[PATCH] Fix error-code overwrite bug

[PATCH] Fix error-code overwrite bug

[Giel van Schijndel]

Don't overwrite the returned error code with the boolean test used by
the if-statement (otherwise it'd be 1 or 0 always, 1 in the if-block).
---
 drivers/mtd/nftlmount.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/mtd/nftlmount.c b/drivers/mtd/nftlmount.c
index 51b9d6a..1cbeb6b 100644
--- a/drivers/mtd/nftlmount.c
+++ b/drivers/mtd/nftlmount.c
@@ -91,7 +91,7 @@ static int find_boot_record(struct NFTLrecord *nftl)
 		/* To be safer with BIOS, also use erase mark as discriminant */
 		if ((ret = nftl_read_oob(mtd, block * nftl->EraseSize +
 					 SECTORSIZE + 8, 8, &retlen,
-					 (char *)&h1) < 0)) {
+					 (char *)&h1)) < 0) {
 			printk(KERN_WARNING "ANAND header found at 0x%x in mtd%d, but OOB data read failed (err %d)\n",
 			       block * nftl->EraseSize, nftl->mbd.mtd->index, ret);
 			continue;
-- 
2.1.4

Re: [PATCH] Fix error-code overwrite bug

[Giel van Schijndel]

[-- Attachment #1: Type: text/plain, Size: 500 bytes --]

On Sun, Jan 04, 2015 at 19:04:22 +0100, Giel van Schijndel wrote:
> Don't overwrite the returned error code with the boolean test used by
> the if-statement (otherwise it'd be 1 or 0 always, 1 in the if-block).
> ---

Forgot to:
Signed-off-by: Giel van Schijndel <me@mortis.eu>

-- 
Met vriendelijke groet,
With kind regards,
Giel van Schijndel
--
"If debugging is the process of removing software bugs, then programming
 must be the process of putting them in."
  -- Edsger Dijkstra

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 181 bytes --]

Re: [PATCH] Fix error-code overwrite bug

[Andy Shevchenko]

On Sun, Jan 4, 2015 at 8:04 PM, Giel van Schijndel <me@mortis.eu> wrote:
> Don't overwrite the returned error code with the boolean test used by
> the if-statement (otherwise it'd be 1 or 0 always, 1 in the if-block).
> ---
>  drivers/mtd/nftlmount.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/mtd/nftlmount.c b/drivers/mtd/nftlmount.c
> index 51b9d6a..1cbeb6b 100644
> --- a/drivers/mtd/nftlmount.c
> +++ b/drivers/mtd/nftlmount.c
> @@ -91,7 +91,7 @@ static int find_boot_record(struct NFTLrecord *nftl)
>                 /* To be safer with BIOS, also use erase mark as discriminant */
>                 if ((ret = nftl_read_oob(mtd, block * nftl->EraseSize +
>                                          SECTORSIZE + 8, 8, &retlen,
> -                                        (char *)&h1) < 0)) {
> +                                        (char *)&h1)) < 0) {

Better to move ret = x(); outside of condition. See here:
http://permalink.gmane.org/gmane.linux.drivers.mtd/56922

P.S.If you are going to fix more, may you talk to me via some channel
like irc, jabber, etc?


>                         printk(KERN_WARNING "ANAND header found at 0x%x in mtd%d, but OOB data read failed (err %d)\n",
>                                block * nftl->EraseSize, nftl->mbd.mtd->index, ret);
>                         continue;
> --
> 2.1.4
>
> --
> 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/



-- 
With Best Regards,
Andy Shevchenko

Re: [PATCH] Fix error-code overwrite bug

[Giel van Schijndel]

[-- Attachment #1: Type: text/plain, Size: 1912 bytes --]

