particularly older ones, don't do a SAVE DATA POINTER before disconnecting.
When the driver reconnects, it does an implicit RESTORE POINTERS, and
restores the data pointer from *before* the last transfer. If the driver
calculates the residual using the data pointer and doesn't account for
this, the residual will be wrong.
What we do is test for the SDEV_AUTOSAVE quirk when we see a disconnect
message, falling through to do a SAVE DATA POINTER if the quirk applies
to this target.
The workaround was suggested by Mycroft, who also provided the explanation
of the problem. Actual code was (mostly) ripped off from the 53C9x driver.
Anne Hutton <hutton@isi.edu>]:
- add support for Adaptec 155 PCI ATM cards (e.g. ANA-5940)
- add sc->is_adaptec to handle differences between cards.
- break out MID_MK_TXQ/MID_MK_RXQ seperate macros to handle
the new Adaptec format TXQ/RXQ.
- adjust en_dqneed to return 1 on ADP (since the Adaptec can
DMA anything in one DRQ/DTQ!)
- add hook for a bus specific reset function (adaptec has
a seperate reset register that needs to be hit when
resettting the midway).
- adjust DMA test to not worry about burst sizes on the
adaptec (since it handles it all for us!) and to handle
the new DTQ/DRQ format.
- add Adaptec DMA support to en_txlaunch() and en_service()
BUG FIXES:
- fixed receiver panic under heavy load ("lost mbuf in slot 0!").
when the reassembly buffer overflows, the T-bit is set in
the RDB and the data field is empty. en_service() sets up
a 4-byte (RDB size) dummy DMA without IF_ENQUEUE. but the
recv intr handling in en_intr() always does IF_DEQUEUE.
as a result, a successive recv intr loses its mbuf and
leads to a panic. the solution is to only IF_DEQUEUE if
the interrupt has non-zero length (indicating that there
is an mbuf to get). in order for this to work, EN_DQ_MK
must always be non-zero. we do this by or'ing in an unused
bit (0x80000).
reported by: Kenjiro Cho <kjc@csl.sony.co.jp>
- fix setting of transmit channel when txspeed[] is non-zero
(e.g. traffic shaping). the old scheme didn't work
properly (it allowed the same VCI to use multiple tx channels
thus defeating the txspeed[] parameter). the new scheme
statically assigns a VC to a channel when txspeed[] is set.
[note that the code to set txspeed[] isn't in the driver right
now since a MI interface to do this hasn't been made yet]
we add sc->txvc2slot[] and sc->txslot[n].nref for this.
reported by: Kenjiro Cho <kjc@csl.sony.co.jp>,
Milind M Buddihikot <milind@ccrc.wustl.edu>,
Dong Lin <dong@eecs.harvard.edu>
- when doing SRAM copies, be sure to round up the length to the next
largest word (otherwise the driver will try to do a byte clean
up DMA and then get an ID error interrupt).
MINOR CLEANUPS:
- clean up loops in DMA test
contributed by: Kenjiro Cho <kjc@csl.sony.co.jp>
- restructure and cleanup of en_read/en_write macros/inlines
- clean up some byte ordering stuff so that we are consistant throughout
the driver
correspond to the ANA numbers... certainly includes ANA-5940.
- add Efficient Nets product 0x0000. according to linux pci.h
the 0x0000 is an FPGA version of the midway card and the 0x0002
is the ASIC version.
into 8 bits generate (bogus) warnings on some architectures, but the
change to pad on "scsi inquiry" is no longer needed, apparently.
Thanks to Matthias Pfaller for pointing the latter out to me.
- Deal with devices that ignore the length specified in the
inquiry command.
- Allow asynchronous requests without using a buf (key off NOSLEEP instead).
Thanks to Michael Smith for reporting, Jason Thorpe for pointing to the
report, doing a quick workaround which pointed me to the right code part and
for testing the final fix.
does not supply media types, default to the "manual" type.
- Clean up the ETHER_CMP stuff slightly.
- Keep track of our carrier status, for media status requests.
Some of the stuff (e.g., rarpd, bootpd, dhcpd etc., libsa) still will
only support Ethernet. Tcpdump itself should be ok, but libpcap needs
lot of work.
For the detailed change history, look at the commit log entries for
the is-newarp branch.
the base address. This allows for a more flexible layout of buffers
in the Lance's memory.
- Add a new element 'sc_saved_csr0' to am7990_softc. It's value is or-ed
with the current csr0 value in the am7990_intr() function. This allowes
for a 'deferred' interrupt sceme.
- Implement a simple quirks framework for changers.
