Skip to contentSkip to conversation

fix the SDK kernel module test

PATCH Superseded Trevor Woerner 2 participants 5 messages Aug 5 – Aug 22

v1 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

cover fix the SDK kernel module test

The kernel module test in the SDK test suite does not pass today.

The cryptodev revision it builds no longer builds against kernel 6.18, and the kernel's host tools are compiled with a pkg-config that the SDK environment has pointed at the target sysroot, so a host compile is handed target include paths and fails under -Werror. OE-Core already solves the second one for its own kernel builds, where kernel.bbclass passes HOSTPKG_CONFIG=pkg-config-native; the SDK simply had no equivalent.

These are independent of any SDK feature work and fix a test that is broken in master today.

Trevor Woerner (2):
oeqa/sdk: update cryptodev to a revision that builds on kernel 6.18 toolchain-scripts, oeqa/sdk: build the kernel's host tools with a host
pkg-config

 meta/classes-recipe/toolchain-scripts.bbclass |  3 +++
 meta/lib/oeqa/sdk/cases/kmod.py               | 11 +++++++----
 2 files changed, 10 insertions(+), 4 deletions(-)
Signature · 1 line
2.50.0.173.g8b6f19ccfc3a
#1 lore raw

PATCH 1/2 oeqa/sdk: update cryptodev to a revision that builds on kernel 6.18

The out-of-tree kernel module test pins a cryptodev-linux revision from before the 1.14 release, which predates the removal of crypto_ahash_alignmask() from the kernel and so no longer compiles:

cryptlib.c:384:28: error: implicit declaration of function 'crypto_ahash_alignmask' [-Wimplicit-function-declaration]

Move to upstream's "Fix build for Linux 6.18-rc1", which is the revision that restores compatibility.

AI-Generated: codex/claude-opus 5 (xhigh)

---
 meta/lib/oeqa/sdk/cases/kmod.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
meta/lib/oeqa/sdk/cases/kmod.py +2 -2
index 0c4d8ddb543f..1be23e994fd2 100644
@@ -31,8 +31,8 @@ class KernelModuleTest(OESDKTestCase):
 
         with tempfile.TemporaryDirectory(prefix="cryptodev", dir=self.tc.sdk_dir) as testdir:
             git_url = "https://github.com/cryptodev-linux/cryptodev-linux"
-            # This is a knnown-good commit post-1.13 that builds with kernel 6.7+
-            git_sha = "bb8bc7cf60d2c0b097c8b3b0e807f805b577a53f"
+            # This is a known-good commit post-1.14 that builds with kernel 6.18+
+            git_sha = "08644db02d43478f802755903212f5ee506af73b"
 
             sourcedir = os.path.join(testdir, "cryptodev-linux")
             subprocess.check_output(["git", "clone", git_url, sourcedir], stderr=subprocess.STDOUT)
Signature · 1 line
2.50.0.173.g8b6f19ccfc3a
#2 lore raw

PATCH 2/2 toolchain-scripts, oeqa/sdk: build the kernel's host tools with a host pkg-config

In reply to Trevor Woerner

The SDK environment sets PKG_CONFIG_SYSROOT_DIR, PKG_CONFIG_PATH and PKG_CONFIG_LIBDIR so that pkg-config answers for the target. That is right for everything built to run there and wrong for a program built to run on the SDK host.

The kernel build compiles host tools of its own, and objtool asks pkg-config where libelf is. It gets the target's include directory, compiles a host tool against the target's C library headers, and stops on warnings that -Wno-system-headers would otherwise have covered:

usr/include/sys/cdefs.h:486: error: "__attribute_const__" redefined

Export a HOSTPKG_CONFIG that undoes those three variables, and pass the same value on the make command line when the test builds an external module. Both halves are needed: the kernel assigns HOSTPKG_CONFIG with '=', so only a command line assignment takes effect today, and the export is there for kernels that assign it with '?=', which is proposed in

https://lore.kernel.org/linux-kbuild/[email protected]/

OE-Core already does the command line half for its own kernel builds, in kernel.bbclass.

