RMKs comments on latest OMAP upstream patches
Just fyi: Tony (thanks!) sent a OMAP1 and a OMAP2 patch to RMK for review and upstream merge. Regarding OMAP1, there are some comments from RMK, nothing yet for OMAP2:
OMAP1: http://www.arm.linux.org.uk/developer/patches/viewpatch.php?id=4570/1
OMAP2: http://www.arm.linux.org.uk/developer/patches/viewpatch.php?id=4571/1
While -rc7 is out [1] and we are short before final 2.6.23 release with merge window opening, we have to see if our patches will be merged.
Cheers
Dirk
Quoted 5 lines · Dirk Behme
Just fyi: Tony (thanks!) sent a OMAP1 and a OMAP2 patch to RMK for review and upstream merge. Regarding OMAP1, there are some comments from RMK, nothing yet for OMAP2:
Quoted 8 lines
OMAP1:
http://www.arm.linux.org.uk/developer/patches/viewpatch.php?id=4570/1OMAP2:
http://www.arm.linux.org.uk/developer/patches/viewpatch.php?id=4571/1While -rc7 is out [1] and we are short before final 2.6.23 release with merge window opening, we have to see if our patches will be merged.
Yeh, I've removed the clock init RMK does not like, and will change them to use just struct clk *clk instead on the fbdevel list.
Regards,
Tony
Quoted 1 line
Quoted 5 lines
Hum, I should Russell's comment reference. Yes there is a clock register function, however, there is no, command to create a new node. Further you are not supposed to know what the node structure is. In the TI tree variant I did add a node create and destroy function. However, it does cheat and exports some parameters which imply internal knowledge of the clock structure.
It would seem the best you could do was to pre-install some dummy nodes. However, you still need a way to fill in the function pointers... Again, need check his pointer.
Regards,
Richard W.
Quoted 5 lines · Woodruff, Richard
Quoted 5 lines
Looks like Russell has applied these two patch series above!
Quoted 3 lines
Hum, I should Russell's comment reference. Yes there is a clock register function, however, there is no, command to create a new node. Further you are not supposed to know what the node structure is. In the TI tree variant I did add a node create and destroy function. However, it does cheat and exports some parameters which imply internal knowledge of the clock structure.
It would seem the best you could do was to pre-install some dummy nodes. However, you still need a way to fill in the function pointers... Again, need check his pointer.
For board specific GPIO clock control clk_register() with custom enable/disable functions should work just fine though.
Tony
Quoted 10 lines · Tony Lindgren
Quoted 5 lines · Woodruff, Richard
Quoted 5 lines
Looks like Russell has applied these two patch series above!
Great!
I'm slightly wondering if this "devel tree" he mentions is any constraint or if it is his "send to Linus" tree. Let us see.
Regards
Dirk
Quoted 9 lines · Tony Lindgren
Quoted 5 lines · Woodruff, Richard
Quoted 5 lines
Looks like Russell has applied these two patch series above!
... and Linus as well :)
See changelog for patch-2.6.23-git9 (search for "ARM: OMAP:"):
http://www.kernel.org/pub/linux/kernel/v2.6/snapshots/patch-2.6.23-git9.log
Dirk
Btw: Do we have already any plans for next merge window which opens with -rc5?
Quoted 7 lines · Dirk Behme
Quoted 6 lines · Tony Lindgren
Quoted 5 lines · Woodruff, Richard
Quoted 5 lines
Quoted 7 lines
Quoted 1 line
Looks like Russell has applied these two patch series above!
... and Linus as well :)
See changelog for patch-2.6.23-git9 (search for "ARM: OMAP:"):
http://www.kernel.org/pub/linux/kernel/v2.6/snapshots/patch-2.6.23-git9.log
Great, I'll tag 2.6.23-omap1 soon, then let's do a pull.
Quoted 2 lines
Btw: Do we have already any plans for next merge window which opens with -rc5?
Yeh, let's try to get plat-omap and mach-omap1 synced.
Tony