[parisc-linux] [PATCH] timer_interrupt and gettimeoffset.

[James Bottomley]

On Tue, 2006-09-05 at 16:49 -0600, Grant Grundler wrote:
> If we only have one TLB entry for kernel text (code), we should
> never have a miss because of code accesses. I thought that was
> the case but don't know for sure. ISTR some discussion about
> use of BTLB (PA1.1) and Large Pages (PA 2.0) hardwired for kernel.
> 
> Kernel data would probably benefit too.

Erm ... no, we never actually implemented this.  We still use 4k TLB
entries for the entire kernel and its data.

> Add trap latency (e.g. sync/RFI) and it's still pretty small (not more
> than "hundreds" of cycles.  I'm worried about orders of magnitutude
> bigger latencies.

Right ... it's a source of potential delay, but not a significant one.

James