PIO/SDMA send logic now uses the hdr_type field to determine
the type of packet that has been constructed. Based on the hdr_type,
certain things such as PBC flags, padding count and the LT extra
trailing bytes are determined.
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Dasaratharaman Chandramouli <dasaratharaman.chandramouli@intel.com>
Signed-off-by: Don Hiatt <don.hiatt@intel.com>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
We introduce a struct hfi1_16b_header to support 16B headers.
16B bypass packets are received by the driver and processed
similar to 9B packets. Add basic support to handle 16B packets.
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Don Hiatt <don.hiatt@intel.com>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
There is a need to forward FW version to user space
application through RDMA netlink. In order to make it safe, there
is need to declare nla_policy and limit the size of FW string.
The new define IB_FW_VERSION_NAME_MAX will limit the size of
FW version string. That define was chosen to be equal to
ETHTOOL_FWVERS_LEN, because many drivers anyway are limited
by that value indirectly.
The introduction of this define allows us to remove the string size
from get_fw_str function signature.
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
When an egress resource(SDMA descriptors, pio credits) is not available,
a sending thread will be put on the resource's wait queue. When the
resource becomes available again, up to a fixed number of sending threads
can be awakened sequentially and removed from the wait queue, depending
on the number of waiting threads and the number of free resources. Since
each awakened sending thread will send as many packets as possible, it
is highly likely that the first sending thread will consume all the
egress resources. Subsequently, it will be put back to the end of the wait
queue. Depending on the timing when the later sending threads wake up,
they may not be able to send any packet and be again put back to the end
of the wait queue sequentially, right behind the first sending thread.
This starvation cycle continues until some sending threads exceed their
retry limit and consequently fail.
This patch fixes the issue by two simple approaches:
(1) Any starved sending thread will be put to the head of the wait queue
while a served sending thread will be put to the tail;
(2) The most starved sending thread will be served first.
Reviewed-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Signed-off-by: Kaike Wan <kaike.wan@intel.com>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
IPOIB is calling free_rdma_netdev even though alloc_rdma_netdev has
returned -EOPNOTSUPP.
Move free_rdma_netdev from ib_device structure to rdma_netdev structure
thus ensuring proper cleanup function is called for the rdma net device.
Fix the following trace:
ib0: Failed to modify QP to ERROR state
BUG: unable to handle kernel paging request at 0000000000001d20
IP: hfi1_vnic_free_rn+0x26/0xb0 [hfi1]
Call Trace:
ipoib_remove_one+0xbe/0x160 [ib_ipoib]
ib_unregister_device+0xd0/0x170 [ib_core]
rvt_unregister_device+0x29/0x90 [rdmavt]
hfi1_unregister_ib_device+0x1a/0x100 [hfi1]
remove_one+0x4b/0x220 [hfi1]
pci_device_remove+0x39/0xc0
device_release_driver_internal+0x141/0x200
driver_detach+0x3f/0x80
bus_remove_driver+0x55/0xd0
driver_unregister+0x2c/0x50
pci_unregister_driver+0x2a/0xa0
hfi1_mod_cleanup+0x10/0xf65 [hfi1]
SyS_delete_module+0x171/0x250
do_syscall_64+0x67/0x150
entry_SYSCALL64_slow_path+0x25/0x25
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
Calls to trace incoming packets will now receive the packet
context as parameter. This enables trace support for future
packet types.
Header trace output is in the format <field>:<value>
which makes parsing easier.
input_ibhdr trace before change:
<idle>-0 [001] d.h. 5904.250925: input_ibhdr: [0000:05:00.0] vl 0
lver 0 sl 0 lnh 2,LRH_BTH dlid 0002 len 18 slid 0001 op
0x64,UD_SEND_ONLY se 0 m 0 pad 0 tver 0 pkey 0xffff f 0 b 0 qpn 0x000001
a 0 psn 0x000001b2 deth qkey 0x80010000 sqpn 0x000001
input_ibhdr trace after change:
<idle>-0 [001] d.h. 6655.714488: input_ibhdr: [0000:05:00.0] (IB)
len:124 sc:0 dlid:0x0001 slid:0x0002 lnh:2,LRH_BTH lver:0 sl:0 age:0
becn:0 fecn:0 l4:0 rc:0 entropy:0 op:0x64,UD_SEND_ONLY se:0 m:0 pad:0
tver:0 pkey:0x7fff f:0 b:0 qpn:0x000001 a:0 psn:0x00000036 hlen:8 deth
qkey:0x80010000 sqpn:0x000001
Reviewed-by: Dasaratharaman Chandramouli <dasaratharaman.chandramouli@intel.com>
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Don Hiatt <don.hiatt@intel.com>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
rdma_ah_attr can now be either ib or roce allowing
core components to use one type or the other and also
to define attributes unique to a specific type. struct
ib_ah is also initialized with the type when its first
created. This ensures that calls such as modify_ah
dont modify the type of the address handle attribute.
