Andre Schwarz [Thu, 10 Jul 2008 09:53:16 +0000 (11:53 +0200)]
powerpc/mpc5200: PCI write combine timer
On MPC5200 the PCI target control register (PCITCR) @ MBAR + 0xD6C is
initialized with only bit 7 (Latrule disable) set. The 8-Bit write
combine timer (Bits 24..31) should be also set to a reasonable value
_greater zero_ (0x08 = default) since setting it to 0x00 leads to
_very poor_ performance as a PCI target since external burst won't be
possible at all.
Setting the WCT to 0x08 (cache-line size) leads to good overall perfomance.
Signed-off-by: Andre Schwarz <andre.schwarz@matrix-vision.de> Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
James Bottomley [Mon, 7 Jul 2008 20:50:01 +0000 (15:50 -0500)]
[SCSI] bsg: fix oops on remove
If you do a modremove of any sas driver, you run into an oops on
shutdown when the host is removed (coming from the host bsg device).
The root cause seems to be that there's a use after free of the
bsg_class_device: In bsg_kref_release_function, this is used (to do a
put_device(bcg->parent) after bcg->release has been called. In sas (and
possibly many other things) bcd->release frees the queue which contains
the bsg_class_device, so we get a put_device on unreferenced memory.
Fix this by taking a copy of the pointer to the parent before releasing
bsg.
Acked-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
[SCSI] sym53c8xx: Fix bogus sym_que_entry re-implementation of container_of
The sym53c8xx driver, for some reason, seems to mostly re-implement
linux/list.h with added bogosity. The main one is it's implementation
of sym_que_entry which spits warnings with recent gcc's on some
64 bits architectures.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
Adrian Bunk [Fri, 4 Jul 2008 06:47:33 +0000 (23:47 -0700)]
[SCSI] make struct scsi_{host,target}_type static
Make the needlessly global struct scsi_{host,target}_type static.
Signed-off-by: Adrian Bunk <bunk@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
James Bottomley [Thu, 3 Jul 2008 16:31:55 +0000 (11:31 -0500)]
[SCSI] fix locking in host use of blk_plug_device()
scsi_lib.c:scsi_host_queue_ready() plugs the device with incorrect
locking. It should actually have the queue lock held, but it's
holding the host lock. Fix this by eliminating the call. The host
ready has no need to plug the queue because if it returns 0 in
scsi_request_function control transfers to not_ready which acquires
the queue lock and plugs the device if its at zero depth.
Reported-by: Elias Oltmanns <eo@nebensachen.de> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
[SCSI] zfcp: consolidate sysfs things into one file.
zfcp was using three files to deal with sysfs representation
for adapters, ports and units. The consolidation into one file
prevents code-duplication and eases maintainability.
[SCSI] zfcp: Move status accessors from zfcp to SCSI include file.
Move the accessor functions for the scsi_cmnd status from zfcp to the
SCSI include file. Change the interface to the functions to pass the
scsi_cmnd pointer instead of the status pointer.
Signed-off-by: Martin Petermann <martin@linux.vnet.ibm.com> Signed-off-by: Christof Schmitt <christof.schmitt@de.ibm.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
[SCSI] zfcp: Adapter reopen for large number of unsolicited status
When zfcp receives 16 unsolicited status messages, this could trigger
an adapter reopen. In this case, first try to send a new status read,
and only if this fails, go through the recovery.
[SCSI] zfcp: wait until adapter is finished with ERP during auto-port
In some situations the auto port attachment task is started
before the ERP is finished. To prevent this unwanted situation
we wait until the adapter is up and running before we start our work.
Brian King [Tue, 1 Jul 2008 18:14:30 +0000 (13:14 -0500)]
[SCSI] ibmvfc: IBM Power Virtual Fibre Channel Adapter Client Driver
This patch adds a new device driver to support the Virtual Fibre Channel
interface on IBM Power based servers. The Virtual I/O Server on IBM Power
servers utilizes N-Port ID Virtualization to export a Virtual Fibre Channel
adapter to the client. This driver is the client device driver.
Signed-off-by: Brian King <brking@linux.vnet.ibm.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
Brian King [Tue, 1 Jul 2008 18:03:19 +0000 (13:03 -0500)]
[SCSI] sg: Add target reset support
Adds support for target reset to SG_SCSI_RESET.
