ci: update musl headers to 6.6 #3921

tammela · 2024-09-13T15:19:35Z

Update the musl headers in CI to use alpine-linux instead of sabotage-linux.
Alpine also uses musl but follows the linux stable releases, providing more up-to-date headers.

I also took the opportunity to clean up libc-test/build.rs a bit.

rustbot · 2024-09-13T15:19:40Z

r? @tgross35

rustbot has assigned @tgross35.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

bors · 2024-11-06T20:48:23Z

☔ The latest upstream changes (presumably #3922) made this pull request unmergeable. Please resolve the merge conflicts.

bors · 2024-11-16T06:14:16Z

☔ The latest upstream changes (presumably #4042) made this pull request unmergeable. Please resolve the merge conflicts.

tgross35

Sorry this hasn't been on my radar. Is the switch to Alpine just because of better maintenance?

Suggested some ways to clean up the shell script, also I think there are a couple other musl dockerfiles that will need to be updated now (loongarch?). libc-test/build.rs is probably pretty out of date, if you'd like we can merge the musl changes first and then clean up build.rs in a followup.

tgross35 · 2024-11-19T06:36:38Z

ci/install-musl.sh

+# Download, configure, build, and install musl-sanitized kernel headers.
+# This routine piggybacks on: https://git.alpinelinux.org/aports/tree/main/linux-headers?h=3.20-stable
+# Alpine follows stable kernel releases, 3.20 uses Linux 6.6 headers.
+git clone -n --depth=1 --filter=tree:0 -b 3.20-stable https://gitlab.alpinelinux.org/alpine/aports


Nit to make use of variables

Suggested change

# Download, configure, build, and install musl-sanitized kernel headers.

# This routine piggybacks on: https://git.alpinelinux.org/aports/tree/main/linux-headers?h=3.20-stable

# Alpine follows stable kernel releases, 3.20 uses Linux 6.6 headers.

git clone -n --depth=1 --filter=tree:0 -b 3.20-stable https://gitlab.alpinelinux.org/alpine/aports

# Download, configure, build, and install musl-sanitized kernel headers.

# Alpine follows stable kernel releases, 3.20 uses Linux 6.6 headers.

alpine_version=3.20

alpine_git=https://gitlab.alpinelinux.org/alpine/aports

# This routine piggybacks on: https://git.alpinelinux.org/aports/tree/main/linux-headers?h=3.20-stable

git clone -n --depth=1 --filter=tree:0 -b "${alpine_version}-stable" "$alpine_git"

tgross35 · 2024-11-19T07:09:09Z

ci/install-musl.sh

+   cp APKBUILD APKBUILD.source
+   cp APKBUILD APKBUILD.sha512
+   {
+   echo "printf \"\$source\""
+   # shellcheck disable=SC2028
+   echo "printf \"\$_kernver\n\""
+   # shellcheck disable=SC2028
+   echo "printf \"\$pkgver\n\""
+   } >> APKBUILD.source
+   echo "printf \"\$sha512sums\"" >> APKBUILD.sha512
+   KERNEL_VER=$(bash APKBUILD.source | tail -2 | head -1 | tr -d "[:space:]")
+   PKGVER=$(bash APKBUILD.source | tail -1 | tr -d "[:space:]")
+   urls=$(bash APKBUILD.source | grep -o 'https.*')
+   kernel=""
+   patch=""
+   for url in $urls; do
+      base=$(basename "$url")
+      curl --retry 5 -L "$url" > "$base"
+      case $base in
+         linux-*) kernel=$base;;
+         patch-*) patch=$base;;
+      esac
+   done
+   bash APKBUILD.sha512 | grep "$kernel" >> sha-check
+   bash APKBUILD.sha512 | grep "$patch" >> sha-check


I think this gets a bit cleaner with heredoc and writing the outputs to files rather than filtering:

Suggested change

cp APKBUILD APKBUILD.source

cp APKBUILD APKBUILD.sha512

{

echo "printf \"\$source\""

