--
Add a workaround for VIA EHCI controllers which, under load, signal qTD
completion before they have performed writeback from the overlay qTD.
This condition would exhibit itself as a umass stall that never recovers.
--
This fixes the problem reported by Thomas Klausner on current-users@:
http://mail-index.netbsd.org/current-users/2006/01/17/0000.html
- initialize it properly in the bus front-ends (all 2, exept in wdc_pcmcia.c
for the "Sandisk CompactFlash Card" where it's set to 1)
- remplace hardcoded '2' by ata_channel->ch_ndrive in MI IDE drivers.
From Christos Zoulas in kern/32501.
a u_int64_t variable). Fixes 1 to 2 TB disks in qemu. Tested with
1200GB and to a lesser extent 2047GB sparse files. For reference
an empty FFS1 filesystem on a 1200GB sparse file uses 19GB of disk.
In this case, the transfer count becomes larger than the reality,
and this fix works around it.
- Add missing bus_dmamap_unload() in attach failure path.
XXX: Linux driver has a hack not to probe the slave drive. We just
create a phantom drive:
wdc0 at pcmcia0 function 0: < , Memory Card Adapter II, V1.00>
wdc0: i/o mapped mode
atabus2 at wdc0 channel 0
wd1 at atabus2 drive 0: <Memory Card Adapter II>
wd1: drive supports 1-sector PIO transfers, LBA addressing
wd1: 241 MB, 965 cyl, 16 head, 32 sec, 512 bytes/sect x 494080 sectors
wd2 at atabus2 drive 1: <>
wd2: drive supports 1-sector PIO transfers, chs addressing
wd2: 0, 0 cyl, 0 head, 0 sec, 512 bytes/sect x 0 sectors
wd2: 32-bit data port
wd2: disk label read failed
wd2: disk label read failed
"private" structure in struct buf that can be used to keep track of
the request associated with this buffer (the buffer used here is one
allocated from rf_CreateDiskQueueData(), so it's ours to do with what
we please). Shrinks code a little, reduces the run-time memory
footprint a bit, and simplifies both rf_DispatchKernelIO() and
KernelWakeupFunc().
Thanks to yamt for his "why is rf_DispatchKernelIO using another buf"
question which prompted me to revisit this code.
is called from pciide_attach() and at this point we don't know which drives
are here. Just assume all drives are there and allocate DMA ressources for
all of them.
Commit fixes problem reported (privately) by Jeff Rizzo.
XXX: Note that we are still are unable to allow a reconstruct of
a failed spare -- it's easy enough to attempt to do it, but the
code in rf_ContinueReconstructFailedDisk() isn't anywhere close to
ready to allow it. Yet another reason why the whole disk/spare thing
needs to be re-worked. *sigh*
