We discovered a problem when running IPoIB applications on multiple
CPUs on an Altix system. Many messages such as:
ib_mthca 0002:01:00.0: SQ 000014 full (19941644 head, 19941707 tail, 64 max, 0 nreq)
appear in syslog, and the driver wedges up.
Apparently this is because writes to the doorbells from different CPUs
reach the device out of order. The following patch adds mmiowb() calls
after doorbell rings to ensure the doorbell writes are ordered.
Signed-off-by: Arthur Kepner <akepner@sgi.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
All HCAs (not just mem-free) need a spare SRQ entry, so bump srq->max
by 1 in all cases.
Noted by Jack Morgenstein <jackm@mellanox.co.il>
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Fill in "max_vl_num" (encoded according to VLCap field in the PortInfo MAD)
and "init_type_reply" values in the ib_query_port() verb.
Signed-off-by: Jack Morgenstein <jackm@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Maintain a per-CPU global "struct pt_regs *" variable which can be used instead
of passing regs around manually through all ~1800 interrupt handlers in the
Linux kernel.
The regs pointer is used in few places, but it potentially costs both stack
space and code to pass it around. On the FRV arch, removing the regs parameter
from all the genirq function results in a 20% speed up of the IRQ exit path
(ie: from leaving timer_interrupt() to leaving do_IRQ()).
Where appropriate, an arch may override the generic storage facility and do
something different with the variable. On FRV, for instance, the address is
maintained in GR28 at all times inside the kernel as part of general exception
handling.
Having looked over the code, it appears that the parameter may be handed down
through up to twenty or so layers of functions. Consider a USB character
device attached to a USB hub, attached to a USB controller that posts its
interrupts through a cascaded auxiliary interrupt controller. A character
device driver may want to pass regs to the sysrq handler through the input
layer which adds another few layers of parameter passing.
I've build this code with allyesconfig for x86_64 and i386. I've runtested the
main part of the code on FRV and i386, though I can't test most of the drivers.
I've also done partial conversion for powerpc and MIPS - these at least compile
with minimal configurations.
This will affect all archs. Mostly the changes should be relatively easy.
Take do_IRQ(), store the regs pointer at the beginning, saving the old one:
struct pt_regs *old_regs = set_irq_regs(regs);
And put the old one back at the end:
set_irq_regs(old_regs);
Don't pass regs through to generic_handle_irq() or __do_IRQ().
In timer_interrupt(), this sort of change will be necessary:
- update_process_times(user_mode(regs));
- profile_tick(CPU_PROFILING, regs);
+ update_process_times(user_mode(get_irq_regs()));
+ profile_tick(CPU_PROFILING);
I'd like to move update_process_times()'s use of get_irq_regs() into itself,
except that i386, alone of the archs, uses something other than user_mode().
Some notes on the interrupt handling in the drivers:
(*) input_dev() is now gone entirely. The regs pointer is no longer stored in
the input_dev struct.
(*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does
something different depending on whether it's been supplied with a regs
pointer or not.
(*) Various IRQ handler function pointers have been moved to type
irq_handler_t.
Signed-Off-By: David Howells <dhowells@redhat.com>
(cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
If a QP has separate send and receive CQs, then the send CQ will never
have receive completions from that QP in it. So when cleaning the
send CQ, there's no need to pass in an SRQ pointer, even if the QP is
attached to an SRQ.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Trigger device remove and then add when a catastrophic error is
detected in hardware. This, in turn, will cause a device reset, which
we hope will recover from the catastrophic condition.
Since this might interefere with debugging the root cause, add a
module option to suppress this behaviour.
Signed-off-by: Jack Morgenstein <jackm@mellanox.co.il>
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Modifications to the existing rdma header files, core files, drivers,
and ulp files to support iWARP, including:
- Hook iWARP CM into the build system and use it in rdma_cm.
- Convert enum ib_node_type to enum rdma_node_type, which includes
the possibility of RDMA_NODE_RNIC, and update everything for this.
