Mark Brown wrote:
On Tue, May 12, 2009 at 12:00:33AM +0200, Karl Beldan wrote:
Then it is perfectly normal since reset enables both REC and RPL, 2/4 needs 1/4. The current tree disables the clocks anytime one function is disabled.
That doesn't seem to tie up - I can see the initialisation changing the behaviour on first run but it seems surprising that this should happen on subsequent runs too. Alternatively, is your initialisation patch safe to apply by itself?
Well 2/4 stops the clocks only if both REC and RPL are disabled. Without 1/4 you end up with REC enabled at startup. In a scenario where you have never used REC you end up RPLing with REC always on. REC being on at shutdown(),clocks won't stop.
I resent 1/4 and will make it 1/5 with the whole serie once everything is Clear.
As previously discussed you need to rework the patch to not do the reset on initial probe not when the module is loaded, you need to address this rather than reposting.
The patch in question is moving the reset in probe rather than module init - with comment updated. What is wrong ?
I'll try to find time to re-review the series but I'm going to need to sit down with the datasheet and check this in much more detail.
For 1/4 and 2/4 there should not be great need, Really.