| commit | ed84a8f5406e35a6f5c88ac30ee003e62f264a14 | |
| author | Julius Werner <jwerner@chromium.org> | |
| Mon, 6 Apr 2015 20:51:46 +0000 (6 13:51 -0700) | ||
| committer | Patrick Georgi <pgeorgi@google.com> | |
| Wed, 22 Apr 2015 06:58:32 +0000 (22 08:58 +0200) | ||
| tree | f9ebebd40ac9f0a9c909b15ad1245a2ab0fb29ec | treesnapshot (tar.gz zip) |
| parent | 5725ea30ddadf35099e5cdd5074663f3ca9b939b | commitdiff |
rockchip/rk3288: Fix operator precedence error in LPDDR init
Upstream coreboot regularly runs Coverity over the code base. Turns out
that's a good idea since it's really easy to screw yourself over with a
missing parenthesis and some unfortunately deceptive line breaking.
This patch fixes a bug in LPDDR3 initialization due to an incorrect
operator precedence assumption ( ?: does not bind stronger than | ). In
effect, instead of setting MR11[1:0] to 0b11 or 0b00 based on ODT, we're
unconditionally setting MR0[1:0] to 0b11. Thankfully, MR0[1:0] seems to
contain read-only bits so this might have not been a problem when ODT is
off (which is currently true for all LPDDR boards).
Also adding a redundant LPDDR_OP() around the 0 to make the intent
clearer and changing 3 and 0 to 0x3 and 0x0 to make it more obvious that
these are bit masks (right?).
BRANCH=veyron
BUG=None
TEST=Running reboot loop on a Minnie, looks good so far...
Change-Id: I06464aaa57e693b1973846a5771162244f7a1c57
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Found-by: Coverity Scan
Original-Commit-Id: 5bd9eba39fb7b0f940fead963bbc1878b031b2cb
Original-Change-Id: I701ce059472078b5de09a45dd31f54b65a51e641
Original-Signed-off-by: Julius Werner <jwerner@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/264135
Original-Reviewed-by: David Hendricks <dhendrix@chromium.org>
Original-Reviewed-by: Jinkun Hong <jinkun.hong@rock-chips.com>
Original-Tested-by: Jinkun Hong <jinkun.hong@rock-chips.com>
Reviewed-on: http://review.coreboot.org/9911
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Upstream coreboot regularly runs Coverity over the code base. Turns out
that's a good idea since it's really easy to screw yourself over with a
missing parenthesis and some unfortunately deceptive line breaking.
This patch fixes a bug in LPDDR3 initialization due to an incorrect
operator precedence assumption ( ?: does not bind stronger than | ). In
effect, instead of setting MR11[1:0] to 0b11 or 0b00 based on ODT, we're
unconditionally setting MR0[1:0] to 0b11. Thankfully, MR0[1:0] seems to
contain read-only bits so this might have not been a problem when ODT is
off (which is currently true for all LPDDR boards).
Also adding a redundant LPDDR_OP() around the 0 to make the intent
clearer and changing 3 and 0 to 0x3 and 0x0 to make it more obvious that
these are bit masks (right?).
BRANCH=veyron
BUG=None
TEST=Running reboot loop on a Minnie, looks good so far...
Change-Id: I06464aaa57e693b1973846a5771162244f7a1c57
Signed-off-by: Patrick Georgi <pgeorgi@chromium.org>
Found-by: Coverity Scan
Original-Commit-Id: 5bd9eba39fb7b0f940fead963bbc1878b031b2cb
Original-Change-Id: I701ce059472078b5de09a45dd31f54b65a51e641
Original-Signed-off-by: Julius Werner <jwerner@chromium.org>
Original-Reviewed-on: https://chromium-review.googlesource.com/264135
Original-Reviewed-by: David Hendricks <dhendrix@chromium.org>
Original-Reviewed-by: Jinkun Hong <jinkun.hong@rock-chips.com>
Original-Tested-by: Jinkun Hong <jinkun.hong@rock-chips.com>
Reviewed-on: http://review.coreboot.org/9911
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
| src/soc/rockchip/rk3288/sdram.c | diffblobblamehistory |