28 Apr, 2007
1 commit
-
* 'for-linus' of master.kernel.org:/pub/scm/linux/kernel/git/roland/infiniband: (49 commits)
IB: Set class_dev->dev in core for nice device symlink
IB/ehca: Implement modify_port
IB/umad: Clarify documentation of transaction ID
IPoIB/cm: spin_lock_irqsave() -> spin_lock_irq() replacements
IB/mad: Change SMI to use enums rather than magic return codes
IB/umad: Implement GRH handling for sent/received MADs
IB/ipoib: Use ib_init_ah_from_path to initialize ah_attr
IB/sa: Set src_path_bits correctly in ib_init_ah_from_path()
IB/ucm: Simplify ib_ucm_event()
RDMA/ucma: Simplify ucma_get_event()
IB/mthca: Simplify CQ cleaning in mthca_free_qp()
IB/mthca: Fix mthca_write_mtt() on HCAs with hidden memory
IB/mthca: Update HCA firmware revisions
IB/ipath: Fix WC format drift between user and kernel space
IB/ipath: Check that a UD work request's address handle is valid
IB/ipath: Remove duplicate stuff from ipath_verbs.h
IB/ipath: Check reserved memory keys
IB/ipath: Fix unit selection when all CPU affinity bits set
IB/ipath: Don't allow QPs 0 and 1 to be opened multiple times
IB/ipath: Disable IB link earlier in shutdown sequence
...
26 Apr, 2007
6 commits
-
To clearly state the intent of copying from linear sk_buffs, _offset being a
overly long variant but interesting for the sake of saving some bytes.Signed-off-by: Arnaldo Carvalho de Melo
-
Now to convert the last one, skb->data, that will allow many simplifications
and removal of some of the offset helpers.Signed-off-by: Arnaldo Carvalho de Melo
Signed-off-by: David S. Miller -
So that it is also an offset from skb->head, reduces its size from 8 to 4 bytes
on 64bit architectures, allowing us to combine the 4 bytes hole left by the
layer headers conversion, reducing struct sk_buff size to 256 bytes, i.e. 4
64byte cachelines, and since the sk_buff slab cache is SLAB_HWCACHE_ALIGN...
:-)Many calculations that previously required that skb->{transport,network,
mac}_header be first converted to a pointer now can be done directly, being
meaningful as offsets or pointers.Signed-off-by: Arnaldo Carvalho de Melo
Signed-off-by: David S. Miller -
For the common, open coded 'skb->h.raw = skb->data' operation, so that we can
later turn skb->h.raw into a offset, reducing the size of struct sk_buff in
64bit land while possibly keeping it as a pointer on 32bit.This one touches just the most simple cases:
skb->h.raw = skb->data;
skb->h.raw = {skb_push|[__]skb_pull}()The next ones will handle the slightly more "complex" cases.
Signed-off-by: Arnaldo Carvalho de Melo
Signed-off-by: David S. Miller -
For the common, open coded 'skb->mac.raw = skb->data' operation, so that we can
later turn skb->mac.raw into a offset, reducing the size of struct sk_buff in
64bit land while possibly keeping it as a pointer on 32bit.This one touches just the most simple case, next will handle the slightly more
"complex" cases.Signed-off-by: Arnaldo Carvalho de Melo
Signed-off-by: David S. Miller -
One less thing for drivers writers to worry about.
