summaryrefslogtreecommitdiff
path: root/.clang-format
diff options
context:
space:
mode:
authorNicolas Boichat <drinkcat@chromium.org>2018-12-12 11:30:51 +0800
committerchrome-bot <chrome-bot@chromium.org>2018-12-18 04:28:56 -0800
commit9f4852b7e9bbceb1fadb04eb68da43037f0b09d5 (patch)
treee3bed99f9248d40ec81aa9020c1439a53594a5d3 /.clang-format
parente5ddc233ebc2285e3d145c6206c86c483bfbae11 (diff)
downloadchrome-ec-9f4852b7e9bbceb1fadb04eb68da43037f0b09d5.tar.gz
core/cortex-m*/task: Record 32-bit exception times
"time in exceptions" looks unreasonable after > 4294s (when 32-bit microsecond timer wraps around). This is because core/cortex-m/task.c:svc_handler records time exc_start_time just before the interrupt handler for 32-bit timer interrupt runs, so the high word of the system clock (clksrc_high) is not updated yet (while the low 32-bits already wrapped around). After the handler runs, clksrc_high is updated, so there appear to be a 4294s gap between the 2 measurements. Fix this by recording the low 32-bit timer value only. There will never be more than 4294s between exceptions, and this fixes the wrap-around issue as well. BRANCH=none BUG=chromium:914208 TEST=Flash kukui (cortex-m0) and kukui_scp (cortex-m), let system run for 4300+s, no more accounting error in "Time in exceptions". Change-Id: If52855ef093ac1a1d38432555694c83742feb8f1 Signed-off-by: Nicolas Boichat <drinkcat@chromium.org> Reviewed-on: https://chromium-review.googlesource.com/1372876
Diffstat (limited to '.clang-format')
0 files changed, 0 insertions, 0 deletions