This is a mix of a grand clean up I did and a reworking of the hotplug
support by Christoph. In testing it works as well as the previous code
did (that is -quite badly-) but we can now read it and as it settles in
actually debug the long standing problems it shares with the old driver.
Signed-off-by: Alan Cox <alan@redhat.com>
From Christoph:
removal of the BSD license is also fine with me, but should really
get a mention in the changelog aswell.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
Back in the beginning of last year we disabled mode page 8 and mode page
3f requests through device quirk bits instead of enhancing the driver to
respond to these mode pages because there was no apparent added value.
The Firmware that supports the new communication commands supports the
ability to force a write around of the adapter cache on a command by
command basis. In the attached patch we enable mode page 8 and 3f and
spoof the results as needed in order to *convince* the layers above to
submit writes with the FUA (Force Unit Attention) bit set if the file
system or application requires it, if the Firmware supports the write
through, or instead to submit a SYNCHRONIZE_CACHE if the Firmware does
not. The added value here is for file systems that benefit from this
functionality and for clustering or redundancy scenarios.
Caveats: By convince, we are responding with a minimal short 3 byte
content mode page 8, with only the data the SCSI layer needs and that we
can fill confidently. Applications that require the customarily larger
mode page 8 results may be confused by this(?). The FUA, or the
SYNCHRONIZE_CACHE only affect the cache on the controller. Our firmware
by default ensure that the underlying physical drives of the array have
their cache turned off so normally this is not a problem.
This attached patch is against current scsi-misc-2.6 and was unit tested
on RHEL5. Since this is a feature enhancement, it should not be
considered for any current stabilization efforts.
Signed-off-by: Mark Salyzyn <aacraid@adaptec.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
Use wait_for_completion_timeout instead a semaphore + timer.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
With libata converted to use sdev->manage_start_stop for suspend and
resume, sht->suspend/resume() has no user left and low level
suspend/ressume should be taken care of by low level driver's
suspend/resume callbacks (e.g. PCI or PCMCIA driver callbacks). This
patch removes sht->suspend/resume() callbacks.
This change is suggested by Christoph Hellwig.
Signed-off-by: Tejun Heo <htejun@gmail.com>
Cc: Christoph Hellwig <hch@infradead.org>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
This is basically a straight conversion. I have one of these things, so
I know it works ... my problem is that it has a wierd SCA like
connector, so I can't connect anything to it (no cables).
However, previously it panic'd in the interrupt, now it completes a bus
scan.
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
This patch provides support for FC virtual ports based on NPIV.
For information on the interfaces and design, please read the
Documentation/scsi/scsi_fc_transport.txt file enclosed within
the patch.
The RFC was originally posted here:
http://marc.info/?l=linux-scsi&m=117226959918393&w=2
Changes from the initial RFC:
- Bug fix: needed a transport_class_unregister() for the vport class
- Create a symlink to the vport in the shost device if it is not the
parent of the vport.
- Made symbolic name writable so it can be set after creation
- Made the temporary fc_vport_identifiers struct private to the
transport.
- Deleted the vport_id field from the vport. I couldn't find any good
use for it (and symname is a good replacement).
- Made the vport_state and vport_last_state "private" attributes.
Added the fc_vport_set_state() helper function to manage state
transitions
- Updated vport_create() to allow a vport to be created in a disabled
state.
- Added INITIALIZING and FAILED vport states
- Added VPCERR_xxx defines for errors to be returned from vport_create()
- Created a Documentation/scsi/scsi_fc_transport.txt file that describes
the interfaces and expected LLDD behaviors.
Signed-off-by: James Smart <James.Smart@emulex.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
.. because video-buf.c requires PCI, and VIDEO_EM28XX selects it.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
stuff that does select USB should depend on USB_ARCH_HAS_HCD, or we'll
end up with unbuildable configs.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
need it for page_private(), not all targets have it pulled indirectly
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Use menuconfigs instead of menus, so the whole menu can be disabled at once
instead of going through all options.
Signed-off-by: Jan Engelhardt <jengelh@gmx.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
Not sure how this one got missed in the great purge some time ago but it did.
Signed-off-by: Alan Cox <alan@redhat.com>
Cc: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
ide_use_dma() duplicates a lot of ide_max_dma_mode() functionality
and as all users of ide_use_dma() were converted to use ide_tune_dma()
now it is possible to add missing checks to ide_tune_dma() and remove
ide_use_dma() completely, so do it.
There should be no functionality changes caused by this patch.
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
* check ->speedproc return value in ide_tune_dma()
* use ide_tune_dma() in cmd64x/cs5530/sc1200/siimage/sl82c105/scc_pata drivers
* remove no longer needed ide_dma_enable()
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
Add the speedproc() method for setting transfer modes, modify config_for_dma()
to call it and use ide_max_dma_mode() to select the best DMA mode.
