| Age | Commit message (Collapse) | Author |
|---|
|
When registering mdiobus children, if we get an -EPROBE_DEFER, we shouldn't
ignore it and continue registering the rest of the mdiobus children. This
would permanently prevent the deferring child mdiobus from working instead
of reattempting it in the future. So, if a child mdiobus needs to be
reattempted in the future, defer the entire mdio-mux initialization.
This fixes the issue where PHYs sitting under the mdio-mux aren't
initialized correctly if the PHY's interrupt controller is not yet ready
when the mdio-mux is being probed. Additional context in the link below.
Fixes: 0ca2997d1452 ("netdev/of/phy: Add MDIO bus multiplexer support.")
Link: https://lore.kernel.org/lkml/CAGETcx95kHrv8wA-O+-JtfH7H9biJEGJtijuPVN0V5dUKUAB3A@mail.gmail.com/#t
Signed-off-by: Saravana Kannan <saravanak@google.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Acked-by: Marc Zyngier <maz@kernel.org>
Tested-by: Marc Zyngier <maz@kernel.org>
Acked-by: Kevin Hilman <khilman@baylibre.com>
Tested-by: Kevin Hilman <khilman@baylibre.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
If we are seeing memory allocation errors, don't try to continue
registering child mdiobus devices. It's unlikely they'll succeed.
Fixes: 342fa1964439 ("mdio: mux: make child bus walking more permissive and errors more verbose")
Signed-off-by: Saravana Kannan <saravanak@google.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Acked-by: Marc Zyngier <maz@kernel.org>
Tested-by: Marc Zyngier <maz@kernel.org>
Acked-by: Kevin Hilman <khilman@baylibre.com>
Tested-by: Kevin Hilman <khilman@baylibre.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
The whole point of devm_* APIs is that you don't have to undo them if you
are returning an error that's going to get propagated out of a probe()
function. So delete unnecessary devm_kfree() call in the error return path.
Fixes: b60161668199 ("mdio: mux: Correct mdio_mux_init error path issues")
Signed-off-by: Saravana Kannan <saravanak@google.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Acked-by: Marc Zyngier <maz@kernel.org>
Tested-by: Marc Zyngier <maz@kernel.org>
Acked-by: Kevin Hilman <khilman@baylibre.com>
Tested-by: Kevin Hilman <khilman@baylibre.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
or worse
It seems that of_find_compatible_node has a weird calling convention in
which it calls of_node_put() on the "from" node argument, instead of
leaving that up to the caller. This comes from the fact that
of_find_compatible_node with a non-NULL "from" argument it only supposed
to be used as the iterator function of for_each_compatible_node(). OF
iterator functions call of_node_get on the next OF node and of_node_put()
on the previous one.
When of_find_compatible_node calls of_node_put, it actually never
expects the refcount to drop to zero, because the call is done under the
atomic devtree_lock context, and when the refcount drops to zero it
triggers a kobject and a sysfs file deletion, which assume blocking
context.
So any driver call to of_find_compatible_node is probably buggy because
an unexpected of_node_put() takes place.
What should be done is to use the of_get_compatible_child() function.
Fixes: 5a8f09748ee7 ("net: dsa: sja1105: register the MDIO buses for 100base-T1 and 100base-TX")
Link: https://lore.kernel.org/netdev/20210814010139.kzryimmp4rizlznt@skbuf/
Suggested-by: Frank Rowand <frowand.list@gmail.com>
Suggested-by: Rob Herring <robh+dt@kernel.org>
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
When adding support for using the skb->hash value as the flow hash in CAKE,
I accidentally introduced a logic error that broke the host-only isolation
modes of CAKE (srchost and dsthost keywords). Specifically, the flow_hash
variable should stay initialised to 0 in cake_hash() in pure host-based
hashing mode. Add a check for this before using the skb->hash value as
flow_hash.
