Re: [PATCH 0/3] ata: Module parameter clean-ups for pata_legacy and pata_platform

[Jens Axboe]

On 3/21/21 1:55 PM, Maciej W. Rozycki wrote:
> Hi,
> 
>  In the course of looking into Christoph's recent proposal to drop legacy 
> IDE drivers I have come across a number of issues with module parameters 
> of the pata_legacy and pata_platform drivers: errors in documentation 
> present in the comment form, missing user-visible documentation, and 
> unconditional poking at ISA I/O ports in pata_legacy that isn't there with 
> the old ide-generic driver (the lack of `probe_mask' parameter).
> 
>  Here's a small patch series that addresses these issues.  Overall I 
> find the design of the pata_legacy driver's options a bit messy, e.g. the 
> `all' vs the `probe_all' parameter, and the interpretation of masks where 
> bits correspond to probed PATA locations in a particular system (rather 
> than either all known or all existing), but it's been there long enough I 
> think we have to keep it, so I merely tried to describe the current 
> semantics.  See the individual change descriptions for details.
> 
>  The changes have been run-time verified with an EISA system and a single 
> ISA PATA adapter at the usual primary I/O location.  They have also been 
> verified (mainly for the correctness of MODULE_PARM_DESC use) with an 
> x86/PC build (for pata_legacy) and a MIPS/SWARM build (for pata_platform).

Applied, thanks.

-- 
Jens Axboe