this is massively off-topic but i'm hoping someone's bored enough to take a look. i have a beagleboard:
and am trying to build the angstrom OS for it. the build (using the openembedded (OE) build utility) should be fairly straightforward but trying this under my current 64-bit install of f11 preview fails each time thusly:
... snip ...
ccache gcc -isystem/home/rpjday/oe/angstrom-dev/staging/x86_64-linux/usr/include -O2 -Wunused -Wall -static -L/home/rpjday/oe/angstrom-dev/staging/x86_64-linux/usr/lib -Wl,-rpath-link,/home/rpjday/oe/angstrom-dev/staging/x86_64-linux/usr/lib -Wl,-rpath,/home/rpjday/oe/angstrom-dev/staging/x86_64-linux/usr/lib -Wl,-O1 -o insmod.static insmod.o mv -f .deps/tables.Tpo .deps/tables.Po /usr/bin/ld: cannot find -lc collect2: ld returned 1 exit status make: *** [insmod.static] Error 1 make: *** Waiting for unfinished jobs.... modinfo.c: In function 'main': modinfo.c:338: warning: 'infosize' may be used uninitialized in this function mv -f .deps/modinfo.Tpo .deps/modinfo.Po make: *** wait: No child processes. Stop. FATAL: oe_runmake failed
i've documented, step by step, what i did here:
http://www.crashcourse.ca/wiki/index.php/Angstrom#Building_Angstrom_from_scr...
and it's frustrating since others have reported normal success (some running under fedora 8). i have no idea why i keep getting this error -- maybe it's related to running 64-bit, i don't know. others are building just fine under other distros as well, so i'm baffled.
thoughts? at this point, i'm really open to suggestions. if you've never played with openembedded, here's your chance. :-)
rday --
======================================================================== Robert P. J. Day Waterloo, Ontario, CANADA
Linux Consulting, Training and Annoying Kernel Pedantry.
Web page: http://crashcourse.ca Linked In: http://www.linkedin.com/in/rpjday Twitter: http://twitter.com/rpjday ========================================================================
Robert P. J. Day wrote:
this is massively off-topic but i'm hoping someone's bored enough to take a look. i have a beagleboard:
ccache gcc -isystem/home/rpjday/oe/angstrom-dev/staging/x86_64-linux/usr/include -O2 -Wunused -Wall -static -L/home/rpjday/oe/angstrom-dev/staging/x86_64-linux/usr/lib -Wl,-rpath-link,/home/rpjday/oe/angstrom-dev/staging/x86_64-linux/usr/lib -Wl,-rpath,/home/rpjday/oe/angstrom-dev/staging/x86_64-linux/usr/lib -Wl,-O1 -o insmod.static insmod.o mv -f .deps/tables.Tpo .deps/tables.Po /usr/bin/ld: cannot find -lc collect2: ld returned 1 exit status make: *** [insmod.static] Error 1 make: *** Waiting for unfinished jobs.... modinfo.c: In function 'main': modinfo.c:338: warning: 'infosize' may be used uninitialized in this function mv -f .deps/modinfo.Tpo .deps/modinfo.Po make: *** wait: No child processes. Stop. FATAL: oe_runmake failed
i've documented, step by step, what i did here:
http://www.crashcourse.ca/wiki/index.php/Angstrom#Building_Angstrom_from_scr...
and it's frustrating since others have reported normal success (some running under fedora 8). i have no idea why i keep getting this error -- maybe it's related to running 64-bit, i don't know. others are building just fine under other distros as well, so i'm baffled.
The gcc call above seems to be trying to link statically against libc (An absurd idea, IMNSHO).
In FC11, the static libs from glibc have been split out into a separate "glibc-static" package.
It's likely getting rid of the static linkage above or installing "glibc-static" (BuildRequires: glibc-static) will help.
Ralf
On Sun, 31 May 2009, Ralf Corsepius wrote:
Robert P. J. Day wrote:
ccache gcc -isystem/home/rpjday/oe/angstrom-dev/staging/x86_64-linux/usr/include -O2 -Wunused -Wall -static -L/home/rpjday/oe/angstrom-dev/staging/x86_64-linux/usr/lib -Wl,-rpath-link,/home/rpjday/oe/angstrom-dev/staging/x86_64-linux/usr/lib -Wl,-rpath,/home/rpjday/oe/angstrom-dev/staging/x86_64-linux/usr/lib -Wl,-O1 -o insmod.static insmod.o mv -f .deps/tables.Tpo .deps/tables.Po /usr/bin/ld: cannot find -lc collect2: ld returned 1 exit status make: *** [insmod.static] Error 1 make: *** Waiting for unfinished jobs.... modinfo.c: In function 'main': modinfo.c:338: warning: 'infosize' may be used uninitialized in this function mv -f .deps/modinfo.Tpo .deps/modinfo.Po make: *** wait: No child processes. Stop. FATAL: oe_runmake failed
The gcc call above seems to be trying to link statically against libc (An absurd idea, IMNSHO).
