Commit Graph

35 Commits

Author SHA1 Message Date
wiz d88f7b1d62 boundary, not boundry. Inspired by Tom Cosgrove. 2003-11-02 10:31:06 +00:00
scw 0dcbc9ae84 Try to drain the Tx pending queues after dealing with Rx/Tx interrupts.
This fixes a bug whereby a fast cpu with a decent cache can easily
outstrip the GT's ability to put packets on the wire, resulting in a
permanent backlog of mbufs in the Tx pending queues under heavy Tx load.

The bug was masked if the packet buffer was mapped non-cached, which
slowed down the cpu to where it couldn't keep up with the GT at 100mbit.
2003-08-05 14:55:06 +00:00
lukem 365cbd9428 add missing __KERNEL_RCSID() 2003-07-14 15:47:00 +00:00
scw 588cfa226b Add basic support for non-PCI host mode. 2003-06-12 19:18:49 +00:00
scw b8ce8543eb Always pass BUS_DMA_COHERENT to bus_dmamem_map() when allocating
descriptor/buffer memory.
2003-06-12 19:18:02 +00:00
scw b751df8807 Tickle the watchdog while doing polled console i/o.
This is especially important if we spend any time in ddb. :)
2003-06-12 19:16:18 +00:00
scw 269144aef4 Disable the watchdog during autoconf of sub-devices.
Some drivers (e.g. pciide(4)) can take long enough to probe that the
watchdog barks before they're done.
2003-06-12 19:14:37 +00:00
scw 6fab97c614 Add an option to control whether the GT's internal PCI Arbiter should
be enabled or not.
2003-05-27 11:39:50 +00:00
scw d1fb17639e defflag GT_ECC. 2003-05-17 08:27:58 +00:00
scw 84d2659d16 In gtmpscinit_stop(), don't wait for abort completion status if the
receiver is disabled. It'll never complete in this case.
2003-05-01 12:13:26 +00:00
matt de36c1091b Update inptr *after* we do the dmasync. 2003-04-30 20:49:49 +00:00
matt 0259e9541c Fix overrun of TX buffer dmamap. 2003-04-30 18:31:30 +00:00
scw dd3614f13a When attaching, make sure the "Reject Broadcast Mode", "Promiscuous Mode"
and "Pass Bad Frames" bits in the EPCR register are clear.

This works around a problem where firmware sets RBM for some reason before
handing control to the NetBSD kernel.
2003-04-10 15:23:19 +00:00
thorpej cf9d21bf8e Use PAGE_SIZE rather than NBPG. 2003-04-08 23:39:14 +00:00
thorpej 106ae5d65e Use PAGE_SIZE rather than NBPG. 2003-04-08 22:33:33 +00:00
matt 46acdfd875 Deal with discovery strangeness/timing issue. Sometimes the Ownership bit
on descriptor doesn't get reset so look beyond to the next descriptor to
see if it's been returned.  If it has, then just process the current one.
2003-04-08 19:37:17 +00:00
matt 6193e8a4a1 Remove #define DEBUG. (this was working right. The problem I encountered
was due to a bug in PPCBOOT.  It was not ordering DIMM space largest to
smallest so that you could badly alignment memory).
2003-04-04 20:09:07 +00:00
matt d914d04522 Cleanup PCI support. 2003-04-01 19:11:44 +00:00
matt 8909c240f2 Add GT_PCI{0,1}_{IO,MEM}{SIZE,START} options 2003-03-27 07:21:30 +00:00
matt caff52554f Only print IRQ is it's not the default value. 2003-03-27 07:20:48 +00:00
matt 0e5392f73f Make reset delay depend on the baud rate. 2003-03-24 20:03:23 +00:00
matt bd800e6684 Fix tyops. Add 2 defines. 2003-03-24 17:03:18 +00:00
matt a69a0868cc Make gtmpsc configurable via config file. (baud rate source, frequency). 2003-03-24 17:02:14 +00:00
matt a09a7e1927 Update bpf interface to current. Make output buffers aligned to a
cacheline bounary.  Only put hash table and descriptors in uncached
memory, buffers are in cached memory.
2003-03-24 17:00:54 +00:00
matt 0e4975df24 Split PCI initialization out of attach hook. 2003-03-18 19:32:46 +00:00
matt 923175a28a remove some unneeded code. 2003-03-18 15:09:28 +00:00
matt ee2ba43d39 Enable I/O & memory access too. 2003-03-18 15:00:00 +00:00
matt 8fb546ff67 Do minimal initialization of PCI. 2003-03-18 05:51:51 +00:00
matt 051e31601c Derive baud rate from cpu_timebase on powerpc. 2003-03-18 05:50:02 +00:00
matt 631dbeba1a Use generic softintr. switch from consinit to explicit cnattach
framework.  Use bus_space for access register access.
2003-03-17 16:42:47 +00:00
matt f27033251e Use a subregion for the mac's ethernet registers. Shaves 512 bytes
from the driver.
2003-03-17 16:41:15 +00:00
matt 08ea2a499c Cleanup. Move much of the code to bus_space instead of using
private mechanisms.
2003-03-16 07:05:33 +00:00
matt 7f87f176b2 Add obio attachment for things like rtcs, com, and flash. 2003-03-16 07:03:30 +00:00
matt 3dc865b386 Relocate Discovery first thing in initppc (if not at expected address).
This implies that if PMON tells us where the GT is, we write that address
in gtbase.
2003-03-06 06:04:21 +00:00
matt ff2281b498 Add preliminary support Marvell (Galileo) Discovery System Controllers.
This code was contributed by Allegro Networks.
2003-03-05 22:08:18 +00:00