RFE: Sync to Linux 6.12 syscall definitions #435

xen0n · 2024-08-20T04:27:34Z

The arch-syscall-validate script updates will need to be backported to v2.5.x later, for continuing to sync aarch64 and riscv64 that got refactored to source the syscall definitions from a syscall.tbl instead of asm-generic/unistd.h.

coveralls · 2024-08-20T04:57:52Z

coverage: 89.454%. remained the same
when pulling d909801 on xen0n:linux-6.11-syscalls
into 9c4eb5e on seccomp:main.

xen0n · 2024-08-20T05:34:39Z

After the change, linux-headers >= 6.11 will be required for ~~native LoongArch~~ all builds, for definitions of new syscalls introduced in this version.

lixing-star · 2024-10-08T03:22:36Z

After the change, linux-headers >= 6.11 will be required for ~~native LoongArch~~ all builds, for definitions of new syscalls introduced in this version.

this is really helpful for LoongArch, thanks.

Dandan336 · 2024-10-17T08:26:06Z

Hi maintainers,

About defining __ARCH_WANT_NEW_STAT in unistd.h
'LoongArch: Define __ARCH_WANT_NEW_STAT in unistd.h' 1 was merged in linux upstream 2.
'LoongArch: Define __ARCH_WANT_NEW_STAT in unistd.h' was supported in the stable version of linux, as follows,

The results of the test cases
Based on main branch of libseccomp, I pulled this pull-request 'RFE: Sync to Linux 6.11 syscall definitions' submitted by xen0n. I built libseccomp project on Debian loong64 rootfs (the verson of linux-libc-dev is 6.11.2-1), there is 0 failed or errored test case.The results of the test cases are as follows,

Regression Test Summary
 tests run: 5084
 tests skipped: 124
 tests passed: 5084
 tests failed: 0
 tests errored: 0
=================================
PASS: regression
=============
1 test passed
=============

Please review.

yetist · 2024-10-28T07:08:07Z

@drakenclimber @pcmoore

Could you review this PR? Thanks.
Also, I would like to know when v2.6.0 will be released.

rusty-snake · 2024-10-28T07:15:35Z

Also, I would like to know when v2.6.0 will be released.

https://github.com/seccomp/libseccomp?tab=readme-ov-file#release-process

TL;DR: When it's ready.

pcmoore · 2024-10-30T18:29:52Z

include/seccomp-syscalls.h

+#else
+#define __SNR_map_shadow_stack		__PNR_map_shadow_stack
+#endif
+


Isn't map_shadow_stack() defined for all the arches?

Yes it is, but in this commit the PNR was allocated anyway and the change is already shipped in v2.5.5. So if we're to consider PNR values as part of the public API we would have to keep __PNR_map_shadow_stack defined, and for stylistic consistency we would have to keep this #ifdef-fery as well.

Thanks for the reminder @xen0n.

In the manpages we "highly recommend" the use of the SCMP_SYS() macro as opposed to directly specifying syscall numbers so it's a bit of a gray area. I think it would be okay for us to modify/remove the PNR syscall values if necessary, but we probably shouldn't do so unless we have a good reason. Considering everything, it looks like you approach of leaving the PNR values as-is is likely the right one. Thanks :)

pcmoore · 2024-10-30T18:33:19Z

Just one small comment/clarification/question above, but otherwise I think this looks good. Thanks for all your work on this @xen0n!

rusty-snake · 2024-10-30T18:47:08Z

include/seccomp-syscalls.h

@@ -280,6 +280,9 @@
 #define __PNR_atomic_barrier			-10246
 #define __PNR_atomic_cmpxchg_32			-10247
 #define __PNR_getpagesize			-10248
+#define __PNR_map_shadow_stack			-10249


FTR: The PNR numbers below -10244 (__PNR_memfd_secret) have already diverged between libseccomp:main and libseccomp:release-2.5.

As a FYI, I'm okay with fixing that sort of thing while merging the PR.

I have preserved the __PNR_map_shadow_stack position while rebasing.

Apart from de-duplication of logic, this refactor is also going to help syncing to the Linux 6.11+ definitions, where all architectures are converted to source their syscall definitions from syscall.tbl files. The change is tested on Linux 6.2 sources to not affect the generated syscalls.csv apart from timestamp changes. Signed-off-by: WANG Xuerui <git@xen0n.name>

The aarch64, loongarch64 and riscv64 architectures have their syscall table sources changed to scripts/syscall.tbl, from the original inclusion of asm-generic/unistd.h. Make the script recognize the new format for these architectures. Signed-off-by: WANG Xuerui <git@xen0n.name>

Due to the addition of fstat & newfstatat to the LoongArch syscall ABI, tests 38 and 55 have to be updated for the changed syscall numbers. As for the PNR additions, normally they are allocated alphabetically for the syscalls introduced between updates of the table, but in the v2.5 release branch -10245 is already assigned to map_shadow_stack in commit 53267af ("all: update the syscall table for Linux v6.7-rc3"). While the map_shadow_stack syscall is in fact available across all architectures, for consistency with v2.5.5 and later it is kept in the same position in this update. Signed-off-by: WANG Xuerui <git@xen0n.name>

xen0n · 2024-10-31T04:11:34Z

This is now rebased and updated to reflect Linux v6.12-rc5 syscall info (no new syscalls since 6.11). I have also incorporated @rusty-snake's review comment of keeping PNR allocations consistent with v2.5.x. Tests still pass on aarch64, loongarch64, riscv64 and x86_64.

drakenclimber · 2024-11-08T22:38:48Z

My apologies for being quiet the last few weeks. I've had some other high-priority work on my plate, but I'm thinking I should have some time in the next few weeks to get back into libseccomp work. Thanks all for your patience :).

pcmoore · 2024-11-08T23:26:42Z

This looked good to me and it passed all my tests locally so I merged the PR at HEAD f01e675, thanks for you work and patience @xen0n!

