HW flushes after QP in error state is not reliable. This can lead to
application hang waiting on a completion for outstanding WRs. Implement a
SW mechanism to generate completions for any outstanding WR's after the QP
is modified to error.
This is accomplished by starting a delayed worker after the QP is modified
to error and the HW flush is performed. The worker will generate
completions that will be returned to the application when it polls the
CQ. This mechanism only applies to Kernel applications.
Link: https://lore.kernel.org/r/20220425181624.1617-1-shiraz.saleem@intel.com
Signed-off-by: Mustafa Ismail <mustafa.ismail@intel.com>
Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
QP destroy is synchronous and waits for its refcnt to be decremented in
irdma_cm_node_free_cb (for iWARP) which fires after the RCU grace period
elapses.
Applications running a large number of connections are exposed to high
wait times on destroy QP for events like SIGABORT.
The long pole for this wait time is the firing of the call_rcu callback
during a CM node destroy which can be slow. It holds the QP reference
count and blocks the destroy QP from completing.
call_rcu only needs to make sure that list walkers have a reference to the
cm_node object before freeing it and thus need to wait for grace period
elapse. The rest of the connection teardown in irdma_cm_node_free_cb is
moved out of the grace period wait in irdma_destroy_connection. Also,
replace call_rcu with a simple kfree_rcu as it just needs to do a kfree on
the cm_node
Fixes: 146b9756f1 ("RDMA/irdma: Add connection manager")
Link: https://lore.kernel.org/r/20220425181703.1634-3-shiraz.saleem@intel.com
Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
When connection establishment fails in iWARP mode, an app can drain the
QPs and hang because flush isn't issued when the QP is modified from RTR
state to error. Issue a flush in this case using function
irdma_cm_disconn().
Update irdma_cm_disconn() to do flush when cm_id is NULL, which is the
case when the QP is in RTR state and there is an error in the connection
establishment.
Fixes: b48c24c2d7 ("RDMA/irdma: Implement device supported verb APIs")
Link: https://lore.kernel.org/r/20220425181703.1634-2-shiraz.saleem@intel.com
Signed-off-by: Tatyana Nikolova <tatyana.e.nikolova@intel.com>
Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
There is a deadlock in irdma_cleanup_cm_core(), which is shown below:
(Thread 1) | (Thread 2)
| irdma_schedule_cm_timer()
irdma_cleanup_cm_core() | add_timer()
spin_lock_irqsave() //(1) | (wait a time)
... | irdma_cm_timer_tick()
del_timer_sync() | spin_lock_irqsave() //(2)
(wait timer to stop) | ...
We hold cm_core->ht_lock in position (1) of thread 1 and use
del_timer_sync() to wait timer to stop, but timer handler also need
cm_core->ht_lock in position (2) of thread 2. As a result,
irdma_cleanup_cm_core() will block forever.
This patch removes the check of timer_pending() in
irdma_cleanup_cm_core(), because the del_timer_sync() function will just
return directly if there isn't a pending timer. As a result, the lock is
redundant, because there is no resource it could protect.
Link: https://lore.kernel.org/r/20220418153322.42524-1-duoming@zju.edu.cn
Signed-off-by: Duoming Zhou <duoming@zju.edu.cn>
Reviewed-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Split out flags from ib_device::device_cap_flags that are only used
internally to the kernel into kernel_cap_flags that is not part of the
uapi. This limits the device_cap_flags to being the same bitmap that will
be copied to userspace.
This cleanly splits out the uverbs flags from the kernel flags to avoid
confusion in the flags bitmap.
Add some short comments describing which each of the kernel flags is
connected to. Remove unused kernel flags.
Link: https://lore.kernel.org/r/0-v2-22c19e565eef+139a-kern_caps_jgg@nvidia.com
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Max Gurtovoy <mgurtovoy@nvidia.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
My static checker complains that:
drivers/infiniband/hw/irdma/ctrl.c:3605 irdma_sc_ceq_init()
warn: can subtract underflow 'info->dev->hmc_fpm_misc.max_ceqs'?
It appears that "info->dev->hmc_fpm_misc.max_ceqs" comes from the firmware
in irdma_sc_parse_fpm_query_buf() so, yes, there is a chance that it could
be zero. Even if we trust the firmware, it's easy enough to change the
condition just as a hardenning measure.