Fixes: b0c19ed6088a ("sch_cake: Take advantage of skb->hash where appropriate")
Reported-by: Pete Heist <pete@heistp.net>
Tested-by: Pete Heist <pete@heistp.net>
Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
No longer required now that userspace can't touch anything that might
need it, and should fix DRM MM operations racing with each other, and
the random hangs/crashes that come with that.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
|
|
Long ago, there had been plans for making use of a bunch of these APIs
from userspace and there's various checks in place to stop misbehaving.
Countless other projects have occurred in the meantime, and the pieces
didn't finish falling into place for that to happen.
They will (hopefully) in the not-too-distant future, but it won't look
quite as insane. The super checks are causing problems right now, and
are going to be removed.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
|
|
I honestly don't even know why... These have never been used.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
|
|
Should fix some initial modeset failures on (at least) Ampere boards.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
|
|
When booted with multiple displays attached, the EFI GOP driver on (at
least) Ampere, can leave DP links powered up that aren't being used to
display anything. This confuses our tracking of SOR routing, with the
likely result being a failed modeset and display engine hang.
Fix this by (ab?)using the DisableLT IED script to power-down the link,
restoring HW to a state the driver expects.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
|
|
Still no GA106 as I don't have HW to verif.
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
|
|
The struct pci_dev uses reference counting but zPCI assumed erroneously
that the last reference would always be the local reference after
calling pci_stop_and_remove_bus_device(). This is usually the case but
not how reference counting works and thus inherently fragile.
In fact one case where this causes a NULL pointer dereference when on an
SRIOV device the function 0 was hot unplugged before another function of
the same multi-function device. In this case the second function's
pdev->sriov->dev reference keeps the struct pci_dev of function 0 alive
even after the unplug. This bug was previously hidden by the fact that
we were leaking the struct pci_dev which in turn means that it always
outlived the struct zpci_dev. This was fixed in commit 0b13525c20fe
("s390/pci: fix leak of PCI device structure") exposing the broken
behavior.
Fix this by accounting for the long living reference a struct pci_dev
has to its underlying struct zpci_dev via the zbus->function[] array and
only release that in pcibios_release_device() ensuring that the struct
pci_dev is not left with a dangling reference. This is a minimal fix in
the future it would probably better to use fine grained reference
counting for struct zpci_dev.
Fixes: 05bc1be6db4b2 ("s390/pci: create zPCI bus")
Cc: stable@vger.kernel.org
Reviewed-by: Matthew Rosato <mjrosato@linux.ibm.com>
Signed-off-by: Niklas Schnelle <schnelle@linux.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
|
|
The memory block occupied by the SCLP early buffer that is allocated
by the decompressor and then handed over to the decompressed kernel,
must be reserved to prevent it from being reused for other purposes.
This is necessary because the SCLP early buffer is still in use
during kernel initialization.
Fixes: f1d3c5323772 ("s390/boot: move sclp early buffer from fixed address in asm to C")
Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Reported-by: Alexander Gordeev <agordeev@linux.ibm.com>
Reviewed-by: Alexander Gordeev <agordeev@linux.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
|
|
Tests showed a mismatch between what the CCA tool reports about
the APKA master key state and what's displayed by the zcrypt dd
in sysfs. After some investigation, we found out that the
documentation which was the source for the zcrypt dd implementation
lacks the listing of 3 fields. So this patch now moves the
evaluation of the APKA master key state to the correct offset.
Signed-off-by: Harald Freudenberger <freude@linux.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
|
|
The last user of arch_set_page_states(), arch_set_page_nodat() and
arch_test_page_nodat() was removed in commit 394216275c7d
("s390: remove broken hibernate / power management support"),
let's remove these functions.
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20210806075430.6103-1-david@redhat.com
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
|
|
Neither of the two drivers provides any SLIB parameter data, so get rid
of the dead code.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Reviewed-by: Benjamin Block <bblock@linux.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
|
|
Move all QIB-related code into qdio_setup_qib(), and slightly re-order
it according to the order of the struct's fields. This makes it easier
to understand what the QIB actually looks like before we send it to HW.
