Re: [PATCH 2/2] Revert "splice: Do splice read from a buffered file without using ITER_PIPE"

From: David Hildenbrand
Date: Thu Feb 16 2023 - 05:58:05 EST


On 16.02.23 10:10, Konrad Dybcio wrote:


On 15.02.2023 14:01, David Howells wrote:
Konrad Dybcio <konrad.dybcio@xxxxxxxxxx> wrote:

next-20230213 introduced commit d9722a475711 ("splice: Do splice read from
a buffered file without using ITER_PIPE") which broke booting on any
Qualcomm ARM64 device I grabbed, dereferencing a null pointer in
generic_filesplice_read+0xf8/x598. Revert it to make the devices
bootable again.

This reverts commit d9722a47571104f7fa1eeb5ec59044d3607c6070.
Signed-off-by: Konrad Dybcio <konrad.dybcio@xxxxxxxxxx>

Commit d9722a47571104f7fa1eeb5ec59044d3607c6070 was part of v13 of my
patches. This got replaced yesterday by a newer version which may or may not
have made it into linux-next.

This is probably a known bug fixed in the v14 by making shmem have its own
splice-read function.

Can you try this?

https://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs.git/log/?h=iov-extract
next-20230216 boots fine again, thanks!


(Also, can you include me in the cc list as I'm the author of the patch you
reverted?)
Ugh.. I thought b4 would have done that for me.. weird..

Right, and usually it's nicer to comment on the problematic patches, asking for a fix or a revert, instead of sending reverts.

My 2 cents.

--
Thanks,

David / dhildenb