Nick Hibma [Sun, 7 May 2000 20:46:49 +0000 (20:46 +0000)]
If PERIPH_1284 is not defined, don't even bother calling the BUS_*_INTR
functions.
If it is defined, check whether bus_alloc_resource has succeeded. If it
hasn't, it is in polled mode. Mike Nowlin reports that this change makes
the geek port (whatever _that_ is :-) work again on his machine.
Warner Losh [Sun, 7 May 2000 18:20:40 +0000 (18:20 +0000)]
Don't claim devices that we fail to probe for. Reverse the sense of a test
so that we return failure when we've failed and succeed when we succeed.
Otherwise, users always get a sn device :-(.
Nick Hibma [Sun, 7 May 2000 18:04:50 +0000 (18:04 +0000)]
*sigh* I must have been on something that night. Make xpt_periph an
extern with the original in cam_xpt.c instead of replicating
xpt_periph in all the sources using it (and hence not initialising it)
Paul Saab [Sun, 7 May 2000 16:41:15 +0000 (16:41 +0000)]
Fix checksum calculations. This should fix the network problems
in current where all packets were returning with bad checksums.
(observed with netstat -s).
Gerard Roudier [Sun, 7 May 2000 09:54:33 +0000 (09:54 +0000)]
Work-around a couple of C1010 quirks:
- Reload SCNTL3 after selection from host (C1010-33).
- Reload SCNTL4 prior to any DATA OUT phase (C1010-66).
- Use max SCSI offset 31 for ST but 62 for DT.
Matt Jacob [Sun, 7 May 2000 04:56:19 +0000 (04:56 +0000)]
Update 'hose' (actually, MCPCIA instance #) for Alpha Rawhide systems.
We have *got* to fix this bogosity of trying to steal part of the PCI
address space for this stuff.
Matt Jacob [Sun, 7 May 2000 04:53:04 +0000 (04:53 +0000)]
Add in a first pass at Alpha 4100 (Rawhide) support. It doesn't quite
boot all the way yet, but it's darn close (blows up somewhere probing
the PS/2 mouse on the EISA bus).
Add a keymap that is the standard ISO PC-101/3 layout, but with the
key marked "Caps Lock" acting as a left control key. Many want a control
key in the Real place, but still want the keymap to match the printed keys
as much as possible.
Jonathan Lemon [Sat, 6 May 2000 15:53:59 +0000 (15:53 +0000)]
Bounce a copy of the mbuf to the bpf listener when we submit the frames
for transmit to the adapter, not when we receive a transmit interrupt
indicating that they were sent. This fix now allows tcpdump to produce
sane results by recording the timestamp at the point where the mbuf was
actually transmitted.
Alexey Zelkin [Sat, 6 May 2000 14:03:26 +0000 (14:03 +0000)]
. clear `.Os' macro value since this tool is not KAME-only anymore
. correctly use .Cm macro
. don't use duplicated arguments for .Nm macro
. use .Er macro for error names
. correctly declare paper reference in SEE ALSO section
. sort Xr's in SEE ALSO section
. add integration note
Alexey Zelkin [Sat, 6 May 2000 13:54:59 +0000 (13:54 +0000)]
. clear `.Os' macro value since this tool is not KAME specific anymore
. compact some macros (.Oo/.Oc -> .Op, .Po/.Pc -> .Pq)
. use .Tn
. avoid using duplicated arguments for .Nm macro
. correct KAME kit pathes (/usr/local/v6 -> /usr)
. add integration note
Alexey Zelkin [Sat, 6 May 2000 12:05:39 +0000 (12:05 +0000)]
mdoc related cleanup:
. use construction ".Aq Pa filename" instead of ".Pa <filename>"
. replace Section Heading macro (.Sh) with Subsection (.Ss) macro for
subsections
Mike Smith [Sat, 6 May 2000 08:54:33 +0000 (08:54 +0000)]
Change the way that scatter/gather list tables are allocated so that we
can use all of the s/g entries available on smaller cards. This is
necessary if we want to be able to handle a non-page-aligned 64k transfer
on 2.x and 3.x firmware.
Fix a missing splx() that may have left us at splbio() for longer than
desired.
Reduce shadowing of controller-supplied parameters a little.
Jonathan Lemon [Sat, 6 May 2000 03:31:09 +0000 (03:31 +0000)]
Implement TCP NewReno, as documented in RFC 2582. This allows
better recovery for multiple packet losses in a single window.
The algorithm can be toggled via the sysctl net.inet.tcp.newreno,
which defaults to "on".
Brian Feldman [Sat, 6 May 2000 01:39:45 +0000 (01:39 +0000)]
Give the "streams" modulea version (1) and depend on it from the
"svr4elf" module. This unbreaks the SVR4 KLD (which had an undefined
function because of thenewly-committed KLD enhancements).
John Baldwin [Fri, 5 May 2000 13:22:10 +0000 (13:22 +0000)]
Switch to using the .code16 as(1) directive and using 16-bit assembly
code instead of using 32-bit code and having to just "know" that it's
really 16-bit instructions when things run. This also allows the code
to use fewer macros and more actual assembly statements, which eases
maintenance. Unfortunately, due to as(1) brokenness, we still use m4
macros for all 16-bit addresses, and all short jumps (i.e., 8-bit
relative addresses in the jump instruction) must be wrapped in .code32
directives to avoid useless bloat by as(1). This also fixes a few
problems that were preventing boot0 from compiling with the latest
and greatest version of as(1).
Bill Paul [Fri, 5 May 2000 12:47:47 +0000 (12:47 +0000)]
Apply patch to rl_rxeof() to really enforce payload alignment in
the case where we receive a packet that wraps from the end of the
RX buffer back to the start. This fixes an unaligned access trap on
the alpha with NFS.