With these rom dumps, the ip6000 now boots. Something is wrong with floppy interrupts, and EDGE isn't good enough to get to the blue screen, but FDM works moderately well.
These changes make sun4 scsi work to some extent, and don't break interpro.
* ncr5390: dma support for disconnected state commands
* ncr5390: fix dma bug with tcounter=0 (meaning 65536)
* ncr5390: improve dma start and counter decrement
* sun4: fix dma control register writes
confusing people for far too long. Yes, this is a change in behaviour.
Add a valdiation check for ROMs with BIOS flag set that are
unselectable, fix the things it uncovers.
(nw) Fix other random stuff.
* Remove vram_rw and ctrl_rw wrapper functions
Signed-off-by: Sven Schnelle <svens@stackframe.org>
* topcat: replace hardcoded display width by m_fb_width
Signed-off-by: Sven Schnelle <svens@stackframe.org>
* topcat: prepare topcat for multi plane operation
framebuffer vram is only 1 bit wide.
Signed-off-by: Sven Schnelle <svens@stackframe.org>
* 98544: set single topcat plane
Signed-off-by: Sven Schnelle <svens@stackframe.org>
* add nereid mapper used on HP98543/98545 graphic cards
Signed-off-by: Sven Schnelle <svens@stackframe.org>
* prepare for multi-topcat operation
Signed-off-by: Sven Schnelle <svens@stackframe.org>
* add HP 98543 4 plane graphics card
* hp9k_3xx: add hp98543 card
* topcat,nerid: disable VERBOSE
Signed-off-by: Sven Schnelle <svens@stackframe.org>
* hp98543: use required_device_array
* 98543: use required_region_ptr
* 98544: use required_region_ptr
* topcat: add save state support
Signed-off-by: Sven Schnelle <svens@stackframe.org>
* nereid: add save state support
* hp98544,hp98543,topcat: use shared memory space
Signed-off-by: Sven Schnelle <svens@stackframe.org>
* nereid: remove redundant variable
Signed-off-by: Sven Schnelle <svens@stackframe.org>
* topcat, nereid: remove trailing underscore from include guard
Signed-off-by: Sven Schnelle <svens@stackframe.org>
* topcat,hp98543: code style fixes proposed in PR
Signed-off-by: Sven Schnelle <svens@stackframe.org>
These changes work with InterPro, but haven't been tested on other machines. Despite this, I'm reasonably confident they are all correct according to the observed behaviour and the documentation.
* don't wait for REQ after initiator complete with NACK
* wait until fifo empty during dma out before command complete
* make sure drq is always cleared on bus/function complete
* reset happens immediately
* always clear TC0 when counter reloaded
* check valid commands separately for 5390 and 5390a
* handle 5390a initiator set attention command
* use logmacro for logging
