Skip to contentSkip to conversation

uninative: Add uninative - a way of reusing native/cross over multiple distros

RFC PATCH In review Richard Purdie 3 participants 3 messages Aug 28, 2014 – Oct 8, 2014

lore
Share & tools
lore
#0 lore raw

RFC PATCH 1/1 uninative: Add uninative - a way of reusing native/cross over multiple distros

These patches illustrate a new idea, a way of allowing a single set of cross/native sstate to work over multiple distros, even old ones.

The assumption is that our own C library is basically up to date. We build and share a small tarball (~2MB) of a prebuilt copy of this along with a patchelf binary (which sadly is C++ based so libstdc++ is in there). This tarball can be generated from our usual SDK generation process through the supplied recipe, uninative-tarball.

At the start of the build, if its not been extracted into the sysroot, this tarball is extracted there and configured for the specified path.

When we install binaries from a "uninative" sstate feed, we change the dynamic loader to point at this dynamic loader and C librbary. This works exactly the same way as our relocatable SDK does. The only real difference is a switch to use patchelf, so even if the interpreter section is too small, it can still adjust the binary.

Right now the patch is a working proof of concept. If you build the tarball and place it at the head of the tree (in COREBASE), you can run a build from sstate and successfully build packages and construct images.

There is some cleanup needed, its hardcoded for x86_64 right now, its trivial to add 32 bit support too. The tarball isn't fetched right now, there is just a harcoded path assumption and there is no error handling. I haven't figured out the best delivery mechanism for that yet. BuildStarted is probably not the right event to hook on either.

I did want to send this out and illustrate how with a small change, we might make the native/cross sstate much more reusable and hence improve the accessibility of lower overhead builds. With this change, its possible the Yocto Project may be able to support a configured sstate mirror out the box. This also has positive implications for our developer workflow/SDK improvements.

meta/classes/uninative.bbclass new file +47 -0
new file mode 100644
index 0000000..862c0de
@@ -0,0 +1,47 @@
+#BUILD_LDFLAGS += "-Wl,-T${LINKERSCRIPT}"
+NATIVELSBSTRING = "universal"
+#LINKERSCRIPT = "${STAGING_DIR_NATIVE}/linkerscript"
+
+addhandler uninative_eventhandler
+uninative_eventhandler[eventmask] = "bb.event.BuildStarted"
+
+python uninative_eventhandler() {
+    loader = e.data.getVar("UNINATIVE_LOADER", True)
+    if not os.path.exists(loader):
+        import subprocess
+        cmd = e.data.expand("mkdir -p ${STAGING_DIR}; cd ${STAGING_DIR}; tar -xjf ${COREBASE}/${BUILD_ARCH}-nativesdk-libc.tar.bz2; ${STAGING_DIR}/relocate_sdk.py ${STAGING_DIR_NATIVE} ${UNINATIVE_LOADER} ${UNINATIVE_LOADER} ${STAGING_BINDIR_NATIVE}/patchelf-uninative")
+        #bb.warn("nativesdk lib extraction: " + cmd)
+        subprocess.check_call(cmd, shell=True)
+}
+
+SSTATEPOSTUNPACKFUNCS_append = " uninative_changeinterp"
+
+UNINATIVE_LOADER = "${STAGING_DIR_NATIVE}/lib/ld-linux-x86-64.so.2"
+
+python uninative_changeinterp () {
+    import subprocess
+    import stat
+    import oe.qa
+
+    if not (bb.data.inherits_class('native', d) or bb.data.inherits_class('crosssdk', d) or bb.data.inherits_class('cross', d)):
+        return
+
+    sstateinst = d.getVar('SSTATE_INSTDIR', True)
+    for walkroot, dirs, files in os.walk(sstateinst):
+        for file in files:
+            f = os.path.join(walkroot, file)
+            if os.path.islink(f):
+                continue
+            s = os.stat(f)
+            if not ((s[stat.ST_MODE] & stat.S_IXUSR) or (s[stat.ST_MODE] & stat.S_IXGRP) or (s[stat.ST_MODE] & stat.S_IXOTH)):
+                continue
+            elf = oe.qa.ELFFile(f)
+            try:
+                elf.open()
+            except:
+                continue
+
+            #bb.warn("patchelf-uninative --set-interpreter %s %s" % (d.getVar("UNINATIVE_LOADER", True), f))
+            subprocess.call("patchelf-uninative --set-interpreter %s %s" % (d.getVar("UNINATIVE_LOADER", True), f), shell=True)
+}
+
meta/recipes-core/meta/uninative-tarball.bb new file +48 -0

48 lines not on this page · raw message

#1 lore raw
Quoted 8 lines · Richard Purdie

These patches illustrate a new idea, a way of allowing a single set of cross/native sstate to work over multiple distros, even old ones.

The assumption is that our own C library is basically up to date. We build
and share a small tarball (~2MB) of a prebuilt copy of this along with a
patchelf binary (which sadly is C++ based so libstdc++ is in there).

Begs for patchelf built against uClibc++ for the build.

Otherwise I like the idea!
Thanks,

#2 lore raw
In reply to Richard Purdie
Quoted 41 lines · Richard Purdie

These patches illustrate a new idea, a way of allowing a single set of cross/native sstate to work over multiple distros, even old ones.

The assumption is that our own C library is basically up to date. We build and share a small tarball (~2MB) of a prebuilt copy of this along with a patchelf binary (which sadly is C++ based so libstdc++ is in there). This tarball can be generated from our usual SDK generation process through the supplied recipe, uninative-tarball.

At the start of the build, if its not been extracted into the sysroot, this tarball is extracted there and configured for the specified path.

When we install binaries from a "uninative" sstate feed, we change the dynamic loader to point at this dynamic loader and C librbary. This works exactly the same way as our relocatable SDK does. The only real difference is a switch to use patchelf, so even if the interpreter section is too small,
it can still adjust the binary.

Right now the patch is a working proof of concept. If you build the tarball and place it at the head of the tree (in COREBASE), you can run a build from
sstate and successfully build packages and construct images.

There is some cleanup needed, its hardcoded for x86_64 right now, its trivial
to add 32 bit support too. The tarball isn't fetched right now, there is just a
harcoded path assumption and there is no error handling. I haven't figured out the best delivery mechanism for that yet. BuildStarted is probably not the right event to hook on either.

I did want to send this out and illustrate how with a small change, we might
make the native/cross sstate much more reusable and hence improve the accessibility
of lower overhead builds. With this change, its possible the Yocto Project may
be able to support a configured sstate mirror out the box. This also has positive implications for our developer workflow/SDK improvements.

Quoted 2 lines

Hey Richard,

This looks extremely promising. I tried using it, however, and ran into all manner of problems. The first one is relatively easy to resolve, and I have a local patch which seems to do so (though I then run into different build failures): binaries which are built and then run, before any relocation patchelf bits are run (e.g. autoconf ./conftest), can fail miserably due to the mismatch between the libc they linked against in the sysroot and the host dynamic linker, unless you add -Wl,--dynamic-linker=${UNINATIVE_LOADER} or so to BUILD_LDFLAGS.

Signature · 5 lines
Christopher Larson
clarson at kergoth dot com
Founder - BitBake, OpenEmbedded, OpenZaurus
Maintainer - Tslib
Senior Software Engineer, Mentor Graphics