- Add a quirk for the Spectra 9000 8mm tape library; it requires a
slightly-longer-than-1-minute delay to take tape inventory. Many
thanks to David Webster at Cygnus Solutions for testing this for me.
to prepare for if_media, and to support EISA 3c59x cards:
* change epconfig() to take a short that encodes the chipset
type (3c509, or Demon/Vortex/Boomerang).
* add distinct 3c509 and Demon/Vortex/Boomerang media-sense
functions to back end.
* Add EISA match/attach support for the 3c592 and 3c597
(Demon) cards as well as the 3c509-comatibl EISA 3c509/3c579.
* Assume that ISA and ISA PnP cards are 3c509-style, until
proof to the contrary (e.g. large packet support).
Tested on 3c595, 3c590, and 3c509b. Not tested on 3c515 or Demon.
(RX_RESET, TX_RESET, GLOBAL_RESET).
Commands issued while a reset is in progress may be dropped. The driver
must wait until resets are complete; ep_complete_cmd() does so.
* Change RX_RESET, TX_RESET, GLOBAL_RESET to use ep_complete_cmd().
* Revert epconfig() to use inline RX_RESET, TX_RESET.
That avoids accidentally setting IFF_UP, which breaks both diskless boot,
and "halt" commands issued before the interface is brought up.
a "Smart Frame Buffer TURBOchannel Module" user's manual(?), and from
some "DECstation 5000 Model 240 Options and Upgrades" guides found
while cleaning out my office.
- Keep a freelist of component buffer headers, defaulting
to 8 headers per component (tunable with the CCDNBUF
kernel compile option).
- When allocating a component buffer, try to pull a header
off the freelist first, falling back on MALLOC() if the
freelist is empty. Use MALLOC() rather than malloc(),
because it will attempt to short-cut the allocation before
actually making a full-blown malloc() call.
- Keep statistics on how many component buffer headers have
been allocated and how many of those allocations have
resulted in freelist misses.
With these changes, I observed measurable decreases in system and
wall clock time on operations on an 8G ccd, as well as a measurable
drop in the time spent in ccdbuffer() (measured with kernel profiling).
isascan (__BROKEN_INDIRECT_CONFIG), and attempt multiple match/attach
of cloning devices in isasearch(). (If a match/attach succeeds for
a cloning device, try it again.)
- fix the parser so that non-compliant isapnp devices work, and we
give up when there are too many errors (some on-board pnp sound/atapi/game
soundblaster compatible card). reported by chuck cranor
- fix the parser so that configuration tags that come after the
end-dependencies tag or before the begin-dependencies tag are merged
in properly (gravis ultrasound). reported by gary duzan
We now build a tree of card -> logical-dev -> conf, and in the end we
convert it to a list of conf...
- if aal5 frame has a CRC error then the length field in the aal5 trailer
may not be valid, so we can not use it [and we must dump the frame]
[Yuhang Sun <sunyh@dworkin.wustl.edu> & chuck@ccrc.wustl.edu]
Dave Huang <khym@bga.com>
Tested on alpha by cgd, tested on several i386 boxes. Certainly causes
no harm to the goddamned mess, but the NCR driver only works when you
perform voodoo rituals on it anyway.
This is what Dave said (in email) has been added to the driver:
----------------------------------------------------------------------
This seems to be the most significant change:
General cleanup and new features for 53c875 based cards, especially the
Tekram DC390W/U/F, whose config EEPROM can now be dumped, if the kernel
is built with option NCR_TEKRAM_EEPROM.
Other changes:
- add brackets to expansion of OUTB/W/L macro arguments.
- remove unused NCB structure element ns_async
- support sync. SCSI offset of 16 (instead of only 8) on 825A and 875
- correctly identify 53c810A and 53c825A chips
- preserve SCSI BIOS settings of PCI performance options
- remove (already disabled) support for NCR reset because of command timeout
- reverse order of reading of SCSI and DMA specific interrupt cause registers
- add definition of Tekram config EEPROM contents (not currently used)
----------------------------------------------------------------------
Newer boards (e.g., Vortex) take time to reset, and wedge on some machines.
Polling on S_COMMAND_IN_PROGRESS would be better, but that doesn't
work with xx_RESET commands on 3c509s.
CONFIG_CNTRL bits by making the "conn" argument to epconfig()
a u_int, not a u_int16_t, and by defining 100mbit media
bits in the upper 16 bits.
A redesigned interface that fixes this properly is in the pipeline.
comments in pr port-i386/2358: "if_ep/3C590 no longer working"
* Only call Vortex media-config code on PCI-attached devices, until
we know if it works on ISA and EISA large-packet devices also.
* Fix typo in printf of MAC address
rather than unknown.
* Look for 100mbit interfaces, and for any present, set corresponding
bits in `conn' argument to epconfig(). epconfig() will now print
100Mbit media by name.
scottr