Re: svn commit: r353937 - in head/share: man/man5 mk
On 27 Oct 2019, at 17:59, Alexey Dokuchaev wrote: > > On Sat, Oct 26, 2019 at 04:34:14PM +0200, Dimitry Andric wrote: ... >> I only tested -j24 on a 32-core system, but I could probably repeat the >> experiment with lower and higher -j values: [...] >> >> So ~2.3% difference in real time, which is not too bad I think. > > Well, I'd say it's acceptable. :-/ I also tested at low (-j4) and high (-j32) levels. It turns out that at low -j levels, the difference is less pronounced, just ~1.1% in real time. And at high -j levels, it is more pronounced, ~4.3% in real time. Note also that at high -j levels, the difference in system time seems to get more influence, e.g. at low -j the difference is ~4.5%, while at high -j the difference is ~7.2%. I guess it is because dynamic linking uses more syscalls. At -j4: Results for real time: --- x j4-static-real.txt + j4-dynamic-real.txt N Min MaxMedian AvgStddev x 7 5508.58 5548.05 5533.25 5530.2829 12.323855 + 7 5514.685660.1 5591.25 5593.5143 59.98042 Difference at 95.0% confidence 63.2314 +/- 50.4309 1.14337% +/- 0.912329% (Student's t, pooled s = 43.2985) Results for user time: --- x j4-static-user.txt + j4-dynamic-user.txt N Min MaxMedian AvgStddev x 7 19525.34 19607.74 19593.1 19582.78 29.037717 + 7 20178.84 20348.11 20214.28 20235.057 57.31409 Difference at 95.0% confidence 652.277 +/- 52.9155 3.33087% +/- 0.272077% (Student's t, pooled s = 45.4318) Results for system time: --- x j4-static-sys.txt + j4-dynamic-sys.txt N Min MaxMedian AvgStddev x 7 1622.63 1634.98 1629.87 1629.1229 4.7234017 + 7 1672.25 1722.23 1704.49 1703.1729 15.724862 Difference at 95.0% confidence 74.05 +/- 13.5224 4.54539% +/- 0.833228% (Student's t, pooled s = 11.6099) At -j32: Results for real time: --- x j32-static-real.txt + j32-dynamic-real.txt N Min MaxMedian AvgStddev x 7 1689.19 1735.36 1707.32 1707.8471 16.500078 + 7 1754.98 1812.57 1777.96 1781.6329 19.554159 Difference at 95.0% confidence 73.7857 +/- 21.0718 4.32039% +/- 1.25627% (Student's t, pooled s = 18.0917) Results for user time: --- x j32-static-user.txt + j32-dynamic-user.txt N Min MaxMedian AvgStddev x 7 38545.36 38802.6 38609.87 38641.009 105.11904 + 7 39430.93 39924.98 39856.91 39769.979 185.3612 Difference at 95.0% confidence 1128.97 +/- 175.5 2.92169% +/- 0.457446% (Student's t, pooled s = 150.68) Results for system time: --- x j32-static-sys.txt + j32-dynamic-sys.txt N Min MaxMedian AvgStddev x 7 2752.81 2809.97 2781.46 2781.4986 21.86523 + 7 2906.22 3151.93 2932.59 2981.5757 103.79842 Difference at 95.0% confidence 200.077 +/- 87.3629 7.19314% +/- 3.15079% (Student's t, pooled s = 75.0073) -Dimitry signature.asc Description: Message signed with OpenPGP
Re: svn commit: r353937 - in head/share: man/man5 mk
On Tue, Oct 29, 2019 at 08:04:07AM +0100, Dimitry Andric wrote: > On 27 Oct 2019, at 17:59, Alexey Dokuchaev wrote: > > > > On Sat, Oct 26, 2019 at 04:34:14PM +0200, Dimitry Andric wrote: > ... > >> I only tested -j24 on a 32-core system, but I could probably repeat the > >> experiment with lower and higher -j values: [...] > >> > >> So ~2.3% difference in real time, which is not too bad I think. > > > > Well, I'd say it's acceptable. :-/ > > I also tested at low (-j4) and high (-j32) levels. It turns out that at > low -j levels, the difference is less pronounced, just ~1.1% in real > time. And at high -j levels, it is more pronounced, ~4.3% in real time. > > Note also that at high -j levels, the difference in system time seems > to get more influence, e.g. at low -j the difference is ~4.5%, while at > high -j the difference is ~7.2%. I guess it is because dynamic linking > uses more syscalls. Yeah, it would be definitely nice if we could optimize runtime linker. Thanks for conducting these tests by the way, much appreciated, Dimitry! ./danfe ___ svn-src-all@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-all To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r354145 - in head/tests/sys/netpfil: common pf
Author: kp Date: Tue Oct 29 09:47:12 2019 New Revision: 354145 URL: https://svnweb.freebsd.org/changeset/base/354145 Log: netpfil tests: Add missing copyright & license statements Modified: head/tests/sys/netpfil/common/pft_ping.py head/tests/sys/netpfil/common/sniffer.py head/tests/sys/netpfil/pf/CVE-2019-5597.py head/tests/sys/netpfil/pf/CVE-2019-5598.py head/tests/sys/netpfil/pf/anchor.sh head/tests/sys/netpfil/pf/echo_inetd.conf head/tests/sys/netpfil/pf/forward.sh head/tests/sys/netpfil/pf/fragmentation.sh head/tests/sys/netpfil/pf/icmp.sh head/tests/sys/netpfil/pf/names.sh head/tests/sys/netpfil/pf/nat.sh head/tests/sys/netpfil/pf/pass_block.sh head/tests/sys/netpfil/pf/pfsync.sh head/tests/sys/netpfil/pf/rdr.sh head/tests/sys/netpfil/pf/route_to.sh head/tests/sys/netpfil/pf/set_skip.sh head/tests/sys/netpfil/pf/set_tos.sh head/tests/sys/netpfil/pf/synproxy.sh head/tests/sys/netpfil/pf/utils.subr Modified: head/tests/sys/netpfil/common/pft_ping.py == --- head/tests/sys/netpfil/common/pft_ping.py Tue Oct 29 04:28:24 2019 (r354144) +++ head/tests/sys/netpfil/common/pft_ping.py Tue Oct 29 09:47:12 2019 (r354145) @@ -1,4 +1,30 @@ #!/usr/bin/env python +# +# SPDX-License-Identifier: BSD-2-Clause +# +# Copyright (c) 2017 Kristof Provost +# +# 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. +# +# 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. +# import argparse import scapy.all as sp Modified: head/tests/sys/netpfil/common/sniffer.py == --- head/tests/sys/netpfil/common/sniffer.pyTue Oct 29 04:28:24 2019 (r354144) +++ head/tests/sys/netpfil/common/sniffer.pyTue Oct 29 09:47:12 2019 (r354145) @@ -1,4 +1,30 @@ # $FreeBSD$ +# +# SPDX-License-Identifier: BSD-2-Clause-FreeBSD +# +# Copyright (c) 2017 Kristof Provost +# +# 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. +# +# 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. +# import threading import scapy.all as sp Modified: head/tests/sys/netpfil/pf/CVE-2019-5597.py == --- head/tests/sys/netpfil/pf/CVE-2019-5597.py Tue Oct 29 04:28:24 2019 (r354144) +++ head/tests/sys/netpfil/pf/CVE-2019-5597.py Tue Oct 29 09:47:12 2019 (r354145) @@ -1,4 +1,29 @@ #!/usr/bin/env python +# +# SPDX-License-Identifier: BSD-2-Clause-FreeBSD +# +# Copyright (c) 2019 Synacktiv +# +# Redistribution and use in source and binary forms
svn commit: r354146 - head/contrib/llvm/lib/TableGen
Author: dim
Date: Tue Oct 29 16:51:12 2019
New Revision: 354146
URL: https://svnweb.freebsd.org/changeset/base/354146
Log:
Pull in r373338 from upstream llvm trunk (by Simon Pilgrim):
Revert rL349624 : Let TableGen write output only if it changed,
instead of doing so in cmake, attempt 2
Differential Revision: https://reviews.llvm.org/D55842
-
As discussed on PR43385 this is causing Visual Studio msbuilds to
perpetually rebuild all tablegen generated files
Pull in r373664 from upstream llvm trunk (by Nico Weber):
Reland r349624: Let TableGen write output only if it changed, instead
of doing so in cmake
Move the write-if-changed logic behind a flag and don't pass it with
the MSVC generator. msbuild doesn't have a restat optimization, so
not doing write-if-change there doesn't have a cost, and it should
fix whatever causes PR43385.
This should fix the scenario where an incremental build from before
r353358 (the clang 9.0.0 upgrade) to r353358 or later fails to update
the timestamp of the generated lib/clang/headers/arm_fp16.h header.
After such a build, installing world from read-only source and object
directories would attempt to generate the header again, leading to
"clang-tblgen: error opening arm_fp16.h.d:Read-only file system".
Reported by: avg, np
PR: 241402
MFC after:1 month
X-MFC-With: r353358
Modified:
head/contrib/llvm/lib/TableGen/Main.cpp
Modified: head/contrib/llvm/lib/TableGen/Main.cpp
==
--- head/contrib/llvm/lib/TableGen/Main.cpp Tue Oct 29 09:47:12 2019
(r354145)
+++ head/contrib/llvm/lib/TableGen/Main.cpp Tue Oct 29 16:51:12 2019
(r354146)
@@ -49,6 +49,9 @@ static cl::list
MacroNames("D", cl::desc("Name of the macro to be defined"),
cl::value_desc("macro name"), cl::Prefix);
+static cl::opt
+WriteIfChanged("write-if-changed", cl::desc("Only write output if it
changed"));
+
static int reportError(const char *ProgName, Twine Msg) {
errs() << ProgName << ": " << Msg;
errs().flush();
@@ -114,12 +117,14 @@ int llvm::TableGenMain(char *argv0, TableGenMainFn *Ma
return Ret;
}
- // Only updates the real output file if there are any differences.
- // This prevents recompilation of all the files depending on it if there
- // aren't any.
- if (auto ExistingOrErr = MemoryBuffer::getFile(OutputFilename))
-if (std::move(ExistingOrErr.get())->getBuffer() == Out.str())
- return 0;
+ if (WriteIfChanged) {
+// Only updates the real output file if there are any differences.
