From: Bjorn Helgaas bhelgaas@google.com
Current default VGA device selection fails in some cases because part of it is done in the vga_arb_device_init() subsys_initcall, and some arches enumerate PCI devices in pcibios_init(), which runs *after* that.
For example:
- On BMC system, the AST2500 bridge [1a03:1150] does not implement PCI_BRIDGE_CTL_VGA. This is perfectly legal but means the legacy VGA resources won't reach downstream devices unless they're included in the usual bridge windows.
- vga_arb_select_default_device() will set a device below such a bridge as the default VGA device as long as it has PCI_COMMAND_IO and PCI_COMMAND_MEMORY enabled.
- vga_arbiter_add_pci_device() is called for every VGA device, either at boot-time or at hot-add time, and it will also set the device as the default VGA device, but ONLY if all bridges leading to it implement PCI_BRIDGE_CTL_VGA.
- This difference between vga_arb_select_default_device() and vga_arbiter_add_pci_device() means that a device below an AST2500 or similar bridge can only be set as the default if it is enumerated before vga_arb_device_init().
- On ACPI-based systems, PCI devices are enumerated by acpi_init(), which runs before vga_arb_device_init().
- On non-ACPI systems, like on MIPS system, they are enumerated by pcibios_init(), which typically runs *after* vga_arb_device_init().
This series consolidates all the default VGA device selection in vga_arbiter_add_pci_device(), which is always called after enumerating a PCI device.
Almost all the work here is Huacai's. I restructured it a little bit and added a few trivial patches on top.
I'd like to move vgaarb.c to drivers/pci eventually, but there's another initcall ordering snag that needs to be resolved first, so this leaves it where it is.
Bjorn
Version history: V0 original implementation as final quirk to set default device. https://lore.kernel.org/r/20210514080025.1828197-6-chenhuacai@loongson.cn
V1 rework vgaarb to do all default device selection in vga_arbiter_add_pci_device(). https://lore.kernel.org/r/20210705100503.1120643-1-chenhuacai@loongson.cn
V2 move arbiter to PCI subsystem, fix nits. https://lore.kernel.org/r/20210722212920.347118-1-helgaas@kernel.org
V3 rewrite the commit log of the last patch (which is also summarized by Bjorn). https://lore.kernel.org/r/20210820100832.663931-1-chenhuacai@loongson.cn
V4 split the last patch to two steps. https://lore.kernel.org/r/20210827083129.2781420-1-chenhuacai@loongson.cn
V5 split Patch-9 again and sort the patches. https://lore.kernel.org/r/20210911093056.1555274-1-chenhuacai@loongson.cn
V6 split Patch-5 again and sort the patches again. https://lore.kernel.org/r/20210916082941.3421838-1-chenhuacai@loongson.cn
V7 stop moving vgaarb to drivers/pci because of ordering issues with misc_init(). https://lore.kernel.org/r/20211015061512.2941859-1-chenhuacai@loongson.cn https://lore.kernel.org/r/CAAhV-H7FhAjM-Ha42Z1dLrE4PvC9frfyeU27KHWcyWKkMftEs...
Bjorn Helgaas (8): vgaarb: Factor out vga_select_framebuffer_device() vgaarb: Factor out default VGA device selection vgaarb: Move framebuffer detection to ADD_DEVICE path vgaarb: Move non-legacy VGA detection to ADD_DEVICE path vgaarb: Move disabled VGA device detection to ADD_DEVICE path vgaarb: Remove empty vga_arb_device_card_gone() vgaarb: Use unsigned format string to print lock counts vgaarb: Replace full MIT license text with SPDX identifier
Huacai Chen (2): vgaarb: Move vga_arb_integrated_gpu() earlier in file vgaarb: Log bridge control messages when adding devices
drivers/gpu/vga/vgaarb.c | 311 +++++++++++++++++++-------------------- 1 file changed, 154 insertions(+), 157 deletions(-)