> On 20 Feb 2026, at 12:32 PM, Nilay Shroff <[email protected]> wrote:
>
> Recent changes replaced the use of no_64bit_msi with msi_addr_mask.
> As a result, msi_addr_mask is now expected to be initialized to
> DMA_BIT_MASK(64) when a pci_dev is set up. However, this initialization
> was missed on powerpc due to differences in the device initialization
> path compared to other (x86) architecture. Due to this, now PCI device
> probe method fails on powerpc system.
>
> On powerpc systems, struct pci_dev instances are created from device
> tree nodes via of_create_pci_dev(). Because msi_addr_mask was not
> initialized there, it remained zero. Later, during MSI setup,
> msi_verify_entries() validates the programmed MSI address agains
> pdev->msi_addr_mask. Since the mask was not set correctly, the
> validation fails, causing PCI driver probe failures for devices on
> powerpc systems.
>
> Initialize pdev->msi_addr_mask to DMA_BIT_MASK(64) in
> of_create_pci_dev() so that MSI address validation succeeds and device
> probe works as expected.
>
> Fixes: 386ced19e9a3 ("PCI/MSI: Convert the boolean no_64bit_msi flag to a DMA
> address mask")
> Signed-off-by: Nilay Shroff <[email protected]>
> ---
> arch/powerpc/kernel/pci_of_scan.c | 7 +++++++
> 1 file changed, 7 insertions(+)
>
> diff --git a/arch/powerpc/kernel/pci_of_scan.c
> b/arch/powerpc/kernel/pci_of_scan.c
> index 756043dd06e9..fb9fbf0d1796 100644
> --- a/arch/powerpc/kernel/pci_of_scan.c
> +++ b/arch/powerpc/kernel/pci_of_scan.c
> @@ -212,6 +212,13 @@ struct pci_dev *of_create_pci_dev(struct device_node
> *node,
> dev->error_state = pci_channel_io_normal;
> dev->dma_mask = 0xffffffff;
>
> + /*
> + * Assume 64-bit addresses for MSI initially. Will be changed to 32-bit
> + * if MSI (rather than MSI-X) capability does not have
> + * PCI_MSI_FLAGS_64BIT. Can also be overridden by driver.
> + */
> + dev->msi_addr_mask = DMA_BIT_MASK(64);
> +
> /* Early fixups, before probing the BARs */
> pci_fixup_device(pci_fixup_early, dev);
>
> --
> 2.52.0
>
Tested-by: Venkat Rao Bagalkote <[email protected]>
Tested on an IBM Power System LPAR (pseries).
On an unpatched kernel, all LPFC functions failed MSI/MSI-X setup with:
[ 0.175139] Running MSI bitmap self-tests ...
[ 1.521630] lpfc 0152:60:00.0: arch assigned 64-bit MSI address 0xffff3840
above device MSI address mask 0x0
[ 1.536203] lpfc 0152:60:00.0: arch assigned 64-bit MSI address 0xffff3a00
above device MSI address mask 0x0
[ 1.536318] lpfc 0152:60:00.0: 14: [ 1.521874] 0:0484 PCI enable MSI-X
failed (-5)
[ 1.536321] lpfc 0152:60:00.0: 15: [ 1.536251] 0:0488 PCI enable MSI mode
failed (-5)
After applying this patch, lpfc driver probe completes successfully and
MSI/MSI-X setup works normally.
[ 0.206925] Running MSI bitmap self-tests ...
[ 3.040737] lpfc 0152:60:00.1: 1:3238 Reducing IO channels to match number
of MSI-X vectors, requested 32 got 2
[ 4.162067] lpfc 0152:60:00.2: 2:3238 Reducing IO channels to match number
of MSI-X vectors, requested 32 got 1
[ 5.270672] lpfc 0152:60:00.3: 3:3238 Reducing IO channels to match number
of MSI-X vectors, requested 32 got 1
Regards,
Venkat.
