2021-08-23 09:41:21

by Keoseong Park

[permalink] [raw]
Subject: [PATCH v2] scsi: ufs: ufshpb: Fix possible memory leak

When HPB pinned region exists and mctx allocation for this region fails,
memory leak is possible because memory is not released for the subregion
table of the current region.

So, change to free memory for the subregion table of the current region.

Signed-off-by: Keoseong Park <[email protected]>
---
v1 -> v2:
* Merge new kvfree() statement with the for-loop below it.
* Change to assign "hpb->rgn_tbl" when no error occurs.

drivers/scsi/ufs/ufshpb.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/scsi/ufs/ufshpb.c b/drivers/scsi/ufs/ufshpb.c
index 9acce92a356b..58db9ab8f0ae 100644
--- a/drivers/scsi/ufs/ufshpb.c
+++ b/drivers/scsi/ufs/ufshpb.c
@@ -1904,8 +1904,6 @@ static int ufshpb_alloc_region_tbl(struct ufs_hba *hba, struct ufshpb_lu *hpb)
if (!rgn_table)
return -ENOMEM;

- hpb->rgn_tbl = rgn_table;
-
for (rgn_idx = 0; rgn_idx < hpb->rgns_per_lu; rgn_idx++) {
int srgn_cnt = hpb->srgns_per_rgn;
bool last_srgn = false;
@@ -1942,10 +1940,12 @@ static int ufshpb_alloc_region_tbl(struct ufs_hba *hba, struct ufshpb_lu *hpb)
rgn->hpb = hpb;
}

+ hpb->rgn_tbl = rgn_table;
+
return 0;

release_srgn_table:
- for (i = 0; i < rgn_idx; i++)
+ for (i = 0; i <= rgn_idx; i++)
kvfree(rgn_table[i].srgn_tbl);

kvfree(rgn_table);
--
2.17.1


2021-08-23 16:22:19

by Bart Van Assche

[permalink] [raw]
Subject: Re: [PATCH v2] scsi: ufs: ufshpb: Fix possible memory leak

On 8/23/21 2:07 AM, Keoseong Park wrote:
> When HPB pinned region exists and mctx allocation for this region fails,
> memory leak is possible because memory is not released for the subregion
> table of the current region.

Reviewed-by: Bart Van Assche <[email protected]>

2021-08-24 03:23:56

by Martin K. Petersen

[permalink] [raw]
Subject: Re: [PATCH v2] scsi: ufs: ufshpb: Fix possible memory leak


Keoseong,

> When HPB pinned region exists and mctx allocation for this region
> fails, memory leak is possible because memory is not released for the
> subregion table of the current region.

Applied to 5.15/scsi-staging, thanks!

--
Martin K. Petersen Oracle Linux Engineering

2021-08-28 02:34:30

by Martin K. Petersen

[permalink] [raw]
Subject: Re: [PATCH v2] scsi: ufs: ufshpb: Fix possible memory leak

On Mon, 23 Aug 2021 18:07:14 +0900, Keoseong Park wrote:

> When HPB pinned region exists and mctx allocation for this region fails,
> memory leak is possible because memory is not released for the subregion
> table of the current region.
>
> So, change to free memory for the subregion table of the current region.
>
>
> [...]

Applied to 5.15/scsi-queue, thanks!

[1/1] scsi: ufs: ufshpb: Fix possible memory leak
https://git.kernel.org/mkp/scsi/c/6c9783e6296e

--
Martin K. Petersen Oracle Linux Engineering