Also get rid of the qebsm_possible() helper - as 31-bit support is long
gone, the comment doesn't make any sense. And while removing some stale
QIB-related comment, also move the clearing of the QDR into its proper
place.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Reviewed-by: Benjamin Block <bblock@linux.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
|
|
Prefer dev_info() over a raw printk. This also adds the device and
driver names into the output, so that we have:
Before:
qdio: 0.0.17c0 ZFCP on SC 17 using [...]
After:
zfcp 0.0.17c0: qdio: ZFCP on SC 17 using [...]
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Reviewed-by: Benjamin Block <bblock@linux.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
|
|
Push the sync check from qdio_inspect_queue() down into the two
get_*_buffer_frontier() code paths, where we actually need the sync to
look at the current queue state. This lets us avoid the check when we
know that there is no work on the queue (ie. when q->nr_buf_used is 0).
While at it introduce the qdio_sync_*_queue() helpers, so that we can
avoid the branch on q->is_input_q when we already know the queue type.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Reviewed-by: Benjamin Block <bblock@linux.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
|
|
Don't bother with translating the SIGA-related capability bits into
our own internal format, just cache the full qdioac1 field instead.
Also adjust the helper macros so that they take a qdio_irq argument
and can be used everywhere, instead of taking a qdio_q and then
internally dereferencing the parent pointer.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Reviewed-by: Benjamin Block <bblock@linux.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
|
|
The queue processing is fully decoupled from any preceding interrupt,
so we're no longer making any use of the sync-after-IRQ HW capabilities.
And as SIGA-sync is a legacy feature, there's also not much point in
re-designing the driver & qdio-layer code just so that we can
potentially avoid a few syncs. So just remove all the leftover code.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Reviewed-by: Benjamin Block <bblock@linux.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
|
|
Clean up yet another path where HW wants an absolute address, and we've
been implicitly relying on V=R.
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Reviewed-by: Benjamin Block <bblock@linux.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
|
|
Introduce dev_busid, which exports the device-id associated with the
io-subchannel (and message-subchannel). The dev_busid indicates that of
the device which may be physically installed on the corrosponding
subchannel. The dev_busid value "none" indicates that the subchannel
is not valid, there is no I/O device currently associated with the
subchannel.
The dev_busid information would be helpful to write device-specific
udev-rules associated with the subchannel. The dev_busid interface would
be available even when the sch is not bound to any driver or if there is
no operational device connected on it. Hence this attribute can be used to
write udev-rules which are specific to the device associated with the
subchannel.
Signed-off-by: Vineeth Vijayan <vneethv@linux.ibm.com>
Reviewed-by: Peter Oberparleiter <oberpar@linux.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
|
|
This patch introduces a new rescan sys-interface for channel-subsystem.
The rescan interface allows the user to invoke an evaluation of all
subchannels defined in the I/O configuration.
The new rescan interface can be found at /sys/devices/css0/rescan
and can be triggered by,
echo > /sys/devices/css0/rescan
Writing to this interface triggers subchannel evaluation. The write
request completes only after scan-related work has completed
This user-invoked subchannel evaluation allows manual recovery in error
situations such as:
- restart of device discovery after resolution of temporary device
error
- inconsistent OS view of subchannel state due to missing state-change
interrupts (CRWs)
Signed-off-by: Vineeth Vijayan <vneethv@linux.ibm.com>
Reviewed-by: Peter Oberparleiter <oberpar@linux.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
|
|
Reported as working here:
https://github.com/t-8ch/linux-gigabyte-wmi-driver/issues/1#issuecomment-900263115
Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
Link: https://lore.kernel.org/r/20210817154628.84992-1-linux@weissschuh.net
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
|
|
is_flush_rq() is called from bt_iter()/bt_tags_iter(), and runs the
following check:
hctx->fq->flush_rq == req
but the passed hctx from bt_iter()/bt_tags_iter() may be NULL because:
1) memory re-order in blk_mq_rq_ctx_init():
rq->mq_hctx = data->hctx;
...
refcount_set(&rq->ref, 1);
OR
2) tag re-use and ->rqs[] isn't updated with new request.
