When a PCIe card is hot-removed, the Presence Detect State and Data Link
Layer Link Active bits often do not clear simultaneously. I've seen delays
of up to 244 msec between the two events with Thunderbolt.
After pciehp has brought down the slot in response to the first event, the
other bit may still be set. It's not discernible whether it's set because
a new card is already in the slot or if it will soon clear. So pciehp
tries to bring up the slot and in the latter case fails with a bunch of
messages, some of them at KERN_ERR severity. If the slot is no longer
occupied, the messages are false positives and annoy users.
Stuart Hayes reports the following splat on hot removal:
KERN_INFO pcieport 0000:3c:06.0: pciehp: Slot(180): Link Up
KERN_INFO pcieport 0000:3c:06.0: pciehp: Timeout waiting for Presence Detect
KERN_ERR pcieport 0000:3c:06.0: pciehp: link training error: status 0x0001
KERN_ERR pcieport 0000:3c:06.0: pciehp: Failed to check link status
Dongdong Liu complains about a similar splat:
KERN_INFO pciehp 0000:80:10.0:pcie004: Slot(36): Link Down
KERN_INFO iommu: Removing device 0000:87:00.0 from group 12
KERN_INFO pciehp 0000:80:10.0:pcie004: Slot(36): Card present
KERN_INFO pcieport 0000:80:10.0: Data Link Layer Link Active not set in 1000 msec
KERN_ERR pciehp 0000:80:10.0:pcie004: Failed to check link status
Users are particularly irritated to see a bringup attempt even though the
slot was explicitly brought down via sysfs. In a perfect world, we could
avoid this by setting Link Disable on slot bringdown and re-enabling it
upon a Presence Detect State change. In reality however, there are broken
hotplug ports which hardwire Presence Detect to zero, see 80696f9914
("PCI: pciehp: Tolerate Presence Detect hardwired to zero"). Conversely,
PCIe r1.0 hotplug ports hardwire Link Active to zero because Link Active
Reporting wasn't specified before PCIe r1.1. On unplug, some ports first
clear Presence then Link (see Stuart Hayes' splat) whereas others use the
inverse order (see Dongdong Liu's splat). To top it off, there are hotplug
ports which flap the Presence and Link bits on slot bringup, see
6c35a1ac3d ("PCI: pciehp: Tolerate initially unstable link").
pciehp is designed to work with all of these variants. Surplus attempts at
slot bringup are a lesser evil than not being able to bring up slots at
all. Although we could try to perfect the behavior for specific hotplug
controllers, we'd risk breaking others or increasing code complexity.
But we can certainly minimize annoyance by emitting only a single message
with KERN_INFO severity if bringup is unsuccessful:
* Drop the "Timeout waiting for Presence Detect" message in
pcie_wait_for_presence(). The sole caller of that function,
pciehp_check_link_status(), ignores the timeout and carries on. It emits
error messages of its own and I don't think this particular message adds
much value.
* There's a single error condition in pciehp_check_link_status() which
does not emit a message. Adding one allows dropping the "Failed to check
link status" message emitted by board_added() if
pciehp_check_link_status() returns a non-zero integer.
* Tone down all messages in pciehp_check_link_status() to KERN_INFO
severity and rephrase them to look as innocuous as possible. To this
end, move the message emitted by pcie_wait_for_link_delay() to its
callers.
As a result, Stuart Hayes' splat becomes:
KERN_INFO pcieport 0000:3c:06.0: pciehp: Slot(180): Link Up
KERN_INFO pcieport 0000:3c:06.0: pciehp: Slot(180): Cannot train link: status 0x0001
Dongdong Liu's splat becomes:
KERN_INFO pciehp 0000:80:10.0:pcie004: Slot(36): Card present
KERN_INFO pciehp 0000:80:10.0:pcie004: Slot(36): No link
The messages now merely serve as information that presence or link bits
were set a little longer than expected. Bringup failures which are not
false positives are still reported, albeit no longer at KERN_ERR severity.
Link: https://lore.kernel.org/linux-pci/20200310182100.102987-1-stuart.w.hayes@gmail.com/
Link: https://lore.kernel.org/linux-pci/1547649064-19019-1-git-send-email-liudongdong3@huawei.com/
Link: https://lore.kernel.org/r/b45e46fd8a6aa6930aaac9d7718c2e4b787a4e5e.1595935071.git.lukas@wunner.de
Reported-by: Stuart Hayes <stuart.w.hayes@gmail.com>
Reported-by: Dongdong Liu <liudongdong3@huawei.com>
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Without this commit, a PCIe hotplug port can stop generating interrupts on
hotplug events, so device adds and removals will not be seen:
The pciehp interrupt handler pciehp_isr() reads the Slot Status register
and then writes back to it to clear the bits that caused the interrupt. If
a different interrupt event bit gets set between the read and the write,
pciehp_isr() returns without having cleared all of the interrupt event
bits. If this happens when the MSI isn't masked (which by default it isn't
in handle_edge_irq(), and which it will never be when MSI per-vector
masking is not supported), we won't get any more hotplug interrupts from
that device.
That is expected behavior, according to the PCIe Base Spec r5.0, section
6.7.3.4, "Software Notification of Hot-Plug Events".
Because the Presence Detect Changed and Data Link Layer State Changed event
bits can both get set at nearly the same time when a device is added or
removed, this is more likely to happen than it might seem. The issue was
found (and can be reproduced rather easily) by connecting and disconnecting
an NVMe storage device on at least one system model where the NVMe devices
were being connected to an AMD PCIe port (PCI device 0x1022/0x1483).
Fix the issue by modifying pciehp_isr() to loop back and re-read the Slot
Status register immediately after writing to it, until it sees that all of
the event status bits have been cleared.
[lukas: drop loop count limitation, write "events" instead of "status",
don't loop back in INTx and poll modes, tweak code comment & commit msg]
Link: https://lore.kernel.org/r/78b4ced5072bfe6e369d20e8b47c279b8c7af12e.1582121613.git.lukas@wunner.de
Tested-by: Stuart Hayes <stuart.w.hayes@gmail.com>
Signed-off-by: Stuart Hayes <stuart.w.hayes@gmail.com>
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Joerg Roedel <jroedel@suse.de>
David Hoyer reports that powering pciehp slots up or down via sysfs may
hang: The call to wait_event() in pciehp_sysfs_enable_slot() and
_disable_slot() does not return because ctrl->ist_running remains true.
This flag, which was introduced by commit 157c1062fc ("PCI: pciehp: Avoid
returning prematurely from sysfs requests"), signifies that the IRQ thread
pciehp_ist() is running. It is set to true at the top of pciehp_ist() and
reset to false at the end. However there are two additional return
statements in pciehp_ist() before which the commit neglected to reset the
flag to false and wake up waiters for the flag.
