIL_CFLAGS}
WARNS?=2
-.if ${CC:T:Mclang} == "clang"
# Unfortunately, clang gives warnings about sendmail code that cannot
# be turned off yet. Since this is contrib code, and we don't really
# care about the warnings, just make them non-fatal for now.
-NO_WERROR=
-.endif
+NO_WERROR.clang=
LIB= sm
Modified: head/lib/libthread_db/Makefile
==
--- head/lib/libthread_db/Makefile Tue Feb 28 18:24:28 2012
(r232262)
+++ head/lib/libthread_db/Makefile Tue Feb 28 18:30:18 2012
(r232263)
@@ -16,10 +16,8 @@ SYM_MAPS+=${.CURDIR}/Symbol.map
SYMBOL_MAPS=${SYM_MAPS}
VERSION_DEF=${.CURDIR}/../libc/Versions.def
-.if ${CC:T:Mclang} == "clang"
# Unfortunately, clang gives an incorrect warning about alignment in
# arch/i386/libpthread_md.c, so turn that off for now.
-NO_WCAST_ALIGN=
-.endif
+NO_WCAST_ALIGN.clang=
.include
Modified: head/lib/libz/Makefile
==
--- head/lib/libz/Makefile Tue Feb 28 18:24:28 2012(r232262)
+++ head/lib/libz/Makefile Tue Feb 28 18:30:18 2012(r232263)
@@ -47,10 +47,6 @@ ACFLAGS+=-Wa,--noexecstack
#SRCS+=gvmat64.S
#CFLAGS+= -DASMV -DNO_UNDERLINE
#ACFLAGS+= -Wa,--noexecstack
-#.if ${CC:T:Mclang} == "clang"
-## XXX: clang integrated-as doesn't grok .intel_syntax directives yet
-#ACFLAGS+= ${.IMPSRC:T:Mgvmat64.S:C/^.+$/-no-integrated-as/}
-#.endif
#.endif
VERSION_DEF= ${.CURDIR}/Versions.def
@@ -71,3 +67,7 @@ test: example minigzip
echo hello world | ./minigzip | ./minigzip -d )
.include
+
+## XXX: clang integrated-as doesn't grok .intel_syntax directives yet
+#ACFLAGS.gvmat64.S=${CLANG_NO_IAS}
+#ACFLAGS+= ${ACFLAGS.${.IMPSRC:T}}
Modified: head/libexec/mail.local/Makefile
==
--- head/libexec/mail.local/MakefileTue Feb 28 18:24:28 2012
(r232262)
+++ head/libexec/mail.local/MakefileTue Feb 28 18:30:18 2012
(r232263)
@@ -12,12 +12,10 @@ CFLAGS+=-I${SENDMAIL_DIR}/include -I.
WARNS?=2
WFORMAT=0
-.if ${CC:T:Mclang} == "clang"
# Unfortunately, clang gives warnings about sendmail code that cannot
# be turned off yet. Since this is contrib code, and we don't really
# care about the warnings, just make them non-fatal for now.
-NO_WERROR=
-.endif
+NO_WERROR.clang=
LIBSMDIR= ${.OBJDIR}/../../lib/libsm
LIBSM= ${LIBSMDIR}/libsm.a
Modified: head/libexec/smrsh/Makefile
==
--- head/libexec/smrsh/Makefile Tue Feb 28 18:24:28 2012(r232262)
+++ head/libexec/smrsh/Makefile Tue Feb 28 18:30:18 2012(r232263)
@@ -17,12 +17,10 @@ LDADD= ${LIBSM}
WARNS?=2
-.if ${CC:T:Mclang} == "clang"
# Unfor
svn commit: r232264 - in head/sys: amd64/include i386/include pc98/include x86/include
Author: tijl Date: Tue Feb 28 18:38:33 2012 New Revision: 232264 URL: http://svn.freebsd.org/changeset/base/232264 Log: Copy amd64 _stdint.h to x86 and merge with i386 _stdint.h. Replace amd64/i386/pc98 _stdint.h with stubs. Added: head/sys/x86/include/_stdint.h - copied, changed from r232259, head/sys/amd64/include/_stdint.h Modified: head/sys/amd64/include/_stdint.h head/sys/i386/include/_stdint.h head/sys/pc98/include/_stdint.h Modified: head/sys/amd64/include/_stdint.h == --- head/sys/amd64/include/_stdint.hTue Feb 28 18:30:18 2012 (r232263) +++ head/sys/amd64/include/_stdint.hTue Feb 28 18:38:33 2012 (r232264) @@ -1,171 +1,6 @@ /*- - * Copyright (c) 2001, 2002 Mike Barcroft - * Copyright (c) 2001 The NetBSD Foundation, Inc. - * All rights reserved. - * - * This code is derived from software contributed to The NetBSD Foundation - * by Klaus Klein. - * - * 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. - * 2. Redistributions in binary form must reproduce the above copyright - *notice, this list of conditions and the following disclaimer in the - *documentation and/or other materials provided with the distribution. - * 3. All advertising materials mentioning features or use of this software - *must display the following acknowledgement: - *This product includes software developed by the NetBSD - *Foundation, Inc. and its contributors. - * 4. Neither the name of The NetBSD Foundation nor the names of its - *contributors may be used to endorse or promote products derived - *from this software without specific prior written permission. - * - * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS - * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED - * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR - * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS - * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, 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 DAMAGE. - * - * $FreeBSD$ + * This file is in the public domain. */ +/* $FreeBSD$ */ -#ifndef _MACHINE__STDINT_H_ -#define_MACHINE__STDINT_H_ - -#if !defined(__cplusplus) || defined(__STDC_CONSTANT_MACROS) - -#defineINT8_C(c) (c) -#defineINT16_C(c) (c) -#defineINT32_C(c) (c) -#defineINT64_C(c) (c ## L) - -#defineUINT8_C(c) (c) -#defineUINT16_C(c) (c) -#defineUINT32_C(c) (c ## U) -#defineUINT64_C(c) (c ## UL) - -#defineINTMAX_C(c) INT64_C(c) -#defineUINTMAX_C(c)UINT64_C(c) - -#endif /* !defined(__cplusplus) || defined(__STDC_CONSTANT_MACROS) */ - -#if !defined(__cplusplus) || defined(__STDC_LIMIT_MACROS) - -/* - * ISO/IEC 9899:1999 - * 7.18.2.1 Limits of exact-width integer types - */ -/* Minimum values of exact-width signed integer types. */ -#defineINT8_MIN(-0x7f-1) -#defineINT16_MIN (-0x7fff-1) -#defineINT32_MIN (-0x7fff-1) -#defineINT64_MIN (-0x7fffL-1) - -/* Maximum values of exact-width signed integer types. */ -#defineINT8_MAX0x7f -#defineINT16_MAX 0x7fff -#defineINT32_MAX 0x7fff -#defineINT64_MAX 0x7fffL - -/* Maximum values of exact-width unsigned integer types. */ -#defineUINT8_MAX 0xff -#defineUINT16_MAX 0x -#defineUINT32_MAX 0xU -#defineUINT64_MAX 0xUL - -/* - * ISO/IEC 9899:1999 - * 7.18.2.2 Limits of minimum-width integer types - */ -/* Minimum values of minimum-width signed integer types. */ -#defineINT_LEAST8_MIN INT8_MIN -#defineINT_LEAST16_MIN INT16_MIN -#defineINT_LEAST32_MIN INT32_MIN -#defineINT_LEAST64_MIN INT64_MIN - -/* Maximum values of minimum-width signed integer types. */ -#defineINT_LEAST8_MAX INT8_MAX -#defineINT_LEAST16_MAX INT16_MAX -#defineINT_LEAST32_MAX INT32_MAX -#defineINT_LEAST64_MAX INT64_MAX - -/* Maximum values of minimum-width unsigned integer types. */ -
svn commit: r232265 - head/sys/kern
Author: jhb
Date: Tue Feb 28 19:16:02 2012
New Revision: 232265
URL: http://svn.freebsd.org/changeset/base/232265
Log:
Properly clear a device's devclass if DEVICE_ATTACH() fails if the device
does not have a fixed devclass.
Reviewed by: imp
MFC after:2 weeks
Modified:
head/sys/kern/subr_bus.c
Modified: head/sys/kern/subr_bus.c
==
--- head/sys/kern/subr_bus.cTue Feb 28 18:38:33 2012(r232264)
+++ head/sys/kern/subr_bus.cTue Feb 28 19:16:02 2012(r232265)
@@ -2732,9 +2732,8 @@ device_attach(device_t dev)
if ((error = DEVICE_ATTACH(dev)) != 0) {
printf("device_attach: %s%d attach returned %d\n",
dev->driver->name, dev->unit, error);
- /* Unset the class; set in device_probe_child */
- if (dev->devclass == NULL)
- (void)device_set_devclass(dev, NULL);
+ if (!(dev->flags & DF_FIXEDCLASS))
+ devclass_delete_device(dev->devclass, dev);
(void)device_set_driver(dev, NULL);
device_sysctl_fini(dev);
dev->state = DS_NOTPRESENT;
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r232266 - in head/sys: amd64/include i386/include pc98/include x86/include
Author: tijl
Date: Tue Feb 28 19:39:54 2012
New Revision: 232266
URL: http://svn.freebsd.org/changeset/base/232266
Log:
Copy amd64 endian.h to x86 and merge with i386 endian.h. Replace
amd64/i386/pc98 endian.h with stubs.
In __bswap64_const(x) the conflict between 0xffUL and 0xffULL has been
resolved by reimplementing the macro in terms of __bswap32(x). As a side
effect __bswap64_var(x) is now implemented using two bswap instructions on
i386 and should be much faster. __bswap32_const(x) has been reimplemented
in terms of __bswap16(x) for consistency.
Added:
head/sys/x86/include/endian.h
- copied, changed from r232259, head/sys/amd64/include/endian.h
Modified:
head/sys/amd64/include/endian.h
head/sys/i386/include/endian.h
head/sys/pc98/include/endian.h
Modified: head/sys/amd64/include/endian.h
==
--- head/sys/amd64/include/endian.h Tue Feb 28 19:16:02 2012
(r232265)
+++ head/sys/amd64/include/endian.h Tue Feb 28 19:39:54 2012
(r232266)
@@ -1,145 +1,6 @@
/*-
- * Copyright (c) 1987, 1991 Regents of the University of California.
- * 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.
- * 2. Redistributions in binary form must reproduce the above copyright
- *notice, this list of conditions and the following disclaimer in the
- *documentation and/or other materials provided with the distribution.
- * 4. Neither the name of the University nor the names of its contributors
- *may be used to endorse or promote products derived from this software
- *without specific prior written permission.
- *
- * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, 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 DAMAGE.
- *
- * @(#)endian.h7.8 (Berkeley) 4/3/91
- * $FreeBSD$
+ * This file is in the public domain.
*/
+/* $FreeBSD$ */
-#ifndef _MACHINE_ENDIAN_H_
-#define_MACHINE_ENDIAN_H_
-
-#include
-#include
-
-#ifdef __cplusplus
-extern "C" {
-#endif
-
-/*
- * Define the order of 32-bit words in 64-bit words.
- */
-#define_QUAD_HIGHWORD 1
-#define_QUAD_LOWWORD 0
-
-/*
- * Definitions for byte order, according to byte significance from low
- * address to high.
- */
-#define_LITTLE_ENDIAN 1234/* LSB first: i386, vax */
-#define_BIG_ENDIAN 4321/* MSB first: 68000, ibm, net */
-#define_PDP_ENDIAN 3412/* LSB first in word, MSW first in long
*/
-
-#define_BYTE_ORDER _LITTLE_ENDIAN
-
-/*
- * Deprecated variants that don't have enough underscores to be useful in more
- * strict namespaces.
