On Tue, Sep 14, 2021 at 08:54:26AM -0700, Linus Torvalds wrote:
On Mon, Sep 13, 2021 at 7:43 PM Michael Schmitz <schmitzmic@xxxxxxxxx> wrote:
having run stress tests on a kernel with Al Viro's signal handling fixes
applied for the past two days with no further errors, I am now quite
confident that the format error I saw in resume_userspace() and a bus
error in setup_frame() were caused by multiple pending signals, and the
resulting stack mangling that Al's patches fix.
Sounds good. Except you say "Al Viro's signal handling fixes" without
giving any context, so I have no idea which patches you are referring
to.
Regardless, if there's some fixes queue that also removes set_fs()
from m68k, I'm willing to take it for 5.15 just to have another
architecture down and done with.
And even if the set_fs() patches might still be under discussion, the
format errors have been going on for some time, so the signal fixes -
wherever they are - seem to be no-brainers if they fix those.
vfs.git #{untested,fixes}.m68k, series posted to m68k folks back in July.
3 commits in there, and I've some local followup cleanups (ifdefs in m68k/signal.c
are bloody confused), but the latter got stalled by ptrace and signal review and
documentation writing...
See https://lkml.kernel.org/lkml/YP2dTQPm1wGPWFgD@xxxxxxxxxxxxxxxxxxxxx/T/
for original thread.
