svn commit: r367581 - head/sys/conf
Author: imp
Date: Tue Nov 10 23:25:16 2020
New Revision: 367581
URL: https://svnweb.freebsd.org/changeset/base/367581
Log:
Add INIT_ALL_ZERO and INIT_ALL_PATTERN to kern.opts.mk
These options need to be in the kern.opts.mk file to be alive for kernel
and module builds. This also reverts r367579 since that's not needed with
this fix: the host's bsd.opts.mk is irrelevant.
Reviewed by: brooks@
Differential Revision: https://reviews.freebsd.org/D27170
Modified:
head/sys/conf/kern.mk
head/sys/conf/kern.opts.mk
Modified: head/sys/conf/kern.mk
==
--- head/sys/conf/kern.mk Tue Nov 10 21:29:10 2020(r367580)
+++ head/sys/conf/kern.mk Tue Nov 10 23:25:16 2020(r367581)
@@ -230,15 +230,15 @@ CFLAGS+= -mretpoline
#
# Initialize stack variables on function entry
#
-.if defined(MK_INIT_ALL_ZERO) && ${MK_INIT_ALL_ZERO} == "yes"
-.if defined(COMPILER_FEATURES) && ${COMPILER_FEATURES:Minit-all}
+.if ${MK_INIT_ALL_ZERO} == "yes"
+.if ${COMPILER_FEATURES:Minit-all}
CFLAGS+= -ftrivial-auto-var-init=zero \
-enable-trivial-auto-var-init-zero-knowing-it-will-be-removed-from-clang
.else
.warning InitAll (zeros) requested but not support by compiler
.endif
-.elif defined(MK_INIT_ALL_PATTERN) && ${MK_INIT_ALL_PATTERN} == "yes"
-.if defined(COMPILER_FEATURES) && ${COMPILER_FEATURES:Minit-all}
+.elif ${MK_INIT_ALL_PATTERN} == "yes"
+.if ${COMPILER_FEATURES:Minit-all}
CFLAGS+= -ftrivial-auto-var-init=pattern
.else
.warning InitAll (pattern) requested but not support by compiler
Modified: head/sys/conf/kern.opts.mk
==
--- head/sys/conf/kern.opts.mk Tue Nov 10 21:29:10 2020(r367580)
+++ head/sys/conf/kern.opts.mk Tue Nov 10 23:25:16 2020(r367581)
@@ -52,6 +52,8 @@ __DEFAULT_YES_OPTIONS = \
__DEFAULT_NO_OPTIONS = \
BHYVE_SNAPSHOT \
EXTRA_TCP_STACKS \
+INIT_ALL_PATTERN \
+INIT_ALL_ZERO \
KERNEL_RETPOLINE \
OFED \
RATELIMIT \
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r367577 - in head: share/mk sys/conf tools/build/options
On Tue, Nov 10, 2020 at 02:44:45PM -0500, Shawn Webb wrote: > On Tue, Nov 10, 2020 at 07:17:29PM +, Brooks Davis wrote: > > On Tue, Nov 10, 2020 at 07:15:14PM +, Brooks Davis wrote: > > > Author: brooks > > > Date: Tue Nov 10 19:15:13 2020 > > > New Revision: 367577 > > > URL: https://svnweb.freebsd.org/changeset/base/367577 > > > > > > Log: > > > Support initializing stack variables on function entry > > > > > > There are two options: > > >- WITH_INIT_ALL_ZERO: Zero all variables on the stack. > > >- WITH_INIT_ALL_PATTERN: Initialize variables with well-defined > > > patterns. > > > > > > The exact pattern are a compiler implementation detail and vary by type. > > > They are somewhat documented in the LLVM commit message: > > > https://reviews.llvm.org/rL349442 > > > I've used WITH_INIT_ALL_* to match Microsoft's InitAll feature rather > > > than naming them after the LLVM specific compiler flags. > > > > > > In a range of consumer products, options like these are used in > > > both debug and production builds with debugs builds using patterns > > > (intended to provoke crashes on use of uninitialized values) and > > > production using zeros (deemed more likely to lead to harmless > > > misbehavior or NULL-pointer dereferences). > > > > We've tested this extensively in CheriBSD on RISC-V, in the wild it's > > probably most tested on Arm64 and x86. > > > > Despite the silly compiler flag you'll spot in the code, the zeroing > > option isn't going away in practice as Apple, Google, and Microsoft all > > ship with this feature in some of their products. > > HardenedBSD's testing of this last year on amd64 have (privately) > shown the feature to really hinder performance on more complex > applications (like when applied to clang/lld). A build of base > without init all zero applied to clang/lld would take around 1.5 > hours on my system. A build with it applied to clang/lld took around > four hours, if my memory serves correctly. I would probably advise > against applying it system-wide. But YMMV. I agree a more nuanced approach is likely useful in practice, but this does work and is part of the configuration we shipped for DARPA's FETT bug bounty. Hopefully this provides a starting point for further exploration. -- Brooks signature.asc Description: PGP signature
Re: svn commit: r367577 - in head: share/mk sys/conf tools/build/options
Sorry about that. I've fixed it in r367579.
-- Brooks
On Tue, Nov 10, 2020 at 12:46:45PM -0800, Matthew Macy wrote:
> These flags aren't defined by default when building external kernel modules:
>
> gmake[2]: Entering directory '/usr/home/matt/devel/ZoF/module'
> env -u MAKEFLAGS make -C /home/matt/devel/ZoF/module -f Makefile.bsd -w
> make[3]: Entering directory `/home/matt/devel/ZoF/module'
> make[3]: "/usr/home/matt/devel/freebsd/sys/conf/kern.mk" line 233:
> Malformed conditional (${MK_INIT_ALL_ZERO} == "yes")
> make[3]: Fatal errors encountered -- cannot continue
> make[3]: stopped in /home/matt/devel/ZoF/module
> gmake[2]: *** [Makefile:53: modules-FreeBSD] Error 1
> gmake[2]: Leaving directory '/usr/home/matt/devel/ZoF/module'
>
> On Tue, Nov 10, 2020 at 11:15 AM Brooks Davis wrote:
> >
> > Author: brooks
> > Date: Tue Nov 10 19:15:13 2020
> > New Revision: 367577
> > URL: https://svnweb.freebsd.org/changeset/base/367577
> >
> > Log:
> > Support initializing stack variables on function entry
> >
> > There are two options:
> >- WITH_INIT_ALL_ZERO: Zero all variables on the stack.
> >- WITH_INIT_ALL_PATTERN: Initialize variables with well-defined patterns.
> >
> > The exact pattern are a compiler implementation detail and vary by type.
> > They are somewhat documented in the LLVM commit message:
> > https://reviews.llvm.org/rL349442
> > I've used WITH_INIT_ALL_* to match Microsoft's InitAll feature rather
> > than naming them after the LLVM specific compiler flags.
