[PATCH] PCI/VGA: Fix boot device detection after framebuffer relocation

[PATCH] PCI/VGA: Fix boot device detection after framebuffer relocation

[wuqianhai]

From: wuqianhai <wuqianhai@loongson.cn>

PCI resource assignment can relocate the firmware framebuffer BAR, which
leaves the base address in screen_info stale.  The VGA arbiter then calls
screen_info_pci_dev() from vga_is_firmware_default(), which fails to match
the framebuffer against any PCI device's BARs.  As a result, the boot VGA
device is not recognized and the arbiter may select the wrong default
device.

Fix this by calling screen_info_apply_fixups() in vga_arb_device_init()
before adding the VGA devices.  This relocates lfb_base to the assigned
BAR address so that vga_is_firmware_default() matches the boot device
again.

Tested on a LoongArch platform, with vga_is_firmware_default()
additionally enabled for CONFIG_LOONGARCH via local changes.

Fixes: 337bf13aa9ddab8ae47f2677ab75cf69ae33437b ("PCI/VGA: Replace vga_is_firmware_default() with a screen info check")
Signed-off-by: wuqianhai <wuqianhai@loongson.cn>
---
 drivers/pci/vgaarb.c | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/drivers/pci/vgaarb.c b/drivers/pci/vgaarb.c
index 3e58b2c05c85..aca594651f58 100644
--- a/drivers/pci/vgaarb.c
+++ b/drivers/pci/vgaarb.c
@@ -1527,6 +1527,14 @@ static int __init vga_arb_device_init(void)
 	int rc;
 	struct pci_dev *pdev;
 
+	/*
+	 * PCI resource assignment may have relocated the firmware
+	 * framebuffer BAR. Update screen_info's lfb_base so that
+	 * vga_is_firmware_default() can match it against the boot
+	 * device's relocated resource.
+	 */
+	screen_info_apply_fixups();
+
 	rc = misc_register(&vga_arb_device);
 	if (rc < 0)
 		pr_err("error %d registering device\n", rc);
-- 
2.34.1

Re: [PATCH] PCI/VGA: Fix boot device detection after framebuffer relocation

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
- [High] PCI/VGA: premature and duplicate screen_info_apply_fixups() in vga_arb_device_init()
- [Medium] PCI/VGA: undefined reference to screen_info_apply_fixups() when !SCREEN_INFO
--

commit da40d2d05299ef123ed4789d2ecb58829165fbe3
Author: wuqianhai <wuqianhai@loongson.cn>

PCI/VGA: Fix boot device detection after framebuffer relocation

This commit calls screen_info_apply_fixups() in vga_arb_device_init() so that
screen_info's lfb_base is updated before VGA devices are registered and
checked by vga_is_firmware_default().

> diff --git a/drivers/pci/vgaarb.c b/drivers/pci/vgaarb.c
> index 3e58b2c05c854..aca594651f580 100644
> --- a/drivers/pci/vgaarb.c
> +++ b/drivers/pci/vgaarb.c
> @@ -1527,6 +1527,14 @@ static int __init vga_arb_device_init(void)
>  	int rc;
>  	struct pci_dev *pdev;
>  
> +	/*
> +	 * PCI resource assignment may have relocated the firmware
> +	 * framebuffer BAR. Update screen_info's lfb_base so that
> +	 * vga_is_firmware_default() can match it against the boot
> +	 * device's relocated resource.
> +	 */
> +	screen_info_apply_fixups();

[Severity: Medium]
Will calling screen_info_apply_fixups() unconditionally in
vga_arb_device_init() fail to link when CONFIG_PCI=y, CONFIG_VGA_ARB=y, and
CONFIG_SCREEN_INFO=n?

In drivers/pci/Kconfig, VGA_ARB only selects SCREEN_INFO on x86:

drivers/pci/Kconfig:
	config VGA_ARB
		bool "VGA Arbitration" if EXPERT
		default y
		depends on !S390
		select SCREEN_INFO if X86

which matches the CONFIG_X86 guard in vga_is_firmware_default():

drivers/pci/vgaarb.c:vga_is_firmware_default() {
#if defined CONFIG_X86
	return pdev == screen_info_pci_dev(&sysfb_primary_display.screen);
#else
	return false;
#endif
}

However, include/linux/screen_info.h declares screen_info_apply_fixups() as an
external function whenever CONFIG_PCI is enabled, while
drivers/video/Makefile only builds screen_info_pci.o when CONFIG_SCREEN_INFO
is enabled:

include/linux/screen_info.h:
#if defined(CONFIG_PCI)
void screen_info_apply_fixups(void);
struct pci_dev *screen_info_pci_dev(const struct screen_info *si);
#else
static inline void screen_info_apply_fixups(void)
{ }

drivers/video/Makefile:
obj-$(CONFIG_SCREEN_INFO)         += screen_info.o
...
screen_info-$(CONFIG_PCI)         += screen_info_pci.o

On non-x86 architectures with CONFIG_PCI=y, CONFIG_VGA_ARB=y, and
CONFIG_SCREEN_INFO=n, doesn't this result in an undefined reference to
screen_info_apply_fixups()?

[Severity: High]
Also, is subsys_initcall_sync(vga_arb_device_init) the right place in the
initcall sequence to call screen_info_apply_fixups()?

Upstream vga_is_firmware_default() in drivers/pci/vgaarb.c only calls
screen_info_pci_dev() under CONFIG_X86. On x86, subsys_initcall(pci_subsys_init)
runs pcibios_resource_survey() -> pcibios_allocate_dev_resources(). When the
firmware framebuffer BAR has a resource conflict and awaits reassignment,
pcibios_allocate_dev_resources() saves the firmware address and resets
r->start to 0 while keeping resource_size(r) intact:

arch/x86/pci/i386.c:pcibios_allocate_dev_resources() {
	...
	/* We'll assign a new address later */
	pcibios_save_fw_addr(dev, idx, r->start);
	r->end -= r->start;
	r->start = 0;
	...
}

The new BAR address on x86 is not assigned until
fs_initcall(pcibios_assign_resources) -> pci_assign_unassigned_resources(),
which runs after subsys_initcall_sync(vga_arb_device_init).

When screen_info_apply_fixups() runs at subsys_initcall_sync while a
conflicting BAR has pr->start == 0, pr->start != screen_info_lfb_res_start is
true and __screen_info_relocation_is_valid(si, pr) succeeds because
resource_size(pr) is unchanged:

drivers/video/screen_info_pci.c:screen_info_apply_fixups() {
	if (screen_info_lfb_pdev) {
		struct resource *pr = &screen_info_lfb_pdev->resource[screen_info_lfb_bar];

		if (pr->start != screen_info_lfb_res_start) {
			if (__screen_info_relocation_is_valid(si, pr)) {
				__screen_info_set_lfb_base(si, pr->start + screen_info_lfb_offset);
				pr_info("Relocating firmware framebuffer to offset %pa[d] within %pr\n",
					&screen_info_lfb_offset, pr);
			}
			...
		}
	}
}

Wouldn't this clobber si->lfb_base to 0 + screen_info_lfb_offset with a
[mem 0x00000000-...] message before fs_initcall(pcibios_assign_resources)
assigns the actual BAR address, while still failing to match in
vga_is_firmware_default()?

Similarly, for PCI host bridges probed at device_initcall via
pci_host_probe(), vga_arb_device_init() runs before the PCI bus is scanned
and before VGA devices are added via pci_notify().

Finally, since screen_info_apply_fixups() does not update
screen_info_lfb_res_start or clear screen_info_lfb_pdev after applying the
fixup, won't calling it in both vga_arb_device_init() and
device_initcall(sysfb_init) in drivers/firmware/sysfb.c cause the relocation
and log messages to run twice?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261028094634.766723-1-wuqianhai@loongson.cn?part=1

Re: [PATCH] PCI/VGA: Fix boot device detection after framebuffer relocation

[Bjorn Helgaas]

On Wed, Oct 28, 2026 at 05:46:34PM +0800, wuqianhai@loongson.cn wrote:
> From: wuqianhai <wuqianhai@loongson.cn>
> 
> PCI resource assignment can relocate the firmware framebuffer BAR, which
> leaves the base address in screen_info stale.  The VGA arbiter then calls
> screen_info_pci_dev() from vga_is_firmware_default(), which fails to match
> the framebuffer against any PCI device's BARs.  As a result, the boot VGA
> device is not recognized and the arbiter may select the wrong default
> device.
> 
> Fix this by calling screen_info_apply_fixups() in vga_arb_device_init()
> before adding the VGA devices.  This relocates lfb_base to the assigned
> BAR address so that vga_is_firmware_default() matches the boot device
> again.
> 
> Tested on a LoongArch platform, with vga_is_firmware_default()
> additionally enabled for CONFIG_LOONGARCH via local changes.
> 
> Fixes: 337bf13aa9ddab8ae47f2677ab75cf69ae33437b ("PCI/VGA: Replace vga_is_firmware_default() with a screen info check")
> Signed-off-by: wuqianhai <wuqianhai@loongson.cn>
> ---
>  drivers/pci/vgaarb.c | 8 ++++++++
>  1 file changed, 8 insertions(+)
> 
> diff --git a/drivers/pci/vgaarb.c b/drivers/pci/vgaarb.c
> index 3e58b2c05c85..aca594651f58 100644
> --- a/drivers/pci/vgaarb.c
> +++ b/drivers/pci/vgaarb.c
> @@ -1527,6 +1527,14 @@ static int __init vga_arb_device_init(void)
>  	int rc;
>  	struct pci_dev *pdev;
>  
> +	/*
> +	 * PCI resource assignment may have relocated the firmware
> +	 * framebuffer BAR. Update screen_info's lfb_base so that
> +	 * vga_is_firmware_default() can match it against the boot
> +	 * device's relocated resource.
> +	 */
> +	screen_info_apply_fixups();

I'm really confused about this.  vga_is_firmware_default() relies on
screen_info_pci_dev(&sysfb_primary_display.screen):

  vga_is_firmware_default
    screen_info_pci_dev(&sysfb_primary_display.screen)
      numres = screen_info_resources(&sysfb_primary_display.screen, res, ...)
      for (i = 0; i < numres; ++i)
        pdev = __screen_info_pci_dev(&res[i])

screen_info_resources() returns resource structs with PCI bus
addresses in them.  These are not valid CPU addresses, so that seems
broken on any system where bus addresses are not the same as CPU
addresses.

screen_info_fixup_lfb() and screen_info_apply_fixups() may fix the
issue you see, but I think the way this happens is questionable.

screen_info_fixup_lfb() captures the framebuffer PCI bus address in
screen_info_lfb_res_start.  Then screen_info_apply_fixups() compares
that PCI bus address with the CPU address in pdev->resource[].  That's
broken on any system where bus addresses are not the same as CPU
addresses.  I think it's also broken because screen_info_lfb_res_start
may not be at the beginning of the BAR.

I think screen_info_fixup_lfb() should only capture the pci_dev, the
BAR number, and the framebuffer offset into the BAR.  When we need the
CPU or bus address, we can always compute it from those values.

>  	rc = misc_register(&vga_arb_device);
>  	if (rc < 0)
>  		pr_err("error %d registering device\n", rc);
> -- 
> 2.34.1
> 

Re: [PATCH] PCI/VGA: Fix boot device detection after framebuffer relocation

[wuqianhai]

Hi Bjorn,

Thanks for the detailed analysis.

You're right that screen_info_fixup_lfb()/screen_info_apply_fixups()
stores a PCI bus address in screen_info_lfb_res_start and compares it
with the CPU address in pdev->resource[].start, which is broken on
systems where bus and CPU addresses differ.

On LoongArch, though, bus and CPU addresses are identical: with modern
firmware, acpi_prepare_root_resources() takes the PCIH path and leaves
the window offset at 0 (see commit 49f33840dcc9).  So the existing
screen_info_apply_fixups() works correctly for LoongArch, and the
address-space confusion you describe does not trigger there.

I've sent a v2 that:
- enables vga_is_firmware_default() for LoongArch,
- restricts the screen_info_apply_fixups() call to LoongArch so x86 and
  other architectures are unaffected,
- makes VGA_ARB select SYSFB on LoongArch to avoid the undefined
  reference when CONFIG_SCREEN_INFO is not set.

The general bus != CPU problem in screen_info_pci.c is still worth
fixing for ARM and friends, but I'd rather keep that separate from this
LoongArch enablement.  Happy to help with that follow-up if useful.

Regards,
wuqianhai

Re: [PATCH] PCI/VGA: Fix boot device detection after framebuffer relocation

[Bjorn Helgaas]

[+ Loongson folks]

On Thu, Oct 29, 2026 at 11:39:14AM +0800, wuqianhai@loongson.cn wrote:
> Hi Bjorn,
> 
> Thanks for the detailed analysis.
> 
> You're right that screen_info_fixup_lfb()/screen_info_apply_fixups()
> stores a PCI bus address in screen_info_lfb_res_start and compares it
> with the CPU address in pdev->resource[].start, which is broken on
> systems where bus and CPU addresses differ.
> 
> On LoongArch, though, bus and CPU addresses are identical: with modern
> firmware, acpi_prepare_root_resources() takes the PCIH path and leaves
> the window offset at 0 (see commit 49f33840dcc9).  

