Re: [PATCH 2/8] Buslogic: remove ISA support

[Khalid Aziz <khalid@gonehiking.org>]

On 3/30/21 11:03 AM, Christoph Hellwig wrote:
> On Mon, Mar 29, 2021 at 02:29:21PM -0600, Khalid Aziz wrote:
>> On 3/25/21 11:58 PM, Christoph Hellwig wrote:
>>> The ISA support in Buslogic has been broken for a long time, as all
>>> the I/O path expects a struct device for DMA mapping that is derived from
>>> the PCI device, which would simply crash for ISA adapters.
>>>
>>> Signed-off-by: Christoph Hellwig <hch@lst.de>
>>> ---
>>>  drivers/scsi/BusLogic.c | 156 ++--------------------------------------
>>>  drivers/scsi/BusLogic.h |   3 -
>>>  drivers/scsi/Kconfig    |   2 +-
>>>  3 files changed, 6 insertions(+), 155 deletions(-)
>>>
>>
>> Hi Chris,
>>
>> This looks good. There is more code that can be removed, for instance
>> all of the code that supports "IO:" driver option to specify ISA port
>> addresses. enum blogic_adapter_bus_type can shrink. "limited_isa" and
>> "probe*" members of struct blogic_probe_options can go away. You could
>> add those to this patch, or if you would like, I can create a follow-on
>> patch to remove that code.
> 
> I've added the above suggestions.  If there is anything more you
> can easily think of let me know.
> 

Awesome! Thanks. Updates to Documentation/scsi/BusLogic.rst to match
these changes would be great. Doc currently lists "IO:" and "NoProbeISA"
which can go away. "Supported Host Adapters: section lists ISA and EISA
adapters that can go away as well. There is reference to ISA in
"QueueDepth:<integer>" - "For Host Adapters that require ISA Bounce
Buffers, the Queue Depth is automatically set by default to
BusLogic_TaggedQueueDepthBB or BusLogic_UntaggedQueueDepthBB to avoid
excessive preallocation of DMA Bounce Buffer memory." which is
irrelevant now.

Thanks,
Khalid