Re: usercopy whitelist woe in scsi_sense_cache

From: Kees Cook
Date: Thu Apr 12 2018 - 14:44:45 EST


On Wed, Apr 11, 2018 at 5:03 PM, Kees Cook <keescook@xxxxxxxxxxxx> wrote:
> On Wed, Apr 11, 2018 at 3:47 PM, Kees Cook <keescook@xxxxxxxxxxxx> wrote:
>> On Tue, Apr 10, 2018 at 8:13 PM, Kees Cook <keescook@xxxxxxxxxxxx> wrote:
>>> I'll see about booting with my own kernels, etc, and try to narrow this down. :)
>>
>> If I boot kernels I've built, I no longer hit the bug in this VM
>> (though I'll keep trying). What compiler are you using?
>
> Ignore that: I've reproduced it with my kernels now. I think I messed
> up the initramfs initially. But with an exact copy of your .config,
> booting under Arch grub with initramfs, I see it. I'll start removing
> variables now... :P

My first bisect attempt gave me commit 5448aca41cd5 ("null_blk: wire
up timeouts"), which seems insane given that null_blk isn't even built
in the .config. I managed to get the testing automated now for a "git
bisect run ...", so I'm restarting again to hopefully get a better
answer. Normally the Oops happens in the first minute of running. I've
set the timeout to 10 minutes for a "good" run...

-Kees

--
Kees Cook
Pixel Security