Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <20201114212215.GA1194074@bjorn-Precision-5520>
 <87v9e6n2b2.fsf@x220.int.ebiederm.org> <87sg9almmg.fsf@x220.int.ebiederm.org>
 <874klqac40.fsf@nanos.tec.linutronix.de> <87lff2ic0h.fsf@x220.int.ebiederm.org>
 <c8238524-4197-c22d-7bae-df61133fcbfe@canonical.com> <87a6vhht71.fsf@x220.int.ebiederm.org>
In-Reply-To: <87a6vhht71.fsf@x220.int.ebiederm.org>
From:   Guilherme Piccoli <gpiccoli@canonical.com>
Date:   Mon, 16 Nov 2020 18:49:08 -0300
Message-ID: <CAHD1Q_zk6Z841E1yAH-GKpDXQ=32YPb-R4Exo7UcNHnTY3bdqQ@mail.gmail.com>
Subject: Re: [PATCH 1/3] x86/quirks: Scan all busses for early PCI quirks
To:     "Eric W. Biederman" <ebiederm@xmission.com>
Cc:     Thomas Gleixner <tglx@linutronix.de>,
        Bjorn Helgaas <helgaas@kernel.org>, lukas@wunner.de,
        linux-pci@vger.kernel.org, Pingfan Liu <kernelfans@gmail.com>,
        andi@firstfloor.org, "H. Peter Anvin" <hpa@zytor.com>,
        Baoquan He <bhe@redhat.com>, x86@kernel.org,
        Sinan Kaya <okaya@kernel.org>, Ingo Molnar <mingo@redhat.com>,
        Jay Vosburgh <jay.vosburgh@canonical.com>,
        Dave Young <dyoung@redhat.com>,
        Gavin Guo <gavin.guo@canonical.com>,
        Borislav Petkov <bp@alien8.de>,
        Bjorn Helgaas <bhelgaas@google.com>,
        Guowen Shan <gshan@redhat.com>,
        "Rafael J. Wysocki" <rjw@rjwysocki.net>,
        "Guilherme G. Piccoli" <kernel@gpiccoli.net>,
        kexec mailing list <kexec@lists.infradead.org>,
        LKML <linux-kernel@vger.kernel.org>,
        Dan Streetman <ddstreet@canonical.com>,
        Vivek Goyal <vgoyal@redhat.com>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Mon, Nov 16, 2020 at 6:45 PM Eric W. Biederman <ebiederm@xmission.com> wrote:
> The way to do that would be to collect the set of pci devices when the
> kexec on panic kernel is loaded, not during crash_kexec.  If someone
> performs a device hotplug they would need to reload the kexec on panic
> kernel.
>
> I am not necessarily endorsing that just pointing out how it can be
> done.
>
> Eric

Thanks Eric, I agree! I think if we use something like PKRAM (a more
dynamic approach) we could have the PCI hotplug path updating the data
to-be-passed to the crash kernel, so the crash kernel doesn't even
need to be loaded again.