PCIe hotplug has a bus per slot, so we can just use a normal
secondary bus reset. However, if a slot supports surprise removal,
a bus reset can be seen as a presence detection change triggering
a hot-remove followed by a hot-add. Disable presence detection from
triggering an interrupt or being polled around the bus reset.
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
We had an inconsistent mix of using and omitting the "extern" keyword
on function declarations in header files. This removes them all.
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
When we have a hotplug-capable PCIe port with a second hotplug-capable
PCIe port below it, removing the device below the upstream port causes
a deadlock.
The deadlock happens because we use the pciehp_wq workqueue to run
pciehp_power_thread(), which uses pciehp_disable_slot() to remove devices
below the upstream port. When we remove the downstream PCIe port, we call
pciehp_remove(), the pciehp driver's .remove() method. That calls
flush_workqueue(pciehp_wq), which deadlocks because the
pciehp_power_thread() work item is still running.
This patch avoids the deadlock by creating a workqueue for every PCIe port
and removing the single shared workqueue.
Here's the call path that leads to the deadlock:
pciehp_queue_pushbutton_work
queue_work(pciehp_wq) # queue pciehp_power_thread
...
pciehp_power_thread
pciehp_disable_slot
remove_board
pciehp_unconfigure_device
pci_stop_and_remove_bus_device
...
pciehp_remove # pciehp driver .remove method
pciehp_release_ctrl
pcie_cleanup_slot
flush_workqueue(pciehp_wq)
This is fairly urgent because it can be caused by simply unplugging a
Thunderbolt adapter, as reported by Daniel below.
[bhelgaas: changelog]
Reference: http://lkml.kernel.org/r/CAMVG2ssiRgcTD1bej2tkUUfsWmpL5eNtPcNif9va2-Gzb2u8nQ@mail.gmail.com
Reported-and-tested-by: Daniel J Blueman <daniel@quora.org>
Reviewed-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Yijing Wang <wangyijing@huawei.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
CC: stable@vger.kernel.org
module_param(bool) used to counter-intuitively take an int. In
fddd5201 (mid-2009) we allowed bool or int/unsigned int using a messy
trick.
It's time to remove the int/unsigned int option. For this version
it'll simply give a warning, but it'll break next kernel version.
Acked-by: Mauro Carvalho Chehab <mchehab@redhat.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Use non-ordered workqueue for attention button events.
Attention button events on each slot can be handled asynchronously. So
we should use non-ordered workqueue. This patch also removes ordered
workqueue in pciehp as a result.
Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
* Rename pciehp_wq to pciehp_ordered_wq and add non-ordered pciehp_wq
which is used instead of the system workqueue. This is to remove
the use of flush_scheduled_work() which is deprecated and scheduled
for removal.
* With cmwq in place, there's no point in creating workqueues lazily.
Create both pciehp_wq and pciehp_ordered_wq upfront.
* Include workqueue.h from pciehp.h.
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
One of the recent changes caused complilation of
drivers/pci/hotplug/pciehp_core.c to fail. Fix this issue.
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
After commit 852972acff (ACPI: Disable
ASPM if the platform won't provide _OSC control for PCIe) control of
the PCIe Capability Structure is unconditionally requested by
acpi_pci_root_add(), which in principle may cause problems to
happen in two ways. First, the BIOS may refuse to give control of
the PCIe Capability Structure if it is not asked for any of the
_OSC features depending on it at the same time. Second, the BIOS may
assume that control of the _OSC features depending on the PCIe
Capability Structure will be requested in the future and may behave
incorrectly if that doesn't happen. For this reason, control of
the PCIe Capability Structure should always be requested along with
control of any other _OSC features that may depend on it (ie. PCIe
native PME, PCIe native hot-plug, PCIe AER).
Rework the PCIe port driver so that (1) it checks which native PCIe
port services can be enabled, according to the BIOS, and (2) it
requests control of all these services simultaneously. In
particular, this causes pcie_portdrv_probe() to fail if the BIOS
refuses to grant control of the PCIe Capability Structure, which
means that no native PCIe port services can be enabled for the PCIe
Root Complex the given port belongs to. If that happens, ASPM is
disabled to avoid problems with mishandling it by the part of the
PCIe hierarchy for which control of the PCIe Capability Structure
has not been received.