+// This prevents recompilation of all the files depending on it if there
+// aren't any.
+if (auto ExistingOrErr = MemoryBuffer::getFile(OutputFilename))
+ if (std::move(ExistingOrErr.get())->getBuffer() == Out.str())
+return 0;
+ }
std::error_code EC;
ToolOutputFile OutFile(OutputFilename, EC, sys::fs::F_Text);
___
svn-src-all@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r354147 - head/sys/sys
Author: glebius
Date: Tue Oct 29 17:19:36 2019
New Revision: 354147
URL: https://svnweb.freebsd.org/changeset/base/354147
Log:
Augment macros that manipulate td_no_sleeping with assertions to check
underleak and overflow of the counter.
Reviewed by: kib
Modified:
head/sys/sys/proc.h
Modified: head/sys/sys/proc.h
==
--- head/sys/sys/proc.h Tue Oct 29 16:51:12 2019(r354146)
+++ head/sys/sys/proc.h Tue Oct 29 17:19:36 2019(r354147)
@@ -246,8 +246,7 @@ struct thread {
u_char td_lend_user_pri; /* (t) Lend user pri. */
/* Cleared during fork1() */
-#definetd_startzero td_epochnest
- u_char td_epochnest; /* (k) Epoch nest counter. */
+#definetd_startzero td_flags
int td_flags; /* (t) TDF_* flags. */
int td_inhibitors; /* (t) Why can not run. */
int td_pflags; /* (k) Private thread (TDP_*) flags. */
@@ -299,7 +298,7 @@ struct thread {
struct vm_map_entry *td_map_def_user; /* (k) Deferred entries. */
pid_t td_dbg_forked; /* (c) Child pid for debugger. */
u_int td_vp_reserv; /* (k) Count of reserved vnodes. */
- int td_no_sleeping; /* (k) Sleeping disabled count. */
+ u_int td_no_sleeping; /* (k) Sleeping disabled count. */
void*td_su; /* (k) FFS SU private */
sbintime_t td_sleeptimo; /* (t) Sleep timeout. */
int td_rtcgen; /* (s) rtc_generation of abs. sleep */
@@ -948,9 +947,15 @@ extern pid_t pid_max;
#definethread_safetoswapout(td)((td)->td_flags & TDF_CANSWAP)
/* Control whether or not it is safe for curthread to sleep. */
-#defineTHREAD_NO_SLEEPING()((curthread)->td_no_sleeping++)
+#defineTHREAD_NO_SLEEPING()do {
\
+ curthread->td_no_sleeping++;\
+ MPASS(curthread->td_no_sleeping > 0); \
+} while (0)
-#defineTHREAD_SLEEPING_OK()((curthread)->td_no_sleeping--)
+#defineTHREAD_SLEEPING_OK()do {
\
+ MPASS(curthread->td_no_sleeping > 0); \
+ curthread->td_no_sleeping--;\
+} while (0)
#defineTHREAD_CAN_SLEEP() ((curthread)->td_no_sleeping ==
0)
___
svn-src-all@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r354148 - head/sys/kern
Author: glebius
Date: Tue Oct 29 17:28:25 2019
New Revision: 354148
URL: https://svnweb.freebsd.org/changeset/base/354148
Log:
Merge td_epochnest with td_no_sleeping.
Epoch itself doesn't rely on the counter and it is provided
merely for sleeping subsystems to check it.
- In functions that sleep use THREAD_CAN_SLEEP() to assert
correctness. With EPOCH_TRACE compiled print epoch info.
- _sleep() was a wrong place to put the assertion for epoch,
right place is sleepq_add(), as there ways to call the
latter bypassing _sleep().
- Do not increase td_no_sleeping in non-preemptible epochs.
The critical section would trigger all possible safeguards,
no sleeping counter is extraneous.
Reviewed by: kib
Modified:
head/sys/kern/genoffset.c
head/sys/kern/kern_malloc.c
head/sys/kern/kern_synch.c
head/sys/kern/subr_epoch.c
head/sys/kern/subr_sleepqueue.c
head/sys/kern/subr_trap.c
Modified: head/sys/kern/genoffset.c
==
--- head/sys/kern/genoffset.c Tue Oct 29 17:19:36 2019(r354147)
+++ head/sys/kern/genoffset.c Tue Oct 29 17:28:25 2019(r354148)
@@ -37,7 +37,6 @@ __FBSDID("$FreeBSD$");
#include
OFFSYM(td_priority, thread, u_char);
-OFFSYM(td_epochnest, thread, u_char);
OFFSYM(td_critnest, thread, u_int);
OFFSYM(td_pinned, thread, int);
OFFSYM(td_owepreempt, thread, u_char);
Modified: head/sys/kern/kern_malloc.c
==
--- head/sys/kern/kern_malloc.c Tue Oct 29 17:19:36 2019(r354147)
+++ head/sys/kern/kern_malloc.c Tue Oct 29 17:28:25 2019(r354148)
@@ -523,12 +523,13 @@ malloc_dbg(caddr_t *vap, size_t *sizep, struct malloc_
if (flags & M_WAITOK) {
KASSERT(curthread->td_intr_nesting_level == 0,
("malloc(M_WAITOK) in interrupt context"));
+ if (__predict_false(!THREAD_CAN_SLEEP())) {
#ifdef EPOCH_TRACE
- if (__predict_false(curthread->td_epochnest > 0))
epoch_trace_list(curthread);
#endif
- KASSERT(curthread->td_epochnest == 0,
- ("malloc(M_WAITOK) in epoch context"));
+ KASSERT(1,
+ ("malloc(M_WAITOK) with sleeping prohibited"));
+ }
}
KASSERT(curthread->td_critnest == 0 || SCHEDULER_STOPPED(),
("malloc: called with spinlock or critical section held"));
Modified: head/sys/kern/kern_synch.c
==
--- head/sys/kern/kern_synch.c Tue Oct 29 17:19:36 2019(r354147)
+++ head/sys/kern/kern_synch.c Tue Oct 29 17:28:25 2019(r354148)
@@ -151,11 +151,6 @@ _sleep(void *ident, struct lock_object *lock, int prio
("sleeping without a lock"));
KASSERT(ident != NULL, ("_sleep: NULL ident"));
KASSERT(TD_IS_RUNNING(td), ("_sleep: curthread not running"));
-#ifdef EPOCH_TRACE
- if (__predict_false(curthread->td_epochnest > 0))
- epoch_trace_list(curthread);
-#endif
- KASSERT(td->td_epochnest == 0, ("sleeping in an epoch section"));
if (priority & PDROP)
KASSERT(lock != NULL && lock != &Giant.lock_object,
("PDROP requires a non-Giant lock"));
Modified: head/sys/kern/subr_epoch.c
==
--- head/sys/kern/subr_epoch.c Tue Oct 29 17:19:36 2019(r354147)
+++ head/sys/kern/subr_epoch.c Tue Oct 29 17:28:25 2019(r354148)
@@ -377,7 +377,7 @@ _epoch_enter_preempt(epoch_t epoch, epoch_tracker_t et
epoch_trace_enter(td, epoch, et, file, line);
#endif
et->et_td = td;
- td->td_epochnest++;
+ THREAD_NO_SLEEPING();
critical_enter();
sched_pin();
td->td_pre_epoch_prio = td->td_priority;
@@ -390,13 +390,10 @@ _epoch_enter_preempt(epoch_t epoch, epoch_tracker_t et
void
epoch_enter(epoch_t epoch)
{
- struct thread *td;
epoch_record_t er;
MPASS(cold || epoch != NULL);
INIT_CHECK(epoch);
- td = curthread;
- td->td_epochnest++;
critical_enter();
er = epoch_currecord(epoch);
ck_epoch_begin(&er->er_record, NULL);
@@ -412,8 +409,7 @@ _epoch_exit_preempt(epoch_t epoch, epoch_tracker_t et
td = curthread;
critical_enter();
sched_unpin();
- MPASS(td->td_epochnest);
- td->td_epochnest--;
+ THREAD_SLEEPING_OK();
er = epoch_currecord(epoch);
MPASS(epoch->e_flags & EPOCH_PREEMPT);
MPASS(et != NULL);
@@ -435,13 +431,9 @@ _epoch_exit_preempt(epoch_t epoch, epoch_tracker_t et
void
epoch_exit(epoch_t epoch)
{
- struct thread *td;
epoch_record_t er;
INIT_CHECK(epoch);
- td = curthread;
- MPASS(td->td_epoch
svn commit: r354149 - head/sys/net
Author: glebius
Date: Tue Oct 29 17:36:06 2019
New Revision: 354149
URL: https://svnweb.freebsd.org/changeset/base/354149
Log:
There is a long standing problem with multicast programming for NICs
and IPv6. With IPv6 we may call if_addmulti() in context of processing
of an incoming packet. Usually this is interrupt context. While most
of the NIC drivers are able to reprogram multicast filters without
sleeping, some of them can't. An example is e1000 family of drivers.
With iflib conversion the problem was somewhat hidden. Iflib processes
packets in private taskqueue, so going to sleep doesn't trigger an
assertion. However, the sleep would block operation of the driver and
following incoming packets would fill the ring and eventually would
start being dropped. Enabling epoch for the full time of a packet
processing again started to trigger assertions for e1000.
Fix this problem once and for all using a general taskqueue to call
if_ioctl() method in all cases when if_addmulti() is called in a
non sleeping context. Note that nobody cares about returned value.
Reviewed by: hselasky, kib
Differential Revision: https://reviews.freebsd.org/D22154
Modified:
head/sys/net/if.c
head/sys/net/if_var.h
Modified: head/sys/net/if.c
==
--- head/sys/net/if.c Tue Oct 29 17:28:25 2019(r354148)
+++ head/sys/net/if.c Tue Oct 29 17:36:06 2019(r354149)
@@ -271,6 +271,7 @@ static int if_getgroupmembers(struct ifgroupreq *);
static voidif_delgroups(struct ifnet *);
static voidif_attach_internal(struct ifnet *, int, struct if_clone *);
static int if_detach_internal(struct ifnet *, int, struct if_clone **);
+static voidif_siocaddmulti(void *, int);
#ifdef VIMAGE
static voidif_vmove(struct ifnet *, struct vnet *);
#endif
@@ -556,6 +557,7 @@ if_alloc_domain(u_char type, int numa_domain)
IF_ADDR_LOCK_INIT(ifp);
TASK_INIT(&ifp->if_linktask, 0, do_link_state_change, ifp);
+ TASK_INIT(&ifp->if_addmultitask, 0, if_siocaddmulti, ifp);
ifp->if_afdata_initialized = 0;
IF_AFDATA_LOCK_INIT(ifp);
CK_STAILQ_INIT(&ifp->if_addrhead);
@@ -1131,6 +1133,7 @@ if_detach_internal(struct ifnet *ifp, int vmove, struc
if_delgroups(ifp);
taskqueue_drain(taskqueue_swi, &ifp->if_linktask);
+ taskqueue_drain(taskqueue_swi, &ifp->if_addmultitask);
/*
* Check if this is a cloned interface or not. Must do even if
@@ -3538,7 +3541,10 @@ if_addmulti(struct ifnet *ifp, struct sockaddr *sa,
* interface to let them know about it.
