Ask the target to grant 4 seconds instead of the standard and minimum of
1 second window after bus reset for reconnection. This accelerates
reconnection if there are more than one targets on the bus: If a login
and inquiry to one target blocks the fw-sbp2 workqueue for more than 1s
after bus reset, we now still can reconnect to the other target.
Before that, fw-sbp2's reconnect attempts would be rejected with "error
status: 0:9" (function rejected), and fw-sbp2 would finally re-login.
All those futile reconnect attemps cost extra time until the target
which needs re-login is ready for I/O again.
The reconnect timeout field in the login ORB doesn't have to be honored
by the target though. I found that we could get up to
- allegedly 32768s from an old OXFW911 firmware
- 256s from LSI bridges
- 4s from OXUF922 and OXFW912 bridges,
- 2s from TI bridges,
- only the standard 1s from Initio and Prolific bridges and from
Apple OpenFirmware in target mode.
We just try to get 4 seconds which already covers the case of a few
HDDs on the same bus quite nicely.
A minor drawback occurs in the following (rare and impractical) border
case:
- two initiators are there, initiator 1 holds an exclusive login to
a target,
- initiator 1 goes off the bus,
- target refuses login attempts from initiator 2 until reconnect_hold
seconds after bus reset.
An alternative approach to the issue at hand would be to parallelize
fw-sbp2's reconnect and login work.
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
Acked-by: Jarod Wilson <jwilson@redhat.com>
Don't attempt to send a logout ORB if the target was already unplugged
or had its link switched off. If two targets are attached, this
enhances the chance to quickly reconnect to the remaining target when
one target is plugged out.
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
Acked-by: Jarod Wilson <jwilson@redhat.com>
SBP2_MAX_SECTORS is nowhere used in fw-sbp2.
It merely got copied over from sbp2 where it played a role in the past.
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
This patch relaxes the default SCSI DMA alignment from 512 bytes to 4
bytes. I remember from previous discussions that usb and firewire have
sector size alignment requirements, so I upped their alignments in the
respective slave allocs.
The reason for doing this is so that we don't get such a huge amount of
copy overhead in bio_copy_user() for udev. (basically all inquiries it
issues can now be directly mapped).
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
Since patch "fw-sbp2: use an own workqueue (fix system responsiveness)"
increased parallelism between fw-sbp2 and fw-core, it was possible that
fw-sbp2 didn't release the SCSI device when the FireWire device was
disconnected.
This happened if sbp2_update() ran during sbp2_login(), because a bus
reset occurred during sbp2_login(). The sbp2_login() work would [try
to] reschedule itself because it failed due to the bus reset, and it
would _not_ drop its reference on the target. However, sbp2_update()
would schedule sbp2_login() too before sbp2_login() rescheduled itself
and hence sbp2_update() would take an additional reference. And then
we would have one reference too many.
The fix is to _always_ drop the reference when leaving the sbp2_login()
work. If the sbp2_login() work reschedules itself, it takes a
reference, but only if it wasn't already rescheduled by sbp2_update().
Ditto in the sbp2_reconnect() work.
The resulting code is actually simpler than before: We _always_ take
a reference when successfully scheduling work. And we _always_ drop
a reference when leaving a workqueue job. No exceptions.
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
Firewire-sbp2 did very uncooperative things in the kernel's shared
workqueue: Sleeping until reception of management status from the
target for up to 2 seconds, and performing SCSI inquiry and all of the
setup of SCSI command set drivers via scsi_add_device. If there were
transient or permanent error conditions, this caused long blockage of
the kernel's events process, noticeable e.g. by blocked keyboard input.
We now allocate a workqueue process exclusive to fw-sbp2. As a side
effect, this also increases parallelism of fw-sbp2's login and reconnect
work versus fw-core's device discovery and device update work which is
performed in the shared workqueue.
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
Signed-off-by: Kristian Høgsberg <krh@redhat.com>
On rare occasions, the ability to set one of the workaround flags at
runtime may save the day.
People who experience I/O errors with firewire-sbp2 while the old sbp2
driver worked for them should try workarounds=1 and report to the devel
mailinglist whether that improves things. Firewire-sbp2 defaults to the
SCSI stack's maximum transfer size per command, while sbp2 limits them
to 128 kBytes. Flag 1 accomplishes just that.
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
On IOMMU-less noncoherent architectures, orb->callback will memcpy the
whole SCSI command buffer for READ-like SCSI commands. It is therefore
friendlier to enable IRQs before the call, like before patch "Add
ref-counting for sbp2 orbs".
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
Acked-by: Kristian Høgsberg <krh@redhat.com>
This handles the case where we get the status write before getting the
complete_transaction callback ("status write for unknown orb"). In
this case, we just assume that the initial orb pointer transaction
succeeded and finish the orb. To prevent the transaction callback
from touching freed memory, we ref-count the orb structures.
