coccinelle: standardize kernel timeout arguments #19562
Conversation
pabigot
requested review from
Mani-Sadhasivam,
MaureenHelm,
Mierunski,
Vudentz,
anangl,
andrewboie,
finikorg,
himanshujha199640,
jakub-uC,
jfischer-no,
jhedberg,
joerchan,
jukkar,
loicpoulain,
nashif,
nordic-krch,
pfalcon,
rlubos and
tbursztyka
as code owners
zephyrbot
added
area: Sensors
zephyrbot
added
area: Tests
|
All checks are passing now. Tip: The bot edits this comment instead of posting a new one, so you can check the comment's history to see earlier messages. |
|
No complaints here beyond whines about a bunch of extra work, I guess. Is the intent to get this in first and then have me rebase all the collisions in #17155 on top? |
pabigot
force-pushed
the
dnm/cocci-timeout
branch
from
87d2bf2 to
211125d
Compare
Short answer: no, the intent is to get this in, then revisit the goals of #17155 and proceed in a more sustainable manner. The longer answer is here. |
drivers/bluetooth/hci/userchan.c
Outdated
There was a problem hiding this comment.
This is calling a libc API for the native posix build (i.e. the host libc), so this change is wrong (it will break once the macro expands to something else than an integer).
There was a problem hiding this comment.
Thanks! Mis-placed parentheses in the regular expression caused the conversion to affect poll instead of k_poll. Fixed now.
Some legacy code still passes integer literals in milliseconds as the value to functions that take a timeout. This usage interferes with plans to replace the millisecond representation with a more generic k_timeout_t value. Add a Coccinelle script to convert call sites to use the proper constants and macros. Signed-off-by: Peter Bigot <[email protected]>
Use the int_literal_to_timeout Coccinelle script to convert literal integer arguments for kernel API timeout parameters to the standard timeout value representations. Signed-off-by: Peter Bigot <[email protected]>
pabigot
force-pushed
the
dnm/cocci-timeout
branch
from
211125d to
b040b66
Compare
|
What's the process for removing the shippable override patch without triggering CI again, which will then fail? |
|
pabigot
force-pushed
the
dnm/cocci-timeout
branch
from
3548645 to
b040b66
Compare
|
@galak Thanks, done. Looks like history of the success has disappeared, but it was https://app.shippable.com/github/zephyrproject-rtos/zephyr/runs/52375/summary/console. |
merged |
10 participants
Legacy code uses integer literal delays measured in milliseconds as arguments to kernel API that is intended to take timeout values. Replace these integers with the corresponding defined constants and macro conversions that represent the timeout values.
This is a demonstration and prototype of using semantic patch technology to repeatably perform the conversions necessary to support #17155, done in a way that is reviewable, verifiable, and easy to re-apply when pending work is merged to master or rebases are required.