[PATCH] linux: Remove shmmax check from tst-sysvshm-linux

[PATCH] linux: Remove shmmax check from tst-sysvshm-linux

[Adhemerval Zanella]

The shmmax expected value is tricky to check because kernel clamps it
to INT_MAX in two cases:

  1. Compat symbols with IPC_64, i.e, 32-bit binaries running on 64-bit
     kernels.

  2. Default symbol without IPC_64 (defined as IPC_OLD within Linux) and
     glibc always use IPC_64 for 32-bit ABIs (to support 64-bit time_t).
     It means that 32-bit binaries running on 32-bit kernels will not see
     shmmax being clamped.

And finding out whether the compat symbol is used would require checking
the underlying kernel against the current ABI.  The shmall and shmmni
already provided enough coverage.

Checked on x86_64-linux-gnu and i686-linux-gnu.  It should fix the
tst-sysvshm-linux failures on 32-bit kernels.

I will commit this shortly if noone opposes it.
---
 sysdeps/unix/sysv/linux/tst-sysvshm-linux.c | 25 +++++++++++----------
 1 file changed, 13 insertions(+), 12 deletions(-)

diff --git a/sysdeps/unix/sysv/linux/tst-sysvshm-linux.c b/sysdeps/unix/sysv/linux/tst-sysvshm-linux.c
index 2f05f21e4a..6986443c8f 100644
--- a/sysdeps/unix/sysv/linux/tst-sysvshm-linux.c
+++ b/sysdeps/unix/sysv/linux/tst-sysvshm-linux.c
@@ -122,18 +122,20 @@ do_test (void)
   if (shmid == -1)
     FAIL_EXIT1 ("shmget failed: %m");
 
+  /* It does check shmmax because kernel clamp its value to INT_MAX for:
+
+     1. Compat symbols with IPC_64, i.e, 32-bit binaries running on 64-bit
+        kernels.
+
+     2. Default symbol without IPC_64 (defined as IPC_OLD within Linux) and
+        glibc always use IPC_64 for 32-bit ABIs (to support 64-bit time_t).
+        It means that 32-bit binaries running on 32-bit kernels will not see
+        shmmax being clamped.
+
+     And finding out whether the compat symbol is used would require checking
+     the underlying kernel against the current ABI.  The shmall and shmmni
+     already provided enough coverage.  */
   struct test_shminfo tipcinfo;
-  {
-    uint64_t v = read_proc_file ("/proc/sys/kernel/shmmax");
-#if LONG_MAX == INT_MAX
-    /* Kernel explicit clamp the value for shmmax on compat symbol (32-bit
-       binaries running on 64-bit kernels).  */
-    if (sizeof (__syscall_ulong_t) == sizeof (unsigned long int)
-        && v > INT_MAX)
-      v = INT_MAX;
-#endif
-    tipcinfo.shmmax = v;
-  }
   tipcinfo.shmall = read_proc_file ("/proc/sys/kernel/shmall");
   tipcinfo.shmmni = read_proc_file ("/proc/sys/kernel/shmmni");
 
@@ -152,7 +154,6 @@ do_test (void)
       FAIL_EXIT1 ("shmctl with IPC_INFO failed: %m");
 
     TEST_COMPARE (ipcinfo.shmall, tipcinfo.shmall);
-    TEST_COMPARE (ipcinfo.shmmax, tipcinfo.shmmax);
     TEST_COMPARE (ipcinfo.shmmni, tipcinfo.shmmni);
   }
 
-- 
2.25.1

Re: [PATCH] linux: Remove shmmax check from tst-sysvshm-linux

[Andreas Schwab]

On Feb 02 2021, Adhemerval Zanella via Libc-alpha wrote:

> +  /* It does check shmmax because kernel clamp its value to INT_MAX for:

does not?

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 7578 EB47 D4E5 4D69 2510  2552 DF73 E780 A9DA AEC1
"And now for something completely different."

Re: [PATCH] linux: Remove shmmax check from tst-sysvshm-linux

[Adhemerval Zanella]

On 02/02/2021 11:09, Andreas Schwab wrote:
> On Feb 02 2021, Adhemerval Zanella via Libc-alpha wrote:
> 
>> +  /* It does check shmmax because kernel clamp its value to INT_MAX for:
> 
> does not?
> 

It should be 'does not'. Fixed locally, thanks.