On 1/4/2022 2:27 pm, Joel Sherrill wrote: > On Thu, Mar 31, 2022, 10:17 PM Chris Johns <chr...@rtems.org > <mailto:chr...@rtems.org>> wrote: > On 1/4/2022 11:14 am, Joel Sherrill wrote: > > This looks ok. Does it impact any array declaration that I missed? I'm > just > > worrying about changes in stack consumption. > > Not that I could see. A buffer allocated on the heap is now bigger to > hold the > larger data block and that has been adjusted. > > Ok. That's there only issue I had. > > I have an application where the JFFS slowed down from 4.11 to 6 and the > only > difference I could see was 6 had a lot more, I would say 50%, small > blocks being > written to the flash device over 4.11. This change makes using untar on > JFFS > much more efficient. > > That's odd by itself to have more smaller blocks but great to get the > performance boost.
The JFFS is compressing and with 512 byte blocks to work with you get a lot of small blocks. In production on 4.11 some of the files are AES encrypted and they do not compress so it is not as noticeable. Chris _______________________________________________ devel mailing list devel@rtems.org http://lists.rtems.org/mailman/listinfo/devel
