Re: [Gumstix-users] Re: Buildroot optional targets that fail - was Re: [Gumstix-users] Program hanging on exit

[Michael Taylor <mike.taylor@ap...>]

Craig,

I turned on LARGEFILE but strace still fails.  Ever seen this one before?

file.c: In function `printstatfs64':
file.c:1534: error: storage size of 'statbuf' isn't known
file.c:1534: warning: unused variable `statbuf'
file.c: In function `sys_statfs64':
file.c:1580: error: invalid application of `sizeof' to incomplete type 
`statfs64'
file.c: In function `sys_fstatfs64':
file.c:1595: error: invalid application of `sizeof' to incomplete type 
`statfs64'
make[2]: *** [file.o] Error 1
make[2]: Leaving directory 
`/opt/gumstix/toolchain/sources/gumstix-buildroot/buildroot/build_arm_nofpu/strace-4.5.6'
make[1]: *** [all] Error 2
make[1]: Leaving directory 
`/opt/gumstix/toolchain/sources/gumstix-buildroot/buildroot/build_arm_nofpu/strace-4.5.6'
make: *** 
[/opt/gumstix/toolchain/sources/gumstix-buildroot/buildroot/build_arm_nofpu/strace-4.5.6/strace] 
Error 2

>>
>> strace builds OK with LARGEFILE support turned on in the top-level  
>> makefile (which required uclibc to have been built that way in the  
>> first place).  The strace makefile just basically assumes you have  
>> LARGEFILE, and it looks like quite a job to modify the source with  
>> #ifdefs to get it to compile w/out largefile.
>

[Gumstix-users] Program hanging on exit

[Darren Gibbs <tsquank@ya...>]

I've got a program that runs and exits fine on v435, but hangs on  
exit in v658.  I scattered some printf()s around and it seems to be  
getting wedged after the return from main().  When I run  ps  it  
shows the process name in square brackets and it is labeled as  
sleeping and swapped out.  It takes 45 seconds to a minute before it  
dies completely and I get back to a command line prompt.  Anyone have  
a clue how to debug this further?  I'm not sure what could have  
changed that would cause this... v435 and v658 are both 2.6.11gum.

Any advice much appreciated,

darren

Re: [Gumstix-users] Program hanging on exit

[Craig Hughes <craig@gu...>]

It's probably getting stuck in a system call which is getting  
triggered on exit; strace or similar would help you figure out which  
syscall is responsible.  Did anyone ever get strace working?

C

On Dec 8, 2005, at 9:19 AM, Darren Gibbs wrote:

> I've got a program that runs and exits fine on v435, but hangs on  
> exit in v658.  I scattered some printf()s around and it seems to be  
> getting wedged after the return from main().  When I run  ps  it  
> shows the process name in square brackets and it is labeled as  
> sleeping and swapped out.  It takes 45 seconds to a minute before  
> it dies completely and I get back to a command line prompt.  Anyone  
> have a clue how to debug this further?  I'm not sure what could  
> have changed that would cause this... v435 and v658 are both  
> 2.6.11gum.
>
> Any advice much appreciated,

Buildroot optional targets that fail - was Re: [Gumstix-users] Program hanging on exit

[Michael Taylor <mike.taylor@ap...>]

Craig/all,

I was unable to build strace myself.  Further, I've found a few make
files with either stale or invalid URLs, or with build failures.  Is
there a list on the Wiki that indicates which of the optional buildroot
targets are stable and which are under development?

Thanks,

Mike

Craig Hughes wrote:

> It's probably getting stuck in a system call which is getting  
> triggered on exit; strace or similar would help you figure out which  
> syscall is responsible.  Did anyone ever get strace working?

Re: Buildroot optional targets that fail - was Re: [Gumstix-users] Program hanging on exit

[Craig Hughes <craig@gu...>]

I think I removed all the ones which weren't working properly -- no,  
there's not a wiki page which break out stable vs unstable.  Perhaps  
actually what I should do is have multiple directories under make/  
which are self-documenting, eg make/stable/*.mk and make/alpha/*.mk  
make/beta/*.mk, etc.

If you have the latest buildroot installed, it should now be somewhat  
more resistant to the underlying tarballs vanishing from their  
original URLs, as the download system should now be trying to fall  
back to our download server which has a cache of more or less  
everything -- though it's possible that I've overlooked adding some  
of the tarballs into the archive.

If there are any specific shortcomings of what's in the latest HEAD  
of svn, please let me know, and I'll try and remedy that.

