[OpenPower-Firmware] NVDIMM support

[Artem Senichev]

On Tue, Jan 14, 2020 at 10:51:35AM -0600, Daniel M Crowell wrote:
> 
> > I can't find corresponded code in Hostboot, the hdatMsAreaAddrRange_t
> structure from HB
> The Hostboot side of these changes is still under development.

If it's not a secret, when do you plan to publish these changes?

-- 
Regards,
Artem Senichev
Software Engineer, YADRO.

> 
> --
> Dan Crowell
> Senior Software Engineer - Power Systems Enablement Firmware
> IBM Rochester: t/l 553-2987
> dcrowell at us.ibm.com
> 
> 
> 
> From:	Artem Senichev <artemsen at gmail.com>
> To:	oohall at gmail.com
> Cc:	openpower-firmware at lists.ozlabs.org
> Date:	01/14/2020 07:00 AM
> Subject:	[EXTERNAL] [OpenPower-Firmware] NVDIMM support
> Sent by:	"OpenPower-Firmware" <openpower-firmware-bounces
>             +dcrowell=us.ibm.com at lists.ozlabs.org>
> 
> Hi Oliver,
> 
> Could you help me with NVDIMM support in OpenPOWER?
> 
> I found your commit to skiboot:
> 4822a7ba9 hdata/memory: Add NVDIMM support
> https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_open-2Dpower_skiboot_commit_4822a7ba9d33468a15d1c90ee8c0823826d33320&d=DwICAg&c=jf_iaSHvJObTbx-siA1ZOg&r=mCj3CQvqek9g0fdziO-GEHyU1m9T3SAh0ZPd5s_AGpo&m=K0Kjqn0MT7HSha62AXLVoFSdLolFOtgo2VsH0bPtH1M&s=-vQb9w72rwrFly1_Ok6pXJNJIneGiMerjRLaY5aLF-E&e=
> 
> 
> You added a new field to HDIF_ms_area_address_range structure - phys_attr,
> its value defines the type and state of the memory, but I can't find
> corresponded code in Hostboot, the hdatMsAreaAddrRange_t structure from HB
> doesn't contain these attributes.
> 
> Skiboot (HDIF_ms_area_address_range):
> https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_open-2Dpower_skiboot_blob_80fd2e963bd4364ee8c3b5a06215d8cbdfe04fcb_hdata_memory.c-23L31&d=DwICAg&c=jf_iaSHvJObTbx-siA1ZOg&r=mCj3CQvqek9g0fdziO-GEHyU1m9T3SAh0ZPd5s_AGpo&m=K0Kjqn0MT7HSha62AXLVoFSdLolFOtgo2VsH0bPtH1M&s=jrkb8AsbKPqtpx6LYuh8okQM4LDrKP21N9vcfvZPvUw&e=
> 
> Hostboot (hdatMsAreaAddrRange_t):
> https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_open-2Dpower_hostboot_blob_f70454687133f8a8b493f44d981ea1673e4ec437_src_usr_hdat_hdatmsarea.H-23L132&d=DwICAg&c=jf_iaSHvJObTbx-siA1ZOg&r=mCj3CQvqek9g0fdziO-GEHyU1m9T3SAh0ZPd5s_AGpo&m=K0Kjqn0MT7HSha62AXLVoFSdLolFOtgo2VsH0bPtH1M&s=Ab8ggCCIXD6lfD_--2SHyywSRLsGIlVXRnM9Y6R7eiw&e=
> 
> 
> How are these attributes populated?
> 
> P.S. Also, I suppose there is a typo in Skiboot's code:
> https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_open-2Dpower_skiboot_blob_80fd2e963bd4364ee8c3b5a06215d8cbdfe04fcb_hdata_memory.c-23L588&d=DwICAg&c=jf_iaSHvJObTbx-siA1ZOg&r=mCj3CQvqek9g0fdziO-GEHyU1m9T3SAh0ZPd5s_AGpo&m=K0Kjqn0MT7HSha62AXLVoFSdLolFOtgo2VsH0bPtH1M&s=jntoeAlCKtAStnmsvgrKLcc6VIsLnAappJUVU10-O5g&e=
> 
> IMHO, there should be a 'greater' condition, not 'greater or equal'.
> The current implementation has equal values of `arr->eactsz` and `offset`
> (36 bytes), so reading uint32_t from offset 36 looks strange.
> 
> --
> Regards,
> Artem Senichev
> Software Engineer, YADRO.