This is a tangent from the framebuffer question, but what in the world
is 49f33840dcc9 ("LoongArch: Use per-root-bridge PCIH flag to skip mem
resource fixup") doing?  I see that if "PCIH" is non-zero, it skips
some mem resource "fixup" that was added by 57fc7323a8e7 ("LoongArch:
Add PCI controller support").

But I don't understand why that fixup is there in the first place.
Evidently it takes the upper 24 bits of the MCFG address and ORs it
into window addresses from host bridge _CRS?  That sounds nuts.  It
sounds like the ACPI _CRS is just broken.

The struct resource_entry.offset is the offset between PCI bus
addresses and CPU addresses, so entry->offset being non-zero implies
they are NOT identical, at least on non-PCIH platforms.  And if you
have to hack up the offset from the MCFG address, it sounds like the
_CRS entries are missing the _TRA that is supposed to provide the
offset.

Re: [PATCH] PCI/VGA: Fix boot device detection after framebuffer relocation

[Thomas Zimmermann]

Hi

Am 29.10.26 um 00:15 schrieb Bjorn Helgaas:
> On Wed, Oct 28, 2026 at 05:46:34PM +0800, wuqianhai@loongson.cn wrote:
>> From: wuqianhai <wuqianhai@loongson.cn>
>>
>> PCI resource assignment can relocate the firmware framebuffer BAR, which
>> leaves the base address in screen_info stale.  The VGA arbiter then calls
>> screen_info_pci_dev() from vga_is_firmware_default(), which fails to match
>> the framebuffer against any PCI device's BARs.  As a result, the boot VGA
>> device is not recognized and the arbiter may select the wrong default
>> device.
>>
>> Fix this by calling screen_info_apply_fixups() in vga_arb_device_init()
>> before adding the VGA devices.  This relocates lfb_base to the assigned
>> BAR address so that vga_is_firmware_default() matches the boot device
>> again.
>>
>> Tested on a LoongArch platform, with vga_is_firmware_default()
>> additionally enabled for CONFIG_LOONGARCH via local changes.
>>
>> Fixes: 337bf13aa9ddab8ae47f2677ab75cf69ae33437b ("PCI/VGA: Replace vga_is_firmware_default() with a screen info check")
>> Signed-off-by: wuqianhai <wuqianhai@loongson.cn>
>> ---
>>   drivers/pci/vgaarb.c | 8 ++++++++
>>   1 file changed, 8 insertions(+)
>>
>> diff --git a/drivers/pci/vgaarb.c b/drivers/pci/vgaarb.c
>> index 3e58b2c05c85..aca594651f58 100644
>> --- a/drivers/pci/vgaarb.c
>> +++ b/drivers/pci/vgaarb.c
>> @@ -1527,6 +1527,14 @@ static int __init vga_arb_device_init(void)
>>   	int rc;
>>   	struct pci_dev *pdev;
>>   
>> +	/*
>> +	 * PCI resource assignment may have relocated the firmware
>> +	 * framebuffer BAR. Update screen_info's lfb_base so that
>> +	 * vga_is_firmware_default() can match it against the boot
>> +	 * device's relocated resource.
>> +	 */
>> +	screen_info_apply_fixups();
> I'm really confused about this.  vga_is_firmware_default() relies on
> screen_info_pci_dev(&sysfb_primary_display.screen):
>
>    vga_is_firmware_default
>      screen_info_pci_dev(&sysfb_primary_display.screen)
>        numres = screen_info_resources(&sysfb_primary_display.screen, res, ...)
>        for (i = 0; i < numres; ++i)
>          pdev = __screen_info_pci_dev(&res[i])
>
> screen_info_resources() returns resource structs with PCI bus
> addresses in them.  These are not valid CPU addresses, so that seems
> broken on any system where bus addresses are not the same as CPU
> addresses.

What we get from the firmware for the framebuffer is a memory range.  
  If the underlying PCI memory gets relocated during the boot, we need 
to update the framebuffer data accordingly. This has worked well for the 
last 20 yrs.


>
> screen_info_fixup_lfb() and screen_info_apply_fixups() may fix the
> issue you see, but I think the way this happens is questionable.

>
> screen_info_fixup_lfb() captures the framebuffer PCI bus address in
> screen_info_lfb_res_start.  Then screen_info_apply_fixups() compares
> that PCI bus address with the CPU address in pdev->resource[].  That's
> broken on any system where bus addresses are not the same as CPU
> addresses.  I think it's also broken because screen_info_lfb_res_start
> may not be at the beginning of the BAR.
>
> I think screen_info_fixup_lfb() should only capture the pci_dev, the
> BAR number, and the framebuffer offset into the BAR.  When we need the
> CPU or bus address, we can always compute it from those values.

This is not sufficient, I think. We did this until a while ago and got 
problems with relocated bars. IIRC aarch64 had a number of affected 
systems. See commit 2f29b5c23101 ("video: screen_info: Relocate 
framebuffers behind PCI bridges") for details.

I'm happy to improve that code though, if we can do something better.

Best regards
Thomas

>
>>   	rc = misc_register(&vga_arb_device);
>>   	if (rc < 0)
>>   		pr_err("error %d registering device\n", rc);
>> -- 
>> 2.34.1
>>

-- 
--
Thomas Zimmermann
Graphics Driver Developer
SUSE Software Solutions Germany GmbH
Frankenstr. 146, 90461 Nürnberg, Germany, www.suse.com
GF: Stefan Gaiser, Jochen Jaser, Abhinav Puri, (HRB 36809, AG Nürnberg)

[PATCH v2] PCI/VGA: Enable firmware default device detection on LoongArch

[wuqianhai]

From: wuqianhai <wuqianhai@loongson.cn>

On LoongArch, PCI resource assignment completes during subsys_initcall
(acpi_init), before vga_arb_device_init() runs.  By that time the
firmware framebuffer BAR may already have been relocated, leaving
screen_info's lfb_base stale.  Enable vga_is_firmware_default() for
LoongArch and refresh screen_info in vga_arb_device_init() before the
boot device is matched.

Unlike x86, sysfb_primary_display is only defined when CONFIG_SYSFB is
enabled on LoongArch, so have VGA_ARB select SYSFB there.

Signed-off-by: wuqianhai <wuqianhai@loongson.cn>
---
Changes in v2:
- Add VGA_ARB's select of SYSFB for LoongArch to avoid an undefined
  reference when CONFIG_SCREEN_INFO is not set.
- Restrict the screen_info_apply_fixups() call to CONFIG_LOONGARCH so
  x86 and other architectures are unaffected.
- Enable vga_is_firmware_default() for LoongArch.
---
 drivers/pci/Kconfig  |  1 +
 drivers/pci/vgaarb.c | 12 +++++++++++-
 2 files changed, 12 insertions(+), 1 deletion(-)

diff --git a/drivers/pci/Kconfig b/drivers/pci/Kconfig
index 0c7408509ba2..55cc60f5d312 100644
--- a/drivers/pci/Kconfig
+++ b/drivers/pci/Kconfig
@@ -256,6 +256,7 @@ config VGA_ARB
 	default y
 	depends on !S390
 	select SCREEN_INFO if X86
+	select SYSFB if LOONGARCH
 	help
 	  Some "legacy" VGA devices implemented on PCI typically have the same
 	  hard-decoded addresses as they did on ISA. When multiple PCI devices
diff --git a/drivers/pci/vgaarb.c b/drivers/pci/vgaarb.c
index 3e58b2c05c85..1b7e915629f9 100644
--- a/drivers/pci/vgaarb.c
+++ b/drivers/pci/vgaarb.c
@@ -566,7 +566,7 @@ EXPORT_SYMBOL(vga_put);
 
 static bool vga_is_firmware_default(struct pci_dev *pdev)
 {
-#if defined CONFIG_X86
+#if defined(CONFIG_X86) || defined(CONFIG_LOONGARCH)
 	return pdev == screen_info_pci_dev(&sysfb_primary_display.screen);
 #else
 	return false;
@@ -1527,6 +1527,16 @@ static int __init vga_arb_device_init(void)
 	int rc;
 	struct pci_dev *pdev;
 
+#if defined(CONFIG_LOONGARCH)
+	/*
+	 * On LoongArch, PCI resource assignment completes during
+	 * subsys_initcall (acpi_init), before the VGA arbiter runs.  Refresh
+	 * screen_info so vga_is_firmware_default() can match the relocated
+	 * BAR.
+	 */
+	screen_info_apply_fixups();
+#endif
+
 	rc = misc_register(&vga_arb_device);
 	if (rc < 0)
 		pr_err("error %d registering device\n", rc);
-- 
2.34.1

Re: [PATCH v2] PCI/VGA: Enable firmware default device detection on LoongArch

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 3 potential issue(s) to consider:
- [High] PCI/VGA: pci_fixup_vgadev() overrides vga_is_firmware_default() on LoongArch
- [Medium] PCI/VGA: duplicate non-idempotent screen_info_apply_fixups() call on LoongArch
- [Medium] PCI/VGA: vga_is_firmware_default() fails to match Loongson integrated DC
--

commit 6de860408bce3c5a20f5a4c23b073eafdc942117
Author: wuqianhai <wuqianhai@loongson.cn>

PCI/VGA: Enable firmware default device detection on LoongArch

This commit enables vga_is_firmware_default() on LoongArch, selects
CONFIG_SYSFB from CONFIG_VGA_ARB for LoongArch, and calls
screen_info_apply_fixups() in vga_arb_device_init() so relocated firmware
framebuffer BARs can be matched when selecting the default boot VGA device.

> diff --git a/drivers/pci/vgaarb.c b/drivers/pci/vgaarb.c
> index 3e58b2c05c854..1b7e915629f97 100644
> --- a/drivers/pci/vgaarb.c
> +++ b/drivers/pci/vgaarb.c
> @@ -566,7 +566,7 @@ EXPORT_SYMBOL(vga_put);
>  
>  static bool vga_is_firmware_default(struct pci_dev *pdev)
>  {
> -#if defined CONFIG_X86
> +#if defined(CONFIG_X86) || defined(CONFIG_LOONGARCH)
>  	return pdev == screen_info_pci_dev(&sysfb_primary_display.screen);

[Severity: High]
On LoongArch systems with a Loongson display controller (PCI device IDs
0x7a06, 0x7a36, or 0x7a46) and one or more discrete VGA devices, will
pci_fixup_vgadev() in arch/loongarch/pci/pci.c later overwrite the firmware
default VGA device selected here?

During ACPI boot on LoongArch, PCI enumeration runs in
subsys_initcall(acpi_init), and vga_arb_device_init() runs next at
subsys_initcall_sync:

vga_arb_device_init()
  vga_arbiter_add_pci_device()
    vga_is_boot_device()
      vga_is_firmware_default()
    vga_set_default_device(pdev)

After vga_arb_device_init() finishes, pci_apply_final_quirks() in
drivers/pci/quirks.c runs at fs_initcall_sync and invokes the
DECLARE_PCI_FIXUP_FINAL hooks in arch/loongarch/pci/pci.c:

static void pci_fixup_vgadev(struct pci_dev *pdev)
{
	struct pci_dev *devp = NULL;

	while ((devp = pci_get_class(PCI_CLASS_DISPLAY_VGA << 8, devp))) {
		if (devp->vendor != PCI_VENDOR_ID_LOONGSON) {
			vga_set_default_device(devp);
			dev_info(&pdev->dev,
				"Overriding boot device as %X:%X\n",
				devp->vendor, devp->device);
		}
	}
}
DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_LOONGSON, PCI_DEVICE_ID_LOONGSON_DC1, pci_fixup_vgadev);
DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_LOONGSON, PCI_DEVICE_ID_LOONGSON_DC2, pci_fixup_vgadev);
DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_LOONGSON, PCI_DEVICE_ID_LOONGSON_DC3, pci_fixup_vgadev);

Because pci_fixup_vgadev() unconditionally calls vga_set_default_device(devp)
for every non-Loongson VGA device without checking whether a firmware default
device was already chosen, won't it clobber vga_default whenever firmware
selected a Loongson display device or the first of multiple non-Loongson
discrete GPUs as the primary display?

[Severity: Medium]
Also in vga_is_firmware_default() in drivers/pci/vgaarb.c, can this pointer
comparison ever match when the firmware framebuffer resides on Loongson
integrated graphics (such as on LS7A1000, LS7A2000, or LS2K2000)?

