diff options
author | Daisuke Nojiri <dnojiri@chromium.org> | 2018-04-26 12:59:58 -0700 |
---|---|---|
committer | Patrick Georgi <pgeorgi@google.com> | 2018-04-30 06:27:24 +0000 |
commit | ccfa18feff26211631be53fe1db1b8b0b534cd79 (patch) | |
tree | 45ea7f928c6931d8b8c9cfa7804ab1cb3418fa89 /gnat.adc | |
parent | c01a9ab56244dbb6d93e3be128ef1fadccc9ed2b (diff) |
cros-ec: Avoid infinitely looping in google_chromeec_pd_get_amode
Currently, google_chromeec_pd_get_amode infinitely loops if a TCPC port
is connected to a device with alternate mode(s) and the call is made
for the mode with the index higher than 0 (e.g. Zinger).
Cros EC manages alternative modes entered in an array (amode[]). The
command is designed to accept a query for an particular index and a
particular SVID.
Zinger has a 'Google' mode. It's stored in amode[0]. When AP queries
first time for DisplayPort with index=0, EC says 'no' as expected.
AP sends the next query with index=1 but EC_CMD_PROTO_VERSION (0x00)
is sent instead because cmd_code is cleared by google_chromeec_command.
res.svid is supposed to be 0 when EC hits the last index + 1 but
res.svid is set to 2 by the EC_CMD_PROTO_VERSION handler because
EC_PROTO_VERSION is currently 2. So, the call succeeds and AP goes to
the next index and this repeats forever.
Any USB-C device with non-DisplayPort alternate mode can cause this
hang unless HDMI port is used.
This patch resets all the fields of chromeec_command in each iteration
in case google_chromeec_command changes them.
BUG=b:78630899
BRANCH=none
TEST=Verify Fizz boots without monitors on Zinger. Verify the svid
enumeration happens as expected.
Change-Id: I388ed4bdfac9176d8e690c429e99674ed267004f
Signed-off-by: Daisuke Nojiri <dnojiri@chromium.org>
Reviewed-on: https://review.coreboot.org/25878
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: Aaron Durbin <adurbin@chromium.org>
Reviewed-by: Furquan Shaikh <furquan@google.com>
Diffstat (limited to 'gnat.adc')
0 files changed, 0 insertions, 0 deletions