On Tue, Feb 14, 2023 at 11:09:48AM +1100, Benjamin Herrenschmidt wrote: > On Thu, 2023-02-09 at 01:24 +0100, Mihai Moldovan wrote: > > > I'm tempted to increase this to speed things up on UEFI especially, > > > either > > > that or maybe changing grub_fshelp_read_file() so that get_block > > > can > > > optionally return a count of contiguous blocks so we can issue > > > larger > > > requests ? > > > > Is that really necessary? Note that this is also used for legacy > > MBR/BIOS > > systems and I faintly remember places using that, which may do so in > > a fixed > > way, so I advise to check it well if you really intend to change it. > > > That's ok as the MBS/BIOS bits will impose limits at other levels. > > The reason I want larger IOs is that it will help with cloud setups > where the EFI "disk" is actually some kind of networked storage. We > are chasing milliseconds in instance launch times :-) > > Anyways, I think the "right" approach is to make get_blocks() called > by grub_fshelp_read_file() optionally return the number of contiguous > blocks, so we can chose to do larger chunks.
It seems to me that this approach or something similar will be better than just increasing disk cache size. > I'll look into it if/when I have time. Cool! Thanks! Daniel _______________________________________________ Grub-devel mailing list Grub-devel@gnu.org https://lists.gnu.org/mailman/listinfo/grub-devel
