[LTP] [PATCH] increasing buffer size for correct alignment

[LTP] [PATCH] increasing buffer size for correct alignment

[Han Pingtian]

Hey there,

I have a patch for aio-stress to align buffer address correctly. Please
have a look. Thanks in advance.

For align buffer address to the number specified by -a option in
aio-stress, it needs to add page_size_mask extra length to total_ram
before call allocation functions. Or it wouldn't be correct in this line
after the allocation:

         p = (char *)((intptr_t) (p + page_size_mask) & ~page_size_mask);

Signed-off-by: Han Pingtian <hanpt@linux.vnet.ibm.com>
---
 testcases/kernel/io/ltp-aiodio/aio-stress.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/testcases/kernel/io/ltp-aiodio/aio-stress.c b/testcases/kernel/io/ltp-aiodio/aio-stress.c
index 0b7148c..24675ec 100644
--- a/testcases/kernel/io/ltp-aiodio/aio-stress.c
+++ b/testcases/kernel/io/ltp-aiodio/aio-stress.c
@@ -1022,7 +1022,7 @@ int setup_shared_mem(int num_threads, int num_files, int depth,
 	if (use_shm == USE_MALLOC) {
 		p = malloc(total_ram + page_size_mask);
 	} else if (use_shm == USE_SHM) {
-		shm_id = shmget(IPC_PRIVATE, total_ram, IPC_CREAT | 0700);
+		shm_id = shmget(IPC_PRIVATE, total_ram + page_size_mask, IPC_CREAT | 0700);
 		if (shm_id < 0) {
 			perror("shmget");
 			drop_shm();
@@ -1046,9 +1046,9 @@ int setup_shared_mem(int num_threads, int num_files, int depth,
 			goto free_buffers;
 		}
 		unlink(mmap_name);
-		ftruncate(fd, total_ram);
+		ftruncate(fd, total_ram + page_size_mask);
 		shm_id = fd;
-		p = mmap((char *)0x50000000, total_ram,
+		p = mmap((char *)0x50000000, total_ram + page_size_mask,
 			 PROT_READ | PROT_WRITE, MAP_SHARED, fd, 0);
 
 		if (p == MAP_FAILED) {
-- 
1.9.3


------------------------------------------------------------------------------
Dive into the World of Parallel Programming. The Go Parallel Website,
sponsored by Intel and developed in partnership with Slashdot Media, is your
hub for all things parallel software development, from weekly thought
leadership blogs to news, videos, case studies, tutorials and more. Take a
look and join the conversation now. http://goparallel.sourceforge.net/
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list

Re: [LTP] [PATCH] increasing buffer size for correct alignment

[Cyril Hrubis]

Hi!
> I have a patch for aio-stress to align buffer address correctly. Please
> have a look. Thanks in advance.
> 
> For align buffer address to the number specified by -a option in
> aio-stress, it needs to add page_size_mask extra length to total_ram
> before call allocation functions. Or it wouldn't be correct in this line
> after the allocation:
> 
>          p = (char *)((intptr_t) (p + page_size_mask) & ~page_size_mask);

You are adding the page_size_mask to all uses of total_ram, wouldn't it
be easier to add it to the total_ram at the start of the function (and
remove the page_size_mask from the malloc() call that has it allready)?

-- 
Cyril Hrubis
chrubis@suse.cz

------------------------------------------------------------------------------
Dive into the World of Parallel Programming. The Go Parallel Website,
sponsored by Intel and developed in partnership with Slashdot Media, is your
hub for all things parallel software development, from weekly thought
leadership blogs to news, videos, case studies, tutorials and more. Take a
look and join the conversation now. http://goparallel.sourceforge.net/
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list

Re: [LTP] [PATCH] increasing buffer size for correct alignment

[Han Pingtian]

On Mon, Jan 26, 2015 at 02:45:56PM +0100, Cyril Hrubis wrote:
> You are adding the page_size_mask to all uses of total_ram, wouldn't it
> be easier to add it to the total_ram at the start of the function (and
> remove the page_size_mask from the malloc() call that has it allready)?
> 
Yes, you're right. This is the new patch. Please review. Thanks.


From 8d0521c7fbee1b46f27d8268fee28e9721c79cc2 Mon Sep 17 00:00:00 2001
From: Han Pingtian <hanpt@linux.vnet.ibm.com>
Date: Mon, 26 Jan 2015 17:29:41 +0800
Subject: [PATCH] increasing buffer size for correct alignment

For align buffer address to the number specified by -a option in
aio-stress, it needs to add page_size_mask extra length to total_ram
before call allocation functions. Or it wouldn't be correct in this line
after the allocation:

         p = (char *)((intptr_t) (p + page_size_mask) & ~page_size_mask);

Signed-off-by: Han Pingtian <hanpt@linux.vnet.ibm.com>
---
 testcases/kernel/io/ltp-aiodio/aio-stress.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/testcases/kernel/io/ltp-aiodio/aio-stress.c b/testcases/kernel/io/ltp-aiodio/aio-stress.c
index 0b7148c..06a4953 100644
--- a/testcases/kernel/io/ltp-aiodio/aio-stress.c
+++ b/testcases/kernel/io/ltp-aiodio/aio-stress.c
@@ -1019,8 +1019,11 @@ int setup_shared_mem(int num_threads, int num_files, int depth,
 	if (verify)
 		total_ram += padded_reclen;
 
+	/* for aligning buffer after the allocation */
+	total_ram += page_size_mask;
+
 	if (use_shm == USE_MALLOC) {
-		p = malloc(total_ram + page_size_mask);
+		p = malloc(total_ram);
 	} else if (use_shm == USE_SHM) {
 		shm_id = shmget(IPC_PRIVATE, total_ram, IPC_CREAT | 0700);
 		if (shm_id < 0) {
-- 
1.9.3


------------------------------------------------------------------------------
Dive into the World of Parallel Programming. The Go Parallel Website,
sponsored by Intel and developed in partnership with Slashdot Media, is your
hub for all things parallel software development, from weekly thought
leadership blogs to news, videos, case studies, tutorials and more. Take a
look and join the conversation now. http://goparallel.sourceforge.net/
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list

Re: [LTP] [PATCH] increasing buffer size for correct alignment

[Cyril Hrubis]

Hi!
Pushed, thanks.

-- 
Cyril Hrubis
chrubis@suse.cz

------------------------------------------------------------------------------
Dive into the World of Parallel Programming. The Go Parallel Website,
sponsored by Intel and developed in partnership with Slashdot Media, is your
hub for all things parallel software development, from weekly thought
leadership blogs to news, videos, case studies, tutorials and more. Take a
look and join the conversation now. http://goparallel.sourceforge.net/
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list