Linus Torvalds writes: (Summary) On Thu, Feb 8, 2018 at 1:47 AM, Ingo Molnar <mingo@kernel.org> wrote:
You make a good argument.
If the only real downside is slightly bigger static footprint (but the actual real dynamic I$ footprint is not really impacted), then let's just go for the simpler code.
just go for the simpler code.
It really would be lovely if our entry asm was understandable to mere mortals. It really would be lovely if our entry asm was understandable to mere mortals.
[...]
want to _complicate it_ beyond this simple, straightforward model You make a good argument.You make a good argument.
If the only real downside is slightly bigger static footprint (but the actual real dynamic I$ footprint is not really impacted), then let's just go for the simpler code.
just go for the simpler code.
It really would be lovely if our entry asm was understandable to mere mortals. It really would be lovely if our entry asm was understandable to mere mortals.