Use a bit in wc_flags rather then a whole integer to hold the
"checksum OK" flag. By itself, this change doesn't reduce the size of
struct ib_wc on 64bit machines -- it stays on 56 bytes because of
padding. However, it will allow to add more fields in the future
without enlarging the struct. Also, it will let us have a unified
approach with future libibverbs checksum offload reporting, because a
bit flag doesn't break the library ABI.
This patch was suggested during conversation with Liran Liss
<liranl@mellanox.com>.
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Reviewed-by: Sean Hefty <sean.hefty@intel.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
* 'modsplit-Oct31_2011' of git://git.kernel.org/pub/scm/linux/kernel/git/paulg/linux: (230 commits)
Revert "tracing: Include module.h in define_trace.h"
irq: don't put module.h into irq.h for tracking irqgen modules.
bluetooth: macroize two small inlines to avoid module.h
ip_vs.h: fix implicit use of module_get/module_put from module.h
nf_conntrack.h: fix up fallout from implicit moduleparam.h presence
include: replace linux/module.h with "struct module" wherever possible
include: convert various register fcns to macros to avoid include chaining
crypto.h: remove unused crypto_tfm_alg_modname() inline
uwb.h: fix implicit use of asm/page.h for PAGE_SIZE
pm_runtime.h: explicitly requires notifier.h
linux/dmaengine.h: fix implicit use of bitmap.h and asm/page.h
miscdevice.h: fix up implicit use of lists and types
stop_machine.h: fix implicit use of smp.h for smp_processor_id
of: fix implicit use of errno.h in include/linux/of.h
of_platform.h: delete needless include <linux/module.h>
acpi: remove module.h include from platform/aclinux.h
miscdevice.h: delete unnecessary inclusion of module.h
device_cgroup.h: delete needless include <linux/module.h>
net: sch_generic remove redundant use of <linux/module.h>
net: inet_timewait_sock doesnt need <linux/module.h>
...
Fix up trivial conflicts (other header files, and removal of the ab3550 mfd driver) in
- drivers/media/dvb/frontends/dibx000_common.c
- drivers/media/video/{mt9m111.c,ov6650.c}
- drivers/mfd/ab3550-core.c
- include/linux/dmaengine.h
The num_free field of mthca_buddy has a type of array of unsigned int
while it was allocated as an array of pointers. On 64-bit platforms
this allocates twice more than required. Fix this by allocating the
correct size for the type.
This is the same bug just fixed in mlx4 by Eli Cohen <eli@mellanox.co.il>.
Signed-off-by: Roland Dreier <roland@purestorage.com>
These were getting it implicitly via device.h --> module.h but
we are going to stop that when we clean up the headers.
Fix these in advance so the tree remains biscect-clean.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
They had been getting it implicitly via device.h but we can't
rely on that for the future, due to a pending cleanup so fix
it now.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
They get it via module.h (via device.h) but we want to clean that up.
When we do, we'll get things like:
CC [M] drivers/infiniband/core/sysfs.o
sysfs.c:361: error: 'S_IRUGO' undeclared here (not in a function)
sysfs.c:654: error: 'S_IWUSR' undeclared here (not in a function)
so add in the stat header it is using explicitly in advance.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Currently, there is only a single ("basic") type of SRQ, but with XRC
support we will add a second. Prepare for this by defining an SRQ type
and setting all current users to IB_SRQT_BASIC.
Signed-off-by: Sean Hefty <sean.hefty@intel.com>
Signed-off-by: Roland Dreier <roland@purestorage.com>
The PCIE capability offset is saved during PCI bus walking. It will
remove an unnecessary search in the PCI configuration space if this
value is referenced instead of reacquiring it. Also, pci_is_pcie is a
better way of determining if the device is PCIE or not (as it uses the
same saved PCIE capability offset).
Signed-off-by: Jon Mason <jdmason@kudzu.us>
Signed-off-by: Roland Dreier <roland@purestorage.com>
By default, each device is assumed to be able only handle 64 KB chunks
during DMA. By giving the segment size a larger value, the block layer
will coalesce more S/G entries together for SRP, allowing larger
requests with the same sg_tablesize setting. The block layer is the
only direct user of it, though a few IOMMU drivers reference it as
well for their *_map_sg coalescing code. pci-gart_64 on x86, and a
smattering on on sparc, powerpc, and ia64.
