[PATCH v2 1/2] sysdeps: tst-bz21269: fix test parameter

[PATCH v2 1/2] sysdeps: tst-bz21269: fix test parameter

[Sam James]

All callers pass 1 or 0x11 anyway (same meaning according to man page),
but still.

Signed-off-by: Sam James <sam@gentoo.org>
---

 sysdeps/unix/sysv/linux/i386/tst-bz21269.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sysdeps/unix/sysv/linux/i386/tst-bz21269.c b/sysdeps/unix/sysv/linux/i386/tst-bz21269.c
index 51d4a1b082..f508ef8f16 100644
--- a/sysdeps/unix/sysv/linux/i386/tst-bz21269.c
+++ b/sysdeps/unix/sysv/linux/i386/tst-bz21269.c
@@ -52,7 +52,7 @@ xset_thread_area (struct user_desc *u_info)
 static void
 xmodify_ldt (int func, const void *ptr, unsigned long bytecount)
 {
-  TEST_VERIFY_EXIT (syscall (SYS_modify_ldt, 1, ptr, bytecount) == 0);
+  TEST_VERIFY_EXIT (syscall (SYS_modify_ldt, func, ptr, bytecount) == 0);
 }
 
 static int
-- 
2.41.0

[PATCH v2 2/2] sysdeps: tst-bz21269: handle ENOSYS & skip appropriately

[Sam James]

SYS_modify_ldt requires CONFIG_MODIFY_LDT_SYSCALL to be set in the kernel, which
some distributions may disable for hardening. Check if that's the case (unset)
and mark the test as UNSUPPORTED if so.

Signed-off-by: Sam James <sam@gentoo.org>
---
Fix the FAIL_EXIT1 error.

 sysdeps/unix/sysv/linux/i386/tst-bz21269.c | 11 ++++++++++-
 1 file changed, 10 insertions(+), 1 deletion(-)

diff --git a/sysdeps/unix/sysv/linux/i386/tst-bz21269.c b/sysdeps/unix/sysv/linux/i386/tst-bz21269.c
index f508ef8f16..28f5359bea 100644
--- a/sysdeps/unix/sysv/linux/i386/tst-bz21269.c
+++ b/sysdeps/unix/sysv/linux/i386/tst-bz21269.c
@@ -52,7 +52,16 @@ xset_thread_area (struct user_desc *u_info)
 static void
 xmodify_ldt (int func, const void *ptr, unsigned long bytecount)
 {
-  TEST_VERIFY_EXIT (syscall (SYS_modify_ldt, func, ptr, bytecount) == 0);
+  long ret = syscall (SYS_modify_ldt, func, ptr, bytecount);
+
+  if (ret == -1)
+    {
+      if (errno == ENOSYS)
+	FAIL_UNSUPPORTED ("modify_ldt not supported");
+      FAIL_EXIT1 ("modify_ldt failed (errno=%d)", errno);
+    }
+
+  return 0;
 }
 
 static int
-- 
2.41.0

Re: [PATCH v2 2/2] sysdeps: tst-bz21269: handle ENOSYS & skip appropriately

[DJ Delorie]

Sam James via Libc-alpha <libc-alpha@sourceware.org> writes:
> SYS_modify_ldt requires CONFIG_MODIFY_LDT_SYSCALL to be set in the kernel, which
> some distributions may disable for hardening. Check if that's the case (unset)
> and mark the test as UNSUPPORTED if so.

