On Sat, Jun 29, 2013 at 3:58 PM, Andrew Fish <af...@apple.com> wrote: > On Jun 29, 2013, at 8:05 AM, "Zeng, Star" <star.z...@intel.com> wrote: >> About whether it is really worth it to allocate the HOB in this case, >> it is because we don't want to assume the potential call times of >> platform. > > I agree that slower FLASH/ROM speeds have a big impact on boot > time. More modern chipsets have improved the NOR FLASH speed, > but it is hard to get a general number on something like this.
Yes, it can be slow, but that doesn't mean we should create a HOB for any data accessed from flash. HOB memory can be tightly constrained, so I don't think we should get into the habit of using it unless we've shown that is really worth the trade off. Star, are you sure this isn't a case of premature optimization? By the way, we already accessing the flash directly for every other variable header. Why is this particular 'merge operation' so important? If we are accessing the each header several times, then maybe it would be faster to copy/merge the header for every variable, and then access the header from the copy on the stack. This discussion is a prime example of how our project could benefit from using git. These two "big" patches really deserve to be broken up somewhat. There is quite a lot happening here, and it would be nice to review and discuss it at a finer granularity. -Jordan ------------------------------------------------------------------------------ This SF.net email is sponsored by Windows: Build for Windows Store. http://p.sf.net/sfu/windows-dev2dev _______________________________________________ edk2-devel mailing list edk2-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/edk2-devel
