| Submitter | Tom Zanussi |
|---|---|
| Date | June 13, 2011, 2:35 p.m. |
| Message ID | <1307975748.10825.2.camel@elmorro> |
| Download | mbox | patch |
| Permalink | /patch/5745/ |
| State | New, archived |
| Headers | show |
Comments
Op 13 jun 2011, om 16:35 heeft Tom Zanussi het volgende geschreven: > -COMPATIBLE_MACHINE = "(qemux86|qemux86-64|qemuppc|emenlow|crownbay|atom-pc|n450)" > +COMPATIBLE_MACHINE = "(qemux86|qemux86-64|qemuppc|atom-pc)" After having looked at the recipe (basically 'inherit autotools', PACKAGE_ARCH = BASE_PACKAGE_ARCH) and reading the systemtap wiki I can't figure out why COMPATIBLE_MACHINE is actually needed. The closest I get is "needs specific kernel config", but that argument quickly falls down when looking at other recipes that need that (e.g. udev) and considering out layer strategy (bbappending it for all your machines). So, what am I missing here? regards, Koen PS: there's no 'atom-pc' in OE-core
Op 13 jun 2011, om 16:47 heeft Koen Kooi het volgende geschreven: > > Op 13 jun 2011, om 16:35 heeft Tom Zanussi het volgende geschreven: >> -COMPATIBLE_MACHINE = "(qemux86|qemux86-64|qemuppc|emenlow|crownbay|atom-pc|n450)" >> +COMPATIBLE_MACHINE = "(qemux86|qemux86-64|qemuppc|atom-pc)" > > After having looked at the recipe (basically 'inherit autotools', PACKAGE_ARCH = BASE_PACKAGE_ARCH) and reading the systemtap wiki I can't figure out why COMPATIBLE_MACHINE is actually needed. The closest I get is "needs specific kernel config", but that argument quickly falls down when looking at other recipes that need that (e.g. udev) and considering out layer strategy (bbappending it for all your machines). > > So, what am I missing here? And PACKAGE_ARCH = BASE_PACKAGE_ARCH coupled with COMPATIBLE_MACHINE doesn't make much sense either.
On Mon, 2011-06-13 at 07:47 -0700, Koen Kooi wrote: > Op 13 jun 2011, om 16:35 heeft Tom Zanussi het volgende geschreven: > > -COMPATIBLE_MACHINE = "(qemux86|qemux86-64|qemuppc|emenlow|crownbay|atom-pc|n450)" > > +COMPATIBLE_MACHINE = "(qemux86|qemux86-64|qemuppc|atom-pc)" > > After having looked at the recipe (basically 'inherit autotools', PACKAGE_ARCH = BASE_PACKAGE_ARCH) and reading the systemtap wiki I can't figure out why COMPATIBLE_MACHINE is actually needed. The closest I get is "needs specific kernel config", but that argument quickly falls down when looking at other recipes that need that (e.g. udev) and considering out layer strategy (bbappending it for all your machines). > > So, what am I missing here? > My testing showed systemtap doesn't actually work on arm, and has no support for mips - the machines listed are the only ones I've been able to verify that work. Tom > regards, > > Koen > > PS: there's no 'atom-pc' in OE-core > _______________________________________________ > Openembedded-core mailing list > Openembedded-core@lists.openembedded.org > http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Op 13 jun 2011, om 16:54 heeft Tom Zanussi het volgende geschreven: > On Mon, 2011-06-13 at 07:47 -0700, Koen Kooi wrote: >> Op 13 jun 2011, om 16:35 heeft Tom Zanussi het volgende geschreven: >>> -COMPATIBLE_MACHINE = "(qemux86|qemux86-64|qemuppc|emenlow|crownbay|atom-pc|n450)" >>> +COMPATIBLE_MACHINE = "(qemux86|qemux86-64|qemuppc|atom-pc)" >> >> After having looked at the recipe (basically 'inherit autotools', PACKAGE_ARCH = BASE_PACKAGE_ARCH) and reading the systemtap wiki I can't figure out why COMPATIBLE_MACHINE is actually needed. The closest I get is "needs specific kernel config", but that argument quickly falls down when looking at other recipes that need that (e.g. udev) and considering out layer strategy (bbappending it for all your machines). >> >> So, what am I missing here? >> > > My testing showed systemtap doesn't actually work on arm, and has no > support for mips - the machines listed are the only ones I've been able > to verify that work. So that sounds like COMPATIBLE_HOST instead of machine, no? regards, Koen
On Mon, 2011-06-13 at 09:54 -0500, Tom Zanussi wrote: > On Mon, 2011-06-13 at 07:47 -0700, Koen Kooi wrote: > > Op 13 jun 2011, om 16:35 heeft Tom Zanussi het volgende geschreven: > > > -COMPATIBLE_MACHINE = "(qemux86|qemux86-64|qemuppc|emenlow|crownbay|atom-pc|n450)" > > > +COMPATIBLE_MACHINE = "(qemux86|qemux86-64|qemuppc|atom-pc)" > > > > After having looked at the recipe (basically 'inherit autotools', PACKAGE_ARCH = BASE_PACKAGE_ARCH) and reading the systemtap wiki I can't figure out why COMPATIBLE_MACHINE is actually needed. The closest I get is "needs specific kernel config", but that argument quickly falls down when looking at other recipes that need that (e.g. udev) and considering out layer strategy (bbappending it for all your machines). > > > > So, what am I missing here? > > > > My testing showed systemtap doesn't actually work on arm, and has no > support for mips - the machines listed are the only ones I've been able > to verify that work. If it's actually architecture-dependent, ie it will work on any x86 (or sparc, or...) platform, the way to deal with that is via COMPATIBLE_HOST. p.
