libsel4muslcsys: Add fstat syscall by dasch8-neutrality · Pull Request #100 · seL4/seL4_libs

dasch8-neutrality · 2025-03-21T16:30:25Z

The implementation only sets the file mode.
It is also only implemented for STDIN, STDOUT and STDERR.

This is to support libraries that insist on calling fstat on file descriptors before doing I/O.

The implementation only sets the file mode. It is also only implemented for STDIN, STDOUT and STDERR. This is to support libraries that insist on calling fstat on file descriptors before doing I/O. Signed-off-by: Daniel Schwyn <[email protected]>

musllibc does not define the fstat syscall number for some architectures (e.g. RISCV32). In addition, also install the syscall for the explicit 64bit version if its syscall number is defined. The implementation does not depend on the word size. Signed-off-by: Daniel Schwyn <[email protected]>

dasch8-neutrality · 2025-04-03T15:08:28Z

I've fixed the Gitlint issue with the 2nd commit.

lsf37

This looks fine from my side.

kent-mcleod · 2025-04-03T22:16:58Z

libsel4muslcsys/src/sys_io.c

+long sys_fstat(va_list ap)
+{
+    int fd = va_arg(ap, int);
+    struct stat *statbuf = va_arg(ap, struct stat *);


I think this type may need to be struct kstat because muslc uses a separate user type than the abi type it uses for the kernel.

Ah yes, I think you are right, thanks. This makes it a bit trickier though as the kstat.h header is not exported by musllibc. The cleanest solution I can think of is to add another interface library to the musllibc CMake build that has arch/${muslc_arch}/ and arch/generic as include directories. Probably copied to the muslc build directory to exclude the bits subdirectories that are already part of musllibc's public includes. What do you think?

Or alternatively to keep the changes local use the MUSLLIBC_CURRENT_DIR variable set in Findmusllibc.cmake to add the include directories.

I implemented the first option and opened the corresponding pull request in the muslc repo: seL4/musllibc#25

musllibc expects fstat to return the internal type struct kstat. Signed-off-by: Daniel Schwyn <[email protected]>

dasch8-neutrality · 2025-05-13T10:29:43Z

Is the library with the internal musllibc headers an acceptable solution? Let me know if there is something I should still do to get this merged.

dasch8-neutrality · 2025-11-11T10:41:22Z

Where are we on this? Let me know if something needs doing to merge seL4/musllibc#25 and this.

Indanz

A bit of a hack to get stdio working to be honest.

But I doubt that propagating this all the way to the UART driver or whoever takes care of stdio is any better/easier though.

Indanz · 2025-11-11T14:58:02Z

Not sure what I'm doing wrong with the "Test-with: seL4/musllibc#25", I thought that would fix the build errors.

lsf37 · 2025-11-11T21:10:12Z

Not sure what I'm doing wrong with the "Test-with: seL4/musllibc#25", I thought that would fix the build errors.

The syntax is Test with not Test-with. Have updated and restarted the build.

Indanz · 2025-11-11T23:46:59Z

Hmm, this fails sometimes (or always on Aarch64?) with:

creating config.mak... done
  cp: will not overwrite just-created '/github/workspace/build/apps/sel4test-driver/musllibc/build-temp/stage/include-internal/fp_arch.h' with '/github/workspace/projects/musllibc/arch/generic/fp_arch.h'
  ninja: build stopped: subcommand failed.

Indanz · 2025-11-11T23:50:15Z

The syntax is Test with not Test-with. Have updated and restarted the build.

Thanks. Is this documented anywhere btw? Now that other commit is merged, I removed the line again.

lsf37 · 2025-11-11T23:55:39Z

The syntax is Test with not Test-with. Have updated and restarted the build.

Thanks. Is this documented anywhere btw? Now that other commit is merged, I removed the line again.

Like all totally intuitive almost natural language commands of course this one didn't need any documentation, because we would never get it wrong :) By which I mean, I've been lazy and haven't found a good place to put it, so deferred it indefinitely.

Where would you first look for docs for it? Would the README of the ci-actions repo be a good place? Or would a page on the docsite be better?

Edit: the "spec", i.e. tests are here: https://github.com/seL4/ci-actions/blob/996e12ca25fde30ef6567c840cfd63697d202cb8/scripts/get-prs#L74

midnightveil · 2025-11-12T00:27:38Z

Where would you first look for docs for it? Would the README of the ci-actions repo be a good place? Or would a page on the docsite be better?

In the past I think I've gone to ci-actions. I don't tend to look at the docsite much, personally... Mostly because also I know that if I search the strings I'll find get-prs.

midnightveil · 2025-11-12T02:16:00Z

Hmm, this fails sometimes (or always on Aarch64?) with:

creating config.mak... done
  cp: will not overwrite just-created '/github/workspace/build/apps/sel4test-driver/musllibc/build-temp/stage/include-internal/fp_arch.h' with '/github/workspace/projects/musllibc/arch/generic/fp_arch.h'
  ninja: build stopped: subcommand failed.