Signed-off-by: Brian King <brking@linux.vnet.ibm.com> Acked-by: Douglas Gilbert <dougg@torque.net> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
[SCSI] lib: Add support for the T10 (SCSI) Data Integrity Field CRC
The SCSI Block Protocol uses this 16-bit CRC to verify the integrity
of each data sector. crc_t10dif() is used by sd_dif.c when performing
I/O to or from disks formatted with protection information.
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
Frans Pop [Fri, 20 Jun 2008 03:20:12 +0000 (20:20 -0700)]
[SCSI] esp: correct module name in Kconfig help for SCSI_SUNESP
The module name was changed from esp to sun_esp some time ago.
Also correct the list of chips supported by the driver.
Signed-off-by: Frans Pop <elendil@planet.nl> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
David S. Miller [Mon, 1 Oct 2007 00:10:42 +0000 (17:10 -0700)]
[SCSI] esp: Correct chip ID probing sequence.
The features enable bit has to be set in the config2 register
before we can be absolutely sure we will probe a correct
part unique ID and family code from the transfer-count-high
register.
Also, reload the CFACT, STP, SOFF, and TIMEO near the end of
esp_reset_esp().
From a patch by Maciej W. Rozycki.
Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
Matthew Wilcox [Thu, 19 Jun 2008 16:02:58 +0000 (10:02 -0600)]
[SCSI] scsi_debug: add support for rotation speed
Add support for VPD page b1 to scsi_debug
SCSI VPD page b1 reports the nominal rotation speed of the device.
Since scsi_debug is ram-based, claim to be a non-rotating medium.
Signed-off-by: Matthew Wilcox <willy@linux.intel.com> Acked-by: Douglas Gilbert <dougg@torque.net> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
Mike Christie [Mon, 16 Jun 2008 15:11:35 +0000 (10:11 -0500)]
[SCSI] iscsi class: fix endpoint leak
class_find_device gets a ref to the device so we must release it.
The class will serialize access to the ep so we do not have to worry
about a remove racing with the callers access, so we can simplify the
use and drop the ref right away.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
Mike Christie [Mon, 16 Jun 2008 15:11:33 +0000 (10:11 -0500)]
[SCSI] libiscsi, iscsi_tcp, ib_iser: fix setting of can_queue with old tools.
This patch fixes two bugs that are related.
1. Old tools did not set can_queue/cmds_max. This patch modifies
libiscsi so that when we add the host we catch this and set it
to the default.
2. iscsi_tcp thought that the scsi command that was passed to
the eh functions needed a iscsi_cmd_task allocated for it. It
only needed a mgmt task, and now it does not matter since it
all comes from the same pool and libiscsi handles this for the
drivers. ib_iser had copied iscsi_tcp's code and set can_queue
to its max - 1 to handle this. So this patch removes the max -1,
and just sets it to the max.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
James Smart [Sun, 15 Jun 2008 02:52:59 +0000 (22:52 -0400)]
[SCSI] lpfc 8.2.7 : Miscellaneous Fixes
Miscellaneous Fixes:
- Fix bug in mbox sysfs interface that locked in EAGAIN if discovery stalled.
- Fix missing error message when npiv and loop are true when link up occurs.
- Fix panic in lpfc_scsi_cmd_iocb_cmpl: scsi_buf was NULL, but created
race conditions with other code paths.
- Fix error in sysfs mailbox structure that didn't rezero on next use.
- Add missing mempool_free() to attachment failure path
- Fix missing put of ndlp structure during driver unload.
- Fix applications unable to send mailbox commands during discovery.
- Remove unused argument (type) from function lpfc_post_buffer() API
- Fix vport name is not shown after hbacmd vportcreate.
- Remove repeated code statements.
Signed-off-by: James Smart <james.smart@emulex.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
James Smart [Sun, 15 Jun 2008 02:52:53 +0000 (22:52 -0400)]
[SCSI] lpfc 8.2.7 : Rework the worker thread
Rework of the worker thread to make it more efficient.
Make a finer-grain notfication of pending work so less time is
spent checking conditions. Also made other general cleanups.
Signed-off-by: James Smart <james.smart@emulex.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
James Smart [Sun, 15 Jun 2008 02:52:47 +0000 (22:52 -0400)]
[SCSI] lpfc 8.2.7 : Discovery Fixes
- Fix ADISC timeout on initiators causing devloss timeout on targets
- Correct FAN processing : port state vs unreg rpi's wasn't consistent
- Correct mismatches between ASICs and PLOGI that would skip PLOGI
Signed-off-by: James Smart <james.smart@emulex.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
James Smart [Sun, 15 Jun 2008 02:52:38 +0000 (22:52 -0400)]
[SCSI] lpfc 8.2.7 : Change device reset behavior
Prior handler was only waiting for I/O on one lun to finish before
returning completion. Now, wait for all LUNs on the target.
