Page MenuHomeFreeBSD

brooks (Brooks Davis)
User

Projects

User Details

User Since
Jul 1 2014, 6:02 PM (515 w, 5 d)

Recent Activity

Fri, May 17

brooks added a comment to D45236: sound: Check user-supplied size passed to SNDSTIOC_ADD_USER_DEVS*.

More comments.

Fri, May 17, 11:11 PM
brooks added a comment to D45236: sound: Check user-supplied size passed to SNDSTIOC_ADD_USER_DEVS*.

You need to decide the type of data based on the command passed before you examine it so you'll need a case statement before the tests. I'd also tend not to assign the arg/arg32 until you've decided the type. Additionally, data will be NULL for at least SNDSTIOC_REFRESH_DEVS and SNDSTIOC_FLUSH_USER_DEVS,

Fri, May 17, 10:20 PM
brooks committed rG61ad1ddc5a81: Improve broken-on-gcc checks (authored by brooks).
Improve broken-on-gcc checks
Fri, May 17, 8:52 PM
brooks closed D45230: Improve broken-on-gcc checks.
Fri, May 17, 8:51 PM
brooks committed rG1062db307002: libcxxrt: allow build with gcc13 and --no-undefined-version (authored by brooks).
libcxxrt: allow build with gcc13 and --no-undefined-version
Fri, May 17, 8:51 PM
brooks closed D45233: libcxxrt: allow build with gcc13 and --no-undefined-version.
Fri, May 17, 8:51 PM
brooks added a comment to D45233: libcxxrt: allow build with gcc13 and --no-undefined-version.
In D45233#1031909, @dim wrote:

This is fine for now, but may be can put some #ifdefs in the version maps? Or some more fine-grained gcc version checks?

Fri, May 17, 8:18 PM
brooks requested review of D45234: libcompiler_rt: gcc13 doesn't support _Float16 on arm.
Fri, May 17, 7:46 PM
brooks requested review of D45233: libcxxrt: allow build with gcc13 and --no-undefined-version.
Fri, May 17, 7:31 PM
brooks requested review of D45232: libpmc: remove tautological assert.
Fri, May 17, 7:07 PM
brooks committed rG25db7bd2fac1: libgcc_s: fix nonsense defines (authored by brooks).
libgcc_s: fix nonsense defines
Fri, May 17, 7:07 PM
brooks added a comment to D44271: bsd.lib.mk: add targets to detect symbol changes.

I noticed we mostly don't use printf in share/mk so switched to the ${:U@} trick. When doing that I noticed that bsd.symver.mk finds it's awk script with a path search so I switched to that model.

Fri, May 17, 6:54 PM
brooks committed R11:83f1b7559314: devel/llvm-devel: new snapshot (authored by brooks).
devel/llvm-devel: new snapshot
Fri, May 17, 6:29 PM
brooks requested review of D45230: Improve broken-on-gcc checks.
Fri, May 17, 5:40 PM
brooks committed rGd316de24faa7: zfs: Avoid a gcc -Wint-to-pointer-cast warning (authored by brooks).
zfs: Avoid a gcc -Wint-to-pointer-cast warning
Fri, May 17, 4:02 PM
brooks committed rG4c0dfd5959bd: arm: fail early on gcc builds (authored by brooks).
arm: fail early on gcc builds
Fri, May 17, 4:02 PM
brooks closed D45193: arm: fail early on gcc builds.
Fri, May 17, 4:02 PM
brooks closed D45089: aarch64: disable LIB32 with gcc.
Fri, May 17, 4:02 PM
brooks committed rG16d8dfde0c0f: aarch64: disable LIB32 with gcc (authored by brooks).
aarch64: disable LIB32 with gcc
Fri, May 17, 4:02 PM

Thu, May 16

brooks added inline comments to D43696: Jail descriptors.
Thu, May 16, 8:42 PM
brooks accepted D45184: makeman: extend duplicate option warning to OPT_ options.