Fix the issue by re-writing is_flush_rq() as:
return rq->end_io == flush_end_io;
which turns out simpler to follow and immune to data race since we have
ordered WRITE rq->end_io and refcount_set(&rq->ref, 1).
Fixes: 2e315dc07df0 ("blk-mq: grab rq->refcount before calling ->fn in blk_mq_tagset_busy_iter")
Cc: "Blank-Burian, Markus, Dr." <blankburian@uni-muenster.de>
Cc: Yufen Yu <yuyufen@huawei.com>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Link: https://lore.kernel.org/r/20210818010925.607383-1-ming.lei@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
In ixgbe_xsk_pool_enable(), if ixgbe_xsk_wakeup() fails,
We should restore the previous state and clean up the
resources. Add the missing clear af_xdp_zc_qps and unmap dma
to fix this bug.
Fixes: d49e286d354e ("ixgbe: add tracking of AF_XDP zero-copy state for each queue pair")
Fixes: 4a9b32f30f80 ("ixgbe: fix potential RX buffer starvation for AF_XDP")
Signed-off-by: Wang Hai <wanghai38@huawei.com>
Acked-by: Magnus Karlsson <magnus.karlsson@intel.com>
Tested-by: Sandeep Penigalapati <sandeep.penigalapati@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Link: https://lore.kernel.org/r/20210817203736.3529939-1-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/kvalo/wireless-drivers
Kalle Valo says:
====================
wireless-drivers fixes for v5.14
First set of fixes for v5.14 and nothing major this time. New devices
for iwlwifi and one fix for a compiler warning.
iwlwifi
* support for new devices
mt76
* fix compiler warning about MT_CIPHER_NONE
* tag 'wireless-drivers-2021-08-17' of git://git.kernel.org/pub/scm/linux/kernel/git/kvalo/wireless-drivers:
mt76: fix enum type mismatch
iwlwifi: add new so-jf devices
iwlwifi: add new SoF with JF devices
iwlwifi: pnvm: accept multiple HW-type TLVs
====================
Link: https://lore.kernel.org/r/20210817171027.EC1E6C43460@smtp.codeaurora.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Pin ring in io_fallback_req_func() by briefly elevating ctx->refs in
case any task_work handler touches ctx after releasing a request.
Fixes: 9011bf9a13e3b ("io_uring: fix stuck fallback reqs")
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/833a494713d235ec144284a9bbfe418df4f6b61c.1629235576.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace
Pull tracing fix from Steven Rostedt:
"Limit the shooting in the foot of tp_printk
The "tp_printk" option redirects the trace event output to printk at
boot up. This is useful when a machine crashes before boot where the
trace events can not be retrieved by the in kernel ring buffer. But it
can be "dangerous" because trace events can be located in high
frequency locations such as interrupts and the scheduler, where a
printk can slow it down that it live locks the machine (because by the
time the printk finishes, the next event is triggered). Thus tp_printk
must be used with care.
It was discovered that the filter logic to trace events does not apply
to the tp_printk events. This can cause a surprise and live lock when
the user expects it to be filtered to limit the amount of events
printed to the console when in fact it still prints everything"
* tag 'trace-v5.14-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace:
tracing: Apply trace filters on all output channels
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/vireshk/pm
Pull ARM cpufreq fixes for v5.14 from Viresh Kumar:
"This contains:
- Addition of SoCs to blocklist for cpufreq-dt driver (Bjorn Andersson
and Thara Gopinath).
