On 13/12/17 21:02, Boris Brezillon wrote:
> Hi Colin,
>
> On Wed, 13 Dec 2017 20:49:09 +0000
> Colin King <colin.k...@canonical.com> wrote:
>
>> From: Colin Ian King <colin.k...@canonical.com>
>>
>> There are two error return paths that are not kfree'ing ep that
>> lead to memory leaks. Fix this by exiting on error via the
>> ext_out exit path that performs the necessary kfree.
>
> Well, given that no one implements the ->exec_op() hook yet, the
> xxxx_op() functions always return 0 right now, but I agree, we
> should fix the generic case.
>
>>
>> Detected by CoverityScan, CID#1462747 ("Resource Leak")
>>
>> Fixes: b83ea87958c5 ("mtd: nand: provide several helpers to do common NAND
>> operations")
>> Signed-off-by: Colin Ian King <colin.k...@canonical.com>
>
> If you don't mind, I'd like to squash these changes in the original
> commit.
Yep, no problem with squashing it.
Colin
>
> Thanks,
>
> Boris
>
>> ---
>> drivers/mtd/nand/nand_base.c | 4 ++--
>> 1 file changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/mtd/nand/nand_base.c b/drivers/mtd/nand/nand_base.c
>> index eb810d5d44e7..1c5126d1db8c 100644
>> --- a/drivers/mtd/nand/nand_base.c
>> +++ b/drivers/mtd/nand/nand_base.c
>> @@ -5046,14 +5046,14 @@ static int nand_flash_detect_ext_param_page(struct
>> nand_chip *chip,
>> /* Send our own NAND_CMD_PARAM. */
>> ret = nand_read_param_page_op(chip, 0, NULL, 0);
>> if (ret)
>> - return ret;
>> + goto ext_out;
>>
>> /* Use the Change Read Column command to skip the ONFI param pages. */
>> ret = nand_change_read_column_op(chip,
>> sizeof(*p) * p->num_of_param_pages,
>> ep, len, true);
>> if (ret)
>> - return ret;
>> + goto ext_out;
>>
>> ret = -EINVAL;
>> if ((onfi_crc16(ONFI_CRC_BASE, ((uint8_t *)ep) + 2, len - 2)
>