That omission opens up the following race when powering up the slot:
* pciehp_ist() runs because a PCI_EXP_SLTSTA_PDC event was requested
by pciehp_sysfs_enable_slot()
* pciehp_ist() turns on slot power via the following call stack:
pciehp_handle_presence_or_link_change() -> pciehp_enable_slot() ->
__pciehp_enable_slot() -> board_added() -> pciehp_power_on_slot()
* after slot power is turned on, the link comes up, resulting in a
PCI_EXP_SLTSTA_DLLSC event
* the IRQ handler pciehp_isr() stores the event in ctrl->pending_events
and returns IRQ_WAKE_THREAD
* the IRQ thread is already woken (it's bringing up the slot), but the
genirq code remembers to re-run the IRQ thread after it has finished
(such that it can deal with the new event) by setting IRQTF_RUNTHREAD
via __handle_irq_event_percpu() -> __irq_wake_thread()
* the IRQ thread removes PCI_EXP_SLTSTA_DLLSC from ctrl->pending_events
via board_added() -> pciehp_check_link_status() in order to deal with
presence and link flaps per commit 6c35a1ac3d ("PCI: pciehp:
Tolerate initially unstable link")
* after pciehp_ist() has successfully brought up the slot, it resets
ctrl->ist_running to false and wakes up the sysfs requester
* the genirq code re-runs pciehp_ist(), which sets ctrl->ist_running
to true but then returns with IRQ_NONE because ctrl->pending_events
is empty
* pciehp_sysfs_enable_slot() is finally woken but notices that
ctrl->ist_running is true, hence continues waiting
The only way to get the hung task going again is to trigger a hotplug
event which brings down the slot, e.g. by yanking out the card.
The same race exists when powering down the slot because remove_board()
likewise clears link or presence changes in ctrl->pending_events per commit
3943af9d01 ("PCI: pciehp: Ignore Link State Changes after powering off a
slot") and thereby may cause a re-run of pciehp_ist() which returns with
IRQ_NONE without resetting ctrl->ist_running to false.
Fix by adding a goto label before the teardown steps at the end of
pciehp_ist() and jumping to that label from the two return statements which
currently neglect to reset the ctrl->ist_running flag.
Fixes: 157c1062fc ("PCI: pciehp: Avoid returning prematurely from sysfs requests")
Link: https://lore.kernel.org/r/cca1effa488065cb055120aa01b65719094bdcb5.1584530321.git.lukas@wunner.de
Reported-by: David Hoyer <David.Hoyer@netapp.com>
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Cc: stable@vger.kernel.org # v4.19+
Some systems have in-band presence detection disabled for hot-plug PCI
slots but do not report this in the slot capabilities 2 (SLTCAP2) register.
On these systems, presence detect can become active well after the link is
reported to be active, which can cause the slots to be disabled after a
device is connected.
Add a DMI table to flag these systems as having in-band presence detect
disabled.
Link: https://lore.kernel.org/r/20191025190047.38130-4-stuart.w.hayes@gmail.com
Signed-off-by: Stuart Hayes <stuart.w.hayes@gmail.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Reviewed-by: Lukas Wunner <lukas@wunner.de>
When in-band presence detect is disabled, PDS may come up at any time or
not at all. PDS being low may indicate that the card is still mating, and
we could expect contact bounce to bring down the link as well.
It is reasonable to assume that most cards will mate in a hotplug slot in
about a second. Thus, when we know PDS only reflects out-of-band presence
detect, it's worthwhile to wait the extra second or so to make sure the
card is properly mated before loading the driver and to prevent the hotplug
code from disabling a device if the presence detect change goes active
after the device is enabled.
Link: https://lore.kernel.org/r/20191025190047.38130-3-stuart.w.hayes@gmail.com
[bhelgaas: use ctrl_info() instead of pci_info()]
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Signed-off-by: Stuart Hayes <stuart.w.hayes@gmail.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Reviewed-by: Lukas Wunner <lukas@wunner.de>
The presence detect state (PDS) is normally a logical OR of in-band and
out-of-band (OOB) presence detect. As of PCIe 4.0, there is the option to
disable in-band presence so that the PDS bit always reflects the state of
the out-of-band presence.
The recommendation of the PCIe spec is to disable in-band presence whenever
supported (PCIe r5.0, appendix I implementation note):
Due to architectural issues, the in-band (Physical-Layer-based) portion
of the PD mechanism is deprecated for use with async hot-plug. One issue
is that in-band PD as architected does not detect adapter removal during
certain LTSSM states, notably the L1 and Disabled States. Another issue
is that when both in-band and OOB PD are being used together, the
Presence Detect State bit and its associated interrupt mechanism always
reflect the logical OR of the inband and OOB PD states, and with some
hot-plug hardware configurations, it is important for software to detect
and respond to in-band and OOB PD events independently. If OOB PD is
being used and the associated DSP supports In-Band PD Disable, it is
recommended that the In-Band PD Disable bit be Set, and the Presence
Detect State bit and its associated interrupt mechanism be used
exclusively for OOB PD. As a substitute for in-band PD with async
hot-plug, the reference model uses either the DPC or the DLL Link Active
mechanism.
Link: https://lore.kernel.org/r/20191025190047.38130-2-stuart.w.hayes@gmail.com
[bhelgaas: move PCI_EXP_SLTCAP2 read earlier & print PCI_EXP_SLTCAP2_IBPD
value (suggested by Lukas)]
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Reviewed-by: Lukas Wunner <lukas@wunner.de>
This addresses deadlocks in these common cases in hierarchies containing
two switches:
- All involved ports are runtime suspended and they are unplugged. This
can happen easily if the drivers involved automatically enable runtime
PM (xHCI for example does that).
- System is suspended (e.g., closing the lid on a laptop) with a dock +
something else connected, and the dock is unplugged while suspended.
These cases lead to the following deadlock:
INFO: task irq/126-pciehp:198 blocked for more than 120 seconds.
irq/126-pciehp D 0 198 2 0x80000000
Call Trace:
schedule+0x2c/0x80
schedule_timeout+0x246/0x350
wait_for_completion+0xb7/0x140
kthread_stop+0x49/0x110
free_irq+0x32/0x70
pcie_shutdown_notification+0x2f/0x50
pciehp_remove+0x27/0x50
pcie_port_remove_service+0x36/0x50
device_release_driver+0x12/0x20
bus_remove_device+0xec/0x160
device_del+0x13b/0x350
device_unregister+0x1a/0x60
remove_iter+0x1e/0x30
device_for_each_child+0x56/0x90
pcie_port_device_remove+0x22/0x40
pcie_portdrv_remove+0x20/0x60
pci_device_remove+0x3e/0xc0
device_release_driver_internal+0x18c/0x250
device_release_driver+0x12/0x20
pci_stop_bus_device+0x6f/0x90
pci_stop_bus_device+0x31/0x90
pci_stop_and_remove_bus_device+0x12/0x20
pciehp_unconfigure_device+0x88/0x140
pciehp_disable_slot+0x6a/0x110
pciehp_handle_presence_or_link_change+0x263/0x400
pciehp_ist+0x1c9/0x1d0
irq_thread_fn+0x24/0x60
irq_thread+0xeb/0x190
kthread+0x120/0x140
INFO: task irq/190-pciehp:2288 blocked for more than 120 seconds.
irq/190-pciehp D 0 2288 2 0x80000000
Call Trace:
__schedule+0x2a2/0x880
schedule+0x2c/0x80
schedule_preempt_disabled+0xe/0x10
mutex_lock+0x2c/0x30
pci_lock_rescan_remove+0x15/0x20
pciehp_unconfigure_device+0x4d/0x140
pciehp_disable_slot+0x6a/0x110
pciehp_handle_presence_or_link_change+0x263/0x400
pciehp_ist+0x1c9/0x1d0
irq_thread_fn+0x24/0x60
irq_thread+0xeb/0x190
kthread+0x120/0x140
What happens here is that the whole hierarchy is runtime resumed and the
parent PCIe downstream port, which got the hot-remove event, starts
removing devices below it, taking pci_lock_rescan_remove() lock. When the
child PCIe port is runtime resumed it calls pciehp_check_presence() which
ends up calling pciehp_card_present() and pciehp_check_link_active(). Both
of these use pcie_capability_read_word(), which notices that the underlying
device is already gone and returns PCIBIOS_DEVICE_NOT_FOUND with the
capability value set to 0. When pciehp gets this value it thinks that its
child device is also hot-removed and schedules its IRQ thread to handle the
event.
