Re: [RFC PATCH 1/3] Unified trace buffer

[Martin Bligh]

> If we just record the TSC unshifted, in 27 bits, at 4GHz, that gives us
> about 1/30 of a second? So we either shift, use > 27 bits, or record
> at least 30 events a second, none of which I like much ...

If we use 32 bits instead of 27, then the timestamp events are only
about once per second, which is probably fine for overhead ... ?

I think we're OK losing 5 bits of precision, that's only 32 cycles,
given all the CPU reordering stuff we've talked about here,
not-quite-synced TSCs, etc. I suspect you thought we were
shifting by much more than this, in reality it was 5-10 bits,
with timestamp events inbetween, though we put wall time in
them, which I think was a mistake.

(note: I'm not suggesting we have to use this compact a format,
at least by default)
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/