Summary: | app-arch/cpio-2.15: emerge hangs on conftest when building cpio | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Robert Marmorstein <robert> |
Component: | Current packages | Assignee: | Gentoo's Team for Core System packages <base-system> |
Status: | RESOLVED FIXED | ||
Severity: | normal | CC: | robert, sam, sandbox |
Priority: | Normal | ||
Version: | unspecified | ||
Hardware: | AMD64 | ||
OS: | Linux | ||
See Also: | https://bugs.gentoo.org/show_bug.cgi?id=447970 | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- |
Description
Robert Marmorstein
2024-01-21 23:35:14 UTC
Killing conftest seems to be a workaround for this -- but not a good one. When building this with: ebuild cpio-2.16.ebuild compile it seems to hang on: checking whether getcwd handles long file names properly... Could you try strace the test when it's stuck? I imagine this is the same as bug 447970. (In reply to Sam James from comment #3) > Could you try strace the test when it's stuck? I imagine this is the same as > bug 447970. It does look similar (funny, I just found that bug myself and was comparing). Here's the tail end of an strace: epoll_ctl(3, EPOLL_CTL_ADD, 12, {events=EPOLLIN, data={u32=12, u64=140338056396812}}) = 0 epoll_ctl(3, EPOLL_CTL_DEL, 10, 0x7ffd11ebb49c) = 0 read(10, "checking whether getcwd handles "..., 4096) = 60 write(11, "checking whether getcwd handles "..., 60checking whether getcwd handles long file names properly... ) = 60 write(13, "checking whether getcwd handles "..., 60) = 60 read(10, 0x565264a6d450, 4096) = -1 EAGAIN (Resource temporarily unavailable) epoll_ctl(3, EPOLL_CTL_ADD, 10, {events=EPOLLIN, data={u32=10, u64=140338056396810}}) = 0 epoll_wait(3, [{events=EPOLLIN, data={u32=12, u64=140338056396812}}], 4, -1) = 1 epoll_wait(3, [{events=EPOLLIN, data={u32=12, u64=140338056396812}}], 4, 0) = 1 epoll_ctl(3, EPOLL_CTL_DEL, 12, 0x7ffd11ebb49c) = 0 read(12, "checking whether getcwd handles "..., 4096) = 60 write(14, "checking whether getcwd handles "..., 60) = 60 read(12, 0x565264a6d450, 4096) = -1 EAGAIN (Resource temporarily unavailable) epoll_ctl(3, EPOLL_CTL_ADD, 12, {events=EPOLLIN, data={u32=12, u64=140338056396812}}) = 0 epoll_wait(3, If it is the same as #44790, the behavior is clearly worse for my case. In 44790, the comments suggest the slowdown never takes more than a minute or two. But I let this run for over 24 hours. I've asked Arsen to have a look, but I think in the meantime, we should set the cache var for it in profiles. This is a huge waste of CPU time. (In reply to Robert Marmorstein from comment #5) > If it is the same as #44790, the behavior is clearly worse for my case. In > 44790, the comments suggest the slowdown never takes more than a minute or > two. But I let this run for over 24 hours. To confirm if it's roughly the same, could you tell me if FEATURES="-sandbox -usersandbox" emerge -v1 cpio is fast or not (does it hang on that test)? FWIW, the speed of the hangs reported in that bug vary a lot. > To confirm if it's roughly the same, could you tell me if FEATURES="-sandbox
> -usersandbox" emerge -v1 cpio is fast or not (does it hang on that test)?
>
Tried this, but it still hangs. So maybe not the same bug after all.
I'm going to run this with the "time" command in a "screen" session to see if it ever finishes. I'll give it a couple days and see what happens... The bug has been closed via the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=39e7dae5ae53f438bb94c464b05aba169938d9fb commit 39e7dae5ae53f438bb94c464b05aba169938d9fb Author: Sam James <sam@gentoo.org> AuthorDate: 2024-01-22 16:11:52 +0000 Commit: Sam James <sam@gentoo.org> CommitDate: 2024-01-23 15:13:47 +0000 profiles: workaround sandbox bug with getcwd() configure test (gl_cv_func_getcwd_path_max) Workaround for sandbox bug which causes this gnulib configure test to take many real hours on slower machines, and certainly a huge amount of CPU hours on others. Spoof the same result as configure gets on a modern glibc & musl system for now. Bug: https://bugs.gentoo.org/447970 Closes: https://bugs.gentoo.org/922652 Signed-off-by: Sam James <sam@gentoo.org> profiles/default/linux/make.defaults | 9 ++++++++- profiles/features/musl/make.defaults | 7 +++++++ 2 files changed, 15 insertions(+), 1 deletion(-) |