The deadlock happens when the child's IRQ thread runs and tries to acquire
pci_lock_rescan_remove() which is already taken by the parent and the
parent waits for the child's IRQ thread to finish.
Prevent this from happening by checking the return value of
pcie_capability_read_word() and if it is PCIBIOS_DEVICE_NOT_FOUND stop
performing any hot-removal activities.
[bhelgaas: add common scenarios to commit log]
Link: https://lore.kernel.org/r/20191029170022.57528-2-mika.westerberg@linux.intel.com
Tested-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Infinite timeout loops are hard to read. Refactor it to plausible 'do {}
while ()'.
Note, the supplied timeout can't be negative for current use, though if
it's not dividable to 10, we may go below 0, that's why type of the
parameter is int. And thus, we may move the check to the loop condition.
No functional change intended.
Link: https://lore.kernel.org/r/20191108111855.85866-1-andriy.shevchenko@linux.intel.com
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Andrew Murray <andrew.murray@arm.com>
A sysfs request to enable or disable a PCIe hotplug slot should not
return before it has been carried out. That is sought to be achieved by
waiting until the controller's "pending_events" have been cleared.
However the IRQ thread pciehp_ist() clears the "pending_events" before
it acts on them. If pciehp_sysfs_enable_slot() / _disable_slot() happen
to check the "pending_events" after they have been cleared but while
pciehp_ist() is still running, the functions may return prematurely
with an incorrect return value.
Fix by introducing an "ist_running" flag which must be false before a sysfs
request is allowed to return.
Fixes: 32a8cef274 ("PCI: pciehp: Enable/disable exclusively from IRQ thread")
Link: https://lore.kernel.org/linux-pci/1562226638-54134-1-git-send-email-wangxiongfeng2@huawei.com
Link: https://lore.kernel.org/r/4174210466e27eb7e2243dd1d801d5f75baaffd8.1565345211.git.lukas@wunner.de
Reported-and-tested-by: Xiongfeng Wang <wangxiongfeng2@huawei.com>
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Cc: stable@vger.kernel.org # v4.19+
Remove pciehp_green_led_{on,off,blink}() and use pciehp_set_indicators()
instead, since the code is mostly the same.
[bhelgaas: drop set_power_indicator() wrapper to reduce the number of
interfaces]
Link: https://lore.kernel.org/r/20190903111021.1559-5-efremov@linux.com
Signed-off-by: Denis Efremov <efremov@linux.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Remove pciehp_set_attention_status() and use pciehp_set_indicators()
instead, since the code is mostly the same.
Link: https://lore.kernel.org/r/20190903111021.1559-4-efremov@linux.com
Signed-off-by: Denis Efremov <efremov@linux.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Combine adjacent updates of power and attention indicators into a single
pciehp_set_indicators() call. This sends one command to the hotplug
controller instead of two.
Link: https://lore.kernel.org/r/20190903111021.1559-3-efremov@linux.com
Signed-off-by: Denis Efremov <efremov@linux.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Reviewed-by: Lukas Wunner <lukas@wunner.de>
Add pciehp_set_indicators() to set power and attention indicators with a
single register write.
This is a minor optimization because we frequently set both indicators and
this can do it with a single command. It also reduces the number of
interfaces related to the indicators and makes them more discoverable
because callers use the PCI_EXP_SLTCTL_ATTN_IND_* and
PCI_EXP_SLTCTL_PWR_IND_* definitions directly.
[bhelgaas: extend commit log, s/PCI_EXP_SLTCTL_.*_IND_NONE/INDICATOR_NOOP/
so they don't look like things defined by the spec, add function doc, mask
commands to make it obvious we only send valid commands
(pcie_do_write_cmd() does mask it, but requires more effort to verify)]
Link: https://lore.kernel.org/r/20190903111021.1559-2-efremov@linux.com
Signed-off-by: Denis Efremov <efremov@linux.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
MY_NAME is only used once and offers no benefit, so remove it.
Link: https://lore.kernel.org/lkml/20190509141456.223614-11-helgaas@kernel.org
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Keith Busch <keith.busch@intel.com>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Replace the last uses of dbg() with the equivalent pr_debug(), then remove
unused dbg(), err(), info(), and warn() wrappers.
Link: https://lore.kernel.org/lkml/20190509141456.223614-9-helgaas@kernel.org
Signed-off-by: Frederick Lawler <fred@fredlawl.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Keith Busch <keith.busch@intel.com>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Log messages with pci_dev, not pcie_device. Factor out common message
prefixes with dev_fmt().
Example output change:
- pciehp 0000:00:06.0:pcie004: Slot(0) Powering on due to button press
+ pcieport 0000:00:06.0: pciehp: Slot(0) Powering on due to button press
Link: https://lore.kernel.org/lkml/20190509141456.223614-8-helgaas@kernel.org
Signed-off-by: Frederick Lawler <fred@fredlawl.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Keith Busch <keith.busch@intel.com>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
We're about to convert pciehp to the dyndbg mechanism, which means we can
eventually remove pciehp_debug.
Replace uses of pciehp_debug with dbg() and ctrl_dbg(), which check
pciehp_debug internally.
Link: https://lore.kernel.org/lkml/20190509141456.223614-6-helgaas@kernel.org
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Keith Busch <keith.busch@intel.com>
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
- Blacklist Gigabyte X299 Root Port power management to fix Thunderbolt
hotplug (Mika Westerberg)
- Revert runtime PM suspend/resume callbacks that broke PME on network
cable plug (Mika Westerberg)
- Disable Data Link State Changed interrupts to prevent wakeup
immediately after suspend (Mika Westerberg)
* pci/pm:
PCI/PME: Fix possible use-after-free on remove
PCI/PME: Fix hotplug/sysfs remove deadlock in pcie_pme_remove()
PCI: pciehp: Disable Data Link Layer State Changed event on suspend
Revert "PCI/PME: Implement runtime PM callbacks"
PCI: Blacklist power management of Gigabyte X299 DESIGNARE EX PCIe ports
- Mark expected switch fall-through (Mathieu Malaterre)
- Use of_node_name_eq() for node name comparisons (Rob Herring)
- Add ACS and pciehp quirks for HXT SD4800 (Shunyong Yang)
- Consolidate Rohm Vendor ID definitions (Andy Shevchenko)
- Use u32 (not __u32) for things not exposed to userspace (Logan
Gunthorpe)
- Fix locking semantics of bus and slot reset interfaces (Alex
Williamson)
- Update PCIEPORTBUS Kconfig help text (Hou Zhiqiang)
* pci/misc:
PCI: Update PCIEPORTBUS Kconfig help text
PCI: Fix "try" semantics of bus and slot reset
PCI: Clean up usage of __u32 type
genirq/msi: Clean up usage of __u8/__u16 types
PCI: Move Rohm Vendor ID to generic list
PCI: pciehp: Add HXT quirk for Command Completed errata
PCI: Add ACS quirk for HXT SD4800
PCI: Add HXT vendor ID
PCI: Use of_node_name_eq() for node name comparisons
PCI: Mark expected switch fall-through
Commit 0e157e5286 ("PCI/PME: Implement runtime PM callbacks") tried to
solve an issue where the hierarchy immediately wakes up when it is
transitioned into D3cold. However, it turns out to prevent PME
propagation on some systems that do not support D3cold.
