On 2/18/26 01:45, Nilay Shroff wrote:
> Recent changes [1] 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 architectures. 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 against
> pdev->msi_addr_mask. Since the mask was not set correctly, the
> validation fails, causing PCI driver probe failures for devices on
> powerpc systems.
Thanks for catching this. I had naively assumed that pci_setup_device()
was the right place for adding this initialization, and didn't think of
other possibilities.
I grep'd for pci_alloc_dev() and found these uses:
* of_create_pci_dev() in arch/powerpc/kernel/pci_of_scan.c (this patch)
* of_create_pci_dev() in arch/sparc/kernel/pci.c (*same missed init*)
* drivers/char/agp/{alpha,parisc}-agp.c (fake pci_dev, should be fine)
* drivers/scsi/megaraid.c (copying from existing pci_dev, should be fine)
So, while we're at it, can we fix the SPARC one as well in v2? The code
seems similar to what we do for powerpc.
> 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.
>
> [1]
> https://lore.kernel.org/all/[email protected]/
Nit: Link seems redundant given the Fixes tag below.
Thanks again,
Vivian "dramforever" Wang
> 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 | 6 ++++++
> 1 file changed, 6 insertions(+)
>
> [...]
