From mboxrd@z Thu Jan 1 00:00:00 1970 From: michal.simek@xilinx.com (Michal Simek) Date: Wed, 30 Nov 2016 14:17:10 +0100 Subject: arasan,sdhci.txt "compatibility" DT binding In-Reply-To: <648bd95b-5be0-b39f-a956-af741c6407e4@laposte.net> References: <56B07BA5.4060407@free.fr> <583C50E7.6030400@free.fr> <14771688.Q8f97gHC2H@wuerfel> <583C60CC.2020408@free.fr> <648bd95b-5be0-b39f-a956-af741c6407e4@laposte.net> Message-ID: <215d077d-cba8-3c95-eb36-9d5a7afb9bcd@xilinx.com> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On 30.11.2016 11:51, Sebastian Frias wrote: > On 29/11/16 08:29, Rameshwar Sahu wrote: >> Hi Mason, >> >> Nowhere in the documentation do they specify an "IP version". >> Some documents do provide a revision number, but that's just >> a *documentation* revision number, e.g. >> >> changes in version 3.6 : fix typos >> changes in version 9.1a : update company logo >> >> That's why Xilinx used "arasan,sdhci-8.9a" and APM used >> "arasan,sdhci-4.9a". These are documentation revisions. >> In my opinion, that information is mostly worthless. >> > > For the record, the important information conveyed by Rameshwar's > email is the following: > > Arasan SD/SDIO/eMMC IP has a register which tells about the SD > specification version and Vendor version number > Reg Name: Host controller version register (offset 0FEh) > bit [15:8] is for vendor version number, > But, I have seen that Arasaan vendor version number is same as > document revision number. > > (At first I had ignored the email because it repeated Mason's email > without quoting, but then I realised it contained some information) > > Values on real HW. ZynqMP device xsdb% mrd 0xFF1600FC FF1600FC: 10020000 Zynq device: xsdb% mrd 0xE01000FC E01000FC: 89010000 Based on docs I have access to. Specification_Version_Number (bits:23:16) 00 - SD Host Specification version 1.0 01 - SD Host Specification version 2.00 including only the feature of the Test Register 02 - SD Host Specification version 3.00 Vendor_Version_Number 31:24 with 32bit access from FC Thanks, Michal