I looked more closely at what might cause the immediate wakeup. It happens
when the ACPI power resource of the root port is turned off. The AML code
associated with the _OFF() method of the ACPI power resource starts a PCIe
L2/L3 Ready transition and waits for it to complete. Right after the L2/L3
Ready transition is started the root port receives a PME from the
downstream port.
The simplest hierarchy where this happens looks like this:
00:1d.0 PCIe Root Port
^
|
v
05:00.0 PCIe switch #1 upstream port
06:01.0 PCIe switch #1 downstream hotplug port
^
|
v
08:00.0 PCIe switch #2 upstream port
It seems that the PCIe link between the two switches, before
PME_Turn_Off/PME_TO_Ack is complete for the whole hierarchy, goes
inactive and triggers PME towards the root port bringing it back to D0.
The L2/L3 Ready sequence is described in PCIe r4.0 spec sections 5.2 and
5.3.3 but unfortunately they do not state what happens if DLLSCE is
enabled during the sequence.
Disabling Data Link Layer State Changed event (DLLSCE) seems to prevent
the issue and still allows the downstream hotplug port to notice when a
device is plugged/unplugged.
Link: https://bugzilla.kernel.org/show_bug.cgi?id=202593
Fixes: 0e157e5286 ("PCI/PME: Implement runtime PM callbacks")
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
CC: stable@vger.kernel.org # v4.20+
The HXT SD4800 PCI controller does not set the Command Completed bit unless
writes to the Slot Command register change "Control" bits.
Add SD4800 to the quirk.
Signed-off-by: Shunyong Yang <shunyong.yang@hxt-semitech.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
CC: Joey Zheng <yu.zheng@hxt-semitech.com>
Shameerali reported that running v4.20-rc1 as QEMU guest, the PCIe hotplug
port times out during boot:
pciehp 0000:00:01.0:pcie004: Timeout on hotplug command 0x03f1 (issued 1016 msec ago)
pciehp 0000:00:01.0:pcie004: Timeout on hotplug command 0x03f1 (issued 1024 msec ago)
pciehp 0000:00:01.0:pcie004: Failed to check link status
pciehp 0000:00:01.0:pcie004: Timeout on hotplug command 0x02f1 (issued 2520 msec ago)
The issue was bisected down to commit 720d6a671a ("PCI: pciehp: Do not
handle events if interrupts are masked") and was further analyzed by the
reporter to be caused by the fact that pciehp first updates the hardware
and only then cache the ctrl->slot_ctrl in pcie_do_write_cmd(). If the
interrupt happens before we cache the value, pciehp_isr() reads value 0 and
decides that the interrupt was not meant for it causing the above timeout
to trigger.
Fix by moving ctrl->slot_ctrl assignment to happen before it is written to
the hardware.
Fixes: 720d6a671a ("PCI: pciehp: Do not handle events if interrupts are masked")
Link: https://lore.kernel.org/linux-pci/5FC3163CFD30C246ABAA99954A238FA8387DD344@FRAEML521-MBX.china.huawei.com
Reported-by: Shameerali Kolothum Thodi <shameerali.kolothum.thodi@huawei.com>
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
PCIe native hotplug shares MSI vector with native PME so the interrupt
handler might get called even the hotplug interrupt is masked. In that case
we should not handle any events because the interrupt was not meant for us.
Modify the PCIe hotplug interrupt handler to check this accordingly and
bail out if it finds out that the interrupt was not about hotplug.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Lukas Wunner <lukas@wunner.de>
When PCIe hotplug port is transitioned into D3hot, the link to the
downstream component will go down. If hotplug interrupt generation is
enabled when that happens, it will trigger immediately, waking up the
system and bringing the link back up.
To prevent this, disable hotplug interrupt generation when system suspend
is entered. This does not prevent wakeup from low power states according
to PCIe 4.0 spec section 6.7.3.4:
Software enables a hot-plug event to generate a wakeup event by
enabling software notification of the event as described in Section
6.7.3.1. Note that in order for software to disable interrupt generation
while keeping wakeup generation enabled, the Hot-Plug Interrupt Enable
bit must be cleared.
So as long as we have set the slot event mask accordingly, wakeup should
work even if slot interrupt is disabled. The port should trigger wake and
then send PME to the root port when the PCIe hierarchy is brought back up.
Limit this to systems using native PME mechanism to make sure older Apple
systems depending on commit e3354628c376 ("PCI: pciehp: Support interrupts
sent from D3hot") still continue working.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
The spec has timing requirements when waiting for a link to become active
after a conventional reset. Implement those hard delays when waiting for
an active link so pciehp and dpc drivers don't need to duplicate this.
For devices that don't support data link layer active reporting, wait the
fixed time recommended by the PCIe spec.
Signed-off-by: Keith Busch <keith.busch@intel.com>
[bhelgaas: changelog]
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Sinan Kaya <okaya@kernel.org>
When the PCI hotplug core and its first user, cpqphp, were introduced in
February 2002 with historic commit a8a2069f432c, cpqphp allocated a slot
struct for its internal use plus a hotplug_slot struct to be registered
with the hotplug core and linked the two with pointers:
https://git.kernel.org/tglx/history/c/a8a2069f432c
Nowadays, the predominant pattern in the tree is to embed ("subclass")
such structures in one another and cast to the containing struct with
container_of(). But it wasn't until July 2002 that container_of() was
introduced with historic commit ec4f214232cf:
https://git.kernel.org/tglx/history/c/ec4f214232cf
pnv_php, introduced in 2016, did the right thing and embedded struct
hotplug_slot in its internal struct pnv_php_slot, but all other drivers
cargo-culted cpqphp's design and linked separate structs with pointers.
Embedding structs is preferrable to linking them with pointers because
it requires fewer allocations, thereby reducing overhead and simplifying
error paths. Casting an embedded struct to the containing struct
becomes a cheap subtraction rather than a dereference. And having fewer
pointers reduces the risk of them pointing nowhere either accidentally
or due to an attack.
Convert all drivers to embed struct hotplug_slot in their internal slot
struct. The "private" pointer in struct hotplug_slot thereby becomes
unused, so drop it.
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Tyrel Datwyler <tyreld@linux.vnet.ibm.com> # drivers/pci/hotplug/rpa*
Acked-by: Sebastian Ott <sebott@linux.ibm.com> # drivers/pci/hotplug/s390*
Acked-by: Andy Shevchenko <andy.shevchenko@gmail.com> # drivers/platform/x86
Cc: Len Brown <lenb@kernel.org>
Cc: Scott Murray <scott@spiteful.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Oliver OHalloran <oliveroh@au1.ibm.com>
Cc: Gavin Shan <gwshan@linux.vnet.ibm.com>
Cc: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Cc: Corentin Chary <corentin.chary@gmail.com>
Cc: Darren Hart <dvhart@infradead.org>
Of the members which were just moved from pciehp's slot struct to the
controller struct, rename "lock" to "state_lock" and rename "work" to
"button_work" for clarity. Perform the rename separately to the
unification of the two structs per Sinan's request.
No functional change intended.
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Cc: Sinan Kaya <okaya@kernel.org>
pciehp was originally introduced together with shpchp in a single
commit, c16b4b14d980 ("PCI Hotplug: Add SHPC and PCI Express hot-plug
drivers"):
https://git.kernel.org/tglx/history/c/c16b4b14d980
shpchp supports up to 31 slots per controller, hence uses separate slot
and controller structs. pciehp has a 1:1 relationship between slot and
controller and therefore never required this separation. Nevertheless,
because much of the code had been copy-pasted between the two drivers,
pciehp likewise uses separate structs to this very day.
