On 11/24/25 16:06, Max Staudt wrote:
On 11/24/25 3:32 PM, Jiri Slaby wrote:
Isn't this fixed already by: commit 8513c154f8ad7097653dd9bf43d6155e5aad4ab3 Author: Abdun Nihaal nihaal@cse.iitm.ac.in Date: Mon Nov 10 22:45:50 2025 +0530
HID: playstation: Fix memory leak in dualshock4_get_calibration_data() ?
As far as I can see, that patch does indeed fix the same issue, and it is already upstream.
Thanks for the hint - Abdun's patch has been upstreamed quite recently, hence I guess Eslam missed it by accident. But maybe I'm wrong and Eslam can chime in himself?
Thank's Max & Jiri, sorry i was sick the past couple of days i missed your replies. yes. that patch fixes it. I guess i missed it because it wasn't merged yet when i submitted v1. So please ignore this patch.
Anyway, this is a typical use-case for __free(). Why not to use that?
Wow, there's been a lot of interesting stuff happening around cleanup.h. I've been out of the kernel for too long, this looks like fun. Thanks for pointing it out :)
Max
Lastly, One question to max, at the beginning of the function dualshock4_get_calibration_data buf = kzalloc(DS4_FEATURE_REPORT_CALIBRATION_SIZE, GFP_KERNEL); if (!buf) { ret = -ENOMEM; goto transfer_failed; } if the allocation fails. can't we just return here . or do we need to go the the end of the function and do sanity checks at the end?