[PATCH v7 03/14] ASoC: Intel: catpt: Add IPC message handlers

Andy Shevchenko andriy.shevchenko at linux.intel.com
Tue Sep 22 11:04:42 CEST 2020


On Mon, Sep 21, 2020 at 08:48:12PM +0000, Rojewski, Cezary wrote:
> On 2020-09-21 8:41 PM, Andy Shevchenko wrote:> On Mon, Sep 21, 2020 at 06:13:59PM +0000, Rojewski, Cezary wrote:
> >> On 2020-09-21 2:59 PM, Andy Shevchenko wrote:
> >>> On Mon, Sep 21, 2020 at 01:54:13PM +0200, Cezary Rojewski wrote:

...

> While this should never happen (means user is somehow not making use of
> officially released firmware binary), coredumps are useful only if you
> have access to debug tools. In cases you'd mentioned, invalid hash would
> have been dumped to coredump and crash reader simply wouldn't have been
> able to navigate to actual build for it. The rest of the coredump is still
> vital though.
> 
> memcpy() could be gated behind an 'if' for safety if needed:
> 
> 	info = cdev->ipc.config.fw_info;
> 	eof = info + FW_INFO_SIZE_MAX;
> 	/* navigate to fifth info segment (fw hash) */
> 	for (i = 0; i < 4 && info < eof; i++, info++)
> 		/* info segments are separated by space each */
> 		if ((info = strnchr(info, eof - info, ' ')) == NULL)
> 			break;

> 	if (i == 4 && info < eof)
> 		memcpy(pos, info, min(eof - info, CATPT_DUMP_HASH_SIZE));

And here basically enough check is info against NULL, right?
Just try to look at different possibilities how to make code simpler and neater.

> Didn't compile this, some typecheck fixes might be in order and so on.

-- 
With Best Regards,
Andy Shevchenko




More information about the Alsa-devel mailing list