Message ID | 20240123160507.31678-1-mdoucha@suse.cz |
---|---|
State | Accepted |
Headers | show |
Series | fallocate06: Increase test loop device size to 1GB | expand |
Hi Martin, > Btrfs on newer kernels may have issues with running out of space > for metadata in the fill FS+CoW subtest if the test device is too > small. Increase minimal device size to avoid the corner case. I hope it's not a btrfs bug :). Given you tested 512 MB and it still failed on Btrfs I agree with 1GB. Reviewed-by: Petr Vorel <pvorel@suse.cz> Kind regards, Petr
Hi! > diff --git a/testcases/kernel/syscalls/fallocate/fallocate06.c b/testcases/kernel/syscalls/fallocate/fallocate06.c > index 124fb7eae..83b4d88e6 100644 > --- a/testcases/kernel/syscalls/fallocate/fallocate06.c > +++ b/testcases/kernel/syscalls/fallocate/fallocate06.c > @@ -260,6 +260,7 @@ static struct tst_test test = { > .test = run, > .tcnt = ARRAY_SIZE(testcase_list), > .needs_root = 1, > + .dev_min_size = 1024, > .mount_device = 1, > .mntpoint = MNTPOINT, > .all_filesystems = 1, Don't we need to add at least .max_runtime = 120 here as well? Looks like I'm getting quite close to the default 30s timeout on a machine with a old good rotating harddisc. I did measure between 10 and 20s for the case that we fill the 1GB filesystem.
On 25. 01. 24 15:10, Cyril Hrubis wrote: > Hi! >> diff --git a/testcases/kernel/syscalls/fallocate/fallocate06.c b/testcases/kernel/syscalls/fallocate/fallocate06.c >> index 124fb7eae..83b4d88e6 100644 >> --- a/testcases/kernel/syscalls/fallocate/fallocate06.c >> +++ b/testcases/kernel/syscalls/fallocate/fallocate06.c >> @@ -260,6 +260,7 @@ static struct tst_test test = { >> .test = run, >> .tcnt = ARRAY_SIZE(testcase_list), >> .needs_root = 1, >> + .dev_min_size = 1024, >> .mount_device = 1, >> .mntpoint = MNTPOINT, >> .all_filesystems = 1, > > Don't we need to add at least .max_runtime = 120 here as well? > > Looks like I'm getting quite close to the default 30s timeout on a > machine with a old good rotating harddisc. I did measure between 10 and > 20s for the case that we fill the 1GB filesystem. OK, let's add that just in case. Should I resubmit?
Hi!
> OK, let's add that just in case. Should I resubmit?
No need, I will add that and push, thanks.
diff --git a/testcases/kernel/syscalls/fallocate/fallocate06.c b/testcases/kernel/syscalls/fallocate/fallocate06.c index 124fb7eae..83b4d88e6 100644 --- a/testcases/kernel/syscalls/fallocate/fallocate06.c +++ b/testcases/kernel/syscalls/fallocate/fallocate06.c @@ -260,6 +260,7 @@ static struct tst_test test = { .test = run, .tcnt = ARRAY_SIZE(testcase_list), .needs_root = 1, + .dev_min_size = 1024, .mount_device = 1, .mntpoint = MNTPOINT, .all_filesystems = 1,
Btrfs on newer kernels may have issues with running out of space for metadata in the fill FS+CoW subtest if the test device is too small. Increase minimal device size to avoid the corner case. Signed-off-by: Martin Doucha <mdoucha@suse.cz> --- Tested on affected SLE beta kernel. testcases/kernel/syscalls/fallocate/fallocate06.c | 1 + 1 file changed, 1 insertion(+)