Re: [PATCH v3 00/51] PCI: Resource allocation cleanup for v4.3

[Bjorn Helgaas <bhelgaas@google.com>]

On Mon, Jul 27, 2015 at 04:29:18PM -0700, Yinghai Lu wrote:
> Hi,
> 
> After 5b28541552ef (PCI: Restrict 64-bit prefetchable bridge windows
> to 64-bit resources), we have several reports on resource allocation
> failure, and we try to fix the problem with resource clip, and get
> more problems.
> 
> One is realloc fail with two graphics cards above 4G.
> One is from sparc that have problem with clip as we don't parse
> mem64 for it.
> 
> Other report is about pci remove/rescan does not work on some setup
> when BIOS tend to allocate small bus size.
> 
> This patchset enhance resource allocation to address those problems.
> 1. optimize bus mmio alignment calculation.
> 2. optimize bus mmio optional alignment calculation.
> 3. add support for alt size to prefer small bus size to small bus alignment.
>    when we have small resource window on parent bridges.
> 4. treat ROM bar as optional resource.
> 5. during allocation, will pick up just fit resource.
> 6. parse MEM64 for sparc and other system with OF.
> 7. treat non-pref mmio64 if parent bridges are all pcie.
> 8. restore old pref allocation logic if hostbridge does not support mmio64 really.
> 9. don't realloc resource if device firmware does not support bar change.
> 10. add pci=assign_pref_bars to clear and assign pref bars.
> 11. don't clear resource when allocation fails.
> 
> I put latest copy at:
>   git://git.kernel.org/pub/scm/linux/kernel/git/yinghai/linux-yinghai.git for-pci-v4.3-next
> 
> That is based on kind of v4.2-rc4.
> 
> Yijing and Wei Yang both test the branch, and the patchset fixs rescan problem,
> and is working well on powerpc setup.

Overall comments:

Thanks for paying attention to your subject lines.  They are
capitalized and punctuated quite consistently.

The one-line subject is a summary.  The body of the changelog is not a
continuation of it.  The body should stand alone even if you don't
read the subject.  Sometimes that means you have to repeat the subject
line in the body.  That's OK.

Please use "()" when referring to functions.  You do it sometimes, but
not always.

I think we should talk about "required/optional" resources instead of
"must/optional" ones.  Is "addon" a third, separate, thing?

Some typo fixes below:

> Yinghai Lu (51):
>   PCI: Cleanup res_to_dev_res() printout for addon resources
>   PCI: Reuse res_to_dev_res in reassign_resources_sorted

s/reassign_resources_sorted/reassign_resources_sorted()/

>   PCI: Use correct align for optional only resources during sorting
>   PCI: Optimize bus align/size calculation during sizing
>   PCI: Optimize bus align/size calculation for optional during sizing
>   PCI: Don't add too much optional size for hotplug bridge mmio

s/mmio/MMIO/

>   PCI: Reorder resources list for must/optional resources
>   PCI: Remove duplicated code for resource sorting
>   PCI: Rename pdev_sort_resources to pdev_check_resources

s/pdev_sort_resources/pdev_sort_resources()/
s/pdev_check_resources/pdev_check_resources()/

>   PCI: Treat ROM resource as optional during realloc
>   PCI: Add debug printout during releasing partial assigned resources
>   PCI: Simplify res reference using in __assign_resourcs_sorted

s/__assign_resourcs_sorted/__assign_resources_sorted()/ (two fixes)

>   PCI: Separate realloc list checking after allocation
>   PCI: Add __add_to_list()
>   PCI: Cache window alignment value
>   PCI: Check if resource is allocated before pci_assign

s/pci_assign/pci_assign()/

>   PCI: Separate out save_resources/restore_resource

s/save_resources/save_resources()/
s/restore_resource/restore_resource()/

>   PCI: Move comment to pci_need_to_release()
>   PCI: Separate must+optional assigning to another function
>   PCI: Skip must+optional if there is no optional addon
>   PCI: Move saved required resource list out of must+optional assigning
>   PCI: Add alt_size allocation support
>   PCI: Add support for more than two alt_size under same bridge
>   PCI: Better support for two alt_size
>   PCI: Fix size calculation with old_size on rescan path
>   PCI: Don't add too much optional size for hotplug bridge io
>   PCI: Move ISA ioport align out of calculate_iosize

s/calculate_iosize/calculate_iosize()/

>   PCI: Unifiy calculate_size for io port and mmio

s/Unifiy/Unify/
s/calculate_size/calculate_size()/

>   PCI: Allow optional only io resource must size to be 0
>   PCI: Unify skip_ioresource_align()
>   PCI: Kill macro checking for bus io port sizing
>   resources: Split out __allocate_resource()
>   resources: Make allocate_resource return just fit resource

I think "just fit" is commonly referred to as "best fit," as in
http://courses.cs.vt.edu/~csonline/OS/Lessons/MemoryAllocation/index.html.