Also performed some rudimentary cleanup while in this code.
Signed-off-by: James Smart <james.smart@emulex.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
Swen Schillig [Tue, 10 Jun 2008 16:21:00 +0000 (18:21 +0200)]
[SCSI] zfcp: Automatically attach remote ports
Automatically attach the remote ports in zfcp when the adapter is set
online. This is done by querying all available ports from the FC
namesever. The scan for remote ports is also triggered by RSCNs and
can be triggered manually with the sysfs attribute 'port_rescan'.
Sven Schuetz [Tue, 10 Jun 2008 16:20:59 +0000 (18:20 +0200)]
[SCSI] zfcp: Add port_state attribute to sysfs
The sysfs attribute /sys/class/fc_host/hostX/port_state was not set by
zfcp so far.
Now, the appropriate members of the fc_function_template struct are
set during its initialziation. The first is a boolean to show the port
state. The second is a function pointer to the function
zfcp_get_host_port_state, which reads the port state from our adapter
status bits and calls fc_host_port_state with the approriate port
state afterwards.
Signed-off-by: Sven Schuetz <sven@linux.vnet.ibm.com> Signed-off-by: Christof Schmitt <christof.schmitt@de.ibm.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
Christof Schmitt [Tue, 10 Jun 2008 16:20:58 +0000 (18:20 +0200)]
[SCSI] zfcp: Message cleanup
Cleanup the messages used in the zfcp driver: Remove unnecessary debug
and trace message and convert the remaining messages to standard
kernel macros. Remove the zfcp message macros and while updating the
whole flie also update the copyright headers.
Christof Schmitt [Tue, 10 Jun 2008 16:20:55 +0000 (18:20 +0200)]
[SCSI] zfcp: Move CFDC code to new file.
zfcp implements a device file to allow Linux guests changing the
Access Control Tables stored in the adapter. The code for the device
file has nothing to do with the other parts of the driver, so move it
to a new file and cleanup the code while doing so.
Mark Salyzyn [Wed, 28 May 2008 19:32:55 +0000 (15:32 -0400)]
[SCSI] aacraid: prevent copy_from_user() BUG!
Seen:
kernel BUG at arch/i386/lib/usercopy.c:872
under a 2.6.18-8.el5 kernel. Traced it to a garbage-in/garbage-out
ioctl condition in the aacraid driver.
Adaptec's special ioctl scb passthrough needs to check the validity of
the individual scatter gather count fields to the maximum the adapter
supports. Doing so will have the side effect of preventing
copy_from_user() from bugging out while populating the dma buffers.
This is a hardening effort, issue was triggered by an errant version
of the management tools and thus the BUG should not be seen in the
field.
[jejb: fixed up compile failure] Signed-off-by: Mark Salyzyn <aacraid@adaptec.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
Make scsi_debug sector size configurable at load time instead of being
a #define. Handy for testing 4KB sectors.
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Acked-by: Douglas Gilbert <dougg@torque.net> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
FUJITA Tomonori [Wed, 28 May 2008 22:56:55 +0000 (07:56 +0900)]
[SCSI] bsg: fix bsg_mutex hang with device removal
We don't need to hold bsg_mutex during bsg_complete_all_commands(). It
leads to a problem that we block bsg_unregister_queue during
bsg_complete_all_commands (untill all the outstanding commands
complete).
Thanks to Pete Wyckoff for finding the bug and testing the patch.
Boaz Harrosh [Mon, 26 May 2008 08:31:19 +0000 (11:31 +0300)]
[SCSI] iscsi_tcp: Enable any size command
Let through upto the largest command of 260 defined by the scsi standard.
iscsi core supports this already. Now that the scsi-ml supports it we can
start using large commands.
[jejb:rejections fixed up] Signed-off-by: Boaz Harrosh <bharrosh@panasas.com> Acked-by: Mike Christie <michaelc@cs.wisc.edu> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
Harvey Harrison [Wed, 21 May 2008 20:54:20 +0000 (15:54 -0500)]
[SCSI] scsi: use get_unaligned_* helpers
Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com> Signed-off-by: Mike Christie <michaelc@cs.wisc.edu> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
Harvey Harrison [Wed, 21 May 2008 20:54:19 +0000 (15:54 -0500)]
[SCSI] Replace __FUNCTION__ with __func__ in iscsi_tcp.