*/
if (ifp->if_ioctl != NULL) {
- (void) (*ifp->if_ioctl)(ifp, SIOCADDMULTI, 0);
+ if (THREAD_CAN_SLEEP())
+ (void )(*ifp->if_ioctl)(ifp, SIOCADDMULTI, 0);
+ else
+ taskqueue_enqueue(taskqueue_swi, &ifp->if_addmultitask);
}
if ((llsa != NULL) && (llsa != (struct sockaddr *)&sdl))
@@ -3553,6 +3559,19 @@ free_llsa_out:
unlock_out:
IF_ADDR_WUNLOCK(ifp);
return (error);
+}
+
+static void
+if_siocaddmulti(void *arg, int pending)
+{
+ struct ifnet *ifp;
+
+ ifp = arg;
+#ifdef DIAGNOSTIC
+ if (pending > 1)
+ if_printf(ifp, "%d SIOCADDMULTI coalesced\n", pending);
+#endif
+ (void )(*ifp->if_ioctl)(ifp, SIOCADDMULTI, 0);
}
/*
Modified: head/sys/net/if_var.h
==
--- head/sys/net/if_var.h Tue Oct 29 17:28:25 2019(r354148)
+++ head/sys/net/if_var.h Tue Oct 29 17:36:06 2019(r354149)
@@ -317,6 +317,7 @@ struct ifnet {
struct ifaltq if_snd; /* output queue (includes altq) */
struct task if_linktask; /* task for link change events */
+ struct task if_addmultitask; /* task for SIOCADDMULTI */
/* Addresses of different protocol families assigned to this if. */
struct mtx if_addr_lock;/* lock to protect address lists */
___
svn-src-all@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r354150 - stable/12
Author: jkim Date: Tue Oct 29 18:06:00 2019 New Revision: 354150 URL: https://svnweb.freebsd.org/changeset/base/354150 Log: MFC: r353907 Belatedly remove stale debug symbols. Modified: stable/12/ObsoleteFiles.inc Directory Properties: stable/12/ (props changed) Modified: stable/12/ObsoleteFiles.inc == --- stable/12/ObsoleteFiles.inc Tue Oct 29 17:36:06 2019(r354149) +++ stable/12/ObsoleteFiles.inc Tue Oct 29 18:06:00 2019(r354150) @@ -516,6 +516,16 @@ OLD_FILES+=usr/include/openssl/kssl.h OLD_FILES+=usr/include/openssl/pqueue.h OLD_FILES+=usr/include/openssl/ssl23.h OLD_FILES+=usr/include/openssl/ui_compat.h +OLD_FILES+=usr/lib/debug/usr/lib/engines/lib4758cca.so.debug +OLD_FILES+=usr/lib/debug/usr/lib/engines/libaep.so.debug +OLD_FILES+=usr/lib/debug/usr/lib/engines/libatalla.so.debug +OLD_FILES+=usr/lib/debug/usr/lib/engines/libcapi.so.debug +OLD_FILES+=usr/lib/debug/usr/lib/engines/libchil.so.debug +OLD_FILES+=usr/lib/debug/usr/lib/engines/libcswift.so.debug +OLD_FILES+=usr/lib/debug/usr/lib/engines/libgost.so.debug +OLD_FILES+=usr/lib/debug/usr/lib/engines/libnuron.so.debug +OLD_FILES+=usr/lib/debug/usr/lib/engines/libsureware.so.debug +OLD_FILES+=usr/lib/debug/usr/lib/engines/libubsec.so.debug OLD_FILES+=usr/share/openssl/man/man1/dss1.1.gz OLD_FILES+=usr/share/openssl/man/man1/md2.1.gz OLD_FILES+=usr/share/openssl/man/man1/md4.1.gz ___ svn-src-all@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-all To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r354151 - in head: contrib/netbsd-tests/lib/libexecinfo lib/libexecinfo/tests
Author: cem
Date: Tue Oct 29 18:24:36 2019
New Revision: 354151
URL: https://svnweb.freebsd.org/changeset/base/354151
Log:
libexecinfo test: Don't strip installed test
It turns out that a test of backtrace symbol resolution and formatting
requires symbols. Another option mightt be building with -rdynamic instead,
but this works for now.
Re-enabled skipped CI test, as it should now pass.
PR: 241562
Submitted by: lwhsu
Reported by: lwhsu
X-MFC-With: r354126, r354135, r354144
Modified:
head/contrib/netbsd-tests/lib/libexecinfo/t_backtrace.c
head/lib/libexecinfo/tests/Makefile
Modified: head/contrib/netbsd-tests/lib/libexecinfo/t_backtrace.c
==
--- head/contrib/netbsd-tests/lib/libexecinfo/t_backtrace.c Tue Oct 29
18:06:00 2019(r354150)
+++ head/contrib/netbsd-tests/lib/libexecinfo/t_backtrace.c Tue Oct 29
18:24:36 2019(r354151)
@@ -151,9 +151,6 @@ ATF_TC_HEAD(backtrace_fmt_basic, tc)
ATF_TC_BODY(backtrace_fmt_basic, tc)
{
- if (atf_tc_get_config_var_as_bool_wd(tc, "ci", false))
- atf_tc_skip("https://bugs.freebsd.org/241562";);
-
myfunc(12);
if (prevent_inline)
Modified: head/lib/libexecinfo/tests/Makefile
==
--- head/lib/libexecinfo/tests/Makefile Tue Oct 29 18:06:00 2019
(r354150)
+++ head/lib/libexecinfo/tests/Makefile Tue Oct 29 18:24:36 2019
(r354151)
@@ -6,6 +6,12 @@ TESTSRC= ${SRCTOP}/contrib/netbsd-tests/lib/libexecinf
.include
+# Perhaps not surprisingly, the backtrace symbol test needs symbols. Leave
+# them in:
+STRIP=
+# It could probably be made to work without symbols by checking pointer values
+# instead, but part of the tested functionality is the symbol resolution logic,
+# so that doesn't seem useful.
NETBSD_ATF_TESTS_C+= backtrace_test
LIBADD.backtrace_test+=execinfo
___
svn-src-all@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r354152 - head/sys/arm64/rockchip
Author: manu
Date: Tue Oct 29 18:36:16 2019
New Revision: 354152
URL: https://svnweb.freebsd.org/changeset/base/354152
Log:
arm64: rockchip: typec_phy: Rename timeout to retry
Declare retry in the function scope.
Rename it to retry as there is a timeout function which was
causing to code to compile.
Reported by: jhibbits
MFC after:1 month
X-MFC-WITH: r354089
Modified:
head/sys/arm64/rockchip/rk_typec_phy.c
Modified: head/sys/arm64/rockchip/rk_typec_phy.c
==
--- head/sys/arm64/rockchip/rk_typec_phy.c Tue Oct 29 18:24:36 2019
(r354151)
+++ head/sys/arm64/rockchip/rk_typec_phy.c Tue Oct 29 18:36:16 2019
(r354152)
@@ -191,7 +191,7 @@ rk_typec_phy_enable(struct phynode *phynode, bool enab
device_t dev;
intptr_t phy;
uint32_t reg;
- int err;
+ int err, retry;
dev = phynode_get_device(phynode);
phy = phynode_get_id(phynode);
@@ -271,13 +271,13 @@ rk_typec_phy_enable(struct phynode *phynode, bool enab
hwreset_deassert(sc->rst_uphy);
- for (int timeout = 1; timeout > 0; timeout--) {
+ for (retry = 1; retry > 0; retry--) {
reg = RK_TYPEC_PHY_READ(sc, PMA_CMN_CTRL1);
if (reg & PMA_CMN_CTRL1_READY)
break;
DELAY(10);
}
- if (timeout == 0) {
+ if (retry == 0) {
device_printf(sc->dev, "Timeout waiting for PMA\n");
return (ENXIO);
}
___
svn-src-all@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r354153 - in stable: 11/contrib/ipfilter/man 12/contrib/ipfilter/man
Author: cy Date: Tue Oct 29 19:36:21 2019 New Revision: 354153 URL: https://svnweb.freebsd.org/changeset/base/354153 Log: MFC r352784: Sync with source: Only a role of "ipf" is currently supported as the other documented (and undocumented) roles are #ifdef'd out. The plan is to complete ippool(8) as it is even in its current state a powerful feature/tool. PR: 218433 Modified: stable/11/contrib/ipfilter/man/ippool.8 Directory Properties: stable/11/ (props changed) Changes in other areas also in this revision: Modified: stable/12/contrib/ipfilter/man/ippool.8 Directory Properties: stable/12/ (props changed) Modified: stable/11/contrib/ipfilter/man/ippool.8 == --- stable/11/contrib/ipfilter/man/ippool.8 Tue Oct 29 18:36:16 2019 (r354152) +++ stable/11/contrib/ipfilter/man/ippool.8 Tue Oct 29 19:36:21 2019 (r354153) @@ -96,11 +96,8 @@ retrieving statistical information. .TP .B -o Sets the role with which this pool is to be used. Currently only -.B ipf, -.B auth -and -.B count -are accepted as arguments to this option. +.B ipf +(the default) is accepted as arguments to this option. .TP .B -S Sets the hashing seed to the number specified. Only for use with ___ svn-src-all@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-all To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r354153 - in stable: 11/contrib/ipfilter/man 12/contrib/ipfilter/man
Author: cy Date: Tue Oct 29 19:36:21 2019 New Revision: 354153 URL: https://svnweb.freebsd.org/changeset/base/354153 Log: MFC r352784: Sync with source: Only a role of "ipf" is currently supported as the other documented (and undocumented) roles are #ifdef'd out. The plan is to complete ippool(8) as it is even in its current state a powerful feature/tool. PR: 218433 Modified: stable/12/contrib/ipfilter/man/ippool.8 Directory Properties: stable/12/ (props changed) Changes in other areas also in this revision: Modified: stable/11/contrib/ipfilter/man/ippool.8 Directory Properties: stable/11/ (props changed) Modified: stable/12/contrib/ipfilter/man/ippool.8 == --- stable/12/contrib/ipfilter/man/ippool.8 Tue Oct 29 18:36:16 2019 (r354152) +++ stable/12/contrib/ipfilter/man/ippool.8 Tue Oct 29 19:36:21 2019 (r354153) @@ -96,11 +96,8 @@ retrieving statistical information. .TP .B -o Sets the role with which this pool is to be used. Currently only -.B ipf, -.B auth -and -.B count -are accepted as arguments to this option. +.B ipf +(the default) is accepted as arguments to this option. .TP .B -S Sets the hashing seed to the number specified. Only for use with ___ svn-src-all@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-all To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r354154 - head/usr.bin/vmstat
Author: trasz
Date: Tue Oct 29 20:28:02 2019
New Revision: 354154
URL: https://svnweb.freebsd.org/changeset/base/354154
Log:
Fix column title alignment.
