[U-Boot] [PATCH] sunxi: reduce Orange Pi Zero DRAM clock speed
Using `fel-boot-lima-memtester-on-orange-pi-pc 672` on an Orange Pi without heatsink results in the following error after a few minutes: WRITE FAILURE: 0x00200000 != 0xffdfffff at offset 0x0137f47c (bitflip).
Also, the constructor repository (github/orangepi-xunlong) seems to contain that 624 Mhz clock speed in its u-boot fork. It may be that 672 Mhz is the advertized overclocked speed.
According to http://linux-sunxi.org/Orange_Pi_PC#DRAM_clock_speed_limit it may be worth decreasing that value with other Orange Pi boards.
See also e7d6aa0b74b7f4d08ee68da8a586c76c761348e2. --- configs/orangepi_zero_defconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/configs/orangepi_zero_defconfig b/configs/orangepi_zero_defconfig index c650ce87d9..16a3929aa5 100644 --- a/configs/orangepi_zero_defconfig +++ b/configs/orangepi_zero_defconfig @@ -2,7 +2,7 @@ CONFIG_ARM=y CONFIG_ARCH_SUNXI=y CONFIG_SPL_SPI_FLASH_SUPPORT=y CONFIG_MACH_SUN8I_H3=y -CONFIG_DRAM_CLK=672 +CONFIG_DRAM_CLK=624 CONFIG_DRAM_ZQ=3881979 CONFIG_DRAM_ODT_EN=y # CONFIG_VIDEO_DE2 is not set
On Mon, Jul 24, 2017 at 1:44 AM, Clément Bœsch u@pkh.me wrote:
Using `fel-boot-lima-memtester-on-orange-pi-pc 672` on an Orange Pi without heatsink results in the following error after a few minutes: WRITE FAILURE: 0x00200000 != 0xffdfffff at offset 0x0137f47c (bitflip).
Also, the constructor repository (github/orangepi-xunlong) seems to contain that 624 Mhz clock speed in its u-boot fork. It may be that 672 Mhz is the advertized overclocked speed.
According to http://linux-sunxi.org/Orange_Pi_PC#DRAM_clock_speed_limit it may be worth decreasing that value with other Orange Pi boards.
See also e7d6aa0b74b7f4d08ee68da8a586c76c761348e2.
Applied to u-boot-sunxi/master
thanks!
On 2017-08-11 12:21, Jagan Teki wrote:
On Mon, Jul 24, 2017 at 1:44 AM, Clément Bœsch u@pkh.me wrote:
Using `fel-boot-lima-memtester-on-orange-pi-pc 672` on an Orange Pi without heatsink results in the following error after a few minutes: WRITE FAILURE: 0x00200000 != 0xffdfffff at offset 0x0137f47c (bitflip).
Also, the constructor repository (github/orangepi-xunlong) seems to contain that 624 Mhz clock speed in its u-boot fork. It may be that 672 Mhz is the advertized overclocked speed.
According to http://linux-sunxi.org/Orange_Pi_PC#DRAM_clock_speed_limit it may be worth decreasing that value with other Orange Pi boards.
See also e7d6aa0b74b7f4d08ee68da8a586c76c761348e2.
Applied to u-boot-sunxi/master
I've a 1.5 years old OPi Zero that was running uboot 2017.01 so far. I just tried to update to 2018.05, also using the new defconfig. The result was that uboot got stuck early during the first lines of output. Then I changed the DRAM clock back to 672, and things seem to work find.
Are we facing different board revisions that need different settings, or could the lowering be incorrect?
Jan
On 2018-05-31 19:36, Jan Kiszka wrote:
On 2017-08-11 12:21, Jagan Teki wrote:
On Mon, Jul 24, 2017 at 1:44 AM, Clément Bœsch u@pkh.me wrote:
Using `fel-boot-lima-memtester-on-orange-pi-pc 672` on an Orange Pi without heatsink results in the following error after a few minutes: WRITE FAILURE: 0x00200000 != 0xffdfffff at offset 0x0137f47c (bitflip).
Also, the constructor repository (github/orangepi-xunlong) seems to contain that 624 Mhz clock speed in its u-boot fork. It may be that 672 Mhz is the advertized overclocked speed.
According to http://linux-sunxi.org/Orange_Pi_PC#DRAM_clock_speed_limit it may be worth decreasing that value with other Orange Pi boards.
See also e7d6aa0b74b7f4d08ee68da8a586c76c761348e2.
Applied to u-boot-sunxi/master
I've a 1.5 years old OPi Zero that was running uboot 2017.01 so far. I just tried to update to 2018.05, also using the new defconfig. The result was that uboot got stuck early during the first lines of output. Then I changed the DRAM clock back to 672, and things seem to work find.
Are we facing different board revisions that need different settings, or could the lowering be incorrect?
Sorry, false alarm. Seems I messed up something when writing out the 624 MHz version and happen to fix that while testing the 672 one. But now also the upstream config works. And I suspect/hope that the lower frequency will resolve the sporadic crashes of the board that I saw in the past...
Thanks, Jan
participants (3)
-
Clément Bœsch -
Jagan Teki -
Jan Kiszka