I don't think this is caused by this, but I think was actually because of seL4/musllibc#25 as I saw it on another PR.

dasch8-neutrality · 2025-11-12T09:36:39Z

Hmm, this fails sometimes (or always on Aarch64?) with:
creating config.mak... done
  cp: will not overwrite just-created '/github/workspace/build/apps/sel4test-driver/musllibc/build-temp/stage/include-internal/fp_arch.h' with '/github/workspace/projects/musllibc/arch/generic/fp_arch.h'
  ninja: build stopped: subcommand failed.
I don't think this is caused by this, but I think was actually because of seL4/musllibc#25 as I saw it on another PR.

Ah, I overlooked that arch/generic contains some default headers that e.g. for aarch64 are replaced with architecture specific ones. I should have done the copy in two steps, first the generic headers and then the architecture specific ones so that the specific version wins. @Indanz should I file a new pull request in the muslc repo?

Indanz · 2025-11-12T10:43:18Z

I don't think this is caused by this, but I think was actually because of seL4/musllibc#25 as I saw it on another PR.

Ah, I overlooked that arch/generic contains some default headers that e.g. for aarch64 are replaced with architecture specific ones. I should have done the copy in two steps, first the generic headers and then the architecture specific ones so that the specific version wins. @Indanz should I file a new pull request in the muslc repo?

Yes please, otherwise I'll revert that other commit.

dasch8-neutrality · 2025-11-12T18:08:56Z

I don't think this is caused by this, but I think was actually because of seL4/musllibc#25 as I saw it on another PR.

Ah, I overlooked that arch/generic contains some default headers that e.g. for aarch64 are replaced with architecture specific ones. I should have done the copy in two steps, first the generic headers and then the architecture specific ones so that the specific version wins. @Indanz should I file a new pull request in the muslc repo?

Yes please, otherwise I'll revert that other commit.

I've opened a pull request with the fix: seL4/musllibc#27

libsel4muslcsys: Add fstat syscall

80aa494

The implementation only sets the file mode. It is also only implemented for STDIN, STDOUT and STDERR. This is to support libraries that insist on calling fstat on file descriptors before doing I/O. Signed-off-by: Daniel Schwyn <[email protected]>

lsf37 added the hw-build do all sel4test hardware builds on this PR label Mar 21, 2025

dasch8-neutrality force-pushed the master branch 2 times, most recently from 3ff410c to b28b020 Compare March 24, 2025 11:35

dasch8-neutrality force-pushed the master branch from b28b020 to a86970c Compare March 25, 2025 09:40

lsf37 approved these changes Apr 3, 2025

View reviewed changes

Merge branch 'master' into master

27f4520

lsf37 self-assigned this Apr 3, 2025

kent-mcleod reviewed Apr 3, 2025

View reviewed changes

libsel4muslcsys: fstat syscall returns kstat

079efd3

musllibc expects fstat to return the internal type struct kstat. Signed-off-by: Daniel Schwyn <[email protected]>

Indanz approved these changes Nov 11, 2025

View reviewed changes

Merge branch 'master' into master

aa97f79

kent-mcleod approved these changes Nov 11, 2025

View reviewed changes

midnightveil mentioned this pull request Nov 12, 2025

cmake: export internal headers seL4/musllibc#25

Merged

Conversation

dasch8-neutrality commented Mar 21, 2025 • edited by Indanz Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

dasch8-neutrality commented Apr 3, 2025

Uh oh!

lsf37 left a comment

Choose a reason for hiding this comment

Uh oh!

kent-mcleod Apr 3, 2025

Choose a reason for hiding this comment

Uh oh!

dasch8-neutrality Apr 4, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

dasch8-neutrality Apr 4, 2025

Choose a reason for hiding this comment

Uh oh!

dasch8-neutrality Apr 8, 2025

Choose a reason for hiding this comment

Uh oh!

dasch8-neutrality commented May 13, 2025

Uh oh!

dasch8-neutrality commented Nov 11, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Indanz left a comment

Choose a reason for hiding this comment

Uh oh!

Indanz commented Nov 11, 2025

Uh oh!

lsf37 commented Nov 11, 2025

Uh oh!

Indanz commented Nov 11, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Indanz commented Nov 11, 2025

Uh oh!

lsf37 commented Nov 11, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

midnightveil commented Nov 12, 2025

Uh oh!

midnightveil commented Nov 12, 2025

Uh oh!

dasch8-neutrality commented Nov 12, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Indanz commented Nov 12, 2025

Uh oh!

dasch8-neutrality commented Nov 12, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

dasch8-neutrality commented Mar 21, 2025 •

edited by Indanz

Loading

dasch8-neutrality Apr 4, 2025 •

edited

Loading

dasch8-neutrality commented Nov 11, 2025 •

edited

Loading

Indanz commented Nov 11, 2025 •

edited

Loading

lsf37 commented Nov 11, 2025 •

edited

Loading

dasch8-neutrality commented Nov 12, 2025 •

edited

Loading