Signed-off-by: Tom Tucker <tom@opengridcomputing.com>
Signed-off-by: Steve Wise <swise@opengridcomputing.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Remove some trailing whitespace that has snuck in despite the best
efforts of whitespace=error-all. Also fix a few other whitespace
bogosities.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Incorrect number of bits was taken for static_rate field.
Signed-off-by: Jack Morgenstein <jackm@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
port_num was not being returned for unconnected QPs.
Signed-off-by: Jack Morgenstein <jackm@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
When default static rate is returned for Tavor, need to translate it
to an ib rate value.
Signed-off-by: Jack Morgenstein <jackm@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Pass a struct ib_udata to the low-level driver's ->modify_srq() and
->modify_qp() methods, so that it can get to the device-specific data
passed in by the userspace driver.
Signed-off-by: Ralph Campbell <ralph.campbell@qlogic.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Make kernel use UAR2 instead of UAR1 for hardware access: this adds
sanity checking from the hardware side, without any performance cost.
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
The SM LID used to send traps to is incorrectly set to port LID. This
is a regression from 2.6.17 -- after a PortInfo MAD is received, no
traps are sent to the SM LID. The traps go to the loopback interface
instead, and are dropped there. The SM LID should be taken from the
sm_lid of the PortInfo response.
The bug was introduced by commit 12bbb2b7be:
IB/mthca: Add client reregister event generation
Signed-off-by: Jack Morgenstein <jackm@dev.mellanox.co.il>
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
It is supposed to be OK to call mthca_create_ah() and mthca_destroy_ah()
from any context. However, for mem-full HCAs, these functions use the
mthca_alloc() and mthca_free() bitmap helpers, and those helpers use
non-IRQ-safe spin_lock() internally. Lockdep correctly warns that
this could lead to a deadlock. Fix this by changing mthca_alloc() and
mthca_free() to use spin_lock_irqsave().
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Update the driver's list of HCA firmware revisions to make sure people
running Sinai firmware older than 1.1.0 get a message suggesting a
firmware upgrade. Update the Arbel versions as well while we are at it.
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Leave all SRQ methods out of the device's uverbs_cmd_mask if the
device doesn't have SRQ support (because of ancient firmware) so that
we don't allow userspace to call the driver's create_srq method. This
fixes a userspace-triggerable oops caused by ib_uverbs_create_srq()
following the device's ->create_srq function pointer, which will be
NULL if the device doesn't support SRQs.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
When destroying a QP, mthca locks both the QP's send CQ and receive
CQ. However, the following scenario is perfectly valid:
QP_a: send_cq == CQ_x, recv_cq == CQ_y
QP_b: send_cq == CQ_y, recv_cq == CQ_x
The old mthca code simply locked send_cq and then recv_cq, which in
this case could lead to an AB-BA deadlock if QP_a and QP_b were
destroyed simultaneously.
We can fix this by changing the locking code to lock the CQ with the
lower CQ number first, which will create a consistent lock ordering.
Also, the second CQ is locked with spin_lock_nested() to tell lockdep
that we know what we're doing with the lock nesting.
This bug was found by lockdep.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
The fence bit needs to be set in the doorbell too, not just the WQE.
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Define a constant MTHCA_ARRAY_MASK to replace repeated uses of
(PAGE_SIZE / sizeof (void *) - 1) in mthca array code.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
mthca_array_clear() does not clear the slot if the used count is
positive. This leads to crashes in mthca_qp_event() since that uses
mthca_array_get() to check that the qp is valid.
Discovered by Ali Ayoub.
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Read the max_cmds value from the response to the QUERY_FW command
before printing out the value, so that the real value goes into the
debug output.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Mem-free HCAs always keep one spare SRQ WQE, so the SRQ limit cannot
be set beyond srq->max - 1.
Signed-off-by: Dotan Barak <dotanb@mellanox.co.il>
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
After recent changes, mthca_wq_init does not actually initialize the WQ as it
used to - it simply resets all index fields to their initial values. So,
let's rename it to mthca_wq_reset.