LGTM
Reviewed-by: DJ Delorie <dj@redhat.com>

> Signed-off-by: Sam James <sam@gentoo.org>
> ---
> Fix the FAIL_EXIT1 error.
>
>  sysdeps/unix/sysv/linux/i386/tst-bz21269.c | 11 ++++++++++-
>  1 file changed, 10 insertions(+), 1 deletion(-)
>
> diff --git a/sysdeps/unix/sysv/linux/i386/tst-bz21269.c b/sysdeps/unix/sysv/linux/i386/tst-bz21269.c
> index f508ef8f16..28f5359bea 100644
> --- a/sysdeps/unix/sysv/linux/i386/tst-bz21269.c
> +++ b/sysdeps/unix/sysv/linux/i386/tst-bz21269.c
> @@ -52,7 +52,16 @@ xset_thread_area (struct user_desc *u_info)
>  static void
>  xmodify_ldt (int func, const void *ptr, unsigned long bytecount)
>  {
> -  TEST_VERIFY_EXIT (syscall (SYS_modify_ldt, func, ptr, bytecount) == 0);
> +  long ret = syscall (SYS_modify_ldt, func, ptr, bytecount);
> +
> +  if (ret == -1)
> +    {
> +      if (errno == ENOSYS)
> +	FAIL_UNSUPPORTED ("modify_ldt not supported");
> +      FAIL_EXIT1 ("modify_ldt failed (errno=%d)", errno);
> +    }
> +
> +  return 0;
>  }
>  
>  static int

Re: [PATCH v2 2/2] sysdeps: tst-bz21269: handle ENOSYS & skip appropriately

[Sam James]

DJ Delorie <dj@redhat.com> writes:

> Sam James via Libc-alpha <libc-alpha@sourceware.org> writes:
>> SYS_modify_ldt requires CONFIG_MODIFY_LDT_SYSCALL to be set in the kernel, which
>> some distributions may disable for hardening. Check if that's the case (unset)
>> and mark the test as UNSUPPORTED if so.
>
> LGTM
> Reviewed-by: DJ Delorie <dj@redhat.com>


Thanks DJ! Pushed.

Re: [PATCH v2 2/2] sysdeps: tst-bz21269: handle ENOSYS & skip appropriately

[Andreas Schwab]

../sysdeps/unix/sysv/linux/i386/tst-bz21269.c: In function 'xmodify_ldt':
../sysdeps/unix/sysv/linux/i386/tst-bz21269.c:64:10: error: 'return' with a value, in function returning void [-Werror=return-type]
   64 |   return 0;
      |          ^
../sysdeps/unix/sysv/linux/i386/tst-bz21269.c:53:1: note: declared here
   53 | xmodify_ldt (int func, const void *ptr, unsigned long bytecount)
      | ^~~~~~~~~~~
cc1: all warnings being treated as errors

-- 
Andreas Schwab, SUSE Labs, schwab@suse.de
GPG Key fingerprint = 0196 BAD8 1CE9 1970 F4BE  1748 E4D4 88E3 0EEA B9D7
"And now for something completely different."

Re: [PATCH v2 2/2] sysdeps: tst-bz21269: handle ENOSYS & skip appropriately

[Sam James]

Andreas Schwab <schwab@suse.de> writes:

> ../sysdeps/unix/sysv/linux/i386/tst-bz21269.c: In function 'xmodify_ldt':
> ../sysdeps/unix/sysv/linux/i386/tst-bz21269.c:64:10: error: 'return' with a value, in function returning void [-Werror=return-type]
>    64 |   return 0;
>       |          ^
> ../sysdeps/unix/sysv/linux/i386/tst-bz21269.c:53:1: note: declared here
>    53 | xmodify_ldt (int func, const void *ptr, unsigned long bytecount)
>       | ^~~~~~~~~~~
> cc1: all warnings being treated as errors

Fixed, thanks.

Re: [PATCH v2 1/2] sysdeps: tst-bz21269: fix test parameter

[DJ Delorie]

Sam James via Libc-alpha <libc-alpha@sourceware.org> writes:
> All callers pass 1 or 0x11 anyway (same meaning according to man page),
> but still.

Confirmed.  LGTM.  I was just wondering about this yesterday, by
coincidence.

Reviewed-by: DJ Delorie <dj@redhat.com>

> -  TEST_VERIFY_EXIT (syscall (SYS_modify_ldt, 1, ptr, bytecount) == 0);
> +  TEST_VERIFY_EXIT (syscall (SYS_modify_ldt, func, ptr, bytecount) == 0);