[poky] Using RPM with Poky

Mon Mar 21 23:22:23 PDT 2011

On Mon, 2011-03-21 at 17:55 -0700, Gary Thomas wrote:
> On 03/21/2011 05:24 PM, Darren Hart wrote:
> >
> >
> > On 03/21/2011 03:58 PM, Gary Thomas wrote:
> >> On 03/21/2011 03:02 PM, Tom Zanussi wrote:
> >>> On Mon, 2011-03-21 at 13:25 -0700, Bruce Ashfield wrote:
> >>>> On 11-03-21 04:20 PM, Gary Thomas wrote:
> >>>>> On 03/21/2011 01:45 PM, Darren Hart wrote:
> >>>>>>
> >>>>>>
> >>>>>> On 03/21/2011 11:33 AM, Gary Thomas wrote:
> >>>>>>> On 03/21/2011 12:27 PM, Darren Hart wrote:
> >>>>>>>>
> >>>>>>>>
> >>>>>>>> On 03/21/2011 10:57 AM, Richard Purdie wrote:
> >>>>>>>>> On Mon, 2011-03-21 at 10:35 -0600, Gary Thomas wrote:
> >>>>>>>>>> On 03/21/2011 09:54 AM, Richard Purdie wrote:
> >>>>>>>>>>> On Mon, 2011-03-21 at 09:25 -0600, Gary Thomas wrote:
> >>>>>>>>>>>> I'm trying to do some testing with Poky on an atom-PC. I thought
> >>>>>>>>>>>> I'd try with a minimal system and then add pieces I need. Sadly,
> >>>>>>>>>>>> I've run aground doing this. I've built a couple of images with
> >>>>>>>>>>>> varying results:
> >>>>>>>>>>>> poky-image-minimal-live - boots on my box
> >>>>>>>>>>>> poky-image-sato-live - boots, but fails to find X display, then
> >>>>>>>>>>>> hangs (*)
> >>>>>>>>>>>>
> >>>>>>>>>>>> There seem to be no RPM tools on the minimal image. Did I miss
> >>>>>>>>>>>> something?
> >>>>>>>>>>>> On my old, ipk based system, I'd build up a minimal image
> >>>>>>>>>>>> then use
> >>>>>>>>>>>> opkg
> >>>>>>>>>>>> to install additional packages.
> >>>>>>>>>>>
> >>>>>>>>>>> Minimal images were never meant to ship with the package manager
> >>>>>>>>>>> and now
> >>>>>>>>>>> don't. You can obviously change the image definitions easily
> >>>>>>>>>>> enough as
> >>>>>>>>>>> needed.
> >>>>>>>>>>>
> >>>>>>>>>>>> How do I do that with the RPM based images?
> >>>>>>>>>>>
> >>>>>>>>>>> The same switch turns on/off the package manager data for rpm and
> >>>>>>>>>>> opkg.
> >>>>>>>>>>
> >>>>>>>>>> I tried adding this to local.conf, but nothing changed when I
> >>>>>>>>>> build
> >>>>>>>>>> minimal image:
> >>>>>>>>>> IMAGE_FEATURES += " package-management ssh-server-dropbear "
> >>>>>>>>>
> >>>>>>>>> There is also this:
> >>>>>>>>>
> >>>>>>>>> # remove not needed ipkg informations
> >>>>>>>>> ROOTFS_POSTPROCESS_COMMAND += "remove_packaging_data_files ; "
> >>>>>>>>>
> >>>>>>>>> in poky-image-minimal.bbclass which has been there for a long
> >>>>>>>>> time and
> >>>>>>>>> removes the package manager data files.
> >>>>>>>>>
> >>>>>>>>>
> >>>>>>>>>> BTW, where is this documented? Not in the Poky handbook from
> >>>>>>>>>> what I
> >>>>>>>>>> can tell.
> >>>>>>>>>>
> >>>>>>>>>> Also, is there documentation on how to use zypper, how to set up
> >>>>>>>>>> repositories, etc?
> >>>>>>>>>
> >>>>>>>>> I think others have helped here. Please do send patches for the
> >>>>>>>>> manual
> >>>>>>>>> when people go to the trouble of finding information to help
> >>>>>>>>> though!
> >>>>>>>>>
> >>>>>>>>>>>> I've checked the documentation and there's basically nothing
> >>>>>>>>>>>> about this, unless I looked in the wrong place.
