Skip to content

vgaarb: Rework default VGA device selection

Bjorn Helgaas dri-devel linux-pci lkml 3 versions – In review

Get series with b4
b4 am [email protected]

Versions newest first · linked by author and title

  1. v9 Latest11 patches 3 repliesIn review
  2. v810 patches 12 repliesSuperseded
  3. v29 patches 13 repliesSuperseded

Patches in v9 No applies recorded yet

#Subject and filesSizeRepliesApplied
1/11PCI/VGA: Move vgaarb to drivers/pcidrivers/gpu/vga/Kconfig drivers/gpu/vga/Makefile drivers/pci/Kconfig +2 more+21 -210 replies
2/11PCI/VGA: Move vga_arb_integrated_gpu() earlier in filedrivers/pci/vgaarb.c+11 -140 replies
3/11PCI/VGA: Factor out vga_select_framebuffer_device()drivers/pci/vgaarb.c+53 -460 replies
4/11PCI/VGA: Factor out default VGA device selectiondrivers/pci/vgaarb.c+39 -60 replies
5/11PCI/VGA: Move firmware default device detection to ADD_DEVICE pathdrivers/pci/vgaarb.c+17 -200 replies
6/11PCI/VGA: Move non-legacy VGA detection to ADD_DEVICE pathdrivers/pci/vgaarb.c+32 -220 replies
7/11PCI/VGA: Move disabled VGA device detection to ADD_DEVICE pathdrivers/pci/vgaarb.c+9 -180 replies
8/11PCI/VGA: Remove empty vga_arb_device_card_gone()drivers/pci/vgaarb.c+1 -150 replies
9/11PCI/VGA: Log bridge control messages when adding devicesdrivers/pci/vgaarb.c+8 -110 replies
10/11PCI/VGA: Use unsigned format string to print lock countsdrivers/pci/vgaarb.c+1 -10 replies
11/11PCI/VGA: Replace full MIT license text with SPDX identifierdrivers/pci/vgaarb.c+1 -220 replies

Reviews trailers per patch

No Reviewed-by, Acked-by or Tested-by trailers for v9 yet.

Read the v9 thread

Cover letter [PATCH v9 00/11] vgaarb: Rework default VGA device selection

From: Bjorn Helgaas <[email protected]>

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.

The big change from the v8 posting is that this moves vgaarb.c from drivers/gpu/vga to drivers/pci because it really has nothing to do with GPUs or DRM.

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.

Bjorn

Version history:
V0 original implementation as final quirk to set default device.
https://lore.kernel.org/r/[email protected]

V1 rework vgaarb to do all default device selection in vga_arbiter_add_pci_device().
https://lore.kernel.org/r/[email protected]

V2 move arbiter to PCI subsystem, fix nits.
https://lore.kernel.org/r/[email protected]

V3 rewrite the commit log of the last patch (which is also summarized by Bjorn).
https://lore.kernel.org/r/[email protected]

V4 split the last patch to two steps.
https://lore.kernel.org/r/[email protected]

V5 split Patch-9 again and sort the patches.
https://lore.kernel.org/r/[email protected]

V6 split Patch-5 again and sort the patches again.
https://lore.kernel.org/r/[email protected]

V7 stop moving vgaarb to drivers/pci because of ordering issues with misc_init().
https://lore.kernel.org/r/[email protected]
https://lore.kernel.org/r/CAAhV-H7FhAjM-Ha42Z1dLrE4PvC9frfyeU27KHWcyWKkMftEsA@mail.gmail.com

V8 expand commit logs, rework boot VGA device selection
https://lore.kernel.org/r/[email protected]

V9 (this posting) move to drivers/pci, resolve initcall ordering, rename vga_is_framebuffer_device() to vga_is_firmware_default()

Bjorn Helgaas (5):
PCI/VGA: Move vgaarb to drivers/pci
PCI/VGA: Factor out vga_select_framebuffer_device() PCI/VGA: Remove empty vga_arb_device_card_gone() PCI/VGA: Use unsigned format string to print lock counts PCI/VGA: Replace full MIT license text with SPDX identifier

Huacai Chen (6):
PCI/VGA: Move vga_arb_integrated_gpu() earlier in file PCI/VGA: Factor out default VGA device selection PCI/VGA: Move firmware default device detection to ADD_DEVICE path PCI/VGA: Move non-legacy VGA detection to ADD_DEVICE path PCI/VGA: Move disabled VGA device detection to ADD_DEVICE path PCI/VGA: Log bridge control messages when adding devices

 drivers/gpu/vga/Kconfig           |  19 --
 drivers/gpu/vga/Makefile          |   1 -
 drivers/pci/Kconfig               |  19 ++
 drivers/pci/Makefile              |   1 +
 drivers/{gpu/vga => pci}/vgaarb.c | 313 +++++++++++++++---------------
 5 files changed, 175 insertions(+), 178 deletions(-)
 rename drivers/{gpu/vga => pci}/vgaarb.c (90%)
Signature · 1 line
2.25.1