> >
> > In a range of consumer products, options like these are used in
> > both debug and production builds with debugs builds using patterns
> > (intended to provoke crashes on use of uninitialized values) and
> > production using zeros (deemed more likely to lead to harmless
> > misbehavior or NULL-pointer dereferences).
> >
> > Reviewed by: emaste
> > Obtained from:CheriBSD
> > Sponsored by: DARPA
> > Differential Revision:https://reviews.freebsd.org/D27131
> >
> > Added:
> > head/tools/build/options/WITH_INIT_ALL_PATTERN (contents, props changed)
> > head/tools/build/options/WITH_INIT_ALL_ZERO (contents, props changed)
> > Modified:
> > head/share/mk/bsd.compiler.mk
> > head/share/mk/bsd.lib.mk
> > head/share/mk/bsd.opts.mk
> > head/share/mk/bsd.prog.mk
> > head/sys/conf/kern.mk
> >
> > Modified: head/share/mk/bsd.compiler.mk
> > ==
> > --- head/share/mk/bsd.compiler.mk Tue Nov 10 19:09:35 2020
> > (r367576)
> > +++ head/share/mk/bsd.compiler.mk Tue Nov 10 19:15:13 2020
> > (r367577)
> > @@ -24,6 +24,7 @@
> > # - c++11: supports full (or nearly full) C++11 programming
> > environment.
> > # - retpoline: supports the retpoline speculative execution vulnerability
> > # mitigation.
> > +# - init-all: supports stack variable initialization.
> > #
> > # These variables with an X_ prefix will also be provided if XCC is set.
> > #
> > @@ -214,7 +215,7 @@ ${X_}COMPILER_FEATURES= c++11 c++14
> > ${X_}COMPILER_FEATURES+= c++17
> > .endif
> > .if ${${X_}COMPILER_TYPE} == "clang"
> > -${X_}COMPILER_FEATURES+= retpoline
> > +${X_}COMPILER_FEATURES+= retpoline init-all
> > .endif
> >
> > .else
> >
> > Modified: head/share/mk/bsd.lib.mk
> > ==
> > --- head/share/mk/bsd.lib.mkTue Nov 10 19:09:35 2020(r367576)
> > +++ head/share/mk/bsd.lib.mkTue Nov 10 19:15:13 2020(r367577)
> > @@ -85,6 +85,25 @@ LDFLAGS+= -Wl,-zretpolineplt
> > .endif
> > .endif
> >
> > +# Initialize stack variables on function entry
> > +.if ${MK_INIT_ALL_ZERO} == "yes"
> > +.if ${COMPILER_FEATURES:Minit-all}
> > +CFLAGS+= -ftrivial-auto-var-init=zero \
> > +
> > -enable-trivial-auto-var-init-zero-knowing-it-will-be-removed-from-clang
> > +CXXFLAGS+= -ftrivial-auto-var-init=zero \
> > +
> > -enable-trivial-auto-var-init-zero-knowing-it-will-be-removed-from-clang
> > +.else
> > +.warning InitAll (zeros) requested but not support by compiler
> > +.endif
> > +.elif ${MK_INIT_ALL_PATTERN} == "yes"
> > +.if ${COMPILER_FEATURES:Minit-all}
> > +CFLAGS+= -ftrivial-auto-var-init=pattern
> > +CXXFLAGS+= -ftrivial-auto-var-init=pattern
> > +.else
> > +.warning InitAll (pattern) requested but not support by compiler
> > +.endif
> > +.endif
> > +
> > .if ${MK_DEBUG_FILES} != "no" && empty(DEBUG_FLAGS:M-g) && \
> > empty(DEBUG_FLAGS:M-gdwarf*)
> > CFLAGS+= ${DEBUG_FILES_CFLAGS}
> >
> > Modified: head/share/mk/bsd.opts.mk
> > ==
> > --- head/share/mk/bsd.opts.mk Tue Nov 10 19:09:35 2020(r367576)
> > +++ head/share/mk/bsd.opts.mk Tue Nov 10 19:15:13 2020(r367577)
> > @@ -71,6 +71,8 @@ __DEFAULT_NO_OPTIONS = \
> > BIND_NOW \
> >
svn commit: r367580 - head/sys/kern
Author: mjg
Date: Tue Nov 10 21:29:10 2020
New Revision: 367580
URL: https://svnweb.freebsd.org/changeset/base/367580
Log:
thread: tidy up r367543
"locked" variable is spurious in the committed version.
Modified:
head/sys/kern/kern_proc.c
Modified: head/sys/kern/kern_proc.c
==
--- head/sys/kern/kern_proc.c Tue Nov 10 21:12:32 2020(r367579)
+++ head/sys/kern/kern_proc.c Tue Nov 10 21:29:10 2020(r367580)
@@ -2748,7 +2748,6 @@ sysctl_kern_proc_kstack(SYSCTL_HANDLER_ARGS)
struct stack *st;
struct sbuf sb;
struct proc *p;
- bool locked;
name = (int *)arg1;
error = pget((pid_t)name[0], PGET_NOTINEXEC | PGET_WANTREAD, );
@@ -2789,14 +2788,12 @@ sysctl_kern_proc_kstack(SYSCTL_HANDLER_ARGS)
i++;
}
PROC_UNLOCK(p);
- locked = false;
numthreads = i;
for (i = 0; i < numthreads; i++) {
td = tdfind(lwpidarray[i], p->p_pid);
if (td == NULL) {
continue;
}
- locked = true;
bzero(kkstp, sizeof(*kkstp));
(void)sbuf_new(, kkstp->kkst_trace,
sizeof(kkstp->kkst_trace), SBUF_FIXEDLEN);
@@ -2810,7 +2807,6 @@ sysctl_kern_proc_kstack(SYSCTL_HANDLER_ARGS)
kkstp->kkst_state = KKST_STATE_RUNNING;
thread_unlock(td);
PROC_UNLOCK(p);
- locked = false;
stack_sbuf_print(, st);
sbuf_finish();
sbuf_delete();
@@ -2818,10 +2814,7 @@ sysctl_kern_proc_kstack(SYSCTL_HANDLER_ARGS)
if (error)
break;
}
- if (!locked)
- PROC_LOCK(p);
- _PRELE(p);
- PROC_UNLOCK(p);
+ PRELE(p);
if (lwpidarray != NULL)
free(lwpidarray, M_TEMP);
stack_destroy(st);
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r367579 - head/sys/conf
Author: brooks
Date: Tue Nov 10 21:12:32 2020
New Revision: 367579
URL: https://svnweb.freebsd.org/changeset/base/367579
Log:
Be more tolerant of share/mk and kern.mk mismatch
When building out-of-tree modules, it appears that the system share/mk
is used, but sys/conf/kern.mk is used. That results in MK_INIT_ALL_ZERO
being undefined. In the interest of maximum compatability, check
that MK_INIT_ALL_* and COMPILER_FEATURES are defined before comparing
their values.