As shown in lsdc_get_dedicated_vram() in drivers/gpu/drm/loongson/lsdc_drv.c,
the integrated graphics subsystem is split across two PCI functions:

	/*
	 * The GPU has 00:06.0 as its BDF, while the DC has 00:06.1
	 * This is true for the LS7A1000, LS7A2000 and LS2K2000.
	 */
	pdev_gpu = pci_get_domain_bus_and_slot(pci_domain_nr(pdev_dc->bus),
					       pdev_dc->bus->number,
					       PCI_DEVFN(6, 0));
	...
	base = pci_resource_start(pdev_gpu, 2);
	size = pci_resource_len(pdev_gpu, 2);

When the firmware framebuffer is placed in integrated VRAM,
screen_info_pci_dev(&sysfb_primary_display.screen) matches BAR 2 of the
companion GPU device at 00:06.0 (PCI_DEVICE_ID_LOONGSON_GPU1/2/3).

However, vga_arb_device_init() and pci_notify() filter devices using
pci_is_vga(pdev), so only the VGA-class Display Controller function at
00:06.1 (PCI_DEVICE_ID_LOONGSON_DC1/2/3, whose BAR 0 only maps MMIO
registers) is passed to vga_is_firmware_default(pdev). Doesn't comparing
pdev (00:06.1) against screen_info_pci_dev() (00:06.0) always evaluate to
false for the integrated display controller?

>  #else
>  	return false;
> @@ -1527,6 +1527,16 @@ static int __init vga_arb_device_init(void)
>  	int rc;
>  	struct pci_dev *pdev;
>  
> +#if defined(CONFIG_LOONGARCH)
> +	/*
> +	 * On LoongArch, PCI resource assignment completes during
> +	 * subsys_initcall (acpi_init), before the VGA arbiter runs.  Refresh
> +	 * screen_info so vga_is_firmware_default() can match the relocated
> +	 * BAR.
> +	 */
> +	screen_info_apply_fixups();
> +#endif

[Severity: Medium]
With CONFIG_VGA_ARB selecting CONFIG_SYSFB on LoongArch, calling
screen_info_apply_fixups() here in vga_arb_device_init() at
subsys_initcall_sync means screen_info_apply_fixups() will be called a second
time when sysfb_init() in drivers/firmware/sysfb.c runs at device_initcall:

drivers/firmware/sysfb.c:sysfb_init() {
	...
	screen_info_apply_fixups();

	mutex_lock(&disable_lock);
	if (disabled)
		goto unlock_mutex;
	...
}

Looking at screen_info_apply_fixups() in drivers/video/screen_info_pci.c:

void screen_info_apply_fixups(void)
{
	struct screen_info *si = &sysfb_primary_display.screen;

	if (screen_info_lfb_pdev) {
		struct resource *pr = &screen_info_lfb_pdev->resource[screen_info_lfb_bar];

		if (pr->start != screen_info_lfb_res_start) {
			if (__screen_info_relocation_is_valid(si, pr)) {
				__screen_info_set_lfb_base(si, pr->start + screen_info_lfb_offset);
				pr_info("Relocating firmware framebuffer to offset %pa[d] within %pr\n",
					&screen_info_lfb_offset, pr);
			} else {
				pr_warn("Invalid relocating, disabling firmware framebuffer\n");
			}
		}
	}
}

Because screen_info_apply_fixups() neither updates screen_info_lfb_res_start
nor clears screen_info_lfb_pdev after applying the fixup, won't the second
call in sysfb_init() re-run the relocation logic and print a duplicate
message whenever the framebuffer BAR was relocated, or emit a spurious
warning if a built-in GPU driver probed earlier in device_initcall and
resized the BAR after calling sysfb_disable()?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261029033440.973077-1-wuqianhai@loongson.cn?part=1

Re: [PATCH v2] PCI/VGA: Enable firmware default device detection on LoongArch

[Thomas Zimmermann]

Hi

Am 29.10.26 um 04:34 schrieb wuqianhai@loongson.cn:
> From: wuqianhai <wuqianhai@loongson.cn>
>
> On LoongArch, PCI resource assignment completes during subsys_initcall
> (acpi_init), before vga_arb_device_init() runs.  By that time the
> firmware framebuffer BAR may already have been relocated, leaving
> screen_info's lfb_base stale.  Enable vga_is_firmware_default() for
> LoongArch and refresh screen_info in vga_arb_device_init() before the
> boot device is matched.
>
> Unlike x86, sysfb_primary_display is only defined when CONFIG_SYSFB is
> enabled on LoongArch, so have VGA_ARB select SYSFB there.
>
> Signed-off-by: wuqianhai <wuqianhai@loongson.cn>
> ---
> Changes in v2:
> - Add VGA_ARB's select of SYSFB for LoongArch to avoid an undefined
>    reference when CONFIG_SCREEN_INFO is not set.
> - Restrict the screen_info_apply_fixups() call to CONFIG_LOONGARCH so
>    x86 and other architectures are unaffected.
> - Enable vga_is_firmware_default() for LoongArch.
> ---
>   drivers/pci/Kconfig  |  1 +
>   drivers/pci/vgaarb.c | 12 +++++++++++-
>   2 files changed, 12 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/pci/Kconfig b/drivers/pci/Kconfig
> index 0c7408509ba2..55cc60f5d312 100644
> --- a/drivers/pci/Kconfig
> +++ b/drivers/pci/Kconfig
> @@ -256,6 +256,7 @@ config VGA_ARB
>   	default y
>   	depends on !S390
>   	select SCREEN_INFO if X86
> +	select SYSFB if LOONGARCH
>   	help
>   	  Some "legacy" VGA devices implemented on PCI typically have the same
>   	  hard-decoded addresses as they did on ISA. When multiple PCI devices
> diff --git a/drivers/pci/vgaarb.c b/drivers/pci/vgaarb.c
> index 3e58b2c05c85..1b7e915629f9 100644
> --- a/drivers/pci/vgaarb.c
> +++ b/drivers/pci/vgaarb.c
> @@ -566,7 +566,7 @@ EXPORT_SYMBOL(vga_put);
>   
>   static bool vga_is_firmware_default(struct pci_dev *pdev)
>   {
> -#if defined CONFIG_X86
> +#if defined(CONFIG_X86) || defined(CONFIG_LOONGARCH)
>   	return pdev == screen_info_pci_dev(&sysfb_primary_display.screen);
>   #else
>   	return false;
> @@ -1527,6 +1527,16 @@ static int __init vga_arb_device_init(void)
>   	int rc;
>   	struct pci_dev *pdev;
>   
> +#if defined(CONFIG_LOONGARCH)
> +	/*
> +	 * On LoongArch, PCI resource assignment completes during
> +	 * subsys_initcall (acpi_init), before the VGA arbiter runs.  Refresh
> +	 * screen_info so vga_is_firmware_default() can match the relocated
> +	 * BAR.
> +	 */
> +	screen_info_apply_fixups();
> +#endif

No per-arch defines in generic code.  All the setup should work the same 
on all architectures.  All global screen-info state should be maintained 
by drivers/firmware/sysfb.c. Otherwise the framebuffer setup gets even 
more convoluted than it already is.

My question is why does vga_is_firmware_default() run before 
sysfb_init()? [1] Does acpi_init call into it? [1] 
https://elixir.bootlin.com/linux/v7.3/source/drivers/firmware/sysfb.c#L225


Apart from that, we can decouple screen_info_apply_fixups() from 
sysfb_init().  After relocation happens, the PCI code could call the 
helper directly and let it update the global screen_info.

Best regards
Thomas


> +
>   	rc = misc_register(&vga_arb_device);
>   	if (rc < 0)
>   		pr_err("error %d registering device\n", rc);

-- 
--
Thomas Zimmermann
Graphics Driver Developer
SUSE Software Solutions Germany GmbH
Frankenstr. 146, 90461 Nürnberg, Germany, www.suse.com
GF: Stefan Gaiser, Jochen Jaser, Abhinav Puri, (HRB 36809, AG Nürnberg)

[PATCH v3 0/6] LoongArch: Identify the firmware/boot display device

[wuqianhai]

From: wuqianhai <wuqianhai@loongson.cn>

On LoongArch the kernel cannot reliably identify the display device that
was set up by the firmware and used for the boot console.  Two code paths
need this information and currently get it wrong:

  * The PCI VGA arbiter: vga_is_firmware_default() is compiled only on
    x86, so on LoongArch vga_default_device() may select the wrong GPU
    when several are present (for example an integrated display
    controller and a discrete card).

  * The framebuffer console: LoongArch does not implement
    video_is_primary_device(), so with
    CONFIG_FRAMEBUFFER_CONSOLE_DETECT_PRIMARY=y the console falls back to
    the generic stub that always returns false and binds to the first
    registered framebuffer instead of the boot display.

This series fixes both by publishing the firmware framebuffer information
through screen_info on LoongArch and consuming it the same way x86 does.

Patch 1 always defines and initializes sysfb_primary_display, and always
provides its EFI stub alias, instead of guarding them behind
SYSFB/EFI_EARLYCON.

Patch 2 refreshes screen_info from the PCI resource-assignment path after
the root-bus resources are assigned, so the framebuffer address is final
before the VGA arbiter and sysfb_init() consume it.  The call is placed
in the LoongArch PCI scan path because only LoongArch has this ordering:
acpi_init() relocates the BAR before vga_arb_device_init() runs.  On x86
the screen_info address is still accurate at that point.  If you would
rather have the helper called from the generic
pci_assign_unassigned_root_bus_resources() path so that all architectures
share it, that can be done instead.

Patch 3 removes pci_fixup_vgadev(), which overrides the VGA default
device to a discrete card after the arbiter has already selected the
firmware framebuffer device, so vga_default_device() consistently
reflects the boot display.

Patch 4 implements video_is_primary_device() for LoongArch by reusing the
VGA default device selected by the arbiter.

Patch 5 lets the EFI stub link directly to sysfb_primary_display on
LoongArch, so the firmware framebuffer information is always populated.

Patch 6 lets the VGA arbiter use CONFIG_SCREEN_INFO instead of an
architecture check and makes VGA_ARB select SCREEN_INFO on LoongArch.

Patches 1-4 are LoongArch-specific.  Patches 5 and 6 touch generic code
(the EFI stub and the PCI VGA arbiter) and are placed last so they can be
reviewed and merged separately.  Patches 3 and 4 depend on patch 6 taking
effect, but are ordered ahead of it only to group the LoongArch changes
together.

Changes since v2:
  - Split the screen_info refresh out of vga_arb_device_init() into the
    LoongArch PCI resource-assignment path (patch 2), following Thomas'
    suggestion to decouple it from sysfb_init().
  - Drop the per-architecture check in vgaarb.c and the select of SYSFB
    in favour of CONFIG_SCREEN_INFO and select SCREEN_INFO (patch 6).
  - Always define and initialize sysfb_primary_display, and let the EFI
    stub link to it directly (patches 1 and 5), instead of guarding them
    behind SYSFB/EFI_EARLYCON.
  - Remove the vga_default_device() override (patch 3) so the arbiter's
    firmware-framebuffer selection takes effect.
  - Implement video_is_primary_device() (patch 4) by reusing the arbiter's
    default device, so FRAMEBUFFER_CONSOLE_DETECT_PRIMARY works.

Build-tested on loongarch64 with GCC 8.3.0.

wuqianhai (6):
  LoongArch: efi: Provide sysfb_primary_display for screen_info users
  LoongArch/PCI: Refresh screen_info after resource assignment
  LoongArch/PCI: Remove vga default device override
  LoongArch: video: Implement primary framebuffer device detection
  efi/libstub: Enable direct sysfb_primary_display access on LoongArch
  PCI/VGA: Enable firmware default device detection on LoongArch

 arch/loongarch/Makefile                       |  3 ++
 arch/loongarch/include/asm/video.h            |  7 +++++
 arch/loongarch/kernel/efi.c                   |  5 +---
 arch/loongarch/kernel/image-vars.h            |  2 --
 arch/loongarch/pci/acpi.c                     |  4 +++
 arch/loongarch/pci/pci.c                      | 18 ------------
 arch/loongarch/video/Makefile                 |  3 ++
 arch/loongarch/video/video-common.c           | 29 +++++++++++++++++++
 drivers/firmware/efi/libstub/efi-stub-entry.c |  1 +
 drivers/pci/Kconfig                           |  2 +-
 drivers/pci/vgaarb.c                          |  2 +-
 11 files changed, 50 insertions(+), 26 deletions(-)
 create mode 100644 arch/loongarch/video/Makefile
 create mode 100644 arch/loongarch/video/video-common.c

-- 
2.34.1

[PATCH v3 1/6] LoongArch: efi: Provide sysfb_primary_display for screen_info users

[wuqianhai]

From: wuqianhai <wuqianhai@loongson.cn>

LoongArch defines and initializes sysfb_primary_display only when
CONFIG_SYSFB or CONFIG_EFI_EARLYCON is enabled.  The PCI VGA arbiter and
the framebuffer console need the firmware framebuffer information through
screen_info whenever CONFIG_SCREEN_INFO is enabled, even if neither of
those options is set.

