[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <878rbuss4s.fsf@minerva.mail-host-address-is-not-set>
Date: Wed, 05 Jul 2023 10:02:27 +0200
From: Javier Martinez Canillas <javierm@...hat.com>
To: Thomas Zimmermann <tzimmermann@...e.de>, arnd@...db.de,
deller@....de, daniel@...ll.ch, airlied@...il.com
Cc: linux-hyperv@...r.kernel.org, linux-efi@...r.kernel.org,
linux-ia64@...r.kernel.org, linux-sh@...r.kernel.org,
linux-fbdev@...r.kernel.org, dri-devel@...ts.freedesktop.org,
linux-mips@...r.kernel.org, sparclinux@...r.kernel.org,
linux-riscv@...ts.infradead.org, linux-arch@...r.kernel.org,
linux-hexagon@...r.kernel.org, linux-staging@...ts.linux.dev,
linux-csky@...r.kernel.org, Antonino Daplas <adaplas@...il.com>,
Maik Broemme <mbroemme@...mpq.org>, loongarch@...ts.linux.dev,
linux-arm-kernel@...ts.infradead.org,
Randy Dunlap <rdunlap@...radead.org>,
linux-kernel@...r.kernel.org,
Thomas Zimmermann <tzimmermann@...e.de>,
linux-alpha@...r.kernel.org, linuxppc-dev@...ts.ozlabs.org
Subject: Re: [PATCH 10/12] fbdev/core: Use fb_is_primary_device() in
fb_firmware_edid()
Thomas Zimmermann <tzimmermann@...e.de> writes:
> Detect the primary VGA device with fb_is_primary_device() in the
> implementation of fb_firmware_edid(). Remove the existing code.
>
An explanation about why this is possible would be useful here.
> Adapt the function to receive an instance of struct fb_info and
> update all callers.
>
[...]
> -const unsigned char *fb_firmware_edid(struct device *device)
> +const unsigned char *fb_firmware_edid(struct fb_info *info)
> {
> - struct pci_dev *dev = NULL;
> - struct resource *res = NULL;
> unsigned char *edid = NULL;
>
> - if (device)
> - dev = to_pci_dev(device);
> -
> - if (dev)
> - res = &dev->resource[PCI_ROM_RESOURCE];
> -
> - if (res && res->flags & IORESOURCE_ROM_SHADOW)
This open codes what used to be the fb_is_primary_device() logic before
commit 5ca1479cd35d ("fbdev: Simplify fb_is_primary_device for x86").
But now after that commit there is functional change since the ROM
shadowing check would be dropped.
I believe that's OK and Sima explains in their commit message that
vga_default_device() should be enough and the check is redundant.
Still, I think that this change should be documented in your commit
message.
With that change,
Reviewed-by: Javier Martinez Canillas <javierm@...hat.com>
--
Best regards,
Javier Martinez Canillas
Core Platforms
Red Hat
Powered by blists - more mailing lists