- Fix error path for scmi driver (Lukasz Luba).
- Temporarily disable highest frequency for armada, its unsafe and
breaks stuff."
* 'cpufreq/arm/fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/vireshk/pm:
cpufreq: armada-37xx: forbid cpufreq for 1.2 GHz variant
cpufreq: blocklist Qualcomm sm8150 in cpufreq-dt-platdev
cpufreq: arm_scmi: Fix error path when allocation failed
cpufreq: blacklist Qualcomm sc8180x in cpufreq-dt-platdev
|
|
[Why]
Userspace should get back a copy of drm_wait_vblank that's been modified
even when drm_wait_vblank_ioctl returns a failure.
Rationale:
drm_wait_vblank_ioctl modifies the request and expects the user to read
it back. When the type is RELATIVE, it modifies it to ABSOLUTE and updates
the sequence to become current_vblank_count + sequence (which was
RELATIVE), but now it became ABSOLUTE.
drmWaitVBlank (in libdrm) expects this to be the case as it modifies
the request to be Absolute so it expects the sequence to would have been
updated.
The change is in compat_drm_wait_vblank, which is called by
drm_compat_ioctl. This change of copying the data back regardless of the
return number makes it en par with drm_ioctl, which always copies the
data before returning.
[How]
Return from the function after everything has been copied to user.
Fixes IGT:kms_flip::modeset-vs-vblank-race-interruptible
Tested on ChromeOS Trogdor(msm)
Reviewed-by: Michel Dänzer <mdaenzer@redhat.com>
Signed-off-by: Mark Yacoub <markyacoub@chromium.org>
Signed-off-by: Sean Paul <seanpaul@chromium.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20210812194917.1703356-1-markyacoub@chromium.org
|
|
Update the comment with the new features.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Acked-by: Josh Poimboeuf <jpoimboe@redhat.com>
Link: https://lore.kernel.org/r/YQwIorQBHEq+s73b@hirez.programming.kicks-ass.net
|
|
On PREEMPT_RT enabled kernels local_lock maps to a per CPU 'sleeping'
spinlock which protects the critical section while staying preemptible. CPU
locality is established by disabling migration.
Provide the necessary types and macros to substitute the non-RT variant.
Co-developed-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20210815211306.023630962@linutronix.de
|
|
Add the static and runtime initializer mechanics to support the RT variant
of local_lock, which requires the lock type in the lockdep map to be set
to LD_LOCK_PERCPU.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20210815211305.967526724@linutronix.de
|
|
Going to sleep when locks are contended can be quite inefficient when the
contention time is short and the lock owner is running on a different CPU.
The MCS mechanism cannot be used because MCS is strictly FIFO ordered while
for rtmutex based locks the waiter ordering is priority based.
Provide a simple adaptive spinwait mechanism which currently restricts the
spinning to the top priority waiter.
[ tglx: Provide a contemporary changelog, extended it to all rtmutex based
locks and updated it to match the other spin on owner implementations ]
Originally-by: Gregory Haskins <ghaskins@novell.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20210815211305.912050691@linutronix.de
|
|
The current logic only allows lock stealing to occur if the current task is
of higher priority than the pending owner.
Significant throughput improvements can be gained by allowing the lock
stealing to include tasks of equal priority when the contended lock is a
spin_lock or a rw_lock and the tasks are not in a RT scheduling task.
The assumption was that the system will make faster progress by allowing
the task already on the CPU to take the lock rather than waiting for the
system to wake up a different task.
This does add a degree of unfairness, but in reality no negative side
effects have been observed in the many years that this has been used in the
RT kernel.
