Re: [PATCH] fs: don't block i_writecount during exec

[Josef Bacik <josef@toxicpanda.com>]

On Thu, Jun 06, 2024 at 04:37:49PM +0100, Mark Brown wrote:
> On Thu, Jun 06, 2024 at 01:45:05PM +0100, Aishwarya TCV wrote:
> > On 31/05/2024 14:01, Christian Brauner wrote:
> 
> > > Back in 2021 we already discussed removing deny_write_access() for
> > > 
> > > executables. Back then I was hesistant because I thought that this might
> > > 
> > > cause issues in userspace. But even back then I had started taking some
> > > 
> > > notes on what could potentially depend on this and I didn't come up with
> > > 
> > > a lot so I've changed my mind and I would like to try this.
> 
> > LTP test "execve04" is failing when run against
> > next-master(next-20240606) kernel with Arm64 on JUNO in our CI.
> 
> It's also causing the LTP creat07 test to fail with basically the same
> bisection (I started from next/pending-fixes rather than the -rc so the
> initial phases were different):
> 
> tst_test.c:1690: TINFO: LTP version: 20230929
> tst_test.c:1574: TINFO: Timeout per run is 0h 01m 30s
> creat07.c:37: TFAIL: creat() succeeded unexpectedly
> Test timeouted, sending SIGKILL!
> tst_test.c:1622: TINFO: Killed the leftover descendant processes
> tst_test.c:1628: TINFO: If you are running on slow machine, try exporting LTP_TIMEOUT_MUL > 1
> tst_test.c:1630: TBROK: Test killed! (timeout?)
> 
> The code in the testcase is below:
> 
> static void verify_creat(void)
> {
>         pid_t pid;
> 
>         pid = SAFE_FORK();
>         if (pid == 0) {
>                 SAFE_EXECL(TEST_APP, TEST_APP, NULL);
>                 exit(1);
>         }
> 
>         TST_CHECKPOINT_WAIT(0);
> 
>         TEST(creat(TEST_APP, O_WRONLY));
> 
>         if (TST_RET != -1) {
>                 tst_res(TFAIL, "creat() succeeded unexpectedly");
>                 return;
>         }
> 
>         if (TST_ERR == ETXTBSY)
>                 tst_res(TPASS, "creat() received EXTBSY");
>         else
>                 tst_res(TFAIL | TTERRNO, "creat() failed unexpectedly");
> 

These tests will have to be updated, as this patch removes that behavior.
Thanks,

Josef