The artificial separation of data structures adds unnecessary complexity
and bloat to pciehp and requires constantly chasing pointers at runtime.
Simplify the driver by merging struct slot into struct controller.
Merge the slot constructor pcie_init_slot() and the destructor
pcie_cleanup_slot() into the controller counterparts.
No functional change intended.
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
The WiGig Bus Extension (WBE) specification allows tunneling PCIe over
IEEE 802.11. A product implementing this spec is the wil6210 from
Wilocity (now part of Qualcomm Atheros). It integrates a PCIe switch
with a wireless network adapter:
00.0-+ [1ae9:0101] Upstream Port
+-00.0-+ [1ae9:0200] Downstream Port
| +-00.0 [168c:0034] Atheros AR9462 Wireless Network Adapter
+-02.0 [1ae9:0201] Downstream Port
+-03.0 [1ae9:0201] Downstream Port
Wirelessly attached devices presumably appear below the hotplug ports
with device ID [1ae9:0201]. Oddly, the Downstream Port [1ae9:0200]
leading to the wireless network adapter is likewise Hotplug Capable,
but has its Presence Detect State bit hardwired to zero. Even if the
Link Active bit is set, Presence Detect is zero, so this cannot be
caused by in-band presence detection but only by broken hardware.
pciehp assumes an empty slot if Presence Detect State is zero,
regardless of Link Active being one. Consequently, up until v4.18 it
removes the wireless network adapter in pciehp_resume(). From v4.19 it
already does so in pciehp_probe().
Be lenient towards broken hardware and assume the slot is occupied if
Link Active is set: Introduce pciehp_card_present_or_link_active()
and use it in lieu of pciehp_get_adapter_status() everywhere, except
in pciehp_handle_presence_or_link_change() whose log messages depend
on which of Presence Detect State or Link Active is set.
Remove the Presence Detect State check from __pciehp_enable_slot()
because it is only called if either of Presence Detect State or Link
Active is set.
Caution: There is a possibility that broken hardware exists which has
working Presence Detect but hardwires Link Active to one. On such
hardware the slot will now incorrectly be considered always occupied.
If such hardware is discovered, this commit can be rolled back and a
quirk can be added which sets is_hotplug_bridge = 0 for [1ae9:0200].
Link: https://bugzilla.kernel.org/show_bug.cgi?id=200839
Reported-and-tested-by: David Yang <mmyangfl@gmail.com>
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Cc: Rajat Jain <rajatja@google.com>
Cc: Ashok Raj <ashok.raj@intel.com>
pciehp's ->enable_slot, ->disable_slot, ->get_attention_status and
->reset_slot callbacks are currently implemented by wrapper functions
that do nothing else but call down to a backend function. The backends
are not called from anywhere else, so drop the wrappers and use the
backends directly as callbacks, thereby shaving off a few lines of
unnecessary code.
No functional change intended.
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Drop the following includes from pciehp source files which no longer use
any of the included symbols:
* <linux/sched/signal.h> in pciehp.h
<linux/signal.h> in pciehp_hpc.c
Added by commit de25968cc8 ("fix more missing includes") to
accommodate for a call to signal_pending().
The call was removed by commit 262303fe32 ("pciehp: fix wait command
completion").
* <linux/interrupt.h> in pciehp_core.c
Added by historic commit f308a2dfbe63 ("PCI: add PCI Express Port Bus
Driver subsystem") to accommodate for a call to free_irq():
https://git.kernel.org/tglx/history/c/f308a2dfbe63
The call was removed by commit 407f452b05 ("pciehp: remove
unnecessary free_irq").
* <linux/time.h> in pciehp_core.c and pciehp_hpc.c
Added by commit 34d03419f0 ("PCIEHP: Add Electro Mechanical
Interlock (EMI) support to the PCIE hotplug driver."),
which was reverted by commit bd3d99c170 ("PCI: Remove untested
Electromechanical Interlock (EMI) support in pciehp.").
* <linux/module.h> in pciehp_ctrl.c, pciehp_hpc.c and pciehp_pci.c
Added by historic commit c16b4b14d980 ("PCI Hotplug: Add SHPC and PCI
Express hot-plug drivers"):
https://git.kernel.org/tglx/history/c/c16b4b14d980
Module-related symbols were neither used back then in those files,
nor are they used today.
* <linux/slab.h> in pciehp_ctrl.c
Added by commit 5a0e3ad6af ("include cleanup: Update gfp.h and
slab.h includes to prepare for breaking implicit slab.h inclusion from
percpu.h") to accommodate for calls to kmalloc().
The calls were removed by commit 0e94916e60 ("PCI: pciehp: Handle
events synchronously").
* "../pci.h" in pciehp_ctrl.c
Added by historic commit 67f4660b72f2 ("PCI: ASPM patch for") to
accommodate for usage of the global variable pcie_mch_quirk:
https://git.kernel.org/tglx/history/c/67f4660b72f2
The global variable was removed by commit 0ba379ec0f ("PCI: Simplify
hotplug mch quirk").
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
If both hot-add and power fault were observed in a single interrupt, we
handled the hot-add first, then the power fault, in this path:
pciehp_ist
if (events & (PDC | DLLSC))
pciehp_handle_presence_or_link_change
case OFF_STATE:
pciehp_enable_slot
__pciehp_enable_slot
board_added
pciehp_power_on_slot
ctrl->power_fault_detected = 0
pcie_write_cmd(ctrl, PCI_EXP_SLTCTL_PWR_ON, PCI_EXP_SLTCTL_PCC)
pciehp_green_led_on(p_slot) # power LED on
pciehp_set_attention_status(p_slot, 0) # attention LED off
if ((events & PFD) && !ctrl->power_fault_detected)
ctrl->power_fault_detected = 1
pciehp_set_attention_status(1) # attention LED on
pciehp_green_led_off(slot) # power LED off
This left the attention indicator on (even though the hot-add succeeded)
and the power indicator off (even though the slot power was on).
Fix this by checking for power faults before checking for new devices.
Prior to 0e94916e60, this was successful because everything was chained
through work queues and the order was:
INT_PRESENCE_ON -> INT_POWER_FAULT -> ENABLE_REQ
The ENABLE_REQ cleared the power fault at the end, but now everything is
handled inline with the interrupt thread, such that the work ENABLE_REQ was
doing happens before power fault handling now.
Fixes: 0e94916e60 ("PCI: pciehp: Handle events synchronously")
Signed-off-by: Keith Busch <keith.busch@intel.com>
[bhelgaas: changelog]
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Lukas Wunner <lukas@wunner.de>
- To avoid bus errors, enable PASID only if entire path supports End-End
TLP prefixes (Sinan Kaya)
- Unify slot and bus reset functions and remove hotplug knowledge from
callers (Sinan Kaya)
- Add Function-Level Reset quirks for Intel and Samsung NVMe devices to
fix guest reboot issues (Alex Williamson)
- Add function 1 DMA alias quirk for Marvell 88SS9183 PCIe SSD Controller
(Bjorn Helgaas)
* pci/virtualization:
PCI: Add function 1 DMA alias quirk for Marvell 88SS9183
PCI: Delay after FLR of Intel DC P3700 NVMe
PCI: Disable Samsung SM961/PM961 NVMe before FLR
PCI: Export pcie_has_flr()
PCI: Rename pci_try_reset_bus() to pci_reset_bus()
PCI: Deprecate pci_reset_bus() and pci_reset_slot() functions
PCI: Unify try slot and bus reset API
PCI: Hide pci_reset_bridge_secondary_bus() from drivers
IB/hfi1: Use pci_try_reset_bus() for initiating PCI Secondary Bus Reset
PCI: Handle error return from pci_reset_bridge_secondary_bus()
PCI/IOV: Tidy pci_sriov_set_totalvfs()
PCI: Enable PASID only if entire path supports End-End TLP prefixes
# Conflicts:
# drivers/pci/hotplug/pciehp_hpc.c
On driver probe and on resume from system sleep, pciehp checks the
Presence Detect State bit in the Slot Status register to bring up an
occupied slot or bring down an unoccupied slot. Both code paths are
identical, so deduplicate them per Mika's request.