Since other IB protocols could potentially see larger segments with
this, let's check those:
- iSER is fine, because you limit your maximum request size to 512
KB, so we'll never overrun the page vector in struct iser_page_vec
(128 entries currently). It is independent of the DMA segment size,
and handles multi-page segments already.
- IPoIB is fine, as it maps each page individually, and doesn't use
ib_dma_map_sg().
- RDS appears to do the right thing and has no dependencies on DMA
segment size, but I don't claim to have done a complete audit.
- NFSoRDMA and 9p are OK -- they do not use ib_dma_map_sg(), so they
doesn't care about the coalescing.
- Lustre's ko2iblnd does not care about coalescing -- it properly
walks the returned sg list.
This patch ups the value on Mellanox hardware to 1 GB, which matches
reported firmware limits on mlx4.
Signed-off-by: David Dillow <dillowda@ornl.gov>
Signed-off-by: Roland Dreier <roland@purestorage.com>
The meaning of CONFIG_EMBEDDED has long since been obsoleted; the option
is used to configure any non-standard kernel with a much larger scope than
only small devices.
This patch renames the option to CONFIG_EXPERT in init/Kconfig and fixes
references to the option throughout the kernel. A new CONFIG_EMBEDDED
option is added that automatically selects CONFIG_EXPERT when enabled and
can be used in the future to isolate options that should only be
considered for embedded systems (RISC architectures, SLOB, etc).
Calling the option "EXPERT" more accurately represents its intention: only
expert users who understand the impact of the configuration changes they
are making should enable it.
Reviewed-by: Ingo Molnar <mingo@elte.hu>
Acked-by: David Woodhouse <david.woodhouse@intel.com>
Signed-off-by: David Rientjes <rientjes@google.com>
Cc: Greg KH <gregkh@suse.de>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Robin Holt <holt@sgi.com>
Cc: <linux-arch@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Some systems have PCI addresses that don't fit in unsigned long (eg some
32-bit PowerPC 440 systems have 36-bit bus addresses). Fix up the driver
by using phys_addr_t where appropriate, so we don't truncate any PCI
resource addresses before ioremapping them.
Signed-off-by: John L. Burr <jlburr@cadence.com>
[ Update to apply to current driver source. - Roland ]
Signed-off-by: Roland Dreier <rolandd@cisco.com>
ib_create_send_mad() can return ERR_PTR(-ENOMEM) here.
Signed-off-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Add 802.1q VLAN support to IBoE. The VLAN tag is encoded within the
GID derived from a link local address in the following way:
GID[11] GID[12] contain the VLAN ID when the GID contains a VLAN.
The 3 bits user priority field of the packets are identical to the 3
bits of the SL.
In case of rdma_cm apps, the TOS field is used to generate the SL
field by doing a shift right of 5 bits effectively taking to 3 MS bits
of the TOS field.
Signed-off-by: Eli Cohen <eli@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Add support for packing IBoE packet headers.
Signed-off-by: Eli Cohen <eli@mellanox.co.il>
[ Clean up and fix ib_ud_header_init() a bit. - Roland ]
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Change abbreviated IB_QPT_RAW_ETY to IB_QPT_RAW_ETHERTYPE to make
the special QP type easier to understand.
cf http://www.mail-archive.com/linux-rdma@vger.kernel.org/msg04530.html
Signed-off-by: Aleksey Senin <alekseys@voltaire.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Add a new parameter to ib_register_device() so that low-level device
drivers can pass in a pointer to a callback function that will be
called for each port that is registered in sysfs. This allows
low-level device drivers to create files in
/sys/class/infiniband/<hca>/ports/<N>/
without having to poke through the internals of the RDMA sysfs handling.
There is no need for an unregister function since the kobject
reference will go to zero when ib_unregister_device() is called.
Signed-off-by: Ralph Campbell <ralph.campbell@qlogic.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
The DMA API is preferred; no functional change.
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files. percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.
percpu.h -> slab.h dependency is about to be removed. Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability. As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.
http://userweb.kernel.org/~tj/misc/slabh-sweep.py
The script does the followings.