Drop the guards so sysfb_primary_display is always defined and
initialized, and its EFI stub alias is always provided, mirroring x86.
This avoids selecting SYSFB just to obtain the symbol.

Signed-off-by: wuqianhai <wuqianhai@loongson.cn>
---
 arch/loongarch/kernel/efi.c        | 5 +----
 arch/loongarch/kernel/image-vars.h | 2 --
 2 files changed, 1 insertion(+), 6 deletions(-)

diff --git a/arch/loongarch/kernel/efi.c b/arch/loongarch/kernel/efi.c
index 69dd83f8082f..e491a82b2e0d 100644
--- a/arch/loongarch/kernel/efi.c
+++ b/arch/loongarch/kernel/efi.c
@@ -74,10 +74,8 @@ bool efi_poweroff_required(void)
 
 unsigned long __initdata primary_display_table = EFI_INVALID_TABLE_ADDR;
 
-#if defined(CONFIG_SYSFB) || defined(CONFIG_EFI_EARLYCON)
 struct sysfb_display_info sysfb_primary_display __section(".data");
 EXPORT_SYMBOL_GPL(sysfb_primary_display);
-#endif
 
 static void __init init_primary_display(void)
 {
@@ -129,8 +127,7 @@ void __init efi_init(void)
 
 	set_bit(EFI_CONFIG_TABLES, &efi.flags);
 
-	if (IS_ENABLED(CONFIG_EFI_EARLYCON) || IS_ENABLED(CONFIG_SYSFB))
-		init_primary_display();
+	init_primary_display();
 
 	if (boot_memmap == EFI_INVALID_TABLE_ADDR)
 		return;
diff --git a/arch/loongarch/kernel/image-vars.h b/arch/loongarch/kernel/image-vars.h
index e557ebd46c2b..4d1d7978553a 100644
--- a/arch/loongarch/kernel/image-vars.h
+++ b/arch/loongarch/kernel/image-vars.h
@@ -11,9 +11,7 @@ __efistub_strcmp		= strcmp;
 __efistub_kernel_entry		= kernel_entry;
 __efistub_kernel_asize		= kernel_asize;
 __efistub_kernel_fsize		= kernel_fsize;
-#if defined(CONFIG_EFI_EARLYCON) || defined(CONFIG_SYSFB)
 __efistub_sysfb_primary_display	= sysfb_primary_display;
-#endif
 
 #endif
 
-- 
2.34.1

Re: [PATCH v3 1/6] LoongArch: efi: Provide sysfb_primary_display for screen_info users

[Thomas Zimmermann]

Hi

Am 10.11.26 um 09:56 schrieb wuqianhai@loongson.cn:
> From: wuqianhai <wuqianhai@loongson.cn>
>
> LoongArch defines and initializes sysfb_primary_display only when
> CONFIG_SYSFB or CONFIG_EFI_EARLYCON is enabled.  The PCI VGA arbiter and
> the framebuffer console need the firmware framebuffer information through
> screen_info whenever CONFIG_SCREEN_INFO is enabled, even if neither of
> those options is set.
>
> Drop the guards so sysfb_primary_display is always defined and
> initialized, and its EFI stub alias is always provided, mirroring x86.
> This avoids selecting SYSFB just to obtain the symbol.
>
> Signed-off-by: wuqianhai <wuqianhai@loongson.cn>
> ---
>   arch/loongarch/kernel/efi.c        | 5 +----
>   arch/loongarch/kernel/image-vars.h | 2 --
>   2 files changed, 1 insertion(+), 6 deletions(-)
>
> diff --git a/arch/loongarch/kernel/efi.c b/arch/loongarch/kernel/efi.c
> index 69dd83f8082f..e491a82b2e0d 100644
> --- a/arch/loongarch/kernel/efi.c
> +++ b/arch/loongarch/kernel/efi.c
> @@ -74,10 +74,8 @@ bool efi_poweroff_required(void)
>   
>   unsigned long __initdata primary_display_table = EFI_INVALID_TABLE_ADDR;
>   
> -#if defined(CONFIG_SYSFB) || defined(CONFIG_EFI_EARLYCON)
>   struct sysfb_display_info sysfb_primary_display __section(".data");
>   EXPORT_SYMBOL_GPL(sysfb_primary_display);
> -#endif

I never understood how linking works. I would assume that this symbol 
interferes with the shared one at [1]. Can't loongarch use that as well?

[1] 
https://elixir.bootlin.com/linux/v7.3.1/source/drivers/firmware/efi/efi-init.c#L64

Best regards
Thomas

>   
>   static void __init init_primary_display(void)
>   {
> @@ -129,8 +127,7 @@ void __init efi_init(void)
>   
>   	set_bit(EFI_CONFIG_TABLES, &efi.flags);
>   
> -	if (IS_ENABLED(CONFIG_EFI_EARLYCON) || IS_ENABLED(CONFIG_SYSFB))
> -		init_primary_display();
> +	init_primary_display();
>   
>   	if (boot_memmap == EFI_INVALID_TABLE_ADDR)
>   		return;
> diff --git a/arch/loongarch/kernel/image-vars.h b/arch/loongarch/kernel/image-vars.h
> index e557ebd46c2b..4d1d7978553a 100644
> --- a/arch/loongarch/kernel/image-vars.h
> +++ b/arch/loongarch/kernel/image-vars.h
> @@ -11,9 +11,7 @@ __efistub_strcmp		= strcmp;
>   __efistub_kernel_entry		= kernel_entry;
>   __efistub_kernel_asize		= kernel_asize;
>   __efistub_kernel_fsize		= kernel_fsize;
> -#if defined(CONFIG_EFI_EARLYCON) || defined(CONFIG_SYSFB)
>   __efistub_sysfb_primary_display	= sysfb_primary_display;
> -#endif
>   
>   #endif
>   

-- 
--
Thomas Zimmermann
Graphics Driver Developer
SUSE Software Solutions Germany GmbH
Frankenstr. 146, 90461 Nürnberg, Germany, www.suse.com
GF: Stefan Gaiser, Jochen Jaser, Abhinav Puri, (HRB 36809, AG Nürnberg)

Re: [PATCH v3 1/6] LoongArch: efi: Provide sysfb_primary_display for screen_info users

[Thomas Zimmermann]

Hi

Am 10.11.26 um 09:56 schrieb wuqianhai@loongson.cn:
> From: wuqianhai <wuqianhai@loongson.cn>
>
> LoongArch defines and initializes sysfb_primary_display only when
> CONFIG_SYSFB or CONFIG_EFI_EARLYCON is enabled.  The PCI VGA arbiter and
> the framebuffer console need the firmware framebuffer information through
> screen_info whenever CONFIG_SCREEN_INFO is enabled, even if neither of
> those options is set.
>
> Drop the guards so sysfb_primary_display is always defined and
> initialized, and its EFI stub alias is always provided, mirroring x86.
> This avoids selecting SYSFB just to obtain the symbol.

The idea is that you select SYSFB is you're interested in that 
information.  I guess we could make sysfb_primary_display available if 
CONFIG_VGA_ARB=y.

Best regards
Thomas

>
> Signed-off-by: wuqianhai <wuqianhai@loongson.cn>
> ---
>   arch/loongarch/kernel/efi.c        | 5 +----
>   arch/loongarch/kernel/image-vars.h | 2 --
>   2 files changed, 1 insertion(+), 6 deletions(-)
>
> diff --git a/arch/loongarch/kernel/efi.c b/arch/loongarch/kernel/efi.c
> index 69dd83f8082f..e491a82b2e0d 100644
> --- a/arch/loongarch/kernel/efi.c
> +++ b/arch/loongarch/kernel/efi.c
> @@ -74,10 +74,8 @@ bool efi_poweroff_required(void)
>   
>   unsigned long __initdata primary_display_table = EFI_INVALID_TABLE_ADDR;
>   
> -#if defined(CONFIG_SYSFB) || defined(CONFIG_EFI_EARLYCON)
>   struct sysfb_display_info sysfb_primary_display __section(".data");
>   EXPORT_SYMBOL_GPL(sysfb_primary_display);
> -#endif
>   
>   static void __init init_primary_display(void)
>   {
> @@ -129,8 +127,7 @@ void __init efi_init(void)
>   
>   	set_bit(EFI_CONFIG_TABLES, &efi.flags);
>   
> -	if (IS_ENABLED(CONFIG_EFI_EARLYCON) || IS_ENABLED(CONFIG_SYSFB))
> -		init_primary_display();
> +	init_primary_display();
>   
>   	if (boot_memmap == EFI_INVALID_TABLE_ADDR)
>   		return;
> diff --git a/arch/loongarch/kernel/image-vars.h b/arch/loongarch/kernel/image-vars.h
> index e557ebd46c2b..4d1d7978553a 100644
> --- a/arch/loongarch/kernel/image-vars.h
> +++ b/arch/loongarch/kernel/image-vars.h
> @@ -11,9 +11,7 @@ __efistub_strcmp		= strcmp;
>   __efistub_kernel_entry		= kernel_entry;
>   __efistub_kernel_asize		= kernel_asize;
>   __efistub_kernel_fsize		= kernel_fsize;
> -#if defined(CONFIG_EFI_EARLYCON) || defined(CONFIG_SYSFB)
>   __efistub_sysfb_primary_display	= sysfb_primary_display;
> -#endif
>   
>   #endif
>   

-- 
--
Thomas Zimmermann
Graphics Driver Developer
SUSE Software Solutions Germany GmbH
Frankenstr. 146, 90461 Nürnberg, Germany, www.suse.com
GF: Stefan Gaiser, Jochen Jaser, Abhinav Puri, (HRB 36809, AG Nürnberg)

Re: [PATCH v3 1/6] LoongArch: efi: Provide sysfb_primary_display for screen_info users

[Thomas Zimmermann]

Hi

Am 10.11.26 um 09:56 schrieb wuqianhai@loongson.cn:
> From: wuqianhai <wuqianhai@loongson.cn>
>
> LoongArch defines and initializes sysfb_primary_display only when
> CONFIG_SYSFB or CONFIG_EFI_EARLYCON is enabled.  The PCI VGA arbiter and
> the framebuffer console need the firmware framebuffer information through
> screen_info whenever CONFIG_SCREEN_INFO is enabled, even if neither of
> those options is set.
>
> Drop the guards so sysfb_primary_display is always defined and
> initialized, and its EFI stub alias is always provided, mirroring x86.
> This avoids selecting SYSFB just to obtain the symbol.

The idea is that you select SYSFB is you're interested in that 
information.  I guess we could make sysfb_primary_display available if 
CONFIG_VGA_ARB=y.

Best regards
Thomas

>
> Signed-off-by: wuqianhai <wuqianhai@loongson.cn>
> ---
>   arch/loongarch/kernel/efi.c        | 5 +----
>   arch/loongarch/kernel/image-vars.h | 2 --
>   2 files changed, 1 insertion(+), 6 deletions(-)
>
> diff --git a/arch/loongarch/kernel/efi.c b/arch/loongarch/kernel/efi.c
> index 69dd83f8082f..e491a82b2e0d 100644
> --- a/arch/loongarch/kernel/efi.c
> +++ b/arch/loongarch/kernel/efi.c
> @@ -74,10 +74,8 @@ bool efi_poweroff_required(void)
>   
>   unsigned long __initdata primary_display_table = EFI_INVALID_TABLE_ADDR;
>   
> -#if defined(CONFIG_SYSFB) || defined(CONFIG_EFI_EARLYCON)
>   struct sysfb_display_info sysfb_primary_display __section(".data");
>   EXPORT_SYMBOL_GPL(sysfb_primary_display);
> -#endif
>   
>   static void __init init_primary_display(void)
>   {
> @@ -129,8 +127,7 @@ void __init efi_init(void)
>   
>   	set_bit(EFI_CONFIG_TABLES, &efi.flags);
>   
> -	if (IS_ENABLED(CONFIG_EFI_EARLYCON) || IS_ENABLED(CONFIG_SYSFB))
> -		init_primary_display();
> +	init_primary_display();
>   
>   	if (boot_memmap == EFI_INVALID_TABLE_ADDR)
>   		return;
> diff --git a/arch/loongarch/kernel/image-vars.h b/arch/loongarch/kernel/image-vars.h
> index e557ebd46c2b..4d1d7978553a 100644
> --- a/arch/loongarch/kernel/image-vars.h
> +++ b/arch/loongarch/kernel/image-vars.h
> @@ -11,9 +11,7 @@ __efistub_strcmp		= strcmp;
>   __efistub_kernel_entry		= kernel_entry;
>   __efistub_kernel_asize		= kernel_asize;
>   __efistub_kernel_fsize		= kernel_fsize;
> -#if defined(CONFIG_EFI_EARLYCON) || defined(CONFIG_SYSFB)
>   __efistub_sysfb_primary_display	= sysfb_primary_display;
> -#endif
>   
>   #endif
>   

-- 
--
Thomas Zimmermann
Graphics Driver Developer
SUSE Software Solutions Germany GmbH
Frankenstr. 146, 90461 Nürnberg, Germany, www.suse.com
GF: Stefan Gaiser, Jochen Jaser, Abhinav Puri, (HRB 36809, AG Nürnberg)

[PATCH v3 2/6] LoongArch/PCI: Refresh screen_info after resource assignment

[wuqianhai]

From: wuqianhai <wuqianhai@loongson.cn>

PCI resource assignment runs from acpi_init() at subsys_initcall time.
The firmware framebuffer BAR can be relocated, leaving screen_info's
lfb_base stale.  vga_arb_device_init() runs later at subsys_initcall_sync
time and reads screen_info to identify the firmware default VGA device.

