On 11/28/14, 22:13, "Arnd Bergmann" [off-list ref] wrote:

On Wednesday 26 November 2014, suravee.suthikulpanit at amd.com wrote:

quoted

From: Suravee Suthikulpanit <Suravee.Suthikulpanit@amd.com>

Initial revision of device tree for AMD Seattle Development platform.

Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Marc Zyngier <redacted>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Will Deacon <redacted>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Suravee Suthikulpanit <Suravee.Suthikulpanit@amd.com>
Signed-off-by: Thomas Lendacky <Thomas.Lendacky@amd.com>
Signed-off-by: Joel Schopp <redacted>
---
V5 Changes:
  * Rebase to arm-soc for-next (per Olof)
  * Restructure the DTS/DTSI into board and SoC configurations (per
Olof)
  * Add model property at the top level (per Olof)
  * Move pcie0 under smb and change smb's ranges property to empty
since pcie
    is not in the same range. (per Olof)
  * Change v2m0's ranges property (per Arnd)
  * Change timer interrupt type to level-trigger (per Marc)

Applied to next/arm64, thanks!

Thank you

Looking at this one more time, I had another question:

quoted

+	smb0: smb {
+		compatible = "simple-bus";
+		#address-cells = <2>;
+		#size-cells = <2>;
+		ranges;
+
+		/* DDR range is 40-bit addressing */
+		dma-ranges = <0x80 0x0 0x80 0x0 0x7f 0xffffffff>;
+

What is a DDR range?

Also, what is special about the last byte? Did you intentionally
leave it out? I think when we calculate the dma mask, we will
use 0x3fffffffff so we don't step on the last byte, which I assume
is not what you intended.

Arnd

Hm..probably not then. What I meant is to specify 40-bit addressing for
the memory range starting
from [0x80_0000_0000 - 0x100_0000_0000).

As, I discussed with you on IRC, it should also cover the V2m MSI register
frame, and should be starting from
0. The fix should then be:

dma-ranges = <0 0 0 0 0x100 0x00000000>


I will send a patch out to fix and add better comment for this.

Thanks,

Suravee