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Reviewed-by: Don Hiatt <don.hiatt@intel.com>
Reviewed-by: Sean Hefty <sean.hefty@intel.com>
Reviewed-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
Signed-off-by: Dasaratharaman Chandramouli <dasaratharaman.chandramouli@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
The driver progress routines can call cond_resched() when
a timeslice is exhausted and irqs are enabled.
If the ULP had been holding a spin lock without disabling irqs and
the post send directly called the progress routine, the cond_resched()
could yield allowing another thread from the same ULP to deadlock
on that same lock.
Correct by replacing the current hfi1_do_send() calldown with a unique
one for post send and adding an argument to hfi1_do_send() to indicate
that the send engine is running in a thread. If the routine is not
running in a thread, avoid calling cond_resched().
CC: <stable@vger.kernel.org> # 4.7.x-
Fixes: Commit 831464ce4b ("IB/hfi1: Don't call cond_resched in atomic mode when sending packets")
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
The Infiniband spec defines "A multicast address is defined by a
MGID and a MLID" (section 10.5).
The current code only uses the MGID for identifying multicast groups.
Update the driver to be compliant with this definition.
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Reviewed-by: Dasaratharaman Chandramouli <dasaratharaman.chandramouli@intel.com>
Signed-off-by: Michael J. Ruhl <michael.j.ruhl@intel.com>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
HFI1 HW specific support for VNIC functionality.
Dynamically allocate a set of contexts for VNIC when the first vnic
port is instantiated. Allocate VNIC contexts from user contexts pool
and return them back to the same pool while freeing up. Set aside
enough MSI-X interrupts for VNIC contexts and assign them when the
contexts are allocated. On the receive side, use an RSM rule to
spread TCP/UDP streams among VNIC contexts.
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
Signed-off-by: Andrzej Kacprowski <andrzej.kacprowski@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
Add ability to fault packets on transmit by opcode.
Dropping by packet can be achieved by setting the mask to 0.
In order to drop non-verbs traffic we set PbcInsertHrc
to NONE (0x2). The packet will still be delivered to
the receiving node but a KHdrHCRCErr (KDETH packet
with a bad HCRC) will be triggered and the packet will
not be delivered to the correct context.
In order to drop regular verbs traffic we set the
PbcTestEbp flag. The packet will still be delivered
to the receiving node but a 'late ebp error' will
be triggered and will be dropped.
A global toggle (/sys/kernel/debug/hfi1/hfi1_X/fault_suppress_err)
has been added to suppress the error messages on the receive
node when a packet was faulted on the sending node.
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Signed-off-by: Don Hiatt <don.hiatt@intel.com>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
Add fault injection capability:
- Drop packets unconditionally (fault_by_packet)
- Drop packets based on opcode (fault_by_opcode)
This feature reacts to the global FAULT_INJECTION
config flag.
The faulting traces have been added:
- misc/fault_opcode
- misc/fault_packet
See 'Documentation/fault-injection/fault-injection.txt'
for details.
Examples:
- Dropping packets by opcode:
/sys/kernel/debug/hfi1/hfi1_X/fault_opcode
# Enable fault
echo Y > fault_by_opcode
# Setprobability of dropping (0-100%)
# echo 25 > probability
# Set opcode
echo 0x64 > opcode
# Number of times to fault
echo 3 > times
# An optional mask allows you to fault
# a range of opcodes
echo 0xf0 > mask
/sys/kernel/debug/hfi1/hfi1_X/fault_stats
contains a value in parentheses to indicate
number of each opcode dropped.
- Dropping packets unconditionally
/sys/kernel/debug/hfi1/hfi1_X/fault_packet
# Enable fault
echo Y > fault_by_packet
/sys/kernel/debug/hfi1/hfi1_X/fault_packet/fault_stats
contains the number of packets dropped.
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Signed-off-by: Don Hiatt <don.hiatt@intel.com>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
The work to create a completion helper moved the translation of send
wqe operations to completion opcodes to rdmvat.
This precludes having driver dependent operations. Make the translation
driver dependent by doing the translation in the driver prior to the
rvt_qp_swqe_complete() call using restored translation tables.
Fixes: Commit f2dc9cdce8 ("IB/rdmavt: Add a send completion helper")
Fixes: Commit 0771da5a6e ("IB/hfi1,IB/qib: Use new send completion helper")
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
Pull rdma DMA mapping updates from Doug Ledford:
"Drop IB DMA mapping code and use core DMA code instead.
Bart Van Assche noted that the ib DMA mapping code was significantly
similar enough to the core DMA mapping code that with a few changes it
was possible to remove the IB DMA mapping code entirely and switch the
RDMA stack to use the core DMA mapping code.
