ARM: BCMRING: fix IO mem static mapping

Fix bcmring IO mem static mapping.

This causes bcmring booting issue since v3.2

---
   arch/arm/mach-bcmring/mm.c |   27 +++++++++++----------------
   1 file changed, 11 insertions(+), 16 deletions(-)

index 1adec78..b46ea07 100644
@@ -31,24 +31,19 @@
       .type = MT_MEMORY }
 
   static struct map_desc bcmring_io_desc[] __initdata = {
-    IO_DESC(MM_IO_BASE_NAND, SZ_64K),    /* phys:0x28000000-0x28000FFF

virt:0xE8000000-0xE8000FFF size:0x00010000 */
- IO_DESC(MM_IO_BASE_UMI, SZ_64K), /* phys:0x2C000000-0x2C000FFF
virt:0xEC000000-0xEC000FFF size:0x00010000 */
-

       IO_DESC(MM_IO_BASE_BROM, SZ_64K),    /*

phys:0x30000000-0x3000FFFF virt:0xF3000000-0xF300FFFF size:0x00010000 */
- MEM_DESC(MM_IO_BASE_ARAM, SZ_1M), /* phys:0x31000000-0x31FFFFFF
virt:0xF3100000-0xF31FFFFF size:0x01000000 */
- IO_DESC(MM_IO_BASE_DMA0, SZ_1M), /* phys:0x32000000-0x32FFFFFF
virt:0xF3200000-0xF32FFFFF size:0x01000000 */
- IO_DESC(MM_IO_BASE_DMA1, SZ_1M), /* phys:0x33000000-0x33FFFFFF
virt:0xF3300000-0xF33FFFFF size:0x01000000 */
- IO_DESC(MM_IO_BASE_ESW, SZ_1M), /* phys:0x34000000-0x34FFFFFF
virt:0xF3400000-0xF34FFFFF size:0x01000000 */
- IO_DESC(MM_IO_BASE_CLCD, SZ_1M), /* phys:0x35000000-0x35FFFFFF
virt:0xF3500000-0xF35FFFFF size:0x01000000 */
- IO_DESC(MM_IO_BASE_APM, SZ_1M), /* phys:0x36000000-0x36FFFFFF
virt:0xF3600000-0xF36FFFFF size:0x01000000 */
- IO_DESC(MM_IO_BASE_SPUM, SZ_1M), /* phys:0x37000000-0x37FFFFFF
virt:0xF3700000-0xF37FFFFF size:0x01000000 */
- IO_DESC(MM_IO_BASE_VPM_PROG, SZ_1M), /*
phys:0x38000000-0x38FFFFFF virt:0xF3800000-0xF38FFFFF size:0x01000000 */
- IO_DESC(MM_IO_BASE_VPM_DATA, SZ_1M), /*
phys:0x3A000000-0x3AFFFFFF virt:0xF3A00000-0xF3AFFFFF size:0x01000000 */ -
+ MEM_DESC(MM_IO_BASE_ARAM, SZ_1M), /* phys:0x31000000-0x31FFFFFF
virt:0xF3100000-0xF31FFFFF size:0x00100000 */
+ IO_DESC(MM_IO_BASE_DMA0, SZ_1M), /* phys:0x32000000-0x32FFFFFF
virt:0xF3200000-0xF32FFFFF size:0x00100000 */
+ IO_DESC(MM_IO_BASE_DMA1, SZ_1M), /* phys:0x33000000-0x33FFFFFF
virt:0xF3300000-0xF33FFFFF size:0x00100000 */
+ IO_DESC(MM_IO_BASE_ESW, SZ_1M), /*
phys:0x34000000-0x34FFFFFF virt:0xF3400000-0xF34FFFFF size:0x00100000 */
+ IO_DESC(MM_IO_BASE_CLCD, SZ_1M), /* phys:0x35000000-0x35FFFFFF
virt:0xF3500000-0xF35FFFFF size:0x00100000 */
+ IO_DESC(MM_IO_BASE_APM, SZ_1M), /*
phys:0x36000000-0x36FFFFFF virt:0xF3600000-0xF36FFFFF size:0x00100000 */
+ IO_DESC(MM_IO_BASE_SPUM, SZ_1M), /* phys:0x37000000-0x37FFFFFF
virt:0xF3700000-0xF37FFFFF size:0x00100000 */
+ IO_DESC(MM_IO_BASE_VPM_PROG, SZ_1M), /*
phys:0x38000000-0x38FFFFFF virt:0xF3800000-0xF38FFFFF size:0x00100000 */
+ IO_DESC(MM_IO_BASE_VPM_DATA, SZ_1M), /*
phys:0x3A000000-0x3AFFFFFF virt:0xF3A00000-0xF3AFFFFF size:0x00100000 */

       IO_DESC(MM_IO_BASE_VRAM, SZ_64K),    /*

phys:0x40000000-0x4000FFFF virt:0xF4000000-0xF400FFFF size:0x00010000 */
- IO_DESC(MM_IO_BASE_CHIPC, SZ_16M), /*
phys:0x80000000-0x80FFFFFF virt:0xF8000000-0xF8FFFFFF size:0x01000000 */
- IO_DESC(MM_IO_BASE_VPM_EXTMEM_RSVD,
- SZ_16M), /* phys:0x0F000000-0x0FFFFFFF
virt:0xF0000000-0xF0FFFFFF size:0x01000000 */
+ IO_DESC(MM_IO_BASE_CHIPC, SZ_8M), /* phys:0x80000000-0x807FFFFF
virt:0xF8000000-0xF87FFFFF size:0x00800000 */
+ IO_DESC(MM_IO_BASE_VPM_EXTMEM_RSVD, SZ_16M), /*
phys:0x0F000000-0x0FFFFFFF virt:0xF0000000-0xF0FFFFFF size:0x01000000 */
};

void __init bcmring_map_io(void)

1.7.9.5

Hi Jiandong Zheng,

There are a few formalities that are wrong with your patches. First of all, the patches are linewrapped. See Documentation/email-clients.txt for how to set up thunderbird so you can send patches correctly, or just use
"git send-email" instead, which simplifies this task. Your commit messages are not well written. The email subject and the first line of the description are almost identical, which mean that one will read the same text twice in the git history.

This long format description however is too short, and it does not explain why it was broken.

And this line for some reason completely breaks the formatting of the email address.

Here you can see how the line wrapping broke the patch. It is not possible to apply a patch like this.

My preferred method of receiving patches from subarch maintainers is using git pull requests, and that would also help solve some of the problems with your email setup. Patches are ok, too, though.

There are two other things I'd like to discuss with you, now that you've shown up.

First of all, I'd like to understand where you are heading with the bcmring platform support. My understanding is that you only support a single SoC and a single board in there, but there are actually a lot of other products in the same SoC family that could get supported here. Are any of them still relevant enough for your company to support upstream? What is your timeline for adding support for those, and how many other SoCs do you work on?

The other question is about the multiplatform kernel I've recently worked on. bcmring is the only subarchitecture in Linux that puts header files in its own directory outside of the mach/*.h hierarchy, which gets in the way and is rather inconsistent. I've prepared a git branch to clean that up, see the "testing/bcmring" branch of
git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc.git for this. I've also added a patch to clean up some API abuse for MMIO accesses that might otherwise cause problems. Can you test this branch and see if everything still works for you?

Arnd

I verified that the changes work fine after I applied the IO mem mapping and UART patches. I also briefly looked into the changes and they seems good to me. Let me know if you want me to do a review.

Will these changes go to v3.5?

Regards,
JD