[KJ] Re: [Patch] kzalloc() conversion in drivers/usb/gadget

[David Brownell <david-b@pacbell.net>]

[-- Attachment #1: Type: text/plain, Size: 963 bytes --]

On Monday 27 February 2006 12:36 pm, Eric Sesterhenn wrote:

> I think there was a bug in drivers/usb/gadget/inode.c because
> it used sizeof(*data) for the kmalloc() and sizeof(data) for
> the memset(), since sizeof(data) just returns the size for a pointer.

Actually in that case it's safe, since all the fields get explicitly
initialized ... most of them right after the memset(), but a couple
of them later on before they're used.  But the same paranoia that
argued for explicit zeroing argues for your fix.  :)


> I assume there is also a bug in drivers/usb/gadget/omap_udc.c
> in omap_alloc_request(), req gets allocated via kmalloc() and
> the function returns &req->req, no matter if req is NULL, which
> looks like a NULL dereference to me, maybe someone who is more familar
> with this code might have a look at it.

That should get fixed too.  Please roll a patch for that.

I'll sign off the conversion patch and forward it to Greg.
Thanks!

- Dave

[-- Attachment #2: Type: text/plain, Size: 168 bytes --]

_______________________________________________
Kernel-janitors mailing list
Kernel-janitors@lists.osdl.org
https://lists.osdl.org/mailman/listinfo/kernel-janitors