Fixes: 3f49d68425 ("RDMA/irdma: Implement HW Admin Queue OPs")
Link: https://lore.kernel.org/r/20220307125928.GE16710@kili
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Acked-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Using PCI_FUNC macro in a VM, when the device is in passthrough mode does
not provide the real function instance. This means that currently, devices
will not probe unless the instance in the VM matches the instance in the
host.
Fix this by getting the pf_id from the LAN during the probe.
Fixes: 8498a30e1b ("RDMA/irdma: Register auxiliary driver and implement private channel OPs")
Link: https://lore.kernel.org/r/20220225163211.127-3-shiraz.saleem@intel.com
Signed-off-by: Mustafa Ismail <mustafa.ismail@intel.com>
Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Pull bitmap updates from Yury Norov:
- introduce for_each_set_bitrange()
- use find_first_*_bit() instead of find_next_*_bit() where possible
- unify for_each_bit() macros
* tag 'bitmap-5.17-rc1' of git://github.com/norov/linux:
vsprintf: rework bitmap_list_string
lib: bitmap: add performance test for bitmap_print_to_pagebuf
bitmap: unify find_bit operations
mm/percpu: micro-optimize pcpu_is_populated()
Replace for_each_*_bit_from() with for_each_*_bit() where appropriate
find: micro-optimize for_each_{set,clear}_bit()
include/linux: move for_each_bit() macros from bitops.h to find.h
cpumask: replace cpumask_next_* with cpumask_first_* where appropriate
tools: sync tools/bitmap with mother linux
all: replace find_next{,_zero}_bit with find_first{,_zero}_bit where appropriate
cpumask: use find_first_and_bit()
lib: add find_first_and_bit()
arch: remove GENERIC_FIND_FIRST_BIT entirely
include: move find.h from asm_generic to linux
bitops: move find_bit_*_le functions from le.h to find.h
bitops: protect find_first_{,zero}_bit properly
find_first{,_zero}_bit is a more effective analogue of 'next' version if
start == 0. This patch replaces 'next' with 'first' where things look
trivial.
Signed-off-by: Yury Norov <yury.norov@gmail.com>
Tested-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Pull rdma updates from Jason Gunthorpe:
"Another small cycle. Mostly cleanups and bug fixes, quite a bit
assisted from bots. There are a few new syzkaller splats that haven't
been solved yet but they should get into the rcs in a few weeks, I
think.
Summary:
- Update drivers to use common helpers for GUIDs, pkeys, bitmaps,
memset_startat, and others
- General code cleanups from bots
- Simplify some of the rxe pool code in preparation for a larger
rework
- Clean out old stuff from hns, including all support for hip06
devices
- Fix a bug where GID table entries could be missed if the table had
holes in it
- Rename paths and sessions in rtrs for better understandability
- Consolidate the roce source port selection code
- NDR speed support in mlx5"
* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma: (83 commits)
RDMA/irdma: Remove the redundant return
RDMA/rxe: Use the standard method to produce udp source port
RDMA/irdma: Make the source udp port vary
RDMA/hns: Replace get_udp_sport with rdma_get_udp_sport
RDMA/core: Calculate UDP source port based on flow label or lqpn/rqpn
IB/qib: Fix typos
RDMA/rtrs-clt: Rename rtrs_clt to rtrs_clt_sess
RDMA/rtrs-srv: Rename rtrs_srv to rtrs_srv_sess
RDMA/rtrs-clt: Rename rtrs_clt_sess to rtrs_clt_path
RDMA/rtrs-srv: Rename rtrs_srv_sess to rtrs_srv_path
RDMA/rtrs: Rename rtrs_sess to rtrs_path
RDMA/hns: Modify the hop num of HIP09 EQ to 1
IB/iser: Align coding style across driver
IB/iser: Remove un-needed casting to/from void pointer
IB/iser: Don't suppress send completions
IB/iser: Rename ib_ret local variable
IB/iser: Fix RNR errors
IB/iser: Remove deprecated pi_guard module param
IB/mlx5: Expose NDR speed through MAD
RDMA/cxgb4: Set queue pair state when being queried
...
