Colin Ian King
Colin Ian King
is this change going to be merged at some point?
Apologies for asking again; is this change going to be merged at some point?
I've not seen this before. The bad-altstack stressor tests various scenarios where an illegal alternated stack is used for signal handling, so it may trigger some weird faults. However, I've...
To add debug, please try: ``` diff --git a/stress-bad-altstack.c b/stress-bad-altstack.c index a49cbc2d..9fd62773 100644 --- a/stress-bad-altstack.c +++ b/stress-bad-altstack.c @@ -285,6 +285,8 @@ again: stress_mwc_reseed(); rnd = stress_mwc32() % 11; + pr_inf("rnd...
I can't reproduce the issue with Debian and 4.15 in a VM. Which kernel are you using?
Was able to reproduce this with Debian unstable with 5.19 kernel. I believe the munmap() symbol for the system call is being lazily fixed up late when we reach the...
Take your time. I won't close this issue until I know it's all OK, no sweat.
Yes, checking for low swap when there is no swap is incorrect, I've now added a check to see if swap is enabled or not for this check in commit...
Seems that the issue is stress-ng-icache: mprotect failed: errno=13 (Permission denied) occurring because mprotect call on the text region with PROT_READ | PROT_EXEC is not allowed.
I normally test this in QEMU with POWER8/POWER9 emulation. I've not seen this occur on emulated systems. I'm not sure why enabling PROT_READ | PROT_EXEC on a process' text page...