Re: [PATCH -mmotm] efi: drop kmemleak_ignore() for page allocator

From: Ard Biesheuvel
Date: Wed Dec 26 2018 - 10:34:48 EST


On Wed, 26 Dec 2018 at 16:13, Qian Cai <cai@xxxxxx> wrote:
>
> On 12/26/18 7:02 AM, Ard Biesheuvel wrote:
> > On Wed, 26 Dec 2018 at 03:35, Qian Cai <cai@xxxxxx> wrote:
> >>
> >> a0fc5578f1d (efi: Let kmemleak ignore false positives) is no longer
> >> needed due to efi_mem_reserve_persistent() uses __get_free_page()
> >> instead where kmemelak is not able to track regardless. Otherwise,
> >> kernel reported "kmemleak: Trying to color unknown object at
> >> 0xffff801060ef0000 as Black"
> >>
> >> Signed-off-by: Qian Cai <cai@xxxxxx>
> >
> > Why are you sending this to -mmotm?
> >
> > Andrew, please disregard this patch. This is EFI/tip material.
>
> Well, I'd like to primarily develop on the -mmotm tree as it fits in a
> sweet-spot where the mainline is too slow and linux-next is too chaotic.
>
> The bug was reproduced and the patch was tested on -mmotm. If for every bugs
> people found in -mmtom, they have to check out the corresponding sub-system tree
> and reproduce/verify the bug over there, that is quite a burden to bear.
>

Yes. But you know what? We all have our burden to bear, and shifting
this burden to someone else, in this case the subsystem maintainer who
typically deals with a sizable workload already, is not a very nice
thing to do.

> That's why sub-system maintainers are copied on those patches, so they can
> decide to fix directly in the sub-system tree instead of -mmotm, and then it
> will propagate to -mmotm one way or another.
>

Please stop sending EFI patches if you can't be bothered to
test/reproduce against the EFI tree.

Thanks,
Ard.