In FC11, the static libs from glibc have been split out into a separate "glibc-static" package.
It's likely getting rid of the static linkage above or installing "glibc-static" (BuildRequires: glibc-static) will help.
ooooh, that could be it. amusingly, i was looking at that earlier this morning and thinking, "is it the static linkage that's doing this?" and, sure enough, that package wasn't installed on my system so i'll test it shortly. thanks.
um ... given your wording, can i assume that f11 is the first fedora where this package has been broken out? which would explain why this builds just fine under f10 for others?
rday --
======================================================================== Robert P. J. Day Waterloo, Ontario, CANADA
Linux Consulting, Training and Annoying Kernel Pedantry.
Web page: http://crashcourse.ca Linked In: http://www.linkedin.com/in/rpjday Twitter: http://twitter.com/rpjday ========================================================================
On Sun, 31 May 2009, Rahul Sundaram wrote:
On 05/31/2009 04:32 PM, Robert P. J. Day wrote:
um ... given your wording, can i assume that f11 is the first fedora where this package has been broken out? which would explain why this builds just fine under f10 for others?
Correct.
that appears to have fixed it. yee ha! buckets of thanks. i can now update the wiki page to cover f11.
rday --
======================================================================== Robert P. J. Day Waterloo, Ontario, CANADA
Linux Consulting, Training and Annoying Kernel Pedantry.
Web page: http://crashcourse.ca Linked In: http://www.linkedin.com/in/rpjday Twitter: http://twitter.com/rpjday ========================================================================
On 05/31/2009 04:42 PM, Robert P. J. Day wrote:
On Sun, 31 May 2009, Rahul Sundaram wrote:
On 05/31/2009 04:32 PM, Robert P. J. Day wrote:
um ... given your wording, can i assume that f11 is the first fedora where this package has been broken out? which would explain why this builds just fine under f10 for others?
Correct.
that appears to have fixed it. yee ha! buckets of thanks. i can now update the wiki page to cover f11.
You might want to tell upstream that static linking isn't the recommended path.
http://people.redhat.com/drepper/no_static_linking.html
Rahul
On Sun, 31 May 2009, Rahul Sundaram wrote:
On 05/31/2009 04:42 PM, Robert P. J. Day wrote:
On Sun, 31 May 2009, Rahul Sundaram wrote:
On 05/31/2009 04:32 PM, Robert P. J. Day wrote:
um ... given your wording, can i assume that f11 is the first fedora where this package has been broken out? which would explain why this builds just fine under f10 for others?
Correct.
that appears to have fixed it. yee ha! buckets of thanks. i can now update the wiki page to cover f11.
You might want to tell upstream that static linking isn't the recommended path.
and yet f11 contains the statically-linked "insmod.static" executable, which is exactly what is being created at that step. so what's the difference?
rday --
======================================================================== Robert P. J. Day Waterloo, Ontario, CANADA
Linux Consulting, Training and Annoying Kernel Pedantry.
Web page: http://crashcourse.ca Linked In: http://www.linkedin.com/in/rpjday Twitter: http://twitter.com/rpjday ========================================================================
On 05/31/2009 04:51 PM, Robert P. J. Day wrote:
and yet f11 contains the statically-linked "insmod.static" executable, which is exactly what is being created at that step. so what's the difference?
The static libraries are provided in Fedora as a workaround to get such applications building but it isn't the recommended path.
https://fedoraproject.org/wiki/Packaging/Guidelines#Packaging_Static_Librari...
Rahul
On Sun, 31 May 2009, Rahul Sundaram wrote:
On 05/31/2009 04:51 PM, Robert P. J. Day wrote:
and yet f11 contains the statically-linked "insmod.static" executable, which is exactly what is being created at that step. so what's the difference?
The static libraries are provided in Fedora as a workaround to get such applications building but it isn't the recommended path.
https://fedoraproject.org/wiki/Packaging/Guidelines#Packaging_Static_Librari...
gotcha. thanks.
rday --
======================================================================== Robert P. J. Day Waterloo, Ontario, CANADA
Linux Consulting, Training and Annoying Kernel Pedantry.