Add support for the multiword DMA modes 0 and 1, using the upper half of the
'drive_data' field to store the DMA timings to program into the drive control
register when DMA is turned on for real.
Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
* add {cs5530,sc1200}_tunepio() for programming PIO timings
* add {cs5530,sc1200}_tune_chipset() (->speedproc method) for setting
transfer mode and convert {cs5530,sc1200}_config_dma() to use it
* bump driver version
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
sc1200.c:
* remove open-coded variant of ide_dma_host_off() (== ->dma_host_off),
it is not needed because ->dma_off_quietly calls ->dma_host_off
* use ->dma_host_on (== ide_dma_host_on() for this driver) instead of
open-coded variant, call it from the users of sc1200_config_dma2()
[ there is no need to call ->dma_host_on in sc1200_config_dma() because
core code takes care of calling ->ide_dma_on on successful execution
of ->ide_dma_check ]
* add comment about ->tuneproc interface abuse
cs5530.c/sc1200.c:
* core code takes care of calling ->dma_off_quietly before calling
->ide_dma_check so there is no need to call it in ->ide_dma_check methods
* bump driver version
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
CS5530/SC1200 specifies that two drives on the same cable cannot mix
UDMA/MDMA. Add {cs5530,sc1200}_udma_filter() to handle this. This also
makes it possible to remove open-coded best DMA mode selection and use
standard ide_use_dma()/ide_max_dma_mode() helpers. While at it bump
version numbers.
There should be no functionality changes caused by this patch.
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
ide_start_power_step() and set_using_dma() were missing ->dma_off_quietly
call (comment in probe_hwif() states that DMA should be always cleared before
tuning is attempted). Fix it.
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
* remove code enabling IORDY and prefetch from config_chipset_for_dma(),
as the comment states it has no real effect because these settings are
overriden when the PIO mode is set (and for this driver ->autotune == 1
so PIO mode is always programmed)
* use ide_tune_dma() in pdcnew_config_drive_xfer_rate() and remove no longer
needed config_chipset_for_dma()
There should be no functionality changes caused by this patch.
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
Use ide_tune_dma() in ali15x3_config_drive_for_dma() and remove all the open
coded DMA tuning code and also config_chipset_for_dma(). Set ->atapi_dma flag
correctly in init_hwif_common_ali15x3() so ide_tune_dma() can take care of
checking if ATAPI DMA is allowed and remove open coded ATAPI DMA check from
ali15x3_config_drive_for_dma().
There should be no functionality changes caused by this patch.
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
* limit max PIO mode to PIO4, this driver doesn't support PIO5 and attempt
to program PIO5 by config_art_rwp_pio() could result in incorrect PIO
timings being programmed and possibly the data corruption (for < ATA100
family chipsets PIO0 timings were used, for ATA100 and ATA100a - the random
content of test1 variable was used, for ATA133 - MWDMA0 timings were used)
* BUG() in sis5513_tune_chipset() if somebody tries to force unsupported PIO5,
also cleanup this function a bit while at it
* add comment about PIO0 timings for < ATA100 family chipsets
* remove open-coded best PIO mode selection from config_art_rwp_pio(),
it contained numerous bugs:
- it didn't check for validity of id->eide_pio_modes and id->eide_pio_iordy
before using them
- it tried to found out maximum PIO mode basing on minimum IORDY cycle time
(moreover wrong cycle times were used for PIO1/5)
- it was overriding PIO blacklist and conservative PIO "downgrade" done
by ide_get_best_pio_mode()
* use sis5513_tune_drive() instead of config_art_rwp_pio()
in sis5513_config_xfer_rate() so the correct PIO mode is also set
on drive even if the device is not IORDY/DMA capable
* config_art_rwp_pio() was always setting the best possible mode and not
the wanted one - fix it and move ide_get_best_pio_mode() call to
config_chipset_for_pio()
* don't use ide_find_best_mode() in config_chipset_for_pio(), it was being
overriden by config_art_rwp_pio() for the host timings anyway + we need to
set the same PIO mode on the device and the host
* pass correct "pio" argument (255 instead of 5) to sis5513_tune_drive() call
in sis5513_config_xfer_rate() so the best PIO mode is set on the drive
and not PIO4
* rename sis5513_tune_drive() to sis5513_tuneproc()
and config_chipset_for_pio() to sis5513_tune_driver()
* bump driver version
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
* limit max PIO mode to PIO4, this driver doesn't support PIO5 and attempt
to program PIO5 by svwks_tune_chipset() could result in incorrect PIO
timings being programmed and possibly the data corruption (it seems that
the minimum possible values were used but I lack the datasheets to be sure)
* select best PIO mode in svwks_tune_drive() and not in svwks_tune_chipset()
when doing PIO autotuning (pio == 255)
* don't try to tune PIO in config_chipset_for_dma() as ide_dma_enable() could
return 1 if DMA was previously enabled (svwks_config_drive_xfer_rate()
takes care of PIO tuning if no suitable DMA mode is found)
* remove config_chipset_for_pio() and use svwks_tune_drive() instead,
config_chipset_for_pio() contained numerous bugs when selecting PIO mode
(luckily it was only used for devices limited to PIO by capabilities/BIOS):
- it didn't check for validity of id->eide_pio_modes and id->eide_pio_iordy
before using them
- it tried to found out maximum PIO mode basing on minimum IORDY cycle time
(moreover wrong cycle times were used for PIO0/1/5)
- it was overriding PIO blacklist and conservative PIO "downgrade" done
by ide_get_best_pio_mode()
- if the max drive PIO was PIO5 then XFER_PIO_0/XFER_PIO_SLOW was selected
(XFER_PIO_SLOW is not supported by svwks_tune_chipset() so the result
was the same as if using XFER_PIO_5 => wrong PIO timings were set)
- it was overriding drive->current_speed
* bump driver version
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
This patch is based on the documentation (I would like to thank Promise
for it) and also partially on the older vendor driver.