Make it possible to override this behavior using 'pcie_ports=native'
(use the PCIe native services regardless of the BIOS response to the
control request), or 'pcie_ports=compat' (do not use the PCIe native
services at all).
Accordingly, rework the existing PCIe port service drivers so that
they don't request control of the services directly.
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Use pci_pcie_cap() instead of pci_find_capability() to get PCIe capability
offset in pciehp driver. This avoids unnecessary search in PCI
configuration space. This patch also removes 'cap_base' field in
struct controller, that was used to hold PCIe capability offset by
pciehp itself.
Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Since slot_cap field in struct controller contains physical slot
number informationq, we don't need number field in struct slot.
Acked-by: Alex Chiang <achiang@hp.com>
Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
The struct hpc_ops seems a set of hooks to controller specific
routines. But, it is meaningless because no hotplug controller driver
follows this framework.
Acked-by: Alex Chiang <achiang@hp.com>
Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Since we have a pointer to pcie_device in struct controller, we don't
need a pointer to pci_dev.
Acked-by: Alex Chiang <achiang@hp.com>
Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
The crit_sect mutex defined in struct controller is to serialize
hot-plug operations against multiple slots under the same bus. But,
since PCIe doesnstream port has only one slot at most, it is
meaningless and we don't need it.
Acked-by: Alex Chiang <achiang@hp.com>
Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
The slot number can be calculated only by physical slot number field
in the slot capabilities register. So the first_slot field in struct
controller is meaningless and we don't need it.
Acked-by: Alex Chiang <achiang@hp.com>
Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Since the device number of the hot-slot under the PCIe downstream port
is always 0, the slot_device_offset field in the slot is meaningless
and we don't need it.
Acked-by: Alex Chiang <achiang@hp.com>
Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
The hp_slot field is to identify the slot under the same
controller. But, since PCIe downstream port has only one slot at most,
it is meaningless and we don't need it.
Acked-by: Alex Chiang <achiang@hp.com>
Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
The device field in the struct slot is not necessary because it is
always 0 in pciehp driver.
Acked-by: Alex Chiang <achiang@hp.com>
Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
The bus field in struct slot is not necessary.
Acked-by: Alex Chiang <achiang@hp.com>
Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
The slot_num_inc field in struct controller is unused and meaningless
in pciehp driver.
Acked-by: Alex Chiang <achiang@hp.com>
Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Since PCIe downstream port has only one slot at most, we don't need
num_slots field in struct controller. Note that struct controller
itself doesn't exist if PCIe downstream port has no slot.
Acked-by: Alex Chiang <achiang@hp.com>
Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Since PCIe downstream port has only one slot at most, we don't need
'slot_list' linked list to manage multiple slots under the port.
Acked-by: Alex Chiang <achiang@hp.com>
Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Use the generic pci_configure_slot() rather than the PCIe-specific
program_fw_provided_values().
Unlike the previous pciehp-specific code, we now walk through subordinate
devices even if there are no settings for the parent. This should be
harmless because we won't change anything unless we discover firmware
settings farther down.
Signed-off-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
Reviewed-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Acked-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
This patch makes acpi_get_hp_params_from_firmware() take a
pci_dev rather than a pci_bus and makes it return a standard
int errno rather than acpi_status.
Signed-off-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
Reviewed-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Acked-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
The EMI support in pciehp is obviously broken. It is implemented using
struct hotplug_slot_attribute, but sysfs_ops for pci_slot_ktype is NOT
for struct hotplug_slot_attribute, but for struct pci_slot_attribute.
This bug had been there for a long time, maybe it was introduced when
PCI slot framework was introduced. The reason why this bug didn't
cause any problem is maybe the EMI support is not tested at all
because of lack of test environment.
As described above, the EMI support in pciehp seems not to be tested
at all. So this patch removes EMI support from pciehp, instead of
fixing the bug.
Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
The cmd_busy field in struct controller takes only two values 0 or
1. So it should be one bit.
Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Current pciehp disables software notification of adapter presence
changed event and MRL changed event when slot is turned off. Because
of this, there is no way to detect those events on empty slots in the
current pciehp implementation.
According to the past discussion(*), this behavior was introduced to
prevent endless loop that could happen if pcie_isr() runs after power
fault is detected on a certain platform whose stickey power-fault bit
remains on till the slot is powered on again.