__FUNCTION__ is gcc-specific, use __func__
(Update diff by Mike Christie)
Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com> Signed-off-by: Mike Christie <michaelc@cs.wisc.edu> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
Mike Christie [Wed, 21 May 2008 20:54:18 +0000 (15:54 -0500)]
[SCSI] libiscsi, iser, tcp: remove recv_lock
The recv lock was defined so the iscsi layer could block
the recv path from processing IO during recovery. It
turns out iser just set a lock to that pointer which was pointless.
We now disconnect the transport connection before doing recovery
so we do not need the recv lock. For iscsi_tcp we still stop
the recv path incase older tools are being used.
This patch also has iscsi_itt_to_ctask user grab the session lock
and has the caller access the task with the lock or get a ref
to it in case the target is broken and sends a tmf success response
then sends data or a response for the command that was supposed to
be affected bty the tmf.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
Mike Christie [Wed, 21 May 2008 20:54:17 +0000 (15:54 -0500)]
[SCSI] libiscsi: fix cmds_max setting
Drivers expect that the cmds_max value they pass to the iscsi layer
is the max scsi commands + mgmt tasks. This patch implements that
and fixes some checks for nr cmd limits.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
Mike Christie [Wed, 21 May 2008 20:54:16 +0000 (15:54 -0500)]
[SCSI] iscsi class: Add session initiatorname and ifacename sysfs attrs.
This adds two new attrs used for creating initiator ports and
binding sessions to hardware.
The session level initiatorname:
Since bnx2i does a scsi_host per host device, we need to add the
iface initiator port settings on the session, so we can create
multiple initiator ports (each with different inames) per device/scsi_host.
The current iname reflects that qla4xxx can have one iname per hba, and we are
allocating a host per session for software. The iname on the host will
remain so we can export and set the hba level qla4xxx setting.
The ifacename attr:
To bind a session to a some peice of hardware in userspace we maintain
some mappings, but during boot or iscsid restart (iscsid contains the user
space part of the driver) we need to be able to figure out which of those
host mappings abstractions maps to certain sessions. This patch adds
a ifacename attr, which userspace can set to id the host side of the
endpoint across pivot_roots and iscsid restarts.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
Mike Christie [Wed, 21 May 2008 20:54:14 +0000 (15:54 -0500)]
[SCSI] iser: Modify iser to take a iscsi_endpoint struct in ep callouts and session setup
This hooks iser into the iscsi endpoint code. Previously it handled the
lookup and allocation. This has been made generic so bnx2i and iser can
share it. It also allows us to pass iser the leading conn's ep, so we
know the ib_deivce being used and can set it as the scsi_host's parent.
And that allows scsi-ml to set the dma_mask based on those values.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
Mike Christie [Wed, 21 May 2008 20:54:13 +0000 (15:54 -0500)]
[SCSI] iscsi class: add endpoint class
Add sysfs representation for the endpoint, so userspace can match the
host and session to the endpoint. This will allow us to set the host's
parent correctly at host creation time.
The next patches will convert tcp and iser, and fix iser's dma_mask
bug.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
Mike Christie [Wed, 21 May 2008 20:54:12 +0000 (15:54 -0500)]
[SCSI] iscsi class: user device_for_each_child instead of duplicating session list
Currently we duplicate the list of sessions, because we were using the
test for if a session was on the host list to indicate if the session
was bound or unbound. We can instead use the target_id and fix up
the class so that drivers like bnx2i do not have to manage the target id
space.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
Mike Christie [Wed, 21 May 2008 20:54:06 +0000 (15:54 -0500)]
[SCSI] libiscsi: merge iscsi_mgmt_task and iscsi_cmd_task
There is no need to have the mgmt and cmd tasks separate
structs. It used to save a lot of memory when we overprealocated
memory for tasks, but the next patches will set up the
driver so in the future they can use a mempool or some other
common scsi command allocator and common tagging.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
Mike Christie [Wed, 21 May 2008 20:54:05 +0000 (15:54 -0500)]
[SCSI] libiscsi: modify libiscsi so it supports offloaded data paths
This patch modifies libiscsi, so drivers like bnx2i and iser can execute
a command from queuecommand/send_pdu instead of having to be queued to
be run in a workq.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
Currently to get a ctask from the session cmd array, you have to
know to use the itt modifier. To make this easier on LLDs and
so in the future we can easilly kill the session array and use
the host shared map instead, this patch adds a nice wrapper
to strip the itt into a session->cmds index and return a ctask.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
Mike Christie [Wed, 21 May 2008 20:54:03 +0000 (15:54 -0500)]
[SCSI] iser: fix handling of scsi cmnds during recovery.