Reported by: mmacy
Sponsored by: DARPA
Modified:
head/sys/conf/kern.mk
Modified: head/sys/conf/kern.mk
==
--- head/sys/conf/kern.mk Tue Nov 10 19:54:39 2020(r367578)
+++ head/sys/conf/kern.mk Tue Nov 10 21:12:32 2020(r367579)
@@ -230,15 +230,15 @@ CFLAGS+= -mretpoline
#
# Initialize stack variables on function entry
#
-.if ${MK_INIT_ALL_ZERO} == "yes"
-.if ${COMPILER_FEATURES:Minit-all}
+.if defined(MK_INIT_ALL_ZERO) && ${MK_INIT_ALL_ZERO} == "yes"
+.if defined(COMPILER_FEATURES) && ${COMPILER_FEATURES:Minit-all}
CFLAGS+= -ftrivial-auto-var-init=zero \
-enable-trivial-auto-var-init-zero-knowing-it-will-be-removed-from-clang
.else
.warning InitAll (zeros) requested but not support by compiler
.endif
-.elif ${MK_INIT_ALL_PATTERN} == "yes"
-.if ${COMPILER_FEATURES:Minit-all}
+.elif defined(MK_INIT_ALL_PATTERN) && ${MK_INIT_ALL_PATTERN} == "yes"
+.if defined(COMPILER_FEATURES) && ${COMPILER_FEATURES:Minit-all}
CFLAGS+= -ftrivial-auto-var-init=pattern
.else
.warning InitAll (pattern) requested but not support by compiler
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r367577 - in head: share/mk sys/conf tools/build/options
These flags aren't defined by default when building external kernel modules:
gmake[2]: Entering directory '/usr/home/matt/devel/ZoF/module'
env -u MAKEFLAGS make -C /home/matt/devel/ZoF/module -f Makefile.bsd -w
make[3]: Entering directory `/home/matt/devel/ZoF/module'
make[3]: "/usr/home/matt/devel/freebsd/sys/conf/kern.mk" line 233:
Malformed conditional (${MK_INIT_ALL_ZERO} == "yes")
make[3]: Fatal errors encountered -- cannot continue
make[3]: stopped in /home/matt/devel/ZoF/module
gmake[2]: *** [Makefile:53: modules-FreeBSD] Error 1
gmake[2]: Leaving directory '/usr/home/matt/devel/ZoF/module'
On Tue, Nov 10, 2020 at 11:15 AM Brooks Davis wrote:
>
> Author: brooks
> Date: Tue Nov 10 19:15:13 2020
> New Revision: 367577
> URL: https://svnweb.freebsd.org/changeset/base/367577
>
> Log:
> Support initializing stack variables on function entry
>
> There are two options:
>- WITH_INIT_ALL_ZERO: Zero all variables on the stack.
>- WITH_INIT_ALL_PATTERN: Initialize variables with well-defined patterns.
>
> The exact pattern are a compiler implementation detail and vary by type.
> They are somewhat documented in the LLVM commit message:
> https://reviews.llvm.org/rL349442
> I've used WITH_INIT_ALL_* to match Microsoft's InitAll feature rather
> than naming them after the LLVM specific compiler flags.
>
> In a range of consumer products, options like these are used in
> both debug and production builds with debugs builds using patterns
> (intended to provoke crashes on use of uninitialized values) and
> production using zeros (deemed more likely to lead to harmless
> misbehavior or NULL-pointer dereferences).
>
> Reviewed by: emaste
> Obtained from:CheriBSD
> Sponsored by: DARPA
> Differential Revision:https://reviews.freebsd.org/D27131
>
> Added:
> head/tools/build/options/WITH_INIT_ALL_PATTERN (contents, props changed)
> head/tools/build/options/WITH_INIT_ALL_ZERO (contents, props changed)
> Modified:
> head/share/mk/bsd.compiler.mk
> head/share/mk/bsd.lib.mk
> head/share/mk/bsd.opts.mk
> head/share/mk/bsd.prog.mk
> head/sys/conf/kern.mk
>
> Modified: head/share/mk/bsd.compiler.mk
> ==
> --- head/share/mk/bsd.compiler.mk Tue Nov 10 19:09:35 2020
> (r367576)
> +++ head/share/mk/bsd.compiler.mk Tue Nov 10 19:15:13 2020
> (r367577)
> @@ -24,6 +24,7 @@
> # - c++11: supports full (or nearly full) C++11 programming environment.
> # - retpoline: supports the retpoline speculative execution vulnerability
> # mitigation.
> +# - init-all: supports stack variable initialization.
> #
> # These variables with an X_ prefix will also be provided if XCC is set.
> #
> @@ -214,7 +215,7 @@ ${X_}COMPILER_FEATURES= c++11 c++14
> ${X_}COMPILER_FEATURES+= c++17
> .endif
> .if ${${X_}COMPILER_TYPE} == "clang"
> -${X_}COMPILER_FEATURES+= retpoline
> +${X_}COMPILER_FEATURES+= retpoline init-all
> .endif
>
> .else
>
> Modified: head/share/mk/bsd.lib.mk
> ==
> --- head/share/mk/bsd.lib.mkTue Nov 10 19:09:35 2020(r367576)
> +++ head/share/mk/bsd.lib.mkTue Nov 10 19:15:13 2020(r367577)
> @@ -85,6 +85,25 @@ LDFLAGS+= -Wl,-zretpolineplt
> .endif
> .endif
>
> +# Initialize stack variables on function entry
> +.if ${MK_INIT_ALL_ZERO} == "yes"
> +.if ${COMPILER_FEATURES:Minit-all}
> +CFLAGS+= -ftrivial-auto-var-init=zero \
> +-enable-trivial-auto-var-init-zero-knowing-it-will-be-removed-from-clang
> +CXXFLAGS+= -ftrivial-auto-var-init=zero \
> +-enable-trivial-auto-var-init-zero-knowing-it-will-be-removed-from-clang
> +.else
> +.warning InitAll (zeros) requested but not support by compiler
> +.endif
> +.elif ${MK_INIT_ALL_PATTERN} == "yes"
> +.if ${COMPILER_FEATURES:Minit-all}
> +CFLAGS+= -ftrivial-auto-var-init=pattern
> +CXXFLAGS+= -ftrivial-auto-var-init=pattern
> +.else
> +.warning InitAll (pattern) requested but not support by compiler
> +.endif
> +.endif
> +
> .if ${MK_DEBUG_FILES} != "no" && empty(DEBUG_FLAGS:M-g) && \
> empty(DEBUG_FLAGS:M-gdwarf*)
> CFLAGS+= ${DEBUG_FILES_CFLAGS}
>
> Modified: head/share/mk/bsd.opts.mk
> ==
> --- head/share/mk/bsd.opts.mk Tue Nov 10 19:09:35 2020(r367576)
> +++ head/share/mk/bsd.opts.mk Tue Nov 10 19:15:13 2020(r367577)
> @@ -71,6 +71,8 @@ __DEFAULT_NO_OPTIONS = \
> BIND_NOW \
> CCACHE_BUILD \
> CTF \
> +INIT_ALL_PATTERN \
> +INIT_ALL_ZERO \
> INSTALL_AS_USER \
> PIE \
> RETPOLINE \
> @@ -84,6 +86,10 @@ __DEFAULT_DEPENDENT_OPTIONS = \
>
>
> .include
> +
> +.if ${MK_INIT_ALL_PATTERN} == "yes" && ${MK_INIT_ALL_ZERO} == "yes"
> +.error WITH_INIT_ALL_PATTERN and WITH_INIT_ALL_ZERO are mutually exclusive.