(*) http://sourceforge.net/mailarchive/message.php?msg_id=20051130135409.A14918%40unix-os.sc.intel.com
I think this endless loop can be avoided using one bit flag that
indicates power fault had been detected, instead of disabling software
notification of adapter present changed event and MRL changed event.
With this patch, we can enable software notification mechanism of
presence changed and MRL changed event on the empty slots again.
Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
According to kerneljanitors todo list all printk calls (beginning
a new line) should have an according KERN_* constant.
Those are the missing pieces here for the pci subsystem.
Signed-off-by: Frank Seidel <frank@f-seidel.de>
Reviewed-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Tested-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Move the enabling of interrupts after all of the data structures
are setup so that we can safely run the interrupt handler as
soon as it is registered.
Reviewed-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Tested-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Eric W. Biederman <ebiederm@aristanetworks.com>
Signed-off-by: Jesse Barnes <jbarnes@hobbes.lan>
There is a problem that some non hot-pluggable PCIe slots are detected
as hot-pluggable by pciehp on some platforms. The immediate cause of
this problem is that hot-plug capable bit in the Slot Capabilities
register is set even for non hot-pluggable slots on those platforms.
It seems a BIOS/hardware problem, but we need workaround about that.
Some of those platforms define hot-pluggable PCIe slots on ACPI
namespace properly, while hot-plug capable bit in the Slot
Capabilities register is set improperly. So using ACPI namespace
information in pciehp to detect PCIe hot-pluggable slots would be a
workaround.
This patch adds 'pciehp_detect_mode' module option. When 'acpi' is
specified, pciehp uses ACPI namespace information to detect PCIe
hot-pluggable slots.
Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
External driver files should not include any private acpica headers.
Signed-off-by: Lin Ming <ming.m.lin@intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
This patch refines messages in pciehp module. The main changes are as
follows:
- remove the trailing "."
- remove __func__ as much as possible
- capitalize the first letter of messages
- show PCI device address including its domain
Signed-off-by: Taku Izumi <izumi.taku@jp.fujitsu.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
This patch adds polling mechanism for Data Link Layer Link Active bit
after turning power on, instead of waiting for 1000 msec. This reduces
reduce the unnecessary long wait.
Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
We do not need to manage our own name parameter, especially since
the PCI core can change it on our behalf, in the case of duplicate
slot names.
Remove 'name' from pciehp's version of struct slot, and remove
unused 'task_list' as well.
Cc: kristen.c.accardi@intel.com
Acked-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Alex Chiang <achiang@hp.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Current pciehp driver gets irq number from pci_dev->irq. But because
pciehp driver is a pci express port service driver, it should get irq
number from pcie_device->irq.
Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Commit 3800345f72 (pciehp: fix slot name)
introduces the pciehp_slot_with_bus module parameter, which was intended
to help work around broken firmware that assigns the same name to multiple
slots.
Commit 9e4f2e8d4d (pciehp: add message about
pciehp_slot_with_bus option) tells the user to use the above parameter
in the event of a name collision.
This approach is sub-optimal because it requires too much work from
the user.
Instead, let's rename the slot on behalf of the user. If firmware
assigns the name N to multiple slots, then:
The first registered slot is assigned N
The second registered slot is assigned N-1
The third registered slot is assigned N-2
The Mth registered slot becomes N-M
In the event we overflow the slot->name parameter, we report an
error to the user.
This is a temporary fix until the entire PCI core can be reworked
such that individual drivers no longer have to manage their own
slot names.
Tested-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Acked-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Alex Chiang <achiang@hp.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Proper memory barriers have been added to order accesses
to ->cmd_busy, so volatile declaration for cmd_busy can
be removed.
Signed-off-by: Ming Lei <tom.leiming@gmail.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Current pciehp driver's intialization sequence is as follows:
(1) initialize controller data structure
(2) install interrupt handler
(3) enable software notification
(4) initialize controller specific slot data structure
(5) initialize generic slot data structure and register it to pci hotplug core
The interrupt handler of pciehp assumes that controller specific slot
data structure is already initialized. However, it is installed at (2)
before initializing controller specific slot data structure at
(4). Because of this, pciehp driver cannot handle the following cases
properly.
- If devices that shares IRQ with pciehp raise interrupts between (2) and (4).
