Re: [PATCH v2 0/2] bump bitbake and wic for python 3.10 support

[Henning Schild <henning.schild@siemens.com>]

Am Fri, 22 Apr 2022 10:26:08 +0200
schrieb Henning Schild <henning.schild@siemens.com>:

> Am Fri, 22 Apr 2022 10:15:22 +0200
> schrieb "Bezdeka, Florian (T CED SES-DE)"
> <florian.bezdeka@siemens.com>:
> 
> > On Fri, 2022-04-22 at 10:59 +0300, Anton Mikanovich wrote:  
> > > 21.04.2022 11:52, Henning Schild wrote:    
> > > > debian bookworm switched to python 3.10 recently, our versions
> > > > of bitbake and wic both have some issues with that new
> > > > interpreter. All these issues have been fixes in bitbake and
> > > > OE, so all we need to do is update the two tools in our tree.
> > > > 
> > > > v1 was send as "[PATCH] bitbake: Update to 1.50.5 release",
> > > > which is p1 here, only later i found that wic also needs a bump
> > > > to v2 added p2
> > > > 
> > > > If this series is applied we do not need "[PATCH] wic-img: use
> > > > python3.9 in bookworm", in case this one is applied before it
> > > > should be reverted after the version bumps.
> > > > 
> > > > Henning Schild (2):
> > > >    bitbake: Update to 1.50.5 release
> > > >    wic: Update to the latest revision
> > > > 
> > > >   bitbake/lib/bb/cache.py                       |  3 +-
> > > >   bitbake/lib/bb/cooker.py                      | 30 +++++++-
> > > >   bitbake/lib/bb/data_smart.py                  |  4 +-
> > > >   bitbake/lib/bb/fetch2/__init__.py             |  4 +
> > > >   bitbake/lib/bb/fetch2/perforce.py             |  2 +-
> > > >   bitbake/lib/bb/fetch2/wget.py                 |  2 +-
> > > >   bitbake/lib/bb/persist_data.py                |  5 +-
> > > >   bitbake/lib/bb/process.py                     |  2 +-
> > > >   bitbake/lib/bb/runqueue.py                    | 34 ++++----
> > > >   bitbake/lib/bb/server/process.py              |  2 +-
> > > >   bitbake/lib/bb/tests/fetch.py                 | 35 +++++----
> > > >   bitbake/lib/bb/utils.py                       | 13 +++-
> > > >   bitbake/lib/hashserv/server.py                |  4 +-
> > > >   bitbake/lib/toaster/tests/builds/buildtest.py |  2 +-
> > > >   .../wic/plugins/source/bootimg-efi-isar.py    | 77
> > > > ++++++++++++++++---
> > > > .../wic/plugins/source/bootimg-pcbios-isar.py |  6 +-
> > > > scripts/lib/wic/canned-wks/common.wks.inc     |  2 +-
> > > > scripts/lib/wic/canned-wks/directdisk-gpt.wks |  2 +-
> > > > scripts/lib/wic/canned-wks/mkefidisk.wks      |  2 +-
> > > > scripts/lib/wic/engine.py                     |  6 +-
> > > > scripts/lib/wic/help.py                       | 10 ++-
> > > > scripts/lib/wic/ksparser.py                   |  8 +-
> > > > scripts/lib/wic/misc.py                       |  4 +-
> > > > scripts/lib/wic/partition.py                  | 25 ++++--
> > > > scripts/lib/wic/pluginbase.py                 |  8 +-
> > > > scripts/lib/wic/plugins/imager/direct.py      | 11 ++-
> > > > scripts/lib/wic/plugins/source/bootimg-efi.py | 74
> > > > +++++++++++++++--- .../lib/wic/plugins/source/bootimg-pcbios.py
> > > > |  6 +- scripts/lib/wic/plugins/source/rawcopy.py     | 35
> > > > ++++++++- scripts/lib/wic/plugins/source/rootfs.py      |  2 +-
> > > > scripts/wic                                   |  9 ++- 31 files
> > > > changed, 320 insertions(+), 109 deletions(-)    
> > > 
> > > Thanks for update, python issue is really fixed now, but it
> > > brakes qemu386 targets booting in QEMU:  
> 
> Thanks, i will look into this.
>  
> > > 00:46:00 [   34.424136] blk_update_request: I/O error, dev fd0,
> > > sector 0 00:46:00 [   34.428049] floppy: error -5 while reading
> > > block 0 00:46:00 [   34.508133] blk_update_request: I/O error, dev
> > > fd0, sector 0 00:46:00 [   34.512050] floppy: error -5 while
> > > reading block 0 00:46:00 Gave up waiting for root file system
> > > device.  Common problems: 00:46:00  - Boot args (cat
> > > /proc/cmdline) 00:46:00    - Check rootdelay= (did the system
> > > wait long enough?) 00:46:00  - Missing modules (cat
> > > /proc/modules; ls /dev) 00:46:00 ALERT!  PARTLABEL=platform does
> > > not exist.  Dropping to a shell! 00:46:00 (initramfs)
> > >     
> > 
> > Might be upstream commit [1] which Henning pulled in now.
> > 
> > Adding Adriaan as author of this commit to CC. I have no idea why
> > this should break exactly this image...  
> 
> We boot all our i386 with BIOS directdisk-isar.wks and all amd64 with
> EFI sdimage-efi.wks, plus a few exceptions from that pattern where we
> use efi-plus-pcbios.wks or systemd-boot/EFI with sdimage-efi-sd.wks
> 
> So maybe that directdisk-isar.wks or common-isar.wks.inc need some
> bits i did not see and good old legacy BIOS is indeed broken in
> general with these patches.

PARTLABEL is a gpt thing, and we use msdos. And our common-isar.wks.inc
does not have --use-uuid while upstream common.wks.inc has it.

Will send a patch. Could not find if we ever dropped that use-uuid, i
think it is more likely it was overlooked while maintaining the fork.

Henning

> Henning
> 
> > [1]
> > https://github.com/openembedded/openembedded-core/commit/2fb247c5ecf057bb96649a3c0234794b4991c050
> > 
> > Regards,
> > Florian
> >   
>