Pull irq updates from Thomas Gleixner:
"Updates for the interrupt subsystem:
Core:
- Provide a new interface for affinity hints to provide a separation
between hint and actual affinity change which has become a hidden
property of the current interface
- Fix up the in tree usage of the affinity hint interfaces
Drivers:
- No new irqchip drivers!
- Fix GICv3 redistributor table reservation with RT across kexec
- Fix GICv4.1 redistributor view of the VPE table across kexec
- Add support for extra interrupts on spear-shirq
- Make obtaining some interrupts optional for the Renesas drivers
- Various cleanups and bug fixes"
* tag 'irq-core-2022-01-13' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (25 commits)
irqchip/renesas-intc-irqpin: Use platform_get_irq_optional() to get the interrupt
irqchip/renesas-irqc: Use platform_get_irq_optional() to get the interrupt
irqchip/gic-v4: Disable redistributors' view of the VPE table at boot time
irqchip/ingenic-tcu: Use correctly sized arguments for bit field
irqchip/gic-v2m: Add const to of_device_id
irqchip/imx-gpcv2: Mark imx_gpcv2_instance with __ro_after_init
irqchip/spear-shirq: Add support for IRQ 0..6
irqchip/gic-v3-its: Limit memreserve cpuhp state lifetime
irqchip/gic-v3-its: Postpone LPI pending table freeing and memreserve
irqchip/gic-v3-its: Give the percpu rdist struct its own flags field
net/mlx4: Use irq_update_affinity_hint()
net/mlx5: Use irq_set_affinity_and_hint()
hinic: Use irq_set_affinity_and_hint()
scsi: lpfc: Use irq_set_affinity()
mailbox: Use irq_update_affinity_hint()
ixgbe: Use irq_update_affinity_hint()
be2net: Use irq_update_affinity_hint()
enic: Use irq_update_affinity_hint()
RDMA/irdma: Use irq_update_affinity_hint()
scsi: mpt3sas: Use irq_set_affinity_and_hint()
...
Pull driver core updates from Greg KH:
"Here is the set of changes for the driver core for 5.17-rc1.
Lots of little things here, including:
- kobj_type cleanups
- auxiliary_bus documentation updates
- auxiliary_device conversions for some drivers (relevant subsystems
all have provided acks for these)
- kernfs lock contention reduction for some workloads
- other tiny cleanups and changes.
All of these have been in linux-next for a while with no reported
issues"
* tag 'driver-core-5.17-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core: (43 commits)
kobject documentation: remove default_attrs information
drivers/firmware: Add missing platform_device_put() in sysfb_create_simplefb
debugfs: lockdown: Allow reading debugfs files that are not world readable
driver core: Make bus notifiers in right order in really_probe()
driver core: Move driver_sysfs_remove() after driver_sysfs_add()
firmware: edd: remove empty default_attrs array
firmware: dmi-sysfs: use default_groups in kobj_type
qemu_fw_cfg: use default_groups in kobj_type
firmware: memmap: use default_groups in kobj_type
sh: sq: use default_groups in kobj_type
headers/uninline: Uninline single-use function: kobject_has_children()
devtmpfs: mount with noexec and nosuid
driver core: Simplify async probe test code by using ktime_ms_delta()
nilfs2: use default_groups in kobj_type
kobject: remove kset from struct kset_uevent_ops callbacks
driver core: make kobj_type constant.
driver core: platform: document registration-failure requirement
vdpa/mlx5: Use auxiliary_device driver data helpers
net/mlx5e: Use auxiliary_device driver data helpers
soundwire: intel: Use auxiliary_device driver data helpers
...
Alexei Starovoitov says:
====================
pull-request: bpf-next 2021-12-30
The following pull-request contains BPF updates for your *net-next* tree.
We've added 72 non-merge commits during the last 20 day(s) which contain
a total of 223 files changed, 3510 insertions(+), 1591 deletions(-).
The main changes are:
1) Automatic setrlimit in libbpf when bpf is memcg's in the kernel, from Andrii.
2) Beautify and de-verbose verifier logs, from Christy.
3) Composable verifier types, from Hao.
4) bpf_strncmp helper, from Hou.