@drakenclimber nothing in here looked controversial and we needed something like this for future syscall updates so I went ahead and merged the PR, if you notice anything awry please let me know and we can revert/fix it.

pcmoore · 2024-11-08T23:28:00Z

While the PR is merged, I'm going to leave this open until I finish the backport to the release-2.5 branch.

pcmoore · 2024-11-11T18:26:17Z

Here is the release-2.5/v2.5.x branch backport: #439

pcmoore · 2024-11-11T18:27:08Z

With the backport PR now posted, I'm closing this PR, thanks again everyone!

xen0n force-pushed the linux-6.11-syscalls branch from aecd5e9 to f36e149 Compare August 20, 2024 04:53

xen0n changed the title ~~Sync to Linux 6.11 syscall definitions~~ RFE: Sync to Linux 6.11 syscall definitions Aug 20, 2024

xen0n force-pushed the linux-6.11-syscalls branch from f36e149 to f4afd16 Compare August 20, 2024 05:13

xen0n mentioned this pull request Aug 20, 2024

Content suggestion for This Week in LoongArch newsletter / 《每周一龙》新闻线索信箱 loongson-community/areweloongyet#16

Open

pcmoore added enhancement pending/review labels Sep 5, 2024

drakenclimber requested review from drakenclimber and pcmoore September 20, 2024 18:03

pcmoore added this to the v2.6.0 milestone Sep 20, 2024

Flat mentioned this pull request Oct 22, 2024

RFE: add the uretprobe() syscall to the list for kernel 6.11.* #438

Closed

pcmoore reviewed Oct 30, 2024

View reviewed changes

rusty-snake reviewed Oct 30, 2024

View reviewed changes

xen0n added 3 commits October 31, 2024 11:26

xen0n force-pushed the linux-6.11-syscalls branch from f4afd16 to d909801 Compare October 31, 2024 03:50

xen0n changed the title ~~RFE: Sync to Linux 6.11 syscall definitions~~ RFE: Sync to Linux 6.12 syscall definitions Oct 31, 2024

pcmoore mentioned this pull request Nov 11, 2024

RFE: update v2.5.x syscall tables to v6.12-rc7 #439

Open

pcmoore closed this Nov 11, 2024

xen0n deleted the linux-6.11-syscalls branch November 12, 2024 02:34

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

RFE: Sync to Linux 6.12 syscall definitions #435

RFE: Sync to Linux 6.12 syscall definitions #435

xen0n commented Aug 20, 2024

coveralls commented Aug 20, 2024 •

edited

Loading

xen0n commented Aug 20, 2024 •

edited

Loading

lixing-star commented Oct 8, 2024

Dandan336 commented Oct 17, 2024

yetist commented Oct 28, 2024

rusty-snake commented Oct 28, 2024

pcmoore Oct 30, 2024

xen0n Oct 31, 2024

pcmoore Nov 8, 2024

pcmoore commented Oct 30, 2024

rusty-snake Oct 30, 2024

pcmoore Oct 30, 2024

xen0n Oct 31, 2024

xen0n commented Oct 31, 2024

drakenclimber commented Nov 8, 2024

pcmoore commented Nov 8, 2024

pcmoore commented Nov 8, 2024

pcmoore commented Nov 11, 2024

pcmoore commented Nov 11, 2024

RFE: Sync to Linux 6.12 syscall definitions #435

RFE: Sync to Linux 6.12 syscall definitions #435

Conversation

xen0n commented Aug 20, 2024

coveralls commented Aug 20, 2024 • edited Loading

xen0n commented Aug 20, 2024 • edited Loading

lixing-star commented Oct 8, 2024

Dandan336 commented Oct 17, 2024

yetist commented Oct 28, 2024

rusty-snake commented Oct 28, 2024

pcmoore Oct 30, 2024

Choose a reason for hiding this comment

xen0n Oct 31, 2024

Choose a reason for hiding this comment

pcmoore Nov 8, 2024

Choose a reason for hiding this comment

pcmoore commented Oct 30, 2024

rusty-snake Oct 30, 2024

Choose a reason for hiding this comment

pcmoore Oct 30, 2024

Choose a reason for hiding this comment

xen0n Oct 31, 2024

Choose a reason for hiding this comment

xen0n commented Oct 31, 2024

drakenclimber commented Nov 8, 2024

pcmoore commented Nov 8, 2024

pcmoore commented Nov 8, 2024

pcmoore commented Nov 11, 2024

pcmoore commented Nov 11, 2024

coveralls commented Aug 20, 2024 •

edited

Loading

xen0n commented Aug 20, 2024 •

edited

Loading