Seems reasonable

Thu, May 16, 5:28 PM

Tue, May 14

brooks requested review of D45193: arm: fail early on gcc builds.
Tue, May 14, 12:11 AM

Mon, May 13

brooks added inline comments to D44372: Allow subset of wait4(2) functionality in Capsicum mode.
Mon, May 13, 6:47 PM

Fri, May 10

brooks committed R11:fdb8b5148cc6: devel/llvm17: openmp Fix child processes affinity (authored by brooks).
devel/llvm17: openmp Fix child processes affinity
Fri, May 10, 8:05 PM
brooks committed R11:db92b24a9b71: devel/llvm16: openmp Fix child processes affinity (authored by brooks).
devel/llvm16: openmp Fix child processes affinity
Fri, May 10, 8:05 PM
brooks accepted D45086: tftpd: silence gcc overflow warnings.

Thanks for picking this up, you went a similarly direction to what I was thinking, but I've been at a conference and traveling home so distracted.

Fri, May 10, 5:00 PM

Thu, May 9

brooks committed R11:ccca8574faa7: devel/llvm18: 18.1.5 release (authored by brooks).
devel/llvm18: 18.1.5 release
Thu, May 9, 9:59 PM
brooks added a comment to D45089: aarch64: disable LIB32 with gcc.

I've created a tracking bug https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=278879.

Thu, May 9, 5:30 PM

Wed, May 8

brooks added a comment to D45086: tftpd: silence gcc overflow warnings.
In D45086#1029238, @des wrote:

Would you like me to commandeer this?

Wed, May 8, 8:21 PM
brooks updated the diff for D45089: aarch64: disable LIB32 with gcc.
  • sort arch list
  • examine X_COMPILER_TYPE if defined and COMPILER_TYPE iff X_COMPILER_TYPE is not defined.
Wed, May 8, 3:41 PM
brooks committed rG2a32b54a5774: ofed: don't expose symbols twice (authored by brooks).
ofed: don't expose symbols twice
Wed, May 8, 2:55 PM
brooks committed rGeb5fd01b53a6: soxstack: slience 32-bit gcc warning (authored by brooks).
soxstack: slience 32-bit gcc warning
Wed, May 8, 2:55 PM
brooks closed D45090: ofed: don't expose symbols twice.
Wed, May 8, 2:55 PM
brooks closed D45087: soxstack: slience 32-bit gcc warning.
Wed, May 8, 2:54 PM
brooks closed D45126: libthr: avoid varargs in fcntl and openat interposers.
Wed, May 8, 2:50 PM
brooks committed rGec2b6b16aac4: libthr: avoid varargs in fcntl and openat interposers (authored by brooks).
libthr: avoid varargs in fcntl and openat interposers
Wed, May 8, 2:50 PM
brooks requested review of D45126: libthr: avoid varargs in fcntl and openat interposers.
Wed, May 8, 1:42 PM

Sat, May 4

brooks added inline comments to D45089: aarch64: disable LIB32 with gcc.
Sat, May 4, 12:21 AM
brooks added a comment to D45086: tftpd: silence gcc overflow warnings.

I think I've addressed everything. There might be argument for just using a pragma push/pop to disable the warnings instead of adding checks, but I've already spent more time then this code is worth.

Sat, May 4, 12:08 AM

Fri, May 3

brooks updated the diff for D45086: tftpd: silence gcc overflow warnings.

Fix strftime log message. Prefer sizeof(buffer) to bare MAXPATHLEN.

Fri, May 3, 11:55 PM
brooks added inline comments to D45086: tftpd: silence gcc overflow warnings.
Fri, May 3, 11:20 PM
brooks updated the diff for D45086: tftpd: silence gcc overflow warnings.

Graturiously check the return of sprintf because glibc's headers
require it when compiling with gcc.

Fri, May 3, 10:59 PM
brooks requested review of D45090: ofed: don't expose symbols twice.
Fri, May 3, 10:34 PM
brooks added reviewers for D45089: aarch64: disable LIB32 with gcc: karels, jrtc27.