- */
-#if __BSD_VISIBLE
-#defineLITTLE_ENDIAN _LITTLE_ENDIAN
-#defineBIG_ENDIAN _BIG_ENDIAN
-#definePDP_ENDIAN _PDP_ENDIAN
-#defineBYTE_ORDER _BYTE_ORDER
-#endif
-
-#if defined(__GNUCLIKE_ASM) && defined(__GNUCLIKE_BUILTIN_CONSTANT_P)
-
-#define__bswap64_const(_x) \
- (((_x) >> 56) | \
- (((_x) >> 40) & (0xffUL << 8)) |\
- (((_x) >> 24) & (0xffUL << 16)) | \
- (((_x) >> 8) & (0xffUL << 24)) |\
- (((_x) << 8) & (0xffUL << 32)) |\
- (((_x) << 24) & (0xffUL << 40)) | \
- (((_x) << 40) & (0xffUL << 48)) | \
- ((_x) << 56))
-
-#define__bswap32_const(_x) \
- (((_x) >> 24) | \
- (((_x) & (0xff << 16)) >> 8) | \
- (((_x) & (0xff << 8)) << 8) | \
- ((_x) << 24))
-
-#define __bswap16_const(_x)(__uint16_t)((_x) << 8 | (_x) >> 8)
-
-static __inline __uint64_t
-__bswap64_var(__uint64_t _x)
-{
-
- __asm ("bswap %0" : "+r" (_x));
- return (_x);
-}
-
-static __inline __uint32_t
-__bswap32_var(__uint32_t _x)
-{
-
- __asm ("bswap %0" : "+r" (_x));
- return (_x);
-}
-
-static __inline __uint16_t
-__bswap16_var(__uint16_t _x)
-{
-
- return (__bswap16_const(_x));
-}
-
-#define__bswap64(_x
svn commit: r232267 - head/sys/x86/x86
Author: emaste Date: Tue Feb 28 19:42:40 2012 New Revision: 232267 URL: http://svn.freebsd.org/changeset/base/232267 Log: Workaround for PCIe 4GB boundary issue Enforce a boundary of no more than 4GB - transfers crossing a 4GB boundary can lead to data corruption due to PCIe limitations. This change is a less-intrusive workaround that can be quickly merged back to older branches; a cleaner implementation will arrive in HEAD later but may require KPI changes. This change is based on a suggestion by jhb@. Reviewed by:scottl, jhb Sponsored by: Sandvine Incorporated MFC after: 3 days Modified: head/sys/x86/x86/busdma_machdep.c Modified: head/sys/x86/x86/busdma_machdep.c == --- head/sys/x86/x86/busdma_machdep.c Tue Feb 28 19:39:54 2012 (r232266) +++ head/sys/x86/x86/busdma_machdep.c Tue Feb 28 19:42:40 2012 (r232267) @@ -227,6 +227,14 @@ bus_dma_tag_create(bus_dma_tag_t parent, bus_dma_tag_t newtag; int error = 0; + /* Always enforce at least a 4GB (2GB for PAE) boundary. */ +#if defined(__amd64__) + if (boundary == 0 || boundary > ((bus_addr_t)1 << 32)) + boundary = (bus_size_t)1 << 32; +#elif defined(PAE) + if (boundary == 0 || boundary > ((bus_addr_t)1 << 31)) + boundary = (bus_size_t)1 << 31; +#endif /* Basic sanity checking */ if (boundary != 0 && boundary < maxsegsz) maxsegsz = boundary; ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r232183 - head/sys/kern
On Sun Feb 26 12, Jilles Tjoelker wrote:
> Author: jilles
> Date: Sun Feb 26 15:14:29 2012
> New Revision: 232183
> URL: http://svn.freebsd.org/changeset/base/232183
>
> Log:
> Fix fchmod() and fchown() on fifos.
>
> The new fifo implementation in r232055 broke fchmod() and fchown() on fifos.
> Postfix needs this.
clang seems to have a problem with this commit:
usr/subversion-src/sys/kern/sys_pipe.c:1556:10: error: promoted type 'int' of
K&R function parameter is not compatible with the parameter type 'mode_t' (aka
'unsigned short') declared in a previous prototype [-Werror]
mode_t mode;
^
/usr/subversion-src/sys/kern/sys_pipe.c:155:19: note: previous declaration is
here
static fo_chmod_t pipe_chmod;
^
1 error generated.
*** Error code 1
cheers.
alex
>
> Submitted by: gianni
> Reported by:dougb
>
> Modified:
> head/sys/kern/sys_pipe.c
>
> Modified: head/sys/kern/sys_pipe.c
> ==
> --- head/sys/kern/sys_pipe.c Sun Feb 26 14:27:34 2012(r232182)
> +++ head/sys/kern/sys_pipe.c Sun Feb 26 15:14:29 2012(r232183)
> @@ -152,6 +152,8 @@ static fo_poll_t pipe_poll;
> static fo_kqfilter_t pipe_kqfilter;
> static fo_stat_t pipe_stat;
> static fo_close_tpipe_close;
> +static fo_chmod_tpipe_chmod;
> +static fo_chown_tpipe_chown;
>
> struct fileops pipeops = {
> .fo_read = pipe_read,
> @@ -162,8 +164,8 @@ struct fileops pipeops = {
> .fo_kqfilter = pipe_kqfilter,
> .fo_stat = pipe_stat,
> .fo_close = pipe_close,
> - .fo_chmod = invfo_chmod,
> - .fo_chown = invfo_chown,
> + .fo_chmod = pipe_chmod,
> + .fo_chown = pipe_chown,
> .fo_flags = DFLAG_PASSABLE
> };
>
> @@ -1548,6 +1550,43 @@ pipe_close(fp, td)
> return (0);
> }
>
> +static int
> +pipe_chmod(fp, mode, active_cred, td)
> + struct file *fp;
> + mode_t mode;
> + struct ucred *active_cred;
> + struct thread *td;
> +{
> + struct pipe *cpipe;
> + int error;
> +
> + cpipe = fp->f_data;
> + if (cpipe->pipe_state & PIPE_NAMED)
> + error = vn_chmod(fp, mode, active_cred, td);
> + else
> + error = invfo_chmod(fp, mode, active_cred, td);
> + return (error);
> +}
> +
> +static int
> +pipe_chown(fp, uid, gid, active_cred, td)
> + struct file *fp;
> + uid_t uid;
> + gid_t gid;
> + struct ucred *active_cred;
> + struct thread *td;
> +{
> + struct pipe *cpipe;
> + int error;
> +
> + cpipe = fp->f_data;
> + if (cpipe->pipe_state & PIPE_NAMED)
> + error = vn_chown(fp, uid, gid, active_cred, td);
> + else
> + error = invfo_chown(fp, uid, gid, active_cred, td);
> + return (error);
> +}
> +
> static void
> pipe_free_kmem(cpipe)
> struct pipe *cpipe;
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r232268 - head/sys/dev/aac
Author: emaste Date: Tue Feb 28 19:50:14 2012 New Revision: 232268 URL: http://svn.freebsd.org/changeset/base/232268 Log: Revert r232260. The problem is now fixed by a general workaround in r232267. Modified: head/sys/dev/aac/aac_pci.c Modified: head/sys/dev/aac/aac_pci.c == --- head/sys/dev/aac/aac_pci.c Tue Feb 28 19:42:40 2012(r232267) +++ head/sys/dev/aac/aac_pci.c Tue Feb 28 19:50:14 2012(r232268) @@ -403,8 +403,7 @@ aac_pci_attach(device_t dev) * Note that some of these controllers are 64-bit capable. */ if (bus_dma_tag_create(NULL,/* parent */ - PAGE_SIZE, /* alignment */ - ((bus_size_t)((uint64_t)1 << 32)), /* boundary*/ + PAGE_SIZE, 0,/* algnmnt, boundary */ BUS_SPACE_MAXADDR, /* lowaddr */ BUS_SPACE_MAXADDR, /* highaddr */ NULL, NULL, /* filter, filterarg */ ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r232266 - in head/sys: amd64/include i386/include pc98/include x86/include
On 2012-02-28 20:39, Tijl Coosemans wrote: > Author: tijl > Date: Tue Feb 28 19:39:54 2012 > New Revision: 232266 > URL: http://svn.freebsd.org/changeset/base/232266 > > Log: > Copy amd64 endian.h to x86 and merge with i386 endian.h. Replace > amd64/i386/pc98 endian.h with stubs. > > In __bswap64_const(x) the conflict between 0xffUL and 0xffULL has been > resolved by reimplementing the macro in terms of __bswap32(x). As a side > effect __bswap64_var(x) is now implemented using two bswap instructions on > i386 and should be much faster. __bswap32_const(x) has been reimplemented > in terms of __bswap16(x) for consistency. ... > +#define __bswap32_const(_x) \ > + (((__uint32_t)__bswap16(_x) << 16) | __bswap16((_x) >> 16)) > + > +#define __bswap32(_x) \ > + (__builtin_constant_p(_x) ? \ > + __bswap32_const((__uint32_t)(_x)) : __bswap32_var(_x)) > + > +#define __bswap64_const(_x) \ > + (((__uint64_t)__bswap32(_x) << 32) | __bswap32((_x) >> 32)) Hmm, shouldn't __bswap32_const() be implemented in terms of __bswap16_const(), and similarly for __bswap64_const(), which should be implemented in terms of __bswap32_const()? The whole reason for the difficult dance with __builtin_constant_p is to allow for compile-time resolving of bswap'd constants. Invoking the regular __bswap() macros doesn't gain you anything here. ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r232238 - in head: share/man/man4 sys/dev/e1000 sys/dev/ixgbe sys/dev/netmap sys/dev/re sys/net tools/tools/netmap
On Tue, Feb 28, 2012 at 08:31:22AM +0100, Luigi Rizzo wrote: > On Mon, Feb 27, 2012 at 07:05:45PM -0500, Ben Kaduk wrote: > ... > > > ?1. add support for devices with different number of rx and tx queues; > > > > > > ?2. add better support for zero-copy operation, adding an extra field > > > ? ? to the netmap ring to indicate how many buffers we have already > > > processed > > > ? ? but not yet released (with help from Eddie Kohler); > > > > > > ?3. The two changes above unfortunately require an API change, so while > > > ? ? at it add a version field and some spares to the ioctl() argument > > > ? ? to help detect mismatches. > > > > Is it worth bumping __FreeBSD_version? > > I don't think it is necessary. > There is basically no code that uses the netmap API except for > the examples in tools/tools/netmap, and i now have a NETMAP_API macro no code in *FreeBSD base system*, yes? because I am write tools uses the netmap API now. > that defines the version without the need for using __FreeBSD_version ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r232266 - in head/sys: amd64/include i386/include pc98/include x86/include
On Tuesday 28 February 2012 21:55:57 Dimitry Andric wrote: > On 2012-02-28 20:39, Tijl Coosemans wrote: >> Author: tijl >> Date: Tue Feb 28 19:39:54 2012 >> New Revision: 232266 >> URL: http://svn.freebsd.org/changeset/base/232266 >> >> Log: >> Copy amd64 endian.h to x86 and merge with i386 endian.h. Replace >> amd64/i386/pc98 endian.h with stubs. >> >> In __bswap64_const(x) the conflict between 0xffUL and 0xffULL has been >> resolved by reimplementing the macro in terms of __bswap32(x). As a side >> effect __bswap64_var(x) is now implemented using two bswap instructions on >> i386 and should be much faster. __bswap32_const(x) has been reimplemented >> in terms of __bswap16(x) for consistency. > ... >> +#define __bswap32_const(_x) \ >> +(((__uint32_t)__bswap16(_x) << 16) | __bswap16((_x) >> 16)) >> + >> +#define __bswap32(_x) \ >> +(__builtin_constant_p(_x) ? \ >> +__bswap32_const((__uint32_t)(_x)) : __bswap32_var(_x)) >> + >> +#define __bswap64_const(_x) \ >> +(((__uint64_t)__bswap32(_x) << 32) | __bswap32((_x) >> 32)) > > Hmm, shouldn't __bswap32_const() be implemented in terms of > __bswap16_const(), and similarly for __bswap64_const(), which should be > implemented in terms of __bswap32_const()? > > The whole reason for the difficult dance with __builtin_constant_p is to > allow for compile-time resolving of bswap'd constants. Invoking the > regular __bswap() macros doesn't gain you anything here. __bswap64_const is also used in __bswap64_var, so its argument isn't always a compile time constant and then __bswap32 becomes __bswap32_var. If it is a constant __bswap32 becomes __bswap32_const (even at -O0). signature.asc Description: This is a digitally signed message part.
Re: svn commit: r232238 - in head: share/man/man4 sys/dev/e1000 sys/dev/ixgbe sys/dev/netmap sys/dev/re sys/net tools/tools/netmap
On Wed, Feb 29, 2012 at 12:35:23AM +0400, Slawa Olhovchenkov wrote: ... > > > > ?3. The two changes above unfortunately require an API change, so while > > > > ? ? at it add a version field and some spares to the ioctl() argument > > > > ? ? to help detect mismatches. > > > > > > Is it worth bumping __FreeBSD_version? > > > > I don't think it is necessary. > > There is basically no code that uses the netmap API except for > > the examples in tools/tools/netmap, and i now have a NETMAP_API macro > > no code in *FreeBSD base system*, yes? > because I am write tools uses the netmap API now. i am glad to hear that, and the NETMAP_API will serve you well because you can use the same also on the Linux version of netmap cheers luigi ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r232269 - head/sys/net80211
Author: adrian
Date: Tue Feb 28 21:41:47 2012
New Revision: 232269
URL: http://svn.freebsd.org/changeset/base/232269
Log:
Print out the bogus beacon interval.
Modified:
head/sys/net80211/ieee80211_input.c
Modified: head/sys/net80211/ieee80211_input.c
==
--- head/sys/net80211/ieee80211_input.c Tue Feb 28 19:50:14 2012
(r232268)
+++ head/sys/net80211/ieee80211_input.c Tue Feb 28 21:41:47 2012
(r232269)
@@ -652,7 +652,8 @@ ieee80211_parse_beacon(struct ieee80211_
scan->bintval <= IEEE80211_BINTVAL_MAX)) {
IEEE80211_DISCARD(vap,
IEEE80211_MSG_ELEMID | IEEE80211_MSG_INPUT,
- wh, NULL, "bogus beacon interval", scan->bintval);
+ wh, NULL, "bogus beacon interval (%d TU)",
+ (int) scan->bintval);
vap->iv_stats.is_rx_badbintval++;
scan->status |= IEEE80211_BPARSE_BINTVAL_INVALID;
}
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r232270 - head/sys/net80211
Author: adrian
Date: Tue Feb 28 21:43:29 2012
New Revision: 232270
URL: http://svn.freebsd.org/changeset/base/232270
Log:
Only increment is_beacon_bad if we're not scanning.