On Mon, 2011-06-13 at 08:12 -0700, Koen Kooi wrote: > Op 13 jun 2011, om 16:54 heeft Tom Zanussi het volgende geschreven: > > > On Mon, 2011-06-13 at 07:47 -0700, Koen Kooi wrote: > >> Op 13 jun 2011, om 16:35 heeft Tom Zanussi het volgende geschreven: > >>> -COMPATIBLE_MACHINE = "(qemux86|qemux86-64|qemuppc|emenlow|crownbay|atom-pc|n450)" > >>> +COMPATIBLE_MACHINE = "(qemux86|qemux86-64|qemuppc|atom-pc)" > >> > >> After having looked at the recipe (basically 'inherit autotools', PACKAGE_ARCH = BASE_PACKAGE_ARCH) and reading the systemtap wiki I can't figure out why COMPATIBLE_MACHINE is actually needed. The closest I get is "needs specific kernel config", but that argument quickly falls down when looking at other recipes that need that (e.g. udev) and considering out layer strategy (bbappending it for all your machines). > >> > >> So, what am I missing here? > >> > > > > My testing showed systemtap doesn't actually work on arm, and has no > > support for mips - the machines listed are the only ones I've been able > > to verify that work. > > So that sounds like COMPATIBLE_HOST instead of machine, no? > Yeah, you're right - that does make more sense in this case. I'll update things with your suggestions and respin, thanks... Tom > regards, > > Koen > _______________________________________________ > Openembedded-core mailing list > Openembedded-core@lists.openembedded.org > http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Patch
diff --git a/meta/recipes-kernel/systemtap/systemtap_git.bb b/meta/recipes-kernel/systemtap/systemtap_git.bb index 8ac4c14..783f30d 100644 --- a/meta/recipes-kernel/systemtap/systemtap_git.bb +++ b/meta/recipes-kernel/systemtap/systemtap_git.bb @@ -5,7 +5,7 @@ LIC_FILES_CHKSUM = "file://COPYING;md5=94d55d512a9ba36caa9b7df079bae19f" DEPENDS = "elfutils" SRCREV = "4ab3a1863bf4f472acae7a809bf2b38d91658aa8" -PR = "r1" +PR = "r2" PV = "1.4+git${SRCPV}" SRC_URI = "git://sources.redhat.com/git/systemtap.git;protocol=git \ @@ -23,7 +23,7 @@ EXTRA_OECONF = "--prefix=${D} --with-libelf=${STAGING_DIR_TARGET} --without-rpm SRC_URI[md5sum] = "cb202866ed704c44a876d041f788bdee" SRC_URI[sha256sum] = "8ffe35caec0d937bd23fd78a3a8d94b58907cc0de0330b35e38f9f764815c459" -COMPATIBLE_MACHINE = "(qemux86|qemux86-64|qemuppc|emenlow|crownbay|atom-pc|n450)" +COMPATIBLE_MACHINE = "(qemux86|qemux86-64|qemuppc|atom-pc)" S = "${WORKDIR}/git"
Move these from here into their respective layers instead. Signed-off-by: Tom Zanussi <tom.zanussi@intel.com> --- meta/recipes-kernel/systemtap/systemtap_git.bb | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-)