Re: BUG: optimized kprobes illegal instructions in v4.19 stable kernels

From: Mathieu Desnoyers
Date: Fri Feb 22 2019 - 15:18:51 EST


----- On Feb 22, 2019, at 3:29 AM, Greg Kroah-Hartman gregkh@xxxxxxxxxxxxxxxxxxx wrote:

> On Fri, Feb 22, 2019 at 07:25:50AM +0100, Greg Kroah-Hartman wrote:
>> On Thu, Feb 21, 2019 at 07:17:11PM -0500, Mathieu Desnoyers wrote:
>> > ----- On Feb 21, 2019, at 7:10 PM, Russell King, ARM Linux linux@xxxxxxxxxxxxxxx
>> > wrote:
>> >
>> > > On Thu, Feb 21, 2019 at 03:02:57PM -0500, Mathieu Desnoyers wrote:
>> > >> Hi Arnd, Russell, Linus,
>> > >>
>> > >> Can we ensure the arm32 kprobes fix I submitted gets upstream before 5.0 final ?
>> > >> It takes care of an illegal instruction issue with optimized kprobes on arm32.
>> > >>
>> > >> Here is the current state of default kprobes configuration on arm32:
>> > >> using them will trigger illegal instruction OOPS on v5.0-rc7, 4.19.24,
>> > >> v4.14.102.
>> > >>
>> > >> My fix is in "accepted" state in the arm patch tracking system:
>> > >>
>> > >> https://www.arm.linux.org.uk/developer/patches/viewpatch.php?id=8834/1
>> > >>
>> > >> Should I send it directly to Linus as well ?
>> > >
>> > > Accepted means it's in my tree pending to be sent to Linus. It should
>> > > now be in mainline. Have you checked?
>> >
>> > Hrm, why did I not see it earlier today... it's embarrassing.
>> >
>> > It's there indeed, all is good!
>> >
>> > Greg, you should be able to pick it into the stable kernels now.
>>
>> What is the git commit id of it?
>
> Oh nevermind, I see it. I'll queue it up for the next round of kernel
> releases after this one.

Sorry about late reply, it's indeed
commit 0ac569bf6a79 ("ARM: 8834/1: Fix: kprobes: optimized kprobes illegal instruction")

upstream.

Thanks!

Mathieu


--
Mathieu Desnoyers
EfficiOS Inc.
http://www.efficios.com