All of lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH v2 00/16] PCI: Serialize bridge window sizing and resource assignment
@ 2026-11-20 13:46 Ilpo Järvinen
  2026-11-20 13:46 ` [PATCH v2 01/16] PCI: Add pci_bus_size_and_assign_resources() Ilpo Järvinen
                   ` (15 more replies)
  0 siblings, 16 replies; 34+ messages in thread
From: Ilpo Järvinen @ 2026-11-20 13:46 UTC (permalink / raw)
  To: linux-pci, Bjorn Helgaas, Lorenzo Pieralisi, Rob Herring,
	Krzysztof Wilczyński, Rafael J. Wysocki (Intel)
  Cc: linux-kernel, Ilpo Järvinen

Hi all,

Bridge window sizing and resource assignment can be started from
multiple entrypoints: rescan and hotplug under pci_rescan_remove_lock,
and BAR resize from sysfs or driver probe under device_lock() only. As
the entrypoints hold pci_bus_sem at most for read, nothing prevents two
sizing and assignment passes from running concurrently on the same
bridge windows and BARs.

Multiple issues due to these races could be reproduced some on a real
host and more in QEMU with a synthetic PCI host bridge (not part of
this series) by looping BAR resize against a rescan through sysfs. The
failures include spurious BAR resize failures, a WARN in
pci_dev_res_restore(), a NULL pointer dereference in
__release_resource(), and a bridge window programmed over RAM. None of
them occurred with this series applied.

The series adds pci_bus_size_and_assign_resources() and
pci_assign_unassigned_slot_resources() so that PCI core can hold the
locks across both phases. It converts the callers (alpha nautilus
retains an extra sizing call to avoid complex refactoring). The
entrypoints are made to consistently hold pci_bus_sem for read and
use guard(). pci_res_setup_lock is added to serialize the sizing and
assignment passes. Last, the series converts the remaining
pci_bus_assign_resources() users and removes it.

Drivers that call pci_assign_resource() or pci_release_resource()
directly are not covered by this series.

I've another series coming up which will add struct pci_setup_data that
is to be held across the entire sizing and assignment algorithm, which
is another reason why those actions have to be kept together going
forward. The goal of that upcoming series is to size also assigned
bridge windows properly and we need a place to keep the alignment
because STARTALIGN is lost when a resource gets assigned.


v2:
- Expand BAR Resize lock coverage to include the parent window acquire
  (sashiko)
- Add pci_assign_unassigned_slot_resources()
- Refactor acpiphp, replaces sizing call in the loop with one
  pci_assign_unassigned_slot_resources() after the loop to not hold
  locks across scan
- Keep pci_res_setup_lock local to setup-bus.c (drops setup-cardbus.c
  assert that was of minor importance anyway)
- Minor language tweaks to changelogs

Ilpo Järvinen (16):
  PCI: Add pci_bus_size_and_assign_resources()
  alpha/PCI: Convert to pci_bus_size_and_assign_resources()
  PCI: Use guard() for pci_bus_sem in
    pci_do_resource_release_and_resize()
  PCI: Add pci_assign_unassigned_slot_resources()
  PCI: acpiphp: Use pci_assign_unassigned_slot_resources()
  PCI: Hold pci_bus_sem across sizing and assignment entrypoints
  PCI: Serialize bridge window sizing and resource assignment
  PCI: Make internal sizing and assignment functions static
  EDAC/i82875p: Convert to pci_assign_unassigned_bus_resources()
  platform/x86: asus-wmi: Convert to
    pci_assign_unassigned_bus_resources()
  platform/x86: eeepc-laptop: Convert to
    pci_assign_unassigned_bus_resources()
  parisc/PCI: Convert to pci_assign_unassigned_bus_resources()
  PCI: octep_hp: Convert to pci_assign_unassigned_bus_resources()
  PCI: ftpci100: Convert to pci_assign_unassigned_bus_resources()
  PCI: hv: Convert to pci_assign_unassigned_bus_resources()
  PCI: Remove pci_bus_assign_resources()