- If hotplug events (e.g. MRL open) happen between (3) and (4).
We already have a workaround for this problem ("pciehp: fix NULL
dereference in interrupt handler: dbd79aed1a).
But we still need fundamental fix.
This patch fix the problem by changing the initilization sequence as follows:
(1) initialize controller data structure
(2) initialize controller specific slot data structure
(3) install interrupt handler
(4) enable software notification
(5) initialize generic slot data structure and register it to pci hotplug core
Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Acked-by: Alex Chiang <achiang@hp.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Fix the following problems of shpchp driver about getting hotplug
control from firmware.
- The shpchp driver must not control the hotplug controller if it
fails to get control from the firmware. But current shpchp
controls the hotplug controller regardless the result, because it
doesn't check the return value of get_hp_hw_control_from_firmware().
- Current shpchp driver doesn't support _OSC.
The pciehp driver already have the code for evaluating _OSC and OSHP
and shpchp and pciehp can share it. So this patch move that code from
pciehp to acpi_pcihp.c.
Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Current pciehp evaluates _OSC/OSHP method after some controller
initialization is done. So if evaluating _OSC/OSHP is failed, we need
to cleanup already initialized data structures or hardware. This
clearly is not robust way. With this patch, _OSC/OSHP evaluation is
done first.
Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Fix the "pciehp probing slow" problem reported from Jan C. Nordholz in
http://bugzilla.kernel.org/show_bug.cgi?id=10751.
The command completed bit in Slot Status register applies only to
commands issued to control the attention indicator, power indicator,
power controller, or electromechanical interlock. However, writes to
other parts of the Slot Control register would end up writing to the
control fields. Hence, any write to Slot Control register is
considered as a command. However, if the controller doesn't support
any of attention indicator, power indicator, power controller and
electromechanical interlock, command completed bit would not set in
writing to Slot Control register. In this case, we should not wait for
command completed bit set, otherwise all commands would be considered
not completed in timeout seconds (1 sec.).
The cause of the problem is pciehp driver didn't take this situation
into account. This patch changes pciehp to take it into account. This
patch also add the check for "No Command Completed Support" bit in
Slot Capability register. If it is set, we should not wait for command
completed bit set as well.
This problem seems to be revealed by the commit
c27fb883df that fixed the bug that
pciehp did not wait for command completed properly (pciehp just
ignored the command completion event).
Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Current pciehp saves only 8bits of Slot Capability registers in
ctrl->ctrlcap. But it refers more than 8bit for checking EMI capability.
It is clearly a bug and EMI would never work. To fix this problem,
this patch saves full Slot Capability contens in ctrl->slot_cap. It also
reduce the redundant reads of Slot Capability register. And this pach
also cleans up the macros to check the slot capabilitys (e.g. MRL_SENS(),
and so on).
Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Current pciehp implementation disables and re-enables hotplug interrupts
in its interrupt handler. This operation might be intend to guarantee
that interrupts for the events newly occured during previous events are
being handled will be successfully generated. But current implementaion
has the following prolems.
- Current interrupt service routin clears status changes without
waiting command completion. Because of this, events might not be
cleared properly.
- Current interrupt service routine clears status changes caused by
disabling or enabling hotplug interrupts itself. This will lose new
events that occurs during previous interrupts are being handled.
- Current implementation doesn't have any serialization mechanism
between the code to wait for command completion and the interrupt
handler that clears the command completion events caused by itself.
There is clearly race conditions between them, and it may cause
the problem that waiting for command completion doesn't work for
example.
To fix those problems, this patch stops disabling/re-enabling hotplug
interrupts in interrupt service routine. Instead of this, this patch
re-inspects Slot Status register after clearing what is presumed to
be the last bending interrupt in order to guarantee that all interrupt
events are serviced.
Signed-off-by: Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
Signed-off-by: Kristen Carlson Accardi <kristen.c.accardi@intel.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
__FUNCTION__ is gcc-specific, use __func__
Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Alex Williamson
Bjorn Helgaas
Yijing Wang
Yinghai Lu
Rusty Russell
Kenji Kaneshige
Tejun Heo
Rafael J. Wysocki
Bjorn Helgaas
Frank Seidel
Eric W. Biederman
Len Brown
Lin Ming
Taku Izumi
Alex Chiang
Ming Lei
Harvey Harrison