C

On Dec 8, 2005, at 1:36 PM, Michael Taylor wrote:

> Craig/all,
>
> I was unable to build strace myself.  Further, I've found a few make
> files with either stale or invalid URLs, or with build failures.  Is
> there a list on the Wiki that indicates which of the optional  
> buildroot
> targets are stable and which are under development?
>
> Thanks,
>
> Mike
>
> Craig Hughes wrote:
>
>> It's probably getting stuck in a system call which is getting   
>> triggered on exit; strace or similar would help you figure out  
>> which  syscall is responsible.  Did anyone ever get strace working?
>
>
>
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc. Do you grep through  
> log files
> for problems?  Stop!  Download the new AJAX search engine that makes
> searching your log files as easy as surfing the  web.  DOWNLOAD  
> SPLUNK!
> http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
> _______________________________________________
> gumstix-users mailing list
> gumstix-users@...
> https://lists.sourceforge.net/lists/listinfo/gumstix-users

Re: Buildroot optional targets that fail - was Re: [Gumstix-users] Program hanging on exit

[Craig Hughes <craig@gu...>]

On Dec 8, 2005, at 2:56 PM, Craig Hughes wrote:

> I think I removed all the ones which weren't working properly

strace builds OK with LARGEFILE support turned on in the top-level  
makefile (which required uclibc to have been built that way in the  
first place).  The strace makefile just basically assumes you have  
LARGEFILE, and it looks like quite a job to modify the source with  
#ifdefs to get it to compile w/out largefile.

C

[Gumstix-users] Re: Buildroot optional targets that fail - was Re: [Gumstix-users] Program hanging on exit

[Michael Taylor <mike.taylor@ap...>]

Craig Hughes wrote:

>> I think I removed all the ones which weren't working properly
>
> strace builds OK with LARGEFILE support turned on in the top-level  
> makefile (which required uclibc to have been built that way in the  
> first place).  The strace makefile just basically assumes you have  
> LARGEFILE, and it looks like quite a job to modify the source with  
> #ifdefs to get it to compile w/out largefile.


Thanks Craig.  I'll try it with LARGEFILE enabled.  BTW, the other 
problem I keep running into is some undefined symbols in building 
ncurses.  I'll send the tail of the build failure as soon as I finish 
this rebuild, assuming it still fails.

Thanks,

Mike

Re: [Gumstix-users] Re: Buildroot optional targets that fail - was Re: [Gumstix-users] Program hanging on exit

[Michael Taylor <mike.taylor@ap...>]

Craig,

I turned on LARGEFILE but strace still fails.  Ever seen this one before?

file.c: In function `printstatfs64':
file.c:1534: error: storage size of 'statbuf' isn't known
file.c:1534: warning: unused variable `statbuf'
file.c: In function `sys_statfs64':
file.c:1580: error: invalid application of `sizeof' to incomplete type 
`statfs64'
file.c: In function `sys_fstatfs64':
file.c:1595: error: invalid application of `sizeof' to incomplete type 
`statfs64'
make[2]: *** [file.o] Error 1
make[2]: Leaving directory 
`/opt/gumstix/toolchain/sources/gumstix-buildroot/buildroot/build_arm_nofpu/strace-4.5.6'
make[1]: *** [all] Error 2
make[1]: Leaving directory 
`/opt/gumstix/toolchain/sources/gumstix-buildroot/buildroot/build_arm_nofpu/strace-4.5.6'
make: *** 
[/opt/gumstix/toolchain/sources/gumstix-buildroot/buildroot/build_arm_nofpu/strace-4.5.6/strace] 
Error 2

>>
>> strace builds OK with LARGEFILE support turned on in the top-level  
>> makefile (which required uclibc to have been built that way in the  
>> first place).  The strace makefile just basically assumes you have  
>> LARGEFILE, and it looks like quite a job to modify the source with  
>> #ifdefs to get it to compile w/out largefile.
>

Re: [Gumstix-users] Re: Buildroot optional targets that fail - was Re: [Gumstix-users] Program hanging on exit

[Craig Hughes <craig@gu...>]

Yup -- that's cos LARGEFILE support isn't built into your uclibc.   
You need to follow the recipe:

cd /path/to/buildroot
rm -rf toolchain* build_arm*
vi Makefile # and turn on LARGEFILE
make # go make coffee
make strace # now it'll work


The key is that uclibc has to be built with largefile support -- you  
can't just twiddle that in the top-level Makefile *after* uclibc is  
already built.