5) bpf.h header dependency cleanup, from Jakub.
6) get_func_[arg|ret|arg_cnt] helpers, from Jiri.
7) Sleepable local storage, from KP.
8) Extend kfunc with PTR_TO_CTX, PTR_TO_MEM argument support, from Kumar.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
sock.h is pretty heavily used (5k objects rebuilt on x86 after
it's touched). We can drop the include of filter.h from it and
add a forward declaration of struct sk_filter instead.
This decreases the number of rebuilt objects when bpf.h
is touched from ~5k to ~1k.
There's a lot of missing includes this was masking. Primarily
in networking tho, this time.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Marc Kleine-Budde <mkl@pengutronix.de>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Acked-by: Nikolay Aleksandrov <nikolay@nvidia.com>
Acked-by: Stefano Garzarella <sgarzare@redhat.com>
Link: https://lore.kernel.org/bpf/20211229004913.513372-1-kuba@kernel.org
The driver uses irq_set_affinity_hint() to update the affinity_hint mask
that is consumed by the userspace to distribute the interrupts. However,
under the hood irq_set_affinity_hint() also applies the provided cpumask
(if not NULL) as the affinity for the given interrupt which is an
undocumented side effect.
To remove this side effect irq_set_affinity_hint() has been marked
as deprecated and new interfaces have been introduced. Hence, replace the
irq_set_affinity_hint() with the new interface irq_update_affinity_hint()
that only updates the affinity_hint pointer.
Signed-off-by: Nitesh Narayan Lal <nitesh@redhat.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Tatyana Nikolova <tatyana.e.nikolova@intel.com>
Link: https://lore.kernel.org/r/20210903152430.244937-7-nitesh@redhat.com
Completion events (CEs) are lost if the application is allowed to arm the
CQ more than two times when no new CE for this CQ has been generated by
the HW.
Check if arming has been done for the CQ and if not, arm the CQ for any
event otherwise promote to arm the CQ for any event only when the last arm
event was solicited.
Fixes: b48c24c2d7 ("RDMA/irdma: Implement device supported verb APIs")
Link: https://lore.kernel.org/r/20211201231509.1930-2-shiraz.saleem@intel.com
Signed-off-by: Tatyana Nikolova <tatyana.e.nikolova@intel.com>
Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
'pchunk->bitmapbuf' is a bitmap. Its size (in number of bits) is stored in
'pchunk->sizeofbitmap'.
When it is allocated, the size (in bytes) is computed by:
size_in_bits >> 3
There are 2 issues (numbers bellow assume that longs are 64 bits):
- there is no guarantee here that 'pchunk->bitmapmem.size' is modulo
BITS_PER_LONG but bitmaps are stored as longs
(sizeofbitmap=8 bits will only allocate 1 byte, instead of 8 (1 long))
- the number of bytes is computed with a shift, not a round up, so we
may allocate less memory than needed
(sizeofbitmap=65 bits will only allocate 8 bytes (i.e. 1 long), when 2
longs are needed = 16 bytes)
Fix both issues by using 'bitmap_zalloc()' and remove the useless
'bitmapmem' from 'struct irdma_chunk'.
While at it, remove some useless NULL test before calling
kfree/bitmap_free.
Fixes: 915cc7ac0f ("RDMA/irdma: Add miscellaneous utility definitions")
Link: https://lore.kernel.org/r/5e670b640508e14b1869c3e8e4fb970d78cbe997.1638692171.git.christophe.jaillet@wanadoo.fr
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Reviewed-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
When irdma_ws_add fails, irdma_ws_remove is used to cleanup the leaf node.
This lead to holding the qos mutex twice in the QP resume path. Fix this
by avoiding the call to irdma_ws_remove and unwinding the error in
irdma_ws_add. This skips the call to irdma_tc_in_use function which is not
needed in the error unwind cases.
Fixes: 3ae331c751 ("RDMA/irdma: Add QoS definitions")
Link: https://lore.kernel.org/r/20211019151654.1943-2-shiraz.saleem@intel.com
Signed-off-by: Mustafa Ismail <mustafa.ismail@intel.com>
Signed-off-by: Shiraz Saleem <shiraz.saleem@intel.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