Web page: http://crashcourse.ca Linked In: http://www.linkedin.com/in/rpjday Twitter: http://twitter.com/rpjday ========================================================================
On 05/31/2009 04:32 PM, Robert P. J. Day wrote:
um ... given your wording, can i assume that f11 is the first fedora where this package has been broken out? which would explain why this builds just fine under f10 for others?
On Sun, 31 May 2009, Rahul Sundaram wrote:
Correct.
On Sun, May 31, 2009 at 1:12 PM, Robert P. J. Day wrote:
that appears to have fixed it. yee ha! buckets of thanks. i can now update the wiki page to cover f11.
Hello,
I think I will add "glibc-static" into FEL's section on the comps.xml. Hence it will reduce pain for embedded designers.
Robert,
Since you are compiling from scratch, is it possible for you to package some of the bits into Fedora ? It would really help many fedora users.
Every Fedora event, someone pops out to ask for more embedded tools on Fedora. We are currently packaging openocd for fedora. But we need an expert in the embedded field to guide us.
Kind regards, Chitlesh
Chitlesh GOORAH wrote:
On 05/31/2009 04:32 PM, Robert P. J. Day wrote:
um ... given your wording, can i assume that f11 is the first fedora where this package has been broken out? which would explain why this builds just fine under f10 for others?
On Sun, 31 May 2009, Rahul Sundaram wrote:
Correct.
On Sun, May 31, 2009 at 1:12 PM, Robert P. J. Day wrote:
that appears to have fixed it. yee ha! buckets of thanks. i can now update the wiki page to cover f11.
Hello,
I think I will add "glibc-static" into FEL's section on the comps.xml.
A mistake.
Hence it will reduce pain for embedded designers.
How so? I don't see this.
Ralf
On Tue, 2 Jun 2009, Chitlesh GOORAH wrote:
On 05/31/2009 04:32 PM, Robert P. J. Day wrote:
um ... given your wording, can i assume that f11 is the first fedora where this package has been broken out? which would explain why this builds just fine under f10 for others?
On Sun, 31 May 2009, Rahul Sundaram wrote:
Correct.
On Sun, May 31, 2009 at 1:12 PM, Robert P. J. Day wrote:
that appears to have fixed it. yee ha! buckets of thanks. i can now update the wiki page to cover f11.
Hello,
I think I will add "glibc-static" into FEL's section on the comps.xml. Hence it will reduce pain for embedded designers.
i'm not convinced that's a good idea. there's no point polluting the group sections with packages that might not be necessary. i think it will be sufficient for tools like openembedded to generate their own warnings. this isn't a fedora problem, let's not make it one.
rday --
======================================================================== Robert P. J. Day Waterloo, Ontario, CANADA
Linux Consulting, Training and Annoying Kernel Pedantry.
Web page: http://crashcourse.ca Linked In: http://www.linkedin.com/in/rpjday Twitter: http://twitter.com/rpjday ========================================================================
Maybe I'll even buy a Beagleboard. This took several hours to do, all day on my Dell Latitude E6400 laptop in fact, but it seems to have worked.
Results under Fedora 11, meaning this version exactly
[rlc@deafeng3 openembedded]$ uname -a Linux deafeng3.signtype.info 2.6.29.4-167.fc11.x86_64 #1 SMP Wed May 27 17:27:08 EDT 2009 x86_64 x86_64 x86_64 GNU/Linux [rlc@deafeng3 openembedded]$ cat /etc/fedora-release Fedora release 11 (Leonidas)
[From `bitbake base-image`:]
NOTE: Tasks Summary: Attempted 2946 tasks of which 0 didn't need to be rerun and 0 failed. NOTE: build 200905310908: completed [rlc@deafeng3 openembedded]$ bitbake console-image NOTE: Handling BitBake files: - (6650/6650) [100 %] NOTE: Parsing finished. 6368 cached, 0 parsed, 282 skipped, 0 masked. NOTE: Cache is clean, not saving. NOTE: build 200905311345: started ..... NOTE: Tasks Summary: Attempted 2928 tasks of which 2890 didn't need to be rerun and 0 failed. NOTE: build 200905311345: completed
[rlc@deafeng3 openembedded]$ bitbake x11-image NOTE: Handling BitBake files: - (6650/6650) [100 %] NOTE: Parsing finished. 6368 cached, 0 parsed, 282 skipped, 0 masked. NOTE: Cache is clean, not saving. NOTE: build 200905311351: started
OE Build Configuration: BB_VERSION = "1.8.12" METADATA_BRANCH = "stable/2009" METADATA_REVISION = "901e7c5b47b9f3695cef841e6c8663d37c932fec" TARGET_ARCH = "arm" TARGET_OS = "linux-gnueabi" MACHINE = "beagleboard" DISTRO = "angstrom" DISTRO_VERSION = "2009.X-test-20090531" TARGET_FPU = "hard"
NOTE: Resolving any missing task queue dependencies NOTE: preferred version git of u-boot not available (for item u-boot) NOTE: preferred version git of u-boot not available (for item u-boot) .... .... NOTE: Tasks Summary: Attempted 6096 tasks of which 2908 didn't need to be rerun and 0 failed. NOTE: build 200905311351: completed
On 05/31/2009 06:43 AM, Robert P. J. Day wrote:
this is massively off-topic but i'm hoping someone's bored enough to take a look. i have a beagleboard:
and am trying to build the angstrom OS for it. the build (using the openembedded (OE) build utility) should be fairly straightforward but trying this under my current 64-bit install of f11 preview fails each time thusly:
... snip ...