Otherwise things such as off-channel probe responses and beacons are also
silently discarded and logged against this error counter.
Modified:
head/sys/net80211/ieee80211_sta.c
Modified: head/sys/net80211/ieee80211_sta.c
==
--- head/sys/net80211/ieee80211_sta.c Tue Feb 28 21:41:47 2012
(r232269)
+++ head/sys/net80211/ieee80211_sta.c Tue Feb 28 21:43:29 2012
(r232270)
@@ -1306,7 +1306,8 @@ sta_recv_mgmt(struct ieee80211_node *ni,
}
/* XXX probe response in sta mode when !scanning? */
if (ieee80211_parse_beacon(ni, m0, &scan) != 0) {
- vap->iv_stats.is_beacon_bad++;
+ if (! (ic->ic_flags & IEEE80211_F_SCAN))
+ vap->iv_stats.is_beacon_bad++;
return;
}
/*
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r232271 - head/sys/kern
Author: dim
Date: Tue Feb 28 21:45:21 2012
New Revision: 232271
URL: http://svn.freebsd.org/changeset/base/232271
Log:
Change definition of pipe_chmod() from K&R to C99, to avoid the
following clang warning:
sys/kern/sys_pipe.c:1556:10: error: promoted type 'int' of K&R function
parameter is not compatible with the parameter type 'mode_t'
(aka 'unsigned short') declared in a previous prototype [-Werror]
mode_t mode;
^
sys/kern/sys_pipe.c:155:19: note: previous declaration is here
static fo_chmod_t pipe_chmod;
^
Modified:
head/sys/kern/sys_pipe.c
Modified: head/sys/kern/sys_pipe.c
==
--- head/sys/kern/sys_pipe.cTue Feb 28 21:43:29 2012(r232270)
+++ head/sys/kern/sys_pipe.cTue Feb 28 21:45:21 2012(r232271)
@@ -1551,11 +1551,7 @@ pipe_close(fp, td)
}
static int
-pipe_chmod(fp, mode, active_cred, td)
- struct file *fp;
- mode_t mode;
- struct ucred *active_cred;
- struct thread *td;
+pipe_chmod(struct file *fp, mode_t mode, struct ucred *active_cred, struct
thread *td)
{
struct pipe *cpipe;
int error;
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r232272 - head/sys/netinet/ipfw
Author: oleg
Date: Tue Feb 28 21:53:39 2012
New Revision: 232272
URL: http://svn.freebsd.org/changeset/base/232272
Log:
lookup_dyn_rule_locked(): style(9) cleanup
MFC after:1 month
Modified:
head/sys/netinet/ipfw/ip_fw_dynamic.c
Modified: head/sys/netinet/ipfw/ip_fw_dynamic.c
==
--- head/sys/netinet/ipfw/ip_fw_dynamic.c Tue Feb 28 21:45:21 2012
(r232271)
+++ head/sys/netinet/ipfw/ip_fw_dynamic.c Tue Feb 28 21:53:39 2012
(r232272)
@@ -390,72 +390,68 @@ ipfw_remove_dyn_children(struct ip_fw *r
IPFW_DYN_UNLOCK();
}
-/**
- * lookup a dynamic rule, locked version
+/*
+ * Lookup a dynamic rule, locked version.
*/
static ipfw_dyn_rule *
lookup_dyn_rule_locked(struct ipfw_flow_id *pkt, int *match_direction,
struct tcphdr *tcp)
{
/*
-* stateful ipfw extensions.
-* Lookup into dynamic session queue
+* Stateful ipfw extensions.
+* Lookup into dynamic session queue.
*/
#define MATCH_REVERSE 0
#define MATCH_FORWARD 1
#define MATCH_NONE 2
#define MATCH_UNKNOWN 3
int i, dir = MATCH_NONE;
- ipfw_dyn_rule *prev, *q=NULL;
+ ipfw_dyn_rule *prev, *q = NULL;
IPFW_DYN_LOCK_ASSERT();
if (V_ipfw_dyn_v == NULL)
- goto done; /* not found */
- i = hash_packet( pkt );
- for (prev=NULL, q = V_ipfw_dyn_v[i] ; q != NULL ; ) {
+ goto done; /* not found */
+ i = hash_packet(pkt);
+ for (prev = NULL, q = V_ipfw_dyn_v[i]; q != NULL;) {
if (q->dyn_type == O_LIMIT_PARENT && q->count)
goto next;
- if (TIME_LEQ( q->expire, time_uptime)) { /* expire entry */
+ if (TIME_LEQ(q->expire, time_uptime)) { /* expire entry */
UNLINK_DYN_RULE(prev, V_ipfw_dyn_v[i], q);
continue;
}
- if (pkt->proto == q->id.proto &&
- q->dyn_type != O_LIMIT_PARENT) {
- if (IS_IP6_FLOW_ID(pkt)) {
- if (IN6_ARE_ADDR_EQUAL(&(pkt->src_ip6),
- &(q->id.src_ip6)) &&
- IN6_ARE_ADDR_EQUAL(&(pkt->dst_ip6),
- &(q->id.dst_ip6)) &&
+ if (pkt->proto != q->id.proto || q->dyn_type == O_LIMIT_PARENT)
+ goto next;
+
+ if (IS_IP6_FLOW_ID(pkt)) {
+ if (IN6_ARE_ADDR_EQUAL(&pkt->src_ip6, &q->id.src_ip6) &&
+ IN6_ARE_ADDR_EQUAL(&pkt->dst_ip6, &q->id.dst_ip6) &&
pkt->src_port == q->id.src_port &&
- pkt->dst_port == q->id.dst_port ) {
+ pkt->dst_port == q->id.dst_port) {
dir = MATCH_FORWARD;
break;
- }
- if (IN6_ARE_ADDR_EQUAL(&(pkt->src_ip6),
- &(q->id.dst_ip6)) &&
- IN6_ARE_ADDR_EQUAL(&(pkt->dst_ip6),
- &(q->id.src_ip6)) &&
- pkt->src_port == q->id.dst_port &&
- pkt->dst_port == q->id.src_port ) {
- dir = MATCH_REVERSE;
- break;
- }
- } else {
- if (pkt->src_ip == q->id.src_ip &&
- pkt->dst_ip == q->id.dst_ip &&
- pkt->src_port == q->id.src_port &&
- pkt->dst_port == q->id.dst_port ) {
- dir = MATCH_FORWARD;
- break;
- }
- if (pkt->src_ip == q->id.dst_ip &&
- pkt->dst_ip == q->id.src_ip &&
- pkt->src_port == q->id.dst_port &&
- pkt->dst_port == q->id.src_port ) {
- dir = MATCH_REVERSE;
- break;
- }
+ }
+ if (IN6_ARE_ADDR_EQUAL(&pkt->src_ip6, &q->id.dst_ip6) &&
+ IN6_ARE_ADDR_EQUAL(&pkt->dst_ip6, &q->id.src_ip6) &&
+ pkt->src_port == q->id.dst_port &&
+ pkt->dst_port == q->id.src_port) {
+ dir = MATCH_REVERSE;
+ break;
+ }
+ } else {
+ if (pkt->src_ip == q->id.src_ip &&
+ pkt->dst_ip == q->id.dst_ip &&
+ pkt->src_port == q->id.
svn commit: r232273 - head/sys/netinet/ipfw
Author: oleg
Date: Tue Feb 28 22:00:41 2012
New Revision: 232273
URL: http://svn.freebsd.org/changeset/base/232273
Log:
- Refresh dynamic tcp rule only if both sides answered keepalive packets.
- Remove some useless assignments.
MFC after:1 month
Modified:
head/sys/netinet/ipfw/ip_fw_dynamic.c
Modified: head/sys/netinet/ipfw/ip_fw_dynamic.c
==
--- head/sys/netinet/ipfw/ip_fw_dynamic.c Tue Feb 28 21:53:39 2012
(r232272)
+++ head/sys/netinet/ipfw/ip_fw_dynamic.c Tue Feb 28 22:00:41 2012
(r232273)
@@ -472,8 +472,12 @@ next:
#define BOTH_SYN (TH_SYN | (TH_SYN << 8))
#define BOTH_FIN (TH_FIN | (TH_FIN << 8))
+#defineTCP_FLAGS (TH_FLAGS | (TH_FLAGS << 8))
+#defineACK_FWD 0x1 /* fwd ack seen */
+#defineACK_REV 0x2 /* rev ack seen */
+
q->state |= (dir == MATCH_FORWARD) ? flags : (flags << 8);
- switch (q->state) {
+ switch (q->state & TCP_FLAGS) {
case TH_SYN:/* opening */
q->expire = time_uptime + V_dyn_syn_lifetime;
break;
@@ -482,24 +486,28 @@ next:
case BOTH_SYN | TH_FIN: /* one side tries to close */
case BOTH_SYN | (TH_FIN << 8):
#define _SEQ_GE(a,b) ((int)(a) - (int)(b) >= 0)
- if (tcp == NULL) {
- q->expire = time_uptime + V_dyn_ack_lifetime;
+ if (tcp == NULL)
break;
- }
ack = ntohl(tcp->th_ack);
if (dir == MATCH_FORWARD) {
- if (q->ack_fwd == 0 || _SEQ_GE(ack, q->ack_fwd))
+ if (q->ack_fwd == 0 ||
+ _SEQ_GE(ack, q->ack_fwd)) {
q->ack_fwd = ack;
- else/* ignore out-of-sequence */
- break;
+ q->state |= ACK_FWD;
+ }
} else {
- if (q->ack_rev == 0 || _SEQ_GE(ack, q->ack_rev))
+ if (q->ack_rev == 0 ||
+ _SEQ_GE(ack, q->ack_rev)) {
q->ack_rev = ack;
- else/* ignore out-of-sequence */
- break;
+ q->state |= ACK_REV;
+ }
+ }
+ if ((q->state & (ACK_FWD | ACK_REV)) ==
+ (ACK_FWD | ACK_REV)) {
+ q->expire = time_uptime + V_dyn_ack_lifetime;
+ q->state &= ~(ACK_FWD | ACK_REV);
}
- q->expire = time_uptime + V_dyn_ack_lifetime;
break;
case BOTH_SYN | BOTH_FIN: /* both sides closed */
@@ -1074,10 +1082,12 @@ ipfw_tick(void * vnetx)
if (TIME_LEQ(q->expire, time_uptime))
continue; /* too late, rule expired */
- m = ipfw_send_pkt(NULL, &(q->id), q->ack_rev - 1,
- q->ack_fwd, TH_SYN);
- mnext = ipfw_send_pkt(NULL, &(q->id), q->ack_fwd - 1,
- q->ack_rev, 0);
+ m = (q->state & ACK_REV) ? NULL :
+ ipfw_send_pkt(NULL, &(q->id), q->ack_rev - 1,
+ q->ack_fwd, TH_SYN);
+ mnext = (q->state & ACK_FWD) ? NULL :
+ ipfw_send_pkt(NULL, &(q->id), q->ack_fwd - 1,
+ q->ack_rev, 0);
switch (q->id.addr_type) {
case 4:
@@ -1103,18 +1113,16 @@ ipfw_tick(void * vnetx)
break;
#endif
}
-
- m = mnext = NULL;
}
}
IPFW_DYN_UNLOCK();
- for (m = mnext = m0; m != NULL; m = mnext) {
+ for (m = m0; m != NULL; m = mnext) {
mnext = m->m_nextpkt;
m->m_nextpkt = NULL;
ip_output(m, NULL, NULL, 0, NULL, NULL);
}
#ifdef INET6
- for (m = mnext = m6; m != NULL; m = mnext) {
+ for (m = m6; m != NULL; m = mnext) {
mnext = m->m_nextpkt;
m->m_nextpkt = NULL;
ip6_output(m, NULL, NULL, 0, NULL, NULL, NULL);
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listi
Re: svn commit: r232272 - head/sys/netinet/ipfw
On Tue, Feb 28, 2012 at 09:53:39PM +, Oleg Bulyzhin wrote:
> Author: oleg
> Date: Tue Feb 28 21:53:39 2012
> New Revision: 232272
> URL: http://svn.freebsd.org/changeset/base/232272
>
> Log:
> lookup_dyn_rule_locked(): style(9) cleanup
>
> MFC after: 1 month
is there a reason for delaying the MFC, such as this is
preparatory work for further changes ?
Otherwise i'd suggest that whitespace changes like this
are quickly pushed to the stable branch so that the code
versions do not diverge.
cheers
luigi
> Modified:
> head/sys/netinet/ipfw/ip_fw_dynamic.c
>
> Modified: head/sys/netinet/ipfw/ip_fw_dynamic.c
> ==
> --- head/sys/netinet/ipfw/ip_fw_dynamic.c Tue Feb 28 21:45:21 2012
> (r232271)
> +++ head/sys/netinet/ipfw/ip_fw_dynamic.c Tue Feb 28 21:53:39 2012
> (r232272)
> @@ -390,72 +390,68 @@ ipfw_remove_dyn_children(struct ip_fw *r
> IPFW_DYN_UNLOCK();
> }
>
> -/**
> - * lookup a dynamic rule, locked version
> +/*
> + * Lookup a dynamic rule, locked version.
