libreboot
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Libreboot] new T400 beta ROMs for testing


From: persmule
Subject: Re: [Libreboot] new T400 beta ROMs for testing
Date: Sun, 21 Aug 2016 00:48:45 +0800
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:45.0) Gecko/20100101 Icedove/45.2.0

   Hi Leah,
   I have just finished testing (1095b5fd5a01c2392bab878fbc41b0c2f5a661d8,
   using cb revision 2a3434757ef425dbdfedf1fc69e1a033a6e7310d) on my
   t400s, which starts EC properly. But after I replaced the
   resources/libreboot/config/grub/t400_8mb/config with the one of the
   newest libreboot revision ($ git checkout master
   resources/libreboot/config/grub/t400_8mb/config ), the EC failed to
   init!
   It seems the regression leading to failure initialization of t400s' EC
   lies in the config file!
   But, if that is true, how can I do "bisect" on the config file?
   Persmule
   在 2016年08月20日 08:57, persmule 写道:

   Hi Leah,
   Till now, what I get is that the revision you update the coreboot
   revision (1a3cfd3ca6c59c49349925b25e30f6308c47fb26, using cb revision
   d83b0e9ac4174cca92ac2c3b83a7e8491a9a1ff4) failed to init EC on t400s
   (if I restart instead of shutdown the computer after flashing the rom
   when EC is functional, EC will be functional after reboot), and the
   revision just before that (1095b5fd5a01c2392bab878fbc41b0c2f5a661d8,
   using cb revision 2a3434757ef425dbdfedf1fc69e1a033a6e7310d) starts EC
   properly. And I failed to build the combination of (1095b5f,d83b0e9)
   due to some compilation error.
   Persmule
   在 2016年08月19日 12:54, Persmule 写道:

   Hi Leah,
   I manage to do bisect myself.
   I have noticed that the coreboot revision used by t400 has changed
   around Jul 17:

     $ git diff 23048a4 HEAD
     resources/libreboot/config/grub/t400_8mb/cbrevision
     diff --git a/resources/libreboot/config/grub/t400_8mb/cbrevision
     b/resources/libreboot/config/grub/t400_8mb/cbrevision
     index 1ab8299..d4e47be 100644
     --- a/resources/libreboot/config/grub/t400_8mb/cbrevision
     +++ b/resources/libreboot/config/grub/t400_8mb/cbrevision
     @@ -1 +1 @@
     -2a3434757ef425dbdfedf1fc69e1a033a6e7310d
     +d83b0e9ac4174cca92ac2c3b83a7e8491a9a1ff4

   Is it possible to do bisect like below?

     start bisect on coreboot's repository between the two revision, then
     write the current coreboot revision number given by $git show -s to
     resources/libreboot/config/grub/t400_?mb/cbrevision, then build
     libreboot as usual and flash and test, then mark the current
     revision on coreboot's repository and recurse.

   Persmule
   在 2016年08月11日 18:27, Leah Rowe 写道:

     Hi Persmule,
     Op 11/08/16 om 03:46 schreef Persmule:
     > Hi all,
     > Sorry, the different byte is actually 0x14 (024 in octal) in the
     > older and 0x24 (044 in octal) in the newer. I used "$cmp -bl" to
     > search the possible difference, which print the different byte in
     > octal.
     I made some changes to ich9gen, relating to the descriptor, some
     time
     ago. We found for instance that the "component density" section was
     wrong, so we fixed that. There were some other minor changes. Users
     have reported working wifi with those newer descriptors in earlier
     snapshots. The 96 snapshot uses a new revision of coreboot and as
     far
     as I know, this is all that has changed, aside from some minor
     patches
     changed which also shouldn't (in my opinion) affect wifi. We think
     upstream (coreboot project) might have broken it, so someone will
     have
     to bisect coreboot to see if there is a regression, and if so on
     what
     commit.

Attachment: signature.asc
Description: OpenPGP digital signature


reply via email to

[Prev in Thread] Current Thread [Next in Thread]