Rewrite mode programming code:
* disable 66MHz clock in pdc202xx_tune_chipset() so it is correctly disabled
even if both devices on the channel are not DMA capable and after reset
* enable/disable IORDY and PREFETCH bits in pdc202xx_tune_chipset()
as they need to be setup correctly also for PIO only devices, plus IORDY
wasn't disabled for non-IORDY devices and PREFETCH wasn't disabled for
ATAPI devices
* remove dead code for setting SYNC_ERDDY_EN bits from config_chipset_for_dma()
(driver sets ->autotune to 1 so PIO modes are always programmed => lower
nibble of register A never equals 4 => "chipset_is_set" is always true)
* enable PIO mode programming for all ATAPI devices
(it was limited to ->media == ide_cdrom devices)
* remove extra reads of registers A/B/C, don't read register D et all
* do clearing / programming of registers A/B/C in one go
(gets rid of extra PCI config space read/write cycle)
* set initial values of drive_conf/AP/BP/CP variables to zero
(paranoia for the case when PCI reads fail)
* remove XFER_UDMA6 to XFER_UDMA5 remapping case - it can't happen
(ide_rate_filter() takes care of it)
* fix XFER_MW_DMA0 timings (they were overclocked, use the official ones)
* fix bitmasks for clearing bits of register B:
- when programming DMA mode bit 0x10 of register B was cleared which
resulted in overclocked PIO timing setting (iff PIO0 was used)
- when programming PIO mode bits 0x18 weren't cleared so suboptimal
timings were used for PIO1-4 if PIO0 was previously set (bit 0x10)
and for PIO0/3/4 if PIO1/2 was previously set (bit 0x08)
* add FIXME comment about missing locking for 66MHz clock register
Also while at it:
* remove unused defines
* do a few cosmetic / CodingStyle fixes
* bump driver version
v2:
* in pdc202xx_tune_chipset() the old content of drive configuration
registers is used only by the debugging code so cover "drive_conf"
PCI registers read by #if PDC202XX_DEBUG_DRIVE_INFO
(Noticed by Sergei Shtylyov)
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
It was agreed that phy-connection-type was a better name for
the interface-type property, so this patch renames it.
Also, the max-speed property name was determined too generic,
and is therefore eliminated in favour of phy-connection-type
derivation logic.
includes corrections to copyright text.
Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
Looks like the new version of this patch has been overlooked,
so I'm resending it.
It just adapts the driver to the new IRQ API
according to what Russell has pointed out.
drivers/net/smc911x.c | 6 ++----
1 files changed, 2 insertions(+), 4 deletions(-)
Signed-off-by: Vitaly Wool <vitalywool@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
Turns out we have an old version of firmware that stores the mac address
in 'mac-address' as a string instead of a byte array. All versions that
use local-mac-address should have it as byte array, so no need to do
string parsing for that case.
Signed-off-by: Olof Johansson <olof@lixom.net>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
This caused some very interesting behaviour depending on what happened to
be built at the same time. Add terminating empty entry to the list of IDs.
Signed-off-by: Olof Johansson <olof@lixom.net>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
Interrupt ack fixes
Fix the packet count resets at interrupt time, using the cacheable
packet count status to set number of processed/received packets, since
the ack count is the cumulative number of packets processed, and not
incremental.
Signed-off-by: Olof Johansson <olof@lixom.net>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
Some shift values were obviously wrong. Fix them to correspond with
the masks.
Signed-off-by: Olof Johansson <olof@lixom.net>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
This reverts commit f64da958df.
Andi Kleen is unhappy with the changes, and they really do not seem
worth it. IPMI could use DIE_NMI_IPI instead of the new callback, even
though that ends up having its own set of problems too, mainly because
the IPMI code cannot really know the NMI was from IPMI or not.