> */
> static ipfw_dyn_rule *
> lookup_dyn_rule_locked(struct ipfw_flow_id *pkt, int *match_direction,
> struct tcphdr *tcp)
> {
> /*
> - * stateful ipfw extensions.
> - * Lookup into dynamic session queue
> + * Stateful ipfw extensions.
> + * Lookup into dynamic session queue.
>*/
> #define MATCH_REVERSE0
> #define MATCH_FORWARD1
> #define MATCH_NONE 2
> #define MATCH_UNKNOWN3
> int i, dir = MATCH_NONE;
> - ipfw_dyn_rule *prev, *q=NULL;
> + ipfw_dyn_rule *prev, *q = NULL;
>
> IPFW_DYN_LOCK_ASSERT();
>
> if (V_ipfw_dyn_v == NULL)
> - goto done; /* not found */
> - i = hash_packet( pkt );
> - for (prev=NULL, q = V_ipfw_dyn_v[i] ; q != NULL ; ) {
> + goto done; /* not found */
> + i = hash_packet(pkt);
> + for (prev = NULL, q = V_ipfw_dyn_v[i]; q != NULL;) {
> if (q->dyn_type == O_LIMIT_PARENT && q->count)
> goto next;
> - if (TIME_LEQ( q->expire, time_uptime)) { /* expire entry */
> + if (TIME_LEQ(q->expire, time_uptime)) { /* expire entry */
> UNLINK_DYN_RULE(prev, V_ipfw_dyn_v[i], q);
> continue;
> }
> - if (pkt->proto == q->id.proto &&
> - q->dyn_type != O_LIMIT_PARENT) {
> - if (IS_IP6_FLOW_ID(pkt)) {
> - if (IN6_ARE_ADDR_EQUAL(&(pkt->src_ip6),
> - &(q->id.src_ip6)) &&
> - IN6_ARE_ADDR_EQUAL(&(pkt->dst_ip6),
> - &(q->id.dst_ip6)) &&
> + if (pkt->proto != q->id.proto || q->dyn_type == O_LIMIT_PARENT)
> + goto next;
> +
> + if (IS_IP6_FLOW_ID(pkt)) {
> + if (IN6_ARE_ADDR_EQUAL(&pkt->src_ip6, &q->id.src_ip6) &&
> + IN6_ARE_ADDR_EQUAL(&pkt->dst_ip6, &q->id.dst_ip6) &&
> pkt->src_port == q->id.src_port &&
> - pkt->dst_port == q->id.dst_port ) {
> + pkt->dst_port == q->id.dst_port) {
> dir = MATCH_FORWARD;
> break;
> - }
> - if (IN6_ARE_ADDR_EQUAL(&(pkt->src_ip6),
> - &(q->id.dst_ip6)) &&
> - IN6_ARE_ADDR_EQUAL(&(pkt->dst_ip6),
> - &(q->id.src_ip6)) &&
> - pkt->src_port == q->id.dst_port &&
> - pkt->dst_port == q->id.src_port ) {
> - dir = MATCH_REVERSE;
> - break;
> - }
> - } else {
> - if (pkt->src_ip == q->id.src_ip &&
> - pkt->dst_ip == q->id.dst_ip &&
> - pkt->src_port == q->id.src_port &&
> - pkt->dst_port == q->id.dst_port ) {
> - dir = MATCH_FORWARD;
> - break;
> - }
> - if (pkt->src_ip == q->id.dst_ip &&
> - pkt->dst_ip == q->id.src_ip &&
> - pkt->src_port == q->id.dst_port &&
> - pkt->dst_port == q->id.src_port ) {
> - dir = MATCH_REVERSE;
> - break;
> - }
> + }
> + if (IN6_ARE_ADDR_EQUAL(&pkt->src_ip6, &q->id.dst_ip6) &&
> + IN6_ARE_ADDR_EQUAL(&pkt->dst_ip6, &q->id.src_ip6) &&
> + pkt->src_port
Re: svn commit: r232271 - head/sys/kern
On Tue, Feb 28, 2012 at 10:45 PM, Dimitry Andric wrote:
> Author: dim
> Date: Tue Feb 28 21:45:21 2012
> New Revision: 232271
> URL: http://svn.freebsd.org/changeset/base/232271
>
> Log:
> Â Change definition of pipe_chmod() from K&R to C99, to avoid the
> Â following clang warning:
>
> Â sys/kern/sys_pipe.c:1556:10: error: promoted type 'int' of K&R function
> parameter is not compatible with the parameter type 'mode_t'
> Â Â Â Â (aka 'unsigned short') declared in a previous prototype [-Werror]
> Â Â Â Â Â mode_t mode;
> Â Â Â Â Â Â Â Â ^
> Â sys/kern/sys_pipe.c:155:19: note: previous declaration is here
>  static fo_chmod_t    pipe_chmod;
> Â Â Â Â Â Â Â Â Â Â Â Â Â ^
>
> Modified:
> Â head/sys/kern/sys_pipe.c
>
> Modified: head/sys/kern/sys_pipe.c
> ==
> --- head/sys/kern/sys_pipe.c   Tue Feb 28 21:43:29 2012     (r232270)
> +++ head/sys/kern/sys_pipe.c   Tue Feb 28 21:45:21 2012     (r232271)
> @@ -1551,11 +1551,7 @@ pipe_close(fp, td)
> Â }
>
> Â static int
> -pipe_chmod(fp, mode, active_cred, td)
> - Â Â Â struct file *fp;
> - Â Â Â mode_t mode;
> - Â Â Â struct ucred *active_cred;
> - Â Â Â struct thread *td;
> +pipe_chmod(struct file *fp, mode_t mode, struct ucred *active_cred, struct
> thread *td)
> Â {
> Â Â Â Â struct pipe *cpipe;
> Â Â Â Â int error;
Thank you.
It introduces a style(9) violation, though.
I think it should be in this way
pipe_chmod(struct file *fp, mode_t mode, struct ucred *active_cred,
struct thread *td)
Anyway that's not a problem I can take care of this and other style violations
during next weeks.
--
Gianni
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r232274 - head/usr.bin/fetch
Author: obrien Date: Tue Feb 28 22:15:46 2012 New Revision: 232274 URL: http://svn.freebsd.org/changeset/base/232274 Log: Document the 'URL' argument. Modified: head/usr.bin/fetch/fetch.1 Modified: head/usr.bin/fetch/fetch.1 == --- head/usr.bin/fetch/fetch.1 Tue Feb 28 22:00:41 2012(r232273) +++ head/usr.bin/fetch/fetch.1 Tue Feb 28 22:15:46 2012(r232274) @@ -1,5 +1,5 @@ .\"- -.\" Copyright (c) 2000-2011 Dag-Erling Smørgrav +.\" Copyright (c) 2000-2012 Dag-Erling Smørgrav .\" All rights reserved. .\" Portions Copyright (c) 1999 Massachusetts Institute of Technology; used .\" by permission. @@ -29,7 +29,7 @@ .\" .\" $FreeBSD$ .\" -.Dd September 27, 2011 +.Dd February 28, 2012 .Dt FETCH 1 .Os .Sh NAME @@ -214,6 +214,11 @@ When the flag is specified, wait this many seconds between successive retries. .El .Pp +.Ar URL +.Bd -literal +:(//((:)?@)?(:)?)?/()? +.Ed +.Pp If .Nm receives a ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r232275 - in head/sys: amd64/include i386/include pc98/include x86/include
Author: tijl
Date: Tue Feb 28 22:17:52 2012
New Revision: 232275
URL: http://svn.freebsd.org/changeset/base/232275
Log:
Copy amd64 setjmp.h to x86 and replace amd64/i386/pc98 setjmp.h with stubs.
Added:
head/sys/x86/include/setjmp.h
- copied unchanged from r232268, head/sys/amd64/include/setjmp.h
Modified:
head/sys/amd64/include/setjmp.h
head/sys/i386/include/setjmp.h
head/sys/pc98/include/setjmp.h
Modified: head/sys/amd64/include/setjmp.h
==
--- head/sys/amd64/include/setjmp.h Tue Feb 28 22:15:46 2012
(r232274)
+++ head/sys/amd64/include/setjmp.h Tue Feb 28 22:17:52 2012
(r232275)
@@ -1,50 +1,6 @@
/*-
- * Copyright (c) 1998 John Birrell .
- * 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.
- * 2. Redistributions in binary form must reproduce the above copyright
- *notice, this list of conditions and the following disclaimer in the
- *documentation and/or other materials provided with the distribution.
- * 3. Neither the name of the author nor the names of any co-contributors
- *may be used to endorse or promote products derived from this software
- *without specific prior written permission.
- *
- * THIS SOFTWARE IS PROVIDED BY JOHN BIRRELL AND CONTRIBUTORS ``AS IS'' AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, 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 DAMAGE.
- *
- * $FreeBSD$
+ * This file is in the public domain.
*/
+/* $FreeBSD$ */
-#ifndef _MACHINE_SETJMP_H_
-#define_MACHINE_SETJMP_H_
-
-#include
-
-#define_JBLEN 12 /* Size of the jmp_buf on AMD64. */
-
-/*
- * jmp_buf and sigjmp_buf are encapsulated in different structs to force
- * compile-time diagnostics for mismatches. The structs are the same
- * internally to avoid some run-time errors for mismatches.
- */
-#if __BSD_VISIBLE || __POSIX_VISIBLE || __XSI_VISIBLE
-typedefstruct _sigjmp_buf { long _sjb[_JBLEN]; } sigjmp_buf[1];
-#endif
-
-typedefstruct _jmp_buf { long _jb[_JBLEN]; } jmp_buf[1];
-
-#endif /* !_MACHINE_SETJMP_H_ */
+#include
Modified: head/sys/i386/include/setjmp.h
==
--- head/sys/i386/include/setjmp.h Tue Feb 28 22:15:46 2012
(r232274)
+++ head/sys/i386/include/setjmp.h Tue Feb 28 22:17:52 2012
(r232275)
@@ -1,50 +1,6 @@
/*-
- * Copyright (c) 1998 John Birrell .
- * 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.
- * 2. Redistributions in binary form must reproduce the above copyright
- *notice, this list of conditions and the following disclaimer in the
- *documentation and/or other materials provided with the distribution.
- * 3. Neither the name of the author nor the names of any co-contributors
- *may be used to endorse or promote products derived from this software
- *without specific prior written permission.
- *
- * THIS SOFTWARE IS PROVIDED BY JOHN BIRRELL AND CONTRIBUTORS ``AS IS'' AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, 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 DAMAGE.
- *
- * $FreeBSD$
+ * This file is in the public domain.