> +.endif
>
> #
svn commit: r367578 - head/sys/dev/cxgbe/tom
Author: jhb Date: Tue Nov 10 19:54:39 2020 New Revision: 367578 URL: https://svnweb.freebsd.org/changeset/base/367578 Log: Clear tp->tod in t4_pcb_detach(). Otherwise, a socket can have a non-NULL tp->tod while TF_TOE is clear. In particular, if a newly accepted socket falls back to non-TOE due to an active open failure, the non-TOE socket will still have tp->tod set even though TF_TOE is clear. Reviewed by: np MFC after:2 weeks Sponsored by: Chelsio Communications Differential Revision:https://reviews.freebsd.org/D27028 Modified: head/sys/dev/cxgbe/tom/t4_tom.c Modified: head/sys/dev/cxgbe/tom/t4_tom.c == --- head/sys/dev/cxgbe/tom/t4_tom.c Tue Nov 10 19:15:13 2020 (r367577) +++ head/sys/dev/cxgbe/tom/t4_tom.c Tue Nov 10 19:54:39 2020 (r367578) @@ -382,6 +382,7 @@ t4_pcb_detach(struct toedev *tod __unused, struct tcpc } #endif + tp->tod = NULL; tp->t_toe = NULL; tp->t_flags &= ~TF_TOE; toep->flags &= ~TPF_ATTACHED; ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r367577 - in head: share/mk sys/conf tools/build/options
On Tue, Nov 10, 2020 at 07:17:29PM +, Brooks Davis wrote: > On Tue, Nov 10, 2020 at 07:15:14PM +, Brooks Davis wrote: > > Author: brooks > > Date: Tue Nov 10 19:15:13 2020 > > New Revision: 367577 > > URL: https://svnweb.freebsd.org/changeset/base/367577 > > > > Log: > > Support initializing stack variables on function entry > > > > There are two options: > >- WITH_INIT_ALL_ZERO: Zero all variables on the stack. > >- WITH_INIT_ALL_PATTERN: Initialize variables with well-defined patterns. > > > > The exact pattern are a compiler implementation detail and vary by type. > > They are somewhat documented in the LLVM commit message: > > https://reviews.llvm.org/rL349442 > > I've used WITH_INIT_ALL_* to match Microsoft's InitAll feature rather > > than naming them after the LLVM specific compiler flags. > > > > In a range of consumer products, options like these are used in > > both debug and production builds with debugs builds using patterns > > (intended to provoke crashes on use of uninitialized values) and > > production using zeros (deemed more likely to lead to harmless > > misbehavior or NULL-pointer dereferences). > > We've tested this extensively in CheriBSD on RISC-V, in the wild it's > probably most tested on Arm64 and x86. > > Despite the silly compiler flag you'll spot in the code, the zeroing > option isn't going away in practice as Apple, Google, and Microsoft all > ship with this feature in some of their products. HardenedBSD's testing of this last year on amd64 have (privately) shown the feature to really hinder performance on more complex applications (like when applied to clang/lld). A build of base without init all zero applied to clang/lld would take around 1.5 hours on my system. A build with it applied to clang/lld took around four hours, if my memory serves correctly. I would probably advise against applying it system-wide. But YMMV. Thanks, -- Shawn Webb Cofounder / Security Engineer HardenedBSD GPG Key ID: 0xFF2E67A277F8E1FA GPG Key Fingerprint: D206 BB45 15E0 9C49 0CF9 3633 C85B 0AF8 AB23 0FB2 https://git-01.md.hardenedbsd.org/HardenedBSD/pubkeys/src/branch/master/Shawn_Webb/03A4CBEBB82EA5A67D9F3853FF2E67A277F8E1FA.pub.asc signature.asc Description: PGP signature
Re: svn commit: r367577 - in head: share/mk sys/conf tools/build/options
On Tue, Nov 10, 2020 at 07:15:14PM +, Brooks Davis wrote: > Author: brooks > Date: Tue Nov 10 19:15:13 2020 > New Revision: 367577 > URL: https://svnweb.freebsd.org/changeset/base/367577 > > Log: > Support initializing stack variables on function entry > > There are two options: >- WITH_INIT_ALL_ZERO: Zero all variables on the stack. >- WITH_INIT_ALL_PATTERN: Initialize variables with well-defined patterns. > > The exact pattern are a compiler implementation detail and vary by type. > They are somewhat documented in the LLVM commit message: > https://reviews.llvm.org/rL349442 > I've used WITH_INIT_ALL_* to match Microsoft's InitAll feature rather > than naming them after the LLVM specific compiler flags. > > In a range of consumer products, options like these are used in > both debug and production builds with debugs builds using patterns > (intended to provoke crashes on use of uninitialized values) and > production using zeros (deemed more likely to lead to harmless > misbehavior or NULL-pointer dereferences). We've tested this extensively in CheriBSD on RISC-V, in the wild it's probably most tested on Arm64 and x86. Despite the silly compiler flag you'll spot in the code, the zeroing option isn't going away in practice as Apple, Google, and Microsoft all ship with this feature in some of their products. -- Brooks signature.asc Description: PGP signature
svn commit: r367577 - in head: share/mk sys/conf tools/build/options
Author: brooks
Date: Tue Nov 10 19:15:13 2020
New Revision: 367577
URL: https://svnweb.freebsd.org/changeset/base/367577
Log:
Support initializing stack variables on function entry
There are two options:
- WITH_INIT_ALL_ZERO: Zero all variables on the stack.
- WITH_INIT_ALL_PATTERN: Initialize variables with well-defined patterns.
The exact pattern are a compiler implementation detail and vary by type.