Manually fix up conflicts in arch/x86_64/kernel/traps.c and
drivers/char/ipmi/ipmi_watchdog.c.
Cc: Andi Kleen <ak@suse.de>
Cc: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Cc: Corey Minyard <minyard@acm.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
In the presence of some running RX connections, we repeat
queue_delayed_work calls each 4 RX WRs, which is a waste. It's enough
to start stale task when a first passive connection is added, and
rerun it every IPOIB_CM_RX_DELAY as long as there are outstanding
passive connections.
This removes some code from RX data path.
Signed-off-by: Michael S. Tsirkin <mst@dev.mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
mthca_cq_clean() updates the CQ consumer index without moving CQEs
back to HW ownership. As a result, the same WRID might get reported
twice, resulting in a use-after-free. This was observed in IPoIB CM.
Fix by moving all freed CQEs to HW ownership.
This fixes <https://bugs.openfabrics.org/show_bug.cgi?id=617>
Signed-off-by: Michael S. Tsirkin <mst@dev.mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Fix posting lists of > 255 receive WRs for Tavor: rq.next_ind must
be updated each doorbell, otherwise the next doorbell will use an
incorrect index.
Found by Ronni Zimmermann at Mellanox.
Signed-off-by: Michael S. Tsirkin <mst@dev.mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Several checks in the rdma_cm check against the state of the
cm_id, but only to validate that the cm_id is bound to an underlying
transport specific CM and an RDMA device. Make the check explicit
in what we're trying to check for, since we're not synchronizing
against the cm_id state.
This will allow a user to disconnect a cm_id or reject a connection
after receiving a device removal event.
Signed-off-by: Sean Hefty <sean.hefty@intel.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
The cma_iw_handler needs to validate the state of the rdma_cm_id before
processing a new connection request to ensure that a device removal is
not already being processed for the same rdma_cm_id. Without the state
check, the user can receive simultaneous callbacks for the same cm_id, or
a callback after they've destroyed the cm_id.
Signed-off-by: Sean Hefty <sean.hefty@intel.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Add a new routine and rename another to encapsulate common code for
synchronizing with device removal.
Signed-off-by: Sean Hefty <sean.hefty@intel.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
- Scaling code is still considered experimental, so disable it by default
- Increase version to SVNEHCA_0023
Signed-off-by: Joachim Fenkes <fenkes@de.ibm.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
eHCA's sysfs attributes are now being created via sysfs_create_group(),
making the process neatly table-driven. The return value is checked, thus
fixing a few compiler warnings.
Signed-off-by: Joachim Fenkes <fenkes@de.ibm.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
- In ehca_process_eq(), we're IRQ safe throughout the whole function, so we
don't need another _irqsave in the middle of flight.
- take_over_work() is only called by comp_pool_callback(), so it can move
into the same #ifdef block.
Signed-off-by: Joachim Fenkes <fenkes@de.ibm.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
AQP0/1 should report qp_num={0|1} and the actual QP# should be stored
in struct ehca_qp, not the other way round.
Signed-off-by: Joachim Fenkes <fenkes@de.ibm.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
The driver needs to always supply the "GRH present" flag to the
hypervisor, whether it's true or false. Not supplying it (i.e. not
setting the corresponding mask bit) amounts to a "perhaps", which we
don't want.
Signed-off-by: Joachim Fenkes <fenkes@de.ibm.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Some pSeries hypervisor versions show a race condition in the allocate
MR hCall. Serialize this call per adapter to circumvent this problem.
Signed-off-by: Joachim Fenkes <fenkes@de.ibm.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Once upon a time, GPIO interrupts were rare. But then a chip bug in
the waldo series forced the use of a GPIO interrupt to signal packet
reception. This greatly increased the frequency of GPIO interrupts
which have the gpio_mask bits set on the waldo chips. Other bits in
the gpio_status register are used for I2C clock and data lines, these
bits are usually on. An "unlikely" annotation leftover from the old
days was improperly applied to these bits, and an unnecessary chip
mmio read was being accessed in the interrupt fast path on waldo.
Remove the stagnant unlikely annotation in the interrupt handler and
keep a shadow copy of the gpio_mask register to avoid the slow mmio
read when testing for interruptable GPIO bits.
Signed-off-by: Arthur Jones <arthur.jones@qlogic.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
uar_lock spinlock was used in mlx4_ib_cq_arm without being initialized
(this only affects 32-bit archs, because uar_lock is not used on
64-bit archs and MLX4_INIT_DOORBELL_LOCK() is a NOP).
Signed-off-by: Jack Morgenstein <jackm@dev.mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
... and make it depend on the response flag instead of the command type.
Signed-off-by: Nicolas Pitre <npitre@mvista.com>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>
Alan Cox