[PATCH] firewire: cdev: fix potential leak of kernel stack due to uninitialized value

[Takashi Sakamoto]

On Tue, Jun 14, 2022 at 03:07:46PM +0200, Takashi Iwai wrote:
> On Tue, 14 Jun 2022 14:30:36 +0200,
> Takashi Sakamoto wrote:
> > 
> > Hi Iwai-san,
> > 
> > I have a moderate request to you for the patch which fixes an issue
> > included in v5.19-rc1. If it's applicable and I can borrow your help
> > again, I'd like you to send the patch to mainline via your tree.
> 
> Do you have the lore URL I can get a patch from?
 
Here it is:

https://lore.kernel.org/alsa-devel/20220512112037.103142-1-o-takashi@sakamocchi.jp/

> > If possible, it's preferable to apply additional three patches I
> > respined[1], but it could be optional since not so critical.
> > 
> > [1] https://lore.kernel.org/alsa-devel/20220512111756.103008-1-o-takashi@sakamocchi.jp/
> 
> I can merge those, but now looking at the patches, I'm afraid that the
> patch 2 ("firewire: use struct_size over open coded arithmetic") is
> wrong; struct_size() takes the number of elements, and the element
> type is u32, hence you're allocating 4 times large data with that
> patch.

Indeed, I overlooked it. The length should be quadlet count instead of
byte count in the case. I'll post revised patches later. Thanks for your
review.


Regards

Takashi Sakamoto