Fixes [YOCTO #16239].

AI-Generated: codex/claude-opus 5 (xhigh)

---
 meta/classes-recipe/toolchain-scripts.bbclass | 3 +++
 meta/lib/oeqa/sdk/cases/kmod.py               | 7 +++++--
 2 files changed, 8 insertions(+), 2 deletions(-)
meta/classes-recipe/toolchain-scripts.bbclass +3 -0
index d94dcdee39b3..54d05c242c4b 100644
@@ -63,6 +63,9 @@ toolchain_create_sdk_env_script () {
 	echo "export PATH=$sdkpathnative$bindir:$sdkpathnative$sbindir:$sdkpathnative$base_bindir:$sdkpathnative$base_sbindir:$sdkpathnative$bindir/../${HOST_SYS}/bin:$sdkpathnative$bindir/${TARGET_SYS}"$EXTRAPATH':"$PATH"' >> $script
 	echo 'export PKG_CONFIG_SYSROOT_DIR=$SDKTARGETSYSROOT' >> $script
 	echo 'export PKG_CONFIG_PATH=$SDKTARGETSYSROOT'"$libdir"'/pkgconfig:$SDKTARGETSYSROOT'"$prefix"'/share/pkgconfig' >> $script
+	# pkg-config here answers for the target, which is wrong for the host
+	# tools the kernel build compiles. Honoured by kernels using '?='.
+	echo 'export HOSTPKG_CONFIG="env -u PKG_CONFIG_SYSROOT_DIR -u PKG_CONFIG_PATH -u PKG_CONFIG_LIBDIR pkg-config"' >> $script
 	echo 'export CONFIG_SITE=${SDKPATH}/site-config-'"${multimach_target_sys}" >> $script
 	echo "export OECORE_NATIVE_SYSROOT=\"$sdkpathnative\"" >> $script
 	echo 'export OECORE_TARGET_SYSROOT="$SDKTARGETSYSROOT"' >> $script
meta/lib/oeqa/sdk/cases/kmod.py +5 -2
index 1be23e994fd2..1d73f076e552 100644
@@ -26,8 +26,11 @@ class KernelModuleTest(OESDKTestCase):
         parallel_make = "-j %d" % (pmv) if pmv else ""
 
         self.ensure_target_package("kernel-devsrc")
+        # the kernel assigns HOSTPKG_CONFIG with '=', so it has to come from
+        # the make command line, as kernel.bbclass does
+        host_pkg_config = 'HOSTPKG_CONFIG="env -u PKG_CONFIG_SYSROOT_DIR -u PKG_CONFIG_PATH -u PKG_CONFIG_LIBDIR pkg-config"'
         # These targets need to be built before kernel modules can be built.
-        self._run("make %s -C $OECORE_TARGET_SYSROOT/usr/src/kernel prepare scripts" % (parallel_make))
+        self._run("make %s %s -C $OECORE_TARGET_SYSROOT/usr/src/kernel prepare scripts" % (parallel_make, host_pkg_config))
 
         with tempfile.TemporaryDirectory(prefix="cryptodev", dir=self.tc.sdk_dir) as testdir:
             git_url = "https://github.com/cryptodev-linux/cryptodev-linux"
@@ -39,5 +42,5 @@ class KernelModuleTest(OESDKTestCase):
             self.assertTrue(os.path.isdir(sourcedir))
             subprocess.check_output(["git", "-C", sourcedir, "checkout", git_sha], stderr=subprocess.STDOUT)
 
-            self._run("make -C %s V=1 KERNEL_DIR=$OECORE_TARGET_SYSROOT/usr/src/kernel" % sourcedir)
+            self._run("make -C %s V=1 %s KERNEL_DIR=$OECORE_TARGET_SYSROOT/usr/src/kernel" % (sourcedir, host_pkg_config))
             self.check_elf(os.path.join(sourcedir, "cryptodev.ko"))
Signature · 1 line
2.50.0.173.g8b6f19ccfc3a
#3 lore raw

Re: [OE-core] [PATCH 2/2] toolchain-scripts, oeqa/sdk: build the kernel's host tools with a host pkg-config

Quoted 24 lines · Trevor Woerner via lists.openembedded.org

The SDK environment sets PKG_CONFIG_SYSROOT_DIR, PKG_CONFIG_PATH and PKG_CONFIG_LIBDIR so that pkg-config answers for the target. That is right for everything built to run there and wrong for a program built to run on the SDK host.

