diff mbox

src/nsexec: fix stack pointer alignment exception

Message ID 20170927055252.12225-1-zlang@redhat.com (mailing list archive)
State New, archived
Headers show

Commit Message

Zorro Lang Sept. 27, 2017, 5:52 a.m. UTC
When test g/317 or g/318 on ARM server, we got a kernel exception:

  kernel: nsexec[8203]: SP Alignment exception: pc=00000000004010a0 sp=00000000005200e8

nsexec gives an unaligned child stack address to clone() system
call sometimes. For making sure it's always aligned, use
"__attribute__((aligned))" extension of GCC (Thanks this suggestion
from Eric sandeen).

Signed-off-by: Zorro Lang <zlang@redhat.com>
---
 src/nsexec.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

Comments

Carlos Maiolino Sept. 27, 2017, 10:01 a.m. UTC | #1
On Wed, Sep 27, 2017 at 05:52:52AM +0000, Zorro Lang wrote:
> When test g/317 or g/318 on ARM server, we got a kernel exception:
> 
>   kernel: nsexec[8203]: SP Alignment exception: pc=00000000004010a0 sp=00000000005200e8
> 
> nsexec gives an unaligned child stack address to clone() system
> call sometimes. For making sure it's always aligned, use
> "__attribute__((aligned))" extension of GCC (Thanks this suggestion
> from Eric sandeen).
> 
> Signed-off-by: Zorro Lang <zlang@redhat.com>
> ---
>  src/nsexec.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/src/nsexec.c b/src/nsexec.c
> index f033b1a4..205dd081 100644
> --- a/src/nsexec.c
> +++ b/src/nsexec.c
> @@ -138,7 +138,8 @@ childFunc(void *arg)
>  
>  #define STACK_SIZE (1024 * 1024)
>  
> -static char child_stack[STACK_SIZE];    /* Space for child's stack */
> +/* Space for child's stack */
> +static char __attribute__((aligned)) child_stack[STACK_SIZE];

I don't really think the comment is necessary, other than that:

Reviewed-by: Carlos Maiolino <cmaiolino@redhat.com>

>  
>  int
>  main(int argc, char *argv[])
> -- 
> 2.13.5
> 
> --
> To unsubscribe from this list: send the line "unsubscribe fstests" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
Eric Sandeen Sept. 29, 2017, 2:44 p.m. UTC | #2
On 9/27/17 12:52 AM, Zorro Lang wrote:
> When test g/317 or g/318 on ARM server, we got a kernel exception:
> 
>   kernel: nsexec[8203]: SP Alignment exception: pc=00000000004010a0 sp=00000000005200e8
> 
> nsexec gives an unaligned child stack address to clone() system
> call sometimes. For making sure it's always aligned, use
> "__attribute__((aligned))" extension of GCC (Thanks this suggestion
> from Eric sandeen).
> 
> Signed-off-by: Zorro Lang <zlang@redhat.com>
> ---
>  src/nsexec.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/src/nsexec.c b/src/nsexec.c
> index f033b1a4..205dd081 100644
> --- a/src/nsexec.c
> +++ b/src/nsexec.c
> @@ -138,7 +138,8 @@ childFunc(void *arg)
>  
>  #define STACK_SIZE (1024 * 1024)
>  
> -static char child_stack[STACK_SIZE];    /* Space for child's stack */
> +/* Space for child's stack */
> +static char __attribute__((aligned)) child_stack[STACK_SIZE];

Does it matter if the __attribute__ goes before or after
the array name (child_stack[])?  I can't quite tell from reading i.e.
https://gcc.gnu.org/onlinedocs/gcc-3.1/gcc/Variable-Attributes.html#Variable%20Attributes
if it does, though most examples I see have the __attribute__ after the variable
name, i.e.

> short array[3] __attribute__ ((aligned));

in the URL above.

-Eric

>  
>  int
>  main(int argc, char *argv[])
> 
--
To unsubscribe from this list: send the line "unsubscribe fstests" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Zorro Lang Sept. 29, 2017, 3:52 p.m. UTC | #3
On Fri, Sep 29, 2017 at 09:44:05AM -0500, Eric Sandeen wrote:
> On 9/27/17 12:52 AM, Zorro Lang wrote:
> > When test g/317 or g/318 on ARM server, we got a kernel exception:
> > 
> >   kernel: nsexec[8203]: SP Alignment exception: pc=00000000004010a0 sp=00000000005200e8
> > 
> > nsexec gives an unaligned child stack address to clone() system
> > call sometimes. For making sure it's always aligned, use
> > "__attribute__((aligned))" extension of GCC (Thanks this suggestion
> > from Eric sandeen).
> > 
> > Signed-off-by: Zorro Lang <zlang@redhat.com>
> > ---
> >  src/nsexec.c | 3 ++-
> >  1 file changed, 2 insertions(+), 1 deletion(-)
> > 
> > diff --git a/src/nsexec.c b/src/nsexec.c
> > index f033b1a4..205dd081 100644
> > --- a/src/nsexec.c
> > +++ b/src/nsexec.c
> > @@ -138,7 +138,8 @@ childFunc(void *arg)
> >  
> >  #define STACK_SIZE (1024 * 1024)
> >  
> > -static char child_stack[STACK_SIZE];    /* Space for child's stack */
> > +/* Space for child's stack */
> > +static char __attribute__((aligned)) child_stack[STACK_SIZE];
> 
> Does it matter if the __attribute__ goes before or after
> the array name (child_stack[])?  I can't quite tell from reading i.e.
> https://gcc.gnu.org/onlinedocs/gcc-3.1/gcc/Variable-Attributes.html#Variable%20Attributes
> if it does, though most examples I see have the __attribute__ after the variable
> name, i.e.

Sorry, I didn't notice this "before/after" problem. Yes, you're right,
everyone use "__attribute__" at the end. I'll change that.

BTW, using it in the middle works (compile and run) fine from my side.

Thanks,
Zorro

> 
> > short array[3] __attribute__ ((aligned));
> 
> in the URL above.
> 
> -Eric
> 
> >  
> >  int
> >  main(int argc, char *argv[])
> > 
--
To unsubscribe from this list: send the line "unsubscribe fstests" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
diff mbox

Patch

diff --git a/src/nsexec.c b/src/nsexec.c
index f033b1a4..205dd081 100644
--- a/src/nsexec.c
+++ b/src/nsexec.c
@@ -138,7 +138,8 @@  childFunc(void *arg)
 
 #define STACK_SIZE (1024 * 1024)
 
-static char child_stack[STACK_SIZE];    /* Space for child's stack */
+/* Space for child's stack */
+static char __attribute__((aligned)) child_stack[STACK_SIZE];
 
 int
 main(int argc, char *argv[])