*/
+/* $FreeBSD$ */
-#ifndef _MACHINE_SETJMP_
svn commit: r232276 - in head/sys: amd64/include i386/include pc98/include x86/include
Author: tijl Date: Tue Feb 28 22:30:58 2012 New Revision: 232276 URL: http://svn.freebsd.org/changeset/base/232276 Log: Copy amd64 stdarg.h to x86 and replace amd64/i386/pc98 stdarg.h with stubs. Added: head/sys/x86/include/stdarg.h - copied unchanged from r232268, head/sys/amd64/include/stdarg.h Modified: head/sys/amd64/include/stdarg.h head/sys/i386/include/stdarg.h head/sys/pc98/include/stdarg.h Modified: head/sys/amd64/include/stdarg.h == --- head/sys/amd64/include/stdarg.h Tue Feb 28 22:17:52 2012 (r232275) +++ head/sys/amd64/include/stdarg.h Tue Feb 28 22:30:58 2012 (r232276) @@ -1,75 +1,6 @@ /*- - * Copyright (c) 2002 David E. O'Brien. 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. - * 2. Redistributions in binary form must reproduce the above copyright - *notice, this list of conditions and the following disclaimer in the - *documentation and/or other materials provided with the distribution. - * 3. Neither the name of the University nor the names of its contributors - *may be used to endorse or promote products derived from this software - *without specific prior written permission. - * - * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND - * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE - * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE - * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE - * FOR ANY DIRECT, INDIRECT, INCIDENTAL, 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 DAMAGE. - * - * $FreeBSD$ + * This file is in the public domain. */ +/* $FreeBSD$ */ -#ifndef _MACHINE_STDARG_H_ -#define_MACHINE_STDARG_H_ - -#include -#include - -#ifndef _VA_LIST_DECLARED -#define_VA_LIST_DECLARED -typedef__va_list va_list; -#endif - -#ifdef __GNUCLIKE_BUILTIN_STDARG - -#defineva_start(ap, last) \ - __builtin_va_start((ap), (last)) - -#defineva_arg(ap, type) \ - __builtin_va_arg((ap), type) - -#define__va_copy(dest, src) \ - __builtin_va_copy((dest), (src)) - -#if __ISO_C_VISIBLE >= 1999 -#defineva_copy(dest, src) \ - __va_copy(dest, src) -#endif - -#defineva_end(ap) \ - __builtin_va_end(ap) - -#elif defined(lint) -/* Provide a fake implementation for lint's benefit */ -#define__va_size(type) \ - (((sizeof(type) + sizeof(long) - 1) / sizeof(long)) * sizeof(long)) -#defineva_start(ap, last) \ - ((ap) = (va_list)&(last) + __va_size(last)) -#defineva_arg(ap, type) \ - (*(type *)((ap) += __va_size(type), (ap) - __va_size(type))) -#defineva_end(ap) - -#else -#error this file needs to be ported to your compiler -#endif - -#endif /* !_MACHINE_STDARG_H_ */ +#include Modified: head/sys/i386/include/stdarg.h == --- head/sys/i386/include/stdarg.h Tue Feb 28 22:17:52 2012 (r232275) +++ head/sys/i386/include/stdarg.h Tue Feb 28 22:30:58 2012 (r232276) @@ -1,90 +1,6 @@ /*- - * Copyright (c) 2002 David E. O'Brien. All rights reserved. - * Copyright (c) 1991, 1993 - * The Regents of the University of California. 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. - * 2. Redistributions in binary form must reproduce the above copyright - *notice, this list of conditions and the following disclaimer in the - *documentation and/or other materials provided with the distribution. - * 3. All advertising materials mentioning features or use of this software - *must display the following acknowledgement: - * This product includes software developed by the University of - * California, Berkeley and its contributors. - * 4. Neither the name of the University nor the names of its contributors - *may be used to endorse or promote products derived from this software - *without specific prior written permission. -
Re: svn commit: r232272 - head/sys/netinet/ipfw
On Wed, Feb 29, 2012 at 02:39:37AM +0400, Oleg Bulyzhin wrote: > On Tue, Feb 28, 2012 at 11:22:52PM +0100, Luigi Rizzo wrote: > > On Tue, Feb 28, 2012 at 09:53:39PM +, Oleg Bulyzhin wrote: > > > Author: oleg > > > Date: Tue Feb 28 21:53:39 2012 > > > New Revision: 232272 > > > URL: http://svn.freebsd.org/changeset/base/232272 > > > > > > Log: > > > lookup_dyn_rule_locked(): style(9) cleanup > > > > > > MFC after: 1 month > > > > is there a reason for delaying the MFC, such as this is > > preparatory work for further changes ? > > Otherwise i'd suggest that whitespace changes like this > > are quickly pushed to the stable branch so that the code > > versions do not diverge. > > > > cheers > > luigi > > Actual commit is r232273, this one used for reduce diff > of meaningful commit. ok, thanks luigi ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r232272 - head/sys/netinet/ipfw
On Tue, Feb 28, 2012 at 11:22:52PM +0100, Luigi Rizzo wrote: > On Tue, Feb 28, 2012 at 09:53:39PM +, Oleg Bulyzhin wrote: > > Author: oleg > > Date: Tue Feb 28 21:53:39 2012 > > New Revision: 232272 > > URL: http://svn.freebsd.org/changeset/base/232272 > > > > Log: > > lookup_dyn_rule_locked(): style(9) cleanup > > > > MFC after:1 month > > is there a reason for delaying the MFC, such as this is > preparatory work for further changes ? > Otherwise i'd suggest that whitespace changes like this > are quickly pushed to the stable branch so that the code > versions do not diverge. > > cheers > luigi Actual commit is r232273, this one used for reduce diff of meaningful commit. -- Oleg. === Oleg Bulyzhin -- OBUL-RIPN -- OBUL-RIPE -- o...@rinet.ru === ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r232261 - in head/sys: amd64/include i386/include pc98/include x86/include
On Tue, 28 Feb 2012, Tijl Coosemans wrote: Log: Copy amd64 _types.h to x86 and merge with i386 _types.h. Replace existing amd64/i386/pc98 _types.h with stubs. I don't like this much. It gives 2 layers of convolutions for readers (humans and compilers) to slowly untangle. There is another layer of include files for compatibility (but both layers are still used by default), and lots of ifdefs. The whole point of 1 file per arch was to avoid such ifdefs. This might be OK if arches were actually identical for the APIs in these files, but for types there are lots of differences between 32-bit and 64-bit machines. The differences can be reduced by spelling 32/64-bit types as [unsigned] long and by always using the basic type instead of a derived type. Old code (e.g., FreeBSD-3) did this, but now almost everything is declared using the derived fixed-width types int32_t/int64_t etc., so there is always a spelling difference for 32/64-bit types. The only exceptions are floating point types, and the broken __clock_t type. clock_t is unsigned long on i386, but is int32_t on amd64. This is backwards at best. I think this brokenness came from NetBSD. _BSD_CLOCK_T_ was correct (unsigned long) for all arches in 4.4BSD, and i386 just didn't break this. But now __clock_t is broken for all arches except i386: it is __uint32_t on arm and powerpc, which is just a different spelling for the 32-bit case and is at least ABI-compatible for the 64-bit case, but for all other arches including all 64-bit ones, it broken to __int32_t. Perhaps the difference is explained by "long" being bad for ABI compatibility. Old code like 4.4BSD-Lite1 used long excessively (since technically, int might be only 16 bits). Even pid_t was long in 4.4BSD-Lite1. NetBSD changed many of these longs to ints or int32_t's for ABI compatibility and/or because 64 bit longs are just too wide, and FreeBSD eventually picked up these changes (mostly via 4.4BSD-Lite2 for general typedefs and directly from NetBSD for 64-bit arches). So pid_t is now int32_t and clock_t is mostly broken. clock_t really needs all 64 bits if they are readily available, but has been reduced to 31, especially when 64 are readily available. OTOH, if you just want ABI and API compatibility for clock_t, then it should have been changed to uint32_t for all arches and not defined in any MD types file. There is now a minor API compatibility for printing clock_t's -- %lu format must be used on i386, %u on others, and %d on most. Except for the gratuitous loss of unsignedness, this is just a special case of printing a typedefed types. clock_t may also be a floating point type, so the only good way to print it is to convert it to [long] double and then worry about the correct floating point format (how much precision should it have?...). Copied and modified: head/sys/x86/include/_types.h (from r232259, head/sys/amd64/include/_types.h) == --- head/sys/amd64/include/_types.h Tue Feb 28 15:52:01 2012 (r232259, copy source) +++ head/sys/x86/include/_types.h Tue Feb 28 18:15:28 2012 (r232261) @@ -54,19 +54,41 @@ typedef short __int16_t; typedef unsigned short __uint16_t; typedef int __int32_t; typedef unsigned int__uint32_t; +#ifdef _LP64 typedef long__int64_t; typedef unsigned long __uint64_t; This is about the only ifdef that is really needed. +#else +#ifndef lint +__extension__ An old bug -- work around broken lints. Although messy, this is not messy enough to be correct -- __extension__ is a hard-coded gccism. Elsewhere, in much less important code than this, there are messy ifdefs to avoid hard-coded gccisms. +#endif +/* LONGLONG */ long long has only been standard for 13 years now, so broken lints still need this messy markup. +typedeflong long __int64_t; +#ifndef lint +__extension__ +#endif +/* LONGLONG */ +typedefunsigned long long __uint64_t; +#endif I ifdefed all this correctly 15+ years ago so that it compiled (but didn't run if *int64_t was used) for a non-gcc K&R compiler. The long long abomination was not used, and __attribute__(()) was used to declare *int64_t, but only under a gcc ifdef. This wasn't broken until 8 Jan 2011 by, erm, us. We also added the __extensions__. The justification was that long long is now standard. But there are still the old messes for lint, and new breakage for non-gcc to support C90. /* * Standard type definitions. */ +#ifdef _LP64 typedef __int32_t __clock_t; /* clock()... */ typedef __int64_t __critical_t; typedef double __double_t; typedef float __float_t; typedef __int64_t __intfptr_t; -typedef__int64
Re: svn commit: r232181 - in head/sys: kern sys
On 2/27/12 11:29 PM, Mikolaj Golub wrote: On Mon, 27 Feb 2012 22:34:25 -0800 Julian Elischer wrote: JE> I don't think this belongs in the kernel by default. It's not exactl a JE> call for backout but It's teh next thing short of that. a call for "do JE> you REALLY think we need this particular specific case catered for?" The main goal of the patch was to provide ability to get another process umask. It looks like usefulness of this is not questioned here. well that's exactly what I AM questioning.. how often will this be used? one person using this once in all of history isn't a real requirement for inclusion. It seems to me that someone is more likely to figure out a sneaky way to use this in a bad way than to want to use it in the way you expect. ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r232278 - in head: sys/compat/linprocfs sys/compat/linsysfs sys/fs/procfs sys/fs/pseudofs sys/kern sys/sys usr.sbin/jail
Author: mm
Date: Wed Feb 29 00:30:18 2012
New Revision: 232278
URL: http://svn.freebsd.org/changeset/base/232278
Log:
Add procfs to jail-mountable filesystems.
Reviewed by: jamie
MFC after:1 week
Modified:
head/sys/compat/linprocfs/linprocfs.c
head/sys/compat/linsysfs/linsysfs.c
head/sys/fs/procfs/procfs.c
head/sys/fs/pseudofs/pseudofs.h
head/sys/kern/kern_jail.c
head/sys/sys/jail.h
head/usr.sbin/jail/jail.8
Modified: head/sys/compat/linprocfs/linprocfs.c
==
--- head/sys/compat/linprocfs/linprocfs.c Tue Feb 28 23:30:19 2012
(r232277)
+++ head/sys/compat/linprocfs/linprocfs.c Wed Feb 29 00:30:18 2012
(r232278)
@@ -1460,7 +1460,7 @@ linprocfs_uninit(PFS_INIT_ARGS)
return (0);
}
-PSEUDOFS(linprocfs, 1);
+PSEUDOFS(linprocfs, 1, 0);
MODULE_DEPEND(linprocfs, linux, 1, 1, 1);
MODULE_DEPEND(linprocfs, procfs, 1, 1, 1);
MODULE_DEPEND(linprocfs, sysvmsg, 1, 1, 1);
Modified: head/sys/compat/linsysfs/linsysfs.c
==
--- head/sys/compat/linsysfs/linsysfs.c Tue Feb 28 23:30:19 2012
(r232277)
+++ head/sys/compat/linsysfs/linsysfs.c Wed Feb 29 00:30:18 2012
(r232278)
@@ -280,5 +280,5 @@ linsysfs_uninit(PFS_INIT_ARGS)
return (0);
}
-PSEUDOFS(linsysfs, 1);
+PSEUDOFS(linsysfs, 1, 0);
MODULE_DEPEND(linsysfs, linux, 1, 1, 1);
Modified: head/sys/fs/procfs/procfs.c
==
--- head/sys/fs/procfs/procfs.c Tue Feb 28 23:30:19 2012(r232277)
+++ head/sys/fs/procfs/procfs.c Wed Feb 29 00:30:18 2012(r232278)
@@ -209,4 +209,4 @@ procfs_uninit(PFS_INIT_ARGS)
return (0);
}
-PSEUDOFS(procfs, 1);
+PSEUDOFS(procfs, 1, PR_ALLOW_MOUNT_PROCFS);
Modified: head/sys/fs/pseudofs/pseudofs.h
==
--- head/sys/fs/pseudofs/pseudofs.h Tue Feb 28 23:30:19 2012
(r232277)
+++ head/sys/fs/pseudofs/pseudofs.h Wed Feb 29 00:30:18 2012
(r232278)
@@ -31,6 +31,8 @@
#ifndef _PSEUDOFS_H_INCLUDED
#define _PSEUDOFS_H_INCLUDED
+#include
+
/*
* Opaque structures
*/
@@ -271,7 +273,7 @@ int pfs_destroy(struct pfs_node *pn);
/*
* Now for some initialization magic...
