linux-kernel - Re: [PATCH 1/3] x86/quirks: Scan all busses for early PCI quirks

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: <20201130202021.GA1106292@bjorn-Precision-5520>
Date:   Mon, 30 Nov 2020 14:20:21 -0600
From:   Bjorn Helgaas <helgaas@...nel.org>
To:     Guilherme Piccoli <gpiccoli@...onical.com>
Cc:     "Eric W. Biederman" <ebiederm@...ssion.com>,
        Thomas Gleixner <tglx@...utronix.de>, lukas@...ner.de,
        linux-pci@...r.kernel.org, Pingfan Liu <kernelfans@...il.com>,
        andi@...stfloor.org, "H. Peter Anvin" <hpa@...or.com>,
        Baoquan He <bhe@...hat.com>, x86@...nel.org,
        Sinan Kaya <okaya@...nel.org>, Ingo Molnar <mingo@...hat.com>,
        Jay Vosburgh <jay.vosburgh@...onical.com>,
        Dave Young <dyoung@...hat.com>,
        Gavin Guo <gavin.guo@...onical.com>,
        Borislav Petkov <bp@...en8.de>,
        Bjorn Helgaas <bhelgaas@...gle.com>,
        Guowen Shan <gshan@...hat.com>,
        "Rafael J. Wysocki" <rjw@...ysocki.net>,
        "Guilherme G. Piccoli" <kernel@...ccoli.net>,
        kexec mailing list <kexec@...ts.infradead.org>,
        LKML <linux-kernel@...r.kernel.org>,
        Dan Streetman <ddstreet@...onical.com>,
        Vivek Goyal <vgoyal@...hat.com>
Subject: Re: [PATCH 1/3] x86/quirks: Scan all busses for early PCI quirks

On Wed, Nov 18, 2020 at 07:36:08PM -0300, Guilherme Piccoli wrote:
> Thanks a lot Bjorn! I confess except for PPC64 Server machines, I
> never saw other "domains" or segments. Is it common in x86 to have
> that? The early_quirks() are restricted to the first segment, no
> matter how many host bridges we have in segment 0000?

I don't know whether it's *common* to have multiple domains on x86,
but they're definitely used on large systems.  This includes some
lspci info from an HPE Superdome Flex system:
https://lore.kernel.org/lkml/5350E02A-6457-41A8-8F33-AF67BFDAEE3E@fb.com/

The early quirks in arch/x86/kernel/early-quirks.c (not the
DECLARE_PCI_FIXUP_EARLY quirks in drivers/pci/quirks.c) are restricted
to segment 0, no matter how many host bridges there are.  This is
because they use read_pci_config_16(), which uses a config access
mechanism that has no provision for a segment number.

Bjorn