Skip to content

Thunderbolt workarounds take 2

Mika Westerberg linux-pci lkml 1 version In review

Get series with b4
b4 am 1372177330-28013-1-git-send-email-mika.westerberg@linux.intel.com

Versions newest first

  1. v1 Latest6 patches 64 repliesIn review

Patches in v1 No applies recorded yet

#Subject and filesSizeRepliesApplied
1/6PCI: acpiphp: do not check for SLOT_ENABLED in enable_device()drivers/pci/hotplug/acpiphp_glue.c+2 -313 replies
2/6PCI: acpiphp: enable_device(): rescan even if no new devices on slotdrivers/pci/hotplug/acpiphp_glue.c+7 -117 replies
3/6PCI: introduce pci_trim_stale_devices()drivers/pci/remove.c include/linux/pci.h+21 -03 replies
4/6PCI: acpiphp: check for new devices on enabled hostdrivers/pci/hotplug/acpiphp_glue.c+27 -296 replies
5/6PCI: acpiphp: look _RMV method a bit deeper in the hierarhcydrivers/pci/hotplug/acpi_pcihp.c+48 -510 replies
6/6x86/PCI: quirk Thunderbolt PCI-to-PCI bridgesarch/x86/pci/fixup.c+51 -019 replies

Reviews trailers per patch

None of the 6 patches has a Reviewed-by or Acked-by of their own; the series has one for all of them.

PatchRafael J. Wysocki
SeriesA-b
1/6
2/6
3/6
4/6
5/6
6/6

Trailers given in replies are filled; those carried in the patches from earlier versions are outlined with dashes.

Read the v1 thread

Cover letter [PATCH 0/6] Thunderbolt workarounds take 2

Hi all,

This is the second iteration of the patches first posted here quite a while ago:

	https://lkml.org/lkml/2012/12/13/589

With Thunderbolt devices we might see pretty complex PCIe hierarchies that the current ACPI hotplug code was not able to handle properly but instead failed to scan the whole hierarchy. This series tries to address these problems. In addition to that there's a possibility that only part of the chain is disconnected so we don't want to leave behind stale PCI devices.

We have tested this on Acer Aspire S5 with the largest chain looking like this:

PC +--+ eSata Hub #0 +--+ eSata Hub #1 +--+ Apple Thunderbolt display +--+ Apple ethernet dongle

There are still problems with different device drivers because of the surprise hotplug nature of Thunderbolt, causing occasional hangs and failures that should be fixed per-driver later on (not addressed in this series).

Currently Thunderbolt on Apple hardware like Macbooks is not yet supported as they use some different mechanism than ACPI events to trigger hotplug events.

The series is based on linux-pm.git/linux-next.

Kirill A. Shutemov (4):
PCI: acpiphp: do not check for SLOT_ENABLED in enable_device()
PCI: acpiphp: enable_device(): rescan even if no new devices on slot
PCI: introduce pci_trim_stale_devices()
PCI: acpiphp: check for new devices on enabled host

Mika Westerberg (2):
PCI: acpiphp: look _RMV method a bit deeper in the hierarhcy x86/PCI: quirk Thunderbolt PCI-to-PCI bridges

 arch/x86/pci/fixup.c               | 51 ++++++++++++++++++++++++
 drivers/pci/hotplug/acpi_pcihp.c   | 53 ++++++++++++++++++++++---
 drivers/pci/hotplug/acpiphp_glue.c | 79 +++++++++++++++++---------------------
 drivers/pci/remove.c               | 20 ++++++++++
 include/linux/pci.h                |  1 +
 5 files changed, 156 insertions(+), 48 deletions(-)
Signature · 1 line
1.8.3.1