[ tglx: Refactored and rewritten several times by Steve Rostedt, Sebastian
Siewior and myself ]
Signed-off-by: Gregory Haskins <ghaskins@novell.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20210815211305.857240222@linutronix.de
|
|
On PREEMPT_RT regular spinlocks and rwlocks are substituted with rtmutex
based constructs. spin/rwlock held regions are preemptible on PREEMPT_RT,
so PREEMPT_LOCK_OFFSET has to be 0 to make the various cond_resched_*lock()
functions work correctly.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20210815211305.804246275@linutronix.de
|
|
On PREEMPT_RT the futex hashbucket spinlock becomes 'sleeping' and rtmutex
based. That causes a lockdep false positive because some of the futex
functions invoke spin_unlock(&hb->lock) with the wait_lock of the rtmutex
associated to the pi_futex held. spin_unlock() in turn takes wait_lock of
the rtmutex on which the spinlock is based which makes lockdep notice a
lock recursion.
Give the futex/rtmutex wait_lock a separate key.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20210815211305.750701219@linutronix.de
|
|
The requeue_pi() operation on RT kernels creates a problem versus the
task::pi_blocked_on state when a waiter is woken early (signal, timeout)
and that early wake up interleaves with the requeue_pi() operation.
When the requeue manages to block the waiter on the rtmutex which is
associated to the second futex, then a concurrent early wakeup of that
waiter faces the problem that it has to acquire the hash bucket spinlock,
which is not an issue on non-RT kernels, but on RT kernels spinlocks are
substituted by 'sleeping' spinlocks based on rtmutex. If the hash bucket
lock is contended then blocking on that spinlock would result in a
impossible situation: blocking on two locks at the same time (the hash
bucket lock and the rtmutex representing the PI futex).
It was considered to make the hash bucket locks raw_spinlocks, but
especially requeue operations with a large amount of waiters can introduce
significant latencies, so that's not an option for RT.
The RT tree carried a solution which (ab)used task::pi_blocked_on to store
the information about an ongoing requeue and an early wakeup which worked,
but required to add checks for these special states all over the place.
The distangling of an early wakeup of a waiter for a requeue_pi() operation
is already looking at quite some different states and the task::pi_blocked_on
magic just expanded that to a hard to understand 'state machine'.
This can be avoided by keeping track of the waiter/requeue state in the
futex_q object itself.
Add a requeue_state field to struct futex_q with the following possible
states:
Q_REQUEUE_PI_NONE
Q_REQUEUE_PI_IGNORE
Q_REQUEUE_PI_IN_PROGRESS
Q_REQUEUE_PI_WAIT
Q_REQUEUE_PI_DONE
Q_REQUEUE_PI_LOCKED
The waiter starts with state = NONE and the following state transitions are
valid:
On the waiter side:
Q_REQUEUE_PI_NONE -> Q_REQUEUE_PI_IGNORE
Q_REQUEUE_PI_IN_PROGRESS -> Q_REQUEUE_PI_WAIT
On the requeue side:
Q_REQUEUE_PI_NONE -> Q_REQUEUE_PI_INPROGRESS
Q_REQUEUE_PI_IN_PROGRESS -> Q_REQUEUE_PI_DONE/LOCKED
Q_REQUEUE_PI_IN_PROGRESS -> Q_REQUEUE_PI_NONE (requeue failed)
Q_REQUEUE_PI_WAIT -> Q_REQUEUE_PI_DONE/LOCKED
Q_REQUEUE_PI_WAIT -> Q_REQUEUE_PI_IGNORE (requeue failed)
The requeue side ignores a waiter with state Q_REQUEUE_PI_IGNORE as this
signals that the waiter is already on the way out. It also means that
the waiter is still on the 'wait' futex, i.e. uaddr1.
The waiter side signals early wakeup to the requeue side either through
setting state to Q_REQUEUE_PI_IGNORE or to Q_REQUEUE_PI_WAIT depending
on the current state. In case of Q_REQUEUE_PI_IGNORE it can immediately
proceed to take the hash bucket lock of uaddr1. If it set state to WAIT,
which means the wakeup is interleaving with a requeue in progress it has
to wait for the requeue side to change the state. Either to DONE/LOCKED
or to IGNORE. DONE/LOCKED means the waiter q is now on the uaddr2 futex
and either blocked (DONE) or has acquired it (LOCKED). IGNORE is set by
the requeue side when the requeue attempt failed via deadlock detection
and therefore the waiter's futex_q is still on the uaddr1 futex.
