Switch to handling bad SCSI status as a sequencer interrupt
instead of having the kernel proccess these failures via
the completion queue. This is done because:
o The old scheme required us to pause the sequencer and clear
critical sections for each SCB. It seems that these pause
actions, if coincident with a sequencer FIFO interrupt, would
result in a FIFO interrupt getting lost or directing to the
wrong FIFO. This caused hangs when the driver was stressed
under high "queue full" loads.
o The completion code assumed that it was always called with
the sequencer running. This may not be the case in timeout
processing where completions occur manually via
ahd_pause_and_flushwork().
o With this scheme, the extra expense of clearing critical
sections is avoided since the sequencer will only self pause
once all pending selections have cleared and it is not in
a critical section.
aic79xx.c
Add code to handle the new BAD_SCB_STATUS sequencer
interrupt code. This just redirects the SCB through
the already existing ahd_complete_scb() code path.
Remove code in ahd_handle_scsi_status() that paused
the sequencer, made sure that no selections where
pending, and cleared critical sections. Bad
status SCBs are now only processed when all of these
conditions are true.
aic79xx.reg:
Add the BAD_SCB_STATUS sequencer interrupt code.
aic79xx.seq:
When completing an SCB upload to the host, if
we are doing this because the SCB contains non-zero
SCSI status, defer completing the SCB until there
are no pending selection events. When completing
these SCBs, use the new BAD_SCB_STATUS sequencer
interrupt. For all other uploaded SCBs (currently
only for underruns), the SCB is completed via the
normal done queue. Additionally, keep the SCB that
is currently being uploaded on the COMPLETE_DMA_SCB
list until the dma is completed, not just until the
DMA is started. This ensures that the DMA is restarted
properly should the host disable the DMA transfer for
some reason.
In our RevA workaround for Maxtor drives, guard against
the host pausing us while trying to pause I/O until the
first data-valid REQ by clearing the current snapshot
so that we can tell if the transfer has completed prior
to us noticing the REQINIT status.
In cfg4data_intr, shave off an instruction before getting
the data path running by adding an entrypoint to the
overrun handler to also increment the FIFO use count.
In the overrun handler, be sure to clear our LONGJMP
address in both exit paths.
Perform a few sequencer optimizations.
aic79xx.c:
Print the full path from the SCB when a packetized
status overrun occurs.
Remove references to LONGJMP_SCB which is being
removed from firmware usage.
Print the new SCB_FIFO_USE_COUNT field in the
per-SCB section of ahd_dump_card_state(). The
SCB_TAG field is now re-used by the sequencer,
so it no longer makes sense to reference this
field in the kernel driver.
aic79xx.h:
Re-arrange fields in the hardware SCB from largest
size type to smallest. This makes it easier to
move fields without changing field alignment.
The hardware scb tag field is now down near the
"spare" portion of the SCB to facilitate reuse
by the sequencer.
aic79xx.reg:
Remove LONGJMP_ADDR.
Rearrange SCB fields to match aic79xx.h.
Add SCB_FIFO_USE_COUNT as the first byte
of the SCB_TAG field.
aic79xx.seq:
Add a per-SCB "Fifos in use count" field and use
it to determine when it is safe (all data posted)
to deliver status back to the host. The old method
involved polling one or both FIFOs to verify that
the current task did not have pending data. This
makes running down the GSFIFO very cheap, so we
will empty the GSFIFO in one idle loop pass in
all cases.
Use this simplification of the completion process
to prune down the data FIFO teardown sequencer for
packetized transfers. Much more code is now shared
between the data residual and transfer complete cases.
Correct some issues in the packetized status handler.
It used to be possible to CLRCHN our FIFO before status
had fully transferred to the host. We also failed to
handle NONPACKREQ phases that could occur should a CRC
error occur during transmission of the status data packet.
Correct a few big endian issues:
aic79xx.c:
aic79xx_inline.h:
aic79xx_pci.c:
aic79xx_osm.c:
o Always get the SCB's tag via the SCB_GET_TAG acccessor
o Add missing use of byte swapping macros when touching
hscb fields.
o Don't double swap SEEPROM data when it is printed.