The easiest way to address this would probably be a cross toolchain wrapper script that picks the right compiler when -m32 is on the command line, but just disabling it for now seems reasonable. (I'm trying to test make tinderbox with gcc13 and this was one of the issues I encountered.)

Fri, May 3, 10:27 PM
brooks requested review of D45089: aarch64: disable LIB32 with gcc.
Fri, May 3, 10:24 PM
brooks requested review of D45087: soxstack: slience 32-bit gcc warning.
Fri, May 3, 9:23 PM
brooks requested review of D45086: tftpd: silence gcc overflow warnings.
Fri, May 3, 9:16 PM

Thu, May 2

brooks committed rGb25ceb97eddc: Revert "Make WITHOUT_UNDEFINED_VERSION the default" (authored by brooks).
Revert "Make WITHOUT_UNDEFINED_VERSION the default"
Thu, May 2, 9:56 PM
brooks added a reverting change for rG4510f2ca9170: Make WITHOUT_UNDEFINED_VERSION the default: rGb25ceb97eddc: Revert "Make WITHOUT_UNDEFINED_VERSION the default".
Thu, May 2, 9:56 PM
brooks added a reverting change for D44216: Make WITHOUT_UNDEFINED_VERSION the default: rGb25ceb97eddc: Revert "Make WITHOUT_UNDEFINED_VERSION the default".
Thu, May 2, 9:56 PM
brooks updated the diff for D44271: bsd.lib.mk: add targets to detect symbol changes.

actually install symref.awk

Thu, May 2, 8:34 PM
brooks updated the diff for D44271: bsd.lib.mk: add targets to detect symbol changes.

Eliminate the use of printf

Thu, May 2, 8:31 PM
brooks updated the diff for D44271: bsd.lib.mk: add targets to detect symbol changes.
  • allow local override of ${SHLIB_NAME}.symbols (requested by @sjg)
  • don't hide symref files (requested by @kib)
Thu, May 2, 7:48 PM
brooks accepted D45040: Allow rfork(2) in capsicum(4) capability mode.
Thu, May 2, 6:11 PM
brooks committed rG4510f2ca9170: Make WITHOUT_UNDEFINED_VERSION the default (authored by brooks).
Make WITHOUT_UNDEFINED_VERSION the default
Thu, May 2, 5:15 PM
brooks committed rG723e60a278b1: libgcc_s: __extendxftf2 and __trunctfxf2 are amd64-only (authored by brooks).
libgcc_s: __extendxftf2 and __trunctfxf2 are amd64-only
Thu, May 2, 5:15 PM
brooks closed D44216: Make WITHOUT_UNDEFINED_VERSION the default.
Thu, May 2, 5:15 PM
brooks closed D45052: libgcc_s: __extendxftf2 and __trunctfxf2 are amd64-only.
Thu, May 2, 5:15 PM
brooks added inline comments to D45052: libgcc_s: __extendxftf2 and __trunctfxf2 are amd64-only.
Thu, May 2, 5:07 PM
brooks updated the diff for D45052: libgcc_s: __extendxftf2 and __trunctfxf2 are amd64-only.

Avoid a couple compiler invocations to create empty object files

Thu, May 2, 5:01 PM
brooks added inline comments to D45052: libgcc_s: __extendxftf2 and __trunctfxf2 are amd64-only.
Thu, May 2, 4:20 PM
brooks updated the diff for D45052: libgcc_s: __extendxftf2 and __trunctfxf2 are amd64-only.

These symbols are amd64-only

Thu, May 2, 4:20 PM
brooks retitled D45052: libgcc_s: __extendxftf2 and __trunctfxf2 are amd64-only from libgcc_s: __extendxftf2 and __trunctfxf2 are x86 symbols to libgcc_s: __extendxftf2 and __trunctfxf2 are amd64-only.
Thu, May 2, 4:19 PM

Wed, May 1

brooks accepted D39829: Allow vfork(2) in capsicum(4) capability mode.
Wed, May 1, 11:14 PM
brooks added reviewers for D44216: Make WITHOUT_UNDEFINED_VERSION the default: dim, arichardson, emaste, kib.

