Re: [PATCH V7 9/9] iommufd: map file selftest

[Nicolin Chen <nicolinc@nvidia.com>]

On Fri, Oct 25, 2024 at 01:58:57PM -0400, Steven Sistare wrote:
> On 10/25/2024 1:04 PM, Nicolin Chen wrote:
> > On Fri, Oct 25, 2024 at 06:11:59AM -0700, Steve Sistare wrote:
> > > 
> > > Add test cases to exercise IOMMU_IOAS_MAP_FILE.
> > > 
> > > Signed-off-by: Steve Sistare <steven.sistare@oracle.com>
> > > Reviewed-by: Nicolin Chen <nicolinc@nvidia.com>

Reran on Jason's linus tree:
...
# PASSED: 204 / 204 tests passed.
# Totals: pass:194 fail:0 xfail:0 xpass:0 skip:10 error:0
...
# PASSED: 8 / 8 tests passed.
# Totals: pass:8 fail:0 xfail:0 xpass:0 skip:0 error:0

Tested-by: Nicolin Chen <nicolinc@nvidia.com>

> > The build passes, but I now see a few failing tests. Any idea?
> 
> Strange.  All tests pass for me, with patches applied to base commit
> e2d8fe9148b79ed1cbf0663edc988db7769173dc
> 
>   # modprobe iomufd
>   # iommufd
>   ...
>   Totals: pass:204 fail:0 xfail:0 xpass:0 skip:0 error:0
> 
> However, I also needed this patch from Yi Liu to make it work:
> [2] https://github.com/yiliu1765/iommufd/commit/ff0772d155c1ae2841e7b7ff0d7ca889271b81b2
> 
> Without that, most tests fail, even before my series.

I run on ARM64, so I don't have that for it fixing an x86 issue.

> I will try to reproduce your failures.
> What is your base commit?

I was using the for-next tree (rc1) from:
git git://git.kernel.org/pub/scm/linux/kernel/git/jgg/iommufd.git

And it was failing, so I guess rc1 is missing something for memfd.
So, it should be a false alarm.

> What is your value of /sys/kernel/mm/transparent_hugepage/shmem_enabled ?
>   (both always and never pass for me)

$ cat /sys/kernel/mm/transparent_hugepage/shmem_enabled
always within_size advise [never] deny force

And I retried with "always" that works too with the linus tree.

Thanks
Nicolin