LGTM
- Queries
- All Stories
- Search
- Advanced Search
- Transactions
- Transaction Logs
Feed Advanced Search
Advanced Search
Advanced Search
Aug 22 2018
Aug 22 2018
That being said, I see no reason not to commit this version now.
It might be worth spending the time to see how the performance differs. If it's not extreme, then just dropping skein_block_asm.s entirely would make sense given we're talking about an SHA3 candidate that didn't win.
Seems fine to me.
Aug 21 2018
Aug 21 2018
Limit LLVM builds to linking a single binary at a time.
brooks committed rS338149: Revert r337978: Rework rtld's TLS Variant I implementation to match r326794.
Revert r337978: Rework rtld's TLS Variant I implementation to match r326794
Aug 17 2018
Aug 17 2018
Not my area of expertise, but looks valid to me.
brooks committed rP477441: Switch from ninja to Makefiles to work around bugs in generated ninja code..
Switch from ninja to Makefiles to work around bugs in generated ninja code.
Rework rtld's TLS Variant I implementation to match r326794
Aug 16 2018
Aug 16 2018
Update a newer snapshow to support code generated by the latest
Update to a new snapshot. LLVM trunk is now 8.0.
Add textproc/py-recommonmark:
brooks accepted D16745: Use MACHINE_CPUARCH/TARGET_CPUARCH in preference to MACHINE/TARGET and minor cleanup.
LGTM
MFC r337727 (rewritten due to unmerged API change in 11+):
Aug 15 2018
Aug 15 2018
- Clarify that explicit approval can take any form that makes sense.
brooks retitled D16730: Committers Guide: Add a section encouraging pre-commit review. from Committers Guild: Add a section encouraging pre-commit review. to Committers Guide: Add a section encouraging pre-commit review..
- Fix unicode apostrophes.
- Convey that reviewers can't usually be any random committer.
- Clarify an edit-o. @imp's large comment to be addressed seperately.
brooks added inline comments to D16730: Committers Guide: Add a section encouraging pre-commit review..
For reference: https://llvm.org/docs/DeveloperPolicy.html#code-reviews
brooks retitled D16730: Committers Guide: Add a section encouraging pre-commit review. from Add a section encouraging pre-commit review. to Committers Guild: Add a section encouraging pre-commit review..
Aug 14 2018
Aug 14 2018
Add a port of LLVM 7.0.0.rc1 based on devel/llvm60.
Aug 13 2018
Aug 13 2018
Garbage collect LLVM and Clang 3.4.
At long last, garbage collect LLVM and Clang 3.3.
Copy out from kernel to data, not the other way around.
Update clang/llvm version to 6.0 (used on FreeBSD 10)
Aug 9 2018
Aug 9 2018
Remove (non-default) support for building with LLVM 3.8 and 3.9.
Aug 8 2018
Aug 8 2018
Terminate filter_create_ext() args with NULL, not 0.
Remove obsolete Clang 3.8 support from bazel.
Depend on the real port, not the removed METAPORT on 10.x.
Remove broken and expired port which also depends on clang38 (on 10).
brooks committed rP476656: Deprecate gps. It is unmaintained and depends on an obsolete LLVM version..
Deprecate gps. It is unmaintained and depends on an obsolete LLVM version.
Remove METAPORTs for obsolete LLVM 3.8.
Jul 30 2018
Jul 30 2018
brooks added reviewers for D16510: Rework rtld's TLS Variant I implementation to match r326794: jhibbits, andrew, br.
Jul 27 2018
Jul 27 2018
Jul 19 2018
Jul 19 2018
LGTM. I'd commit the netbsd ones separately and forward the commit some appropriate netbsd list (probably tech-kern).
Jul 16 2018
Jul 16 2018
Jul 13 2018
Jul 13 2018
Round down the location of execpathp to slightly improve copyout speed.
Jul 10 2018
Jul 10 2018
Looks good to me. Some may complain about the Xr references to the removed numa_getaffinity(2) and numa_setaffinity(2) in numa(4), but I think it's fine given that they are removed from the SEE ALSO.
Get rid of netbsd_lchown and netbsd_msync syscall entries.
Jul 9 2018
Jul 9 2018
LGTM
Jul 6 2018
Jul 6 2018
One more 32-bit fix for r335979.
Correct breakage on 32-bit platforms from r335979.
Jul 5 2018
Jul 5 2018
Work around lame warnings in ancient gcc on 32-bit platforms.
It looks like usr.bin/Makefile also needs an update to unhook the removed numactl.
In D16141#342277, @emaste wrote:Although we still need to make sure that everything still works with this change.