They are somewhat documented in the LLVM commit message:
https://reviews.llvm.org/rL349442
I've used WITH_INIT_ALL_* to match Microsoft's InitAll feature rather
than naming them after the LLVM specific compiler flags.
In a range of consumer products, options like these are used in
both debug and production builds with debugs builds using patterns
(intended to provoke crashes on use of uninitialized values) and
production using zeros (deemed more likely to lead to harmless
misbehavior or NULL-pointer dereferences).
Reviewed by: emaste
Obtained from:CheriBSD
Sponsored by: DARPA
Differential Revision:https://reviews.freebsd.org/D27131
Added:
head/tools/build/options/WITH_INIT_ALL_PATTERN (contents, props changed)
head/tools/build/options/WITH_INIT_ALL_ZERO (contents, props changed)
Modified:
head/share/mk/bsd.compiler.mk
head/share/mk/bsd.lib.mk
head/share/mk/bsd.opts.mk
head/share/mk/bsd.prog.mk
head/sys/conf/kern.mk
Modified: head/share/mk/bsd.compiler.mk
==
--- head/share/mk/bsd.compiler.mk Tue Nov 10 19:09:35 2020
(r367576)
+++ head/share/mk/bsd.compiler.mk Tue Nov 10 19:15:13 2020
(r367577)
@@ -24,6 +24,7 @@
# - c++11: supports full (or nearly full) C++11 programming environment.
# - retpoline: supports the retpoline speculative execution vulnerability
# mitigation.
+# - init-all: supports stack variable initialization.
#
# These variables with an X_ prefix will also be provided if XCC is set.
#
@@ -214,7 +215,7 @@ ${X_}COMPILER_FEATURES= c++11 c++14
${X_}COMPILER_FEATURES+= c++17
.endif
.if ${${X_}COMPILER_TYPE} == "clang"
-${X_}COMPILER_FEATURES+= retpoline
+${X_}COMPILER_FEATURES+= retpoline init-all
.endif
.else
Modified: head/share/mk/bsd.lib.mk
==
--- head/share/mk/bsd.lib.mkTue Nov 10 19:09:35 2020(r367576)
+++ head/share/mk/bsd.lib.mkTue Nov 10 19:15:13 2020(r367577)
@@ -85,6 +85,25 @@ LDFLAGS+= -Wl,-zretpolineplt
.endif
.endif
+# Initialize stack variables on function entry
+.if ${MK_INIT_ALL_ZERO} == "yes"
+.if ${COMPILER_FEATURES:Minit-all}
+CFLAGS+= -ftrivial-auto-var-init=zero \
+-enable-trivial-auto-var-init-zero-knowing-it-will-be-removed-from-clang
+CXXFLAGS+= -ftrivial-auto-var-init=zero \
+-enable-trivial-auto-var-init-zero-knowing-it-will-be-removed-from-clang
+.else
+.warning InitAll (zeros) requested but not support by compiler
+.endif
+.elif ${MK_INIT_ALL_PATTERN} == "yes"
+.if ${COMPILER_FEATURES:Minit-all}
+CFLAGS+= -ftrivial-auto-var-init=pattern
+CXXFLAGS+= -ftrivial-auto-var-init=pattern
+.else
+.warning InitAll (pattern) requested but not support by compiler
+.endif
+.endif
+
.if ${MK_DEBUG_FILES} != "no" && empty(DEBUG_FLAGS:M-g) && \
empty(DEBUG_FLAGS:M-gdwarf*)
CFLAGS+= ${DEBUG_FILES_CFLAGS}
Modified: head/share/mk/bsd.opts.mk
==
--- head/share/mk/bsd.opts.mk Tue Nov 10 19:09:35 2020(r367576)
+++ head/share/mk/bsd.opts.mk Tue Nov 10 19:15:13 2020(r367577)
@@ -71,6 +71,8 @@ __DEFAULT_NO_OPTIONS = \
BIND_NOW \
CCACHE_BUILD \
CTF \
+INIT_ALL_PATTERN \
+INIT_ALL_ZERO \
INSTALL_AS_USER \
PIE \
RETPOLINE \
@@ -84,6 +86,10 @@ __DEFAULT_DEPENDENT_OPTIONS = \
.include
+
+.if ${MK_INIT_ALL_PATTERN} == "yes" && ${MK_INIT_ALL_ZERO} == "yes"
+.error WITH_INIT_ALL_PATTERN and WITH_INIT_ALL_ZERO are mutually exclusive.
+.endif
#
# Supported NO_* options (if defined, MK_* will be forced to "no",
Modified: head/share/mk/bsd.prog.mk
==
--- head/share/mk/bsd.prog.mk Tue Nov 10 19:09:35 2020(r367576)
+++ head/share/mk/bsd.prog.mk Tue Nov 10 19:15:13 2020(r367577)
@@ -60,6 +60,25 @@ LDFLAGS+= -Wl,-zretpolineplt
.endif
.endif
+# Initialize stack variables on function entry
+.if ${MK_INIT_ALL_ZERO} == "yes"
+.if ${COMPILER_FEATURES:Minit-all}
+CFLAGS+= -ftrivial-auto-var-init=zero \
+-enable-trivial-auto-var-init-zero-knowing-it-will-be-removed-from-clang
+CXXFLAGS+= -ftrivial-auto-var-init=zero \
+-enable-trivial-auto-var-init-zero-knowing-it-will-be-removed-from-clang
+.else
+.warning InitAll (zeros) requested but not support by compiler
+.endif
+.elif ${MK_INIT_ALL_PATTERN} ==
svn commit: r367576 - in head/lib/csu/tests: . dynamic dynamicpie
Author: jhb
Date: Tue Nov 10 19:09:35 2020
New Revision: 367576
URL: https://svnweb.freebsd.org/changeset/base/367576
Log:
Add C startup code tests for PIE binaries.
- Force dynamic to be a non-PIE binary.
- Add a dynamicpie test which uses a PIE binary.