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Acked-by: Zach Brown <zach.brown@oracle.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
mthca_ah_query returs the static rate of the address handle in internal mthc
format. fix it to use rate encoding from enum ib_rate, which is what users
expect.
Signed-off-by: Jack Morgenstein <jackm@mellanox.co.il>
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
mthca: initialize send and receive queue locks separately
lockdep identifies a lock by the call site of its initialization. By
initializing the send and receive queue locks in mthca_wq_init() we confuse
lockdep. It warns that that the ordered acquiry of both locks in
mthca_modify_qp() is recursive acquiry of one lock:
=============================================
[ INFO: possible recursive locking detected ]
---------------------------------------------
modprobe/1192 is trying to acquire lock:
(&wq->lock){....}, at: [<f892b4db>] mthca_modify_qp+0x60/0xa7b [ib_mthca]
but task is already holding lock:
(&wq->lock){....}, at: [<f892b4ce>] mthca_modify_qp+0x53/0xa7b [ib_mthca]
Initializing the locks separately in mthca_alloc_qp_common() stops the
warning and will let lockdep enforce proper ordering on paths that acquire
both locks.
Signed-off-by: Zach Brown <zach.brown@oracle.com>
Cc: Roland Dreier <rolandd@cisco.com>
Cc: Arjan van de Ven <arjan@linux.intel.com>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This is needed if we wish to change the size of the resource structures.
Based on an original patch from Vivek Goyal <vgoyal@in.ibm.com>
Cc: Vivek Goyal <vgoyal@in.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Documentation/infiniband/core_locking.txt says:
All of the methods in struct ib_device exported by a low-level
driver must be fully reentrant. The low-level driver is required to
perform all synchronization necessary to maintain consistency, even
if multiple function calls using the same object are run
simultaneously.
However, mthca's modify_qp, modify_srq and resize_cq methods are
currently not reentrant. Add a mutex to the QP, SRQ and CQ structures
so that these calls can be properly serialized.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Some error paths after the mthca_alloc_mailbox() call in mthca_modify_qp()
just do a "return -EINVAL" without freeing the mailbox. Convert these
returns to "goto out" to avoid leaking the mailbox storage.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Report the true max_map_per_fmr value from mthca_query_device(),
taking into account the change in FMR remapping introduced by the
Sinai performance optimization.
Signed-off-by: Or Gerlitz <ogerlitz@voltaire.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Change the mthca snoop of MADs that set PortInfo to check if the SM
has set the client reregister bit, and if it has, generate a client
reregister event. If the bit is not set, just generate a LID change
event as usual.
Signed-off-by: Leonid Arsh <leonida@voltaire.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
The kernel has had wait_for_completion_timeout() for a long time now.
mthca should use it to handle FW commands timing out, instead of
implementing the same thing in a much more complicated way by using
wait_for_completion() along with a timer that does complete().
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Memfree firmware is in rare cases reporting WQE index == base - 1 in
receive completion with error, instead of (rq size - 1); base is 0 in
mthca. Here is a patch to avoid kernel crash and report a correct WR
id in this case.
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
mthca does not restore the following PCI-X/PCI Express registers after reset:
PCI-X device: PCI-X command register
PCI-X bridge: upstream and downstream split transaction registers
PCI Express : PCI Express device control and link control registers
This causes instability and/or bad performance on systems where one of
these registers is set to a non-default value by BIOS.
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
If we post a list of length exactly a multiple of 256, nreq in
doorbell gets set to 256 which is wrong: it should be encoded by 0.
This is because we only zero it out on the next WR, which may not be
there. The solution is to ring the doorbell after posting a WQE, not
before posting the next one.
This is the same bug that we just fixed for QPs with non-shared RQ.
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
If we post a list of length 256 exactly, nreq in doorbell gets set to
256 which is wrong: it should be encoded by 0. This is because we
only zero it out on the next WR, which may not be there. The solution
is to ring the doorbell after posting a WQE, not before posting the
next one.
