On 2/25/20 10:43 AM, Roman Bolshakov wrote:
I'm not an expert of FDMI feature but it seems to introduce an
inconsistency with regards to structure definition. IMO, All multi-byte
binary fields and bitmasks in RPA request should be made __be32 rather
than only one. Probably that should go into one patch series where all
fields in the structure are fixed patch-by-patch or one patch that fixes
all fields at once (the latter is harder to review though).
Hi Roman,
I agree with what you wrote. I have a (huge) patch pending that fixes
all endianness issues in the qla2xxx driver. I plan to post that patch
after agreement has been achieved about this patch series.
Thanks,
Bart.
