posix: options: add custom Zephyr POSIX subprofile #88547

cfriedt · 2025-04-12T12:20:15Z

eventfd: bring config to top-level of posix dir, since it is not posix
posix: separate option groups into c library ext and system interfaces
posix: profiles: add custom Zephyr POSIX subprofile
posix: profiles: make POSIX_AEP_CHOICE_ZEPHYR the default
posix: options: add keep-sorted-start and -stop

Should only be merged after

Note

The sources used when CONFIG_POSIX_C_LANG_SUPPORT_R=y and CONFIG_POSIX_C_LIB_EXT=y can still be
overridden with CONFIG_TC_PROVIDES_POSIX_C_LANG_SUPPORT_R=y an CONFIG_TC_PROVIDES_POSIX_C_LIB_EXT=y (which must be used with select <option> in Kconfig), as they are primarily meant for toolchain maintainers.

A comparison when building samples/hello_world for native_sim.

Before:

wc -l build-main/zephyr/.config
   1295 build-main/zephyr/.config

After:

wc -l build-branch/zephyr/.config
   1220 build-branch/zephyr/.config

diff -Nua build-{main,branch}/zephyr/.config

--- build-main/zephyr/.config   2025-05-24 11:33:53.697011215 -0400
+++ build-branch/zephyr/.config 2025-05-24 11:34:27.140913656 -0400
@@ -900,98 +900,23 @@
 #
 # POSIX API Support
 #
-
-#
-# POSIX Options
-#
 # CONFIG_POSIX_API is not set
-CONFIG_POSIX_AEP_CHOICE_NONE=y
-# CONFIG_POSIX_AEP_CHOICE_BASE is not set
-# CONFIG_POSIX_AEP_CHOICE_PSE51 is not set
-# CONFIG_POSIX_AEP_CHOICE_PSE52 is not set
-# CONFIG_POSIX_AEP_CHOICE_PSE53 is not set
-# CONFIG_POSIX_ASYNCHRONOUS_IO is not set
-# CONFIG_POSIX_BARRIERS is not set
-# CONFIG_POSIX_C_LANG_SUPPORT_R is not set
-# CONFIG_POSIX_C_LIB_EXT is not set
-
-#
-# POSIX device I/O
-#
-# CONFIG_POSIX_DEVICE_IO is not set
-CONFIG_POSIX_OPEN_MAX=4
-# end of POSIX device I/O
-
-# CONFIG_POSIX_FD_MGMT is not set
-# CONFIG_POSIX_FILE_SYSTEM_R is not set
-# CONFIG_POSIX_FILE_SYSTEM is not set
-
-#
-# POSIX memory
-#
-CONFIG_POSIX_PAGE_SIZE=0x40
-# CONFIG_POSIX_SHARED_MEMORY_OBJECTS is not set
-# CONFIG_POSIX_MAPPED_FILES is not set
-# CONFIG_POSIX_MEMORY_PROTECTION is not set
-# end of POSIX memory
-
-# CONFIG_POSIX_MESSAGE_PASSING is not set
-# CONFIG_POSIX_SINGLE_PROCESS is not set
-# CONFIG_POSIX_MULTI_PROCESS is not set
-# CONFIG_POSIX_THREADS is not set
-# CONFIG_POSIX_READER_WRITER_LOCKS is not set
-
-#
-# POSIX scheduler options
-#
-# CONFIG_POSIX_PRIORITY_SCHEDULING is not set
-# end of POSIX scheduler options
-
-# CONFIG_POSIX_SEMAPHORES is not set
-
-#
-# POSIX signals
-#
-# CONFIG_POSIX_REALTIME_SIGNALS is not set
-# CONFIG_POSIX_SIGNALS is not set
-# end of POSIX signals
-
-# CONFIG_POSIX_SPIN_LOCKS is not set
-
-#
-# POSIX synchronized I/O
-#
-# CONFIG_POSIX_FSYNC is not set
-# CONFIG_POSIX_SYNCHRONIZED_IO is not set
-# end of POSIX synchronized I/O
-
-# CONFIG_POSIX_TIMERS is not set
-
-#
-# X/Open system interfaces
-#
-# CONFIG_XSI_REALTIME is not set
-# CONFIG_XSI_SINGLE_PROCESS is not set
-# CONFIG_XOPEN_STREAMS is not set
-# CONFIG_XSI_SYSTEM_LOGGING is not set
-# CONFIG_XSI_THREADS_EXT is not set
-# end of X/Open system interfaces
-
-#
-# Miscellaneous POSIX-related options
-#
-# CONFIG_EVENTFD is not set
-# end of Miscellaneous POSIX-related options
-
+# CONFIG_POSIX_AEP_CHOICE_NONE is not set
+CONFIG_POSIX_AEP_CHOICE_ZEPHYR=y
 CONFIG_TC_PROVIDES_POSIX_C_LANG_SUPPORT_R=y
-# end of POSIX Options
 