This resulted in a nice set of cleanups, but touched the entire tree
and has been kept separate for that reason."
* tag 'for-next-dma_ops' of git://git.kernel.org/pub/scm/linux/kernel/git/dledford/rdma: (37 commits)
IB/rxe, IB/rdmavt: Use dma_virt_ops instead of duplicating it
IB/core: Remove ib_device.dma_device
nvme-rdma: Switch from dma_device to dev.parent
RDS: net: Switch from dma_device to dev.parent
IB/srpt: Modify a debug statement
IB/srp: Switch from dma_device to dev.parent
IB/iser: Switch from dma_device to dev.parent
IB/IPoIB: Switch from dma_device to dev.parent
IB/rxe: Switch from dma_device to dev.parent
IB/vmw_pvrdma: Switch from dma_device to dev.parent
IB/usnic: Switch from dma_device to dev.parent
IB/qib: Switch from dma_device to dev.parent
IB/qedr: Switch from dma_device to dev.parent
IB/ocrdma: Switch from dma_device to dev.parent
IB/nes: Remove a superfluous assignment statement
IB/mthca: Switch from dma_device to dev.parent
IB/mlx5: Switch from dma_device to dev.parent
IB/mlx4: Switch from dma_device to dev.parent
IB/i40iw: Remove a superfluous assignment statement
IB/hns: Switch from dma_device to dev.parent
...
Pull Mellanox rdma updates from Doug Ledford:
"Mellanox specific updates for 4.11 merge window
Because the Mellanox code required being based on a net-next tree, I
keept it separate from the remainder of the RDMA stack submission that
is based on 4.10-rc3.
This branch contains:
- Various mlx4 and mlx5 fixes and minor changes
- Support for adding a tag match rule to flow specs
- Support for cvlan offload operation for raw ethernet QPs
- A change to the core IB code to recognize raw eth capabilities and
enumerate them (touches non-Mellanox code)
- Implicit On-Demand Paging memory registration support"
* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dledford/rdma: (40 commits)
IB/mlx5: Fix configuration of port capabilities
IB/mlx4: Take source GID by index from HW GID table
IB/mlx5: Fix blue flame buffer size calculation
IB/mlx4: Remove unused variable from function declaration
IB: Query ports via the core instead of direct into the driver
IB: Add protocol for USNIC
IB/mlx4: Support raw packet protocol
IB/mlx5: Support raw packet protocol
IB/core: Add raw packet protocol
IB/mlx5: Add implicit MR support
IB/mlx5: Expose MR cache for mlx5_ib
IB/mlx5: Add null_mkey access
IB/umem: Indicate that process is being terminated
IB/umem: Update on demand page (ODP) support
IB/core: Add implicit MR flag
IB/mlx5: Support creation of a WQ with scatter FCS offload
IB/mlx5: Enable QP creation with cvlan offload
IB/mlx5: Enable WQ creation and modification with cvlan offload
IB/mlx5: Expose vlan offloads capabilities
IB/uverbs: Enable QP creation with cvlan offload
...
gcc-7 produces a mysterious warning about the size argument being potentially out
of range:
drivers/infiniband/hw/hfi1/verbs.c: In function 'init_cntr_names':
drivers/infiniband/hw/hfi1/verbs.c:1644:2: error: 'memcpy': specified size between 18446744071562067968 and 18446744073709551615 exceeds maximum object size 9223372036854775807 [-Werror=stringop-overflow=]
This seems to refer to a the case where an 64-bit size_t gets truncated
into a negative 'int' and subsequently turned into a high 64-bit number
again.
The fix is clearly to use size_t here, which matches the type that gets
used for this value elsewhere.
Fixes: b7481944b0 ("IB/hfi1: Show statistics counters under IB stats interface")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
Convert copy_sge and related SGE state functions to use boolean.
For determining if QP is in user mode, add helper function in rdmavt_qp.h.
This is used to determine if QP needs the last byte ordering.
While here, change rvt_pd.user to a boolean.
Reviewed-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Reviewed-by: Dean Luick <dean.luick@intel.com>
Signed-off-by: Brian Welty <brian.welty@intel.com>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
Receive code paths use the QP's device and port
number to access the struct hfi1_ibport. When an
instance of struct hfi1_ctxtdata is present, it can
be used to access struct hfi1_ibport through a pointer.
This makes struct hfi1_ibport lookup time faster as an
array doesn't have to be indexed and access fields in
other cache-lines.
Reviewed-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Signed-off-by: Sebastian Sanchez <sebastian.sanchez@intel.com>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
Change the drivers to call ib_query_port in their get port
immutable handler instead of their own query port handler.
Doing this required to set the core cap flags of this device
before the ib_query_port call is made, since the IB core might
need these caps to serve the port query.
