| Age | Commit message (Collapse) | Author |
|---|
|
attachment and interface, one screen-attached for the real work.
The attachment code is now required to decide whether sti_end_attach() is
run immediately, or as a startuphook.
This allows hp300 to initialize sti early, and use it as a console; hppa
is functionally unchanged, as it uses the PROM console until the root device
is mounted.
|
|
console-dependent routines know which pass they are being invoked in.
The first pass is the usual console detection code. The second pass runs
after uvm is initialized, and will allow devices which need VM services to
run to claim the console.
|
|
switch, which is a physical switch on this model, unlike other frodo-based
hp300.
Service mode will now select serial console, while normal mode will select
graphics console.
|
|
|
|
Prevents monkeys pounding on your serial terminal from panicing your kernel
before getty gets spawned on the serial ports.
dcm(4) looks safe from this problem from code inspection.
|
|
lack of hw
|
|
|
|
returning, so that rasops bursts including non-accelerated operations
do not have side effects.
|
|
Features:
- coloured kernel messages
- color support in console (use TERM=wsvt25)
- pretty Gallant wsfont, rather than the acceptable or ugly (depending upon
the frame buffer) ROM font
- colormap ioctl support for Gatorbox and Topcat; DaVinci and Renaissance
not done due to lack of hardware and/or documentation
- even more code factorization
tested on hyperion, and 4/6/8bpp catseye/topcat/kathmandu hardware, by
millert@ and I
|
|
|
|
|
|
|
|
|
|
and collateral changes.
Because this driver requires us_dma (and as such, vm services) to work, it
can not be selected in consinit(). Instead, add a hook to the arm
cpu_startup() which will, on zaurus, switch console from serial (selected
in consinit()) to lcd.
This also makes the zaurus-specific early pxa2x0_clkman() substitute code
cleaner.
While there, move boot -c handling later, after the glass console is set up.
Tested by drahn@ and uwe@
|
|
|
|
|
|
|
|
|
|
large enough constant.
|
|
the video registers size, this value can easily be computed from the standard
fb registers.
|
|
an Hyperion in internal space when looking for the console device.
|
|
|
|
devices for which diodevs has the wrong span value.
Before this, we would get:
tiger display at dio0 scode 157 not configured
internal HP-IB at dio0 scode 158 not configured
internal HP-IB at dio0 scode 159 not configured
Now, we will get:
tiger display at dio0 scode 157 not configured
dio0: warning: select code 158 is likely a span of a previous unsupported device
dio0: warning: select code 159 is likely a span of a previous unsupported device
|
|
|
|
|
|
gatorbox during console initialization. Oops.
|
|
|
|
|
|
Main features:
- MI hil drivers, allowing live plug/unplug of devices, and the keyboard
does not need to be at id 1 anymore.
- SGC bus support on 4xx models. SGC frame buffers are supported too, but
not as console, though you can use them as a regular glass terminal if
you run a getty on ttyC0. Currently not compiled-in, awaiting for an sti(4)
fix to be commited.
- HP-UX compatibility interfaces are removed. grfinfo and hilinfo are removed.
X11 support is currently broken by these changes; X.org X11R6 support will
be available very soon.
Tested on hyperion and 4 different flavours of catseye/topcat/kathmandu; other,
older frame buffers (davinci, gatorbox and renaissance) could not been tested
but should work; please report any regression.
|
|
its out-of-DIO-range memory when sizing the external iomap.
|
|
|
|
|
|
rather than <include/cpu.h>.
|
|
|
|
|
|
no functional change.
|
|
and if so, resume to regular phase change processing (due to the way this
SCSI controller works, it is not possible to rely upon DMA completion
interrupts).
While there, bring back openings to 2 in all cases, and move timeout_del()
to spc_done().
Tested by deraadt@, millert@ and I.
|
|
- more delay in loops (1.36)
- on manual xfer via TEMP register, set PCTL_BFINT_ENAB and check
bus free by INTS register (1.35)
|
|
transferred correctly.
|
|
|
|
|
|
|
|
|
|
not busy wait for the transfer to complete at the very beginning, as
this interrupt could be unrelated to the transfer; busywaiting could
cause the scsi command timeout to be triggered, and the code would not
recover correctly.
Instead, make the DMA check simply return, leaving the "busy" flag set;
and perform the usual interrupt handler duties, except for the phase
change.
test&ok millert@
|
|
DMA.
|
|
completes; from the old driver, forgotten during the replacement here.
|
|
|
|
and do not enable parity in the scsi chip in this case.
From the old scsi code.
Tested by millert@ and I, though none of us has non-parity capable 98658.
|
|
Tested by millert@ and I.
|
|
|
