drivers: sdhc: fix sdhc caps field offset

[natto1784]

The first field in the capabilities register, "Timeout Clock Frequency" takes up the first 6 bits (5:0), not 5.

Reference: 2.2.26, SD Specifications, Part A2, SD Host Controller
Simplified Specification, Version 4.20
URL: https://www.sdcard.org/downloads/pls/pdf/?p=PartA2_SD%20Host_Controller_Simplified_Specification_Ver4.20.jpg

Side note: I am writing an SDHCI driver and found that most drivers write their own struct/macros for the SDHC registers, is it possible to have a common definition for these registers to avoid duplicated code?

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[natto1784]

Actually on further inspection, seems like the struct was written as a bitfield to fit the capabilities register like a fiddle (hence the names of the reserved fields). However, some fields were added (and some were also omitted) that makes it very weird to look at. I will see if this needs more changes.

We can do two things

1. Get rid of reserved field
2. Make two sub-structs inside this, one that fits the standard spec and another for extra configurations (like hs200/hs400 support)

[natto1784]

Created a more comprehensive patch here: #91701 that splits standard capabilities from non standard capabilities. Please do let me know if this is desirable before I undraft it and ping every affected maintainer.

[danieldegrasse]

Created a more comprehensive patch here: #91691 that splits standard capabilities from non standard capabilities. Please do let me know if this is desirable before I undraft it and ping every affected maintainer.

This looks like the same PR- did you intend to link to a different draft?

[danieldegrasse]

Actually on further inspection, seems like the struct was written as a bitfield to fit the capabilities register like a fiddle (hence the names of the reserved fields). However, some fields were added (and some were also omitted) that makes it very weird to look at. I will see if this needs more changes.

We can do two things

Get rid of reserved field
Make two sub-structs inside this, one that fits the standard spec and another for extra configurations (like hs200/hs400 support)

Yup, the struct was originally written to follow the capabilities register in the SD host specification (although the offset error this PR fixes always existed), but when MMC support was added the struct deviated from that format. Can you update the link to your more comprehensive patch? I think making this struct match the SDHCI specification would be best

Side note: I am writing an SDHCI driver and found that most drivers write their own struct/macros for the SDHC registers, is it possible to have a common definition for these registers to avoid duplicated code?

Sure- the only reason common definitions don't exist right now is because not every SDHC IP I've encountered works well with the SDHCI specification, some need vendor specific registers

[natto1784]

Created a more comprehensive patch here: #91691 that splits standard capabilities from non standard capabilities. Please do let me know if this is desirable before I undraft it and ping every affected maintainer.

This looks like the same PR- did you intend to link to a different draft?

Oops my bad, edited: #91701

[natto1784]

Closing in favor of #91701