Skip to content

RFC 64 bit resources

Vivek Goyal lkml 1 version NAK

Status from this message: NAK - if the resource is 64bit then the pci_write_config_dword is also

Get series with b4
b4 am [email protected]

Versions newest first

  1. v1 RFC Latest10 patches 15 repliesNAK

Patches in v1

#Subject and filesSizeRepliesApplied
1/1064 bit resources core changesinclude/linux/ioport.h kernel/resource.c+37 -349 replies
2/1064 bit resources drivers pci changesdrivers/pci/bus.c drivers/pci/hotplug/cpcihp_zt5550.c drivers/pci/hotplug/cpqphp_core.c +9 more+72 -600 replies
3/1064 bit resources drivers ide changesdrivers/ide/pci/aec62xx.c drivers/ide/pci/cmd64x.c drivers/ide/pci/hpt34x.c +2 more+10 -82 replies
4/1064 bit resources drivers media changesdrivers/media/video/bttv-driver.c drivers/media/video/cx88/cx88-alsa.c drivers/media/video/cx88/cx88-core.c +3 more+20 -180 replies
5/1064 bit resources drivers net changesdrivers/net/3c59x.c drivers/net/8139cp.c drivers/net/8139too.c +8 more+39 -340 replies
6/1064 bit resources drivers pcmcia changesdrivers/pcmcia/i82365.c drivers/pcmcia/pd6729.c drivers/pcmcia/rsrc_nonstatic.c +1 more+21 -170 replies
7/1064 bit resources drivers others changesdrivers/amba/bus.c drivers/atm/ambassador.c drivers/atm/firestream.c +21 more+99 -740 replies
8/1064 bit resources sound changessound/arm/aaci.c sound/drivers/mpu401/mpu401.c sound/isa/es18xx.c +7 more+27 -200 replies
9/1064 bit resources arch changesarch/arm/kernel/bios32.c arch/arm/kernel/setup.c arch/i386/kernel/efi.c +4 more+78 -360 replies
10/10i386: export memory more than 4G through /proc/iomemarch/i386/kernel/setup.c+0 -20 replies

Reviews trailers per patch

No Reviewed-by, Acked-by or Tested-by trailers for v1 yet.

Read the v1 thread

Cover letter [RFC][PATCH 0/10] 64 bit resources

Hi,

Here is an attempt to implement support for 64 bit resources. This will enable memory more than 4G to be exported through /proc/iomem, which is used by kexec/kdump to determine the physical memory layout of the system.

As suggested in the last week's discussion, I picked up greg's patch and made changes on top of that.

http://www.kernel.org/pub/linux/kernel/people/gregkh/pci/2.6/2.6.11-rc3/bk-resource-2.6.11-rc3-mm1.patch

Last discussion witnessed a divided opinion upon whether to use u64 or something like dma_addr_t as data type. The argument is usage of dma_addr_t will reduce code size bloat for the people who don't want to use 64 bit resources on 32 bit platforms. But probably name dma_addr_t is not very appropriate for usage at various places.

Implementing u64 seems to be simpler and code bloat might not be significant.

This implementation is based on u64 and I have measured code bloat on i386.

- Greg's patch already had core changes. We have maninly fixed the warnings. Most of them originating from printk's().

- This patch does not try to modify the users of resources who think that on 32 bit platforms resources can be 32 bit only. They will continue to work as usual as they will silently truncate the 64 bit value to 32bit.

- We did the compilation for i386, arm, sparc, sparc64, x86_64 and ppc64 platforms and fixed the warnings.

Test Results:
------------

We used "make allyesconfig" with CONFIG_DEBUG_INFO=n on 2.6.16-mm1.

i386
----

vmlinux size without patch: 40191425
vmlinux size with path: 40244677
vmlinux size bloat: 52K (.13%)

x86_64
------

vmlinux size without patch: 42387898
vmlinux size with path: 42387945
vmlinux size bloat: 47 bytes

Thanks
Vivek