 arch/alpha/kernel/sys_nautilus.c      |   2 +-
 arch/arm/kernel/bios32.c              |   5 +-
 arch/m68k/coldfire/pci.c              |   3 +-
 arch/mips/pci/pci-legacy.c            |   5 +-
 arch/sh/drivers/pci/pci.c             |   3 +-
 drivers/edac/i82875p_edac.c           |   2 +-
 drivers/parisc/dino.c                 |   2 +-
 drivers/parisc/lba_pci.c              |   7 +-
 drivers/pci/controller/pci-ftpci100.c |   2 +-
 drivers/pci/controller/pci-hyperv.c   |   2 +-
 drivers/pci/hotplug/acpiphp_glue.c    |   8 +-
 drivers/pci/hotplug/octep_hp.c        |   2 +-
 drivers/pci/pci.h                     |   6 +-
 drivers/pci/setup-bus.c               | 128 +++++++++++++++++++++-----
 drivers/pcmcia/cardbus.c              |   3 +-
 drivers/platform/x86/asus-wmi.c       |   2 +-
 drivers/platform/x86/eeepc-laptop.c   |   2 +-
 include/linux/pci.h                   |   2 +-
 18 files changed, 126 insertions(+), 60 deletions(-)


base-commit: 6628e7d4c6b25c8d18f1345e3a20d21829aae7e4
-- 
2.47.3


^ permalink raw reply	[flat|nested] 34+ messages in thread

end of thread, other threads:[~2026-11-20 14:33 UTC | newest]

Thread overview: 34+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-11-20 13:46 [PATCH v2 00/16] PCI: Serialize bridge window sizing and resource assignment Ilpo Järvinen
2026-11-20 13:46 ` [PATCH v2 01/16] PCI: Add pci_bus_size_and_assign_resources() Ilpo Järvinen
2026-11-20 13:53   ` sashiko-bot
2026-11-20 13:46 ` [PATCH v2 02/16] alpha/PCI: Convert to pci_bus_size_and_assign_resources() Ilpo Järvinen
2026-11-20 13:55   ` sashiko-bot
2026-11-20 13:46 ` [PATCH v2 03/16] PCI: Use guard() for pci_bus_sem in pci_do_resource_release_and_resize() Ilpo Järvinen
2026-11-20 13:54   ` sashiko-bot
2026-11-20 13:46 ` [PATCH v2 04/16] PCI: Add pci_assign_unassigned_slot_resources() Ilpo Järvinen
2026-11-20 13:53   ` sashiko-bot
2026-11-20 13:47 ` [PATCH v2 05/16] PCI: acpiphp: Use pci_assign_unassigned_slot_resources() Ilpo Järvinen
2026-11-20 13:54   ` sashiko-bot
2026-11-20 14:33   ` Rafael J. Wysocki (Intel)
2026-11-20 13:47 ` [PATCH v2 06/16] PCI: Hold pci_bus_sem across sizing and assignment entrypoints Ilpo Järvinen
2026-11-20 13:55   ` sashiko-bot
2026-11-20 13:47 ` [PATCH v2 07/16] PCI: Serialize bridge window sizing and resource assignment Ilpo Järvinen
2026-11-20 14:00   ` sashiko-bot
2026-11-20 13:47 ` [PATCH v2 08/16] PCI: Make internal sizing and assignment functions static Ilpo Järvinen
2026-11-20 13:52   ` sashiko-bot
2026-11-20 13:47 ` [PATCH v2 09/16] EDAC/i82875p: Convert to pci_assign_unassigned_bus_resources() Ilpo Järvinen
2026-11-20 13:56   ` sashiko-bot
2026-11-20 13:47 ` [PATCH v2 10/16] platform/x86: asus-wmi: " Ilpo Järvinen
2026-11-20 13:53   ` sashiko-bot
2026-11-20 13:47 ` [PATCH v2 11/16] platform/x86: eeepc-laptop: " Ilpo Järvinen
2026-11-20 13:55   ` sashiko-bot
2026-11-20 13:47 ` [PATCH v2 12/16] parisc/PCI: " Ilpo Järvinen
2026-11-20 13:59   ` sashiko-bot
2026-11-20 13:47 ` [PATCH v2 13/16] PCI: octep_hp: " Ilpo Järvinen
2026-11-20 13:58   ` sashiko-bot
2026-11-20 13:47 ` [PATCH v2 14/16] PCI: ftpci100: " Ilpo Järvinen
2026-11-20 13:57   ` sashiko-bot
2026-11-20 13:47 ` [PATCH v2 15/16] PCI: hv: " Ilpo Järvinen
2026-11-20 13:56   ` sashiko-bot
2026-11-20 13:47 ` [PATCH v2 16/16] PCI: Remove pci_bus_assign_resources() Ilpo Järvinen
2026-11-20 13:55   ` sashiko-bot

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.