linux-kernel - Re: Legacy PCI interrupt support in PCIe host driver

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]

Date:   Thu, 16 Mar 2017 18:47:33 +0100 (CET)
From:   Thomas Gleixner <tglx@...utronix.de>
To:     Mason <slash.tmp@...e.fr>
cc:     LKML <linux-kernel@...r.kernel.org>,
        Linux ARM <linux-arm-kernel@...ts.infradead.org>,
        Marc Zyngier <marc.zyngier@....com>,
        Jason Cooper <jason@...edaemon.net>,
        linux-pci <linux-pci@...r.kernel.org>,
        Bjorn Helgaas <helgaas@...nel.org>,
        Thibaud Cornic <thibaud_cornic@...madesigns.com>,
        Phuong Nguyen <phuong_nguyen@...madesigns.com>,
        Robin Murphy <robin.murphy@....com>,
        Liviu Dudau <liviu.dudau@....com>,
        Lorenzo Pieralisi <lorenzo.pieralisi@....com>,
        Uwe Kleine-Konig <u.kleine-koenig@...gutronix.de>,
        Rob Herring <robh@...nel.org>
Subject: Re: Legacy PCI interrupt support in PCIe host driver

On Thu, 16 Mar 2017, Mason wrote:
> About shared ISRs. Are they supposed to return IRQ_HANDLED
> if and only if they handled something?

Every interrupt handler is supposed to return IRQ_NONE if it did not handle
it. That does not depend on shared or not. The handler does not know if
it's on a shared interrupt line or not. IRQF_SHARED only tells, that the
handler is capable of sharing the interrupt line with another device.

> Will that stop the next ISR from being called?

No.

> I guess if two interrupts fire at the same time, we'll just take two
> separate exceptions?

Wrong guess. That might work with level interupts, but with other types
nothing will raise another exception. Sharing interrupts on edge types is a
stupid idea, but hardware folks insist on implementing stupid ideas.

Thanks,

	tglx