Reviewed by: andrew
Obtained from:CheriBSD
MFC after:2 weeks
Sponsored by: DARPA
Differential Revision:https://reviews.freebsd.org/D27127
Added:
head/lib/csu/tests/dynamicpie/
head/lib/csu/tests/dynamicpie/Makefile
- copied, changed from r367573, head/lib/csu/tests/dynamic/Makefile
Modified:
head/lib/csu/tests/Makefile
head/lib/csu/tests/dynamic/Makefile
Modified: head/lib/csu/tests/Makefile
==
--- head/lib/csu/tests/Makefile Tue Nov 10 19:07:30 2020(r367575)
+++ head/lib/csu/tests/Makefile Tue Nov 10 19:09:35 2020(r367576)
@@ -3,6 +3,7 @@
SUBDIR=dso
TESTS_SUBDIRS= dynamic
TESTS_SUBDIRS+=dynamiclib
+TESTS_SUBDIRS+=dynamicpie
TESTS_SUBDIRS+=static
SUBDIR_DEPEND_dynamiclib=dso
Modified: head/lib/csu/tests/dynamic/Makefile
==
--- head/lib/csu/tests/dynamic/Makefile Tue Nov 10 19:07:30 2020
(r367575)
+++ head/lib/csu/tests/dynamic/Makefile Tue Nov 10 19:09:35 2020
(r367576)
@@ -2,5 +2,8 @@
.PATH: ${.CURDIR:H}
+.include
+MK_PIE=no
+
.include "../Makefile.tests"
.include
Copied and modified: head/lib/csu/tests/dynamicpie/Makefile (from r367573,
head/lib/csu/tests/dynamic/Makefile)
==
--- head/lib/csu/tests/dynamic/Makefile Tue Nov 10 18:12:09 2020
(r367573, copy source)
+++ head/lib/csu/tests/dynamicpie/Makefile Tue Nov 10 19:09:35 2020
(r367576)
@@ -2,5 +2,8 @@
.PATH: ${.CURDIR:H}
+.include
+MK_PIE=yes
+
.include "../Makefile.tests"
.include
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r367575 - head/lib/csu/tests
Author: jhb
Date: Tue Nov 10 19:07:30 2020
New Revision: 367575
URL: https://svnweb.freebsd.org/changeset/base/367575
Log:
Fix dso_handle_check for PIE executables.
PIE executables use crtbeginS.o and have a non-NULL dso_handle as a
result.
Reviewed by: andrew, emaste
MFC after:2 weeks
Sponsored by: DARPA
Differential Revision:https://reviews.freebsd.org/D27126
Modified:
head/lib/csu/tests/fini_test.c
Modified: head/lib/csu/tests/fini_test.c
==
--- head/lib/csu/tests/fini_test.c Tue Nov 10 19:04:54 2020
(r367574)
+++ head/lib/csu/tests/fini_test.c Tue Nov 10 19:07:30 2020
(r367575)
@@ -141,9 +141,9 @@ dso_handle_check(void)
{
void *dso = __dso_handle;
-#ifdef DSO_LIB
+#if defined(DSO_LIB) || defined(__PIE__)
ATF_REQUIRE_MSG(dso != NULL,
- "Null __dso_handle in DSO");
+ "Null __dso_handle in DSO/PIE");
#else
ATF_REQUIRE_MSG(dso == NULL,
"Invalid __dso_handle in non-DSO");
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r367574 - head/lib/csu/common
Author: jhb
Date: Tue Nov 10 19:04:54 2020
New Revision: 367574
URL: https://svnweb.freebsd.org/changeset/base/367574
Log:
Rename __JCR_LIST__ to __JCR_END__ in crtend.c.
This is more consistent with the names used for .ctor and .dtor
symbols and better reflects __JCR_END__'s role.
Reviewed by: andrew
Obtained from:CheriBSD
MFC after:2 weeks
Sponsored by: DARPA
Differential Revision:https://reviews.freebsd.org/D27125
Modified:
head/lib/csu/common/crtend.c
Modified: head/lib/csu/common/crtend.c
==
--- head/lib/csu/common/crtend.cTue Nov 10 18:12:09 2020
(r367573)
+++ head/lib/csu/common/crtend.cTue Nov 10 19:04:54 2020
(r367574)
@@ -28,7 +28,7 @@ __FBSDID("$FreeBSD$");
typedef void (*crt_func)(void);
-static crt_func __JCR_LIST__[] __section(".jcr") __used = {
+static crt_func __JCR_END__[] __section(".jcr") __used = {
(crt_func)0
};
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r367573 - head/sys/vm
Author: jtl
Date: Tue Nov 10 18:12:09 2020
New Revision: 367573
URL: https://svnweb.freebsd.org/changeset/base/367573
Log:
When destroying a UMA zone which has a reserve (set with
uma_zone_reserve()), messages like the following appear on the console:
"Freed UMA keg (Test zone) was not empty (0 items). Lost 528 pages of
memory."
When keg_drain_domain() is draining the zone, it tries to keep the number
of items specified in the reservation. However, when we are destroying the
UMA zone, we do not need to keep those items. Therefore, when destroying a
non-secondary and non-cache zone, we should reset the keg reservation to 0
prior to draining the zone.
Reviewed by: markj
Sponsored by: Netflix
Differential Revision:https://reviews.freebsd.org/D27129
Modified:
head/sys/vm/uma_core.c
Modified: head/sys/vm/uma_core.c
==
--- head/sys/vm/uma_core.c Tue Nov 10 18:10:50 2020(r367572)
+++ head/sys/vm/uma_core.c Tue Nov 10 18:12:09 2020(r367573)
@@ -2791,6 +2791,10 @@ zone_dtor(void *arg, int size, void *udata)
rw_wlock(_rwlock);
LIST_REMOVE(zone, uz_link);
rw_wunlock(_rwlock);
+ if ((zone->uz_flags & (UMA_ZONE_SECONDARY | UMA_ZFLAG_CACHE)) == 0) {
+ keg = zone->uz_keg;
+ keg->uk_reserve = 0;
+ }
zone_reclaim(zone, M_WAITOK, true);
/*
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r367572 - head/sys/kern
Author: mjg
Date: Tue Nov 10 18:10:50 2020
New Revision: 367572
URL: https://svnweb.freebsd.org/changeset/base/367572
Log:
Allow rtprio_thread to operate on threads of any process
This in particular unbreaks rtkit.
The limitation was a leftover of previous state, to quote a
comment:
/*
* Though lwpid is unique, only current process is supported
* since there is no efficient way to look up a LWP yet.
*/
Long since then a global tid hash was introduced to remedy
the problem.
Permission checks still apply.
Submitted by: greg_unrelenting.technology (Greg V)
Differential Revision:https://reviews.freebsd.org/D27158
Modified:
head/sys/kern/kern_resource.c
Modified: head/sys/kern/kern_resource.c
==
--- head/sys/kern/kern_resource.c Tue Nov 10 18:07:13 2020
(r367571)
+++ head/sys/kern/kern_resource.c Tue Nov 10 18:10:50 2020
(r367572)
@@ -315,8 +315,7 @@ sys_rtprio_thread(struct thread *td, struct rtprio_thr
td1 = td;
PROC_LOCK(p);
} else {
- /* Only look up thread in current process */
- td1 = tdfind(uap->lwpid, curproc->p_pid);
+ td1 = tdfind(uap->lwpid, -1);
if (td1 == NULL)
return (ESRCH);
p = td1->td_proc;
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r367570 - head/share/mk
Author: bdrewery
Date: Tue Nov 10 18:05:17 2020
New Revision: 367570
URL: https://svnweb.freebsd.org/changeset/base/367570
Log:
makeman: Don't require filemon with MK_DIRDEPS_BUILD.