 #
-# POSIX Shell Utilities
+# POSIX C Library Extensions
 #
-# end of POSIX Shell Utilities
+CONFIG_POSIX_C_LANG_SUPPORT_R=y
+CONFIG_POSIX_C_LIB_EXT=y
+# CONFIG_GETOPT_LONG is not set
+# end of POSIX C Library Extensions
+
+# CONFIG_POSIX_SYSTEM_INTERFACES is not set
 # end of POSIX API Support
 
+# CONFIG_EVENTFD is not set
 # CONFIG_OPENAMP_RSC_TABLE is not set
 # CONFIG_SMF is not set
 CONFIG_LIBGCC_RTLIB=y

cfriedt · 2025-04-12T12:42:04Z

Probably one or two follow-up PRs can be made to ensure that the POSIX_C_LIB_EXT Option Group is complete. I think there are some string manipulation functions missing (not ISO C).

Currently, all of the implementation of POSIX_C_LANG_SUPPORT_R is done in libc/common, which ... doesn't matter too much, but that could be corrected later if necessary.

cfriedt · 2025-04-14T13:41:59Z

Shell utilities should probably also be gated by the new Kconfig.

lib/posix/Kconfig

jukkar

Thanks Chris, this helps to understand what is going on with the POSIX options.

jukkar · 2025-04-15T09:47:34Z

lib/posix/Kconfig.profile

+
+	  These functions are divided into two standalone Option Groups that may be enabled
+	  independently of the remainder of the POSIX API implementation; namely POSIX_C_LIB_EXT and
+	  POSIX_C_LANG_SUPPORT_R. If not referenced by the Zephyr kernel or application, There are no


Suggested change

POSIX_C_LANG_SUPPORT_R. If not referenced by the Zephyr kernel or application, There are no

POSIX_C_LANG_SUPPORT_R. If not referenced by the Zephyr kernel or application, there are no

cfriedt · 2025-05-14T10:10:46Z

I'll give this a rebase shorty

Indentation was fixed

cfriedt · 2025-05-25T03:35:35Z

Removed commits that should be part of a different PR.

cfriedt · 2025-05-25T13:36:18Z

Hmm... this was passing. Now SonarQube seems to be refusing to run. Maybe if I force push again?

ycsin · 2025-05-30T09:26:49Z

Hmm... this was passing. Now SonarQube seems to be refusing to run. Maybe if I force push again?

The commit message for tests: libc: thrd: use pthread barriers for qemu_x86_64/atom is too long

Commit dc81ee98cc:
35: UC4 Commit message body line exceeds max length (100>75): " twister -i --test-only -p qemu_x86_64/atom -s libraries.libc.c11_threads.picolibc >/dev/null 2>&1"

cfriedt · 2025-05-30T10:30:44Z

I fixed it locally, just haven't pushed. Thought I would give CI a break for a while

