[LTP] [PATCH 1/1] hugeshmat04: Restore the correct shmmax value

[LTP] [PATCH 1/1] hugeshmat04: Restore the correct shmmax value

[Leo Yu-Chi Liang]

The type of shmmax is unsinged long in Linux kernel and the default value
is ((~0UL) - (1UL << 24)). If we try to access this value with %ld format
string in scanf-like function, we end up getting different incorrect values
depends on the implementation of different libc.

Fix this by accessing shmmax with the correct %lu format string.

Signed-off-by: Leo Yu-Chi Liang <ycliang@andestech.com>
---
 testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat04.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat04.c b/testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat04.c
index 50efa8a52..b3e7c272a 100644
--- a/testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat04.c
+++ b/testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat04.c
@@ -83,12 +83,12 @@ static void setup(void)
 	long hpage_size, orig_hugepages;
 
 	orig_hugepages = get_sys_tune("nr_hugepages");
-	SAFE_FILE_SCANF(PATH_SHMMAX, "%ld", &orig_shmmax);
-	SAFE_FILE_PRINTF(PATH_SHMMAX, "%ld", (long)SIZE);
-	SAFE_FILE_SCANF(PATH_SHMMAX, "%ld", &new_shmmax);
+	SAFE_FILE_SCANF(PATH_SHMMAX, "%lu", &orig_shmmax);
+	SAFE_FILE_PRINTF(PATH_SHMMAX, "%lu", (long)SIZE);
+	SAFE_FILE_SCANF(PATH_SHMMAX, "%lu", &new_shmmax);
 
 	if (new_shmmax < SIZE)
-		tst_brk(TCONF,	"shmmax too low, have: %ld", new_shmmax);
+		tst_brk(TCONF,	"shmmax too low, have: %lu", new_shmmax);
 
 	hpage_size = SAFE_READ_MEMINFO("Hugepagesize:") * 1024;
 
@@ -99,7 +99,7 @@ static void setup(void)
 static void cleanup(void)
 {
 	if (orig_shmmax != -1)
-		SAFE_FILE_PRINTF(PATH_SHMMAX, "%ld", orig_shmmax);
+		SAFE_FILE_PRINTF(PATH_SHMMAX, "%lu", orig_shmmax);
 }
 
 static struct tst_test test = {
-- 
2.34.1


-- 
Mailing list info: https://lists.linux.it/listinfo/ltp

Re: [LTP] [PATCH 1/1] hugeshmat04: Restore the correct shmmax value

[Cyril Hrubis]

Hi!
> Signed-off-by: Leo Yu-Chi Liang <ycliang@andestech.com>
> ---
>  testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat04.c | 10 +++++-----
>  1 file changed, 5 insertions(+), 5 deletions(-)
> 
> diff --git a/testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat04.c b/testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat04.c
> index 50efa8a52..b3e7c272a 100644
> --- a/testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat04.c
> +++ b/testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat04.c
> @@ -83,12 +83,12 @@ static void setup(void)
>  	long hpage_size, orig_hugepages;
>  
>  	orig_hugepages = get_sys_tune("nr_hugepages");
> -	SAFE_FILE_SCANF(PATH_SHMMAX, "%ld", &orig_shmmax);
> -	SAFE_FILE_PRINTF(PATH_SHMMAX, "%ld", (long)SIZE);
> -	SAFE_FILE_SCANF(PATH_SHMMAX, "%ld", &new_shmmax);
> +	SAFE_FILE_SCANF(PATH_SHMMAX, "%lu", &orig_shmmax);
> +	SAFE_FILE_PRINTF(PATH_SHMMAX, "%lu", (long)SIZE);
> +	SAFE_FILE_SCANF(PATH_SHMMAX, "%lu", &new_shmmax);
>  
>  	if (new_shmmax < SIZE)
> -		tst_brk(TCONF,	"shmmax too low, have: %ld", new_shmmax);
> +		tst_brk(TCONF,	"shmmax too low, have: %lu", new_shmmax);
>  
>  	hpage_size = SAFE_READ_MEMINFO("Hugepagesize:") * 1024;
>  
> @@ -99,7 +99,7 @@ static void setup(void)
>  static void cleanup(void)
>  {
>  	if (orig_shmmax != -1)
> -		SAFE_FILE_PRINTF(PATH_SHMMAX, "%ld", orig_shmmax);
> +		SAFE_FILE_PRINTF(PATH_SHMMAX, "%lu", orig_shmmax);
>  }

Can we please move the PATH_SHMMAX restoration to the .save_restore part
of the tst_test structure?

https://github.com/linux-test-project/ltp/wiki/C-Test-API#127-saving--restoring-procsys-values

-- 
Cyril Hrubis
chrubis@suse.cz

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp

Re: [LTP] [PATCH 1/1] hugeshmat04: Restore the correct shmmax value

[Leo Liang]

Hi Cyril,

On Thu, Aug 03, 2023 at 12:42:26PM +0200, Cyril Hrubis wrote:
> Hi!
> > Signed-off-by: Leo Yu-Chi Liang <ycliang@andestech.com>
> > ---
> >  testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat04.c | 10 +++++-----
> >  1 file changed, 5 insertions(+), 5 deletions(-)
> > 
> > diff --git a/testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat04.c b/testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat04.c
> > index 50efa8a52..b3e7c272a 100644
> > --- a/testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat04.c
> > +++ b/testcases/kernel/mem/hugetlb/hugeshmat/hugeshmat04.c
> > @@ -83,12 +83,12 @@ static void setup(void)
> >  	long hpage_size, orig_hugepages;
> >  
> >  	orig_hugepages = get_sys_tune("nr_hugepages");
> > -	SAFE_FILE_SCANF(PATH_SHMMAX, "%ld", &orig_shmmax);
> > -	SAFE_FILE_PRINTF(PATH_SHMMAX, "%ld", (long)SIZE);
> > -	SAFE_FILE_SCANF(PATH_SHMMAX, "%ld", &new_shmmax);
> > +	SAFE_FILE_SCANF(PATH_SHMMAX, "%lu", &orig_shmmax);
> > +	SAFE_FILE_PRINTF(PATH_SHMMAX, "%lu", (long)SIZE);
> > +	SAFE_FILE_SCANF(PATH_SHMMAX, "%lu", &new_shmmax);
> >  
> >  	if (new_shmmax < SIZE)
> > -		tst_brk(TCONF,	"shmmax too low, have: %ld", new_shmmax);
> > +		tst_brk(TCONF,	"shmmax too low, have: %lu", new_shmmax);
> >  
> >  	hpage_size = SAFE_READ_MEMINFO("Hugepagesize:") * 1024;
> >  
> > @@ -99,7 +99,7 @@ static void setup(void)
> >  static void cleanup(void)
> >  {
> >  	if (orig_shmmax != -1)
> > -		SAFE_FILE_PRINTF(PATH_SHMMAX, "%ld", orig_shmmax);
> > +		SAFE_FILE_PRINTF(PATH_SHMMAX, "%lu", orig_shmmax);
> >  }
> 
> Can we please move the PATH_SHMMAX restoration to the .save_restore part
> of the tst_test structure?
> 
> https://github.com/linux-test-project/ltp/wiki/C-Test-API#127-saving--restoring-procsys-values

Thanks for the pointer.
I will send a v2 patch soon!

Best regards,
Leo

> 
> -- 
> Cyril Hrubis
> chrubis@suse.cz

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp