linux-kernel - RE: Re: "oneshot" interrupt causes another interrupt to be fired erroneously in Haswell system

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 for Android: free password hash cracker in your pocket

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

Message-ID: <MN2PR04MB6255F9C4B0449BA0CE36CFA2C34E0@MN2PR04MB6255.namprd04.prod.outlook.com>
Date:   Thu, 21 Nov 2019 11:22:19 +0000
From:   Kar Hin Ong <kar.hin.ong@...com>
To:     Thomas Gleixner <tglx@...utronix.de>,
        Bjorn Helgaas <helgaas@...nel.org>
CC:     linux-rt-users <linux-rt-users@...r.kernel.org>,
        LKML <linux-kernel@...r.kernel.org>,
        "linux-x86_64@...r.kernel.org" <linux-x86_64@...r.kernel.org>,
        "linux-pci@...r.kernel.org" <linux-pci@...r.kernel.org>,
        "H. Peter Anvin" <hpa@...or.com>,
        Dave Hansen <dave.hansen@...ux.intel.com>,
        Julia Cartwright <julia.cartwright@...com>,
        Keng Soon Cheah <keng.soon.cheah@...com>
Subject: RE: Re: "oneshot" interrupt causes another interrupt to be fired
 erroneously in Haswell system

> > The workarounds for this are enabled by PCI quirls and either
> > CONFIG_X86_REROUTE_FOR_BROKEN_BOOT_IRQS=y or 'ioapicreroute' on
> the
> > command line.
> >
> > It might be wortha try to add the PCI ID of that box to the quirk
> > list, i.e. the PCI ID matches in drivers/pci/quirks.c which belong to
> > the
> > function: quirk_reroute_to_boot_interrupts_intel().
> 
> Do you mean adding the PCI ID of the device that actually fires interrupt? It
> can be any PCI card though (example: external ETH controller, data
> acquisition module, etc).
> Or you mean to add the ID of all PCIe root ports that routed to IOAPIC 2?
> 
> Based on Haswell specification, seems like every entry on IOAPIC 2 will
> experience this problem.
> If to reroute every entry on IOAPIC 2 to IOAPIC 1, probably we should just
> disable IOAPIC 2 instead?

Hi Thomas, Bjorn, 
Any thoughts on this?

Thanks.
Kar Hin Ong