On 4/28/21 12:21 AM, Changheun Lee wrote:

quoted

Actually I don't know why NULL pointer dereference is occurred with Bart's
patch in blk_rq_map_kern(). And same problem have not occured yet in my
test environment with Bart's patch.
Maybe I missed something, or missunderstood?

The call trace that I shared on the linux-block mailing list was
encountered inside a VM that has a SATA disk attached to it and also a
SATA CD-ROM. Does replicating that setup allow to reproduce the NULL
pointer dereference that I reported?

Bart.

I can reproduce and see NULL pointer dereference in same call path without
your patch. But I can't reproduced it with your patch.
I tested in ubuntu via grub. SATA disk is attached only in my environment.
Actually I didn't test in VM yet. I'll try.

Thanks,

Changheun.