*/
-#define PSEUDOFS(name, version)
\
+#define PSEUDOFS(name, version, jflag) \
\
static struct pfs_info name##_info = { \
#name, \
@@ -281,6 +283,8 @@ static struct pfs_info name##_info = {
\
static int \
_##name##_mount(struct mount *mp) {\
+if (jflag && !prison_allow(curthread->td_ucred, jflag))
\
+return (EPERM);
\
return pfs_mount(&name##_info, mp); \
} \
\
@@ -303,7 +307,7 @@ static struct vfsops name##_vfsops = {
.vfs_uninit = _##name##_uninit, \
.vfs_unmount = pfs_unmount,\
}; \
-VFS_SET(name##_vfsops, name, VFCF_SYNTHETIC); \
+VFS_SET(name##_vfsops, name, VFCF_SYNTHETIC | (jflag ? VFCF_JAIL : 0));
\
MODULE_VERSION(name, version); \
MODULE_DEPEND(name, pseudofs, 1, 1, 1);
Modified: head/sys/kern/kern_jail.c
==
--- head/sys/kern/kern_jail.c Tue Feb 28 23:30:19 2012(r232277)
+++ head/sys/kern/kern_jail.c Wed Feb 29 00:30:18 2012(r232278)
@@ -204,6 +204,7 @@ static char *pr_allow_names[] = {
"allow.mount.devfs",
"allow.mount.nullfs",
"allow.mount.zfs",
+ "allow.mount.procfs",
};
const size_t pr_allow_names_size = sizeof(pr_allow_names);
@@ -218,6 +219,7 @@ static char *pr_allow_nonames[] = {
"allow.mount.nodevfs",
"allow.mount.nonullfs",
"allow.mount.nozfs",
+ "allow.mount.noprocfs",
};
const size_t pr_allow_nonames_size = sizeof(pr_allow_nonames);
@@ -4206,6 +4208,10 @@ SYSCTL_PROC(_security_jail, OID_AUTO, mo
CTLTYPE_INT | CTLFLAG_RW | CTLFLAG_MPSAFE,
NULL, PR_ALLOW_MOUNT_NULLFS, sysctl_jail_default_allow, "I",
"Processes in jail
Re: svn commit: r232278 - in head: sys/compat/linprocfs sys/compat/linsysfs sys/fs/procfs sys/fs/pseudofs sys/kern sys/sys usr.sbin/jail
On 29. Feb 2012, at 00:30 , Martin Matuska wrote: > Author: mm > Date: Wed Feb 29 00:30:18 2012 > New Revision: 232278 > URL: http://svn.freebsd.org/changeset/base/232278 > > Log: > Add procfs to jail-mountable filesystems. > The man page lacks a .Dd update? I also think this one should come with a very big red warning in the man page that you can easily compromise your host security I fear unless things changed in "proc" land. > Reviewed by: jamie > MFC after: 1 week > > Modified: > head/sys/compat/linprocfs/linprocfs.c > head/sys/compat/linsysfs/linsysfs.c > head/sys/fs/procfs/procfs.c > head/sys/fs/pseudofs/pseudofs.h > head/sys/kern/kern_jail.c > head/sys/sys/jail.h > head/usr.sbin/jail/jail.8 .. > Modified: head/usr.sbin/jail/jail.8 > == > --- head/usr.sbin/jail/jail.8 Tue Feb 28 23:30:19 2012(r232277) > +++ head/usr.sbin/jail/jail.8 Wed Feb 29 00:30:18 2012(r232278) > @@ -428,6 +428,14 @@ This permission is effective only togeth > and if > .Va enforce_statfs > is set to a value lower than 2. > +.It Va allow.mount.procfs > +privileged users inside the jail will be able to mount and unmount the > +procfs file system. > +This permission is effective only together with > +.Va allow.mount > +and if > +.Va enforce_statfs > +is set to a value lower than 2. > .It Va allow.mount.zfs > privileged users inside the jail will be able to mount and unmount the > ZFS file system. -- Bjoern A. Zeeb You have to have visions! It does not matter how good you are. It matters what good you do! ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r232266 - in head/sys: amd64/include i386/include pc98/include x86/include
On Tue, 28 Feb 2012, Tijl Coosemans wrote:
Log:
Copy amd64 endian.h to x86 and merge with i386 endian.h. Replace
amd64/i386/pc98 endian.h with stubs.
In __bswap64_const(x) the conflict between 0xffUL and 0xffULL has been
resolved by reimplementing the macro in terms of __bswap32(x). As a side
effect __bswap64_var(x) is now implemented using two bswap instructions on
i386 and should be much faster. __bswap32_const(x) has been reimplemented
in terms of __bswap16(x) for consistency.
The files now have enough in common for me to be happy with centralizing
them. The trip through would have to be avoided to complete
the cleanup. I don't know how to do that. Maybe install can produce
a direct symlink. We should try harder to not install both x86/foo.h
and machine/foo.h, so pathnames and don't start being
used in applications and then having to be supported for compatibility.
The patch is unreadable due to significant reimplementations, even without
the move. So I looked at the resulting x86/endian.h:
% ...
% #ifdef __cplusplus
% extern "C" {
% #endif
This ifdef is old style bug:
- it is a hard-coded spelling of __BEGIN_DECLS
- it is placed around the whole file. Most of the file consists of macro
definitions. __BEGIN_DECLS would obviously make no sense around macro
definitions.
- the rest of the file consists of inline function declarations. Since
these aren't extern, 'extern "C"' makes no sense for them, though it
isn't so clear that __BEGIN_DECLS makes no sense for them.
- the log message for the commit that added it is not good, but at least
says that C++ support is added.
% #if defined(__GNUCLIKE_ASM) && defined(__GNUCLIKE_BUILTIN_CONSTANT_P)
This ifdef is still misplaced. Defining the 'const' versions doesn't
require the gcc builtin, and it is further from requiring gcc asm.
%
% #define __bswap64_const(_x) \
% (((_x) >> 56) | \
% (((_x) >> 40) & (0xffULL << 8)) | \
% (((_x) >> 24) & (0xffULL << 16)) | \
% (((_x) >> 8) & (0xffULL << 24)) | \
% (((_x) << 8) & (0xffULL << 32)) | \
% (((_x) << 24) & (0xffULL << 40)) | \
% (((_x) << 40) & (0xffULL << 48)) | \
% ((_x) << 56))
I think this can be simplified in the same way as the `var' case should
have been simplified, by building it up out of 2 __bswap32_const()s...
%
% #define __bswap32_const(_x) \
% (((_x) >> 24) | \
% (((_x) & (0xff << 16)) >> 8) | \
% (((_x) & (0xff << 8)) << 8) | \
% ((_x) << 24))
... and this can be built out of 2 __bswap16_const()s', though it is
short enough already. Building up would be essential for __bswap_2^N
and good for N > 64.
%
% #define __bswap16_const(_x) (__uint16_t)((_x) << 8 | (_x) >> 8)
%
% static __inline __uint64_t
% __bswap64_var(__uint64_t __x)
% {
%
% return __bswap64_const(__x);
% }
This seems to be a pessimization for amd64, and doesn't match your
claim that it now uses 2 bswap instructions on i386. It is the old
i386 version. amd64 used to use a single 64-bit bswap in asm here.
Last time I looked (only on i386), gcc was incapable of turning either
__bswap64_const() or __bswap32_const() into 2 or 1 bswap instructions,
but generated pessimial code using shifts and masks. Only clang
generated the 2 or 1 bswaps.
There is a PR about bswap64 being pessimal on i386. The following seems
to be optimal for gcc on i386. It uses the building-up method. See the
PR for other details.
@ #define _KERNEL
@ #define I486_CPU
@
@ #include
@ #include
@
@ uint64_t x, y;
@
@ int
@ main(void)
@ {
@ /*
@* Follow is a better __bswap64(). It doesn't even need const and
@* var subcases.
@*/
@ y = __bswap32(x) | (uint64_t)__bswap32(x >> 32) << 32;
@ return (0);
@ }
Expanding on its comment, we probably don't even need the const subcase
for consts, but can just use the above expression for consts too.
However, this only works right for i386. For amd64, there are the
following cases:
- const case, compiler either gcc or clang: the above expression should
work
- variable case, gcc: needs the inline with a 64-bit bswap, unless gcc
has become smarter
- variable case, clang: the above expression won't work if __bswap32()
uses an inline asm with a 32-bit bswap like it does now, since clang
will be forced to generate 2 32-bit bswaps. To recover to single
64-bit bswap, either write it in inline asm for clang too, or maybe
use the general 64-bit 'const' expression (written like it is now,
not as in my version above). I think clang translates the big const
expression into a single 64-bit bswap, not 2 32-bit ones like it does
on i386.
%
%
Extra blank line (old style bug duplicated).
% ...
% #define __bswap64(_x) \
% (__builtin_constant_p(_x) ? \
% __bswap64_const((__uint64_t)(_x)) : __bsw
Re: svn commit: r232266 - in head/sys: amd64/include i386/include pc98/include x86/include
On Wed, 29 Feb 2012, Bruce Evans wrote:
On Tue, 28 Feb 2012, Tijl Coosemans wrote:
The patch is unreadable due to significant reimplementations, even without
the move. So I looked at the resulting x86/endian.h:
Oops. I somehow looked at the old i386 version. The committed version
is mostly OK, and fixes or avoids most of the bugs that I pointed out:
% #ifdef __cplusplus
% extern "C" {
% #endif
Still have this.
% #define __bswap16_const(_x) (__uint16_t)((_x) << 8 | (_x) >> 8)
%
% #define __bswap16(_x) \
% (__builtin_constant_p(_x) ? \
% __bswap16_const((__uint16_t)(_x)) : __bswap16_var(_x))
There seems to be no need for the const and var cases for 16 bits (use
the expression for the const case unconditionally, with more casts in
it).
We haven't remove the bogus underscores in macro parameter names yet.
%
% #define __bswap32_const(_x) \
% (((__uint32_t)__bswap16(_x) << 16) | __bswap16((_x) >> 16))
Now uses "building-up".
%
% #define __bswap32(_x) \
% (__builtin_constant_p(_x) ? \
% __bswap32_const((__uint32_t)(_x)) : __bswap32_var(_x))
The technical considerations described in my previous mail prevent
killing the separate cases and using pure building up from 16 bits to
32 and 64 bits.
This is quite confusing. __bswap32_const() is perfectly general since
it uses the general __bswap16() macros, but it is spelled with a `const'.
OTOH, __bswap32() is perfectly general by name and functionality, but is
only useful for the `var' case.
% ...
% static __inline __uint16_t
% __bswap16_var(__uint16_t _x)
% {
%
% return (__bswap16_const(_x));
% }
Not needed, since we killed the asm for the 16-bit case.
%
% static __inline __uint32_t
% __bswap32_var(__uint32_t _x)
% {
%
% __asm ("bswap %0" : "+r" (_x));
% return (_x);
% }
%
% static __inline __uint64_t
% __bswap64_var(__uint64_t _x)
% {
% #ifdef_LP64
Tab instead of space after ifdef.
I think it is gcc and not _LP64 that needs this help.
% __asm ("bswap %0" : "+r" (_x));
% return (_x);
% #else
% return (__bswap64_const(_x));
The use of `const' here confused me into thinking that this still uses
the old slow `const' code.
% #endif
% }
% ...
% #ifdef __cplusplus
% }
% #endif
Still have this.
The best cleanups seem to be:
- use general building up with no asms and no inline functions for the
default case (put this in a MI header)
- use asms for gcc. I think the following works for minimizing the
ifdefs for this:
- start with bswap32(), etc., defined directly to the default using
building up (I'm trying to avoid having the __bswap32() layer too)
- to override this, just #undef it and redefine it with a gcc asm in
it (it seems best to use a statement expression and not an inline)
or:
- start with __bswap32() defined to the default using building up
- normally define bswap32 to __bswap32. Otherwise, define bswap32
to our special asm or C implementation.
or:
- some combination of the above. We can use __bswap32 in an inline
function with the same name, and then #undef the macro:
% #define __bswap32(x)(...)
%
% int
% (__bswap32)(int _x)
% {
% #ifdef asm_case
% {( asm(...
% #else
% return (__bswap32(_x));
% #endif
% }
%
% #undef __bswap32
This technique permits the default macros to use the best names and
not require adorning them with _const or _var or maybe even __ and
then needing more layers to #define them back to the best names.
I don't know how to do this without turning the macro into an inline
function. There is related magic for macros that I forget.
Bruce
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r232286 - head/sys/kern
Author: davidxu
Date: Wed Feb 29 02:01:48 2012
New Revision: 232286
URL: http://svn.freebsd.org/changeset/base/232286
Log:
initialize clock ID and flags only when copying timespec, a _umtx_time
copy already contains these fields.
