Message ID | c61285e27134c0e55b4c16a4c488c8fddfb500a2.1409033664.git.marcel@ziswiler.com |
---|---|
State | Accepted |
Delegated to: | Tom Warren |
Headers | show |
On 08/26/2014 03:49 AM, Marcel Ziswiler wrote: > Use architecture specific memcpy to speed up things. Tested-by: Stephen Warren <swarren@nvidia.com> On Jetson TK1, this makes my "dhcp zImage" on my USB ethernet dongle got from ~1.7MiB/s to ~2.8MiB/s :-)
On Tue, 2014-08-26 at 10:27 -0600, Stephen Warren wrote: > On Jetson TK1, this makes my "dhcp zImage" on my USB ethernet dongle got > from ~1.7MiB/s to ~2.8MiB/s :-) Very well indeed. As mentioned before Stefan discovered this working on Vybrid where we saw similar speed-up when booting from NAND flash.
On Tue, 2014-08-26 at 10:27 -0600, Stephen Warren wrote: > On Jetson TK1, this makes my "dhcp zImage" on my USB ethernet dongle got > from ~1.7MiB/s to ~2.8MiB/s :-) I just noticed this not having been applied anywhere yet. Is the tegra tree eventually gona be re-based and this being applied?
On Mon, 2014-10-20 at 15:40 +0000, Tom Warren wrote: > Sorry, Marcel. Must have missed it. I'll rebase u-boot-tegra/master against ARM master and apply this, and if no other Tegra patches are pending, I'll send a PR later this week. Thanks, Tom. According to my notes the following 10 patches are still pending: [PATCH] arm: tegra: use architecture specific memcpy Tested-by: Stephen Warren <swarren@nvidia.com> [PATCH v2] ARM: tegra: Use mem size from MC in combination with get_ram_size() Acked-by: Stephen Warren <swarren@nvidia.com> [PATCH] tegra: colibri_t30: clean-up spurious new line Acked-by: Simon Glass <sjg@chromium.org> [PATCH] tegra: gpio: fix null label regression Acked-by: Simon Glass <sjg@chromium.org> [PATCH] tegra: apalis_t30: master revamp [PATCH] tegra: dts: colibri_t30 add serial port details [PATCH] tegra: clean-up useless define Acked-by: Simon Glass <sjg@chromium.org> [PATCH] gpio: header file comment spelling fixes Acked-by: Simon Glass <sjg@chromium.org> [PATCH] tegra: colibri_t30: asix usb ethernet reset regression Acked-by: Simon Glass <sjg@chromium.org> [PATCH] mmc: Tegra: Fix timeout issue seen on certain eMMC parts Are you OK with those or should I resend any of them? BTW: Maybe acknowledging patch emails with applied messages like some other maintainers do would ease the tracking of an individual patch's status.
diff --git a/include/configs/tegra-common.h b/include/configs/tegra-common.h index 16f45f5..be1c7f5 100644 --- a/include/configs/tegra-common.h +++ b/include/configs/tegra-common.h @@ -109,6 +109,8 @@ #define CONFIG_SYS_MEMTEST_START (NV_PA_SDRC_CS0 + 0x600000) #define CONFIG_SYS_MEMTEST_END (CONFIG_SYS_MEMTEST_START + 0x100000) +#define CONFIG_USE_ARCH_MEMCPY + /*----------------------------------------------------------------------- * Physical Memory Map */
Use architecture specific memcpy to speed up things. Signed-off-by: Marcel Ziswiler <marcel@ziswiler.com> --- include/configs/tegra-common.h | 2 ++ 1 file changed, 2 insertions(+)