> >>>>>>>>>>>>
> >>>>>>>>>>>> Any ideas or pointers gladly accepted. Thanks
> >>>>>>>>>>>>
> >>>>>>>>>>>> (*) I had success with this in the past when I booted with
> >>>>>>>>>>>> MACHINE=emenlow
> >>>>>>>>>>>> but that hardware isn't in the primary list.
> >>>>>>>>>>>
> >>>>>>>>>>> I'm not sure why this isn't working or what the problem might
> >>>>>>>>>>> be...
> >>>>>>>>>>
> >>>>>>>>>> It complains it can't find /dev/fb0. See the boot log at
> >>>>>>>>>> http://www.mlbassoc.com/poky/boot_2011-03-21.log
> >>>>>>>>>
> >>>>>>>>> Does your board have Intel graphics or PVR/SGX? emenlow has PVR/SGX
> >>>>>>>>> which might explain this...
> >>>>>>>>
> >>>>>>>> And if you need emenlow, that bsp is now part of the meta-intel
> >>>>>>>> layer:
> >>>>>>>>
> >>>>>>>> http://git.pokylinux.org/cgit.cgi/meta-intel/
> >>>>>>>
> >>>>>>> Do you have instructions/suggestions on how to add this to my extant
> >>>>>>> Poky tree?
> >>>>>>>
> >>>>>>
> >>>>>> It's just another layer, so you checkout it out somewhere and add the
> >>>>>> meta-intel/meta-$MACHINE directory to your conf/bblayer.conf file.
> >>>>>
> >>>>> I did that, changed MACHINE=emenlow, but I got this error:
> >>>>>
> >>>>> | cc1: warnings being treated as errors
> >>>>> | cc1: error: include location "/usr/local/include" is unsafe for
> >>>>> cross-compilation
> >>>>> | make: *** [util/scripting-engines/trace-event-perl.o] Error 1
> >>>>> | make: Leaving directory
> >>>>> `/home/local/pc_test/tmp/work/emenlow-poky-linux/linux-yocto-2.6.37+git2+29047c254624e0bd8a0ac6da92862f7c6357cb0b_1+c6299ae5bece8e3a6e1bc2c236862ae004629aae-r16/linux/tools/perf'
> >>>>>
> >>>>>
> >>>>> | FATAL: oe_runmake failed
> >>>>> | ERROR: Function 'do_compile_perf' failed (see
> >>>>> /home/local/pc_test/tmp/work/emenlow-poky-linux/linux-yocto-2.6.37+git2+29047c254624e0bd8a0ac6da92862f7c6357cb0b_1+c6299ae5bece8e3a6e1bc2c236862ae004629aae-r16/temp/log.do_compile_perf.10359
> >>>>>
> >>>>>
> >>>>> for further information)
> >>>>>
> >>>>> This from yesterday's master (aeaa356a5ee77b4596c479451a9db289381a4d16)
> >>>>>
> >>>>> I recall seeing similar errors recently, but not the solution. Is there
> >>>>> an easy fix?
> >>>>
> >>>> It means that the SRCREV for the branch doesn't have the
> >>>> following commit covered:
> >>>>
> >>>> ----------
> >>>>
> >>>> commit 2b412826bbeb4a16abe2ea74f2456ab880c6e3c1
> >>>> Author: Bruce Ashfield<bruce.ashfield at windriver.com>
> >>>> Date: Fri Feb 25 13:18:51 2011 -0500
> >>>>
> >>>> perf: hard-code NO_LIBPERL/NO_LIBPYTHON
> >>>>
> >>>> ExtUtils::Embed ccopts is getting the host's -I/usr/local/include and
> >>>> using it to compile perf, which results in a compilation error that
> >>>> started appearing just recently.
> >>>>
> >>>> This turns the code that makes use of ExtUtils::Embed off and simply
> >>>> hard-codes NO_LIBPERL.
> >>>>
> >>>> It does the same for LIBPYTHON while we're at it, since it probably
> >>>> suffers from a similar underlying problem and just by chance hasn't
> >>>> broken anything yet.
> >>>>
> >>>> This will be re-enabled after I familiarize myself with the perf