Signed-off-by: Arnaldo Carvalho de Melo
Signed-off-by: David S. Miller
25 Apr, 2007
11 commits
-
All RDMA drivers except ehca set class_dev->dev to their dma_device
value (ehca leaves this unset). dma_device is the only value that
makes any sense, so move this assignment to core/sysfs.c. This reduce
the duplicated code in the rest of the drivers and gives ehca a nice
/sys/class/infiniband/ehcaX/device symlink.Signed-off-by: Joachim Fenkes
Signed-off-by: Roland Dreier -
Add "Modify Port" verb support to eHCA driver. The IB communication
manager needs this to set the IsCM port capability bit when
initializing.Signed-off-by: Joachim Fenkes
Signed-off-by: Roland Dreier -
There are quite a few places in ipoib_cm.c where we know IRQs are
enabled because we do something that sleeps in the same function, so
we can convert several occurrences of spin_lock_irqsave() to a plain
spin_lock_irq(). This cleans up the source a little and makes the
code smaller too:add/remove: 0/0 grow/shrink: 1/5 up/down: 3/-51 (-48)
function old new delta
ipoib_cm_tx_reap 403 406 +3
ipoib_cm_stale_task 146 145 -1
ipoib_cm_dev_stop 173 172 -1
ipoib_cm_tx_handler 964 956 -8
ipoib_cm_rx_handler 956 937 -19
ipoib_cm_skb_reap 212 190 -22Signed-off-by: Roland Dreier
-
Clarify code by changing return values from SMI functions to named
enum values instead of magic 0/1 values.Signed-off-by: Hal Rosenstock
Signed-off-by: Roland Dreier -
We need to set the SGID index for routed MADs and pass received
GRH information to userspace when a MAD is received.Signed-off-by: Sean Hefty
-
To support destinations that are not on the local IB subnet, IPoIB
should include the GRH information when constructing an address
handle. Using the existing ib_init_ah_from_path() call will do this
for us.Signed-off-by: Sean Hefty
-
src_path_bits needs to mask off the base LID value.
Signed-off-by: Sean Hefty
-
Use wait_event_interruptible() instead of a more complicated
open-coded equivalent.Signed-off-by: Sean Hefty
-
Use wait_event_interruptible() instead of a more complicated
open-coded equivalent.Signed-off-by: Sean Hefty
-
mthca_free_qp() already has local variables to hold the QP's send_cq
and recv_cq, so we can slightly clean up the calls to mthca_cq_clean()
by using those local variables instead of expressions like
to_mcq(qp->ibqp.send_cq).Also, by cleaning the recv_cq first, we can avoid worrying about
whether the QP is attached to an SRQ for the second call, because we
would only clean send_cq if send_cq is not equal to recv_cq, and that
means send_cq cannot have any receive completions from the QP being
destroyed.All this work even improves the generated code a bit:
add/remove: 0/0 grow/shrink: 0/1 up/down: 0/-5 (-5)
function old new delta
mthca_free_qp 510 505 -5Signed-off-by: Roland Dreier
-
Commit b2875d4c ("IB/mthca: Always fill MTTs from CPU") causes a crash
in mthca_write_mtt() with non-memfree HCAs that have their memory
hidden (that is, have only two PCI BARs instead of having a third BAR
that allows access to the RAM attached to the HCA) on 64-bit
architectures. This is because the commit just before, c20e20ab
("IB/mthca: Merge MR and FMR space on 64-bit systems") makes
dev->mr_table.fmr_mtt_buddy equal to &dev->mr_table.mtt_buddy and
hence mthca_write_mtt() tries to write directly into the HCA's MTT
table. However, since that table is in the HCA's memory, this is
impossible without the PCI BAR that gives access to that memory.This causes a crash because mthca_tavor_write_mtt_seg() basically
tries to dereference some offset of a NULL pointer. Fix this by
adding a test of MTHCA_FLAG_FMR in mthca_write_mtt() so that we always
use the WRITE_MTT firmware command rather than writing directly if
FMRs are not enabled.Signed-off-by: Roland Dreier
19 Apr, 2007
22 commits
-
Update the driver's list of current firmware versions with Mellanox's
latest releases.Signed-off-by: Roland Dreier
-
The kernel ib_wc structure now uses a QP pointer, but the user space
equivalent uses a QP number instead. This means we can no longer use
a simple structure copy to copy stuff into user space.Signed-off-by: Bryan O'Sullivan
Signed-off-by: Roland Dreier -
Signed-off-by: Bryan O'Sullivan
Signed-off-by: Roland Dreier -
Signed-off-by: Bryan O'Sullivan
Signed-off-by: Roland Dreier -
Don't let userspace use the direct-physical-map L_key or R_key.