* Scan files for gfp and slab usages and update includes such that
only the necessary includes are there. ie. if only gfp is used,
gfp.h, if slab is used, slab.h.
* When the script inserts a new include, it looks at the include
blocks and try to put the new include such that its order conforms
to its surrounding. It's put in the include block which contains
core kernel includes, in the same order that the rest are ordered -
alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
doesn't seem to be any matching order.
* If the script can't find a place to put a new include (mostly
because the file doesn't have fitting include block), it prints out
an error message indicating which .h file needs to be added to the
file.
The conversion was done in the following steps.
1. The initial automatic conversion of all .c files updated slightly
over 4000 files, deleting around 700 includes and adding ~480 gfp.h
and ~3000 slab.h inclusions. The script emitted errors for ~400
files.
2. Each error was manually checked. Some didn't need the inclusion,
some needed manual addition while adding it to implementation .h or
embedding .c file was more appropriate for others. This step added
inclusions to around 150 files.
3. The script was run again and the output was compared to the edits
from #2 to make sure no file was left behind.
4. Several build tests were done and a couple of problems were fixed.
e.g. lib/decompress_*.c used malloc/free() wrappers around slab
APIs requiring slab.h to be added manually.
5. The script was run on all .h files but without automatically
editing them as sprinkling gfp.h and slab.h inclusions around .h
files could easily lead to inclusion dependency hell. Most gfp.h
inclusion directives were ignored as stuff from gfp.h was usually
wildly available and often used in preprocessor macros. Each
slab.h inclusion directive was examined and added manually as
necessary.
6. percpu.h was updated not to include slab.h.
7. Build test were done on the following configurations and failures
were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my
distributed build env didn't work with gcov compiles) and a few
more options had to be turned off depending on archs to make things
build (like ipr on powerpc/64 which failed due to missing writeq).
* x86 and x86_64 UP and SMP allmodconfig and a custom test config.
* powerpc and powerpc64 SMP allmodconfig
* sparc and sparc64 SMP allmodconfig
* ia64 SMP allmodconfig
* s390 SMP allmodconfig
* alpha SMP allmodconfig
* um on x86_64 SMP allmodconfig
8. percpu.h modifications were reverted so that it could be applied as
a separate patch and serve as bisection point.
Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
ib_ud_header_init() first clears header and then fills up the various
fields. Later on, it tests header->immediate_present, which it has
already cleared, so the condition is always false. Fix this by adding
an immediate_present parameter and setting header->immediate_present
as is done with grh_present. Also remove unused calculation of
header_len.
Signed-off-by: Eli Cohen <eli@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
catas_reset() uses a pointer to mthca_dev, but mthca_dev is not valid
after the call to __mthca_restart_one().
Based on a similar patch for mlx4 (634354d7, "mlx4: Fix access to
freed memory") by Vitaliy Gusev <vgusev@openvz.org>
Signed-off-by: Jack Morgenstein <jackm@dev.mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Userspace apps are supposed to release all ib device resources if they
receive a fatal async event (IBV_EVENT_DEVICE_FATAL). However, the
app has no way of knowing when the device has come back up, except to
repeatedly attempt ibv_open_device() until it succeeds.
However, currently there is no protection against the open succeeding
while the device is in being removed following the fatal event. In
this case, the open will succeed, but as a result the device waits in
the middle of its removal until the new app releases its resources --
and the new app will not do so, since the open succeeded at a point
following the fatal event generation.
This patch adds an "active" flag to the device. The active flag is set
to false (in the fatal event flow) before the "fatal" event is
generated, so any subsequent ibv_dev_open() call to the device will
fail until the device comes back up, thus preventing the above
deadlock.
Signed-off-by: Jack Morgenstein <jackm@dev.mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
When the mthca driver uses the same name for interrupts for every
device in the system. This can make it very confusing trying to work
out exactly which device MSI-X interrupts are for. Change the driver
to add the PCI name of the device to the interrupt name.
Signed-off-by: Arputham Benjamin <abenjamin@sgi.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
mthca_ib_lock_cqs()/mthca_ib_unlock_cqs() are helper functions that
lock/unlock both CQs attached to a QP in the proper order to avoid
AB-BA deadlocks. Annotate this so sparse can understand what's going
on (and warn us if we misuse these functions).
