** Description changed:

+ [SRU Justification]
+ 
+ Impact: Reported regression to PCI pass-through in QEMU/KVM
+ (https://lists.ubuntu.com/archives/kernel-team/2025-July/161565.html)
+ which was caused by changes to fix suspend/resume on disks with BIOS
+ password set (bug #2110090).
+ 
+ Fix: Upstream fix from 6.16-rc3 addressing the issue (see below)
+ 
+ Testcase: Requires to set up a VM with PCI pass-through.
+ 
+ Regression potential: Things related to PCI power management (devices
+ not being initialized or not properly resuming after suspend).
+ 
+ --- original description ---
+ 
  From: Thomas Lamprecht <t.lampre...@proxmox.com>
  Subject: Re: [PATCH 1/1][SRU][N] PCI: Explicitly put devices into D0 when 
initializing
  
  ...
  
  FYI: This causes regressions on PCI passthrough to QEMU/KVM, among other
  things. A colleague found that the following patch on top fixes it here:
  
  commit 907a7a2e5bf40c6a359b2f6cc53d6fdca04009e0
  Author: Mario Limonciello <mario.limoncie...@amd.com>
  Date:   Wed Jun 11 18:31:16 2025 -0500
  
-     PCI/PM: Set up runtime PM even for devices without PCI PM
-     
-     4d4c10f763d7 ("PCI: Explicitly put devices into D0 when initializing")
-     intended to put PCI devices into D0, but in doing so unintentionally
-     changed runtime PM initialization not to occur on devices that don't
-     support PCI PM.  This caused a regression in vfio-pci due to an imbalance
-     with its use.
-     
-     Adjust the logic in pci_pm_init() so that even if PCI PM isn't supported
-     runtime PM is still initialized.
-     
-     Fixes: 4d4c10f763d7 ("PCI: Explicitly put devices into D0 when 
initializing")
-     Reported-by: Giovanni Cabiddu <giovanni.cabi...@intel.com>
-     Closes: 
https://lore.kernel.org/linux-pci/20250424043232.1848107-1-supe...@kernel.org/T/#m7e8929d6421690dc8bd6dc639d86c2b4db27cbc4
-     Reported-by: Nicolas Dichtel <nicolas.dich...@6wind.com>
-     Closes: 
https://lore.kernel.org/linux-pci/20250424043232.1848107-1-supe...@kernel.org/T/#m40d277dcdb9be64a1609a82412d1aa906263e201
-     Signed-off-by: Mario Limonciello <mario.limoncie...@amd.com>
-     Signed-off-by: Bjorn Helgaas <bhelg...@google.com>
-     Tested-by: Giovanni Cabiddu <giovanni.cabi...@intel.com>
-     Tested-by: Nicolas Dichtel <nicolas.dich...@6wind.com>
-     Reviewed-by: Kuppuswamy Sathyanarayanan 
<sathyanarayanan.kuppusw...@linux.intel.com>
-     Acked-by: Rafael J. Wysocki <raf...@kernel.org>
-     Cc: Alex Williamson <alex.william...@redhat.com>
-     Link: https://patch.msgid.link/20250611233117.61810-1-supe...@kernel.org
+     PCI/PM: Set up runtime PM even for devices without PCI PM
+ 
+     4d4c10f763d7 ("PCI: Explicitly put devices into D0 when initializing")
+     intended to put PCI devices into D0, but in doing so unintentionally
+     changed runtime PM initialization not to occur on devices that don't
+     support PCI PM.  This caused a regression in vfio-pci due to an imbalance
+     with its use.
+ 
+     Adjust the logic in pci_pm_init() so that even if PCI PM isn't supported
+     runtime PM is still initialized.
+ 
+     Fixes: 4d4c10f763d7 ("PCI: Explicitly put devices into D0 when 
initializing")
+     Reported-by: Giovanni Cabiddu <giovanni.cabi...@intel.com>
+     Closes: 
https://lore.kernel.org/linux-pci/20250424043232.1848107-1-supe...@kernel.org/T/#m7e8929d6421690dc8bd6dc639d86c2b4db27cbc4
+     Reported-by: Nicolas Dichtel <nicolas.dich...@6wind.com>
+     Closes: 
https://lore.kernel.org/linux-pci/20250424043232.1848107-1-supe...@kernel.org/T/#m40d277dcdb9be64a1609a82412d1aa906263e201
+     Signed-off-by: Mario Limonciello <mario.limoncie...@amd.com>
+     Signed-off-by: Bjorn Helgaas <bhelg...@google.com>
+     Tested-by: Giovanni Cabiddu <giovanni.cabi...@intel.com>
+     Tested-by: Nicolas Dichtel <nicolas.dich...@6wind.com>
+     Reviewed-by: Kuppuswamy Sathyanarayanan 
<sathyanarayanan.kuppusw...@linux.intel.com>
+     Acked-by: Rafael J. Wysocki <raf...@kernel.org>
+     Cc: Alex Williamson <alex.william...@redhat.com>
+     Link: https://patch.msgid.link/20250611233117.61810-1-supe...@kernel.org
  
  Same holds for the Plucky backport in patch mail id:
  <20250529010730.1863186-2-acelan....@canonical.com>

-- 
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/2117494

Title:
  [Regression Updates] "PCI: Explicitly put devices into D0 when
  initializing" breaks pci-pass-through in QEMU/KVM

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2117494/+subscriptions


-- 
ubuntu-bugs mailing list
ubuntu-bugs@lists.ubuntu.com
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs

Reply via email to