After the stop_conn callback has returned the LLD should not
touch the scsi cmds. iscsi_tcp and libiscsi use the
conn->recv_lock and suspend_rx field to halt recv path
processing, but iser does not have any protection.
This patch modifies iser so that userspace can just
call the ep_disconnect callback, which will halt
all recv IO, before calling the stop_conn callback so
we do not have to worry about the conn->recv_lock and
suspend rx field. iser just needs to stop the send side
from accessing the ib conn.
Fixup to handle when the ep poll fails and ep disconnect
is called from Erez.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
Mike Christie [Wed, 21 May 2008 20:54:02 +0000 (15:54 -0500)]
[SCSI] iscsi: modify iscsi printk so it can take driver data pointers
Some drivers want to be able to just pass in the driver data pointers
to the iscsi objects. To enable this we need the iscsi printk macro
to cast the object.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
Mike Christie [Wed, 21 May 2008 20:54:01 +0000 (15:54 -0500)]
[SCSI] iscsi: remove session/conn_data_size from iscsi_transport
This removes the session and conn data_size fields from the iscsi_transport.
Just pass in the value like with host allocation. This patch also makes
it so the LLD iscsi_conn data is allocated with the iscsi_cls_conn.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
Mike Christie [Wed, 21 May 2008 20:53:56 +0000 (15:53 -0500)]
[SCSI] iscsi class, iscsi_tcp/iser: add host arg to session creation
iscsi offload (bnx2i and qla4xx) allocate a scsi host per hba,
so the session creation path needs a shost/host_no argument.
Software iscsi/iser will follow the same behabior as before
where it allcoates a host per session, but in the future iser
will probably look more like bnx2i where the host's parent is
the hardware (rnic for iser and for bnx2i it is the nic), because
it does not use a socket layer like how iscsi_tcp does.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
Prakash, Sathya [Tue, 20 May 2008 19:32:18 +0000 (01:02 +0530)]
[SCSI] mpt fusion : Adding FAULT Reset polling work
When the firmware is in Fault state it will be identifed only when the next time
the driver access the IOC state.
This patch includes a polling function in the driver which will be executed in
regular interval to check the status of the firmware and if it is in Fault
state, then the firmware will be reset by the driver.
Signed-off-by: Sathya Prakash <sathya.prakash@lsi.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
James Bottomley [Fri, 11 Jul 2008 03:10:55 +0000 (22:10 -0500)]
[SCSI] fusion: default MSI to disabled for SPI and FC controllers
There's a fault on the FC controllers that makes them not respond
correctly to MSI. The SPI controllers are fine, but are likely to be
onboard on older motherboards which don't handle MSI correctly, so
default both these cases to disabled. Enable by setting the module
parameter mpt_msi_enable=1.
For the SAS case, enable MSI by default, but it can be disabled by
setting the module parameter mpt_msi_enable=0.
Cc: "Prakash, Sathya" <sathya.prakash@lsi.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
Jan Glauber [Sat, 12 Jul 2008 07:42:11 +0000 (15:42 +0800)]
crypto: s390 - Respect STFL bit
Bevore issuing any s390 crypto operation check whether the
CPACF facility is enabled in the facility list. That way a
virtualization layer can prevent usage of the CPACF facility
regardless of the availability of the crypto instructions.
Signed-off-by: Jan Glauber <jang@linux.vnet.ibm.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
x64, x2apic/intr-remap: Interrupt-remapping and x2apic support
On Thu, Jul 10, 2008 at 12:53:20PM -0700, Ingo Molnar wrote:
>
> Btw., i threw it at the -tip test-cluster and got back a quick build
> bugreport:
>
> arch/x86/xen/enlighten.c: In function 'xen_patch':
> arch/x86/xen/enlighten.c:1084: warning: label 'patch_site' defined but not used
> arch/x86/xen/enlighten.c: At top level:
> arch/x86/xen/enlighten.c:1272: error: expected identifier before '(' token
> arch/x86/xen/enlighten.c:1273: error: expected '}' before '.' token
> arch/x86/kernel/paravirt.c:376:2: error: invalid preprocessing directive
> #ifndedarch/x86/kernel/paravirt.c:384:2: error: #endif without #if
>
> with this config:
>
> http://redhat.com/~mingo/misc/config-Thu_Jul_10_21_43_28_CEST_2008.bad
x64, x2apic/intr-remap: add x2apic support, including enabling interrupt-remapping
x2apic support. Interrupt-remapping must be enabled before enabling x2apic,
this is needed to ensure that IO interrupts continue to work properly after the
cpu mode is changed to x2apic(which uses 32bit extended physical/cluster
apic id).