MFC after:2 weeks
Modified:
head/usr.bin/vmstat/vmstat.c
Modified: head/usr.bin/vmstat/vmstat.c
==
--- head/usr.bin/vmstat/vmstat.cTue Oct 29 19:36:21 2019
(r354153)
+++ head/usr.bin/vmstat/vmstat.cTue Oct 29 20:28:02 2019
(r354154)
@@ -867,11 +867,11 @@ printhdr(int maxid, u_long cpumask)
num_shown = MIN(num_selected, maxshowdevs);
if (hflag)
- xo_emit("{T:procs} {T:memory} {T:/page%*s}", 19, "");
+ xo_emit("{T:procs} {T:memory} {T:/page%*s}", 19, "");
else
- xo_emit("{T:procs} {T:memory}{T:/page%*s}", 19, "");
+ xo_emit("{T:procs} {T:memory} {T:/page%*s}", 19, "");
if (num_shown > 1)
- xo_emit(" {T:/disks %*s}", num_shown * 4 - 7, "");
+ xo_emit("{T:/disks %*s}", num_shown * 4 - 7, "");
else if (num_shown == 1)
xo_emit(" {T:disks}");
xo_emit(" {T:faults} ");
___
svn-src-all@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r354155 - head/sys/kern
Author: jeff
Date: Tue Oct 29 20:37:59 2019
New Revision: 354155
URL: https://svnweb.freebsd.org/changeset/base/354155
Log:
Drop the object lock in vfs_bio and cluster where it is now safe to do so.
Recent changes to busy/valid/dirty have enabled page based synchronization
and the object lock is no longer required in many cases.
Reviewed by: kib
Sponsored by: Netflix, Intel
Differential Revision:https://reviews.freebsd.org/D21597
Modified:
head/sys/kern/vfs_bio.c
head/sys/kern/vfs_cluster.c
Modified: head/sys/kern/vfs_bio.c
==
--- head/sys/kern/vfs_bio.c Tue Oct 29 20:28:02 2019(r354154)
+++ head/sys/kern/vfs_bio.c Tue Oct 29 20:37:59 2019(r354155)
@@ -162,7 +162,7 @@ static void vfs_page_set_valid(struct buf *bp, vm_ooff
static void vfs_page_set_validclean(struct buf *bp, vm_ooffset_t off,
vm_page_t m);
static void vfs_clean_pages_dirty_buf(struct buf *bp);
-static void vfs_setdirty_locked_object(struct buf *bp);
+static void vfs_setdirty_range(struct buf *bp);
static void vfs_vmio_invalidate(struct buf *bp);
static void vfs_vmio_truncate(struct buf *bp, int npages);
static void vfs_vmio_extend(struct buf *bp, int npages, int size);
@@ -955,8 +955,6 @@ vfs_buf_test_cache(struct buf *bp, vm_ooffset_t foff,
vm_offset_t size, vm_page_t m)
{
- VM_OBJECT_ASSERT_LOCKED(m->object);
-
/*
* This function and its results are protected by higher level
* synchronization requiring vnode and buf locks to page in and
@@ -2865,7 +2863,6 @@ vfs_vmio_iodone(struct buf *bp)
bogus = false;
iosize = bp->b_bcount - bp->b_resid;
- VM_OBJECT_WLOCK(obj);
for (i = 0; i < bp->b_npages; i++) {
resid = ((foff + PAGE_SIZE) & ~(off_t)PAGE_MASK) - foff;
if (resid > iosize)
@@ -2876,7 +2873,10 @@ vfs_vmio_iodone(struct buf *bp)
*/
m = bp->b_pages[i];
if (m == bogus_page) {
- bogus = true;
+ if (bogus == false) {
+ bogus = true;
+ VM_OBJECT_RLOCK(obj);
+ }
m = vm_page_lookup(obj, OFF_TO_IDX(foff));
if (m == NULL)
panic("biodone: page disappeared!");
@@ -2900,8 +2900,9 @@ vfs_vmio_iodone(struct buf *bp)
foff = (foff + PAGE_SIZE) & ~(off_t)PAGE_MASK;
iosize -= resid;
}
+ if (bogus)
+ VM_OBJECT_RUNLOCK(obj);
vm_object_pip_wakeupn(obj, bp->b_npages);
- VM_OBJECT_WUNLOCK(obj);
if (bogus && buf_mapped(bp)) {
BUF_CHECK_MAPPED(bp);
pmap_qenter(trunc_page((vm_offset_t)bp->b_data),
@@ -3029,7 +3030,6 @@ vfs_vmio_extend(struct buf *bp, int desiredpages, int
* are not valid for the range covered by the buffer.
*/
obj = bp->b_bufobj->bo_object;
- VM_OBJECT_WLOCK(obj);
if (bp->b_npages < desiredpages) {
/*
* We must allocate system pages since blocking
@@ -3041,11 +3041,13 @@ vfs_vmio_extend(struct buf *bp, int desiredpages, int
* deadlocks once allocbuf() is called after
* pages are vfs_busy_pages().
*/
+ VM_OBJECT_WLOCK(obj);
(void)vm_page_grab_pages(obj,
OFF_TO_IDX(bp->b_offset) + bp->b_npages,
VM_ALLOC_SYSTEM | VM_ALLOC_IGN_SBUSY |
VM_ALLOC_NOBUSY | VM_ALLOC_WIRED,
&bp->b_pages[bp->b_npages], desiredpages - bp->b_npages);
+ VM_OBJECT_WUNLOCK(obj);
bp->b_npages = desiredpages;
}
@@ -3076,7 +3078,6 @@ vfs_vmio_extend(struct buf *bp, int desiredpages, int
toff += tinc;
tinc = PAGE_SIZE;
}
- VM_OBJECT_WUNLOCK(obj);
/*
* Step 3, fixup the KVA pmap.
@@ -3656,9 +3657,8 @@ vfs_clean_pages_dirty_buf(struct buf *bp)
KASSERT(bp->b_offset != NOOFFSET,
("vfs_clean_pages_dirty_buf: no buffer offset"));
- VM_OBJECT_WLOCK(bp->b_bufobj->bo_object);
vfs_busy_pages_acquire(bp);
- vfs_setdirty_locked_object(bp);
+ vfs_setdirty_range(bp);
for (i = 0; i < bp->b_npages; i++) {
noff = (foff + PAGE_SIZE) & ~(off_t)PAGE_MASK;
eoff = noff;
@@ -3670,69 +3670,57 @@ vfs_clean_pages_dirty_buf(struct buf *bp)
foff = noff;
}
vfs_busy_pages_release(bp);
- VM_OBJECT_WUNLOCK(bp->b_bufobj->bo_object);
}
static void
-vfs_setdirty_locked_object(struct buf *bp)
+vfs_setdirty_range(struct buf *bp)
{
- vm_object_t object;
+ vm_offset_t boffset;
+ vm_offset_t eoffset;
int i;
-
svn commit: r354156 - head/sys/vm
Author: jeff
Date: Tue Oct 29 20:46:25 2019
New Revision: 354156
URL: https://svnweb.freebsd.org/changeset/base/354156
Log:
Drop the object lock earlier in fault and don't relock it after pmap_enter().
Recent changes in object and page locking have enabled more lock pushdown.
Reviewed by: kib
Tested by:pho
Differential Revision:https://reviews.freebsd.org/D22036
Modified:
head/sys/vm/vm_fault.c
Modified: head/sys/vm/vm_fault.c
==
--- head/sys/vm/vm_fault.c Tue Oct 29 20:37:59 2019(r354155)
+++ head/sys/vm/vm_fault.c Tue Oct 29 20:46:25 2019(r354156)
@@ -184,11 +184,10 @@ unlock_vp(struct faultstate *fs)
}
static void
-unlock_and_deallocate(struct faultstate *fs)
+fault_deallocate(struct faultstate *fs)
{
vm_object_pip_wakeup(fs->object);
- VM_OBJECT_WUNLOCK(fs->object);
if (fs->object != fs->first_object) {
VM_OBJECT_WLOCK(fs->first_object);
vm_page_free(fs->first_m);
@@ -202,6 +201,14 @@ unlock_and_deallocate(struct faultstate *fs)
}
static void
+unlock_and_deallocate(struct faultstate *fs)
+{
+
+ VM_OBJECT_WUNLOCK(fs->object);
+ fault_deallocate(fs);
+}
+
+static void
vm_fault_dirty(vm_map_entry_t entry, vm_page_t m, vm_prot_t prot,
vm_prot_t fault_type, int fault_flags, bool set_wd)
{
@@ -1261,10 +1268,12 @@ readrest:
fs.object, OFF_TO_IDX(
fs.first_object->backing_object_offset));
#endif
+ VM_OBJECT_WUNLOCK(fs.object);
fs.first_m = fs.m;
fs.m = NULL;
VM_CNT_INC(v_cow_optim);
} else {
+ VM_OBJECT_WUNLOCK(fs.object);
/*
* Oh, well, lets copy it.