With D45052 committed, this passes a make CROSS_TOOLCHAIN=llvm18 -DWITHOUT_TOOLCHAIN -DMAKE_JUST_WORLDS -j40 tinderbox run. I'll kick off a full run over night in case there are issues in the toolchain bits, but I think this is ready to land. We have had several regressions since I first posted this review so getting this in is clearly valuable (admittedly, a number of which were my fault in merging patches I failed to notice had gotten out of date.)

Wed, May 1, 10:17 PM
brooks updated the diff for D44216: Make WITHOUT_UNDEFINED_VERSION the default.

Rebase, bump doc date

Wed, May 1, 10:11 PM
brooks abandoned D45034: libgcc_s: Don't build __extendxftf2 and __trunctfxf2.

This seems to be the result of botched testing. D45052 replaces it.

Wed, May 1, 10:02 PM
brooks requested review of D45052: libgcc_s: __extendxftf2 and __trunctfxf2 are amd64-only.
Wed, May 1, 10:00 PM
brooks committed rG98c8caafffb6: libgcc_s: 80-bit long double function are x86-only (authored by brooks).
libgcc_s: 80-bit long double function are x86-only
Wed, May 1, 8:50 PM
brooks closed D45028: libgcc_s: 80-bit long double function are x86-only.
Wed, May 1, 8:50 PM
brooks added inline comments to D41564: Add fchroot(2).
Wed, May 1, 8:36 PM
brooks updated subscribers of D45040: Allow rfork(2) in capsicum(4) capability mode.

If we allow rfork we should presumably allow vfork too. Have you asked @rwatson or @jonathan why they weren't allowed in the first place?

Wed, May 1, 4:51 PM

Tue, Apr 30

brooks added inline comments to D44271: bsd.lib.mk: add targets to detect symbol changes.
Tue, Apr 30, 10:39 PM
brooks added reviewers for D45034: libgcc_s: Don't build __extendxftf2 and __trunctfxf2: dim, arichardson.
Tue, Apr 30, 10:09 PM
brooks updated the diff for D45028: libgcc_s: 80-bit long double function are x86-only.

Move extendxftf2 and trunctfxf2 removal to D45034

Tue, Apr 30, 10:09 PM
brooks added a comment to D45028: libgcc_s: 80-bit long double function are x86-only.

I've moved the extendxftf2 and trunctfxf2 removal to D45034.

Tue, Apr 30, 10:08 PM
brooks updated the summary of D45028: libgcc_s: 80-bit long double function are x86-only.
Tue, Apr 30, 10:07 PM
brooks requested review of D45034: libgcc_s: Don't build __extendxftf2 and __trunctfxf2.
Tue, Apr 30, 10:06 PM
brooks updated subscribers of D45028: libgcc_s: 80-bit long double function are x86-only.

@arichardson made a big change upstream that impacted the upstream conditional for extendxftf2 and trunctfxf2 prior to llvm18's release. I might want to make that change a separate commit.

Tue, Apr 30, 8:59 PM
brooks committed rG397d8f3c3648: libsys: don't try to expose freebsd7___semctl (authored by brooks).
libsys: don't try to expose freebsd7___semctl
Tue, Apr 30, 6:06 PM
brooks requested review of D45028: libgcc_s: 80-bit long double function are x86-only.
Tue, Apr 30, 6:05 PM
brooks updated the diff for D44271: bsd.lib.mk: add targets to detect symbol changes.

Only complain about missing symref file in the current directory.
Otherwise, it's too noisy in lib/libc due to libraries in the tests tree.

Tue, Apr 30, 5:52 PM
brooks updated the diff for D44271: bsd.lib.mk: add targets to detect symbol changes.