On systems where apicid's are > 255, BIOS can handover the control to OS in
x2apic mode. Or if the OS handover was in legacy xapic mode, check
if it is capable of x2apic mode. And if we succeed in enabling
Interrupt-remapping, then we can enable x2apic mode in the CPU.
x64, x2apic/intr-remap: MSI and MSI-X support for interrupt remapping infrastructure
MSI and MSI-X support for interrupt remapping infrastructure.
MSI address register will be programmed with interrupt-remapping table
entry(IRTE) index and the IRTE will contain information about the vector,
cpu destination, etc.
For MSI-X, all the IRTE's will be consecutively allocated in the table,
and the address registers will contain the starting index to the block
and the data register will contain the subindex with in that block.
This also introduces a new irq_chip for cleaner irq migration (in the process
context as opposed to the current irq migration in the context of an interrupt.
interrupt-remapping infrastructure will help us achieve this).
As MSI is edge triggered, irq migration is a simple atomic update(of vector
and cpu destination) of IRTE and flushing the hardware cache.
x64, x2apic/intr-remap: IO-APIC support for interrupt-remapping
IO-APIC support in the presence of interrupt-remapping infrastructure.
IO-APIC RTE will be programmed with interrupt-remapping table entry(IRTE)
index and the IRTE will contain information about the vector, cpu destination,
trigger mode etc, which traditionally was present in the IO-APIC RTE.
Introduce a new irq_chip for cleaner irq migration (in the process
context as opposed to the current irq migration in the context of an interrupt.
interrupt-remapping infrastructure will help us achieve this cleanly).
For edge triggered, irq migration is a simple atomic update(of vector
and cpu destination) of IRTE and flush the hardware cache.
For level triggered, we need to modify the io-apic RTE aswell with the update
vector information, along with modifying IRTE with vector and cpu destination.
So irq migration for level triggered is little bit more complex compared to
edge triggered migration. But the good news is, we use the same algorithm
for level triggered migration as we have today, only difference being,
we now initiate the irq migration from process context instead of the
interrupt context.
In future, when we do a directed EOI (combined with cpu EOI broadcast
suppression) to the IO-APIC, level triggered irq migration will also be
as simple as edge triggered migration and we can do the irq migration
with a simple atomic update to IO-APIC RTE.
TBD: some tests/changes needed in the presence of fixup_irqs() for
level triggered irq migration.
x64, x2apic/intr-remap: disable DMA-remapping if Interrupt-remapping is detected (temporary quirk)
Interrupt-remapping enables queued invalidation. And once queued invalidation
is enabled, IOTLB invalidation also needs to use the queued invalidation
mechanism and the register based IOTLB invalidation doesn't work.
For now, Support for IOTLB invalidation using queued invalidation is
missing. Meanwhile, disable DMA-remapping, if Interrupt-remapping
support is detected.
For the meanwhile, if someone wants to really enable DMA-remapping, they
can use nox2apic, which will disable interrupt-remapping and as such
doesn't enable queued invalidation.
And given that none of the release platforms support intr-remapping yet,
we should be ok for this temporary hack.
Introduce basic apic operations which handle the apic programming. This
will be used later to introduce another specific operations for x2apic.
For the perfomance critial accesses like IPI's, EOI etc, we use the
native operations as they are already referenced by different
indirections like genapic, irq_chip etc.
64bit Paravirt ops can also define their apic operations accordingly.
genapic's read_apic_id() returns the actual apic id extracted from
the APIC_ID register. And in some cases like UV, read_apic_id()
returns completely different values from APIC ID register.
Use the native apic register read, rather than genapic read_apic_id()
in verify_local_APIC()
And also, lapic_suspend() should also use native apic register read.
projects / linux-2.6-omap-h63xx.git / log