linux-kernel - Re: [RT] MPIC edge sensitive issues with hardirq preemption

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <486A09CF.1040405@ru.mvista.com>
Date:	Tue, 01 Jul 2008 14:41:19 +0400
From:	Sergei Shtylyov <sshtylyov@...mvista.com>
To:	avorontsov@...mvista.com
Cc:	Benjamin Herrenschmidt <benh@...nel.crashing.org>,
	Ingo Molnar <mingo@...e.hu>,
	Bartlomiej Zolnierkiewicz <bzolnier@...il.com>,
	Alan Cox <alan@...rguk.ukuu.org.uk>,
	linux-kernel@...r.kernel.org, Thomas Gleixner <tglx@...utronix.de>,
	Steven Rostedt <rostedt@...dmis.org>,
	Daniel Walker <dwalker@...sta.com>
Subject: Re: [RT] MPIC edge sensitive issues with hardirq preemption

Hello.

Anton Vorontsov wrote:

> But how this could be a bug in the PIC code? IMO this is a bug in the
> kernel/irq code, since it assumes that fasteoi PIC will retrigger masked
> edge sources... This isn't true for at least MPIC. To make this work for
> all fasteoi PICs, we should mask edge sensitive interrupts very very
> carefully.
>   

   I guess it assumed this based on 8259's behavior (not sure about I/O 
APIC).
   Hm, but the 8259 code never used "fasteoi" path for some obscure 
reason...

> jammed with the idea that MPIC irq type 0 is low level sensitive, but the
> true thing is that it is rising edge sensitive. (Ah, I know where I got
> confused, type 0 is active-low for ISA PICs).
>   

    You mean in the device tree?

> So in all my previous emails I was wrong when I was saying "mpic is
> programmed to low level sensitive". It was programmed for rising edge
> sensitive. An all my further reasonings were flawed because of this.
>   

   Gah.  I'm surprised how it could work at all then...

> Re-programming MPIC to high level sensitive also makes IDE work. But
> this doesn't mean that IRQ code is correct.
>   

   I wonder why. :-O
   Your ULi IDE is in native mode, so it should be generating a PCI 
interrupt -- which is *low* level sensitive.

MBR, Sergei


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/