[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <422fbbe3-6f22-b16d-7af2-23efeb6dd125@embeddedor.com>
Date: Mon, 20 Sep 2021 18:57:06 -0500
From: "Gustavo A. R. Silva" <gustavo@...eddedor.com>
To: Len Baker <len.baker@....com>,
Matthew Wilcox <willy@...radead.org>,
Hannes Reinecke <hare@...e.com>,
"James E.J. Bottomley" <jejb@...ux.ibm.com>,
"Martin K. Petersen" <martin.petersen@...cle.com>
Cc: "Gustavo A. R. Silva" <gustavoars@...nel.org>,
Kees Cook <keescook@...omium.org>,
linux-hardening@...r.kernel.org, linux-scsi@...r.kernel.org,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH] scsi: advansys: Prefer struct_size over open coded
arithmetic
On 9/18/21 06:18, Len Baker wrote:
> As noted in the "Deprecated Interfaces, Language Features, Attributes,
> and Conventions" documentation [1], size calculations (especially
> multiplication) should not be performed in memory allocator (or similar)
> function arguments due to the risk of them overflowing. This could lead
> to values wrapping around and a smaller allocation being made than the
> caller was expecting. Using those allocations could lead to linear
> overflows of heap memory and other misbehaviors.
>
> So, refactor the code a bit to use the struct_size() helper instead of
> the argument "size + count * size" in the kzalloc() function.
>
> [1] https://www.kernel.org/doc/html/latest/process/deprecated.html#open-coded-arithmetic-in-allocator-arguments
>
> Signed-off-by: Len Baker <len.baker@....com>
> ---
> drivers/scsi/advansys.c | 5 +++--
> 1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/scsi/advansys.c b/drivers/scsi/advansys.c
> index ffb391967573..fe882502e7bf 100644
> --- a/drivers/scsi/advansys.c
> +++ b/drivers/scsi/advansys.c
> @@ -7465,6 +7465,7 @@ static int asc_build_req(struct asc_board *boardp, struct scsi_cmnd *scp,
> return ASC_BUSY;
> } else if (use_sg > 0) {
> int sgcnt;
> + size_t size;
I don't think a new variable is needed.
> struct scatterlist *slp;
> struct asc_sg_head *asc_sg_head;
>
> @@ -7477,8 +7478,8 @@ static int asc_build_req(struct asc_board *boardp, struct scsi_cmnd *scp,
> return ASC_ERROR;
> }
>
> - asc_sg_head = kzalloc(sizeof(asc_scsi_q->sg_head) +
> - use_sg * sizeof(struct asc_sg_list), GFP_ATOMIC);
> + size = struct_size(asc_scsi_q->sg_head, sg_list, use_sg);
> + asc_sg_head = kzalloc(size, GFP_ATOMIC);
You can go with this:
diff --git a/drivers/scsi/advansys.c b/drivers/scsi/advansys.c
index ffb391967573..e341b3372482 100644
--- a/drivers/scsi/advansys.c
+++ b/drivers/scsi/advansys.c
@@ -7477,8 +7477,8 @@ static int asc_build_req(struct asc_board *boardp, struct scsi_cmnd *scp,
return ASC_ERROR;
}
- asc_sg_head = kzalloc(sizeof(asc_scsi_q->sg_head) +
- use_sg * sizeof(struct asc_sg_list), GFP_ATOMIC);
+ asc_sg_head = kzalloc(struct_size(asc_sg_head, sg_list, use_sg),
+ GFP_ATOMIC);
if (!asc_sg_head) {
scsi_dma_unmap(scp);
set_host_byte(scp, DID_SOFT_ERROR);
It perfectly fits withing the 80 columns, which by the way is deprecated
but still good-to-have.
Also, if you are finding these instances with the help of Coccinelle, it'd be nice
if you mention that in the changelog text. :)
Thanks!
--
Gustavo
> if (!asc_sg_head) {
> scsi_dma_unmap(scp);
> set_host_byte(scp, DID_SOFT_ERROR);
> --
> 2.25.1
>
Powered by blists - more mailing lists