[poky] What's taking so long??

[Gary Thomas]

On 01/21/2011 07:10 AM, Richard Purdie wrote:
> On Thu, 2011-01-20 at 08:11 -0700, Gary Thomas wrote:
>> My target platform/distro uses IPK and I have a local kernel
>> recipe which is based on the pre linux-yocto recipe that was
>> used by Poky.  When I build&  package the kernel, I have
>> terrible performance - this is new, it used to be just fine.
>>
>> Here's what I mean.  Building for my kernel and timing the
>> various steps, e.g.
>>     % time bitbake virtual/kernel -c compile
>>
>> My kernel:
>>     compile            1m49s
>>     deploy             0m16s
>>     populate_sysroot   0m38s
>>     package           23m20s
>>
>> Just to compare, using OTS Poky for BeagleBoard:
>>     compile            5m07s
>>     deploy             0m36s
>>     populate_sysroot   0m29s
>>     package            5m02s
>>
>> That's a five-fold difference in the time to build&  package
>> the kernel.
>
> This is a serious and worrying number. I just tried a test where I ran:
>
> bitbake virtual/kernel -c package | ../predate.sh
>
> where predate.sh contains:
>
> #!/bin/bash
> while read line ; do
>      echo "$(date): ${line}"
> done
>
> and I added a bb.error to the package functions interation loop in
> package.bbclass.
>
> I saw:
>
> Fri Jan 21 13:58:00 GMT 2011: NOTE: Running task 667 of 667 (ID: 14, /media/build2/builds/rptest/meta/recipes-kernel/linux/linux-yocto_git.bb, do_package)
> Fri Jan 21 13:58:00 GMT 2011: NOTE: package linux-yocto-2.6.37+git0+e3895db37bd65ea97e59b70ea936c84b398d3f3e_0+a9d833fda90e2f1257888a97e092135610b5f259-r15: task do_package: Started
> Fri Jan 21 13:58:00 GMT 2011: ERROR: Running perform_packagecopy
> Fri Jan 21 13:58:02 GMT 2011: ERROR: Running package_do_split_locales
> Fri Jan 21 13:58:02 GMT 2011: ERROR: Running populate_packages
> Fri Jan 21 13:58:02 GMT 2011: WARNING:<string>:5: RuntimeWarning: tmpnam is a potential security risk to your program
> Fri Jan 21 13:58:02 GMT 2011:
> Fri Jan 21 13:58:18 GMT 2011: ERROR: Running package_do_filedeps
> Fri Jan 21 13:59:08 GMT 2011: ERROR: Running package_do_shlibs
> Fri Jan 21 13:59:10 GMT 2011: ERROR: Running package_do_pkgconfig
> Fri Jan 21 13:59:10 GMT 2011: ERROR: Running read_shlibdeps
> Fri Jan 21 13:59:11 GMT 2011: ERROR: Running package_depchains
> Fri Jan 21 13:59:11 GMT 2011: ERROR: Running emit_pkgdata
> Fri Jan 21 13:59:11 GMT 2011: ERROR: Running do_package_qa
> Fri Jan 21 14:00:42 GMT 2011: NOTE: package linux-yocto-2.6.37+git0+e3895db37bd65ea97e59b70ea936c84b398d3f3e_0+a9d833fda90e2f1257888a97e092135610b5f259-r15: task do_package: Succeeded
> Fri Jan 21 14:00:42 GMT 2011: NOTE: Tasks Summary: Attempted 667 tasks of which 615 didn't need to be rerun and 0 failed.
>
> so it did this in 2.6 minutes which doesn't seem unreasonable.
>
> Are you seeing the slow package step in current master and when you made
> the timings above, were they the only task running or were there others
> active too?

This particular test was isolated, just that one recipe.  I noticed
it before when building a complete image, so running it isolated does not
seem to make any difference.  There is also nothing else running on my build
machine (it's a separate server that is used only for builds)

This is against master as of 2011-01-14 (897530621fabf3c345c302050a5c21e16097a2ba)
with my local layers.

What can I look at (alternatively what data can I provide) to help
understand this?

Thanks


-- 
------------------------------------------------------------
Gary Thomas                 |  Consulting for the
MLB Associates              |    Embedded world
------------------------------------------------------------