Previously, log messages would generate warnings (escalated to errors in CI) when CONFIG_TIMEOUT_64BIT=n. For example, ``` west build -p auto -b qemu_cortex_m0 -t run tests/posix/timers/ \ -- -DCONFIG_TIMEOUT_64BIT=n .. warning: format '%llx' expects argument of type \ 'long long unsigned int', but argument 3 has type 'k_ticks_t' \ {aka 'unsigned int'} ``` Use portable print specifiers and cast the argument to int64_t. Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>

Additional entries for the sys_clock API, comprised of: * sys_clock_gettime() * sys_clock_settime() * sys_clock_nanosleep() along with the constants * SYS_CLOCK_REALTIME * SYS_CLOCK_MONOTONIC * SYS_TIMER_ABSTIME The primary motivation for this API is so that libc and other libraries have a familiar-enough API to reach to when POSIX is not available, since POSIX is optional in Zephyr. By adding this API to lib/os, we also eliminate dependency cycles between libc and posix, as lib/os is a mutual dependency. Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>

Add bullets to release notes for sys_clock_gettime(), sys_clock_settime(), and sys_clock_nanosleep(). These changes were significant, as they make timekeeping dependencies more independent between libc and posix, and remove several unnecessary dependencies on posix for ISO C library routines. Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>

Use the newly added sys_clock_gettime() to avoid the unnecessary dependency on POSIX APIs. Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>

Remove POSIX clock_gettime() from the common libc time implementation, since POSIX should not be a dependency for ISO C. Instead, use the newly added lib/os sys_clock API. Specifically, sys_clock_gettime(). Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>

Use the implementation of time() from the common libc, since there it no longer pulls in POSIX. Use is implied for minimal, newlib, and picolibc, and selected for IAR. Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>

Reduce the dependency on POSIX by taking advantage of the newly added sys_clock_nanosleep(). Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>

The C standard does not guarantee that `thrd_success` is equal to zero, so ensure that the test takes that into account. Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>

The specification does not say that the first struct timespec passed to thrd_sleep() may be NULL. Preserving the testpoint so that a possible future testsuite can be run that expects faults to occur when undefined behaviour is used. Faults are one possible solution to handling undefined behaviour, but it would be good to come to a concensus about how that should be handled in Zephyr for standard function calls belonging to ISO C or POSIX. Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>

Use `timespec_from_timeout(K_MSEC(msec), &ts)` instead of leaning on lazily-crafted timespecs with invalid tv_nsec fields. Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>

A bit of a race condition seems to exist, but only for `qemu_x86_64/atom`. It's the same race that was attributed to scheduler noise or something along those lines when running under qemu. Most other architectures no longer see this issue, and it has not been required for some time in the `tests/benchmark/posix/threads` app, but here it seems to be pronounced again. Testing done: Before: We would see failures very reproducible, but only on this platform. Also, the failure has existed in `main` for some time, so is independent of the current PR. https://github.com/zephyrproject-rtos/zephyr/actions/runs/\ 15238463423/job/42862652065 After: ```shell COUNT=100 NOKAY=0 NFAIL=0 for i in $(seq 0 $((COUNT - 1))); do if [ $i -ne 0 ] && [ $((i % 10)) -eq 0 ]; then echo "Running test iteration $i" fi twister -i --test-only -p qemu_x86_64/atom -s \ libraries.libc.c11_threads.picolibc >/dev/null 2>&1 if [ $? -eq 0 ]; then NOKAY=$((NOKAY + 1)) else NFAIL=$((NFAIL + 1)) echo "Test failed on iteration $i" fi done PCT="$(echo "scale=0; ((100 * $NFAIL) / $COUNT) / 1" | bc)" echo "$NFAIL / $COUNT = $PCT % test failure rate" Running test iteration 10 Running test iteration 20 Running test iteration 30 Running test iteration 40 Running test iteration 50 Running test iteration 60 Running test iteration 70 Running test iteration 80 Running test iteration 90 0 / 100 = 0 % test failure rate ``` Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>

Presumably the time testsuite was separate from the c library set of testsuites because it had a depedency on POSIX. Since that dependency no longer exists, colocate the time testsuite with the other c library testsuites. Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>

Use the newly added sys_clock API in lib/os for * clock_gettime() * clock_settime() * clock_nanosleep() and nanosleep() * gettimeofday() Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>