On Wed, Jan 07, 2015 at 20:37:29 +0200, Andy Shevchenko wrote:
> On Sun, Jan 4, 2015 at 8:04 PM, Giel van Schijndel <me@mortis.eu> wrote:
>> Don't overwrite the returned error code with the boolean test used by
>> the if-statement (otherwise it'd be 1 or 0 always, 1 in the if-block).
>> ---
>>  drivers/mtd/nftlmount.c | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/drivers/mtd/nftlmount.c b/drivers/mtd/nftlmount.c
>> index 51b9d6a..1cbeb6b 100644
>> --- a/drivers/mtd/nftlmount.c
>> +++ b/drivers/mtd/nftlmount.c
>> @@ -91,7 +91,7 @@ static int find_boot_record(struct NFTLrecord *nftl)
>>                 /* To be safer with BIOS, also use erase mark as discriminant */
>>                 if ((ret = nftl_read_oob(mtd, block * nftl->EraseSize +
>>                                          SECTORSIZE + 8, 8, &retlen,
>> -                                        (char *)&h1) < 0)) {
>> +                                        (char *)&h1)) < 0) {
> 
> Better to move ret = x(); outside of condition. See here:
> http://permalink.gmane.org/gmane.linux.drivers.mtd/56922

In the sense that this bug wouldn't have occurred when using separate
assignment and condition checking you're right. It's a style issue
though, but a relevant one.

So your approach is probably better, though incomplete (like mine), just
look for the exact same (ret = x() < 0) pattern about 20 lines further
down the same file. (Yes that's disabled code, but I still believe the
bug should be fixed considering it's exactly the same class of bug).

So I suggest you resend that ^^ patch you link to with a fix for the
other instance of the bug fixed as well.

-- 
Met vriendelijke groet,
With kind regards,
Giel van Schijndel
--
"If debugging is the process of removing software bugs, then programming
 must be the process of putting them in."
  -- Edsger Dijkstra

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 181 bytes --]

Re: [PATCH] Fix error-code overwrite bug

[Andy Shevchenko]

On Wed, Jan 7, 2015 at 9:52 PM, Giel van Schijndel <me@mortis.eu> wrote:
> On Wed, Jan 07, 2015 at 20:37:29 +0200, Andy Shevchenko wrote:
>> On Sun, Jan 4, 2015 at 8:04 PM, Giel van Schijndel <me@mortis.eu> wrote:
>>> Don't overwrite the returned error code with the boolean test used by
>>> the if-statement (otherwise it'd be 1 or 0 always, 1 in the if-block).
>>> ---
>>>  drivers/mtd/nftlmount.c | 2 +-
>>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/drivers/mtd/nftlmount.c b/drivers/mtd/nftlmount.c
>>> index 51b9d6a..1cbeb6b 100644
>>> --- a/drivers/mtd/nftlmount.c
>>> +++ b/drivers/mtd/nftlmount.c
>>> @@ -91,7 +91,7 @@ static int find_boot_record(struct NFTLrecord *nftl)
>>>                 /* To be safer with BIOS, also use erase mark as discriminant */
>>>                 if ((ret = nftl_read_oob(mtd, block * nftl->EraseSize +
>>>                                          SECTORSIZE + 8, 8, &retlen,
>>> -                                        (char *)&h1) < 0)) {
>>> +                                        (char *)&h1)) < 0) {
>>
>> Better to move ret = x(); outside of condition. See here:
>> http://permalink.gmane.org/gmane.linux.drivers.mtd/56922
>
> In the sense that this bug wouldn't have occurred when using separate
> assignment and condition checking you're right. It's a style issue
> though, but a relevant one.
>
> So your approach is probably better, though incomplete (like mine), just
> look for the exact same (ret = x() < 0) pattern about 20 lines further
> down the same file. (Yes that's disabled code, but I still believe the
> bug should be fixed considering it's exactly the same class of bug).
>
> So I suggest you resend that ^^ patch you link to with a fix for the
> other instance of the bug fixed as well.

Sounds reasonable. Will do.


-- 
With Best Regards,
Andy Shevchenko