Signed-off-by: Kristian Høgsberg <krh@redhat.com>
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
As far as I know, all CardBus FireWire 400 adapters have a maximum
payload of 1024 bytes which is less than the speed-dependent limit of
2048 bytes. Fw-sbp2 has to take the host adapter's limit into account.
This apparently fixes Juju's incompatibility with my CardBus cards, a
NEC based card and a VIA based card.
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
Acked-by: Kristian Høgsberg <krh@redhat.com>
Signed-off-by: Kristian Høgsberg <krh@redhat.com>
collapsed with fw-sbp2 patch "Drop cast to non-const char * in host
template initialization." from Kristian Høgsberg
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
The CPU must not touch the buffer after it was DMA-mapped.
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
Signed-off-by: Kristian Høgsberg <krh@redhat.com>
The CPU must not touch the buffer after it was DMA-mapped.
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
Signed-off-by: Kristian Høgsberg <krh@redhat.com>
- The CPU must not touch the buffer after it was DMA-mapped.
- The size argument of dma_unmap_single(...page_table...) was bogus.
- Move a comment closer to the code to which it refers to.
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
Signed-off-by: Kristian Høgsberg <krh@redhat.com>
Add rudimentary check for the case that the page table overflows due to
merging of s/g elements by the IOMMU. This would have lead to
overwriting of arbitrary memory.
After this change I expect that an offending command will be
unsuccessfully retried until the scsi_device is taken offline by SCSI
core. It's a border case and not worth to implement a recovery
strategy.
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
Acked-by: Kristian Høgsberg <krh@redhat.com>
Replace a cast with a container_of(). As long as nobody reorders the
structure elements, they do the same thing, but container_of() is more
readable.
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de> (added complete_command_orb)
Acked-by: Kristian Høgsberg <krh@redhat.com>
This affects of course only the "soft shutdown" case, e.g. "modprobe -r
firewire-sbp2", while it doesn't matter for hot unplug.
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
Signed-off-by: Kristian Høgsberg <krh@redhat.com>
This currently only affects one bridge in the hardwired blacklist.
I don't own one of those, hence haven't tested it.
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
Use a speed probe to determine the speed over 1394b buses and of nodes
which report a link speed less than their PHY speed.
Log the effective maximum speed of newly created nodes in dmesg.
Also, read the config ROM (except bus info block) at the maximum speed
rather than S100. This isn't a real optimization though because we
still only use quadlet read requests for the entire ROM.
The patch also adds support for S1600 and S3200, although such hardware
does not exist yet.
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
Signed-off-by: Kristian Høgsberg <krh@redhat.com>
cleanup after support of single-buffer requests was dropped
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
Signed-off-by: Kristian Hoegsberg <krh@redhat.com>
The attribute /sys/bus/scsi/devices/*:*:*:*/ieee1394_id, as generated by
the old sbp2 driver, is typically used to create persistently named
links in /dev/disk/by-id.
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
Signed-off-by: Kristian Høgsberg <krh@redhat.com>
The SCSI layer only passes sg requests down, so drop the
use_sg == 0, request_bufflen != 0 case.
Signed-off-by: Kristian Hoegsberg <krh@redhat.com>
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
Avoids an extra allocation and simplifies lifetime rules for the scsi_host.
Signed-off-by: Kristian Hoegsberg <krh@redhat.com>
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
This patch loads fw-sbp2 if sbp2 is still in the config file. So one can
go back and forth between releases without worry about the root
filesystem drivers.
Signed-off-by: Kristian Hoegsberg <krh@redhat.com>
Existing mkinitrd scripts still have to be adapted, unless they grok
module aliases.
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
Drop filenames from file preamble, drop editor annotations and
use standard indent style for block comments.
Signed-off-by: Kristian Hoegsberg <krh@redhat.com>
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de> (fixed typo)
Signed-off-by: Kristian Høgsberg <krh@redhat.com>
(Reverted part which moved it from eh_abort_handler to eh_host_reset_handler)
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
In some situations we can receive the ORB status write before we
have received the ORB pointer write response. When this happens,
we assume that the fw_transaction is finished and free the ORB
struct containing the fw_transaction.
This fix make the status write logic only accept status writes
for ORBs where the initial ORB pointer write transaction finished.
Signed-off-by: Kristian Høgsberg <krh@redhat.com>
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
Or the SAM status codes from the device sense data into the
command error code.
Signed-off-by: Kristian Høgsberg <krh@redhat.com>
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
The mod_timer based timing out of orb was a little to agressive and
would time out legit, but long-lived scsi cmds. Besides, the scsi
stack keeps track of this already. Since we're only timing out
management orbs, go back to wait_for_completion_timeout.
Signed-off-by: Kristian Høgsberg <krh@redhat.com>
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
Use PREPARE_DELAYED_WORK to just change the function pointer.
Signed-off-by: Kristian Høgsberg <krh@redhat.com>
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
Stefan Richter
James Bottomley
Kristian Høgsberg
Linus Torvalds
Jay Fenlason
Olaf Hering