*/
@@ -1285,7 +1294,6 @@ readrest:
* conditional
*/
vm_object_pip_wakeup(fs.object);
- VM_OBJECT_WUNLOCK(fs.object);
/*
* We only try to prefault read-only mappings to the
@@ -1405,7 +1413,6 @@ readrest:
vm_fault_prefault(&fs, vaddr,
faultcount > 0 ? behind : PFBAK,
faultcount > 0 ? ahead : PFFOR, false);
- VM_OBJECT_WLOCK(fs.object);
/*
* If the page is not wired down, then put it where the pageout daemon
@@ -1427,7 +1434,7 @@ readrest:
/*
* Unlock everything, and return
*/
- unlock_and_deallocate(&fs);
+ fault_deallocate(&fs);
if (hardfault) {
VM_CNT_INC(v_io_faults);
curthread->td_ru.ru_majflt++;
___
svn-src-all@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r354157 - in head/sys: sys vm
Author: jeff
Date: Tue Oct 29 20:58:46 2019
New Revision: 354157
URL: https://svnweb.freebsd.org/changeset/base/354157
Log:
Use atomics and a shared object lock to protect the object reference count.
Certain consumers still need to guarantee a stable reference so we can not
switch entirely to atomics yet. Exclusive lock holders can still modify
and examine the refcount without using the ref api.
Reviewed by: kib
Tested by:pho
Sponsored by: Netflix, Intel
Differential Revision:https://reviews.freebsd.org/D21598
Modified:
head/sys/sys/refcount.h
head/sys/vm/vm_object.c
head/sys/vm/vm_object.h
head/sys/vm/vnode_pager.c
Modified: head/sys/sys/refcount.h
==
--- head/sys/sys/refcount.h Tue Oct 29 20:46:25 2019(r354156)
+++ head/sys/sys/refcount.h Tue Oct 29 20:58:46 2019(r354157)
@@ -175,4 +175,22 @@ refcount_release_if_not_last(volatile u_int *count)
}
}
+static __inline __result_use_check bool
+refcount_release_if_gt(volatile u_int *count, u_int n)
+{
+ u_int old;
+
+ KASSERT(n > 0,
+ ("refcount_release_if_gt: Use refcount_release for final ref"));
+ old = *count;
+ for (;;) {
+ if (REFCOUNT_COUNT(old) <= n)
+ return (false);
+ if (__predict_false(REFCOUNT_SATURATED(old)))
+ return (true);
+ if (atomic_fcmpset_int(count, &old, old - 1))
+ return (true);
+ }
+}
+
#endif /* ! __SYS_REFCOUNT_H__ */
Modified: head/sys/vm/vm_object.c
==
--- head/sys/vm/vm_object.c Tue Oct 29 20:46:25 2019(r354156)
+++ head/sys/vm/vm_object.c Tue Oct 29 20:58:46 2019(r354157)
@@ -224,8 +224,8 @@ vm_object_zinit(void *mem, int size, int flags)
/* These are true for any object that has been freed */
object->type = OBJT_DEAD;
- object->ref_count = 0;
vm_radix_init(&object->rtree);
+ refcount_init(&object->ref_count, 0);
refcount_init(&object->paging_in_progress, 0);
refcount_init(&object->busy, 0);
object->resident_page_count = 0;
@@ -282,7 +282,7 @@ _vm_object_allocate(objtype_t type, vm_pindex_t size,
object->size = size;
object->domain.dr_policy = NULL;
object->generation = 1;
- object->ref_count = 1;
+ refcount_init(&object->ref_count, 1);
object->memattr = VM_MEMATTR_DEFAULT;
object->cred = NULL;
object->charge = 0;
@@ -444,9 +444,9 @@ vm_object_reference(vm_object_t object)
{
if (object == NULL)
return;
- VM_OBJECT_WLOCK(object);
+ VM_OBJECT_RLOCK(object);
vm_object_reference_locked(object);
- VM_OBJECT_WUNLOCK(object);
+ VM_OBJECT_RUNLOCK(object);
}
/*
@@ -461,8 +461,8 @@ vm_object_reference_locked(vm_object_t object)
{
struct vnode *vp;
- VM_OBJECT_ASSERT_WLOCKED(object);
- object->ref_count++;
+ VM_OBJECT_ASSERT_LOCKED(object);
+ refcount_acquire(&object->ref_count);
if (object->type == OBJT_VNODE) {
vp = object->handle;
vref(vp);
@@ -477,24 +477,16 @@ vm_object_vndeallocate(vm_object_t object)
{
struct vnode *vp = (struct vnode *) object->handle;
- VM_OBJECT_ASSERT_WLOCKED(object);
KASSERT(object->type == OBJT_VNODE,
("vm_object_vndeallocate: not a vnode object"));
KASSERT(vp != NULL, ("vm_object_vndeallocate: missing vp"));
-#ifdef INVARIANTS
- if (object->ref_count == 0) {
- vn_printf(vp, "vm_object_vndeallocate ");
- panic("vm_object_vndeallocate: bad object reference count");
- }
-#endif
- if (!umtx_shm_vnobj_persistent && object->ref_count == 1)
+ if (refcount_release(&object->ref_count) &&
+ !umtx_shm_vnobj_persistent)
umtx_shm_object_terminated(object);
- object->ref_count--;
-
+ VM_OBJECT_RUNLOCK(object);
/* vrele may need the vnode lock. */
- VM_OBJECT_WUNLOCK(object);
vrele(vp);
}
@@ -513,24 +505,32 @@ void
vm_object_deallocate(vm_object_t object)
{
vm_object_t temp;
+ bool released;
while (object != NULL) {
- VM_OBJECT_WLOCK(object);
+ VM_OBJECT_RLOCK(object);
if (object->type == OBJT_VNODE) {
vm_object_vndeallocate(object);
return;
}
- KASSERT(object->ref_count != 0,
- ("vm_object_deallocate: object deallocated too many
times: %d", object->type));
-
/*
* If the reference count goes to 0 we start calling
-* vm_object_terminate() on the object chain.
-* A ref count of
svn commit: r354158 - in head/sys: fs/nfsclient fs/nfsserver fs/tmpfs kern ufs/ffs vm
Author: jeff
Date: Tue Oct 29 21:06:34 2019
New Revision: 354158
URL: https://svnweb.freebsd.org/changeset/base/354158
Log:
Replace OBJ_MIGHTBEDIRTY with a system using atomics. Remove the TMPFS_DIRTY
flag and use the same system.
This enables further fault locking improvements by allowing more faults to
proceed with a shared lock.
Reviewed by: kib
Tested by:pho
Differential Revision:https://reviews.freebsd.org/D22116
Modified:
head/sys/fs/nfsclient/nfs_clvnops.c
head/sys/fs/nfsserver/nfs_nfsdport.c
head/sys/fs/tmpfs/tmpfs_subr.c
head/sys/fs/tmpfs/tmpfs_vfsops.c
head/sys/fs/tmpfs/tmpfs_vnops.c
head/sys/kern/vfs_subr.c
head/sys/ufs/ffs/ffs_rawread.c
head/sys/vm/vm_fault.c
head/sys/vm/vm_object.c
head/sys/vm/vm_object.h
head/sys/vm/vm_page.c
Modified: head/sys/fs/nfsclient/nfs_clvnops.c
==
--- head/sys/fs/nfsclient/nfs_clvnops.c Tue Oct 29 20:58:46 2019
(r354157)
+++ head/sys/fs/nfsclient/nfs_clvnops.c Tue Oct 29 21:06:34 2019
(r354158)
@@ -715,7 +715,7 @@ nfs_open(struct vop_open_args *ap)
*/
if (vp->v_writecount <= -1) {
if ((obj = vp->v_object) != NULL &&
- (obj->flags & OBJ_MIGHTBEDIRTY) != 0) {
+ vm_object_mightbedirty(obj)) {
VM_OBJECT_WLOCK(obj);
vm_object_page_clean(obj, 0, 0, OBJPC_SYNC);
VM_OBJECT_WUNLOCK(obj);
Modified: head/sys/fs/nfsserver/nfs_nfsdport.c
==
--- head/sys/fs/nfsserver/nfs_nfsdport.cTue Oct 29 20:58:46 2019
(r354157)
+++ head/sys/fs/nfsserver/nfs_nfsdport.cTue Oct 29 21:06:34 2019
(r354158)
@@ -1498,8 +1498,7 @@ nfsvno_fsync(struct vnode *vp, u_int64_t off, int cnt,
/*
* Give up and do the whole thing
*/
- if (vp->v_object &&
- (vp->v_object->flags & OBJ_MIGHTBEDIRTY)) {
+ if (vp->v_object && vm_object_mightbedirty(vp->v_object)) {
VM_OBJECT_WLOCK(vp->v_object);
vm_object_page_clean(vp->v_object, 0, 0, OBJPC_SYNC);
VM_OBJECT_WUNLOCK(vp->v_object);
@@ -1529,8 +1528,7 @@ nfsvno_fsync(struct vnode *vp, u_int64_t off, int cnt,
}
lblkno = off / iosize;
- if (vp->v_object &&
- (vp->v_object->flags & OBJ_MIGHTBEDIRTY)) {
+ if (vp->v_object && vm_object_mightbedirty(vp->v_object)) {
VM_OBJECT_WLOCK(vp->v_object);
vm_object_page_clean(vp->v_object, off, off + cnt,
OBJPC_SYNC);
Modified: head/sys/fs/tmpfs/tmpfs_subr.c
==
--- head/sys/fs/tmpfs/tmpfs_subr.c Tue Oct 29 20:58:46 2019
(r354157)
+++ head/sys/fs/tmpfs/tmpfs_subr.c Tue Oct 29 21:06:34 2019
(r354158)
@@ -1477,10 +1477,10 @@ tmpfs_check_mtime(struct vnode *vp)
KASSERT((obj->flags & (OBJ_TMPFS_NODE | OBJ_TMPFS)) ==
(OBJ_TMPFS_NODE | OBJ_TMPFS), ("non-tmpfs obj"));
/* unlocked read */
- if ((obj->flags & OBJ_TMPFS_DIRTY) != 0) {
+ if (obj->generation != obj->cleangeneration) {
VM_OBJECT_WLOCK(obj);
- if ((obj->flags & OBJ_TMPFS_DIRTY) != 0) {
- obj->flags &= ~OBJ_TMPFS_DIRTY;
+ if (obj->generation != obj->cleangeneration) {
+ obj->cleangeneration = obj->generation;
node = VP_TO_TMPFS_NODE(vp);
node->tn_status |= TMPFS_NODE_MODIFIED |
TMPFS_NODE_CHANGED;
Modified: head/sys/fs/tmpfs/tmpfs_vfsops.c
==
--- head/sys/fs/tmpfs/tmpfs_vfsops.cTue Oct 29 20:58:46 2019
(r354157)
+++ head/sys/fs/tmpfs/tmpfs_vfsops.cTue Oct 29 21:06:34 2019
(r354158)
@@ -172,7 +172,7 @@ tmpfs_update_mtime(struct mount *mp, bool lazy)
* For non-lazy case, we must flush all pending
* metadata changes now.