Correct a big-endian bug. We cannot assign a
o When assigning a 32bit LE variable to a 64bit LE
variable, we must be explict about how the words
of the 64bit LE variable are initialized. Cast to
(uint32_t*) to do this.
aic79xx.c:
In ahd_clear_critical_section(), hit CRLSCSIINT
after restoring the interrupt masks to avoid what
appears to be a glitch on SCSIINT. Any real SCSIINT
status will be persistent and will immidiately
reset SCSIINT. This clear should only get rid of
spurious SCSIINTs.
This glitch was the cause of the "Unexpected PKT busfree"
status that occurred under high queue full loads
Call ahd_fini_scbdata() after shutdown so that
any ahd_chip_init() routine that might access
SCB data will not access free'd memory.
Reset the bus on an IOERR since the chip doesn't
seem to reset to the new voltage level without
this.
Change offset calculation for scatter gather maps
so that the calculation is correct if an integral
multiple of sg lists does not fit in the allocation
size.
Adjust bus dma tag for data buffers based on 39BIT
addressing flag in our softc.
Use the QFREEZE count to simplify ahd_pause_and_flushworkd().
We can thus rely on the sequencer eventually clearing ENSELO.
In ahd_abort_scbs(), fix a bug that could potentially
corrupt sequencer state. The saved SCB was being
restored in the SCSI mode instead of the saved mode.
It turns out that the SCB did not need to be saved at all
as the scbptr is already restored by all subroutines
called during this function that modify that register.
aic79xx.c:
aic79xx.h:
aic79xx_pci.c:
Add support for parsing the seeprom vital product
data. The VPD data are currently unused.
aic79xx.h:
aic79xx.seq:
aic79xx_pci.c:
Add a firmware workaround to make the LED blink
brighter during packetized operations on the H2A.
aic79xx_inline.h:
The host does not use timer interrupts, so don't
gate our decision on whether or not to unpause
the sequencer on whether or not a timer interrupt
is pending.
aic7xxx.h:
Split out core chip initialization into ahc_chip_init().
This will allow us to reset the chip correctly at times
other than initial chip setup.
aic7770.c
aic7xxx_pci.c:
Flesh out bus chip init methods for our two
bus attachments and use these, in addition to
bus suspend/resume hooks to get the core in
better shape for handling these events.
When disabling PCI parity error checking, use FAILDIS.
Although the chip docs indicate that clearing PERRESPEN
should also work, it does not.
Auto-disable pci parity error checking after informing
the user of AHC_PCI_TARGET_PERR_THRESH number of parity
errors observed as a target.
aic7xxx.h:
aic7xxx_pci.c
aic7770.c
aic7xxx.c
Add the instruction_ram_size softc field.
Remove the now unused stack_size softc field.
Modify ahc_loadseq to return a failure code
and to actually check the downloaded instruction
count against the limit set in our softc.
Modify callers of ahc_loadseq to handle load
failures as appropriate.
Set instruction RAM sizes for each chip type.
aic7xxx_pci.c:
Add some delay in the aic785X termination
control code. This may fix problems with
the 2930.
Be consistent in how we access config space
registers. 16bit registers are accessed using
16bit ops.
aic7xxx.c:
Correct spelling errors.
Have ahc_force_renegotiation() take a devinfo as is done
in the U320 driver. Use this argument to correct a bug
in the selection timeout handler where we forced a renegotiation
with the last device that had set SAVED_SCSIID. SAVED_SCSIID
is only updated once a selection is *sucessfull* and so is
stale for any selection timeout.
Cleanup the setup of the devinfo for busfree events. We
now use this devinfo for a call to ahc_force_renegotiation()
at the bottom of the routine, so it must be initialized in
all cases.
In ahc_pause_and_flushwork(), adjust the loop so that it
will exit in the hot-eject case even if the INT_PEND mask
is something other than 0xFF (as it is in this driver).
Correct a wrapping string constant.
Call ahc_fini_scbdata() after shutdown so that
any ahc_chip_init() routine that might access
SCB data will not access free'd memory.
Correctly setup our buffer tag to indicate that 39bit
addressing is available if in 39bit addressing mode.
