Re: [PATCH 0/6] IRQ handling patches backport to 4.14 stable

["Bhatnagar, Rishabh" <risbhat@amazon.com>]

On 10/6/22 8:07 PM, Herrenschmidt, Benjamin wrote:
> (putting my @amazon.com hat on)
>
> On Sun, 2022-10-02 at 17:30 +0200, Greg KH wrote:
>
>
>> On Thu, Sep 29, 2022 at 09:06:45PM +0000, Rishabh Bhatnagar wrote:
>>> This patch series backports a bunch of patches related IRQ handling
>>> with respect to freeing the irq line while IRQ is in flight at CPU
>>> or at the hardware level.
>>> Recently we saw this issue in serial 8250 driver where the IRQ was
>>> being
>>> freed while the irq was in flight or not yet delivered to the CPU.
>>> As a
>>> result the irqchip was going into a wedged state and IRQ was not
>>> getting
>>> delivered to the cpu. These patches helped fixed the issue in 4.14
>>> kernel.
>> Why is the serial driver freeing an irq while the system is running?
>> Ah, this could happen on a tty hangup, right?
> Right. Rishabh answered that separately.
>
>>> Let us know if more patches need backporting.
>> What hardware platform were these patches tested on to verify they
>> work properly?  And why can't they move to 4.19 or newer if they
>> really need this fix?  What's preventing that?
>>
>> As Amazon doesn't seem to be testing 4.14.y -rc releases, I find it
>> odd that you all did this backport.  Is this a kernel that you all
>> care about?
> These were tested on a collection of EC2 instances, virtual and metal I
> believe (Rishabh, please confirm).
Yes these patches were tested on multiple virt/metal EC2 instances.
>
> Amazon Linux 2 runs 4.14 or 5.10. Unfortunately we still have to
> support customers running the former.
>
> We'll be including these patches in our releases, we thought it would
> be nice to have them in -stable as well for the sake of whoever else
> might be still using this kernel. No huge deal if they don't.
>
> As for testing -rc's, yes, we need to get better at that (and publish
> what we test). Point taken :-)
>
> Cheers,
> Ben.
>