*/
- if (!lazy || (obj->flags & OBJ_TMPFS_DIRTY) != 0) {
+ if (!lazy || obj->generation != obj->cleangeneration) {
if (vget(vp, LK_EXCLUSIVE | LK_INTERLOCK,
curthread) != 0)
continue;
Modified: head/sys/fs/tmpfs/tmpfs_vnops.c
==
--- head/sys/fs/tmpfs/tmpfs_vnops.c Tue Oct 29 20:58:46 2019
(r354157)
+++ head/sys/fs/tmpfs/tmpfs_vnops.c Tue Oct 29 21:06:34 2019
(r354158)
@@ -1323,7 +1323,7 @@ tmpfs_nee
svn commit: r354159 - head/sys/cddl/contrib/opensolaris/uts/common/zmod
Author: mav
Date: Tue Oct 29 21:25:19 2019
New Revision: 354159
URL: https://svnweb.freebsd.org/changeset/base/354159
Log:
FreeBSD'fy ZFS zlib zalloc/zfree callbacks.
The previous code came from OpenSolaris, which in my understanding require
allocation size to be known to free memory. To store that size previous
code allocated additional 8 byte header. But I have noticed that zlib
with present settings allocates 64KB context buffers for each call, that
could be efficiently cached by UMA, but addition of those 8 bytes makes
them fall back to physical RAM allocations, that cause huge overhead and
lock congestion on small blocks. Since FreeBSD's free() does not have
the size argument, switching to it solves the problem, increasing write
speed to ZVOLs with 4KB block size and GZIP compression on my 40-threads
test system from ~60MB/s to ~600MB/s.
MFC after:1 week
Sponsored by: iXsystems, Inc.
Modified:
head/sys/cddl/contrib/opensolaris/uts/common/zmod/zmod.c
Modified: head/sys/cddl/contrib/opensolaris/uts/common/zmod/zmod.c
==
--- head/sys/cddl/contrib/opensolaris/uts/common/zmod/zmod.cTue Oct 29
21:06:34 2019(r354158)
+++ head/sys/cddl/contrib/opensolaris/uts/common/zmod/zmod.cTue Oct 29
21:25:19 2019(r354159)
@@ -27,45 +27,28 @@
#include
#include
#include
-#include
+#include
#include
#include
#include
-struct zchdr {
- uint_t zch_magic;
- uint_t zch_size;
-};
-
-#defineZCH_MAGIC 0x3cc13cc1
-
/*ARGSUSED*/
static void *
zfs_zcalloc(void *opaque, uint_t items, uint_t size)
{
- size_t nbytes = sizeof (struct zchdr) + items * size;
- struct zchdr *z = kobj_zalloc(nbytes, KM_NOWAIT|KM_TMP);
+ void *ptr;
- if (z == NULL)
- return (NULL);
-
- z->zch_magic = ZCH_MAGIC;
- z->zch_size = nbytes;
-
- return (z + 1);
+ ptr = malloc((size_t)items * size, M_SOLARIS, M_NOWAIT);
+ return ptr;
}
/*ARGSUSED*/
static void
zfs_zcfree(void *opaque, void *ptr)
{
- struct zchdr *z = ((struct zchdr *)ptr) - 1;
- if (z->zch_magic != ZCH_MAGIC)
- panic("zcfree region corrupt: hdr=%p ptr=%p", (void *)z, ptr);
-
- kobj_free(z, z->zch_size);
+ free(ptr, M_SOLARIS);
}
/*
___
svn-src-all@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r354160 - stable/12/libexec/rc
Author: cy
Date: Wed Oct 30 01:24:28 2019
New Revision: 354160
URL: https://svnweb.freebsd.org/changeset/base/354160
Log:
MFC r354129:
Remove trailing tab.
Modified:
stable/12/libexec/rc/rc.subr
Directory Properties:
stable/12/ (props changed)
Modified: stable/12/libexec/rc/rc.subr
==
--- stable/12/libexec/rc/rc.subrTue Oct 29 21:25:19 2019
(r354159)
+++ stable/12/libexec/rc/rc.subrWed Oct 30 01:24:28 2019
(r354160)
@@ -1266,7 +1266,7 @@ $command $rc_flags $command_args"
${PROTECT} -p ${pid}
;;
esac
- fi
+ fi
;;
esac
___
svn-src-all@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r354161 - in stable/12: sys/fs/fuse tests/sys/fs/fusefs
Author: asomers
Date: Wed Oct 30 01:35:00 2019
New Revision: 354161
URL: https://svnweb.freebsd.org/changeset/base/354161
Log:
MFC r352404, r352413-r352414
r352404:
fusefs: fix some minor issues with fuse_vnode_setparent
* When unparenting a vnode, actually clear the flag. AFAIK this is basically
a no-op because we only unparent a vnode when reclaiming it or when
unlinking.
* There's no need to call fuse_vnode_setparent during reclaim, because we're
about to free the vnode data anyway.
Reviewed by: emaste
Sponsored by: The FreeBSD Foundation
Differential Revision:https://reviews.freebsd.org/D21630
r352413:
fusefs: fix some minor Coverity CIDs in the tests
Where open(2) is expected to fail, the tests should assert or expect that
its return value is -1. These tests all accepted too much but happened to
pass anyway.
Reported by: Coverity
Coverity CID: 1404512, 1404378, 1404504, 1404483
Sponsored by: The FreeBSD Foundation
r352414:
fusefs: initialize C++ classes the Coverity way
Coverity complained that I wasn't initializing some class members until the
SetUp method. Do it in the constructor instead.
Reported by: Coverity
Coverity CIDs:1404352, 1404378
Sponsored by: The FreeBSD Foundation
Modified:
stable/12/sys/fs/fuse/fuse_node.h
stable/12/sys/fs/fuse/fuse_vnops.c
stable/12/tests/sys/fs/fusefs/create.cc
stable/12/tests/sys/fs/fusefs/default_permissions.cc
stable/12/tests/sys/fs/fusefs/io.cc
stable/12/tests/sys/fs/fusefs/mknod.cc
stable/12/tests/sys/fs/fusefs/opendir.cc
Directory Properties:
stable/12/ (props changed)
Modified: stable/12/sys/fs/fuse/fuse_node.h
==
--- stable/12/sys/fs/fuse/fuse_node.h Wed Oct 30 01:24:28 2019
(r354160)
+++ stable/12/sys/fs/fuse/fuse_node.h Wed Oct 30 01:35:00 2019
(r354161)
@@ -174,6 +174,8 @@ fuse_vnode_setparent(struct vnode *vp, struct vnode *d
MPASS(dvp->v_type == VDIR);
VTOFUD(vp)->parent_nid = VTOI(dvp);
VTOFUD(vp)->flag |= FN_PARENT_NID;
+ } else {
+ VTOFUD(vp)->flag &= ~FN_PARENT_NID;
}
}
Modified: stable/12/sys/fs/fuse/fuse_vnops.c
==
--- stable/12/sys/fs/fuse/fuse_vnops.c Wed Oct 30 01:24:28 2019
(r354160)
+++ stable/12/sys/fs/fuse/fuse_vnops.c Wed Oct 30 01:35:00 2019
(r354161)
@@ -1526,11 +1526,10 @@ fuse_vnop_reclaim(struct vop_reclaim_args *ap)
fuse_filehandle_close(vp, fufh, td, NULL);
}
- if ((!fuse_isdeadfs(vp)) && (fvdat->nlookup)) {
+ if (!fuse_isdeadfs(vp) && fvdat->nlookup > 0) {
fuse_internal_forget_send(vnode_mount(vp), td, NULL, VTOI(vp),
fvdat->nlookup);
}
- fuse_vnode_setparent(vp, NULL);
cache_purge(vp);
vfs_hash_remove(vp);
vnode_destroy_vobject(vp);
Modified: stable/12/tests/sys/fs/fusefs/create.cc
==
--- stable/12/tests/sys/fs/fusefs/create.cc Wed Oct 30 01:24:28 2019
(r354160)
+++ stable/12/tests/sys/fs/fusefs/create.cc Wed Oct 30 01:35:00 2019
(r354161)
@@ -204,7 +204,7 @@ TEST_F(Create, eexist)
EXPECT_LOOKUP(FUSE_ROOT_ID, RELPATH)
.WillOnce(Invoke(ReturnErrno(ENOENT)));
expect_create(RELPATH, mode, ReturnErrno(EEXIST));
- EXPECT_NE(0, open(FULLPATH, O_CREAT | O_EXCL, mode));
+ EXPECT_EQ(-1, open(FULLPATH, O_CREAT | O_EXCL, mode));
EXPECT_EQ(EEXIST, errno);
}
@@ -342,7 +342,7 @@ TEST_F(Create, eperm)
.WillOnce(Invoke(ReturnErrno(ENOENT)));
expect_create(RELPATH, mode, ReturnErrno(EPERM));
- EXPECT_NE(0, open(FULLPATH, O_CREAT | O_EXCL, mode));
+ EXPECT_EQ(-1, open(FULLPATH, O_CREAT | O_EXCL, mode));
EXPECT_EQ(EPERM, errno);
}
Modified: stable/12/tests/sys/fs/fusefs/default_permissions.cc
==
--- stable/12/tests/sys/fs/fusefs/default_permissions.ccWed Oct 30
01:24:28 2019(r354160)
+++ stable/12/tests/sys/fs/fusefs/default_permissions.ccWed Oct 30
01:35:00 2019(r354161)
@@ -749,7 +749,7 @@ TEST_F(Open, eacces)