Rearrange some variable declarations based on
type size.
aic7xxx.c
aic7xxx.h:
aic7xxx.reg:
Consistently use MAX_OFFSET for the user max syncrate
set from non-volatile storage. This ensures that the
offset does not conflict with AHC_OFFSET_UNKNOWN.
Change AHC_OFFSET_UNKNOWN to 0xFF. This is
a value that the curr->offset can never be,
unlike '0' which we previously used. This
fixes code that only checks for a non-zero
offset to determine if a sync negotiation
is required since it will fire in the unknown
case even if the goal is async.
Change MAX_OFFSET to 0x7f which is the max
offset U160 aic7xxx controllers can negotiate.
This ensures that curr->offset will not
match AHC_OFFSET_UNKNOWN.
aic7xxx_inline.h:
Have our inline interrupt handler return with a value
indicating whether we serviced a real interrupt. This
is required for Linux support.
Return earlier if the interrupt is not for us.
patch workarounds for each phy revision.
Obtained from: NetBSD & Broadcom Linux driver
- Disable AUTOPOLL when accessing the PHY as it may cause PCI errors.
Obtained from: NetBSD
- Check the UPDATED bit in the status block so the driver knows
that the status block as indeed changed since the last access.
Broadcom documentation states drivers should unset the UPDATED/CHANGED
bits after reading them.
- When changing media types, first loop the phy then set the media.
Broadcom documentation and Linux drivers do this and I observed
much better handling of link after this change.
- Broadcom documentation states that for 1000BaseT operation,
autonegotiation must be enabled. Fix hard coding of media so that
the driver only advertises 1000BaseT as the supported media type
and enable autonegotition.
- Only set Master/Slave on the 5701.
Obtained from Broadcom Linux driver.
- Use htole* macros where appropriate so that the driver could work on non-x86 architectures
- Use m_getcl() instead of MGETHDR/MCLGET macros
Submitted by: sam (Sam Leffler)
Note: this might print failure messages on some systems, unfortunatly
the info from the device, stating if flushing is supported, cannot be trusted
so the operation is always issued on all devices, just in case...
just return it. Don't try to reinitialize it. This should fix a
number of inconsistencies that some people encountered with "vinum
start".
PR: 30588
PR: 43475
of <machine/pc/bios.h> specific to i386 and added a conditional define
for BIOS_PADDRTOVADDR that depends on ISA_HOLE_START. The latter is
undefined on alpha and ia64. Since the former is defined the same on
both alpha and ia64, assume the ISA_HOLE_START dependent definition
is specific to amd64 and use the identity-mapping in all other cases.
This of course is getting uglier every day...
project by providing documentation (under NDA) and hardware for
testing. This commit is the first result of the cooperation, and
adds support for several of their new controllers that we didn't
support before (and probably newer would have without this arrangement).
Add support for the Promise SATA150 TX2/TX4 and the Promise TX4000
controllers. This also adds support for various motherboard fitted
Promise SATA/ATA chips.
Note that this code uses memory mapped registers to minimize overhead.
I belive FreeBSD has made another first in the Open Source world
by being able to release support for this :)
object: subdisks, plexes and volumes. The encoding for plexes and
subdisks no longer reflects the object to which they belong. The
super devices are high-order volume numbers. This gives vastly more
potential volumes (4 million instead of 256).
Rewrite minor number decoding. Now we have only three types of
object: subdisks, plexes and volumes. The encoding for plexes and
subdisks no longer reflects the object to which they belong. The
super devices are high-order volume numbers. This gives vastly more
potential volumes (4 million instead of 256).
object: subdisks, plexes and volumes. The encoding for plexes and
subdisks no longer reflects the object to which they belong. The
super devices are high-order volume numbers. This gives vastly more
potential volumes (4 million instead of 256).
Correct formats for some error messages. Don't cast the value to
match the format.
object: subdisks, plexes and volumes. The encoding for plexes and
subdisks no longer reflects the object to which they belong. The
super devices are high-order volume numbers. This gives vastly more
potential volumes (4 million instead of 256).
Tidy up comments.
Check for null rqgs. This continue to be reported, though I can't
work out why.
Correct formats for some error messages. Don't cast the value to
match the format.
