[Bug 60852] Random memory corruption reported after writing to /sys/module/drm_kms_helper/parameters/edid_firmware

[bugzilla-daemon at bugzilla.kernel.org]

https://bugzilla.kernel.org/show_bug.cgi?id=60852

--- Comment #8 from Jani Nikula <jani.nikula at intel.com> ---
(In reply to Alex Villacis Lasso from comment #7)
> It does. I thought I made a point of basing the compiled kernel config off
> the distro kernel config except in the places I wanted to change, and I did
> not intend to set CONFIG_FW_LOADER_USER_HELPER in my config. If I set
> CONFIG_FW_LOADER_USER_HELPER=n in my config and recompile, could that
> influence whether the bug might trigger, even before applying the debug
> patch?

With CONFIG_FW_LOADER_USER_HELPER=n the kernel doesn't try to ask the user
space to supply the firmware (=edid in this case). Which also means there's no
60 second timeout for the userspace reply which apparently doesn't happen. And,
*if* the problem is in the firmware loader code, that might vastly influence
whether the bug might trigger.

Now, the 60 second timeout isn't this bug. (It might be considered a different
bug.) This bug is the list corruption you see. All else being equal, if you can
reproduce it with the patch, it's in DRM. If the patch makes the problem go
away, it's in the firmware loader. So we'd appreciate you trying the debug
patch, thanks. :)

-- 
You are receiving this mail because:
You are on the CC list for the bug.
You are the assignee for the bug.