Update screen_info through the generic screen_info_apply_fixups() helper
right after the root bus resources have been assigned, so both the VGA
arbiter and sysfb_init() observe the final framebuffer address.

Signed-off-by: wuqianhai <wuqianhai@loongson.cn>
---
 arch/loongarch/pci/acpi.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/arch/loongarch/pci/acpi.c b/arch/loongarch/pci/acpi.c
index ccbcea61fcd9..cb313ff67f7e 100644
--- a/arch/loongarch/pci/acpi.c
+++ b/arch/loongarch/pci/acpi.c
@@ -9,6 +9,7 @@
 #include <linux/slab.h>
 #include <linux/pci-acpi.h>
 #include <linux/pci-ecam.h>
+#include <linux/screen_info.h>
 
 #include <asm/pci.h>
 #include <asm/numa.h>
@@ -256,6 +257,9 @@ struct pci_bus *pci_acpi_scan_root(struct acpi_pci_root *root)
 		 */
 		pci_assign_unassigned_root_bus_resources(bus);
 
+		if (IS_ENABLED(CONFIG_SCREEN_INFO))
+			screen_info_apply_fixups();
+
 		list_for_each_entry(child, &bus->children, node)
 			pcie_bus_configure_settings(child);
 	}
-- 
2.34.1

Re: [PATCH v3 2/6] LoongArch/PCI: Refresh screen_info after resource assignment

[Thomas Zimmermann]

Hi

Am 10.11.26 um 09:56 schrieb wuqianhai@loongson.cn:
> From: wuqianhai <wuqianhai@loongson.cn>
>
> PCI resource assignment runs from acpi_init() at subsys_initcall time.
> The firmware framebuffer BAR can be relocated, leaving screen_info's
> lfb_base stale.  vga_arb_device_init() runs later at subsys_initcall_sync
> time and reads screen_info to identify the firmware default VGA device.
>
> Update screen_info through the generic screen_info_apply_fixups() helper
> right after the root bus resources have been assigned, so both the VGA
> arbiter and sysfb_init() observe the final framebuffer address.

I don't know much about the internals of the PCI subsystem.  But I would 
assume that there's a common function that sets up PCI resources on all 
architectures. And this is the place where we'd want to call 
screen_info_apply_fixups().

Best regards
Thomas


>
> Signed-off-by: wuqianhai <wuqianhai@loongson.cn>
> ---
>   arch/loongarch/pci/acpi.c | 4 ++++
>   1 file changed, 4 insertions(+)
>
> diff --git a/arch/loongarch/pci/acpi.c b/arch/loongarch/pci/acpi.c
> index ccbcea61fcd9..cb313ff67f7e 100644
> --- a/arch/loongarch/pci/acpi.c
> +++ b/arch/loongarch/pci/acpi.c
> @@ -9,6 +9,7 @@
>   #include <linux/slab.h>
>   #include <linux/pci-acpi.h>
>   #include <linux/pci-ecam.h>
> +#include <linux/screen_info.h>
>   
>   #include <asm/pci.h>
>   #include <asm/numa.h>
> @@ -256,6 +257,9 @@ struct pci_bus *pci_acpi_scan_root(struct acpi_pci_root *root)
>   		 */
>   		pci_assign_unassigned_root_bus_resources(bus);
>   
> +		if (IS_ENABLED(CONFIG_SCREEN_INFO))
> +			screen_info_apply_fixups();
> +
>   		list_for_each_entry(child, &bus->children, node)
>   			pcie_bus_configure_settings(child);
>   	}

-- 
--
Thomas Zimmermann
Graphics Driver Developer
SUSE Software Solutions Germany GmbH
Frankenstr. 146, 90461 Nürnberg, Germany, www.suse.com
GF: Stefan Gaiser, Jochen Jaser, Abhinav Puri, (HRB 36809, AG Nürnberg)

[PATCH v3 3/6] LoongArch/PCI: Remove vga default device override

[wuqianhai]

From: wuqianhai <wuqianhai@loongson.cn>

pci_fixup_vgadev() forces the VGA default device to a non-Loongson VGA
device whenever a Loongson display controller is present.  This was
added to prefer a discrete graphics card as the default boot device, but
it makes vga_default_device() inconsistent with the firmware boot
display: the firmware framebuffer lives on the integrated controller,
while the override points the default at a discrete card (or leaves it
unset when there is no discrete card).

The fixup runs from pci_bus_add_device(), i.e. after the VGA arbiter has
already selected the default device through the firmware framebuffer, so
it overrides that choice.  Drop it so vga_default_device() consistently
reflects the boot display.

Signed-off-by: wuqianhai <wuqianhai@loongson.cn>
---
 arch/loongarch/pci/pci.c | 18 ------------------
 1 file changed, 18 deletions(-)

diff --git a/arch/loongarch/pci/pci.c b/arch/loongarch/pci/pci.c
index f33c7ea1443d..b756a44272ab 100644
--- a/arch/loongarch/pci/pci.c
+++ b/arch/loongarch/pci/pci.c
@@ -8,7 +8,6 @@
 #include <linux/delay.h>
 #include <linux/types.h>
 #include <linux/pci.h>
-#include <linux/vgaarb.h>
 #include <linux/io-64-nonatomic-lo-hi.h>
 #include <asm/cacheflush.h>
 #include <asm/loongson.h>
@@ -88,23 +87,6 @@ int pcibios_alloc_irq(struct pci_dev *dev)
 	return acpi_pci_irq_enable(dev);
 }
 
-static void pci_fixup_vgadev(struct pci_dev *pdev)
-{
-	struct pci_dev *devp = NULL;
-
-	while ((devp = pci_get_class(PCI_CLASS_DISPLAY_VGA << 8, devp))) {
-		if (devp->vendor != PCI_VENDOR_ID_LOONGSON) {
-			vga_set_default_device(devp);
-			dev_info(&pdev->dev,
-				"Overriding boot device as %X:%X\n",
-				devp->vendor, devp->device);
-		}
-	}
-}
-DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_LOONGSON, PCI_DEVICE_ID_LOONGSON_DC1, pci_fixup_vgadev);
-DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_LOONGSON, PCI_DEVICE_ID_LOONGSON_DC2, pci_fixup_vgadev);
-DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_LOONGSON, PCI_DEVICE_ID_LOONGSON_DC3, pci_fixup_vgadev);
-
 #define CRTC_NUM_MAX		2
 #define CRTC_OUTPUT_ENABLE	0x100
 
-- 
2.34.1

[PATCH v3 0/6] LoongArch: Identify the firmware/boot display device

[wuqianhai]

From: wuqianhai <wuqianhai@loongson.cn>

On LoongArch the kernel cannot reliably identify the display device that
was set up by the firmware and used for the boot console.  Two code paths
need this information and currently get it wrong:

  * The PCI VGA arbiter: vga_is_firmware_default() is compiled only on
    x86, so on LoongArch vga_default_device() may select the wrong GPU
    when several are present (for example an integrated display
    controller and a discrete card).

  * The framebuffer console: LoongArch does not implement
    video_is_primary_device(), so with
    CONFIG_FRAMEBUFFER_CONSOLE_DETECT_PRIMARY=y the console falls back to
    the generic stub that always returns false and binds to the first
    registered framebuffer instead of the boot display.

This series fixes both by publishing the firmware framebuffer information
through screen_info on LoongArch and consuming it the same way x86 does.

Patch 1 always defines and initializes sysfb_primary_display, and always
provides its EFI stub alias, instead of guarding them behind
SYSFB/EFI_EARLYCON.

Patch 2 refreshes screen_info from the PCI resource-assignment path after
the root-bus resources are assigned, so the framebuffer address is final
before the VGA arbiter and sysfb_init() consume it.  The call is placed
in the LoongArch PCI scan path because only LoongArch has this ordering:
acpi_init() relocates the BAR before vga_arb_device_init() runs.  On x86
the screen_info address is still accurate at that point.  If you would
rather have the helper called from the generic
pci_assign_unassigned_root_bus_resources() path so that all architectures
share it, that can be done instead.

Patch 3 removes pci_fixup_vgadev(), which overrides the VGA default
device to a discrete card after the arbiter has already selected the
firmware framebuffer device, so vga_default_device() consistently
reflects the boot display.

Patch 4 implements video_is_primary_device() for LoongArch by reusing the
VGA default device selected by the arbiter.

Patch 5 lets the EFI stub link directly to sysfb_primary_display on
LoongArch, so the firmware framebuffer information is always populated.

Patch 6 lets the VGA arbiter use CONFIG_SCREEN_INFO instead of an
architecture check and makes VGA_ARB select SCREEN_INFO on LoongArch.

Patches 1-4 are LoongArch-specific.  Patches 5 and 6 touch generic code
(the EFI stub and the PCI VGA arbiter) and are placed last so they can be
reviewed and merged separately.  Patches 3 and 4 depend on patch 6 taking
effect, but are ordered ahead of it only to group the LoongArch changes
together.

Changes since v2:
  - Split the screen_info refresh out of vga_arb_device_init() into the
    LoongArch PCI resource-assignment path (patch 2), following Thomas'
    suggestion to decouple it from sysfb_init().
  - Drop the per-architecture check in vgaarb.c and the select of SYSFB
    in favour of CONFIG_SCREEN_INFO and select SCREEN_INFO (patch 6).
  - Always define and initialize sysfb_primary_display, and let the EFI
    stub link to it directly (patches 1 and 5), instead of guarding them
    behind SYSFB/EFI_EARLYCON.
  - Remove the vga_default_device() override (patch 3) so the arbiter's
    firmware-framebuffer selection takes effect.
  - Implement video_is_primary_device() (patch 4) by reusing the arbiter's
    default device, so FRAMEBUFFER_CONSOLE_DETECT_PRIMARY works.

Build-tested on loongarch64 with GCC 8.3.0.

wuqianhai (6):
  LoongArch: efi: Provide sysfb_primary_display for screen_info users
  LoongArch/PCI: Refresh screen_info after resource assignment
  LoongArch/PCI: Remove vga default device override
  LoongArch: video: Implement primary framebuffer device detection
  efi/libstub: Enable direct sysfb_primary_display access on LoongArch
  PCI/VGA: Enable firmware default device detection on LoongArch

 arch/loongarch/Makefile                       |  3 ++
 arch/loongarch/include/asm/video.h            |  7 +++++
 arch/loongarch/kernel/efi.c                   |  5 +---
 arch/loongarch/kernel/image-vars.h            |  2 --
 arch/loongarch/pci/acpi.c                     |  4 +++
 arch/loongarch/pci/pci.c                      | 18 ------------
 arch/loongarch/video/Makefile                 |  3 ++
 arch/loongarch/video/video-common.c           | 29 +++++++++++++++++++
 drivers/firmware/efi/libstub/efi-stub-entry.c |  1 +
 drivers/pci/Kconfig                           |  2 +-
 drivers/pci/vgaarb.c                          |  2 +-
 11 files changed, 50 insertions(+), 26 deletions(-)
 create mode 100644 arch/loongarch/video/Makefile
 create mode 100644 arch/loongarch/video/video-common.c

-- 
2.34.1

[PATCH v3 1/6] LoongArch: efi: Provide sysfb_primary_display for screen_info users

[wuqianhai]

From: wuqianhai <wuqianhai@loongson.cn>

LoongArch defines and initializes sysfb_primary_display only when
CONFIG_SYSFB or CONFIG_EFI_EARLYCON is enabled.  The PCI VGA arbiter and
the framebuffer console need the firmware framebuffer information through
screen_info whenever CONFIG_SCREEN_INFO is enabled, even if neither of
those options is set.

Drop the guards so sysfb_primary_display is always defined and
initialized, and its EFI stub alias is always provided, mirroring x86.
This avoids selecting SYSFB just to obtain the symbol.

