[PATCH] staging: zcache: fix serialization bug in zv stats

December 30th, 2011 - 11:50 am ET by Seth Jennings | Report spam
In a multithreaded workload, the zv_curr_dist_counts
and zv_cumul_dist_counts statistics are being corrupted
because the increments and decrements in zv_create
and zv_free are not atomic.

This patch converts these statistics and their corresponding
increments/decrements/reads to atomic operations.

Based on v3.2-rc7

Signed-off-by: Seth Jennings <sjenning@linux.vnet.ibm.com>

drivers/staging/zcache/zcache-main.c | 14 +++++++-
1 files changed, 7 insertions(+), 7 deletions(-)

diff --git a/drivers/staging/zcache/zcache-main.c b/drivers/staging/zcache/zcache-main.c
index 56c1f9c..d39bb51 100644
a/drivers/staging/zcache/zcache-main.c
+++ b/drivers/staging/zcache/zcache-main.c
@@ -655,8 +655,8 @@ static unsigned int zv_max_zsize = (PAGE_SIZE / 8) * 7;
*/
static unsigned int zv_max_mean_zsize = (PAGE_SIZE / 8) * 5;

-static unsigned long zv_curr_dist_counts[NCHUNKS];
-static unsigned long zv_cumul_dist_counts[NCHUNKS];
+static atomic_t zv_curr_dist_counts[NCHUNKS];
+static atomic_t zv_cumul_dist_counts[NCHUNKS];

static struct zv_hdr *zv_create(struct xv_pool *xvpool, uint32_t pool_id,
struct tmem_oid *oid, uint32_t index,
@@ -675,8 +675,8 @@ static struct zv_hdr *zv_create(struct xv_pool *xvpool, uint32_t pool_id,
&page, &offset, ZCACHE_GFP_MASK);
if (unlikely(ret))
goto out;
- zv_curr_dist_counts[chunks]++;
- zv_cumul_dist_counts[chunks]++;
+ atomic_inc(&zv_curr_dist_counts[chunks]);
+ atomic_inc(&zv_cumul_dist_counts[chunks]);
zv = kmap_atomic(page, KM_USER0) + offset;
zv->index = index;
zv->oid = *oid;
@@ -698,7 +698,7 @@ static void zv_free(struct xv_pool *xvpool, struct zv_hdr *zv)

ASSERT_SENTINEL(zv, ZVH);
BUG_ON(chunks >= NCHUNKS);
- zv_curr_dist_counts[chunks]--;
+ atomic_dec(&zv_curr_dist_counts[chunks]);
size -= sizeof(*zv);
BUG_ON(size == 0);
INVERT_SENTINEL(zv, ZVH);
@@ -738,7 +738,7 @@ static int zv_curr_dist_counts_show(char *buf)
char *p = buf;

for (i = 0; i < NCHUNKS; i++) {
- n = zv_curr_dist_counts[i];
+ n = atomic_read(&zv_curr_dist_counts[i]);
p += sprintf(p, "%lu ", n);
chunks += n;
sum_total_chunks += i * n;
@@ -754,7 +754,7 @@ static int zv_cumul_dist_counts_show(char *buf)
char *p = buf;

for (i = 0; i < NCHUNKS; i++) {
- n = zv_cumul_dist_counts[i];
+ n = atomic_read(&zv_cumul_dist_counts[i]);
p += sprintf(p, "%lu ", n);
chunks += n;
sum_total_chunks += i * n;
1.7.5.4

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
email Follow the discussionReplies 1 replyReplies Make a reply

Replies

#1 Konrad Rzeszutek Wilk
January 03rd, 2012 - 04:50 pm ET | Report spam
On Fri, Dec 30, 2011 at 11:35:45AM -0600, Seth Jennings wrote:
On 12/30/2011 11:02 AM, Dan Magenheimer wrote:
>> From: Seth Jennings [mailto:]
>> Sent: Friday, December 30, 2011 9:42 AM
>> To: Greg Kroah-Hartman
>> Cc: Seth Jennings; Dan Magenheimer; Brian King; ; linux-
>>
>> Subject: [PATCH] staging: zcache: fix serialization bug in zv stats
>>
>> In a multithreaded workload, the zv_curr_dist_counts
>> and zv_cumul_dist_counts statistics are being corrupted
>> because the increments and decrements in zv_create
>> and zv_free are not atomic.
>>
>> This patch converts these statistics and their corresponding
>> increments/decrements/reads to atomic operations.
>>
>> Based on v3.2-rc7
>>
>> Signed-off-by: Seth Jennings
>
> I'm inclined to nack this change, at least unless inside an #ifdef DEBUG,

I guess I didn't respond to this suggestion. We could put #ifdef DEBUG
around the updates in zv_create and zv_free and around the two sysfs
*_show functions. Kinda messy though. I guess it might not matter
if we'll be re-evaluating this whole thing soon.



Perhaps also add in the TODO file that we want to move the stats collection
out of the code and only be enabled if CONFIG_DEBUGFS is set? (Which granted means
we need to implement the stats output in debugfs instead of sysfs).
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Similar topics