linux-kernel - Re: [Intel-gfx] [PATCH v3 4/4] PCI/VGA: introduce is_boot_device function callback to vga_client

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]

Date:   Fri, 9 Jun 2023 11:48:50 -0500
From:   Bjorn Helgaas <helgaas@...nel.org>
To:     Sui Jingfeng <suijingfeng@...ngson.cn>
Cc:     Sui Jingfeng <15330273260@....cn>,
        Somalapuram Amaranath <Amaranath.Somalapuram@....com>,
        Karol Herbst <kherbst@...hat.com>,
        nouveau@...ts.freedesktop.org,
        Joonas Lahtinen <joonas.lahtinen@...ux.intel.com>,
        dri-devel@...ts.freedesktop.org, YiPeng Chai <YiPeng.Chai@....com>,
        Mario Limonciello <mario.limonciello@....com>,
        David Airlie <airlied@...il.com>,
        Ville Syrjala <ville.syrjala@...ux.intel.com>,
        Yi Liu <yi.l.liu@...el.com>, kvm@...r.kernel.org,
        amd-gfx@...ts.freedesktop.org, Jason Gunthorpe <jgg@...pe.ca>,
        Ben Skeggs <bskeggs@...hat.com>, linux-pci@...r.kernel.org,
        Andrey Grodzovsky <andrey.grodzovsky@....com>,
        Kevin Tian <kevin.tian@...el.com>,
        Lijo Lazar <lijo.lazar@....com>,
        Daniel Vetter <daniel@...ll.ch>,
        Bokun Zhang <Bokun.Zhang@....com>,
        intel-gfx@...ts.freedesktop.org,
        Maarten Lankhorst <maarten.lankhorst@...ux.intel.com>,
        Maxime Ripard <mripard@...nel.org>,
        loongson-kernel@...ts.loongnix.cn,
        Alex Williamson <alex.williamson@...hat.com>,
        Abhishek Sahu <abhsahu@...dia.com>,
        Jani Nikula <jani.nikula@...ux.intel.com>,
        Rodrigo Vivi <rodrigo.vivi@...el.com>,
        Bjorn Helgaas <bhelgaas@...gle.com>,
        Tvrtko Ursulin <tvrtko.ursulin@...ux.intel.com>,
        Yishai Hadas <yishaih@...dia.com>, Li Yi <liyi@...ngson.cn>,
        Pan Xinhui <Xinhui.Pan@....com>, linux-kernel@...r.kernel.org,
        Thomas Zimmermann <tzimmermann@...e.de>,
        Cornelia Huck <cohuck@...hat.com>,
        Alex Deucher <alexander.deucher@....com>,
        Christian Konig <christian.koenig@....com>,
        Hawking Zhang <Hawking.Zhang@....com>
Subject: Re: [Intel-gfx] [PATCH v3 4/4] PCI/VGA: introduce is_boot_device
 function callback to vga_client_register

On Fri, Jun 09, 2023 at 10:27:39AM +0800, Sui Jingfeng wrote:
> On 2023/6/9 03:19, Bjorn Helgaas wrote:
> > On Thu, Jun 08, 2023 at 07:43:22PM +0800, Sui Jingfeng wrote:
> > > From: Sui Jingfeng <suijingfeng@...ngson.cn>
> > > 
> > > The vga_is_firmware_default() function is arch-dependent, which doesn't
> > > sound right. At least, it also works on the Mips and LoongArch platforms.
> > > Tested with the drm/amdgpu and drm/radeon drivers. However, it's difficult
> > > to enumerate all arch-driver combinations. I'm wrong if there is only one
> > > exception.
> > > 
> > > With the observation that device drivers typically have better knowledge
> > > about which PCI bar contains the firmware framebuffer, which could avoid
> > > the need to iterate all of the PCI BARs.
> > > 
> > > But as a PCI function at pci/vgaarb.c, vga_is_firmware_default() is
> > > probably not suitable to make such an optimization for a specific device.
> > > 
> > > There are PCI display controllers that don't have a dedicated VRAM bar,
> > > this function will lose its effectiveness in such a case. Luckily, the
> > > device driver can provide an accurate workaround.
> > > 
> > > Therefore, this patch introduces a callback that allows the device driver
> > > to tell the VGAARB if the device is the default boot device. This patch
> > > only intends to introduce the mechanism, while the implementation is left
> > > to the device driver authors. Also honor the comment: "Clients have two
> > > callback mechanisms they can use"
> > s/bar/BAR/ (several)
> > 
> > Nothing here uses the callback.  I don't want to merge this until we
> > have a user.
> 
> This is chicken and egg question.
> 
> If you could help get this merge first, I will show you the first user.
> 
> > I'm not sure why the device driver should know whether its device is
> > the default boot device.
> 
> It's not that the device driver should know,
> 
> but it's about that the device driver has the right to override.
> 
> Device driver may have better approach to identify the default boot
> device.

The way we usually handle this is to include the new callback in the
same series as the first user of it.  That has two benefits:
(1) everybody can review the whole picture and possibly suggest
different approaches, and (2) when we merge the infrastructure,
we also merge a user of it at the same time, so the whole thing can be
tested and we don't end up with unused code.

Bjorn