MFC after:2 weeks
Reviewed by: sjg, dim (tested earlier version)
Sponsored by: Dell EMC
Differential Revision:https://reviews.freebsd.org/D27134
Modified:
head/share/mk/local.meta.sys.mk
Modified: head/share/mk/local.meta.sys.mk
==
--- head/share/mk/local.meta.sys.mk Tue Nov 10 14:23:46 2020
(r367569)
+++ head/share/mk/local.meta.sys.mk Tue Nov 10 18:05:17 2020
(r367570)
@@ -178,6 +178,10 @@ LDFLAGS_LAST+= -L${STAGE_LIBDIR}
# we can use this but should not update it.
UPDATE_DEPENDFILE= NO
.endif
+# Don't require filemon for makeman.
+.if make(showconfig)
+UPDATE_DEPENDFILE= NO
+.endif
# define the list of places that contain files we are responsible for
.MAKE.META.BAILIWICK = ${SB} ${OBJROOT} ${STAGE_ROOT}
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r367569 - head/sys/contrib/openzfs/module/zfs
Author: mjg
Date: Tue Nov 10 14:23:46 2020
New Revision: 367569
URL: https://svnweb.freebsd.org/changeset/base/367569
Log:
zfs: combine zio caches if possible
This deduplicates 2 sets of caches using the same sizes.
Memory savings fluctuate a lot, one sample result is buildworld on zfs
saving ~180MB RAM in reduced page count associated with zio caches.
Modified:
head/sys/contrib/openzfs/module/zfs/zio.c
Modified: head/sys/contrib/openzfs/module/zfs/zio.c
==
--- head/sys/contrib/openzfs/module/zfs/zio.c Tue Nov 10 14:21:23 2020
(r367568)
+++ head/sys/contrib/openzfs/module/zfs/zio.c Tue Nov 10 14:23:46 2020
(r367569)
@@ -204,6 +204,19 @@ zio_init(void)
if (align != 0) {
char name[36];
+ if (cflags == data_cflags) {
+ /*
+* Resulting kmem caches would be identical.
+* Save memory by creating only one.
+*/
+ (void) snprintf(name, sizeof (name),
+ "zio_buf_comb_%lu", (ulong_t)size);
+ zio_buf_cache[c] = kmem_cache_create(name,
+ size, align, NULL, NULL, NULL, NULL, NULL,
+ cflags);
+ zio_data_buf_cache[c] = zio_buf_cache[c];
+ continue;
+ }
(void) snprintf(name, sizeof (name), "zio_buf_%lu",
(ulong_t)size);
zio_buf_cache[c] = kmem_cache_create(name, size,
@@ -234,37 +247,55 @@ zio_init(void)
void
zio_fini(void)
{
- size_t c;
- kmem_cache_t *last_cache = NULL;
- kmem_cache_t *last_data_cache = NULL;
+ size_t i, j, n;
+ kmem_cache_t *cache;
- for (c = 0; c < SPA_MAXBLOCKSIZE >> SPA_MINBLOCKSHIFT; c++) {
-#ifdef _ILP32
- /*
-* Cache size limited to 1M on 32-bit platforms until ARC
-* buffers no longer require virtual address space.
-*/
- if (((c + 1) << SPA_MINBLOCKSHIFT) > zfs_max_recordsize)
- break;
-#endif
+ n = SPA_MAXBLOCKSIZE >> SPA_MINBLOCKSHIFT;
+
#if defined(ZFS_DEBUG) && !defined(_KERNEL)
- if (zio_buf_cache_allocs[c] != zio_buf_cache_frees[c])
+ for (i = 0; i < n; i++) {
+ if (zio_buf_cache_allocs[i] != zio_buf_cache_frees[i])
(void) printf("zio_fini: [%d] %llu != %llu\n",
- (int)((c + 1) << SPA_MINBLOCKSHIFT),
- (long long unsigned)zio_buf_cache_allocs[c],
- (long long unsigned)zio_buf_cache_frees[c]);
+ (int)((i + 1) << SPA_MINBLOCKSHIFT),
+ (long long unsigned)zio_buf_cache_allocs[i],
+ (long long unsigned)zio_buf_cache_frees[i]);
+ }
#endif
- if (zio_buf_cache[c] != last_cache) {
- last_cache = zio_buf_cache[c];
- kmem_cache_destroy(zio_buf_cache[c]);
+
+ /*
+* The same kmem cache can show up multiple times in both zio_buf_cache
+* and zio_data_buf_cache. Do a wasteful but trivially correct scan to
+* sort it out.
+*/
+ for (i = 0; i < n; i++) {
+ cache = zio_buf_cache[i];
+ if (cache == NULL)
+ continue;
+ for (j = i; j < n; j++) {
+ if (cache == zio_buf_cache[j])
+ zio_buf_cache[j] = NULL;
+ if (cache == zio_data_buf_cache[j])
+ zio_data_buf_cache[j] = NULL;
}
- zio_buf_cache[c] = NULL;
+ kmem_cache_destroy(cache);
+ }
- if (zio_data_buf_cache[c] != last_data_cache) {
- last_data_cache = zio_data_buf_cache[c];
- kmem_cache_destroy(zio_data_buf_cache[c]);
+ for (i = 0; i < n; i++) {
+ cache = zio_data_buf_cache[i];
+ if (cache == NULL)
+ continue;
+ for (j = i; j < n; j++) {
+ if (cache == zio_data_buf_cache[j])
+ zio_data_buf_cache[j] = NULL;
}
- zio_data_buf_cache[c] = NULL;
+ kmem_cache_destroy(cache);
+ }
+
+ for (i = 0; i < n; i++) {
+ if (zio_buf_cache[i] != NULL)
+ panic("zio_fini: zio_buf_cache[%d] != NULL", (int)i);
+ if (zio_data_buf_cache[i] != NULL)
+ panic("zio_fini: zio_data_buf_cache[%d] !=
svn commit: r367568 - head/sys/contrib/openzfs/module/zfs
Author: mjg
Date: Tue Nov 10 14:21:23 2020
New Revision: 367568
URL: https://svnweb.freebsd.org/changeset/base/367568
Log:
zfs: g/c unused data_alloc_arena
Modified:
head/sys/contrib/openzfs/module/zfs/zio.c
Modified: head/sys/contrib/openzfs/module/zfs/zio.c
==
--- head/sys/contrib/openzfs/module/zfs/zio.c Tue Nov 10 14:17:05 2020
(r367567)
+++ head/sys/contrib/openzfs/module/zfs/zio.c Tue Nov 10 14:21:23 2020
(r367568)
@@ -144,7 +144,6 @@ void
zio_init(void)
{
size_t c;
- vmem_t *data_alloc_arena = NULL;
zio_cache = kmem_cache_create("zio_cache",
sizeof (zio_t), 0, NULL, NULL, NULL, NULL, NULL, 0);
@@ -213,8 +212,7 @@ zio_init(void)
(void) snprintf(name, sizeof (name), "zio_data_buf_%lu",
(ulong_t)size);
zio_data_buf_cache[c] = kmem_cache_create(name, size,
- align, NULL, NULL, NULL, NULL,
- data_alloc_arena, data_cflags);
+ align, NULL, NULL, NULL, NULL, NULL, data_cflags);
}
}
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r367567 - head/sbin/reboot
Author: 0mp (doc,ports committer) Date: Tue Nov 10 14:17:05 2020 New Revision: 367567 URL: https://svnweb.freebsd.org/changeset/base/367567 Log: Address a mandoc warning MFC after:3 days Modified: head/sbin/reboot/boot_i386.8 Modified: head/sbin/reboot/boot_i386.8 == --- head/sbin/reboot/boot_i386.8Tue Nov 10 13:36:07 2020 (r367566) +++ head/sbin/reboot/boot_i386.8Tue Nov 10 14:17:05 2020 (r367567) @@ -89,7 +89,7 @@ from partition of either the floppy or the hard disk. This boot may be aborted by typing any character on the keyboard at the -.Ql boot: +.Ql boot\&: prompt. At this time, the following input will be accepted: .Bl -tag -width indent ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r367555 - head/sys/dev/mlx4/mlx4_ib
Author: hselasky
Date: Tue Nov 10 12:58:25 2020
New Revision: 367555
URL: https://svnweb.freebsd.org/changeset/base/367555
Log:
Include GID type when deleting GIDs from HW table under RoCE in mlx4ib.
