Hello Andrew,

So, to come to a conclusion about this complex patch series, I
attached all the latest versions to this mail. The latest patches from
yesterday including inline are also included to make the set complete.

So, this is the order in which the patches should be installed:
1) atmel_serial_cleanup.patch
2) atmel_serial_irq_splitup.patch
3) NEW: optional:  add-atmel-serial-dma.patch, this merged the DMA
code (from Chip Coldwell) in your 2.6.23 patch back on top of this
series. Because the AT32 bug is not been fixed for a very long time, I
do not expect it to be fixed soon, so I think a reordering is better
to make preempt-RT work on AT91.
4) For RT only: atmel_serial_irqf_nodelay.patch can be applied
anywhere after 1 and 2

Steven, for the RT-patch, patches 1,2 and 4 are the _only_ patches
needed to make RT work stable on AT91 (based on 2.6.23.9-rt13), where
I think that 1 and 2 should go to mainline during time. I tested this
and it works properly, so we can, at least create a working/bootable
preempt-RT system on AT91. Notice that without the other patches from
Andrew the clock granularity is bound to 10ms. (Partly solved in
2.6.24-rc* by mainlining of the clocksource/events patches form David
Brownell)

If any of you have any comments, or suggestions, please let me know.

Kind Regards,

Remy Bohmer