Patch 1: Reserves the Domain IDs allocated by the host kernel while reusing its
	 device table inside kdump kernel
Patch 2: Cleanup around pdom allocation

-------------------------------------------------------------------------------

Changes since v1:
https://lore.kernel.org/all/20251114104442.23296-1-sarunkod@amd.com/
P1:
  - Use FIELD_GET to read the DTE fields [Ankit]
  - Use GFP_KERNEL instead of GFP_ATOMIC [Jason]
  - Remove print on ENOMEM [Jason]

P2:
  - Use u32 instead of int [Jason]
  - pass gfp flags as argument to 'amd_iommu_pdom_id_reserve' [Jason]

-------------------------------------------------------------------------------

Base Commit: 91920a9d87f5192c56ba5bf3e133aeb3576b705d (iommu-next)

-------------------------------------------------------------------------------

Fixes: 38e5f33ee359 ("iommu/amd: Reuse device table for kdump")
Signed-off-by: Sairaj Kodilkar <sarunkod@amd.com>

Sairaj Kodilkar (2):
  amd/iommu: Preserve domain ids inside the kdump kernel
  amd/iommu: Make protection domain ID functions non-static

 drivers/iommu/amd/amd_iommu.h |  5 +++++
 drivers/iommu/amd/init.c      | 24 +++++++++++++++++++++---
 drivers/iommu/amd/iommu.c     | 27 ++++++++++++++++++---------
 3 files changed, 44 insertions(+), 12 deletions(-)

-- 
2.34.1

Hi Joerg,


On 11/21/2025 2:41 PM, Sairaj Kodilkar wrote:
> Patch 1: Reserves the Domain IDs allocated by the host kernel while reusing its
> 	 device table inside kdump kernel
> Patch 2: Cleanup around pdom allocation

Can you please pick this series as it fixes regression (may be for rc)?

Jason's suggestion are valid. We will address them along with some cleanup in
kdump path as separate series.


-Vasant

On Tue, Dec 02, 2025 at 10:53:45AM +0530, Vasant Hegde wrote:
> Hi Joerg,
> 
> 
> On 11/21/2025 2:41 PM, Sairaj Kodilkar wrote:
> > Patch 1: Reserves the Domain IDs allocated by the host kernel while reusing its
> > 	 device table inside kdump kernel
> > Patch 2: Cleanup around pdom allocation
> 
> Can you please pick this series as it fixes regression (may be for rc)?
> 
> Jason's suggestion are valid. We will address them along with some cleanup in
> kdump path as separate series.

Now applied, thanks.