Messages in this thread | | | Date | Tue, 23 Dec 2014 18:33:29 +0100 | From | Arend van Spriel <> | Subject | Re: [PATCH] brcmfmac: Do not crash if platform data is not populated |
| |
On 12/23/14 18:00, Kalle Valo wrote: > Arend van Spriel<arend@broadcom.com> writes: > >> On 12/23/14 16:47, Mika Westerberg wrote: >>> On Tue, Dec 23, 2014 at 04:37:21PM +0100, Arend van Spriel wrote: >>>> On 12/23/14 15:48, Mika Westerberg wrote: >>>>> The driver looks for pdata->oob_irq_supported to find out if wowl can be >>>>> supported. However, not all platforms populate pdata in which case we crash >>>>> the kernel because of NULL pointer dereference. >>>> >>>> Thanks, Mika >>>> >>>> However, this was already reported by Dan Carpenter and I submitted a patch >>>> for that a couple of days ago: "[PATCH 02/10] brcmfmac: Fix possible >>>> dereference of NULL pointer." [1]. >>>> >>>> Regards, >>>> Arend >>>> >>>> [1] >>>> http://mid.gmane.org/1419162233-19492-3-git-send-email-arend@broadcom.com >>> >>> Oh, good. I didn't notice that one. >>> >>> Thanks for fixing it :) >>> >>> BTW, that patch seems to miss brcmf_ops_sdio_resume(), perhaps it is >>> fixed in another patch? >> >> Aargh, a nice "BTW" this is as there is no other patch for the resume >> code. Maybe I should ask Kalle to apply your patch instead of ours. If >> it did not already got applied. > > I have been waiting for net-next to open so I haven't applied any -next > patches yet. So if you want I can take Mika's patch as well. Just send a > reply to your patch I should drop. > > And should this patch go to 3.19?
Yeah, the issue was introduced in 3.19 so that makes sense.
Regards, Arend
| |