# shellcheck disable=SC2028

echo "printf \"\$_kernver\n\""

# shellcheck disable=SC2028

echo "printf \"\$pkgver\n\""

} >> APKBUILD.source

echo "printf \"\$sha512sums\"" >> APKBUILD.sha512

KERNEL_VER=$(bash APKBUILD.source | tail -2 | head -1 | tr -d "[:space:]")

PKGVER=$(bash APKBUILD.source | tail -1 | tr -d "[:space:]")

urls=$(bash APKBUILD.source | grep -o 'https.*')

kernel=""

patch=""

for url in $urls; do

base=$(basename "$url")

curl --retry 5 -L "$url" > "$base"

case $base in

linux-*) kernel=$base;;

patch-*) patch=$base;;

esac

done

bash APKBUILD.sha512 | grep "$kernel" >> sha-check

bash APKBUILD.sha512 | grep "$patch" >> sha-check

# Create a version of APKBUILD to extract variables

cp APKBUILD APKBUILD.vars

cat <<- 'EOF' >> APKBUILD.vars

printf "$source" > alpine-source

printf "$_kernver" > alpine-kernver

printf "$pkgver" > alpine-pkgver

printf "$sha512sums" > alpine-sha512sums

EOF

# Execute the build script to write the output files

sh APKBUILD.vars

alpine_source="$(cat alpine-source)"

kernel_version="$(cat alpine-kernver)"

pkg_version="$(cat alpine-pkgver)"

kernel=""

patch=""

urls="$(echo "$alpine_source" | grep -o 'https.*')"

# Download

for url in $urls; do

base=$(basename "$url")

curl --retry 5 -L "$url" > "$base"

case "$base" in

linux-*) kernel="$base";;

patch-*) patch="$base";;

*)

echo "Unexpected basename $base"

exit 1

;;

esac

done

grep "$kernel" alpine-sha512sums >> sha-check

grep "$patch" alpine-sha512sums >> sha-check

(I haven't tested this)

Also, does CARCH need to be set before running APKBUILD?

Also, does CARCH need to be set before running APKBUILD?

Not in this case, we force the arch to our own $kernel_arch

tammela · 2024-11-19T15:37:08Z

Sorry this hasn't been on my radar. Is the switch to Alpine just because of better maintenance?

Sabotage is pretty much stuck on 4.19.88 for years now. Alpine follows the stable releases from www.kernel.org and uses musl.
As you saw, this is pretty much a hack to piggy back on Alpine's patches for the kernel headers.
I wouldn't claim it's better maintenance over sabotage, but rather it's a better fit since we want to test and add the shiniest new features :)

I will update the PR to take into account your suggestions

tgross35

Sabotage is pretty much stuck on 4.19.88 for years now. Alpine follows the stable releases from www.kernel.org and uses musl. As you saw, this is pretty much a hack to piggy back on Alpine's patches for the kernel headers. I wouldn't claim it's better maintenance over sabotage, but rather it's a better fit since we want to test and add the shiniest new features :)

Thanks for the context, that makes a lot of sense 👍

Two small suggestions for the script then this looks good to me. One other nit - it seems like your bash is indented by three spaces, mind changing to 4 like the rest of the scripts?

tgross35 · 2024-11-19T23:26:24Z

ci/install-musl.sh

+   {
+   echo "printf \"\$source\" > alpine-source"
+   # shellcheck disable=SC2028
+   echo "printf \"\$_kernver\n\" > alpine-kernver"
+   # shellcheck disable=SC2028
+   echo "printf \"\$pkgver\n\" > alpine-pkgver"
+   # shellcheck disable=SC2028
+   echo "printf \"\$sha512sums\n\" > alpine-sha512sums"
+   } >> APKBUILD.vars


Could this just use heredoc rather than the nested quoting? Which shouldn't hit the shellcheck warnings

tgross35 · 2024-11-19T23:42:20Z

ci/install-musl.sh

+   urls=$(grep -o 'https.*' alpine-source)
+   kernel=""
+   patch=""
+   for url in $urls; do
+      base=$(basename "$url")
+      curl --retry 5 -L "$url" > "$base"
+      case $base in
+         linux-*) kernel=$base;;
+         patch-*) patch=$base;;
+      esac
+   done


Does the following work instead?