The ISO C function time() should not depend on POSIX and this was corrected recently via the common libc time() implementation. Remove this dependency from the network subsystem where it has been unduly needed for some time. Similarly, XSI_SINGLE_PROCESS was a dependency for time() via picolibc, because the time() implementation in picolibc relies on the POSIX gettimeofday() call. However, since Zephyr's common libc time() implementation no longer depends on that, it can be removed as a dependency in the network subsystem as well. Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>

Remove POSIX_TIMERS and XSI_SINGLE_PROCESS dependencies from the aws cloud sample and the lwm2m client sample, as they are no longer required. Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>

The eventfd configuration does not need to be so deeply nested within POSIX since it does not depend on POSIX completely. Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>

Separate the POSIX implementation into two categories: - Extensions to ISO C - System Interfaces The first category include standalone functions that generally do not require OS support or depend on any other features within the POSIX specification. The Option Groups that comprise this category include - POSIX_C_LIB_EXT: e.g. strnlen(), fnmatch() - POSIX_C_LANG_SUPPORT_R: e.g. gmtime_r(), strtok_r() The second category includes the majority of other POSIX Option Groups that do require OS support. The latter group may also be categorized generally as being NATIVE_LIBC_INCOMPATIBLE, although that might eventually become more granular. Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>

Add a custom Zephyr POSIX subprofile specifically for enabling the default features that Zephyr requires as per the coding guidelines. Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>

Default POSIX_AEP_CHOICE to POSIX_AEP_CHOICE_ZEPHYR Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>

Add zephyr-keep-sorted-start and zephyr-keep-sorted-stop comments. Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>

sonarqubecloud · 2025-05-31T12:52:35Z

Quality Gate passed

Issues
35 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

cfriedt force-pushed the add-zephyr-posix-aep-choice branch from 0d193e5 to f68b0fb Compare April 12, 2025 12:31

cfriedt requested review from nashif, JordanYates and ycsin April 12, 2025 12:37

cfriedt added the area: POSIX POSIX API Library label Apr 12, 2025

cfriedt assigned ycsin Apr 12, 2025

cfriedt force-pushed the add-zephyr-posix-aep-choice branch 4 times, most recently from e891f53 to 10c619d Compare April 12, 2025 15:29

JordanYates previously requested changes Apr 14, 2025

View reviewed changes

lib/posix/Kconfig Show resolved Hide resolved

jukkar reviewed Apr 15, 2025

View reviewed changes

cfriedt force-pushed the add-zephyr-posix-aep-choice branch 5 times, most recently from db9804b to 2ca347b Compare April 25, 2025 03:31

cfriedt force-pushed the add-zephyr-posix-aep-choice branch 2 times, most recently from aeb78be to 7ac5797 Compare April 29, 2025 01:15

cfriedt mentioned this pull request May 3, 2025

posix: implement xsi_single_process and posix_clock_selection #89068

Merged

cfriedt force-pushed the add-zephyr-posix-aep-choice branch 3 times, most recently from 04b8fa8 to df478b2 Compare May 18, 2025 03:52

cfriedt force-pushed the add-zephyr-posix-aep-choice branch from 02a72f0 to b1e87fc Compare May 25, 2025 03:25

cfriedt force-pushed the add-zephyr-posix-aep-choice branch 2 times, most recently from 6f20e45 to 3b2d0af Compare May 26, 2025 22:51

cfriedt added 20 commits May 31, 2025 08:22

logging: use sys_clock_gettime()

0bba722

Use the newly added sys_clock_gettime() to avoid the unnecessary dependency on POSIX APIs. Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>

libc: use the common libc time() implementation for most libcs

cbfc8ab

Use the implementation of time() from the common libc, since there it no longer pulls in POSIX. Use is implied for minimal, newlib, and picolibc, and selected for IAR. Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>

libc: common: thrd: use sys_clock_nanosleep() instead of nanosleep()