ccache gcc -isystem/home/rpjday/oe/angstrom-dev/staging/x86_64-linux/usr/include -O2 -Wunused -Wall -static -L/home/rpjday/oe/angstrom-dev/staging/x86_64-linux/usr/lib -Wl,-rpath-link,/home/rpjday/oe/angstrom-dev/staging/x86_64-linux/usr/lib -Wl,-rpath,/home/rpjday/oe/angstrom-dev/staging/x86_64-linux/usr/lib -Wl,-O1 -o insmod.static insmod.o mv -f .deps/tables.Tpo .deps/tables.Po /usr/bin/ld: cannot find -lc collect2: ld returned 1 exit status make: *** [insmod.static] Error 1 make: *** Waiting for unfinished jobs.... modinfo.c: In function 'main': modinfo.c:338: warning: 'infosize' may be used uninitialized in this function mv -f .deps/modinfo.Tpo .deps/modinfo.Po make: *** wait: No child processes. Stop. FATAL: oe_runmake failed
i've documented, step by step, what i did here:
http://www.crashcourse.ca/wiki/index.php/Angstrom#Building_Angstrom_from_scr...
and it's frustrating since others have reported normal success (some running under fedora 8). i have no idea why i keep getting this error -- maybe it's related to running 64-bit, i don't know. others are building just fine under other distros as well, so i'm baffled.
thoughts? at this point, i'm really open to suggestions. if you've never played with openembedded, here's your chance. :-)
rday
======================================================================== Robert P. J. Day Waterloo, Ontario, CANADA
Linux Consulting, Training and Annoying Kernel Pedantry.
Web page: http://crashcourse.ca Linked In: http://www.linkedin.com/in/rpjday Twitter: http://twitter.com/rpjday ========================================================================
On Sun, 31 May 2009, Robert L Cochran wrote:
Maybe I'll even buy a Beagleboard. This took several hours to do, all day on my Dell Latitude E6400 laptop in fact, but it seems to have worked.
Results under Fedora 11, meaning this version exactly
[rlc@deafeng3 openembedded]$ uname -a Linux deafeng3.signtype.info 2.6.29.4-167.fc11.x86_64 #1 SMP Wed May 27 17:27:08 EDT 2009 x86_64 x86_64 x86_64 GNU/Linux [rlc@deafeng3 openembedded]$ cat /etc/fedora-release Fedora release 11 (Leonidas)
[From `bitbake base-image`:]
NOTE: Tasks Summary: Attempted 2946 tasks of which 0 didn't need to be rerun and 0 failed. NOTE: build 200905310908: completed [rlc@deafeng3 openembedded]$ bitbake console-image NOTE: Handling BitBake files: - (6650/6650) [100 %] NOTE: Parsing finished. 6368 cached, 0 parsed, 282 skipped, 0 masked. NOTE: Cache is clean, not saving. NOTE: build 200905311345: started ..... NOTE: Tasks Summary: Attempted 2928 tasks of which 2890 didn't need to be rerun and 0 failed. NOTE: build 200905311345: completed
[rlc@deafeng3 openembedded]$ bitbake x11-image NOTE: Handling BitBake files: - (6650/6650) [100 %] NOTE: Parsing finished. 6368 cached, 0 parsed, 282 skipped, 0 masked. NOTE: Cache is clean, not saving. NOTE: build 200905311351: started
OE Build Configuration: BB_VERSION = "1.8.12" METADATA_BRANCH = "stable/2009" METADATA_REVISION = "901e7c5b47b9f3695cef841e6c8663d37c932fec" TARGET_ARCH = "arm" TARGET_OS = "linux-gnueabi" MACHINE = "beagleboard" DISTRO = "angstrom" DISTRO_VERSION = "2009.X-test-20090531" TARGET_FPU = "hard"
NOTE: Resolving any missing task queue dependencies NOTE: preferred version git of u-boot not available (for item u-boot) NOTE: preferred version git of u-boot not available (for item u-boot) .... .... NOTE: Tasks Summary: Attempted 6096 tasks of which 2908 didn't need to be rerun and 0 failed. NOTE: build 200905311351: completed
please tell me you had the glibc-static package installed, cuz if you didn't, then i'm going to be confused again. at the moment, mine's still building but i'm past the point where it was failing before because of that issue.
