[Prism54-devel] Patches and controversial change

[Margit Schubert-While]

OK, patch coming to get 2.4 upto 2.6 after Linus's sparse
changes. Thereafter, we should be back to common changes.
Patch coming to further reduce sparse warnings, so that there
only remains 2 warnings from the union iwreq_data.data.pointer assignment
(defined with attribute __user in wireless.h)

Patch coming to enable device in resume code.

And now something which is probably controversial.
Recently I have access to a couple of cards/boxen that have
the same symptons as other people have reported -
Dropped connections, bouncy bitrate, association problems etc.
This somehow seems related to sleep mode (Note not PSM)
The only way I can get these reliably working is to disable the
SLEEP interrupt. (This anyway must be done for mode Master
and arguably for Adhoc; I would say definitely without knowlege of
what the firmware is doing)
Firmware 1.0.4.3 seems more susceptible to this problem than
1.0.3.0.

Unless someone screams no, I am going to generally disable the
SLEEP interrupt (Don't assert in enable_common_interrupts)
Until we know the ins and outs of this, lets just get the driver
motoring along.

Margit