Drivers are ensured by the IB core that the port attributes passed
to the port query verb implementation are zero, and hence we
removed the zeroing from the drivers.
This patch doesn't add any new functionality.
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Reviewed-by: Matan Barak <matanb@mellanox.com>
Signed-off-by: Leon Romanovsky <leon@kernel.org>
Reviewed-by: Steve Wise <swise@opengridcomputing.com>
Acked-by: Adit Ranadive <aditr@vmware.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
The s_cur_sge field in the qp structure holds a pointer to the
SGE of the currently processed WQE. It assumes the protection
of the RVT_S_BUSY flag to prevent the changing of this field
while the send engine is using it. This scheme works as long
as there is only one instance of the send engine running at a
time.
Scaling of the send engine to multiple cores would break this
assumption as there could be multiple instances of the send engine
running on different CPUs. This opens a window where the QP's
RVT_S_BUSY flag is not set but the send engine is still running.
To prevent accidental changing of the s_cur_sge pointer, the QP's
dependence on it is removed. The SGE pointer is now stored in the
verbs_txreq, which is a per-packet data structure. This ensures
that each individual packet has it's own pointer, which is setup
while the RVT_S_BUSY flag is set.
Reviewed-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Mitko Haralanov <mitko.haralanov@intel.com>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
The qp->s_cur_size field assumes that the S_BUSY bit protects
the field from modification after the slock is dropped. Scaling the
send engine to multiple cores would break that assumption.
Correct the issue by carrying the payload size in the txreq structure.
Reviewed-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Signed-off-by: Don Hiatt <don.hiatt@intel.com>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
Previously tools like hfi1stats had to access these counters through
debugfs, which often caused permission issue for non-root users. It is
not always acceptable to change the debugfs mounting permission due
to security concerns. When exposed under the IB stats interface, the
counters are universally readable by default.
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Jianxin Xiong <jianxin.xiong@intel.com>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
Profiling suggests that the read_seqbegin() in
the txreq put logic is colliding with other uses
of the iowait lock.
The packet at a time use of this lock dictates a unique
lock to avoid reader/writer collisions when the number
of vTxWait events is low.
In order to support a unique lock the iowait struct embedded
in the QP is extended to remember the lock that protects the queue
head.
The QP destroy removes that QP from any wait list. It doesn't
need to know the head because of the linked list API, but it does
need to know the lock required to protect the head.
This also opens up the wait logic to have unique per resources locks
which needs to be in future refinement.
Reviewed-by: Sebastian Sanchez <sebastian.sanchez@intel.com>
Signed-off-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
Pull main rdma updates from Doug Ledford:
"This is the main pull request for the rdma stack this release. The
code has been through 0day and I had it tagged for linux-next testing
for a couple days.
Summary:
- updates to mlx5
- updates to mlx4 (two conflicts, both minor and easily resolved)
- updates to iw_cxgb4 (one conflict, not so obvious to resolve,
proper resolution is to keep the code in cxgb4_main.c as it is in
Linus' tree as attach_uld was refactored and moved into
cxgb4_uld.c)
- improvements to uAPI (moved vendor specific API elements to uAPI
area)
- add hns-roce driver and hns and hns-roce ACPI reset support
- conversion of all rdma code away from deprecated
create_singlethread_workqueue
- security improvement: remove unsafe ib_get_dma_mr (breaks lustre in
staging)"
* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dledford/rdma: (75 commits)
staging/lustre: Disable InfiniBand support
iw_cxgb4: add fast-path for small REG_MR operations
cxgb4: advertise support for FR_NSMR_TPTE_WR
IB/core: correctly handle rdma_rw_init_mrs() failure
IB/srp: Fix infinite loop when FMR sg[0].offset != 0
IB/srp: Remove an unused argument
IB/core: Improve ib_map_mr_sg() documentation
IB/mlx4: Fix possible vl/sl field mismatch in LRH header in QP1 packets
IB/mthca: Move user vendor structures
IB/nes: Move user vendor structures
IB/ocrdma: Move user vendor structures
IB/mlx4: Move user vendor structures
IB/cxgb4: Move user vendor structures
IB/cxgb3: Move user vendor structures
IB/mlx5: Move and decouple user vendor structures
IB/{core,hw}: Add constant for node_desc
ipoib: Make ipoib_warn ratelimited
IB/mlx4/alias_GUID: Remove deprecated create_singlethread_workqueue
IB/ipoib_verbs: Remove deprecated create_singlethread_workqueue
IB/ipoib: Remove deprecated create_singlethread_workqueue
...
The length is incorrect, causing the trace data to
be truncated.
Add the additional 8 bytes that should have been there.
Also trace out the atomic ack in hex to aid debugging.
Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Signed-off-by: Dennis Dalessandro <dennis.dalessandro@intel.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