Suggested change

urls=$(grep -o 'https.*' alpine-source)

kernel=""

patch=""

for url in $urls; do

base=$(basename "$url")

curl --retry 5 -L "$url" > "$base"

case $base in

linux-*) kernel=$base;;

patch-*) patch=$base;;

esac

done

urls=$(grep -o 'https.*' alpine-source)

for url in $urls; do

base=$(basename "$url")

# We only need to download `linux` and the patch archive, not individual

# patches.

case "$base" in

linux-*) kernel="$base";;

patch-*) patch="$base";;

*) continue;;

esac

curl --retry 5 -L "$url" -o "$base"

grep "$base" alpine-sha512sums >> sha-check

done

Then remove the grep "$kernel/patch" alpine-sha512sums >> sha-check lines below. This makes sure we don't download/use anything that we don't verify the sha of.

If we do actually need to download all the files (revert-broken-uapi.patch and 0003-remove-inclusion-of-sysinfo.h-in-kernel.h.patch), then remove *) continue and >> sha-check, instead just sha512sum -c alpine-sha512sums.

We need the latter. I added an explicit check to avoid downloading files that do not contain a checksum.

bors · 2024-11-20T09:00:34Z

☔ The latest upstream changes (presumably #4120) made this pull request unmergeable. Please resolve the merge conflicts.

Update the musl headers in CI to use alpine-linux instead of sabotage-linux. Alpine also uses musl but follows the linux stable releases, providing more up-to-date headers. Signed-off-by: Pedro Tammela <[email protected]>

Now that we have Linux 6.6 we can clean up some of the test exceptions. Not all of them can be removed, so the comments are updated if needed. Signed-off-by: Pedro Tammela <[email protected]>

tgross35

Great, thanks for getting this working!

Update the musl headers in CI to use alpine-linux instead of sabotage-linux. Alpine also uses musl but follows the linux stable releases, providing more up-to-date headers. Signed-off-by: Pedro Tammela <[email protected]> (backport <rust-lang#3921>) (cherry picked from commit acc75e7) Signed-off-by: Trevor Gross <[email protected]>

Now that we have Linux 6.6 we can clean up some of the test exceptions. Not all of them can be removed, so the comments are updated if needed. Signed-off-by: Pedro Tammela <[email protected]> (backport <rust-lang#3921>) (cherry picked from commit 55f9451) Signed-off-by: Trevor Gross <[email protected]>

Update the musl headers in CI to use alpine-linux instead of sabotage-linux. Alpine also uses musl but follows the linux stable releases, providing more up-to-date headers. Signed-off-by: Pedro Tammela <[email protected]> (backport <rust-lang#3921>) (cherry picked from commit acc75e7) Signed-off-by: Trevor Gross <[email protected]>

Now that we have Linux 6.6 we can clean up some of the test exceptions. Not all of them can be removed, so the comments are updated if needed. Signed-off-by: Pedro Tammela <[email protected]> (backport <rust-lang#3921>) (cherry picked from commit 55f9451) Signed-off-by: Trevor Gross <[email protected]>

rustbot assigned tgross35 Sep 13, 2024

rustbot added the S-waiting-on-review label Sep 13, 2024

tammela force-pushed the musl-headers branch 5 times, most recently from 3d166ca to d14a03b Compare September 13, 2024 19:00

tammela marked this pull request as ready for review September 13, 2024 19:01

tammela mentioned this pull request Sep 30, 2024

Add SO_PREFER_BUSY_POLL and SO_BUSY_POLL_BUDGET #3917

Open

tgross35 reviewed Nov 19, 2024

View reviewed changes

tgross35 added S-waiting-on-author and removed S-waiting-on-review labels Nov 19, 2024

