summaryrefslogtreecommitdiff
path: root/src/mainboard/ibm/e326
diff options
context:
space:
mode:
authorjinkun.hong <jinkun.hong@rock-chips.com>2015-01-21 16:03:43 +0800
committerStefan Reinauer <stefan.reinauer@coreboot.org>2015-04-15 22:09:55 +0200
commitd4a227ba05aadf049cec0b4e4edc8345f476c2f4 (patch)
tree2a6b6b260b8d0cae28a4094c6dcbe080e5bb13aa /src/mainboard/ibm/e326
parent29bd9e2c74b7d4c381cae240d561bad291a67c71 (diff)
rk3288: Fix failing DDR3 reboot test
We want a reset signal to last 200us. The length of a reset signal is represented by BIT0~BIT16 in DDR_PUBL_PTR2. When DDR memory runs at 667MHz, the calculated value for the reset signal is 0x20850, which is bigger than the maximum value that can be described with 17 bits (0x1ffff). As a result, the memory controller only sees 0x850, which generates a 3.5us reset cycle instead, which violates the standard and negatively impacts memory stability. So instead, we now set it to the maximum value (0x1ffff) to prevent this overflow, resulting in a reset signal of 196us for 667MHz DDR memory. BUG=chrome-os-partner:34875 TEST=loop reboot BRANCH=veyron Change-Id: Ia01f8a0414b49fa3ecf4d543cfa1822e29ee4cc4 Signed-off-by: Stefan Reinauer <reinauer@chromium.org> Original-Commit-Id: 767a4a3cb8dff47cb15064d335b78ffa5815914d Original-Change-Id: I9b410e1605c87f12a5ca96ead12f8527ca4f417f Original-Signed-off-by: jinkun.hong <jinkun.hong@rock-chips.com> Original-Reviewed-on: https://chromium-review.googlesource.com/242175 Original-Reviewed-by: Julius Werner <jwerner@chromium.org> Reviewed-on: http://review.coreboot.org/9653 Tested-by: build bot (Jenkins) Reviewed-by: Patrick Georgi <pgeorgi@google.com>
Diffstat (limited to 'src/mainboard/ibm/e326')
0 files changed, 0 insertions, 0 deletions