pthread_condattr_setclock: Invalid argument

pthread_condattr_setclock: Invalid argument

[Johann Lombardi]

Hi there,

While rebasing PR #1189 (DFS engine) against latest master, I ran into
the following CI failure on appveyor/msys2:

make.exe test
./fio --minimal --thread --exitall_on_error --runtime=1s --name=nulltest --ioengine=null --rw=randrw --iodepth=2 --norandommap --random_generator=tausworthe64 --size=16T --name=verifyfstest --filename=fiotestfile.tmp --unlink=1 --rw=write --verify=crc32c --verify_state_save=0 --size=16K
pthread_condattr_setclock: Invalid argument
pthread_condattr_setclock: Invalid argument
smalloc: failed adding pool
Assertion failed: i, file smalloc.c, line 235
make: *** [Makefile:622: test] Error 3
Command exited with code 2

https://ci.appveyor.com/project/axboe/fio/builds/38248012/job/kk740eysg1bd1om9

It seems to have been introduced by a recent commit. Any thoughts?

Cheers,
Johann

Re: pthread_condattr_setclock: Invalid argument

[Johann Lombardi]

Looking at the logs, it sounds like the pthread_condattr_setclock
configure check now passes on appveyor/msys2 while it used to fail before.
Then pthread_condattr_setclock(CLOCK_MONOTONIC) fails with EINVAL on
this platform. Shall we just force pthread_condattr_setclock to no on
cygwin?

Johann

On Tue, Mar 16, 2021 at 06:24:36PM +0100, Johann Lombardi wrote:
> Hi there,
> 
> While rebasing PR #1189 (DFS engine) against latest master, I ran into
> the following CI failure on appveyor/msys2:
> 
> make.exe test
> ./fio --minimal --thread --exitall_on_error --runtime=1s --name=nulltest --ioengine=null --rw=randrw --iodepth=2 --norandommap --random_generator=tausworthe64 --size=16T --name=verifyfstest --filename=fiotestfile.tmp --unlink=1 --rw=write --verify=crc32c --verify_state_save=0 --size=16K
> pthread_condattr_setclock: Invalid argument
> pthread_condattr_setclock: Invalid argument
> smalloc: failed adding pool
> Assertion failed: i, file smalloc.c, line 235
> make: *** [Makefile:622: test] Error 3
> Command exited with code 2
> 
> https://ci.appveyor.com/project/axboe/fio/builds/38248012/job/kk740eysg1bd1om9
> 
> It seems to have been introduced by a recent commit. Any thoughts?
> 
> Cheers,
> Johann