>   PCI: Check pref compatible bit for mem64 resource of pcie device

s/pcie/PCIe/

>   PCI: Only treat non-pef mmio64 as pref if all bridges has MEM_64

s/non-pef/non-pref/
s/bridges has/bridges have/

>   PCI: Add has_mem64 for host_bridge

s/host_bridge/struct pci_host_bridge/

>   PCI: Only treat non-pef mmio64 as pref if host-bridge has_mem64

s/non-pef/non-pref/
s/host-bridge/host bridge/

>   PCI: Restore pref mmio allocation logic for hostbridge without mmio64

s/hostbridge/host bridge/

>   sparc/PCI: Add mem64 resource parsing for root bus
>   sparc/PCI: Add IORESOURCE_MEM_64 for 64-bit resource in of parsing

s/of/OF/

>   powerpc/PCI: Add IORESOURCE_MEM_64 for 64-bit resource in of parsing

s/of/OF/

>   of/PCI: Add IORESOURCE_MEM_64 for 64-bit resource
>   PCI: Treat optional as must in first try for bridge rescan
>   PCI: Get new realloc size for bridge for last try
>   PCI: Don't release sibiling bridge resources during hotplug

s/sibiling/sibling/

>   PCI: Don't release fixed resource for realloc
>   PCI: Claim fixed resource during remove/rescan path
>   PCI: Set resource to FIXED for lsi devices

s/lsi/LSI/

>   PCI, x86: Add pci=assign_pref_bars to re-allocate pref bars

s/re-allocate/reallocate/
s/bars/BARs/

>   PCI: Introduce resource_disabled()
>   PCI: Don't set flags to 0 when assign resource fail
> 
>  arch/alpha/kernel/pci.c                   |    2 +-
>  arch/ia64/pci/pci.c                       |    4 +-
>  arch/microblaze/pci/pci-common.c          |   23 +-
>  arch/mn10300/unit-asb2305/pci-asb2305.c   |    4 +-
>  arch/mn10300/unit-asb2305/pci.c           |    4 +-
>  arch/powerpc/kernel/pci-common.c          |   27 +-
>  arch/powerpc/kernel/pci_of_scan.c         |    4 +-
>  arch/powerpc/platforms/powernv/pci-ioda.c |   12 +-
>  arch/s390/pci/pci.c                       |    2 +-
>  arch/sparc/kernel/of_device_32.c          |    5 +-
>  arch/sparc/kernel/of_device_64.c          |    5 +-
>  arch/sparc/kernel/pci.c                   |    9 +-
>  arch/sparc/kernel/pci_common.c            |   15 +-
>  arch/sparc/kernel/pci_impl.h              |    1 +
>  arch/x86/include/asm/pci_x86.h            |    2 +-
>  arch/x86/pci/common.c                     |    7 +-
>  arch/x86/pci/i386.c                       |   72 +-
>  arch/xtensa/kernel/pci.c                  |    4 +-
>  drivers/iommu/intel-iommu.c               |    3 +-
>  drivers/of/address.c                      |    4 +-
>  drivers/pci/bus.c                         |    6 +-
>  drivers/pci/host/pcie-rcar.c              |    2 +-
>  drivers/pci/hotplug/acpiphp_glue.c        |    1 +
>  drivers/pci/iov.c                         |    2 +-
>  drivers/pci/pci.c                         |    3 +-
>  drivers/pci/pci.h                         |    4 +
>  drivers/pci/probe.c                       |   44 +-
>  drivers/pci/quirks.c                      |   37 +
>  drivers/pci/rom.c                         |    2 +-
>  drivers/pci/setup-bus.c                   | 1321 +++++++++++++++++++++--------
>  drivers/pci/setup-res.c                   |   18 +-
>  include/asm-generic/pci-bridge.h          |    2 +
>  include/linux/ioport.h                    |    6 +-
>  include/linux/pci.h                       |    2 +
>  kernel/resource.c                         |  175 +++-
>  35 files changed, 1361 insertions(+), 473 deletions(-)
> 
> -- 
> 1.8.4.5
>