expect_getattr(FUSE_ROOT_ID, S_IFDIR | 0755, UINT64_MAX, 1);
expect_lookup(RELPATH, ino, S_IFREG | 0644, UINT64_MAX);
- EXPECT_NE(0, open(FULLPATH, O_RDWR));
+ EXPECT_EQ(-1, open(FULLPATH, O_RDWR));
EXPECT_EQ(EACCES, errno);
}
Modified: stable/12/tests/sys/fs/fusefs/io.cc
==
--- stable/12/tests/sys/fs/fusefs/io.cc Wed Oct 30 01:24:28 2019
(r354160)
+++ stable/12/tests/sys/fs/fusefs/io.cc Wed Oct 30 01:35:0
svn commit: r354162 - in head/sys/x86: include x86
Author: cem Date: Wed Oct 30 01:41:14 2019 New Revision: 354162 URL: https://svnweb.freebsd.org/changeset/base/354162 Log: amd64: Define and decode new AMD64 feature bits These are documented in revisions 3.32 of the public AMD64 Vol. 2 and revision 3.28 of Vol. 3, published October and September 2019, respectively. Modified: head/sys/x86/include/specialreg.h head/sys/x86/x86/identcpu.c Modified: head/sys/x86/include/specialreg.h == --- head/sys/x86/include/specialreg.h Wed Oct 30 01:35:00 2019 (r354161) +++ head/sys/x86/include/specialreg.h Wed Oct 30 01:41:14 2019 (r354162) @@ -71,6 +71,7 @@ #defineCR4_PCE 0x0100 /* Performance monitoring counter enable */ #defineCR4_FXSR 0x0200 /* Fast FPU save/restore used by OS */ #defineCR4_XMM 0x0400 /* enable SIMD/MMX2 to use except 16 */ +#defineCR4_UMIP 0x0800 /* User Mode Instruction Prevention */ #defineCR4_VMXE 0x2000 /* enable VMX operation (Intel-specific) */ #defineCR4_FSGSBASE 0x0001 /* Enable FS/GS BASE accessing instructions */ #defineCR4_PCIDE 0x0002/* Enable Context ID */ @@ -90,6 +91,7 @@ #defineEFER_LMSLE 0x02000 /* Long Mode Segment Limit Enable */ #defineEFER_FFXSR 0x04000 /* Fast FXSAVE/FSRSTOR */ #defineEFER_TCE 0x08000 /* Translation Cache Extension */ +#defineEFER_MCOMMIT0x0002 /* Enable MCOMMIT (AMD) */ /* * Intel Extended Features registers @@ -384,6 +386,9 @@ #defineAMDFEID_CLZERO 0x0001 #defineAMDFEID_IRPERF 0x0002 #defineAMDFEID_XSAVEERPTR 0x0004 +#defineAMDFEID_RDPRU 0x0004 +#defineAMDFEID_MCOMMIT 0x0100 +#defineAMDFEID_WBNOINVD0x0200 #defineAMDFEID_IBPB0x1000 #defineAMDFEID_IBRS0x4000 #defineAMDFEID_STIBP 0x8000 Modified: head/sys/x86/x86/identcpu.c == --- head/sys/x86/x86/identcpu.c Wed Oct 30 01:35:00 2019(r354161) +++ head/sys/x86/x86/identcpu.c Wed Oct 30 01:41:14 2019(r354162) @@ -1067,6 +1067,9 @@ printcpuinfo(void) "\001CLZERO" "\002IRPerf" "\003XSaveErPtr" + "\005RDPRU" + "\011MCOMMIT" + "\012WBNOINVD" "\015IBPB" "\017IBRS" "\020STIBP" @@ -2355,7 +2358,7 @@ print_svm_info(void) "\017" "\020V_VMSAVE_VMLOAD" "\021vGIF" - "\022" + "\022GMET" /* Guest Mode Execute Trap */ "\023" "\024" "\025" ___ svn-src-all@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-all To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r354163 - stable/12/usr.bin/tftp
Author: asomers
Date: Wed Oct 30 01:41:49 2019
New Revision: 354163
URL: https://svnweb.freebsd.org/changeset/base/354163
Log:
MFC r353068:
tftp: fix two minor Coverity CIDs
Reported by: Coverity
CID 1394842: file descriptor leak in an error path
CID 1007603: single byte array overflow
Differential Revision:https://reviews.freebsd.org/D21695
Modified:
stable/12/usr.bin/tftp/main.c
Directory Properties:
stable/12/ (props changed)
Modified: stable/12/usr.bin/tftp/main.c
==
--- stable/12/usr.bin/tftp/main.c Wed Oct 30 01:41:14 2019
(r354162)
+++ stable/12/usr.bin/tftp/main.c Wed Oct 30 01:41:49 2019
(r354163)
@@ -491,6 +491,7 @@ put(int argc, char *argv[])
if (fstat(fd, &sb) < 0) {
warn("%s", cp);
+ close(fd);
return;
}
asprintf(&options[OPT_TSIZE].o_request, "%ju", sb.st_size);
@@ -746,7 +747,7 @@ command(bool interactive, EditLine *el, History *hist,
exit(0);
len = MIN(MAXLINE, num);
memcpy(line, bp, len);
- line[len] = '\0';
+ line[len - 1] = '\0';
history(hist, hep, H_ENTER, bp);
} else {
line[0] = 0;
___
svn-src-all@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r354164 - in stable/12/sys/fs: nfs nfsclient
Author: rmacklem
Date: Wed Oct 30 01:57:40 2019
New Revision: 354164
URL: https://svnweb.freebsd.org/changeset/base/354164
Log:
MFC: r352736
Replace all mtx_assert() calls for n_mtx and ncl_iod_mutex with macros.
To be consistent with replacing the mtx_lock()/mtx_unlock() calls on
the NFS node mutex (n_mtx) and ncl_iod_mutex, this patch replaces
all mtx_assert() calls on these mutexes with macros as well.
This will simplify changing these locks to sx locks in a future commit.
However, this change may be delayed indefinitely, since it appears there
is a deadlock when vnode_pager_setsize() is called to shrink the size
and the NFS node lock is held.
There is no semantic change as a result of this commit.
Modified:
stable/12/sys/fs/nfs/nfsport.h
stable/12/sys/fs/nfsclient/nfs_clnfsiod.c
stable/12/sys/fs/nfsclient/nfs_clnode.c
Directory Properties:
stable/12/ (props changed)
Modified: stable/12/sys/fs/nfs/nfsport.h
==
--- stable/12/sys/fs/nfs/nfsport.h Wed Oct 30 01:41:49 2019
(r354163)
+++ stable/12/sys/fs/nfs/nfsport.h Wed Oct 30 01:57:40 2019
(r354164)
@@ -688,10 +688,12 @@ void nfsrvd_rcv(struct socket *, void *, int);
#defineNFSUNLOCKV4ROOTMUTEX() mtx_unlock(&nfs_v4root_mutex)
#defineNFSLOCKNODE(n) mtx_lock(&((n)->n_mtx))
#defineNFSUNLOCKNODE(n)mtx_unlock(&((n)->n_mtx))
+#defineNFSASSERTNODE(n)mtx_assert(&((n)->n_mtx), MA_OWNED)
#defineNFSLOCKMNT(m) mtx_lock(&((m)->nm_mtx))
#defineNFSUNLOCKMNT(m) mtx_unlock(&((m)->nm_mtx))
#defineNFSLOCKIOD()mtx_lock(&ncl_iod_mutex)
#defineNFSUNLOCKIOD() mtx_unlock(&ncl_iod_mutex)
+#defineNFSASSERTIOD() mtx_assert(&ncl_iod_mutex, MA_OWNED)
#defineNFSLOCKREQUEST(r) mtx_lock(&((r)->r_mtx))
#defineNFSUNLOCKREQUEST(r) mtx_unlock(&((r)->r_mtx))
#defineNFSPROCLISTLOCK() sx_slock(&allproc_lock)
Modified: stable/12/sys/fs/nfsclient/nfs_clnfsiod.c
==
--- stable/12/sys/fs/nfsclient/nfs_clnfsiod.c Wed Oct 30 01:41:49 2019
(r354163)
+++ stable/12/sys/fs/nfsclient/nfs_clnfsiod.c Wed Oct 30 01:57:40 2019
(r354164)
@@ -169,7 +169,7 @@ nfs_nfsiodnew_sync(void)
{
int error, i;
- mtx_assert(&ncl_iod_mutex, MA_OWNED);
+ NFSASSERTIOD();
for (i = 0; i < ncl_iodmax; i++) {
if (nfs_asyncdaemon[i] == 0) {
nfs_asyncdaemon[i] = 1;
@@ -206,7 +206,7 @@ void
ncl_nfsiodnew(void)
{
- mtx_assert(&ncl_iod_mutex, MA_OWNED);
+ NFSASSERTIOD();
taskqueue_enqueue(taskqueue_thread, &ncl_nfsiodnew_task);
}
Modified: stable/12/sys/fs/nfsclient/nfs_clnode.c
==
--- stable/12/sys/fs/nfsclient/nfs_clnode.c Wed Oct 30 01:41:49 2019
(r354163)
+++ stable/12/sys/fs/nfsclient/nfs_clnode.c Wed Oct 30 01:57:40 2019
(r354164)
@@ -212,7 +212,7 @@ ncl_releasesillyrename(struct vnode *vp, struct thread
ASSERT_VOP_ELOCKED(vp, "releasesillyrename");
np = VTONFS(vp);
- mtx_assert(&np->n_mtx, MA_OWNED);
+ NFSASSERTNODE(np);
if (vp->v_type != VDIR) {
sp = np->n_sillyrename;
np->n_sillyrename = NULL;
___
svn-src-all@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
svn commit: r354165 - in stable/12: . tests/sys/cddl/zfs/include tests/sys/cddl/zfs/tests/cli_root/zdb tests/sys/cddl/zfs/tests/cli_root/zpool_add tests/sys/cddl/zfs/tests/cli_root/zpool_create tes...