While this is not strictly required on !RT making this unconditional has
the benefit of common code and it also allows the waiter to avoid taking
the hash bucket lock on the way out in certain cases, which reduces
contention.
Add the required helpers required for the state transitions, invoke them at
the right places and restructure the futex_wait_requeue_pi() code to handle
the return from wait (early or not) based on the state machine values.
On !RT enabled kernels the waiter spin waits for the state going from
Q_REQUEUE_PI_WAIT to some other state, on RT enabled kernels this is
handled by rcuwait_wait_event() and the corresponding wake up on the
requeue side.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20210815211305.693317658@linutronix.de
|
|
Move the futex key match out of handle_early_requeue_pi_wakeup() which
allows to simplify that function. The upcoming state machine for
requeue_pi() will make that go away.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20210815211305.638938670@linutronix.de
|
|
No point in allocating memory when the input parameters are bogus.
Validate all parameters before proceeding.
Suggested-by: Davidlohr Bueso <dave@stgolabs.net>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20210815211305.581789253@linutronix.de
|
|
The comment about the restriction of the number of waiters to wake for the
REQUEUE_PI case is confusing at best. Rewrite it.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20210815211305.524990421@linutronix.de
|
|
No point in taking two more 'requeue_pi' conditionals just to get to the
requeue. Same for the requeue_pi case just the other way round.
No functional change.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20210815211305.468835790@linutronix.de
|
|
The accounting is wrong when either the PI sanity check or the
requeue PI operation fails. Adjust it in the failure path.
Will be simplified in the next step.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20210815211305.416427548@linutronix.de
|
|
For requeue PI it's required to establish PI state for the PI futex to
which waiters are requeued. This either acquires the user space futex on
behalf of the top most waiter on the inner 'waitqueue' futex, or attaches to
the PI state of an existing waiter, or creates on attached to the owner of
the futex.
This code can retry in case of failure, but retry can never happen when the
pi state was successfully created. The condition to run this code is:
(task_count - nr_wake) < nr_requeue
which is always true because:
task_count = 0
nr_wake = 1
nr_requeue >= 0
Remove it completely.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20210815211305.362730187@linutronix.de
|
|
When requeuing to a PI futex, then the requeue code tries to trylock the PI
futex on behalf of the topmost waiter on the inner 'waitqueue' futex. If
that succeeds, then PI state has to be allocated in order to requeue further
waiters to the PI futex.
The comment and the code are confusing, as the PI state allocation uses
lookup_pi_state(), which either attaches to an existing waiter or to the
owner. As the PI futex was just acquired, there cannot be a waiter on the
PI futex because the hash bucket lock is held.
Clarify the comment and use attach_to_pi_owner() directly. As the task on
which behalf the PI futex has been acquired is guaranteed to be alive and
not exiting, this call must succeed. Add a WARN_ON() in case that fails.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20210815211305.305142462@linutronix.de
|
|
The futex key reference mechanism is long gone. Clean up the stale comments
which still mention it.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20210815211305.249178312@linutronix.de
|
|
The loop in futex_requeue() has a sanity check for the waiter, which is
missing in futex_proxy_trylock_atomic(). In theory the key2 check is
sufficient, but futexes are cursed so add it for completeness and paranoia
sake.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20210815211305.193767519@linutronix.de
|
|
The inner parts of certain locks (mutex, rwlocks) changed due to a rework for
RT and non RT code. Most users remain unaffected, but those who fiddle around
in the inner parts need to be updated.
Match the struct names to the new layout.
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20210815211305.137982730@linutronix.de
|