Signed-off-by: wuqianhai <wuqianhai@loongson.cn>
---
 arch/loongarch/kernel/efi.c        | 5 +----
 arch/loongarch/kernel/image-vars.h | 2 --
 2 files changed, 1 insertion(+), 6 deletions(-)

diff --git a/arch/loongarch/kernel/efi.c b/arch/loongarch/kernel/efi.c
index 69dd83f8082f..e491a82b2e0d 100644
--- a/arch/loongarch/kernel/efi.c
+++ b/arch/loongarch/kernel/efi.c
@@ -74,10 +74,8 @@ bool efi_poweroff_required(void)
 
 unsigned long __initdata primary_display_table = EFI_INVALID_TABLE_ADDR;
 
-#if defined(CONFIG_SYSFB) || defined(CONFIG_EFI_EARLYCON)
 struct sysfb_display_info sysfb_primary_display __section(".data");
 EXPORT_SYMBOL_GPL(sysfb_primary_display);
-#endif
 
 static void __init init_primary_display(void)
 {
@@ -129,8 +127,7 @@ void __init efi_init(void)
 
 	set_bit(EFI_CONFIG_TABLES, &efi.flags);
 
-	if (IS_ENABLED(CONFIG_EFI_EARLYCON) || IS_ENABLED(CONFIG_SYSFB))
-		init_primary_display();
+	init_primary_display();
 
 	if (boot_memmap == EFI_INVALID_TABLE_ADDR)
 		return;
diff --git a/arch/loongarch/kernel/image-vars.h b/arch/loongarch/kernel/image-vars.h
index e557ebd46c2b..4d1d7978553a 100644
--- a/arch/loongarch/kernel/image-vars.h
+++ b/arch/loongarch/kernel/image-vars.h
@@ -11,9 +11,7 @@ __efistub_strcmp		= strcmp;
 __efistub_kernel_entry		= kernel_entry;
 __efistub_kernel_asize		= kernel_asize;
 __efistub_kernel_fsize		= kernel_fsize;
-#if defined(CONFIG_EFI_EARLYCON) || defined(CONFIG_SYSFB)
 __efistub_sysfb_primary_display	= sysfb_primary_display;
-#endif
 
 #endif
 
-- 
2.34.1

Re: [PATCH v3 1/6] LoongArch: efi: Provide sysfb_primary_display for screen_info users

[sashiko-bot]

> From: wuqianhai <wuqianhai@loongson.cn>
> 
> LoongArch defines and initializes sysfb_primary_display only when
> CONFIG_SYSFB or CONFIG_EFI_EARLYCON is enabled.  The PCI VGA arbiter and
> the framebuffer console need the firmware framebuffer information through
> screen_info whenever CONFIG_SCREEN_INFO is enabled, even if neither of
> those options is set.
> 
> Drop the guards so sysfb_primary_display is always defined and
> initialized, and its EFI stub alias is always provided, mirroring x86.
> This avoids selecting SYSFB just to obtain the symbol.
> 
> Signed-off-by: wuqianhai <wuqianhai@loongson.cn>

Sashiko has reviewed this patch and found no issues.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261110090203.2631481-1-wuqianhai@loongson.cn?part=1

[PATCH v3 2/6] LoongArch/PCI: Refresh screen_info after resource assignment

[wuqianhai]

From: wuqianhai <wuqianhai@loongson.cn>

PCI resource assignment runs from acpi_init() at subsys_initcall time.
The firmware framebuffer BAR can be relocated, leaving screen_info's
lfb_base stale.  vga_arb_device_init() runs later at subsys_initcall_sync
time and reads screen_info to identify the firmware default VGA device.

Update screen_info through the generic screen_info_apply_fixups() helper
right after the root bus resources have been assigned, so both the VGA
arbiter and sysfb_init() observe the final framebuffer address.

Signed-off-by: wuqianhai <wuqianhai@loongson.cn>
---
 arch/loongarch/pci/acpi.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/arch/loongarch/pci/acpi.c b/arch/loongarch/pci/acpi.c
index ccbcea61fcd9..cb313ff67f7e 100644
--- a/arch/loongarch/pci/acpi.c
+++ b/arch/loongarch/pci/acpi.c
@@ -9,6 +9,7 @@
 #include <linux/slab.h>
 #include <linux/pci-acpi.h>
 #include <linux/pci-ecam.h>
+#include <linux/screen_info.h>
 
 #include <asm/pci.h>
 #include <asm/numa.h>
@@ -256,6 +257,9 @@ struct pci_bus *pci_acpi_scan_root(struct acpi_pci_root *root)
 		 */
 		pci_assign_unassigned_root_bus_resources(bus);
 
+		if (IS_ENABLED(CONFIG_SCREEN_INFO))
+			screen_info_apply_fixups();
+
 		list_for_each_entry(child, &bus->children, node)
 			pcie_bus_configure_settings(child);
 	}
-- 
2.34.1

Re: [PATCH v3 2/6] LoongArch/PCI: Refresh screen_info after resource assignment

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 1 potential
issue to consider:
- [Medium] LoongArch/PCI: repeated fixup and UAF via
  screen_info_apply_fixups()
--

commit 1413644b1b0b80d135ecb2502366b5c2eb670d52
Author: wuqianhai <wuqianhai@loongson.cn>

LoongArch/PCI: Refresh screen_info after resource assignment

This commit calls screen_info_apply_fixups() in pci_acpi_scan_root() after
assigning root bus resources so that screen_info reflects any relocated
firmware framebuffer BAR before vga_arb_device_init() and sysfb_init() run.

> diff --git a/arch/loongarch/pci/acpi.c b/arch/loongarch/pci/acpi.c
> index ccbcea61fcd9c..cb313ff67f7e7 100644
> --- a/arch/loongarch/pci/acpi.c
> +++ b/arch/loongarch/pci/acpi.c
[ ... ]
> @@ -256,6 +257,9 @@ struct pci_bus *pci_acpi_scan_root(struct acpi_pci_root *root)
>  		 */
>  		pci_assign_unassigned_root_bus_resources(bus);
>  
> +		if (IS_ENABLED(CONFIG_SCREEN_INFO))
> +			screen_info_apply_fixups();

[Severity: Medium]
Is screen_info_apply_fixups() safe to call from pci_acpi_scan_root()?

screen_info_apply_fixups() was previously called only once from
sysfb_init().  When it updates the framebuffer address, it neither updates
screen_info_lfb_res_start nor clears screen_info_lfb_pdev:

drivers/video/screen_info_pci.c:screen_info_apply_fixups() {
	...
	if (screen_info_lfb_pdev) {
		struct resource *pr = &screen_info_lfb_pdev->resource[screen_info_lfb_bar];

		if (pr->start != screen_info_lfb_res_start) {
			if (__screen_info_relocation_is_valid(si, pr)) {
				__screen_info_set_lfb_base(si, pr->start + screen_info_lfb_offset);
				pr_info("Relocating firmware framebuffer to offset %pa[d] within %pr\n",
					&screen_info_lfb_offset, pr);
	...
}

Because pr->start != screen_info_lfb_res_start remains true after a
relocation, every subsequent root bridge scan on multi-root systems and the
later call in sysfb_init() will re-run the relocation branch and emit
duplicate messages.  If a built-in DRM driver probes before sysfb_init(),
calls sysfb_disable(), and resizes the BAR, sysfb_init() will also
re-evaluate the resized BAR before checking whether sysfb was disabled.

In addition, screen_info_lfb_pdev is an unreferenced pointer that is never
cleared on PCI device removal, while pci_acpi_scan_root() can be called
after boot via acpi_pci_root_add() hotplug.  Removing the boot display PCI
device and hot-adding an ACPI PCI root bridge would dereference the freed
screen_info_lfb_pdev in screen_info_apply_fixups().

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261110090203.2631481-1-wuqianhai@loongson.cn?part=2

[PATCH v3 3/6] LoongArch/PCI: Remove vga default device override

[wuqianhai]

From: wuqianhai <wuqianhai@loongson.cn>

pci_fixup_vgadev() forces the VGA default device to a non-Loongson VGA
device whenever a Loongson display controller is present.  This was
added to prefer a discrete graphics card as the default boot device, but
it makes vga_default_device() inconsistent with the firmware boot
display: the firmware framebuffer lives on the integrated controller,
while the override points the default at a discrete card (or leaves it
unset when there is no discrete card).

The fixup runs from pci_bus_add_device(), i.e. after the VGA arbiter has
already selected the default device through the firmware framebuffer, so
it overrides that choice.  Drop it so vga_default_device() consistently
reflects the boot display.

Signed-off-by: wuqianhai <wuqianhai@loongson.cn>
---
 arch/loongarch/pci/pci.c | 18 ------------------
 1 file changed, 18 deletions(-)

diff --git a/arch/loongarch/pci/pci.c b/arch/loongarch/pci/pci.c
index f33c7ea1443d..b756a44272ab 100644
--- a/arch/loongarch/pci/pci.c
+++ b/arch/loongarch/pci/pci.c
@@ -8,7 +8,6 @@
 #include <linux/delay.h>
 #include <linux/types.h>
 #include <linux/pci.h>
-#include <linux/vgaarb.h>
 #include <linux/io-64-nonatomic-lo-hi.h>
 #include <asm/cacheflush.h>
 #include <asm/loongson.h>
@@ -88,23 +87,6 @@ int pcibios_alloc_irq(struct pci_dev *dev)
 	return acpi_pci_irq_enable(dev);
 }
 
-static void pci_fixup_vgadev(struct pci_dev *pdev)
-{
-	struct pci_dev *devp = NULL;
-
-	while ((devp = pci_get_class(PCI_CLASS_DISPLAY_VGA << 8, devp))) {
-		if (devp->vendor != PCI_VENDOR_ID_LOONGSON) {
-			vga_set_default_device(devp);
-			dev_info(&pdev->dev,
-				"Overriding boot device as %X:%X\n",
-				devp->vendor, devp->device);
-		}
-	}
-}
-DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_LOONGSON, PCI_DEVICE_ID_LOONGSON_DC1, pci_fixup_vgadev);
-DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_LOONGSON, PCI_DEVICE_ID_LOONGSON_DC2, pci_fixup_vgadev);
-DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_LOONGSON, PCI_DEVICE_ID_LOONGSON_DC3, pci_fixup_vgadev);
-
 #define CRTC_NUM_MAX		2
 #define CRTC_OUTPUT_ENABLE	0x100
 
-- 
2.34.1

Re: [PATCH v3 3/6] LoongArch/PCI: Remove vga default device override

[sashiko-bot]

> From: wuqianhai <wuqianhai@loongson.cn>
> 
> pci_fixup_vgadev() forces the VGA default device to a non-Loongson VGA
> device whenever a Loongson display controller is present.  This was
> added to prefer a discrete graphics card as the default boot device, but
> it makes vga_default_device() inconsistent with the firmware boot
> display: the firmware framebuffer lives on the integrated controller,
> while the override points the default at a discrete card (or leaves it
> unset when there is no discrete card).
> 
> The fixup runs from pci_bus_add_device(), i.e. after the VGA arbiter has
> already selected the default device through the firmware framebuffer, so
> it overrides that choice.  Drop it so vga_default_device() consistently
> reflects the boot display.
> 
> Signed-off-by: wuqianhai <wuqianhai@loongson.cn>

Sashiko has reviewed this patch and found no issues.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261110090203.2631481-1-wuqianhai@loongson.cn?part=3

[PATCH v3 4/6] LoongArch: video: Implement primary framebuffer device detection

[wuqianhai]

From: wuqianhai <wuqianhai@loongson.cn>

The LoongArch port does not implement video_is_primary_device(), so it
falls back to the generic stub that always returns false.  As a result,
with CONFIG_FRAMEBUFFER_CONSOLE_DETECT_PRIMARY=y the framebuffer console
cannot map itself to the primary display device and instead binds to the
first registered framebuffer, which is not necessarily the one the
firmware used for the boot console.

Implement video_is_primary_device() by returning whether the device is
the VGA default device.  This reuses the VGA arbiter's choice instead of
matching the firmware framebuffer by address, so it is stable across BAR
resizing: the default device is fixed before any driver runs.