Modified:
head/sys/kern/kern_umtx.c
Modified: head/sys/kern/kern_umtx.c
==
--- head/sys/kern/kern_umtx.c Wed Feb 29 01:39:39 2012(r232285)
+++ head/sys/kern/kern_umtx.c Wed Feb 29 02:01:48 2012(r232286)
@@ -2937,11 +2937,11 @@ umtx_copyin_umtx_time(const void *addr,
{
int error;
- tp->_clockid = CLOCK_REALTIME;
- tp->_flags = 0;
- if (size <= sizeof(struct timespec))
+ if (size <= sizeof(struct timespec)) {
+ tp->_clockid = CLOCK_REALTIME;
+ tp->_flags = 0;
error = copyin(addr, &tp->_timeout, sizeof(struct timespec));
- else
+ } else
error = copyin(addr, tp, sizeof(struct _umtx_time));
if (error != 0)
return (error);
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r232183 - head/sys/kern
On Tue, 28 Feb 2012, Alexander Best wrote: On Sun Feb 26 12, Jilles Tjoelker wrote: Log: Fix fchmod() and fchown() on fifos. The new fifo implementation in r232055 broke fchmod() and fchown() on fifos. Postfix needs this. clang seems to have a problem with this commit: usr/subversion-src/sys/kern/sys_pipe.c:1556:10: error: promoted type 'int' of K&R function parameter is not compatible with the parameter type 'mode_t' (aka 'unsigned short') declared in a previous prototype [-Werror] mode_t mode; ^ /usr/subversion-src/sys/kern/sys_pipe.c:155:19: note: previous declaration is here static fo_chmod_t pipe_chmod; ^ 1 error generated. *** Error code 1 Indeed, sys_pipe.c still uses mostly K&R-style function definitions, and the recent change was compatible with this to a fault. It is an old bug in gcc that it accepts function definitions that don't match their prototype. The most common error is the one here -- the function definition is K&R and has a type smaller than its promotion; the prototype must give the promoted type to match, but this breaks it for use with function definitions that are not K&R and use the narrow type. It is an old design error to have APIs that pass narrowed types. Passing narrowed types is normally just a pessimization. The ABI normally requires passing a widened value. The widened value may have garbage in the extra bits, but for safety and to support missing prototypes, in practice callers must sign-extend or zero-extend the narrow value. The for safety and to support missing prototypes, in practice callees can't trust callers to have done this, so they force to a narrow value. Then to actually use the value, they often have to widen it again. The result is often a couple of extra instructions after optimizing away only half of the logical extra conversions for this. So the supposed optimization of passing narrow types is usually a pessimization. Maybe this is a good pessimization for private APIs, but it shouldn't be required for external ones. Narrow types in syscalls are especially bad. Now the C API doesn't really extend across the syscall boundary, so both sides have to force the above conversions or depend on the compiler doing them because it fears missing prototypes. In practice, FreeBSD depends on the compiler doing them for most or all arches. The narrow mode_t for open(2) is worst. It is the only narrow type in in a core syscall, and has further problems. The prototype for open() is "int open(const char *path, int flags, ...);", with the only option for "..." being a mode_t. open() is thus variadic, but variadic functions really can't be passed a narrowed type like mode_t, and syscalls can't really pass this either. FreeBSD used to depend on various type puns near here, and still has my XXX comments in syscalls.master about them, but some of them have been fixed. The largest one is that the kernel assumes that the mode_t arg is always there. It copies stack garbage if it is not there, and only later checks the flags and avoids using the stack garbage when there is no arg there. This is not too bad if done intentionally. It must be arranged that the stack garbage is in accessible memory, but it always is in practice. The kernel is carefully to access the arg as an int and not as a mode_t, and to pass it down through too many layers as an int. Bruce ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r232275 - in head/sys: amd64/include i386/include pc98/include x86/include
On Tue, 28 Feb 2012, Tijl Coosemans wrote:
Log:
Copy amd64 setjmp.h to x86 and replace amd64/i386/pc98 setjmp.h with stubs.
This may be correct (except for comment), but it is confusing.
Added:
head/sys/x86/include/setjmp.h
- copied unchanged from r232268, head/sys/amd64/include/setjmp.h
Modified:
head/sys/amd64/include/setjmp.h
head/sys/i386/include/setjmp.h
head/sys/pc98/include/setjmp.h
Modified: head/sys/amd64/include/setjmp.h
==
--- head/sys/amd64/include/setjmp.h Tue Feb 28 22:15:46 2012
(r232274)
+++ head/sys/amd64/include/setjmp.h Tue Feb 28 22:17:52 2012
(r232275)
@@ -1,50 +1,6 @@
...
-#define_JBLEN 12 /* Size of the jmp_buf on AMD64. */
-
-/*
- * jmp_buf and sigjmp_buf are encapsulated in different structs to force
- * compile-time diagnostics for mismatches. The structs are the same
- * internally to avoid some run-time errors for mismatches.
- */
-#if __BSD_VISIBLE || __POSIX_VISIBLE || __XSI_VISIBLE
-typedefstruct _sigjmp_buf { long _sjb[_JBLEN]; } sigjmp_buf[1];
-#endif
-
-typedefstruct _jmp_buf { long _jb[_JBLEN]; } jmp_buf[1];
On amd64, the comment was specific to amd64 (but with amd64 misspelled
as AMD64), and actually matched the code.
Modified: head/sys/i386/include/setjmp.h
==
--- head/sys/i386/include/setjmp.h Tue Feb 28 22:15:46 2012
(r232274)
+++ head/sys/i386/include/setjmp.h Tue Feb 28 22:17:52 2012
(r232275)
@@ -1,50 +1,6 @@
-#define_JBLEN 11 /* Size of the jmp_buf on x86. */
-
-/*
- * jmp_buf and sigjmp_buf are encapsulated in different structs to force
- * compile-time diagnostics for mismatches. The structs are the same
- * internally to avoid some run-time errors for mismatches.
- */
-#if __BSD_VISIBLE || __POSIX_VISIBLE || __XSI_VISIBLE
-typedefstruct _sigjmp_buf { int _sjb[_JBLEN + 1]; } sigjmp_buf[1];
-#endif
-
-typedefstruct _jmp_buf { int _jb[_JBLEN + 1]; } jmp_buf[1];
On i386, the comment had a differently wrong name for the old code, and
mismatched the old code, but is correct for the new code.
_JBLEN was apparently 1 less than what it was claimed to be on i386.
I forget what the extra 1 is for.
Copied: head/sys/x86/include/setjmp.h (from r232268,
head/sys/amd64/include/setjmp.h)
==
--- /dev/null 00:00:00 1970 (empty, because file is newly added)
+++ head/sys/x86/include/setjmp.h Tue Feb 28 22:17:52 2012
(r232275, copy of r232268, head/sys/amd64/include/setjmp.h)
@@ -0,0 +1,50 @@
...
+#define_JBLEN 12 /* Size of the jmp_buf on AMD64. */
This code looks wrong, since _JBLEN was only 11 for i386, but binary
compatibility of the jmp_buf is preserved by the extra 1 in it.
This comment is wrong, since the code is no longer just for AMD64 (sic),
and the comment makes it look more like the code has an off-by-1 error
for !AMD64.
+
+/*
+ * jmp_buf and sigjmp_buf are encapsulated in different structs to force
+ * compile-time diagnostics for mismatches. The structs are the same
+ * internally to avoid some run-time errors for mismatches.
+ */
+#if __BSD_VISIBLE || __POSIX_VISIBLE || __XSI_VISIBLE
+typedefstruct _sigjmp_buf { long _sjb[_JBLEN]; } sigjmp_buf[1];
+#endif
+
+typedefstruct _jmp_buf { long _jb[_JBLEN]; } jmp_buf[1];
No extra 1 for either now.
The extra 1 was apparently a placeholder for the signal mask and became
obsolete 15-20 years ago. 4.4BSD-Lite2 claims not to support the signal
mask, and doesn't have space for it in the plain jmp_buf. setjmp.h
was in src/include and was moved to MD files to clean up its ifdefs.
Now we're barely escaping putting the ifdefs back in the x86 version,
sigh. Ifdefs are really needed, since the i386 version is too small
to have space for mxcsr, and mxcsr is only hacked into the amd64 version
by packing it using the doubled space given by long entries.
From the 4.4BSD-Lite2 version:
% #if defined(hp300) || defined(__hp300__) || defined(luna68k) ||
defined(__luna68k__)
% #define _JBLEN17
% #endif
%
% #if defined(i386) || defined(__i386__)
% #define _JBLEN10
% #endif
%
% #if defined(mips) || defined(__mips__)
% #define _JBLEN83
% #endif
%
% #if defined(sparc) || defined(__sparc__)
% #define _JBLEN10
% #endif
%
% #if defined(tahoe) || defined(__tahoe__)
% #define _JBLEN10
% #endif
%
% #if defined(vax) || defined(__vax__)
% #define _JBLEN10
% #endif
%
% #ifndef _ANSI_SOURCE
% /*
% * WARNING: sigsetjmp() isn't supported yet, this is a placeholder.
% */
% typedef int sigjmp_buf[_JBLEN + 1];
% #endif /* not ANSI */
%
% typedef int jmp_buf[_JBLEN];
The extra 1 seems to have been nonsense there too. BSD has always (?)
saved the signal m
Re: svn commit: r232266 - in head/sys: amd64/include i386/include pc98/include x86/include
On Tue, 28 Feb 2012, Tijl Coosemans wrote:
On Tuesday 28 February 2012 21:55:57 Dimitry Andric wrote:
On 2012-02-28 20:39, Tijl Coosemans wrote:
Log:
Copy amd64 endian.h to x86 and merge with i386 endian.h. Replace
amd64/i386/pc98 endian.h with stubs.
In __bswap64_const(x) the conflict between 0xffUL and 0xffULL has been
resolved by reimplementing the macro in terms of __bswap32(x). As a side
effect __bswap64_var(x) is now implemented using two bswap instructions on
i386 and should be much faster. __bswap32_const(x) has been reimplemented
in terms of __bswap16(x) for consistency.
...
+#define__bswap32_const(_x) \
+ (((__uint32_t)__bswap16(_x) << 16) | __bswap16((_x) >> 16))
+
+#define__bswap32(_x) \
+ (__builtin_constant_p(_x) ? \
+ __bswap32_const((__uint32_t)(_x)) : __bswap32_var(_x))
+
+#define__bswap64_const(_x) \
+ (((__uint64_t)__bswap32(_x) << 32) | __bswap32((_x) >> 32))
Hmm, shouldn't __bswap32_const() be implemented in terms of
__bswap16_const(), and similarly for __bswap64_const(), which should be
implemented in terms of __bswap32_const()?
The whole reason for the difficult dance with __builtin_constant_p is to
allow for compile-time resolving of bswap'd constants. Invoking the
regular __bswap() macros doesn't gain you anything here.
__bswap64_const is also used in __bswap64_var, so its argument isn't
always a compile time constant and then __bswap32 becomes __bswap32_var.
If it is a constant __bswap32 becomes __bswap32_const (even at -O0).
I cleaned this up a bit according to ideas in my previous mails, and
added a comment about the confusing use of __bswap64_const() (now
named __bswap64_gen()) in __bswap64_var():
% Index: endian.h
% ===
% RCS file: /home/ncvs/src/sys/x86/include/endian.h,v
% retrieving revision 1.1
% diff -u -2 -r1.1 endian.h
% --- endian.h 28 Feb 2012 19:39:54 - 1.1
% +++ endian.h 29 Feb 2012 03:10:33 -
% @@ -37,8 +37,4 @@
% #include
%
% -#ifdef __cplusplus
% -extern "C" {
% -#endif
% -
% /*
% * Define the order of 32-bit words in 64-bit words.
% @@ -68,32 +64,24 @@
% #endif
%
% -#if defined(__GNUCLIKE_ASM) && defined(__GNUCLIKE_BUILTIN_CONSTANT_P)
% -
% -#define __bswap16_const(_x) (__uint16_t)((_x) << 8 | (_x) >> 8)
% -
% -#define __bswap16(_x) \
% - (__builtin_constant_p(_x) ? \
% - __bswap16_const((__uint16_t)(_x)) : __bswap16_var(_x))
% -
% -#define __bswap32_const(_x) \
% - (((__uint32_t)__bswap16(_x) << 16) | __bswap16((_x) >> 16))
% -
% -#define __bswap32(_x) \
% - (__builtin_constant_p(_x) ? \
% - __bswap32_const((__uint32_t)(_x)) : __bswap32_var(_x))
% +#define ___bswap16(x) (__uint16_t)((x) << 8 | (x) >> 8)
% +#define __bswap16(x)(___bswap16((__uint16_t)(x)))
% +#define __bswap32_gen(x)\
% + (((__uint32_t)__bswap16(x) << 16) | __bswap16((x) >> 16))
% +#define __bswap64_gen(x)\
% + (((__uint64_t)__bswap32(x) << 32) | __bswap32((x) >> 32))
%
% -#define __bswap64_const(_x) \
% - (((__uint64_t)__bswap32(_x) << 32) | __bswap32((_x) >> 32))
% -
% -#define __bswap64(_x) \
% - (__builtin_constant_p(_x) ? \
% - __bswap64_const((__uint64_t)(_x)) : __bswap64_var(_x))
% +#if defined(__GNUCLIKE_ASM) && defined(__GNUCLIKE_BUILTIN_CONSTANT_P)
%
% -static __inline __uint16_t
% -__bswap16_var(__uint16_t _x)
% -{
% +/*
% + * The following messes are old optimizations for gcc. The messes have
% + * been reduced significantly, but I don't know how to implement the
% + * preferred way of defining defaults above and overriding them cleanly
% + * here. Even clang needs help for the 64-bit case, so to keep the
% + * ifdefs sane we use this for the 32-bit case with clang too.
% + */
%
% - return (__bswap16_const(_x));
% -}
% +#define __bswap32(x)\
% + (__builtin_constant_p(x) ? \
% + __bswap32_gen((__uint32_t)(x)) : __bswap32_var(x))
%
% static __inline __uint32_t
% @@ -105,34 +93,35 @@
% }
%
% +#define __bswap64(x) \
% + (__builtin_constant_p(x) ? \
% + __bswap64_gen((__uint64_t)(x)) : __bswap64_var(x))
% +
% static __inline __uint64_t
% __bswap64_var(__uint64_t _x)
% {
% -#ifdef _LP64
% +
% +#ifdef _LP64
% __asm ("bswap %0" : "+r" (_x));
% return (_x);
% #else
% - return (__bswap64_const(_x));
% + /*
% + * It is important for the optimizations that the following is not
% + * really generic, but expands to 2 __bswap32_var()'s.