> >>>> recipe and am able to create a proper fix.
> >>>>
> >>>> Signed-off-by: Tom Zanussi<tom.zanussi at intel.com>
> >>>> Signed-off-by: Bruce Ashfield<bruce.ashfield at windriver.com>
> >>>>
> >>>> ----------
> >>>>
> >>>> The kernel repo has this, and I thought that the meta-intel
> >>>> emenlow had an updated SRCREV .. but your build error indicates
> >>>> otherwise.
> >>>>
> >>>
> >>> Hmm, strange - the SRCREV does look like it needed to be updated but yet
> >>> I nor the autobuilder have been seeing this.
> >>>
> >>> Anyway, I just updated the kernel SRCREVs in master, please try again...
> >>
> >> It work now, thanks. It also found my display and SATA devices. Sadly, when
> >> it's running X, I can't use the mouse. It seems to be found, but doesn't
> >> work.
> >>
> >> I get these messages during boot:
> >> usb 1-1.2: USB disconnect, address 6
> >> usb 1-1.2: new low speed USB device using ehci_hcd and address 7
> >> input: Logitech USB-PS/2 Optical Mouse as
> >> /devices/pci0000:00/0000:00:1d.7/usb1/1-1/1-1.2/1-1.2:1.0/input/input4
> >> generic-usb 0003:046D:C03D.0005: input: USB HID v1.10 Mouse [Logitech
> >> USB-PS/2 Optical Mouse] on usb-0000:00:1d.7-1.2/input0
> >> And there is a device created:
> >> root at emenlow:~# ls -l /dev/input/
> >> drwxr-xr-x 2 root root 60 Mar 21 22:27 by-id
> >> drwxr-xr-x 2 root root 60 Mar 21 22:27 by-path
> >> crw-r----- 1 root root 13, 63 Mar 21 06:31 mice
> >> crw-r----- 1 root root 13, 32 Mar 21 22:27 mouse0
> >>
> >> Ideas?
> >
> >
> > By "doesn't work" do you mean that you don't see the cursor? Or do you see the cursor, but it doesn't move? Or it moves but clicking has no effect?
> >
> > If "I don't see a cursor" try looking at the formfactor file and check the value of "HAVE_TOUCHSCREEN":
> >
> > # grep -r TOUCHSCREEN /etc/formfactor/
> > /etc/formfactor/config:if [ -z "$HAVE_TOUCHSCREEN" ]; then
> > /etc/formfactor/config: HAVE_TOUCHSCREEN=0
> > /etc/formfactor/machconfig:HAVE_TOUCHSCREEN=0
> >
> > You want that to be 0, if it's 1, then matchbox will hide the mouse. See /etc/matchbox/session
> 
> I had already checked that file - it's correct & I do see a cursor, it just doesn't move.
> 

The mouse on my emenlow system, a WEBS-2120 box, works ok with a fresh
build from master.  Here's my dmesg output:

usb 2-1: new low speed USB device using uhci_hcd and address 2
input: USB Optical Mouse
as /devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1:1.0/input/input2
generic-usb 0003:0461:4D65.0001: input: USB HID v1.11 Mouse [USB Optical
Mouse] on usb-0000:00:1d.0-1/input0

The difference seems to be uhci_hcd vs ehci_hcd in your case...

Tom

> >>
> >> Now, if I can just figure out how to make the [live] image larger...
> >> (free space)
> >> I can get on with the real testing I need to do.
> >> Thanks for the help
> >>
> >
> 
> -- 
> ------------------------------------------------------------
> Gary Thomas                 |  Consulting for the
> MLB Associates              |    Embedded world
> ------------------------------------------------------------