Hi Nico, The fact that the prior release on GitHub did not include release notes is why I assumed it would work for the -U, when, in fact, it did not - and instead managed to turn the CRBs into boat anchors. First time I've ever seen an FSP be able to completely take down a CRB to where it can't be resurrected by restoring the original factory UEFI BIOS image.
Some FSPs have only listed what specific variants they support because those were the only variants tested/validated, even though the FSP might work on other variants that were not ever validated. But in this case, the KabyLake FSP on GitHub apparently only supports -H. It took the past three months to get this all sorted out, and to finally get a bootloader that boots on the RVP7 and RVP15 CRBs - something I should have been able to have up and running in a couple of hours - or worst case, a couple of days. At least now I can finally start focusing on porting from the RVP7/RVP15 coreboot implementation to my client's board design... - Jay Jay Talbott Principal Consulting Engineer SysPro Consulting, LLC 3057 E. Muirfield St. Gilbert, AZ 85298 (480) 704-8045 (480) 445-9895 (FAX) [email protected] http://www.sysproconsulting.com > -----Original Message----- > From: Nico Huber [mailto:[email protected]] > Sent: Monday, January 08, 2018 3:19 PM > To: Jay Talbott; 'Nico Huber'; 'Toan Le manh'; 'Naresh G. Solanki'; > 'coreboot'; > Stefan Reinauer > Cc: 'Youness Alaoui' > Subject: Re: [coreboot] Can coreboot for KabyLake rvp11 work on Intel Sasby > Island kit? > > Hi Jay, > > On 08.01.2018 19:35, Jay Talbott wrote: > > The release notes for the Kaby Lake FSP on github says it's specifically > > just for KabyLake-H. It makes no mention of supporting KabyLake-U. > > sorry, I might be just blind. But there don't seem to be release notes > for the version from June: > > [icon@bob FSP]$ git checkout KabylakeFsp0001 > HEAD is now at d88078a... initial Kabylake FSP check-in. > [icon@bob FSP]$ find . -iname *release* > [icon@bob FSP]$ git grep -i release > KabylakeFspBinPkg/Fsp.bsf: Help "Size of SMRAM memory reserved. > 0x400000 for Release build and 0x1000000 for Debug build" > KabylakeFspBinPkg/Include/FspmUpd.h: Size of SMRAM memory > reserved. > 0x400000 for Release build and 0x1000000 for Debug build > KabylakeFspBinPkg/SampleCode/Vbt/Vbt.bsf:$RelStage > 1 byte ; Release status > [icon@bob FSP]$ > > The release notes of the Gold release mention it, though. But, earlier > versions often listed only one variant while they supported all. Doesn't > matter any more, I guess. If they force people to reverse engineer their > stuff, that's ok for me (I'm pretty sure now that this is less effort, > than ensuring their undocumented blobs work). > > > After opening an IPS case on the issue, I discovered that the KabyLake > > FSP on github was developed by IoTG for KabyLake-H, while a different > > KabyLake FSP was developed by a completely different team at Intel for > > KabyLake-U (which was the FSP that was used for validating the kblrvp > > mainboard support in coreboot). This other KabyLake FSP is current not > > available on github (and I don't know if it ever will be). > > Now this is what I call a Google Support Package. > > Nico -- coreboot mailing list: [email protected] https://mail.coreboot.org/mailman/listinfo/coreboot