Refer to the Linux commit mentioned below for a more detailed description.
Linux commit:
a18177925c252da7801149abe217c05b80884798
Requested by: Isilon
MFC after:1 week
Sponsored by: Mellanox Technologies // NVIDIA Networking
Modified:
head/sys/dev/mlx4/mlx4_ib/mlx4_ib_main.c
Modified: head/sys/dev/mlx4/mlx4_ib/mlx4_ib_main.c
==
--- head/sys/dev/mlx4/mlx4_ib/mlx4_ib_main.cTue Nov 10 12:45:59 2020
(r367554)
+++ head/sys/dev/mlx4/mlx4_ib/mlx4_ib_main.cTue Nov 10 12:58:25 2020
(r367555)
@@ -371,8 +371,13 @@ static int mlx4_ib_del_gid(struct ib_device *device,
if (!gids) {
ret = -ENOMEM;
} else {
- for (i = 0; i < MLX4_MAX_PORT_GIDS; i++)
- memcpy([i].gid,
_gid_table->gids[i].gid, sizeof(union ib_gid));
+ for (i = 0; i < MLX4_MAX_PORT_GIDS; i++) {
+ memcpy([i].gid,
+ _gid_table->gids[i].gid,
+ sizeof(union ib_gid));
+ gids[i].gid_type =
+ port_gid_table->gids[i].gid_type;
+ }
}
}
spin_unlock_bh(>lock);
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r367552 - head/share/man/man7
Author: 0mp (doc,ports committer) Date: Tue Nov 10 11:32:01 2020 New Revision: 367552 URL: https://svnweb.freebsd.org/changeset/base/367552 Log: Do not document MOTIFLIB in ports(7) Perhaps it made sense in 1998 (r32836), but now it feels a bit out of place. We tend to avoid documenting non-essential ports variables in the manual page (we try to document them in the Porter's Handbook instead). MFC after:1 week Modified: head/share/man/man7/ports.7 Modified: head/share/man/man7/ports.7 == --- head/share/man/man7/ports.7 Tue Nov 10 10:40:44 2020(r367551) +++ head/share/man/man7/ports.7 Tue Nov 10 11:32:01 2020(r367552) @@ -25,7 +25,7 @@ .\" .\" $FreeBSD$ .\" -.Dd November 3, 2020 +.Dd November 10, 2020 .Dt PORTS 7 .Os .Sh NAME @@ -509,9 +509,6 @@ Normally .Xr fetch 1 . .It Va FORCE_PKG_REGISTER If set, overwrite any existing package registration on the system. -.It Va MOTIFLIB -Location of -.Pa "libXm\&." Ns Brq Pa a , Ns Pa so . .It Va INTERACTIVE If defined, only operate on a port if it requires interaction. .It Va BATCH ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r367551 - head
Author: 0mp (doc,ports committer)
Date: Tue Nov 10 10:40:44 2020
New Revision: 367551
URL: https://svnweb.freebsd.org/changeset/base/367551
Log:
Add an entry for r351863 (honoring ${name}_env in rc(8) scripts)
PR: 239692
Requested by: koobs
Modified:
head/RELNOTES
Modified: head/RELNOTES
==
--- head/RELNOTES Tue Nov 10 10:19:55 2020(r367550)
+++ head/RELNOTES Tue Nov 10 10:40:44 2020(r367551)
@@ -157,6 +157,13 @@ r352304:
ntpd is no longer by default locked in memory. rlimit memlock 32
or rlimit memlock 0 can be used to restore this behaviour.
+r351863:
+ rc.subr(8) now honors ${name}_env in all rc(8) scripts. Previously,
+ environment variables set by a user via ${name}_env were ignored
+ if the service defined a custom *_cmd variable to control the behavior
+ of the run_rc_command function, e.g., start_cmd, instead of relying on
+ the variables like command and command_args,
+
r351770,r352920,r352922,r352923:
dd(1) now supports conv=fsync, conv=fdatasync, oflag=fsync, oflag=sync,
and iflag=fullblock flags, compatible with illumos and GNU.
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r367549 - head
Author: 0mp (doc,ports committer) Date: Tue Nov 10 10:17:11 2020 New Revision: 367549 URL: https://svnweb.freebsd.org/changeset/base/367549 Log: Add an entry to RELNOTES about renaming ACPI_DMAR to IOMMU Reviewed by: br (earlier version) Differential Revision:https://reviews.freebsd.org/D26813 Modified: head/RELNOTES Modified: head/RELNOTES == --- head/RELNOTES Tue Nov 10 08:59:55 2020(r367548) +++ head/RELNOTES Tue Nov 10 10:17:11 2020(r367549) @@ -10,6 +10,11 @@ newline. Entries should be separated by a newline. Changes to this file should not be MFCed. +r366267: +Kernel option ACPI_DMAR was renamed to IOMMU. amd64's IOMMU subsystem +was split out from amd64 DMAR support and is now generic, i.e., it can +be used by all architectures. + r364896: A series of commits ending with r364896 added NFS over TLS to the kernel. This is believed to be compatible with ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