The kernel build compiles host tools of its own, and objtool asks pkg-config where libelf is. It gets the target's include directory, compiles a host tool against the target's C library headers, and stops on warnings that -Wno-system-headers would otherwise have covered:

  usr/include/sys/cdefs.h:486: error: "__attribute_const__" redefined

Export a HOSTPKG_CONFIG that undoes those three variables, and pass the same value on the make command line when the test builds an external module. Both halves are needed: the kernel assigns HOSTPKG_CONFIG with '=', so only a command line assignment takes effect today, and the export is there for kernels that assign it with '?=', which is proposed in

  https://lore.kernel.org/linux-kbuild/[email protected]/

OE-Core already does the command line half for its own kernel builds, in kernel.bbclass.

Quoted 8 lines

Fixes [YOCTO #16239].

AI-Generated: codex/claude-opus 5 (xhigh)

---
 meta/classes-recipe/toolchain-scripts.bbclass | 3 +++ meta/lib/oeqa/sdk/cases/kmod.py               | 7 +++++--

 2 files changed, 8 insertions(+), 2 deletions(-)

The kmod test piece of this is probably ok, as you say, we do that in the kernel classes. I do note that the kernel class redirects to a pkg- config wrapper and perhaps the kmod class should match, rather than doing something different though?

The toolchain piece is a but more of a concern since we're encoding kernel specific variables into the generic toolchain environment file and in general we don't want to do recipe specific things there...

Cheers,

Richard

#4 lore raw

Re: [OE-core] [PATCH 2/2] toolchain-scripts, oeqa/sdk: build the kernel's host tools with a host pkg-config

Quoted 34 lines · Richard Purdie
Quoted 24 lines · Trevor Woerner via lists.openembedded.org

The SDK environment sets PKG_CONFIG_SYSROOT_DIR, PKG_CONFIG_PATH and PKG_CONFIG_LIBDIR so that pkg-config answers for the target. That is right for everything built to run there and wrong for a program built to run on the SDK host.

The kernel build compiles host tools of its own, and objtool asks pkg-config where libelf is. It gets the target's include directory, compiles a host tool against the target's C library headers, and stops on warnings that -Wno-system-headers would otherwise have covered:

  usr/include/sys/cdefs.h:486: error: "__attribute_const__" redefined

Export a HOSTPKG_CONFIG that undoes those three variables, and pass the same value on the make command line when the test builds an external module. Both halves are needed: the kernel assigns HOSTPKG_CONFIG with '=', so only a command line assignment takes effect today, and the export is there for kernels that assign it with '?=', which is proposed in

  https://lore.kernel.org/linux-kbuild/[email protected]/

OE-Core already does the command line half for its own kernel builds, in kernel.bbclass.

Quoted 8 lines

Fixes [YOCTO #16239].

AI-Generated: codex/claude-opus 5 (xhigh)

---
 meta/classes-recipe/toolchain-scripts.bbclass | 3 +++ meta/lib/oeqa/sdk/cases/kmod.py               | 7 +++++--

 2 files changed, 8 insertions(+), 2 deletions(-)
Quoted 4 lines

The kmod test piece of this is probably ok, as you say, we do that in the kernel classes. I do note that the kernel class redirects to a pkg- config wrapper and perhaps the kmod class should match, rather than doing something different though?

Agreed. I just sent a v3 that does that, so the test now passes the same wrapper rather than spelling it out.

Quoted 3 lines

The toolchain piece is a but more of a concern since we're encoding kernel specific variables into the generic toolchain environment file and in general we don't want to do recipe specific things there...

Yes, and it is dropped. My v2 crossed with this email and already does this.

Quoted 3 lines

Cheers,

Richard