Signed-off-by: Ralph Campbell
Signed-off-by: Roland Dreier -
At some point things changed so that all the affinity bits can be set,
but cpus_full() macro is not true. This caused problems with the unit
selection logic on multi-unit (board) configurations.Signed-off-by: Dave Olson
Signed-off-by: Bryan O'Sullivan
Signed-off-by: Roland Dreier -
Signed-off-by: Robert Walsh
Signed-off-by: Bryan O'Sullivan
Signed-off-by: Roland Dreier -
Move the code that shuts down the IB link earlier in the unload
process, to be sure no new packets can arrive while we are unloading.Signed-off-by: Dave Olson
Signed-off-by: Bryan O'Sullivan
Signed-off-by: Roland Dreier -
To prevent random utility reads and writes of the diag interface to the
chip, we first require a handshake of reading from offset 0 and writing
to offset 0 before any other reads or writes can be done through the
diags device. Otherwise chip errors can be triggered.Signed-off-by: Dave Olson
Signed-off-by: Bryan O'Sullivan
Signed-off-by: Roland Dreier -
If the chip is no longer usable, LEDs should be turned off so system
can be found easily in the cluster.Also some minor reorganizing so both chips print hardware error
message at same point and only if there were unrecovered errorsSigned-off-by: Dave Olson
Signed-off-by: Bryan O'Sullivan
Signed-off-by: Roland Dreier -
Re-init of the kernel structures after a chip reset was leaving the
portdata structure for port zero in an inconsistent state, and a
pointer to it either stale (in re-init code) or NULL (in devdata)
Fixing the order of operations on this struct, and the condition for
interrupt access, prevents the crashes.Signed-off-by: Bryan O'Sullivan
Signed-off-by: Roland Dreier -
Mostly cleanup.
Signed-off-by: Dave Olson
Signed-off-by: Bryan O'Sullivan
Signed-off-by: Roland Dreier -
Signed-off-by: Bryan O'Sullivan
Signed-off-by: Roland Dreier -
Due to a chip bug, the PIOAvail register is not always updated to
memory. This patch allows userspace to force an update.Signed-off-by: Bryan O'Sullivan
Signed-off-by: Roland Dreier -
In initialization, if we bailed at chip specific initialization, we
forgot to clean up the irq we had requested.Signed-off-by: Bryan O'Sullivan
Signed-off-by: Roland Dreier -
This patch fixes a bug where multicast packets without a GRH were not
being dropped as per the IB spec.Signed-off-by: Ralph Campbell
Signed-off-by: Bryan O'Sullivan
Signed-off-by: Roland Dreier -
If the module parameter "kpiobufs" is set too high, the calculation to
reset it to a sane value was incorrect.Signed-off-by: Ralph Campbell
Signed-off-by: Bryan O'Sullivan
Signed-off-by: Roland Dreier -
Signed-off-by: Dave Olson
Signed-off-by: Bryan O'Sullivan
Signed-off-by: Roland Dreier -
Fix RDMA read response length checking for RDMA_READ_RESPONSE_ONLY to
allow a zero length response. RDMA read responses which don't match
the expected length or occur in response to some other operation
should generate a completion queue error (see table 56, ch. 9.9.2.3 in
the IB spec).Signed-off-by: Bryan O'Sullivan
Signed-off-by: Roland Dreier -
Improve port-sharing performance by allowing any process to receive
packets from the shared hardware port under a spin lock for mutual
exclusion. Previously, one process was nominated as the master and
that process was responsible for receiving all packets from the shared
hardware port and either consuming them or forwarding them to their
destination. This led to starvation problems for other processes when
the master process was busy in computation phases.Signed-off-by: Bryan O'Sullivan
Signed-off-by: Roland Dreier -
The port sharing feature mixed kernel virtual addresses as well as
physical addresses for the offset used to describe the mmap address to
map the InfiniPath hardware into user space. This had a conflict on
powerpc. The new scheme converts it to a physical address so it
doesn't conflict with chip addresses and yet still fits in 40/44 bits
so it isn't truncated by 32-bit applications calling mmap64().Signed-off-by: Bryan O'Sullivan
Signed-off-by: Roland Dreier -
If a receive work request has been removed from the queue but has not
had a CQ entry generated for it and the QP is modified to the error
state, the completion entry generated is incorrect. This patch fixes
the problem.Signed-off-by: Ralph Campbell
Signed-off-by: Bryan O'Sullivan
Signed-off-by: Roland Dreier