Get rid of netbsd_lchown and netbsd_msync syscall entries.
Make struct xinpcb and friends word-size independent.
readelf -a libc.so.7.full | grep netbsd 905: 000000000008cf10 11 FUNC GLOBAL DEFAULT 12 netbsd_msync@FBSD_1.0 (2) 3141: 000000000008cf10 11 FUNC GLOBAL DEFAULT 12 netbsd_lchown@FBSD_1.0 (2) 165: 000000000008c0a0 17 FUNC LOCAL DEFAULT 12 __sys_netbsd_msync 166: 000000000008c0a0 17 FUNC LOCAL DEFAULT 12 _netbsd_msync 167: 000000000008c0a0 17 FUNC LOCAL DEFAULT 12 netbsd_msync 168: 000000000008c0e0 17 FUNC LOCAL DEFAULT 12 __sys_netbsd_lchown 169: 000000000008c0e0 17 FUNC LOCAL DEFAULT 12 _netbsd_lchown 170: 000000000008c0e0 17 FUNC LOCAL DEFAULT 12 netbsd_lchown 2040: 000000000008cf10 11 FUNC GLOBAL DEFAULT 12 netbsd_lchown 2041: 000000000008cf10 11 FUNC GLOBAL DEFAULT 12 netbsd_msync
- Export under the correct versions.
- Spell enosys consistantly.
Jul 4 2018
Jul 4 2018
I don't understand why there isn't an exported symbol. I've got a:
Jun 29 2018
Jun 29 2018
$readelf -a ~/obj/home/bed22/git/freebsd/amd64.amd64/lib/libc/libc.so.7 | grep netbsd_lchown 168: 000000000008c0e0 17 FUNC LOCAL DEFAULT 12 __sys_netbsd_lchown 169: 000000000008c0e0 17 FUNC LOCAL DEFAULT 12 _netbsd_lchown 170: 000000000008c0e0 17 FUNC LOCAL DEFAULT 12 netbsd_lchown
- Fix a botched kptr_t change.
- Make kvaddr_t unsigned to match the defintion in kvm.h and remove the
Correct link metadata created when installing with -DNO_ROOT.
Jun 28 2018
Jun 28 2018
- Rebase.
In D15814#338187, @kib wrote:In D15814#338185, @brooks wrote:In D15814#337257, @kib wrote:You can create symbols which are exported but not linkable, since they do not provide a default version. Such symbol can be only created by asm '@' syntax, it should be removed from the version map. Also I do not see a sense in leaving the private symbols around.
Is there an example of this?
This is a duplicate of D15936.
Jun 27 2018
Jun 27 2018
Update to LLVM 6.0.1 release.
In D16019#339574, @emaste wrote:I wonder if we also want to restore ibcs2's int return value?
Would that be implementing a ibcs2-specific wrapper? I suspect it's not worth it.
Jun 26 2018
Jun 26 2018
Looks good to me. I wonder if we also want to restore ibcs2's int return value?
In D15687#339170, @emaste wrote:For reference, BSDL __stack_chk_fail in Android https://android.googlesource.com/platform/bionic/+/ics-mr1-release/libc/bionic/ssp.c and NetBSD http://cvsweb.netbsd.org/bsdweb.cgi/src/lib/libc/misc/stack_protector.c?annotate=1.9&only_with_tag=MAIN.
That makes sense. If in fact nothing ever linked to it then we should just relegate it to the compat11 port and kill it along with gcc.
In terms of exposed symbols (see objdump -T), libssp definitely is a FORTIFY_SOURCE implementation. I've not looked at the code to avoid contamination in case we need a cleanroom one.
Jun 25 2018
Jun 25 2018
Normalize the g(eom,cache,part,...) build.
brooks added a comment to D15953: Kernel environments: use any and all provided environments, evict hintmode and envmode.
I don't see anything that's going to cause us obvious problems on BERI.
Fix a stack overflow in mount_smbfs when hostname is too long.
Jun 22 2018
Jun 22 2018
In D15814#337257, @kib wrote:You can create symbols which are exported but not linkable, since they do not provide a default version. Such symbol can be only created by asm '@' syntax, it should be removed from the version map. Also I do not see a sense in leaving the private symbols around.
Jun 20 2018
Jun 20 2018
In D15936#337219, @jpaetzel wrote:I'm not sure what the consequences of returning an error here are versus silently truncating. Before this change you'd have hostname "longerthan16chars.local" and it would get passed through as "longerthan16cha" but now mount_smbfs would fail with an error? If that assertion is true maybe it would be better just to copy and truncate at 16 than to return an error.
- Per discussion in the PR, truncate the host name to 15 bytes.