[Skiboot] [PATCH] Self Save: Fixed bugs pertaining to SPR self save.

Stewart Smith stewart at linux.vnet.ibm.com
Thu May 2 18:07:20 AEST 2019


Akshay Adiga <akshay.adiga at linux.vnet.ibm.com> writes:
> From: Prem Shanker Jha <premjha2 at in.ibm.com>
>
> Commit fixes some issues with code found during integration test
>   -  replacement of addi with xor instruction during self save API.
>   -  fixing instruction generation for MFMSR during self save
>   -  data struct updates in STOP API
>   -  error RC updates for hcode image build
>   -  HOMER parser updates.
>   -  removed self save support for URMOR and HRMOR
>   -  code changes for compilation with OPAL
>   -  populating CME Image header with unsecure HOMER address.
>
> Key_Cronus_Test=PM_REGRESS
>
> Change-Id: I7cedcc466267c4245255d8d75c01ed695e316720
> Reviewed-on: http://rchgit01.rchland.ibm.com/gerrit1/66580
> Tested-by: FSP CI Jenkins <fsp-CI-jenkins+hostboot at us.ibm.com>
> Tested-by: HWSV CI <hwsv-ci+hostboot at us.ibm.com>
> Tested-by: PPE CI <ppe-ci+hostboot at us.ibm.com>
> Tested-by: Jenkins Server <pfd-jenkins+hostboot at us.ibm.com>
> Tested-by: Cronus HW CI <cronushw-ci+hostboot at us.ibm.com>
> Tested-by: Hostboot CI <hostboot-ci+hostboot at us.ibm.com>
> Reviewed-by: Gregory S. Still <stillgs at us.ibm.com>
> Reviewed-by: RAHUL BATRA <rbatra at us.ibm.com>
> Reviewed-by: Jennifer A. Stofer <stofer at us.ibm.com>
> Reviewed-on: http://rchgit01.rchland.ibm.com/gerrit1/66587
> Reviewed-by: Christian R. Geddes <crgeddes at us.ibm.com>
> Signed-off-by: Akshay Adiga <akshay.adiga at linux.vnet.ibm.com>

So, we had a bit of a HCODE explosion around all of this.

Observations:
- op-build had new hcode merged that broke booting.
- op-build now has this commit:
commit e79abb32662de8d4afd5bca581487b4ab3e9827a
Author: Stewart Smith <stewart at linux.ibm.com>
Date:   Wed May 1 20:08:58 2019 -0500

    Revert HCODE back to hw040319a.940
    
    The HCODE updates break booting without disabling stop4/stop5.
    
    This should never have been merged as it solidly breaks booting,
    even the simple test of "boot to petitboot and exit to shell".
    
    The CI system had stop4/stop5 disabled via NVRAM.
    
    This partially reverts commit 3c6b318e43ffdf552351b99a19e6c78f6ac0c366
    and 5731c281fb8f8e5a20177d19bd3a503c50bc5748.
    
    Fixes: 3c6b318e43ffdf552351b99a19e6c78f6ac0c366
    Fixes: 5731c281fb8f8e5a20177d19bd3a503c50bc5748
    Signed-off-by: Stewart Smith <stewart at linux.ibm.com>

 which means that P9 will boot again

- Even with this patch applied, the newer hcode that hit op-build
  *still* failed to IPL.

So, I'm pretty skeptical of it all by this point, especially as I got to
waste a non-trivial amount of my day chasing after all that.

In talking with Vasant, we're still pretty sure this patch does fix
things on what hcode was in a recent (IBM internal) ZZ-L build.

But as is, I'm not going to take this patch until:
- there's a HCODE bump that actually works put up as a PR on op-bulid
  - This *explicitly* means that stop4/5 *MUST NOT* be disabled in order
    to get it to pass CI.
- some assurance that carrying this patch doesn't break
  forwards/backwards compat.

I'm rather inclined to add "and hcode project commits start going
through at the *VERY LEAST* a boot-to-an-os test with op-test (with a
view towards expanding that to run our full OpTestEM suite)" as a
requirement as well.


-- 
Stewart Smith
OPAL Architect, IBM.


More information about the Skiboot mailing list