On probe, an additional check is performed to disable power of an
unoccupied slot. This can e.g. happen if power was enabled by BIOS.
It cannot happen once pciehp has taken control, hence is not necessary
on resume: The Slot Control register is set to the same value that it
had on suspend by pci_restore_state(), so if the slot was occupied,
power is enabled and if it wasn't, power is disabled. Should occupancy
have changed during the system sleep transition, power is adjusted by
bringing up or down the slot per the paragraph above.
To allow for deduplication of the presence check, move the power check
to pcie_init(). This seems safer anyway, because right now it is
performed while interrupts are already enabled, and although I can't
think of a scenario where pciehp_power_off_slot() and the IRQ thread
collide, it does feel brittle.
However this means that pcie_init() may now write to the Slot Control
register before the IRQ is requested. If both the CCIE and HPIE bits
happen to be set, pcie_wait_cmd() will wait for an interrupt (instead
of polling the Command Completed bit) and eventually emit a timeout
message. Additionally, if a level-triggered INTx interrupt is used,
the user may see a spurious interrupt splat. Avoid by disabling
interrupts before disabling power. (Normally the HPIE and CCIE bits
should be clear on probe, but conceivably they may already have been
set e.g. by BIOS.)
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Ensure accessibility of a hotplug port's config space when accessed via
sysfs by resuming its parent to D0.
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Cc: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Cc: Mika Westerberg <mika.westerberg@linux.intel.com>
Cc: Ashok Raj <ashok.raj@intel.com>
Cc: Keith Busch <keith.busch@intel.com>
Cc: Yinghai Lu <yinghai@kernel.org>
If a hotplug port is able to send an interrupt, one would naively assume
that it is accessible at that moment. After all, if it wouldn't be
accessible, i.e. if its parent is in D3hot and the link to the hotplug
port is thus down, how should an interrupt come through?
It turns out that assumption is wrong at least for Thunderbolt: Even
though its parents are in D3hot, a Thunderbolt hotplug port is able to
signal interrupts. Because the port's config space is inaccessible and
resuming the parents may sleep, the hard IRQ handler has to defer
runtime resuming the parents and reading the Slot Status register to the
IRQ thread.
If the hotplug port uses a level-triggered INTx interrupt, it needs to
be masked until the IRQ thread has cleared the signaled events. For
simplicity, this commit also masks edge-triggered MSI/MSI-X interrupts.
Note that if the interrupt is shared (which can only happen for INTx),
other devices are starved from receiving interrupts until the IRQ thread
is scheduled, has runtime resumed the hotplug port's parents and has
read and cleared the Slot Status register.
That delay is dominated by the 10 ms D3hot->D0 transition time of each
parent port. The worst case is a Thunderbolt downstream port at the
end of a daisy chain: There may be up to six Thunderbolt controllers
in-between it and the root port, each comprising an upstream and
downstream port, plus its own upstream port. That's 13 x 10 = 130 ms.
Possible mitigations are polling the interrupt while it's disabled or
reducing the d3_delay of Thunderbolt ports if possible.
Open code masking of the interrupt instead of requesting it with the
IRQF_ONESHOT flag to minimize the period during which it is masked.
(IRQF_ONESHOT unmasks the IRQ only after the IRQ thread has finished.)
PCIe r4.0 sec 6.7.3.4 states that "If wake generation is required by the
associated form factor specification, a hotplug capable Downstream Port
must support generation of a wakeup event (using the PME mechanism) on
hotplug events that occur when the system is in a sleep state or the
Port is in device state D1, D2, or D3Hot."
This would seem to imply that PME needs to be enabled on the hotplug
port when it is runtime suspended. pci_enable_wake() currently doesn't
enable PME on bridges, it may be necessary to add an exemption for
hotplug bridges there. On "Light Ridge" Thunderbolt controllers, the
PME_Status bit is not set when an interrupt occurs while the hotplug
port is in D3hot, even if PME is enabled. (I've tested this on a Mac
and we hardcode the OSC_PCI_EXPRESS_PME_CONTROL bit to 0 on Macs in
negotiate_os_control(), modifying it to 1 didn't change the behavior.)
(Side note: Section 6.7.3.4 also states that "PME and Hot-Plug Event
interrupts (when both are implemented) always share the same MSI or
MSI-X vector". That would only seem to apply to Root Ports, however
the section never mentions Root Ports, only Downstream Ports. This is
explained in the definition of "Downstream Port" in the "Terms and
Acronyms" section of the PCIe Base Spec: "The Ports on a Switch that
are not the Upstream Port are Downstream Ports. All Ports on a Root
Complex are Downstream Ports.")
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Cc: Mika Westerberg <mika.westerberg@linux.intel.com>
Cc: Ashok Raj <ashok.raj@intel.com>
Cc: Keith Busch <keith.busch@intel.com>
Cc: Yinghai Lu <yinghai@kernel.org>
Thunderbolt hotplug ports that were occupied before system sleep resume
with their downstream link in "off" state. Only after the Thunderbolt
controller has reestablished the PCIe tunnels does the link go up.
As a result, a spurious Presence Detect Changed and/or Data Link Layer
State Changed event occurs.
The events are not immediately acted upon because tunnel reestablishment
happens in the ->resume_noirq phase, when interrupts are still disabled.
Also, notification of events may initially be disabled in the Slot
Control register when coming out of system sleep and is reenabled in the
->resume_noirq phase through:
pci_pm_resume_noirq()
pci_pm_default_resume_early()
pci_restore_state()
pci_restore_pcie_state()
It is not guaranteed that the events are acted upon at all: PCIe r4.0,
sec 6.7.3.4 says that "a port may optionally send an MSI when there are
hot-plug events that occur while interrupt generation is disabled, and
interrupt generation is subsequently enabled." Note the "optionally".
If an MSI is sent, pciehp will gratuitously turn the slot off and back
on once the ->resume_early phase has commenced.
If an MSI is not sent, the extant, unacknowledged events in the Slot
Status register will prevent future notification of presence or link
changes.
Commit 13c65840fe ("PCI: pciehp: Clear Presence Detect and Data Link
Layer Status Changed on resume") fixed the latter by clearing the events
in the ->resume phase. Move this to the ->resume_noirq phase to also
fix the gratuitous disable/enablement of the slot.
The commit further restored the Slot Control register in the ->resume
phase, but that's dispensable because as shown above it's already been
done in the ->resume_noirq phase.
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Cc: Mika Westerberg <mika.westerberg@linux.intel.com>
The ->reset_slot callback introduced by commits:
2e35afaefe ("PCI: pciehp: Add reset_slot() method") and
06a8d89af5 ("PCI: pciehp: Disable link notification across slot reset")
disables notification of Presence Detect Changed and Data Link Layer
State Changed events for the duration of a secondary bus reset.