Author: asomers Date: Wed Oct 30 02:03:37 2019 New Revision: 354165 URL: https://svnweb.freebsd.org/changeset/base/354165 Log: MFC r353117-r353118, r353281-r353282, r353284-r353289, r353309-r353310, r353360-r353361, r353366, r353379 r353117: ZFS: the hotspare_add_004_neg test needs at least two disks Sponsored by: Axcient r353118: ZFS: fix several of the "zpool create" tests * Remove zpool_create_013_neg. FreeBSD doesn't have an equivalent of Solaris's metadevices. GEOM would be the equivalent, but since all geoms are the same from ZFS's perspective, this test would be redundant with zpool_create_012_neg * Remove zpool_create_014_neg. FreeBSD does not support swapping to regular files. * Remove zpool_create_016_pos. This test is redundant with literally every other test that creates a disk-backed pool. * s:/etc/vfstab:/etc/fstab in zpool_create_011_neg * Delete the VTOC-related portion of zpool_create_008_pos. FreeBSD doesn't use VTOC. * Replace dumpadm with dumpon and swap with swapon in multiple tests. * In zpool_create_015_neg, don't require "zpool create -n" to fail. It's reasonable for that variant to succeed, because it doesn't actually open the zvol. * Greatly simplify zpool_create_012_neg. Make it safer, too, but not interfering with the system's regular swap devices. * Expect zpool_create_011_neg to fail (PR 241070) * Delete some redundant cleanup steps in various tests * Remove some unneeeded ATF timeout specifications. The default is fine. PR: 241070 Sponsored by: Axcient r353281: ZFS: fix several zvol_misc tests * Adapt zvol_misc_001_neg to use dumpon instead of Solaris's dumpadm * Disable zvol_misc_003_neg, zvol_misc_005_neg, and zvol_misc_006_pos, because they involve using a zvol as a dump device, which FreeBSD does not yet support. Sponsored by: Axcient r353282: zfs: fix the slog_012_neg test This test attempts to corrupt a file-backed vdev by deleting it and then recreating it with truncate. But that doesn't work, because the pool already has the vdev open, and it happily hangs on to the open-but-deleted file. Fix by truncating the file without deleting it. Sponsored by: Axcient r353284: ZFS: fix the zpool_get_002_pos test ZFS has grown some additional properties that hadn't been added to the config file yet. While I'm here, improve the error message, and remove a superfluous command. Sponsored by: Axcient r353285: zfs: fix the zdb_001_neg test The test needed to be updated for r331701 (MFV illumos 8671400), which added a "-k" option. Sponsored by: Axcient r353286: zfs: skip the zfsd tests if zfsd is not running Sponsored by: Axcient Differential Revision:https://reviews.freebsd.org/D21878 r353287: ZFS: fix the delegate tests These tests have never worked correctly * Replace runwattr with sudo * Fix a scoping bug with the "dtst" variable * Cleanup user properties created during tests * Eliminate the checks for refreservation and send support. They will always be supported. * Fix verify_fs_snapshot. It seemed to assume that permissions would not yet be delegated, but that's not how it's actually used. * Combine verify_fs_promote with verify_vol_promote * Remove some useless sleeps * Fix backwards condition in verify_vol_volsize * Remove some redundant cleanup steps in the tests. cleanup.ksh will handle everything. * Disable some parts of the tests that FreeBSD doesn't support: * Creating snapshots with mkdir * devices * shareisci * sharenfs * xattr * zoned The sharenfs parts could probably be reenabled with more work to remove the Solarisms. Sponsored by: Axcient Differential Revision:https://reviews.freebsd.org/D21898 r353288: ZFS: mark hotspare_scrub_002_pos as an expected failure "zpool scrub" doesn't detect all errors on active spares in raidz arrays PR: 241069 Sponsored by: Axcient r353289: ZFS: fix the redundancy tests * Fix force_sync_path, which ensures that a file is fully flushed to disk. Apparently "zpool history"'s performance has improved, but exporting and importing the pool still works. * Fix file_dva by using undocumented zdb syntax to clarify that we're interested in the pool's root file system, not the pool itself. This should also fix the zpool_clear_001_pos test. * Remove a redundant cleanup step Sponsored by: Axcient Differential Revision:https://reviews.freebsd.org/D21901 r353309: zfs: fix the zfsd_autoreplace_003_pos test The test declared that it only needed 5 disks, but actually tried to use 6. Fix it to use just 5, which is all it really needs. Sponsored by: Axcient r353310: zfs: fix the zfsd_hotspare_007_pos tes
svn commit: r354166 - stable/12/sys/cddl/contrib/opensolaris/uts/common/fs/zfs
Author: asomers
Date: Wed Oct 30 02:33:43 2019
New Revision: 354166
URL: https://svnweb.freebsd.org/changeset/base/354166
Log:
MFC r353439:
MFZol: Fix performance of "zfs recv" with many deletions
This patch fixes 2 issues with the DMU free throttle implemented
in dmu_free_long_range(). The first issue is that get_next_chunk()
was calculating the number of L1 blocks the free would dirty
incorrectly. In some cases involving extremely large files, this
code would greatly overestimate the number of affected L1 blocks,
causing excessive calls to txg_wait_open(). This patch corrects
the calculation.
The second issue is that the free throttle uses the total number
of free'd blocks in all (open, quiescing, and syncing) txgs to
determine whether to throttle. This causes large frees (such as
those created by the first issue) to cause 4 txg syncs before
any further frees were allowed to proceed. This patch ensures
that the accounting is done entirely in a per-txg fashion, so
that frees from a given txg don't affect those that immediately
follow it.
Reviewed-by: Brian Behlendorf
Reviewed-by: Matthew Ahrens
Signed-off-by: Tom Caputi
zfsonlinux/zfs@f4c594da94d856c422512a54e48070f890b2685b
Freeing throttle should account for holes
Deletion throttle currently does not account for holes in a file.
This means that it can activate when it shouldn't.
To fix it we switch the throttle to be based on the number of
L1 blocks we will have to dirty when freeing
Reviewed-by: Tom Caputi
Reviewed-by: Matt Ahrens
Reviewed-by: Brian Behlendorf
Signed-off-by: Alek Pinchuk
zfsonlinux/zfs@65282ee9e06b130f1f0169baf5d9bf0dd8fc1ef9
Submitted by: Alek Pinchuk
Reviewed by: allanjude
Sponsored by: Axcient
Differential Revision:https://reviews.freebsd.org/D21895
Modified:
stable/12/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu.c
Directory Properties:
stable/12/ (props changed)
Modified: stable/12/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu.c
==
--- stable/12/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu.c Wed Oct
30 02:03:37 2019(r354165)
+++ stable/12/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu.c Wed Oct
30 02:33:43 2019(r354166)
@@ -21,6 +21,7 @@
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2011, 2017 by Delphix. All rights reserved.
+ * Copyright (c) 2019 Datto Inc.
*/
/* Copyright (c) 2013 by Saso Kiselkov. All rights reserved. */
/* Copyright (c) 2013, Joyent, Inc. All rights reserved. */
@@ -62,14 +63,15 @@ SYSCTL_INT(_vfs_zfs, OID_AUTO, nopwrite_enabled, CTLFL
&zfs_nopwrite_enabled, 0, "Enable nopwrite feature");
/*
- * Tunable to control percentage of dirtied blocks from frees in one TXG.
- * After this threshold is crossed, additional dirty blocks from frees
- * wait until the next TXG.
+ * Tunable to control percentage of dirtied L1 blocks from frees allowed into
+ * one TXG. After this threshold is crossed, additional dirty blocks from frees
+ * will wait until the next TXG.
* A value of zero will disable this throttle.
*/
-uint32_t zfs_per_txg_dirty_frees_percent = 30;
+uint32_t zfs_per_txg_dirty_frees_percent = 5;
SYSCTL_INT(_vfs_zfs, OID_AUTO, per_txg_dirty_frees_percent, CTLFLAG_RWTUN,
- &zfs_per_txg_dirty_frees_percent, 0, "Percentage of dirtied blocks from
frees in one txg");
+ &zfs_per_txg_dirty_frees_percent, 0,
+ "Percentage of dirtied indirect blocks from frees allowed in one txg");
/*
* This can be used for testing, to ensure that certain actions happen
@@ -683,11 +685,13 @@ dmu_prefetch(objset_t *os, uint64_t object, int64_t le
*
* On input, *start should be the first offset that does not need to be
* freed (e.g. "offset + length"). On return, *start will be the first
- * offset that should be freed.
+ * offset that should be freed and l1blks is set to the number of level 1
+ * indirect blocks found within the chunk.
*/
static int
-get_next_chunk(dnode_t *dn, uint64_t *start, uint64_t minimum)
+get_next_chunk(dnode_t *dn, uint64_t *start, uint64_t minimum, uint64_t
*l1blks)
{
+ uint64_t blks;
uint64_t maxblks = DMU_MAX_ACCESS >> (dn->dn_indblkshift + 1);
/* bytes of data covered by a level-1 indirect block */
uint64_t iblkrange =
@@ -695,13 +699,23 @@ get_next_chunk(dnode_t *dn, uint64_t *start, uint64_t
ASSERT3U(minimum, <=, *start);
- if (*start - minimum <= iblkrange * maxblks) {
+ /*
+* Check if we can free the entire range assuming that all of the
+* L1 blocks in this range have data. If we can, we use this
+* worst case value as an estimate so we can avoid having to look
+* at the object's actual data.
+*/
+ uint64_t total_l1blks =
+ (roundup(*start, iblkr
svn commit: r354167 - head/sys/x86/include
Author: cem Date: Wed Oct 30 04:00:44 2019 New Revision: 354167 URL: https://svnweb.freebsd.org/changeset/base/354167 Log: amd64: Fix typo: RDPRU bit is 0x10, not 0x04 Bit 4 != 4, of course. X-MFC-With: r354162 Modified: head/sys/x86/include/specialreg.h Modified: head/sys/x86/include/specialreg.h == --- head/sys/x86/include/specialreg.h Wed Oct 30 02:33:43 2019 (r354166) +++ head/sys/x86/include/specialreg.h Wed Oct 30 04:00:44 2019 (r354167) @@ -386,7 +386,7 @@ #defineAMDFEID_CLZERO 0x0001 #defineAMDFEID_IRPERF 0x0002 #defineAMDFEID_XSAVEERPTR 0x0004 -#defineAMDFEID_RDPRU 0x0004 +#defineAMDFEID_RDPRU 0x0010 #defineAMDFEID_MCOMMIT 0x0100 #defineAMDFEID_WBNOINVD0x0200 #defineAMDFEID_IBPB0x1000 ___ svn-src-all@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-all To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