tammela force-pushed the musl-headers branch 4 times, most recently from ae76331 to 9575471 Compare November 19, 2024 16:59

tgross35 requested changes Nov 19, 2024

View reviewed changes

tammela force-pushed the musl-headers branch 2 times, most recently from 77523c0 to 4c6fc50 Compare November 21, 2024 15:59

ci: update musl headers to Linux 6.6

acc75e7

Update the musl headers in CI to use alpine-linux instead of sabotage-linux. Alpine also uses musl but follows the linux stable releases, providing more up-to-date headers. Signed-off-by: Pedro Tammela <[email protected]>

tammela force-pushed the musl-headers branch from 4c6fc50 to b87db4d Compare November 21, 2024 16:03

ci: clean up musl exceptions

55f9451

Now that we have Linux 6.6 we can clean up some of the test exceptions. Not all of them can be removed, so the comments are updated if needed. Signed-off-by: Pedro Tammela <[email protected]>

tammela force-pushed the musl-headers branch from b87db4d to 55f9451 Compare November 21, 2024 16:13

tgross35 approved these changes Nov 21, 2024

View reviewed changes

tgross35 added this pull request to the merge queue Nov 21, 2024

tgross35 added the stable-nominated This PR should be considered for cherry-pick to libc's stable release branch label Nov 21, 2024

tgross35 mentioned this pull request Nov 21, 2024

add ptp_clock_caps #4128

Merged

3 tasks

Merged via the queue into rust-lang:main with commit 0a8f04a Nov 21, 2024
45 checks passed

This was referenced Nov 25, 2024

[0.2] Backports #4139

Merged

[0.2] ci: update musl headers to 6.6 #4140

Closed

[0.2] Backports #4142

Merged

tgross35 added stable-applied This PR has been cherry-picked to libc's stable release branch and removed stable-nominated This PR should be considered for cherry-pick to libc's stable release branch labels Nov 25, 2024

tgross35 mentioned this pull request Nov 25, 2024

Expose len8_dlc field of can_frame struct. #3357

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

ci: update musl headers to 6.6 #3921

ci: update musl headers to 6.6 #3921

tammela commented Sep 13, 2024 •

edited

Loading

rustbot commented Sep 13, 2024

bors commented Nov 6, 2024

bors commented Nov 16, 2024

tgross35 left a comment

tgross35 Nov 19, 2024

tgross35 Nov 19, 2024 •

edited

Loading

tgross35 Nov 19, 2024

tammela Nov 19, 2024

tammela commented Nov 19, 2024 •

edited

Loading

tgross35 left a comment

tgross35 Nov 19, 2024

tgross35 Nov 19, 2024

tammela Nov 21, 2024

bors commented Nov 20, 2024

tgross35 left a comment

ci: update musl headers to 6.6 #3921

ci: update musl headers to 6.6 #3921

Conversation

tammela commented Sep 13, 2024 • edited Loading

rustbot commented Sep 13, 2024

bors commented Nov 6, 2024

bors commented Nov 16, 2024

tgross35 left a comment

Choose a reason for hiding this comment

tgross35 Nov 19, 2024

Choose a reason for hiding this comment

tgross35 Nov 19, 2024 • edited Loading

Choose a reason for hiding this comment

tgross35 Nov 19, 2024

Choose a reason for hiding this comment

tammela Nov 19, 2024

Choose a reason for hiding this comment

tammela commented Nov 19, 2024 • edited Loading

tgross35 left a comment

Choose a reason for hiding this comment

tgross35 Nov 19, 2024

Choose a reason for hiding this comment

tgross35 Nov 19, 2024

Choose a reason for hiding this comment

tammela Nov 21, 2024

Choose a reason for hiding this comment

bors commented Nov 20, 2024

tgross35 left a comment

Choose a reason for hiding this comment

tammela commented Sep 13, 2024 •

edited

Loading

tgross35 Nov 19, 2024 •

edited

Loading

tammela commented Nov 19, 2024 •

edited

Loading