Signed-off-by: Roland Dreier <rolandd@cisco.com>
mthca_config_reg.h was including <asm/page.h> for no reason -- the whole
file is just defines of constants, so it's entirely self-contained.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Replace open-coded reimplementations with printk_once().
Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
dma_sync_single() is deprecated now, and the use in mthca is wrong:
there should be a dma_sync_single_for_cpu() before touching the memory
from the CPU, and a dma_sync_single_for_device() afterwards. Fix
this, prompted by a kick in the pants from a patch from FUJITA
Tomonori <fujita.tomonori@lab.ntt.co.jp>.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
When both MSI-X and legacy INTx fail to generate an interrupt, the
driver frees the MSI-X interrupts twice. Fix this by clearing the
have_irq flag for the MSI-X interrupts when they are freed the first
time.
Reported-by: Yinghai Lu <yhlu.kernel@gmail.com>
Tested-by: Yinghai Lu <yhlu.kernel@gmail.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
The current MTT allocator uses kmalloc() to allocate a buffer for its
buddy allocator, and thus is limited in the amount of MTT segments
that it can control. As a result, the size of memory that can be
registered is limited too. This patch uses a module parameter to
control the number of MTT entries that each segment represents,
allowing more memory to be registered with the same number of
segments.
Signed-off-by: Eli Cohen <eli@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Commands INIT_HCA, CLOSE_HCA, SYS_EN, SYS_DIS, and CLOSE_IB all have 1
second timeouts. For INIT_HCA this causes problems when had more than
2^18 are QPs configured, since the command takes more than 1 second to
complete.
All other commands have 60-second timeouts. This patch makes the
above commands consistent with the rest of the commands (and with the
chip documentation).
This patch is an expansion of a patch from Arthur Kepner
<akepner@sgi.com> fixing just the INIT_HCA timeout.
Signed-off-by: Jack Morgenstein <jackm@dev.mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Replace all DMA_32BIT_MASK macro with DMA_BIT_MASK(32)
Signed-off-by: Yang Hongyang<yanghy@cn.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Replace all DMA_64BIT_MASK macro with DMA_BIT_MASK(64)
Signed-off-by: Yang Hongyang<yanghy@cn.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
When snooping a PortInfo MAD, its client_reregister bit is checked.
If the bit is ON then a CLIENT_REREGISTER event is dispatched,
otherwise a LID_CHANGE event is dispatched. This way of decision
ignores the cases where the MAD changes the LID along with an
instruction to reregister (so a necessary LID_CHANGE event won't be
dispatched) or the MAD is neither of these (and an unnecessary
LID_CHANGE event will be dispatched).
This causes problems at least with IPoIB, which will do a "light"
flush on reregister, rather than the "heavy" flush required due to a
LID change.
Fix this by dispatching a CLIENT_REREGISTER event if the
client_reregister bit is set, but also compare the LID in the MAD to
the current LID. If and only if they are not identical then a
LID_CHANGE event is dispatched.
Signed-off-by: Moni Shoua <monis@voltaire.com>
Signed-off-by: Jack Morgenstein <jackm@dev.mellanox.co.il>
Signed-off-by: Yossi Etigin <yosefe@voltaire.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Back in prehistoric (pre-git!) days, the kernel's MSI-X support did
request_mem_region() on a device's MSI-X tables, which meant that a
driver that enabled MSI-X couldn't use pci_request_regions() (since
that would clash with the PCI layer's MSI-X request).
However, that was removed (by me!) years ago, so mthca can just use
pci_request_regions() and pci_release_regions() instead of its own
much more complicated code that avoids requesting the MSI-X tables.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Add per-device dma_mapping_ops support for CONFIG_X86_64 as POWER
architecture does:
This enables us to cleanly fix the Calgary IOMMU issue that some devices
are not behind the IOMMU (http://lkml.org/lkml/2008/5/8/423).
I think that per-device dma_mapping_ops support would be also helpful for
KVM people to support PCI passthrough but Andi thinks that this makes it
difficult to support the PCI passthrough (see the above thread). So I
CC'ed this to KVM camp. Comments are appreciated.
A pointer to dma_mapping_ops to struct dev_archdata is added. If the
pointer is non NULL, DMA operations in asm/dma-mapping.h use it. If it's
NULL, the system-wide dma_ops pointer is used as before.