Signed-off-by: wuqianhai <wuqianhai@loongson.cn>
---
 arch/loongarch/Makefile             |  3 +++
 arch/loongarch/include/asm/video.h  |  7 +++++++
 arch/loongarch/video/Makefile       |  3 +++
 arch/loongarch/video/video-common.c | 29 +++++++++++++++++++++++++++++
 4 files changed, 42 insertions(+)
 create mode 100644 arch/loongarch/video/Makefile
 create mode 100644 arch/loongarch/video/video-common.c

diff --git a/arch/loongarch/Makefile b/arch/loongarch/Makefile
index 4b22f95aaafb..7fca93298cef 100644
--- a/arch/loongarch/Makefile
+++ b/arch/loongarch/Makefile
@@ -166,6 +166,9 @@ bootvars-y	= VMLINUX_LOAD_ADDRESS=$(load-y)
 
 drivers-$(CONFIG_PCI)		+= arch/loongarch/pci/
 
+# framebuffer console primary device detection
+drivers-$(CONFIG_VIDEO)		+= arch/loongarch/video/
+
 KBUILD_AFLAGS	+= $(cflags-y)
 KBUILD_CFLAGS	+= $(cflags-y)
 KBUILD_CPPFLAGS += -DVMLINUX_LOAD_ADDRESS=$(load-y)
diff --git a/arch/loongarch/include/asm/video.h b/arch/loongarch/include/asm/video.h
index 9f76845f2d4f..7b38881ea1d6 100644
--- a/arch/loongarch/include/asm/video.h
+++ b/arch/loongarch/include/asm/video.h
@@ -8,6 +8,13 @@
 #include <linux/compiler.h>
 #include <linux/string.h>
 
+struct device;
+
+#ifdef CONFIG_VIDEO
+bool video_is_primary_device(struct device *dev);
+#define video_is_primary_device video_is_primary_device
+#endif
+
 static inline void fb_memcpy_fromio(void *to, const volatile void __iomem *from, size_t n)
 {
 	memcpy(to, (void __force *)from, n);
diff --git a/arch/loongarch/video/Makefile b/arch/loongarch/video/Makefile
new file mode 100644
index 000000000000..f1b2a7e54033
--- /dev/null
+++ b/arch/loongarch/video/Makefile
@@ -0,0 +1,3 @@
+# SPDX-License-Identifier: GPL-2.0-only
+
+obj-$(CONFIG_VIDEO)		+= video-common.o
diff --git a/arch/loongarch/video/video-common.c b/arch/loongarch/video/video-common.c
new file mode 100644
index 000000000000..6fea46661cb6
--- /dev/null
+++ b/arch/loongarch/video/video-common.c
@@ -0,0 +1,29 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (C) 2026 Loongson Technology Corporation Limited
+ */
+
+#include <linux/module.h>
+#include <linux/pci.h>
+#include <linux/vgaarb.h>
+
+#include <asm/video.h>
+
+bool video_is_primary_device(struct device *dev)
+{
+	struct pci_dev *pdev;
+
+	if (!dev_is_pci(dev))
+		return false;
+
+	pdev = to_pci_dev(dev);
+
+	if (!pci_is_display(pdev))
+		return false;
+
+	return pdev == vga_default_device();
+}
+EXPORT_SYMBOL(video_is_primary_device);
+
+MODULE_DESCRIPTION("LoongArch video helpers");
+MODULE_LICENSE("GPL");
-- 
2.34.1

Re: [PATCH v3 4/6] LoongArch: video: Implement primary framebuffer device detection

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 1 potential
issue to consider:
- [Low] LoongArch: video: missing linux/export.h include in
  video-common.c
--

commit 3e46d96932a555bf0346972f71660950e04f755d
Author: wuqianhai <wuqianhai@loongson.cn>

LoongArch: video: Implement primary framebuffer device detection

This commit implements video_is_primary_device() for LoongArch by checking
whether a PCI display device matches vga_default_device().

> diff --git a/arch/loongarch/video/video-common.c b/arch/loongarch/video/video-common.c
> new file mode 100644
> index 0000000000000..6fea46661cb64
> --- /dev/null
> +++ b/arch/loongarch/video/video-common.c
> @@ -0,0 +1,29 @@
> +// SPDX-License-Identifier: GPL-2.0
> +/*
> + * Copyright (C) 2026 Loongson Technology Corporation Limited
> + */
> +
> +#include <linux/module.h>
> +#include <linux/pci.h>
> +#include <linux/vgaarb.h>
> +
> +#include <asm/video.h>
> +
> +bool video_is_primary_device(struct device *dev)
> +{
> +	struct pci_dev *pdev;
> +
> +	if (!dev_is_pci(dev))
> +		return false;
> +
> +	pdev = to_pci_dev(dev);
> +
> +	if (!pci_is_display(pdev))
> +		return false;
> +
> +	return pdev == vga_default_device();
> +}
> +EXPORT_SYMBOL(video_is_primary_device);
> +
> +MODULE_DESCRIPTION("LoongArch video helpers");
> +MODULE_LICENSE("GPL");

[Severity: Low]
Should this file include <linux/export.h> instead of <linux/module.h>?

Using EXPORT_SYMBOL() without a direct #include <linux/export.h> triggers
a scripts/misc-check warning when building with W=2:

  arch/loongarch/video/video-common.c: warning: EXPORT_SYMBOL() is used, but #include <linux/export.h> is missing

Since CONFIG_VIDEO is a bool Kconfig symbol, video-common.o is only ever
built-in and does not need <linux/module.h>, MODULE_DESCRIPTION(), or
MODULE_LICENSE().

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261110090203.2631481-1-wuqianhai@loongson.cn?part=4

[PATCH v3 5/6] efi/libstub: Enable direct sysfb_primary_display access on LoongArch

[wuqianhai]

From: wuqianhai <wuqianhai@loongson.cn>

alloc_primary_display() returns NULL on LoongArch when neither
CONFIG_EFI_EARLYCON nor CONFIG_SYSFB is enabled, so the EFI stub does not
populate the core kernel's sysfb_primary_display.  Allow LoongArch to link
directly to sysfb_primary_display, like x86, so the firmware framebuffer
information is always available.

Signed-off-by: wuqianhai <wuqianhai@loongson.cn>
---
 drivers/firmware/efi/libstub/efi-stub-entry.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/firmware/efi/libstub/efi-stub-entry.c b/drivers/firmware/efi/libstub/efi-stub-entry.c
index efa3ab8ad1c6..2ac22720675c 100644
--- a/drivers/firmware/efi/libstub/efi-stub-entry.c
+++ b/drivers/firmware/efi/libstub/efi-stub-entry.c
@@ -20,6 +20,7 @@ struct sysfb_display_info *alloc_primary_display(void)
 		return __alloc_primary_display();
 
 	if (IS_ENABLED(CONFIG_X86) ||
+	    IS_ENABLED(CONFIG_LOONGARCH) ||
 	    IS_ENABLED(CONFIG_EFI_EARLYCON) ||
 	    IS_ENABLED(CONFIG_SYSFB))
 		return kernel_image_addr(&sysfb_primary_display);
-- 
2.34.1

Re: [PATCH v3 5/6] efi/libstub: Enable direct sysfb_primary_display access on LoongArch

[sashiko-bot]

> From: wuqianhai <wuqianhai@loongson.cn>
> 
> alloc_primary_display() returns NULL on LoongArch when neither
> CONFIG_EFI_EARLYCON nor CONFIG_SYSFB is enabled, so the EFI stub does not
> populate the core kernel's sysfb_primary_display.  Allow LoongArch to link
> directly to sysfb_primary_display, like x86, so the firmware framebuffer
> information is always available.
> 
> Signed-off-by: wuqianhai <wuqianhai@loongson.cn>

Sashiko has reviewed this patch and found no issues.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261110090203.2631481-1-wuqianhai@loongson.cn?part=5

[PATCH v3 6/6] PCI/VGA: Enable firmware default device detection on LoongArch

[wuqianhai]

From: wuqianhai <wuqianhai@loongson.cn>

Use CONFIG_SCREEN_INFO rather than an architecture-specific condition to
guard vga_is_firmware_default().  LoongArch now provides
sysfb_primary_display whenever CONFIG_SCREEN_INFO is enabled, so have
VGA_ARB select SCREEN_INFO on LoongArch as well.

The screen_info staleness is handled by the PCI resource-assignment path
in the previous patch, so vga_arb_device_init() no longer has to refresh
screen_info itself.

Signed-off-by: wuqianhai <wuqianhai@loongson.cn>
---
 drivers/pci/Kconfig  | 2 +-
 drivers/pci/vgaarb.c | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/pci/Kconfig b/drivers/pci/Kconfig
index 0c7408509ba2..01e6378effae 100644
--- a/drivers/pci/Kconfig
+++ b/drivers/pci/Kconfig
@@ -255,7 +255,7 @@ config VGA_ARB
 	bool "VGA Arbitration" if EXPERT
 	default y
 	depends on !S390
-	select SCREEN_INFO if X86
+	select SCREEN_INFO if X86 || LOONGARCH
 	help
 	  Some "legacy" VGA devices implemented on PCI typically have the same
 	  hard-decoded addresses as they did on ISA. When multiple PCI devices
diff --git a/drivers/pci/vgaarb.c b/drivers/pci/vgaarb.c
index 3e58b2c05c85..4528ad038610 100644
--- a/drivers/pci/vgaarb.c
+++ b/drivers/pci/vgaarb.c
@@ -566,7 +566,7 @@ EXPORT_SYMBOL(vga_put);
 
 static bool vga_is_firmware_default(struct pci_dev *pdev)
 {
-#if defined CONFIG_X86
+#ifdef CONFIG_SCREEN_INFO
 	return pdev == screen_info_pci_dev(&sysfb_primary_display.screen);
 #else
 	return false;
-- 
2.34.1

Re: [PATCH v3 6/6] PCI/VGA: Enable firmware default device detection on LoongArch

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 1 potential
issue to consider:
- [Medium] PCI/VGA: stale screen_info on DT/non-x86 paths in
  vga_is_firmware_default()
--

commit cc9285f73294dc9c72f6466431882028a26deb92
Author: wuqianhai <wuqianhai@loongson.cn>

PCI/VGA: Enable firmware default device detection on LoongArch

This commit selects SCREEN_INFO for VGA_ARB on LoongArch and switches
vga_is_firmware_default() from a CONFIG_X86 guard to CONFIG_SCREEN_INFO.

> diff --git a/drivers/pci/vgaarb.c b/drivers/pci/vgaarb.c
> index 3e58b2c05c854..4528ad0386102 100644
> --- a/drivers/pci/vgaarb.c
> +++ b/drivers/pci/vgaarb.c
> @@ -566,7 +566,7 @@ EXPORT_SYMBOL(vga_put);
>  
>  static bool vga_is_firmware_default(struct pci_dev *pdev)
>  {
> -#if defined CONFIG_X86
> +#ifdef CONFIG_SCREEN_INFO
>  	return pdev == screen_info_pci_dev(&sysfb_primary_display.screen);
>  #else
>  	return false;

[Severity: Medium]
Does guarding this with CONFIG_SCREEN_INFO cause vga_is_firmware_default()
to read a stale sysfb_primary_display.screen.lfb_base on LoongArch
DeviceTree systems and other non-x86 EFI architectures (such as ARM64, ARM
and RISC-V with CONFIG_SYSFB=y)?

Commit "LoongArch/PCI: Refresh screen_info after resource assignment" only
adds screen_info_apply_fixups() to LoongArch's ACPI pci_acpi_scan_root().
When PCI probing runs via pci_host_probe() (e.g. loongson_pci_probe()) or
generic pci_acpi_scan_root() in drivers/pci/pci-acpi.c,
vga_arb_device_init() runs at subsys_initcall_sync time before
sysfb_init() calls screen_info_apply_fixups() at device_initcall time.  If
the firmware framebuffer BAR was relocated during resource assignment or
sits behind a host bridge with a non-zero CPU-to-PCI offset,
screen_info_pci_dev() will not match pdev and fails to select the firmware
default VGA device.

Also, the second paragraph of the commit message says the previous patch
handled screen_info staleness so vga_arb_device_init() no longer needs to
refresh screen_info itself, which looks like leftover commentary from an
earlier version of the series.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261110090203.2631481-1-wuqianhai@loongson.cn?part=6

Re: [PATCH v3 0/6] LoongArch: Identify the firmware/boot display device

[Huacai Chen]