558501a

Reduce the dependency on POSIX by taking advantage of the newly added sys_clock_nanosleep(). Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>

tests: libc: thrd: compare with thrd_success rather than ok or zero

ae6101a

The C standard does not guarantee that `thrd_success` is equal to zero, so ensure that the test takes that into account. Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>

tests: libc: thrd: use timespec_from_timeout()

e9c8ce5

Use `timespec_from_timeout(K_MSEC(msec), &ts)` instead of leaning on lazily-crafted timespecs with invalid tv_nsec fields. Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>

posix: use sys_clock implementation

6565eaf

Use the newly added sys_clock API in lib/os for * clock_gettime() * clock_settime() * clock_nanosleep() and nanosleep() * gettimeofday() Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>

samples: net: remove POSIX_TIMERS and XSI_SINGLE_PROCESS

1094d07

Remove POSIX_TIMERS and XSI_SINGLE_PROCESS dependencies from the aws cloud sample and the lwm2m client sample, as they are no longer required. Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>

eventfd: bring config to top-level of posix dir, since it is not posix

d1f678a

The eventfd configuration does not need to be so deeply nested within POSIX since it does not depend on POSIX completely. Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>

posix: profiles: add custom Zephyr POSIX subprofile

e704b4a

Add a custom Zephyr POSIX subprofile specifically for enabling the default features that Zephyr requires as per the coding guidelines. Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>

posix: profiles: make POSIX_AEP_CHOICE_ZEPHYR the default

bb33d26

Default POSIX_AEP_CHOICE to POSIX_AEP_CHOICE_ZEPHYR Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>

posix: options: add keep-sorted-start and -stop

764da0b

Add zephyr-keep-sorted-start and zephyr-keep-sorted-stop comments. Signed-off-by: Chris Friedt <cfriedt@tenstorrent.com>

cfriedt force-pushed the add-zephyr-posix-aep-choice branch from 3b2d0af to 764da0b Compare May 31, 2025 12:22

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

posix: options: add custom Zephyr POSIX subprofile #88547

posix: options: add custom Zephyr POSIX subprofile #88547

cfriedt commented Apr 12, 2025 •

edited

Loading

Uh oh!

cfriedt commented Apr 12, 2025

Uh oh!

cfriedt commented Apr 14, 2025

Uh oh!

Uh oh!

jukkar left a comment

Uh oh!

jukkar Apr 15, 2025

Uh oh!

cfriedt Apr 18, 2025

Uh oh!

cfriedt commented May 14, 2025

Uh oh!

cfriedt commented May 25, 2025

Uh oh!

cfriedt commented May 25, 2025

Uh oh!

ycsin commented May 30, 2025

Uh oh!

cfriedt commented May 30, 2025

Uh oh!

sonarqubecloud bot commented May 31, 2025

Uh oh!

Uh oh!

	POSIX_C_LANG_SUPPORT_R. If not referenced by the Zephyr kernel or application, There are no
	POSIX_C_LANG_SUPPORT_R. If not referenced by the Zephyr kernel or application, there are no

posix: options: add custom Zephyr POSIX subprofile #88547

Are you sure you want to change the base?

posix: options: add custom Zephyr POSIX subprofile #88547

Conversation

cfriedt commented Apr 12, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

cfriedt commented Apr 12, 2025

Uh oh!

cfriedt commented Apr 14, 2025

Uh oh!

Uh oh!

jukkar left a comment

Choose a reason for hiding this comment

Uh oh!

jukkar Apr 15, 2025

Choose a reason for hiding this comment

Uh oh!

cfriedt Apr 18, 2025

Choose a reason for hiding this comment

Uh oh!

cfriedt commented May 14, 2025

Uh oh!

cfriedt commented May 25, 2025

Uh oh!

cfriedt commented May 25, 2025

Uh oh!

ycsin commented May 30, 2025

Uh oh!

cfriedt commented May 30, 2025

Uh oh!

sonarqubecloud bot commented May 31, 2025

Quality Gate passed

Uh oh!

Uh oh!

cfriedt commented Apr 12, 2025 •

edited

Loading