If it's useful for KVM people, I plan to implement a mechanism to register
a hook called when a new pci (or dma capable) device is created (it works
with hot plugging). It enables IOMMUs to set up an appropriate
dma_mapping_ops per device.
The major obstacle is that dma_mapping_error doesn't take a pointer to the
device unlike other DMA operations. So x86 can't have dma_mapping_ops per
device. Note all the POWER IOMMUs use the same dma_mapping_error function
so this is not a problem for POWER but x86 IOMMUs use different
dma_mapping_error functions.
The first patch adds the device argument to dma_mapping_error. The patch
is trivial but large since it touches lots of drivers and dma-mapping.h in
all the architecture.
This patch:
dma_mapping_error() doesn't take a pointer to the device unlike other DMA
operations. So we can't have dma_mapping_ops per device.
Note that POWER already has dma_mapping_ops per device but all the POWER
IOMMUs use the same dma_mapping_error function. x86 IOMMUs use device
argument.
[akpm@linux-foundation.org: fix sge]
[akpm@linux-foundation.org: fix svc_rdma]
[akpm@linux-foundation.org: build fix]
[akpm@linux-foundation.org: fix bnx2x]
[akpm@linux-foundation.org: fix s2io]
[akpm@linux-foundation.org: fix pasemi_mac]
[akpm@linux-foundation.org: fix sdhci]
[akpm@linux-foundation.org: build fix]
[akpm@linux-foundation.org: fix sparc]
[akpm@linux-foundation.org: fix ibmvscsi]
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Cc: Muli Ben-Yehuda <muli@il.ibm.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Avi Kivity <avi@qumranet.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
MTT entries are allocated with a buddy allocator, which just keeps
bitmaps for each level of the buddy table. However, all free space
starts out at the highest order, and small allocations start scanning
from the lowest order. When the lowest order tables have no free
space, this can lead to scanning potentially millions of bits before
finding a free entry at a higher order.
We can avoid this by just keeping a count of how many free entries
each order has, and skipping the bitmap scan when an order is
completely empty. This provides a nice performance boost for a
negligible increase in memory usage.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
The MLX transport requires two extra gather entries for sends (one for
the header and one for the checksum at the end, as the comment says).
However the code checked that max_recv_sge was not too big, instead of
checking max_send_sge as it should have. Fix the code to check the
correct condition.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Exactly when the catastrophic error polling timer function runs is not
important, so use round_jiffies() to save unnecessary wakeups.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Since we use del_timer_sync() anyway, there's no need for an
additional flag to tell the timer not to rearm.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Commit b18aad71 ("IB/mthca: Fix RESET to ERROR transition") added some
extra code to handle a QP state transition from RESET to ERROR.
However, the latest 1.2.1 version of the IB spec has clarified that
this transition is actually not allowed, so we can remove this extra
code again.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
This patch adds support for the IB "base memory management extension"
(BMME) and the equivalent iWARP operations (which the iWARP verbs
mandates all devices must implement). The new operations are:
- Allocate an ib_mr for use in fast register work requests.
- Allocate/free a physical buffer lists for use in fast register work
requests. This allows device drivers to allocate this memory as
needed for use in posting send requests (eg via dma_alloc_coherent).
- New send queue work requests:
* send with remote invalidate
* fast register memory region
* local invalidate memory region
* RDMA read with invalidate local memory region (iWARP only)
Consumer interface details:
- A new device capability flag IB_DEVICE_MEM_MGT_EXTENSIONS is added
to indicate device support for these features.
- New send work request opcodes IB_WR_FAST_REG_MR, IB_WR_LOCAL_INV,
IB_WR_RDMA_READ_WITH_INV are added.
- A new consumer API function, ib_alloc_mr() is added to allocate
fast register memory regions.
- New consumer API functions, ib_alloc_fast_reg_page_list() and
ib_free_fast_reg_page_list() are added to allocate and free
device-specific memory for fast registration page lists.
- A new consumer API function, ib_update_fast_reg_key(), is added to
allow the key portion of the R_Key and L_Key of a fast registration
MR to be updated. Consumers call this if desired before posting
a IB_WR_FAST_REG_MR work request.
Consumers can use this as follows:
- MR is allocated with ib_alloc_mr().