However a bus reset not only triggers these events, but may also clear
the Presence Detect State bit in the Slot Status register and the Data
Link Layer Link Active bit in the Link Status register momentarily.
According to Sinan Kaya:
"I know for a fact that bus reset clears the Data Link Layer Active bit
as soon as link goes down. It gets set again following link up.
Presence detect depends on the HW implementation. QDT root ports
don't change presence detect for instance since nobody actually
removed the card. If an implementation supports in-band presence
detect, the answer is yes. As soon as the link goes down, presence
detect bit will get cleared until recovery."
https://lkml.kernel.org/r/42e72f83-3b24-f7ef-e5bc-290fae99259a@codeaurora.org
In-band presence detect is also covered in Table 4-15 in PCIe r4.0,
sec 4.2.6.
pciehp should therefore ensure that any parts of the driver that access
those bits do not run concurrently to a bus reset. The only precaution
the commits took to that effect was to halt interrupt polling. They
made no effort to drain the slot workqueue, cancel an outstanding
Attention Button work, or block slot enable/disable requests via sysfs
and in the ->probe hook.
Now that pciehp is converted to enable/disable the slot exclusively from
the IRQ thread, the only places accessing the two above-mentioned bits
are the IRQ thread and the ->probe hook. Add locking to serialize them
with a bus reset. This obviates the need to halt interrupt polling.
Do not add locking to the ->get_adapter_status sysfs callback to afford
users unfettered access to that bit. Use an rw_semaphore in lieu of a
regular mutex to allow parallel execution of the non-reset code paths
accessing the critical bits, i.e. the IRQ thread and the ->probe hook.
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Cc: Rajat Jain <rajatja@google.com>
Cc: Alex Williamson <alex.williamson@redhat.com>
Cc: Sinan Kaya <okaya@kernel.org>
Per PCIe r4.0, sec 6.7.3.4, a "port may optionally send an MSI when
there are hot-plug events that occur while interrupt generation is
disabled, and interrupt generation is subsequently enabled."
On probe, we currently clear all event bits in the Slot Status register
with the notable exception of the Presence Detect Changed bit. Thereby
we seek to receive an interrupt for an already occupied slot once event
notification is enabled.
But because the interrupt is optional, users may have to specify the
pciehp_force parameter on the command line, which is inconvenient.
Moreover, now that pciehp's event handling has become resilient to
missed events, a Presence Detect Changed interrupt for a slot which is
powered on is interpreted as removal of the card. If the slot has
already been brought up by the BIOS, receiving such an interrupt on
probe causes the slot to be powered off and immediately back on, which
is likewise undesirable.
Avoid both issues by making the behavior of pciehp_force the default and
clearing the Presence Detect Changed bit on probe.
Note that the stated purpose of pciehp_force per the MODULE_PARM_DESC
("Force pciehp, even if OSHP is missing") seems nonsensical because the
OSHP control method is only relevant for SHCP slots according to the
PCI Firmware specification r3.0, sec 4.8.
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Cc: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Cc: Mika Westerberg <mika.westerberg@linux.intel.com>
A hotplug port's Slot Status register does not count how often each type
of event occurred, it only records the fact *that* an event has occurred.
Previously pciehp queued a work item for each event. But if it missed
an event, e.g. removal of a card in-between two back-to-back insertions,
it queued up the wrong work item or no work item at all. Commit
fad214b0aa ("PCI: pciehp: Process all hotplug events before looking
for new ones") sought to improve the situation by shrinking the window
during which events may be missed.
But Stefan Roese reports unbalanced Card present and Link Up events,
suggesting that we're still missing events if they occur very rapidly.
Bjorn Helgaas responds that he considers pciehp's event handling
"baroque" and calls for its simplification and rationalization:
https://lkml.kernel.org/r/20180202192045.GA53759@bhelgaas-glaptop.roam.corp.google.com
It gets worse once a hotplug port is runtime suspended: The port can
signal an interrupt while it and its parents are in D3hot, i.e. while
it is inaccessible. By the time we've runtime resumed all parents to D0
and read the port's Slot Status register, we may have missed an arbitrary
number of events. Event handling therefore needs to be reworked to
become resilient to missed events.
Assume that a Presence Detect Changed event has occurred.
Consider the following truth table:
- Slot is in OFF_STATE and is currently empty. => Do nothing.
(The event is trailing a Link Down or we've
missed an insertion and subsequent removal.)
- Slot is in OFF_STATE and is currently occupied. => Turn the slot on.
- Slot is in ON_STATE and is currently empty. => Turn the slot off.
- Slot is in ON_STATE and is currently occupied. => Turn the slot off,
(Be cautious and assume the card in then back on.
the slot isn't the same as before.)
This leads to the following simple algorithm:
1 If the slot is in ON_STATE, turn it off unconditionally.
2 If the slot is currently occupied, turn it on.
Because those actions are now carried out synchronously, rather than by
scheduled work items, pciehp reacts to the *current* situation and
missed events no longer matter.
Data Link Layer State Changed events can be handled identically to
Presence Detect Changed events. Note that in the above truth table,
a Link Up trailing a Card present event didn't have to be accounted for:
It is filtered out by pciehp_check_link_status().
As for Attention Button Pressed events, PCIe r4.0, sec 6.7.1.5 says:
"Once the Power Indicator begins blinking, a 5-second abort interval
exists during which a second depression of the Attention Button cancels
the operation." In other words, the user can only expect the system to
react to a button press after it starts blinking. Missed button presses
that occur in-between are irrelevant.
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Cc: Stefan Roese <sr@denx.de>
Cc: Mayurkumar Patel <mayurkumar.patel@intel.com>
Cc: Mika Westerberg <mika.westerberg@linux.intel.com>
Cc: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
When a device is hotplugged, Presence Detect and Link Up events often do
not occur simultaneously, but with a lag of a few milliseconds. Only
the first event received is relevant, the other one can be disregarded.
Moreover, Stefan Roese reports that on certain platforms, Link State and
Presence Detect may flap for up to 100 ms before stabilizing, suggesting
that such events should be disregarded for at least this long:
https://lkml.kernel.org/r/20180130084121.18653-1-sr@denx.de
On slot enablement, pciehp_check_link_status() waits for 100 ms per
PCIe r4.0, sec 6.7.3.3, then probes the hotplugged device's vendor
register for up to 1 second.
If this succeeds, the link is definitely up, so ignore any Presence
Detect or Link State events that occurred up to this point.
pciehp_check_link_status() then checks the Link Training bit in the
Link Status register. This is the final opportunity to detect
inaccessibility of the device and abort slot enablement. Any link
or presence change that occurs afterwards will cause the slot to be
disabled again immediately after attempting to enable it.
The astute reviewer may appreciate that achieving this behavior would be
more complicated had pciehp not just been converted to enable/disable
the slot exclusively from the IRQ thread: When the slot is enabled via
sysfs, each link or presence flap would otherwise cause the IRQ thread
to run and it would have to sense that those events are belonging to a
concurrent slot enablement operation and disregard them. It would be
much more difficult than this mere 3 line change.
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Cc: Stefan Roese <sr@denx.de>
Previously slot enablement and disablement could happen concurrently.
But now it's under the exclusive control of the IRQ thread, rendering
the locking obsolete. Drop it.
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Besides the IRQ thread, there are several other places in the driver
which enable or disable the slot:
- pciehp_probe() enables the slot if it's occupied and the pciehp_force
module parameter is used.