Exclude private symbols

Tue, Apr 30, 5:28 PM
brooks added a comment to D44271: bsd.lib.mk: add targets to detect symbol changes.
In D44271#1026327, @kib wrote:
In D44271#1025583, @kib wrote:

Private versioning namespace should be filtered out, I think.

I think we want to know when that set changes even if we don't consider it part of the ABI, but I probably could be persuaded otherwise. I might want to make it optional so libraries can opt in and others opt out.

At least it should be split into two parts somehow, either textually in the same file, or putting the FBSDprivate into aux file at all.

Tue, Apr 30, 5:28 PM
brooks updated the diff for D44271: bsd.lib.mk: add targets to detect symbol changes.

Explain why we can't use readelf -D and add symbol visibility.

Tue, Apr 30, 5:11 PM

Mon, Apr 29

brooks added a comment to D44271: bsd.lib.mk: add targets to detect symbol changes.
In D44271#1025583, @kib wrote:

Private versioning namespace should be filtered out, I think.

Mon, Apr 29, 9:28 PM
brooks updated the diff for D44271: bsd.lib.mk: add targets to detect symbol changes.

Review suggestions from @sjg

Mon, Apr 29, 9:25 PM

Sat, Apr 27

brooks updated the diff for D44271: bsd.lib.mk: add targets to detect symbol changes.

typo fix

Sat, Apr 27, 12:06 AM
brooks updated the diff for D44271: bsd.lib.mk: add targets to detect symbol changes.
  • Rewrite with readelf and an awk script. Weak vs global and symbol type are now tracked. It's also a bit less grotty as readelf implementations are more similar.
  • Make the symref files dot-files so they clutter the tree less.
Sat, Apr 27, 12:04 AM

Fri, Apr 26

brooks abandoned D44275: Restore ${OBJDUMP}.

I'm going to switch to using readelf

Fri, Apr 26, 11:52 PM
brooks committed R11:95ad99efb39c: devel/llvm-devel: new snapshot (authored by brooks).
devel/llvm-devel: new snapshot
Fri, Apr 26, 8:51 PM
brooks committed rGd5ad60b75f47: libgcc_s: expose __divmoddi4 on i386 (authored by brooks).
libgcc_s: expose __divmoddi4 on i386
Fri, Apr 26, 7:27 PM

Thu, Apr 25

brooks added a reverting change for rG22ca6db50f4e: config.mk: Add MK_VIMAGE knob: rG68cbb072fd1c: Revert "config.mk: Add MK_VIMAGE knob".
Thu, Apr 25, 9:45 PM
brooks added a reverting change for D39636: config.mk: Add MK_VIMAGE knob: rG68cbb072fd1c: Revert "config.mk: Add MK_VIMAGE knob".
Thu, Apr 25, 9:44 PM
brooks committed rG68cbb072fd1c: Revert "config.mk: Add MK_VIMAGE knob" (authored by brooks).
Revert "config.mk: Add MK_VIMAGE knob"
Thu, Apr 25, 9:44 PM

Wed, Apr 24

brooks committed rG78101d437a92: syscalls.master: correct return type of {read,write}v (authored by brooks).
syscalls.master: correct return type of {read,write}v
Wed, Apr 24, 7:49 PM
brooks closed D44930: syscalls.master: correct return type of {read,write}v.
Wed, Apr 24, 7:49 PM
brooks requested review of D44930: syscalls.master: correct return type of {read,write}v.
Wed, Apr 24, 7:04 PM

Tue, Apr 23

brooks committed R11:349c87302036: devel/llvm18: 18.1.4 release (authored by brooks).
devel/llvm18: 18.1.4 release
Tue, Apr 23, 10:15 PM
brooks committed R11:3645478a263b: devel/llvm*: Remove unnecessary C/CFLAGS definitions (authored by diizzy).
devel/llvm*: Remove unnecessary C/CFLAGS definitions
Tue, Apr 23, 10:14 PM