- Page list memory is allocated with ib_alloc_fast_reg_page_list().
- MR R_Key/L_Key "key" field is updated with ib_update_fast_reg_key().
- MR made VALID and bound to a specific page list via
ib_post_send(IB_WR_FAST_REG_MR)
- MR made INVALID via ib_post_send(IB_WR_LOCAL_INV),
ib_post_send(IB_WR_RDMA_READ_WITH_INV) or an incoming send with
invalidate operation.
- MR is deallocated with ib_dereg_mr()
- page lists dealloced via ib_free_fast_reg_page_list().
Applications can allocate a fast register MR once, and then can
repeatedly bind the MR to different physical block lists (PBLs) via
posting work requests to a send queue (SQ). For each outstanding
MR-to-PBL binding in the SQ pipe, a fast_reg_page_list needs to be
allocated (the fast_reg_page_list is owned by the low-level driver
from the consumer posting a work request until the request completes).
Thus pipelining can be achieved while still allowing device-specific
page_list processing.
The 32-bit fast register memory key/STag is composed of a 24-bit index
and an 8-bit key. The application can change the key each time it
fast registers thus allowing more control over the peer's use of the
key/STag (ie it can effectively be changed each time the rkey is
rebound to a page list).
Signed-off-by: Steve Wise <swise@opengridcomputing.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Current memfree FW has a bug which in some cases, assumes that ICM
pages passed to it are cleared. This patch uses __GFP_ZERO to
allocate all ICM pages passed to the FW. Once firmware with a fix is
released, we can make the workaround conditional on firmware version.
This fixes the bug reported by Arthur Kepner <akepner@sgi.com> here:
http://lists.openfabrics.org/pipermail/general/2008-May/050026.html
Cc: <stable@kernel.org>
Signed-off-by: Eli Cohen <eli@mellanox.co.il>
[ Rewritten to be a one-liner using __GFP_ZERO instead of vmap()ing
ICM memory and memset()ing it to 0. - Roland ]
Signed-off-by: Roland Dreier <rolandd@cisco.com>
The mthca driver returns the maximum number of scatter/gather entries
returned by the firmware as the max_sge value when device properties
are queried. However, the firmware also reports a limit on the
maximum descriptor size allowed, and because mthca takes into account
the worst case send request overhead when checking whether to allow a
QP to be created, the largest number of scatter/gather entries that
can be used with mthca may be limited by the maximum descriptor size
rather than just by the actual s/g entry limit.
This means that applications cannot actually create QPs with
max_send_sge equal to the limit returned by ib_query_device(). Fix
this by checking if the maximum descriptor size imposes a lower limit
and if so returning that lower limit.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Commit cb9fbc5c ("IB: expand ib_umem_get() prototype") changed the
mthca userspace ABI to provide a way for userspace to indicate which
memory regions need the DMA write barrier attribute. However, it is
possible to handle this without breaking existing userspace, by having
the mthca kernel driver recognize whether it is talking to old or new
userspace, depending on the size of the register MR structure passed in.
The only potential drawback of this is that is allows old userspace
(which has a bug with DMA ordering on large SGI Altix systems) to
continue to run on new kernels, but the advantage of allowing old
userspace to continue to work on unaffected systems seems to outweigh
this, and we can print a warning to push people to upgrade their
userspace.
Signed-off-by: Roland Dreier <rolandd@cisco.com>
When a FMR is unmapped, mthca resets the map count to 0, and clears
the upper part of the R_Key which is used as the sequence counter.
This poses a problem for RDS, which uses ib_fmr_unmap as a fence
operation. RDS assumes that after issuing an unmap, the old R_Keys
will be invalid for a "reasonable" period of time. For instance,
Oracle processes uses shared memory buffers allocated from a pool of
buffers. When a process dies, we want to reclaim these buffers -- but
we must make sure there are no pending RDMA operations to/from those
buffers. The only way to achieve that is by using unmap and sync the
TPT.
However, when the sequence count is reset on unmap, there is a high
likelihood that a new mapping will be given the same R_Key that was
issued a few milliseconds ago.
To prevent this, don't reset the sequence count when unmapping a FMR.
Signed-off-by: Olaf Kirch <olaf.kirch@oracle.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