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Setting fw_cmd_doorbell allows FW command to be queued using posted
writes instead of requiring polling on a "go" bit, so it should be a
performance boost. However, the option causes problems with at least
some device/firmware combinations, so set the default to 0 until we
understand what's going on better.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Addresses for ioremap must be calculated off of pci_resource_start;
we can't directly use the bus address as seen by the HCA. Fix the
code that remaps device memory for FMR access.
Based on patch by Klaus Smolin.
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Fix races in in destroying various objects. If a destroy routine
waits for an object to become free by doing
wait_event(&obj->wait, !atomic_read(&obj->refcount));
/* now clean up and destroy the object */
and another place drops a reference to the object by doing
if (atomic_dec_and_test(&obj->refcount))
wake_up(&obj->wait);
then this is susceptible to a race where the wait_event() and final
freeing of the object occur between the atomic_dec_and_test() and the
wake_up(). And this is a use-after-free, since wake_up() will be
called on part of the already-freed object.
Fix this in mthca by replacing the atomic_t refcounts with plain old
integers protected by a spinlock. This makes it possible to do the
decrement of the reference count and the wake_up() so that it appears
as a single atomic operation to the code waiting on the wait queue.
While touching this code, also simplify mthca_cq_clean(): the CQ being
cleaned cannot go away, because it still has a QP attached to it. So
there's no reason to be paranoid and look up the CQ by number; it's
perfectly safe to use the pointer that the callers already have.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
GuidInfo records have 8 byte GUIDs in them, so an index should be
multiplied by 8 to get an offset. mthca_query_gid() was incorrectly
multiplying by 16.
Noticed by Leonid Keller <leonid@mellanox.co.il>.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
This patch makes the needlessly global mthca_update_rate() static.
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
The driver allocates SRQ WQEs size with a power of 2 size both for
Tavor and for memfree. For Tavor, however, the hardware only requires
the WQE size to be a multiple of 16, not a power of 2, and the max
number of scatter-gather allowed is reported accordingly by the
firmware (and this is the value currently returned by
ib_query_device() and ibv_query_device()).
If the max number of scatter/gather entries reported by the FW is used
when creating an SRQ, the creation will fail for Tavor, since the
required WQE size will be increased to the next power of 2, which
turns out to be larger than the device permitted max WQE size (which
is not a power of 2).
This patch reduces the reported SRQ max wqe size so that it can be used
successfully in creating an SRQ on Tavor HCAs.
Signed-off-by: Jack Morgenstein <jackm@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
The PCI spec recommends against drivers playing with a device's PCI
read burst size, and says that systems software should configure it.
And we actually have users that report that changing it from the
default set by BIOS hurts performance and/or stability for them. On
the other hand, the Mellanox Programmer's Reference Manual recommends
turning it up all the way to the maximum value. Some tests conducted
here in the lab do not show performance improvement from this tuning,
but this might be just me.
As a work-around, make this tuning an option, off by default (safe
value), with an eye towards removing it completely one day if no one
complains.
Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Push translation of static rate to HCA format into low-level drivers,
where it belongs. For static rate encoding, use encoding of rate
field from IB standard PathRecord, with addition of value 0, for
backwards compatibility with current usage. The changes are:
- Add enum ib_rate to midlayer includes.
- Get rid of static rate translation in IPoIB; just use static rate
directly from Path and MulticastGroup records.
- Update mthca driver to translate absolute static rate into the
format used by hardware. This also fixes mthca's static rate
handling for HCAs that are capable of 4X DDR.
Signed-off-by: Jack Morgenstein <jackm@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Change the mthca debugging trace output code so that it can enabled
and disabled at runtime with the debug_level module parameter in
sysfs. Also, don't allow CONFIG_INFINIBAND_MTHCA_DEBUG to be disabled
unless CONFIG_EMBEDDED is selected. We want users (and especially
distros) to have this turned on unless they really need to save space,
because by the time we want debugging output, it's usually too late to
rebuild a kernel.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Quite a few cleanup functions in mthca were marked as __devexit.
However, they could also be called from error paths during
initialization, so they cannot be marked that way. Just delete all of
the incorrect annotations.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
