Re: [PATCH V3 04/25] smartpqi: add support for raid5 and raid6 writes

On Fri, 2021-01-08 at 22:56 +0000, Don.Brace@microchip.com wrote:

quoted

+               }
+               if (rmd->blocks_per_row == 0)
+                       return PQI_RAID_BYPASS_INELIGIBLE; #if 
+BITS_PER_LONG == 32
+               tmpdiv = rmd->first_block;
+               do_div(tmpdiv, rmd->blocks_per_row);
+               rmd->row = tmpdiv;
+#else
+               rmd->row = rmd->first_block / rmd->blocks_per_row; 
+#endif

Why not always use do_div()?

Don: I had removed the BITS_PER_LONG check, was an attempt to clean
up the code, but forgot we still need to support 32bit and I just re-
added BITS_PER_LONG HUNKS. These HUNKS were there before I refactored
the code so it predates me. Any chance I can leave this in? It's been
through a lot of regression testing already...

My suggestion was to rather do the opposite, use the 32bit code (with
do_div()) for both 32bit and 64bit. AFAIK, this would work just fine 
(but not vice-versa). 

You can leave this in. It was just a suggestion how to improve
readability. Perhaps consider cleaning it up sometime later.

Regards,
Martin