Re: [PATCH v5 0/4] Introduce pci_enable_msi/msix_range() interfaces

[Bjorn Helgaas]

On Mon, Dec 30, 2013 at 08:28:12AM +0100, Alexander Gordeev wrote:
> This series is against "next" branch in Bjorn's repo:
> git://git.kernel.org/pub/scm/linux/kernel/git/helgaas/pci.git
> 
> Changes from v4 to v5:
>   - pci_auto_enable_msi* functions renamed to pci_enable_msi* ones;
>   - pci_enable_msi_range() and pci_enable_msix_range() are the only
>     basic interfaces left to enable MSI/MSI-X operation;
>   - pci_enable_msi(), pci_enable_msi_block() and pci_enable_msix() old
>     interfaces deprecated in favour of pci_enable_msi/msix_range() ones;
>     description of the old interfaces is removed from the documentation;
>   - pci_get_msi_vec_count()  renamed to pci_msi_vec_count() and
>     pci_get_msix_vec_count() renamed to pci_msix_vec_count();
> 
> Changes from v3 to v4:
>   - pcim_enable_msi* functions renamed to pci_auto_enable_msi* ones;
>   - PowerPC patches dropped;
>   - pci_get_msi_cap()     renamed to pci_get_msi_vec_count() and
>     pci_msix_table_size() renamed to pci_get_msix_vec_count();
> 
> Changes from v2 to v3:
>   - new public interfaces commented in drivers/pci/msi.c;
>   - patch "Make quota traversing and requesting race-safe" explained;
>   - pci_enable_msi/msix() 'nvec' arg type changed from 'unsigned int' to 'int';
>   - pcim_enable_msi*() arg 'nvec' renamed to 'maxvec' when upper limit passed;
>   - pcim_enable_msi*(..., maxvec, minvec) arg order swapped to minvec, maxvec;
>   - "PCI: Fail MSI/MSI-X initialization if device is not in PCI_D0" commit
>     869a161 and "PCI/MSI: Factor out pci_get_msi_cap() interface" patch
>     conflicts resolved;
> 
> This update introduces pci_enable_msi_range() function which
> supersedes pci_enable_msi() and pci_enable_msi_block()
> interfaces for MSI and pci_enable_msi_range() function which
> supersedes pci_enable_msix() interface for MSI-X interrupts.
> 
> The plan is to convert all drivers to the new interfaces and
> then remove existing pci_enable_msi(), pci_enable_msi_block()
> and pci_enable_msix() functions.
> 
> The newly introduced functions return either a negative error
> code or a number of successfully allocated MSI/MSI-X interrupts.
> 
> Unlike the old interfaces where the "third" state indicates
> a possible number of MSI interrupts that could have been
> allocated, the new interfaces allow device drivers not to deal
> with tri-state return values and hide fallback logic behind
> the implementation.
> 
> As result, device drivers are expected to have more clearer
> and straight code.
> 
> The tree could be found in "pci-next-msi-v5" branch in repo:
> https://github.com/a-gordeev/linux.git
> 
> Alexander Gordeev (4):
>   PCI/MSI: Factor out pci_msi_vec_count() interface
>   PCI/MSI: Get rid of pci_enable_msi_block_auto() interface
>   PCI/MSI: Introduce pci_msix_vec_count() interface
>   PCI/MSI: Introduce pci_enable_msi/msix_range() interfaces

I applied all four of these to my pci/msi branch for v3.14.  Thanks
a lot for all your work on this!  I'm looking forward to the driver
conversions and bug fixes :)

Bjorn

>  Documentation/PCI/MSI-HOWTO.txt |  308 ++++++++++++++++++++++++++-------------
>  drivers/ata/ahci.c              |   56 +++++---
>  drivers/pci/msi.c               |  150 +++++++++++++++-----
>  drivers/pci/pcie/portdrv_core.c |    7 +-
>  include/linux/pci.h             |   27 +++-
>  5 files changed, 385 insertions(+), 163 deletions(-)
> 
> -- 
> 1.7.7.6
> 
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/