- pciehp_resume() enables or disables the slot after system sleep.
- pciehp_queue_pushbutton_work() enables or disables the slot after the
5 second delay following an Attention Button press.
- pciehp_sysfs_enable_slot() and pciehp_sysfs_disable_slot() enable or
disable the slot on sysfs write.
This requires locking and complicates pciehp's state machine.
A simplification can be achieved by enabling and disabling the slot
exclusively from the IRQ thread.
Amend the functions listed above to request slot enable/disablement from
the IRQ thread by either synthesizing a Presence Detect Changed event or,
in the case of a disable user request (via sysfs or an Attention Button
press), submitting a newly introduced force disable request. The latter
is needed because the slot shall be forced off despite being occupied.
For this force disable request, avoid colliding with Slot Status register
bits by using a bit number greater than 16.
For synchronous execution of requests (on sysfs write), wait for the
request to finish and retrieve the result. There can only ever be one
sysfs write in flight due to the locking in kernfs_fop_write(), hence
there is no risk of returning the result of a different sysfs request to
user space.
The POWERON_STATE and POWEROFF_STATE is now no longer entered by the
above-listed functions, but solely by the IRQ thread when it begins a
power transition. Afterwards, it moves to STATIC_STATE. The same
applies to canceling the Attention Button work, it likewise becomes an
IRQ thread only operation.
An immediate consequence is that the POWERON_STATE and POWEROFF_STATE is
never observed by the IRQ thread itself, only by functions called in a
different context, such as pciehp_sysfs_enable_slot(). So remove
handling of these states from pciehp_handle_button_press() and
pciehp_handle_link_change() which are exclusively called from the IRQ
thread.
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
handle_button_press_event() currently determines whether the slot has
been turned on or off by looking at the Power Controller Control bit in
the Slot Control register. This assumes that an attention button
implies presence of a power controller even though that's not mandated
by the spec. Moreover the Power Controller Control bit is unreliable
when a power fault occurs (PCIe r4.0, sec 6.7.1.8). This issue has
existed since the driver was introduced in 2004.
Fix by replacing STATIC_STATE with ON_STATE and OFF_STATE and tracking
whether the slot has been turned on or off. This is also a required
ingredient to make pciehp resilient to missed events, which is the
object of an upcoming commit.
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Previously the slot workqueue was used to handle events and enable or
disable the slot. That's no longer the case as those tasks are done
synchronously in the IRQ thread. The slot workqueue is thus merely used
to handle a button press after the 5 second delay and only one such work
item may be in flight at any given time. A separate workqueue isn't
necessary for this simple task, so use the system workqueue instead.
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Up until now, pciehp's IRQ handler schedules a work item for each event,
which in turn schedules a work item to enable or disable the slot. This
double indirection was necessary because sleeping wasn't allowed in the
IRQ handler.
However it is now that pciehp has been converted to threaded IRQ handling
and polling, so handle events synchronously in pciehp_ist() and remove
the work item infrastructure (with the exception of work items to handle
a button press after the 5 second delay).
For link or presence change events, move the register read to determine
the current link or presence state behind acquisition of the slot lock
to prevent it from becoming stale while the lock is contended.
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
We've just converted pciehp to threaded IRQ handling, but still cannot
sleep in pciehp_ist() because the function is also called in poll mode,
which runs in softirq context (from a timer).
Convert poll mode to a kthread so that pciehp_ist() always runs in task
context.
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
pciehp's IRQ handler queues up a work item for each event signaled by
the hardware. A more modern alternative is to let a long running
kthread service the events. The IRQ handler's sole job is then to check
whether the IRQ originated from the device in question, acknowledge its
receipt to the hardware to quiesce the interrupt and wake up the kthread.
One benefit is reduced latency to handle the IRQ, which is a necessity
for realtime environments. Another benefit is that we can make pciehp
simpler and more robust by handling events synchronously in process
context, rather than asynchronously by queueing up work items. pciehp's
usage of work items is a historic artifact, it predates the introduction
of threaded IRQ handlers by two years. (The former was introduced in
2007 with commit 5d386e1ac4 ("pciehp: Event handling rework"), the
latter in 2009 with commit 3aa551c9b4 ("genirq: add threaded interrupt
handler support").)
Convert pciehp to threaded IRQ handling by retrieving the pending events
in pciehp_isr(), saving them for later consumption by the thread handler
pciehp_ist() and clearing them in the Slot Status register.
By clearing the Slot Status (and thereby acknowledging the events) in
pciehp_isr(), we can avoid requesting the IRQ with IRQF_ONESHOT, which
would have the unpleasant side effect of starving devices sharing the
IRQ until pciehp_ist() has finished.
pciehp_isr() does not count how many times each event occurred, but
merely records the fact *that* an event occurred. If the same event
occurs a second time before pciehp_ist() is woken, that second event
will not be recorded separately, which is problematic according to
commit fad214b0aa ("PCI: pciehp: Process all hotplug events before
looking for new ones") because we may miss removal of a card in-between
two back-to-back insertions. We're about to make pciehp_ist() resilient
to missed events. The present commit regresses the driver's behavior
temporarily in order to separate the changes into reviewable chunks.
This doesn't affect regular slow-motion hotplug, only plug-unplug-plug
operations that happen in a timespan shorter than wakeup of the IRQ
thread.
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Mayurkumar Patel <mayurkumar.patel@intel.com>
Cc: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Commit b440bde74f ("PCI: Add pci_ignore_hotplug() to ignore hotplug
events for a device") iterates over the devices on a hotplug port's
subordinate bus in pciehp's IRQ handler without acquiring pci_bus_sem.
It is thus possible for a user to cause a crash by concurrently
manipulating the device list, e.g. by disabling slot power via sysfs
on a different CPU or by initiating a remove/rescan via sysfs.
This can't be fixed by acquiring pci_bus_sem because it may sleep.
The simplest fix is to avoid the list iteration altogether and just
check the ignore_hotplug flag on the port itself. This works because
pci_ignore_hotplug() sets the flag both on the device as well as on its
parent bridge.
We do lose the ability to print the name of the device blocking hotplug
in the debug message, but that's probably bearable.
Fixes: b440bde74f ("PCI: Add pci_ignore_hotplug() to ignore hotplug events for a device")
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Cc: stable@vger.kernel.org
When pciehp is unbound (e.g. on unplug of a Thunderbolt device), the
hotplug_slot struct is deregistered and thus freed before freeing the
IRQ. The IRQ handler and the work items it schedules print the slot
name referenced from the freed structure in various informational and
debug log messages, each time resulting in a quadruple dereference of
freed pointers (hotplug_slot -> pci_slot -> kobject -> name).
At best the slot name is logged as "(null)", at worst kernel memory is
exposed in logs or the driver crashes:
pciehp 0000:10:00.0:pcie204: Slot((null)): Card not present
An attacker may provoke the bug by unplugging multiple devices on a
Thunderbolt daisy chain at once. Unplugging can also be simulated by
powering down slots via sysfs. The bug is particularly easy to trigger
in poll mode.
It has been present since the driver's introduction in 2004:
https://git.kernel.org/tglx/history/c/c16b4b14d980
Fix by rearranging teardown such that the IRQ is freed first. Run the
work items queued by the IRQ handler to completion before freeing the
hotplug_slot struct by draining the work queue from the ->release_slot
callback which is invoked by pci_hp_deregister().
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Cc: stable@vger.kernel.org # v2.6.4
Lukas Wunner