Use microtime, not getmicrotime, for timing debug entries.
object: subdisks, plexes and volumes. The encoding for plexes and
subdisks no longer reflects the object to which they belong. The
super devices are high-order volume numbers. This gives vastly more
potential volumes (4 million instead of 256).
As a result of the minor number changes, split out the superdevice
handling into a separate function, vinum_super_ioctl. This was most
of the code of vinumioctl.
attachobject: Improve error checking.
init_drive: Rephrase error message text.
Remove dead code (inside #if 0).
Change name of find_drive_by_dev to the more descriptive
find_drive_by_name.
Tidy up comments.
get_emppty_drive: Fix a day one bug with strcpy parameters.
Change name of find_drive_by_dev to the more descriptive
find_drive_by_name.
Rewrite minor number decoding. Now we have only three types of
object: subdisks, plexes and volumes. The encoding for plexes and
subdisks no longer reflects the object to which they belong. The
super devices are high-order volume numbers. This gives vastly more
potential volumes (4 million instead of 256).
object: subdisks, plexes and volumes. The encoding for plexes and
subdisks no longer reflects the object to which they belong. The
super devices are high-order volume numbers. This gives vastly more
potential volumes (4 million instead of 256).
Remove an unnecessary goto.
vinumopen: Return EINVAL, not ENXIO, on an attempt to open a
referenced plex.
at all (ie reads yield constant values). Display the width as the
difference between max and min so that constant timers have width
zero.
o Get the address of the timer from the XPmTmrBlk field instead of
the V1_PmTmrBlk field. The former is a generic address and can
specify a memory mapped I/O address. Remove <machine/bus_pio.h>
to account for this. The timer is now properly configured on
machines with ACPI v2 tables, whether PIO or MEMIO. Note that
the acpica code converts v1 tables into v2 tables so the address
is always present in XPmTmrBlk.
o Replace the TIMER_READ macro with a call to the read_counter()
function and add a barrier to make sure that we observe proper
ordering of the reads.
Check for suspend before the device polling, rather than after it.
Check to see if the current thread owns the lock in ioctl and return
EBUSY if it does.
This advances the locking to the point that I can eject my fxp card 10
times in a row, but I agree with Jeff Hsu that we need to get the
network layer locking finished before chasing more of the races here
(actually, he doesn't think this set is worth it even). There's a
number of races between FXP_LOCK in detach and all other users of
FXP_LOCK, and this gets back to the 'device with sleepers being
forcibly detached' problem as well...
* AcpiOsDerivePciId(): finds a bus number, given the slot/func and the
acpi parse tree.
* AcpiOsPredefinedOverride(): use the sysctl hw.acpi.os_name to
override the value for _OS.
Ideas from: takawata, jhb
Reviewed by: takawata, marcel
Tested on: i386, ia64
(currently) only consumer (en).
Add a sysctl node hw.atm where the atm drivers will hook on their hardware
sysctl sub-trees.
Make atm_ifattach call if_attach and remove the corresponding call to if_attach
from en. Create atm_ifdetach and use that in en.
While the last change actually changes the interface this is not a problem in
practice because the only other consumer of this API is an older LANAI driver
on the net, that is not ready for current anyway.
Reviewed by: -atm
ia64 by defining them in terms of newbus. Add a static inline for
fillw(), which doesn't have anything to do with I/O.
It's still ugly, but now the ugliness can be removed from ia64
specific headers.
1) always call fxp_stop in fxp_detach. Since we don't read from
the card, there's no need to carefully look at things with
bus_child_present.
2) Call FXP_UNLOCK() before calling bus_teardown_intr to avoid
a possible deadlock reported by jhb.
3) add gone to the softc. Set it to true in detach.
4) Return immediately if gone is true in fxp_ioctl
5) Return immediately if gone is true in fxp_intr
that one cannot generally hold a lock and call bus_teardown_intr.
This is race free with wi_intr because bus_teardown_intr won't allow
wi_intr to be called after it returns.
