Summary: | net-proxy/haproxy-{2.4.25,2.6.16,2.8.7,2.9.6} stable request | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Christian Ruppert (idl0r) <idl0r> |
Component: | Stabilization | Assignee: | Christian Ruppert (idl0r) <idl0r> |
Status: | RESOLVED FIXED | ||
Severity: | normal | CC: | bertrand, matoro_gentoo |
Priority: | Normal | Keywords: | CC-ARCHES, STABLEREQ |
Version: | unspecified | Flags: | nattka:
sanity-check+
|
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: |
net-proxy/haproxy-2.6.16
net-proxy/haproxy-2.8.7
net-proxy/haproxy-2.9.6
|
Runtime testing required: | No |
Bug Depends on: | |||
Bug Blocks: | 900737 |
Description
Christian Ruppert (idl0r)
2024-04-08 08:40:18 UTC
ppc: 2.4.25 fails, everything else is good, can I just drop keywords from 2.4? (In reply to matoro from comment #1) > ppc: 2.4.25 fails, everything else is good, can I just drop keywords from > 2.4? That's fine to me :) This also doesn't work anywhere on 32-bit arm unfortunately due to unaligned access (SIGBUS), OK to drop as well? Note that this is actually undefined behavior on all platforms, it's just that it only results in a SIGBUS crash on sparc and 32-bit arm. I'll see if I can get a UBSAN report for it. The bug has been referenced in the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=097e142c6f471684cc5b224b0fc2be2565445067 commit 097e142c6f471684cc5b224b0fc2be2565445067 Author: Matoro Mahri <matoro_gentoo@matoro.tk> AuthorDate: 2024-04-29 17:01:58 +0000 Commit: Ionen Wolkens <ionen@gentoo.org> CommitDate: 2024-04-30 02:28:31 +0000 net-proxy/haproxy: unkeyword all for arm Bug: https://bugs.gentoo.org/928927 Signed-off-by: Matoro Mahri <matoro_gentoo@matoro.tk> Signed-off-by: Ionen Wolkens <ionen@gentoo.org> net-proxy/haproxy/haproxy-2.4.23.ebuild | 2 +- net-proxy/haproxy/haproxy-2.4.25.ebuild | 2 +- net-proxy/haproxy/haproxy-2.4.26.ebuild | 2 +- net-proxy/haproxy/haproxy-2.4.9999.ebuild | 2 +- net-proxy/haproxy/haproxy-2.6.14.ebuild | 2 +- net-proxy/haproxy/haproxy-2.6.16.ebuild | 2 +- net-proxy/haproxy/haproxy-2.6.17.ebuild | 2 +- net-proxy/haproxy/haproxy-2.6.9999.ebuild | 2 +- net-proxy/haproxy/haproxy-2.7.11.ebuild | 2 +- net-proxy/haproxy/haproxy-2.7.9.ebuild | 2 +- net-proxy/haproxy/haproxy-2.7.9999.ebuild | 2 +- net-proxy/haproxy/haproxy-2.8.7.ebuild | 2 +- net-proxy/haproxy/haproxy-2.8.9.ebuild | 2 +- net-proxy/haproxy/haproxy-2.8.9999.ebuild | 2 +- net-proxy/haproxy/haproxy-2.9.6.ebuild | 2 +- net-proxy/haproxy/haproxy-2.9.7.ebuild | 2 +- net-proxy/haproxy/haproxy-2.9.9999.ebuild | 2 +- net-proxy/haproxy/haproxy-9999.ebuild | 2 +- 18 files changed, 18 insertions(+), 18 deletions(-) https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=4067c6dadfdbad2c70fc4c7fa8646467c748a4a2 commit 4067c6dadfdbad2c70fc4c7fa8646467c748a4a2 Author: Matoro Mahri <matoro_gentoo@matoro.tk> AuthorDate: 2024-04-29 14:59:11 +0000 Commit: Ionen Wolkens <ionen@gentoo.org> CommitDate: 2024-04-30 02:28:29 +0000 net-proxy/haproxy: unkeyword SLOT=2.4 for ppc Bug: https://bugs.gentoo.org/928927 Signed-off-by: Matoro Mahri <matoro_gentoo@matoro.tk> Signed-off-by: Ionen Wolkens <ionen@gentoo.org> net-proxy/haproxy/haproxy-2.4.23.ebuild | 2 +- net-proxy/haproxy/haproxy-2.4.25.ebuild | 2 +- net-proxy/haproxy/haproxy-2.4.26.ebuild | 2 +- net-proxy/haproxy/haproxy-2.4.9999.ebuild | 2 +- 4 files changed, 4 insertions(+), 4 deletions(-) arm64 done (In reply to matoro from comment #3) > This also doesn't work anywhere on 32-bit arm unfortunately due to unaligned > access (SIGBUS), OK to drop as well? > > Note that this is actually undefined behavior on all platforms, it's just > that it only results in a SIGBUS crash on sparc and 32-bit arm. I'll see if > I can get a UBSAN report for it. 2.4? That's fine. Or all? That's probably not fine :D (In reply to Christian Ruppert (idl0r) from comment #6) > (In reply to matoro from comment #3) > > This also doesn't work anywhere on 32-bit arm unfortunately due to unaligned > > access (SIGBUS), OK to drop as well? > > > > Note that this is actually undefined behavior on all platforms, it's just > > that it only results in a SIGBUS crash on sparc and 32-bit arm. I'll see if > > I can get a UBSAN report for it. > > 2.4? That's fine. > Or all? That's probably not fine :D Unfortunately the tests are revealing that this does not work (and likely never worked, since I tested all versions in tree) on 32-bit arm. I think this is a less important target to keep and was probably just keyworded without real testing. Same goes for ppc to be honest but that should be fine to keep since it works. And we still have arm64 which realistically if anyone is running this on ARM hardware this is going to be what they use. ppc done all arches done |