rday --
======================================================================== Robert P. J. Day Waterloo, Ontario, CANADA
Linux Consulting, Training and Annoying Kernel Pedantry.
Web page: http://crashcourse.ca Linked In: http://www.linkedin.com/in/rpjday Twitter: http://twitter.com/rpjday ========================================================================
please tell me you had the glibc-static package installed, cuz if you didn't, then i'm going to be confused again. at the moment, mine's still building but i'm past the point where it was failing before because of that issue.
rday
Yes I installed glibc-static and texi2html after carefully reading your posts, your wiki, the Angstrom web page and the initial `bitbake base-image` output. Success in building software comes from reading the docs, initial failure, and trying again. The build took a long, long time for me and all worked well.
Since you have a Beagle Board can you post photos of it in use on your wiki? Or send me photos offlist, if possible? I'm very curious as to your results. If my guess is correct you even had to pay additional taxes to get the Beagle Board from Digi-Key...unless, of course, you drove to Digi-Key's location and took delivery at their sales counter. Not that I know if they do that. So you must have a strong reason to get this board.
Bob
======================================================================== Robert P. J. Day Waterloo, Ontario, CANADA
Linux Consulting, Training and Annoying Kernel Pedantry.
Web page: http://crashcourse.ca Linked In: http://www.linkedin.com/in/rpjday Twitter: http://twitter.com/rpjday ========================================================================
On Sun, 31 May 2009, Robert L Cochran wrote:
Yes I installed glibc-static and texi2html after carefully reading your posts, your wiki, the Angstrom web page and the initial `bitbake base-image` output. Success in building software comes from reading the docs, initial failure, and trying again. The build took a long, long time for me and all worked well.
yes, it takes a while. the "start the build and go out drinking" type of while.
Since you have a Beagle Board can you post photos of it in use on your wiki? Or send me photos offlist, if possible? I'm very curious as to your results.
there's plenty of BB-related stuff at the main web site: http://beagleboard.org. lots of cool pix, lots of links, lots of people doing wicked cool stuff with it -- audio, video, etc.
If my guess is correct you even had to pay additional taxes to get the Beagle Board from Digi-Key...unless, of course, you drove to Digi-Key's location and took delivery at their sales counter. Not that I know if they do that.
if you're in canada, http://digikey.ca. no big deal.
So you must have a strong reason to get this board.
just playing -- it's a cheap option for getting into embedded linux. and, just to be clear, obviously this board isn't technically related to fedora test releases -- i only asked here since it was clear that there was *something* fedora-related to the build failing, i just couldn't figure out what. when others are saying, "hey, i'm using ubuntu, it builds just fine," there's some incentive to figure out what it is about fedora that's mucking things up. and the solution turned out to be fairly trivial. (i might even put in a request that the build process check for that and issue a warning to f11 users.)
in any event, it's a cool toy, particularly for folks who want to get into android, given the growing popularity of android on linux-based phones. there's an android port for the beagleboard so you have a cheap way of jumping in. the only slight drawback is that, while there's buckets of online docs, sometimes it's hard to track it all down, which is why i wrote this:
http://www.crashcourse.ca/wiki/index.php/BeagleBoard
if you follow what's there on your f11 system, it should just work. if it doesn't, then i'll fix the page.
rday
p.s. i even have one of these to go with it:
http://focus.ti.com/dlpdmd/docs/dlpdiscovery.tsp?sectionId=60&tabId=2235
in any event, enjoy. i'm just going to make sure that everything that's supposed to work continues to work under fedora.
--
======================================================================== Robert P. J. Day Waterloo, Ontario, CANADA
Linux Consulting, Training and Annoying Kernel Pedantry.
Web page: http://crashcourse.ca Linked In: http://www.linkedin.com/in/rpjday Twitter: http://twitter.com/rpjday ========================================================================