# jeff hsu points out that there might be a race between this unlock
# and wi_start. While that may be true also, it won't impact this commit.
Submitted by: jhb
as 64-bit architectures won't like this. Use virtual array indexes
instead. This *should* allow the driver to work on 64-bit platforms,
though it's still not endian clean.
firmware 1.50.12, but 2.20.1 and 3.10.4 work. The 1.50.12 card gets
past doing dhclient, but hangs on transmit a little after the ip
address is set. The 1.50.12 card has always been 'cranky' and Bill
Paul's tearing it apart at FreeBSD '99 hasn't helped.
sc_reset and sc_enable are subtlely different things. sc_reset means
exactly "WI_CMD_INI has happened." sc_enabled means "WI_CMD_ENABLE
has been sent to the card without a WI_CMD_DISABLE following." This
is a little different than what they mean on NetBSD (where both of
these concepts are comingled). NetBSD will try to only enable symbol
cards once, while FreeBSD only sends the WI_CMD_INI once.
Also, only try once to reset the card on a symbol.
This makes the lucent cards no worse than before, but apparently not
much better either. I got fewer hangs in my testing than I have in
the past, but I don't know if it is statistically significant or not.
"Ian gave me this patch" I assume this is iadowse@.
Also, pull in a few devices NetBSD's table. More are there, but I need
to properly move them to pccarddev before including them here.
- Remove getcsrdata().
- Don't print device type, this can be obtained by fwcontrol.
Tested with: Maxtor 5000XT
Tested by: Daniel O'Connor <doconnor@gsoft.com.au>
following changes have been done:
- stylify. The original code was too hard to read.
- get rid of a number of compilation options (Adaptec-only, Eni-only, no-DMA).
- more debugging features.
- locking. This is not correct yet in the absence of interface layer locking,
but is correct enough to not to cause lock order reversals.
- remove RAW mode. There are no users of this in the tree and I doubt that
there are any.
- remove NetBSD compatibility code. There was no way to keep NetBSD non-busdma
and FreeBSD busdma code together.
- if_en now buildable as a module.
This has been actively tested on sparc64 and i386 with ENI server and
client cards and an Adaptec card (thanks to kjc).
Reviewed by: mdodd, arr
- Add fxp_start_body() and change fxp_start() to just acquire locks and
then call fxp_start_body(). Places that would call fxp_start() with
locks held (mutex recursion) now call fxp_start_body() directly.
Remove MTX_RECURSE flag from sc_mtx. [gallatin]
- Change fxp_attach() to work without the softc lock, saving interrupt
hooking until the head of fxp_attach().
- Call ether_ifattach() before overriding ifp parameters. This reverts
part of 1.155.
- Remove multiple error paths in fxp_attach().
- Teardown interrupt in fxp_detach() before unlocking the softc.
- Make sure mutex is not held in fxp_release()
- Delete the miibus instance and/or self in fxp_release(), not in
fxp_detach(). This can happen if attach fails partway through.
- Move ifmedia_removeall to fxp_release() since attach may fail after
media have been allocated.
- Add locking to fxp_suspend, fxp_resume, fxp_start, fxp_intr,
fxp_poll, fxp_tick, fxp_ioctl, fxp_watchdog.
- Pass in ifp to fxp_intr_body since its callers sometimes already use
it.
- Add compatibility define for INTR_MPSAFE for 4.x. [gallatin]
- You don't need to bzero softc.
Ideas from: gallatin, mux
Tested by: >400M packets of dd/ssh, NFS, ping on i386 UP
driver at long last!
Many thanks to vaidas.damosevicius@if.lt for keeping this issue alive
and pursuing Intel for a fix, Intel/ICP for working on the driver, and
Sergey Osokin for bringing the original patches up to 5-CURRENT.
of the infrastructure for the gamma driver which was removed a while back.
The DRM_LINUX option is removed because the handler is now provided by the
linux compat code itself.
(with other names) in the USB driver sources, but I felt that pcireg.h
should have a complete list - at least of classes and interfaces that we
know about and use.
or sched_lock are sufficient to test this flag.
XXX: vinum should really be using a kernel process via kthread_create()
instead of this hack. I'm not even sure PS_INMEM can be clear at this
point anyways.
if attach succeeded. device_is_alive just tells us that probe
succeeded. Since we were using it to do things like detach net
interfaces, this caused problems when there were errors in the attach
routine.