Hi, Qianhai,

On Tue, Nov 10, 2026 at 5:04 PM <wuqianhai@loongson.cn> wrote:
>
> From: wuqianhai <wuqianhai@loongson.cn>
>
> On LoongArch the kernel cannot reliably identify the display device that
> was set up by the firmware and used for the boot console.  Two code paths
> need this information and currently get it wrong:
>
>   * The PCI VGA arbiter: vga_is_firmware_default() is compiled only on
>     x86, so on LoongArch vga_default_device() may select the wrong GPU
>     when several are present (for example an integrated display
>     controller and a discrete card).
>
>   * The framebuffer console: LoongArch does not implement
>     video_is_primary_device(), so with
>     CONFIG_FRAMEBUFFER_CONSOLE_DETECT_PRIMARY=y the console falls back to
>     the generic stub that always returns false and binds to the first
>     registered framebuffer instead of the boot display.
>
> This series fixes both by publishing the firmware framebuffer information
> through screen_info on LoongArch and consuming it the same way x86 does.
>
> Patch 1 always defines and initializes sysfb_primary_display, and always
> provides its EFI stub alias, instead of guarding them behind
> SYSFB/EFI_EARLYCON.
>
> Patch 2 refreshes screen_info from the PCI resource-assignment path after
> the root-bus resources are assigned, so the framebuffer address is final
> before the VGA arbiter and sysfb_init() consume it.  The call is placed
> in the LoongArch PCI scan path because only LoongArch has this ordering:
> acpi_init() relocates the BAR before vga_arb_device_init() runs.  On x86
> the screen_info address is still accurate at that point.  If you would
> rather have the helper called from the generic
> pci_assign_unassigned_root_bus_resources() path so that all architectures
> share it, that can be done instead.
>
> Patch 3 removes pci_fixup_vgadev(), which overrides the VGA default
> device to a discrete card after the arbiter has already selected the
> firmware framebuffer device, so vga_default_device() consistently
> reflects the boot display.
>
> Patch 4 implements video_is_primary_device() for LoongArch by reusing the
> VGA default device selected by the arbiter.
>
> Patch 5 lets the EFI stub link directly to sysfb_primary_display on
> LoongArch, so the firmware framebuffer information is always populated.
>
> Patch 6 lets the VGA arbiter use CONFIG_SCREEN_INFO instead of an
> architecture check and makes VGA_ARB select SCREEN_INFO on LoongArch.
>
> Patches 1-4 are LoongArch-specific.  Patches 5 and 6 touch generic code
> (the EFI stub and the PCI VGA arbiter) and are placed last so they can be
> reviewed and merged separately.  Patches 3 and 4 depend on patch 6 taking
> effect, but are ordered ahead of it only to group the LoongArch changes
> together.
>
> Changes since v2:
>   - Split the screen_info refresh out of vga_arb_device_init() into the
>     LoongArch PCI resource-assignment path (patch 2), following Thomas'
>     suggestion to decouple it from sysfb_init().
>   - Drop the per-architecture check in vgaarb.c and the select of SYSFB
>     in favour of CONFIG_SCREEN_INFO and select SCREEN_INFO (patch 6).
>   - Always define and initialize sysfb_primary_display, and let the EFI
>     stub link to it directly (patches 1 and 5), instead of guarding them
>     behind SYSFB/EFI_EARLYCON.
>   - Remove the vga_default_device() override (patch 3) so the arbiter's
>     firmware-framebuffer selection takes effect.
>   - Implement video_is_primary_device() (patch 4) by reusing the arbiter's
>     default device, so FRAMEBUFFER_CONSOLE_DETECT_PRIMARY works.
>
> Build-tested on loongarch64 with GCC 8.3.0.
>
> wuqianhai (6):
>   LoongArch: efi: Provide sysfb_primary_display for screen_info users
>   LoongArch/PCI: Refresh screen_info after resource assignment
>   LoongArch/PCI: Remove vga default device override
>   LoongArch: video: Implement primary framebuffer device detection
>   efi/libstub: Enable direct sysfb_primary_display access on LoongArch
Please fold this patch to the first one.

And your signature should be:
Qianhai Wu <wuqianhai@loongson.cn>

Huacai

>   PCI/VGA: Enable firmware default device detection on LoongArch
>
>  arch/loongarch/Makefile                       |  3 ++
>  arch/loongarch/include/asm/video.h            |  7 +++++
>  arch/loongarch/kernel/efi.c                   |  5 +---
>  arch/loongarch/kernel/image-vars.h            |  2 --
>  arch/loongarch/pci/acpi.c                     |  4 +++
>  arch/loongarch/pci/pci.c                      | 18 ------------
>  arch/loongarch/video/Makefile                 |  3 ++
>  arch/loongarch/video/video-common.c           | 29 +++++++++++++++++++
>  drivers/firmware/efi/libstub/efi-stub-entry.c |  1 +
>  drivers/pci/Kconfig                           |  2 +-
>  drivers/pci/vgaarb.c                          |  2 +-
>  11 files changed, 50 insertions(+), 26 deletions(-)
>  create mode 100644 arch/loongarch/video/Makefile
>  create mode 100644 arch/loongarch/video/video-common.c
>
> --
> 2.34.1
>

Re: [PATCH] PCI/VGA: Fix boot device detection after framebuffer relocation

[kernel test robot]

Hi,

kernel test robot noticed the following build errors:

[auto build test ERROR on pci/next]
[also build test ERROR on pci/for-linus linus/master v7.3 next-20261028]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]

url:    https://github.com/intel-lab-lkp/linux/commits/wuqianhai-loongson-cn/PCI-VGA-Fix-boot-device-detection-after-framebuffer-relocation/20261028-174634
base:   https://git.kernel.org/pub/scm/linux/kernel/git/pci/pci.git next
patch link:    https://lore.kernel.org/r/20261028094634.766723-1-wuqianhai%40loongson.cn
patch subject: [PATCH] PCI/VGA: Fix boot device detection after framebuffer relocation
config: loongarch-allnoconfig (https://download.01.org/0day-ci/archive/20261029/202610291418.tlsEIHSu-lkp@intel.com/config)
compiler: clang version 20.1.8 (https://github.com/llvm/llvm-project 87f0227cb60147a26a1eeb4fb06e3b505e9c7261)
reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20261029/202610291418.tlsEIHSu-lkp@intel.com/reproduce)

If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <lkp@intel.com>
| Closes: https://lore.kernel.org/oe-kbuild-all/202610291418.tlsEIHSu-lkp@intel.com/

All errors (new ones prefixed by >>):

>> ld.lld: error: undefined symbol: screen_info_apply_fixups
   >>> referenced by vgaarb.c
   >>>               drivers/pci/vgaarb.o:(vga_arb_device_init) in archive vmlinux.a

--
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki

Re: [PATCH] PCI/VGA: Fix boot device detection after framebuffer relocation

[kernel test robot]

Hi,

kernel test robot noticed the following build errors:

[auto build test ERROR on pci/next]
[also build test ERROR on pci/for-linus linus/master v7.3 next-20261028]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]

url:    https://github.com/intel-lab-lkp/linux/commits/wuqianhai-loongson-cn/PCI-VGA-Fix-boot-device-detection-after-framebuffer-relocation/20261028-174634
base:   https://git.kernel.org/pub/scm/linux/kernel/git/pci/pci.git next
patch link:    https://lore.kernel.org/r/20261028094634.766723-1-wuqianhai%40loongson.cn
patch subject: [PATCH] PCI/VGA: Fix boot device detection after framebuffer relocation
config: microblaze-defconfig (https://download.01.org/0day-ci/archive/20261029/202610292159.Ls5pkJAS-lkp@intel.com/config)
compiler: microblaze-linux-gcc (GCC) 16.2.0
reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20261029/202610292159.Ls5pkJAS-lkp@intel.com/reproduce)

If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <lkp@intel.com>
| Closes: https://lore.kernel.org/oe-kbuild-all/202610292159.Ls5pkJAS-lkp@intel.com/

All errors (new ones prefixed by >>):

   microblaze-linux-ld: drivers/pci/vgaarb.o: in function `vga_arb_device_init':
>> drivers/pci/vgaarb.c:1536:(.init.text+0x10): undefined reference to `screen_info_apply_fixups'


vim +1536 drivers/pci/vgaarb.c

  1524	
  1525	static int __init vga_arb_device_init(void)
  1526	{
  1527		int rc;
  1528		struct pci_dev *pdev;
  1529	
  1530		/*
  1531		 * PCI resource assignment may have relocated the firmware
  1532		 * framebuffer BAR. Update screen_info's lfb_base so that
  1533		 * vga_is_firmware_default() can match it against the boot
  1534		 * device's relocated resource.
  1535		 */
> 1536		screen_info_apply_fixups();

--
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki

[PATCH v3 4/6] LoongArch: video: Implement primary framebuffer device detection

[wuqianhai]

From: wuqianhai <wuqianhai@loongson.cn>

The LoongArch port does not implement video_is_primary_device(), so it
falls back to the generic stub that always returns false.  As a result,
with CONFIG_FRAMEBUFFER_CONSOLE_DETECT_PRIMARY=y the framebuffer console
cannot map itself to the primary display device and instead binds to the
first registered framebuffer, which is not necessarily the one the
firmware used for the boot console.

Implement video_is_primary_device() by returning whether the device is
the VGA default device.  This reuses the VGA arbiter's choice instead of
matching the firmware framebuffer by address, so it is stable across BAR
resizing: the default device is fixed before any driver runs.

Signed-off-by: wuqianhai <wuqianhai@loongson.cn>
---
 arch/loongarch/Makefile             |  3 +++
 arch/loongarch/include/asm/video.h  |  7 +++++++
 arch/loongarch/video/Makefile       |  3 +++
 arch/loongarch/video/video-common.c | 29 +++++++++++++++++++++++++++++
 4 files changed, 42 insertions(+)
 create mode 100644 arch/loongarch/video/Makefile
 create mode 100644 arch/loongarch/video/video-common.c

diff --git a/arch/loongarch/Makefile b/arch/loongarch/Makefile
index 4b22f95aaafb..7fca93298cef 100644
--- a/arch/loongarch/Makefile
+++ b/arch/loongarch/Makefile
@@ -166,6 +166,9 @@ bootvars-y	= VMLINUX_LOAD_ADDRESS=$(load-y)
 
 drivers-$(CONFIG_PCI)		+= arch/loongarch/pci/
 
+# framebuffer console primary device detection
+drivers-$(CONFIG_VIDEO)		+= arch/loongarch/video/
+
 KBUILD_AFLAGS	+= $(cflags-y)
 KBUILD_CFLAGS	+= $(cflags-y)
 KBUILD_CPPFLAGS += -DVMLINUX_LOAD_ADDRESS=$(load-y)
diff --git a/arch/loongarch/include/asm/video.h b/arch/loongarch/include/asm/video.h
index 9f76845f2d4f..7b38881ea1d6 100644
--- a/arch/loongarch/include/asm/video.h
+++ b/arch/loongarch/include/asm/video.h
@@ -8,6 +8,13 @@
 #include <linux/compiler.h>
 #include <linux/string.h>
 
+struct device;
+
+#ifdef CONFIG_VIDEO
+bool video_is_primary_device(struct device *dev);
+#define video_is_primary_device video_is_primary_device
+#endif
+
 static inline void fb_memcpy_fromio(void *to, const volatile void __iomem *from, size_t n)
 {
 	memcpy(to, (void __force *)from, n);
diff --git a/arch/loongarch/video/Makefile b/arch/loongarch/video/Makefile
new file mode 100644
index 000000000000..f1b2a7e54033
--- /dev/null
+++ b/arch/loongarch/video/Makefile
@@ -0,0 +1,3 @@
+# SPDX-License-Identifier: GPL-2.0-only
+
+obj-$(CONFIG_VIDEO)		+= video-common.o
diff --git a/arch/loongarch/video/video-common.c b/arch/loongarch/video/video-common.c
new file mode 100644
index 000000000000..6fea46661cb6
--- /dev/null
+++ b/arch/loongarch/video/video-common.c
@@ -0,0 +1,29 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (C) 2026 Loongson Technology Corporation Limited
+ */
+
+#include <linux/module.h>
+#include <linux/pci.h>
+#include <linux/vgaarb.h>
+
+#include <asm/video.h>
+
+bool video_is_primary_device(struct device *dev)
+{
+	struct pci_dev *pdev;
+
+	if (!dev_is_pci(dev))
+		return false;
+
+	pdev = to_pci_dev(dev);
+
+	if (!pci_is_display(pdev))
+		return false;
+
+	return pdev == vga_default_device();
+}
+EXPORT_SYMBOL(video_is_primary_device);
+
+MODULE_DESCRIPTION("LoongArch video helpers");
+MODULE_LICENSE("GPL");
-- 
2.34.1

[PATCH v3 6/6] PCI/VGA: Enable firmware default device detection on LoongArch

[wuqianhai]

From: wuqianhai <wuqianhai@loongson.cn>

Use CONFIG_SCREEN_INFO rather than an architecture-specific condition to
guard vga_is_firmware_default().  LoongArch now provides
sysfb_primary_display whenever CONFIG_SCREEN_INFO is enabled, so have
VGA_ARB select SCREEN_INFO on LoongArch as well.

The screen_info staleness is handled by the PCI resource-assignment path
in the previous patch, so vga_arb_device_init() no longer has to refresh
screen_info itself.

Signed-off-by: wuqianhai <wuqianhai@loongson.cn>
---
 drivers/pci/Kconfig  | 2 +-
 drivers/pci/vgaarb.c | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/pci/Kconfig b/drivers/pci/Kconfig
index 0c7408509ba2..01e6378effae 100644
--- a/drivers/pci/Kconfig
+++ b/drivers/pci/Kconfig
@@ -255,7 +255,7 @@ config VGA_ARB
 	bool "VGA Arbitration" if EXPERT
 	default y
 	depends on !S390
-	select SCREEN_INFO if X86
+	select SCREEN_INFO if X86 || LOONGARCH
 	help
 	  Some "legacy" VGA devices implemented on PCI typically have the same
 	  hard-decoded addresses as they did on ISA. When multiple PCI devices
diff --git a/drivers/pci/vgaarb.c b/drivers/pci/vgaarb.c
index 3e58b2c05c85..4528ad038610 100644
--- a/drivers/pci/vgaarb.c
+++ b/drivers/pci/vgaarb.c
@@ -566,7 +566,7 @@ EXPORT_SYMBOL(vga_put);
 
 static bool vga_is_firmware_default(struct pci_dev *pdev)
 {
-#if defined CONFIG_X86
+#ifdef CONFIG_SCREEN_INFO
 	return pdev == screen_info_pci_dev(&sysfb_primary_display.screen);
 #else
 	return false;
-- 
2.34.1