% + */
% + return (__bswap64_gen(_x));
% #endif
% }
%
% -#define __htonl(x) __bswap32(x)
% -#define __htons(x) __bswap16(x)
% -#define __ntohl(x) __bswap32(x)
% -#define __ntohs(x) _
svn commit: r232288 - head/sys/vm
Author: alc
Date: Wed Feb 29 05:41:29 2012
New Revision: 232288
URL: http://svn.freebsd.org/changeset/base/232288
Log:
Simplify kmem_alloc() by eliminating code that existed on account of
external pagers in Mach. FreeBSD doesn't implement external pagers.
Moreover, it don't pageout the kernel object. So, the reasons for
having code don't hold.
Reviewed by: kib
MFC after:6 weeks
Modified:
head/sys/vm/vm_kern.c
Modified: head/sys/vm/vm_kern.c
==
--- head/sys/vm/vm_kern.c Wed Feb 29 02:04:41 2012(r232287)
+++ head/sys/vm/vm_kern.c Wed Feb 29 05:41:29 2012(r232288)
@@ -161,7 +161,6 @@ kmem_alloc(map, size)
{
vm_offset_t addr;
vm_offset_t offset;
- vm_offset_t i;
size = round_page(size);
@@ -187,35 +186,6 @@ kmem_alloc(map, size)
vm_map_unlock(map);
/*
-* Guarantee that there are pages already in this object before
-* calling vm_map_wire. This is to prevent the following
-* scenario:
-*
-* 1) Threads have swapped out, so that there is a pager for the
-* kernel_object. 2) The kmsg zone is empty, and so we are
-* kmem_allocing a new page for it. 3) vm_map_wire calls vm_fault;
-* there is no page, but there is a pager, so we call
-* pager_data_request. But the kmsg zone is empty, so we must
-* kmem_alloc. 4) goto 1 5) Even if the kmsg zone is not empty: when
-* we get the data back from the pager, it will be (very stale)
-* non-zero data. kmem_alloc is defined to return zero-filled memory.
-*
-* We're intentionally not activating the pages we allocate to prevent a
-* race with page-out. vm_map_wire will wire the pages.
-*/
- VM_OBJECT_LOCK(kernel_object);
- for (i = 0; i < size; i += PAGE_SIZE) {
- vm_page_t mem;
-
- mem = vm_page_grab(kernel_object, OFF_TO_IDX(offset + i),
- VM_ALLOC_NOBUSY | VM_ALLOC_ZERO | VM_ALLOC_RETRY);
- mem->valid = VM_PAGE_BITS_ALL;
- KASSERT((mem->oflags & VPO_UNMANAGED) != 0,
- ("kmem_alloc: page %p is managed", mem));
- }
- VM_OBJECT_UNLOCK(kernel_object);
-
- /*
* And finally, mark the data as non-pageable.
*/
(void) vm_map_wire(map, addr, addr + size,
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r232289 - in head/sys/mips/cavium: . octe
Author: gonzo
Date: Wed Feb 29 05:48:29 2012
New Revision: 232289
URL: http://svn.freebsd.org/changeset/base/232289
Log:
Revert part of old logic of assigning MAC addressess:
- Reserver respective number of addresses for managment port
- octm uses base address directly
- other drivers get MACs on "first come first served" basis
Reviewed by: juli
Modified:
head/sys/mips/cavium/files.octeon1
head/sys/mips/cavium/if_octm.c
head/sys/mips/cavium/octe/ethernet-common.c
Modified: head/sys/mips/cavium/files.octeon1
==
--- head/sys/mips/cavium/files.octeon1 Wed Feb 29 05:41:29 2012
(r232288)
+++ head/sys/mips/cavium/files.octeon1 Wed Feb 29 05:48:29 2012
(r232289)
@@ -24,7 +24,6 @@ mips/cavium/cryptocteon/cryptocteon.c o
mips/mips/octeon_cop2_swtch.S standard
mips/mips/octeon_cop2.cstandard
-# octm must be first, so management ports get the first MAC addresses
mips/cavium/if_octm.c optional octm
contrib/octeon-sdk/cvmx-mgmt-port.coptional octm
Modified: head/sys/mips/cavium/if_octm.c
==
--- head/sys/mips/cavium/if_octm.c Wed Feb 29 05:41:29 2012
(r232288)
+++ head/sys/mips/cavium/if_octm.c Wed Feb 29 05:48:29 2012
(r232289)
@@ -63,7 +63,6 @@
#include
#include
#include
-#include "octe/ethernet-common.h"
struct octm_softc {
struct ifnet *sc_ifp;
@@ -177,10 +176,10 @@ octm_attach(device_t dev)
/*
* Set MAC address for this management port.
*/
- if (cvm_assign_mac_address(&mac, NULL) != 0) {
- device_printf(dev, "unable to allocate MAC address.\n");
- return (ENXIO);
- }
+ mac = 0;
+ memcpy((u_int8_t *)&mac + 2, cvmx_sysinfo_get()->mac_addr_base, 6);
+ mac += sc->sc_port;
+
cvmx_mgmt_port_set_mac(sc->sc_port, mac);
/* No watermark for input ring. */
Modified: head/sys/mips/cavium/octe/ethernet-common.c
==
--- head/sys/mips/cavium/octe/ethernet-common.c Wed Feb 29 05:41:29 2012
(r232288)
+++ head/sys/mips/cavium/octe/ethernet-common.c Wed Feb 29 05:48:29 2012
(r232289)
@@ -46,8 +46,8 @@ __FBSDID("$FreeBSD$");
extern int octeon_is_simulation(void);
-static uint64_t mac_addr = 0;
-static uint32_t mac_offset = 0;
+static uint64_t cvm_oct_mac_addr = 0;
+static uint32_t cvm_oct_mac_addr_offset = 0;
/**
* Set the multicast list. Currently unimplemented.
@@ -102,22 +102,42 @@ void cvm_oct_common_set_multicast_list(s
int cvm_assign_mac_address(uint64_t *macp, uint8_t *octets)
{
/* Initialize from global MAC address base; fail if not set */
- if (mac_addr == 0) {
- memcpy((uint8_t *)&mac_addr + 2,
cvmx_sysinfo_get()->mac_addr_base, 6);
- if (mac_addr == 0)
+ if (cvm_oct_mac_addr == 0) {
+ memcpy((uint8_t *)&cvm_oct_mac_addr + 2,
+ cvmx_sysinfo_get()->mac_addr_base, 6);
+
+ if (cvm_oct_mac_addr == 0)
return ENXIO;
+
+ /*
+* The offset from mac_addr_base that should be used for the
next port
+* that is configured. By convention, if any mgmt ports exist
on the
+* chip, they get the first mac addresses. The ports
controlled by
+* driver that use this function are numbered sequencially
following
+* any mgmt addresses that may exist.
+*
+* XXX Would be nice if __cvmx_mgmt_port_num_ports() were
+* not static to cvmx-mgmt-port.c.
+*/
+ if (OCTEON_IS_MODEL(OCTEON_CN56XX))
+ cvm_oct_mac_addr_offset = 1;
+ else if (OCTEON_IS_MODEL(OCTEON_CN52XX) ||
OCTEON_IS_MODEL(OCTEON_CN63XX))
+ cvm_oct_mac_addr_offset = 2;
+ else
+ cvm_oct_mac_addr_offset = 0;
+ cvm_oct_mac_addr += cvm_oct_mac_addr_offset;
}
- if (mac_offset >= cvmx_sysinfo_get()->mac_addr_count)
+ if (cvm_oct_mac_addr_offset >= cvmx_sysinfo_get()->mac_addr_count)
return ENXIO; /* Out of addresses to assign */
if (macp)
- *macp = mac_addr;
+ *macp = cvm_oct_mac_addr;
if (octets)
- memcpy(octets, (u_int8_t *)&mac_addr + 2, 6);
+ memcpy(octets, (u_int8_t *)&cvm_oct_mac_addr + 2, 6);
- mac_addr++;
- mac_offset++;
+ cvm_oct_mac_addr++;
+ cvm_oct_mac_addr_offset++;
return 0;
}
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mai
svn commit: r232291 - head/usr.sbin/jail
Author: mm Date: Wed Feb 29 07:33:07 2012 New Revision: 232291 URL: http://svn.freebsd.org/changeset/base/232291 Log: Bump .Dd to reflect latest update Reported by: bz MFC after:1 week Modified: head/usr.sbin/jail/jail.8 Modified: head/usr.sbin/jail/jail.8 == --- head/usr.sbin/jail/jail.8 Wed Feb 29 06:19:00 2012(r232290) +++ head/usr.sbin/jail/jail.8 Wed Feb 29 07:33:07 2012(r232291) @@ -34,7 +34,7 @@ .\" .\" $FreeBSD$ .\" -.Dd February 26, 2012 +.Dd February 29, 2012 .Dt JAIL 8 .Os .Sh NAME ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r232266 - in head/sys: amd64/include i386/include pc98/include x86/include
On Wed, 29 Feb 2012, Bruce Evans wrote:
I cleaned this up a bit according to ideas in my previous mails, and
added a comment about the confusing use of __bswap64_const() (now
named __bswap64_gen()) in __bswap64_var():
A minor problem with only having a macro version for __bswap64() turned
up:
% -#define __bswap16_const(_x) (__uint16_t)((_x) << 8 | (_x) >> 8)
% -
% -#define __bswap16(_x) \
% - (__builtin_constant_p(_x) ? \
% - __bswap16_const((__uint16_t)(_x)) : __bswap16_var(_x))
...
% +#define ___bswap16(x) (__uint16_t)((x) << 8 | (x) >> 8)
% +#define __bswap16(x)(___bswap16((__uint16_t)(x)))
When x a non-volatile variable, gcc and clang produce the good code
"rolw $8,x" for "x = __bswap16(x);" on short x. But when x a a volatile
variable, gcc and clang produce fairly horrible code, with 2 loads of
x corresponding to the 2 accesses to x. This is probably required by
volatile semantics, and is a problem for all unsafe macros, especially
when their name says that they are safe (oops). When __bswap16 is
implemented as an inline function for the var case like it used to be,
it only loads x once and there are no problems with volatile variables.
Optimizing to "rolw $8,x" might still be possible iff x is not volatile,
but load-modify-store is probably better anyway.
So any macro version must use gcc features to be safe. The following
seems to work:
#define __bswap16(x)__extension__ ({ __uint16_t __x = x;
(__uint16_t)(__x << 8 | __x >> 8); })
clang now produces "rolw $8,x" when x is volatile. This seems to
violate volatile semantics. gcc produces load-rolw-store then. Both
produce "rolw $8,x" when x is not volatile.
Bruce
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r232271 - head/sys/kern
On 2012-02-28 23:11, Giovanni Trematerra wrote: > On Tue, Feb 28, 2012 at 10:45 PM, Dimitry Andric wrote: >> Author: dim >> Date: Tue Feb 28 21:45:21 2012 >> New Revision: 232271 >> URL: http://svn.freebsd.org/changeset/base/232271 ... > It introduces a style(9) violation, though. > I think it should be in this way > > pipe_chmod(struct file *fp, mode_t mode, struct ucred *active_cred, > struct thread *td) > > Anyway that's not a problem I can take care of this and other style violations > during next weeks. Is there any interest in changing the other K&R definitions in this file to C99? Or does that cause too much churn? I tried to make this change as minimal as possible to make it compile. Btw, style(9) doesn't say anything about a maximum line width, maybe that should be added explicitly...:) ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r232181 - in head/sys: kern sys
On Tue, 28 Feb 2012 16:00:36 -0800 Julian Elischer wrote: JE> On 2/27/12 11:29 PM, Mikolaj Golub wrote: >> On Mon, 27 Feb 2012 22:34:25 -0800 Julian Elischer wrote: >> >> JE> I don't think this belongs in the kernel by default. It's not exactl >> a >> JE> call for backout but It's teh next thing short of that. a call for >> "do >> JE> you REALLY think we need this particular specific case catered for?" >> >> The main goal of the patch was to provide ability to get another process >> umask. It looks like usefulness of this is not questioned here. JE> well that's exactly what I AM questioning.. how often will this be used? JE> one person using this once in all of history isn't a real requirement JE> for inclusion. This information may be very useful when troubleshooting unexpected behavior of the application. Dmitry Banschikov, who was asking for this functionality and eventually provided the patch, said that it needed this investigating an issue with an application which created files with unexpected permissions. It turned out the issue was with wrong usage of su(1), which may interpret '-c' option as a login class or as a command to run, so the umask specified in the login class was not applied. Then it wrote an utility to read a process umask via kvm to troubleshoot this. I don't think this situation is in the class "one person using this once in all of history". In my practice I have not face a situation when I need to know umask of another process and it will be good if I never need this. But if I need it eventually I would like to have a quick and easy way to do this. Also for me after applying the patch 'procstat -sa' output on my hosts was rather educational. JE> It seems to me that someone is more likely to figure out a sneaky way JE> to use this in a bad way than to want to use it in the way you expect. Being this someone I would use much easier sneaky ways to make a mess for processes running with my uid. -- Mikolaj Golub ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