Symptoms of problem reported by: martin blapp
ethernet controller. The driver has been tested with the LinkSys
USB200M adapter. I know for a fact that there are other devices out
there with this chip but don't have all the USB vendor/device IDs.
Note: I'm not sure if this will force the driver to end up in the
install kernel image or not. Special magic needs to be done to exclude
it to keep the boot floppies from bloating again, someone please
advise.
other initializations succeeded.
- Initialize the TX and RX rings in epic_attach() rather than in
epic_init() where we're not supposed to fail. Similarly, free
the TX and RX rings in epic_detach() rather than in epic_stop().
- Change epic_init() to be a void function now that it can't fail.
Also change its parameter to a void * so that we have a correct
prototype for if_init.
- Now that epic_init() has a correct prototype, don't cast the
function pointer when initializing if_init.
- Fix nearby style bugs.
- Don't initialize if_output, ether_ifattach() does this for us.
- Use pci_enable_busmaster() instead of using pci_read_config()
and pci_write_config() directly.
- Don't try to enable I/O, bus_alloc_resource() does this for us.
- Unconditionally call *_stop() if device is in the tree. This is to
prevent callouts from happening after the device is gone. Checks for
bus_child_present() should be added in the future to keep from touching
potentially non-existent hardware in *_detach(). Found by iedowse@.
- Always check for and free miibus children, even if the device is not in
the tree since some failure cases could have gotten here.
- Call ether_ifdetach() in the irq setup failure case
- ti(4), xl(4): move ifmedia_init() calls to the beginning of attach so
that ifmedia_removeall() can be unconditionally called on detach. There
is no way to detect whether ifmedia has been initialized without using
a separate variable (as tl(4) does).
- Add comments to indicate assumptions of code path
bytes_in_mbuf rather than MCLBYTES. Add the ethertnet header to the
front of the mbuf. Adjust bytes_in_mbuf inside the loop that reads
the packet out of the card.
codec during initialization. This code mirrors the reset code used on
the VIA82c686 and fixes a codec initialization failure (SoundMAX
AD1885) reported by Matthias Schuendehuette.
in dc_detach() instead of only calling it if the hardware is preset.
This is a workaround for page faults in softclock() after a `dc'
device was detached, caused by not disabling a timer before freeing
its memory. The bus_child_present() checks should probably be
re-added later, but only to avoid the hardware accesses and not the
other resource cleanups in dc_stop().
Approved by: njl
Many internal structure changes for the FireWire driver.
- Compute CRC in CROM parsing.
- Add support for configuration ROM build.
- Simplify dummy buffer handling.
- busdma conversion
- Use swi_taskqueue_giant for -current. Mark the interrupt routine as MPSAFE.
- AR buffer handling.
Don't reallocate AR buffer but just recycle it.
Don't malloc and copy per packet in fwohci_arcv().
Pass packet to fw_rcv() using iovec.
Application must prepare receiving buffer in advance.
- Change fw_bind API so that application should pre-allocate xfer structure.
- Add fw_xfer_unload() for recycling struct fw_xfer.
- Add post_busreset hook
- Remove unused 'sub' and 'act_type' in struct fw_xfer.
- Remove npacket from struct fw_bulkxfer.
- Don't call back handlers in fwochi_arcv() if the packet has
not drained in AT queue
- Make firewire works on big endian platform.
- Use native endian for packet header and remove unnecessary ntohX/htonX.
- Remove FWXFERQ_PACKET mode. We don't use it anymore.
- Remove unnecessary restriction of FWSTMAXCHUNK.
- Don't set root node for phy config packet if the root node is
not cycle master capable but set myself for root node.
We should be the root node after next bus reset.
Spotted by: Yoshihiro Tabira <tabira@scd.mei.co.jp>
- Improve self id handling
Tested on: i386, sparc64 and i386 with forced bounce buffer
command config register. At the present, this represents a nop
because these bits should have been set earlier in the process. In
the future, we'll only set these bits when the driver requests the
resource, not when the bus code detects the resource.
Reviewed by: mdodd
network layer (ether).
- Don't abuse module names to facilitate ifconfig module loading;
such abuse isn't really needed. (And if we do need type information
associated with a module then we should make it explicit and not
use hacks.)