C


On Dec 9, 2005, at 11:11 AM, Michael Taylor wrote:

> Craig,
>
> I turned on LARGEFILE but strace still fails.  Ever seen this one  
> before?
>
> file.c: In function `printstatfs64':
> file.c:1534: error: storage size of 'statbuf' isn't known
> file.c:1534: warning: unused variable `statbuf'
> file.c: In function `sys_statfs64':
> file.c:1580: error: invalid application of `sizeof' to incomplete  
> type `statfs64'
> file.c: In function `sys_fstatfs64':
> file.c:1595: error: invalid application of `sizeof' to incomplete  
> type `statfs64'
> make[2]: *** [file.o] Error 1
> make[2]: Leaving directory `/opt/gumstix/toolchain/sources/gumstix- 
> buildroot/buildroot/build_arm_nofpu/strace-4.5.6'
> make[1]: *** [all] Error 2
> make[1]: Leaving directory `/opt/gumstix/toolchain/sources/gumstix- 
> buildroot/buildroot/build_arm_nofpu/strace-4.5.6'
> make: *** [/opt/gumstix/toolchain/sources/gumstix-buildroot/ 
> buildroot/build_arm_nofpu/strace-4.5.6/strace] Error 2
>
>>>
>>> strace builds OK with LARGEFILE support turned on in the top- 
>>> level  makefile (which required uclibc to have been built that  
>>> way in the  first place).  The strace makefile just basically  
>>> assumes you have  LARGEFILE, and it looks like quite a job to  
>>> modify the source with  #ifdefs to get it to compile w/out  
>>> largefile.
>>
>
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc. Do you grep through  
> log files
> for problems?  Stop!  Download the new AJAX search engine that makes
> searching your log files as easy as surfing the  web.  DOWNLOAD  
> SPLUNK!
> http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
> _______________________________________________
> gumstix-users mailing list
> gumstix-users@...
> https://lists.sourceforge.net/lists/listinfo/gumstix-users

[Gumstix-users] Re: Buildroot optional targets that fail - was Re: [Gumstix-users] Program hanging on exit

[Michael Taylor <mike.taylor@ap...>]

Craig Hughes wrote:

> If there are any specific shortcomings of what's in the latest HEAD  
> of svn, please let me know, and I'll try and remedy that.

I had problems a couple of days ago with openobex, dropbear, and 
hostap.  hostap and dropbear sites were down and did not fallback as you 
suggested they should in a separate posting on this thread.  The 
openobex.mk file lists umn.dl.sourceforge.net, which at least from here 
is an unreachable machine.  Not sure what to say about that one, except 
that the kent.dl.sourceforge.net mirror seems to work (patch enclosed).

Thanks,

Mike
--

Index: make/openobex.mk
===================================================================
--- make/openobex.mk    (revision 674)
+++ make/openobex.mk    (working copy)
@@ -3,7 +3,8 @@
 # openobex
 #
 #############################################################
-OPENOBEX_SITE:=http://umn.dl.sourceforge.net/sourceforge/openobex
+#OPENOBEX_SITE:=http://umn.dl.sourceforge.net/sourceforge/openobex
+OPENOBEX_SITE:=http://kent.dl.sourceforge.net/sourceforge/openobex/
 OPENOBEX_CAT:=zcat
 OPENOBEX_LIBS_SOURCE:=openobex-1.0.1.tar.gz
 OPENOBEX_APPS_SOURCE:=openobex-apps-1.0.0.tar.gz
Index: make/dropbear_sshd.mk
===================================================================
--- make/dropbear_sshd.mk       (revision 674)
+++ make/dropbear_sshd.mk       (working copy)
@@ -4,7 +4,8 @@
 #
 #############################################################
 DROPBEAR_SSHD_SOURCE:=dropbear-0.44.tar.bz2
-DROPBEAR_SSHD_SITE:=http://matt.ucc.asn.au/dropbear/releases/
+#DROPBEAR_SSHD_SITE:=http://matt.ucc.asn.au/dropbear/releases/
+DROPBEAR_SSHD_SITE:=http://buildroot.uclibc.org/downloads/buildroot-sources/
 DROPBEAR_SSHD_DIR:=$(BUILD_DIR)/dropbear-0.44
 DROPBEAR_SSHD_CAT:=bzcat
 DROPBEAR_BINARY:=dropbearmulti