[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <6B00421C-8B00-45FB-BE54-EED082494D8C@zytor.com>
Date: Tue, 11 May 2021 08:55:24 -0700
From: "H. Peter Anvin" <hpa@...or.com>
To: Thomas Gleixner <tglx@...utronix.de>,
Ingo Molnar <mingo@...hat.com>, Borislav Petkov <bp@...en8.de>,
Andy Lutomirski <luto@...nel.org>
CC: Steve Wahl <steve.wahl@....com>, Mike Travis <mike.travis@....com>,
Dimitri Sivanich <dimitri.sivanich@....com>,
Russ Anderson <russ.anderson@....com>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 5/6] x86/irq: WARN_ONCE() if irq_move_cleanup is called on a pending interrupt
You are, of course, correct – or 2 and 3 can be reversed, I believe.
On May 11, 2021 7:23:53 AM PDT, Thomas Gleixner <tglx@...utronix.de> wrote:
>On Mon, May 10 2021 at 17:55, H. Peter Anvin wrote:
>> From: "H. Peter Anvin (Intel)" <hpa@...or.com>
>>
>> The current IRQ vector allocation code should be "clean" and never
>> issue a IRQ_MOVE_CLEANUP_VECTOR IPI for an interrupt that could still
>> be pending. This should make it possible to move it to the "normal"
>> system IRQ vector range. This should probably be a three-step
>process:
>>
>> 1. Introduce this WARN_ONCE() on this event ever occurring.
>> 2. Move the IRQ_MOVE_CLEANUP_VECTOR to the sysvec range.
>> 3. Remove the self-IPI hack.
>
>Actually 2+3 must be combined because _if_ this ever happens then the
>self-IPI loops forever.
>
>Thanks,
>
> tglx
--
Sent from my Android device with K-9 Mail. Please excuse my brevity.
Powered by blists - more mailing lists