This improves AEN handler for Host Network Controller Driver Status
Change (HNCDSC):
* The channel's lock should be hold when accessing its state.
* Do failover when host driver isn't ready.
* Configure channel when host driver becomes ready.
Signed-off-by: Gavin Shan <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
The issue was found on BCM5718 which has two NCSI channels in one
package: C0 and C1. C0 is in link-up state while C1 is in link-down
state. C0 is chosen as active channel until unplugging and plugging
C0's cable: On unplugging C0's cable, LSC (Link State Change) AEN
packet received on C0 to report link-down event. After that, C1 is
chosen as active channel. LSC AEN for link-up event is lost on C0
when plugging C0's cable back. We lose the network even C0 is usable.
This resolves the issue by recording the (hot) channel that was ever
chosen as active one. The hot channel is chosen to be active one
if none of available channels in link-up state. With this, C0 is still
the active one after unplugging C0's cable. LSC AEN packet received
on C0 when plugging its cable back.
Signed-off-by: Gavin Shan <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
The issue was found on BCM5718 which has two NCSI channels in one
package: C0 and C1. Both of them are connected to different LANs,
means they are in link-up state and C0 is chosen as the active one
until resetting BCM5718 happens as below.
Resetting BCM5718 results in LSC (Link State Change) AEN packet
received on C0, meaning LSC AEN is missed on C1. When LSC AEN packet
received on C0 to report link-down, it fails over to C1 because C1
is in link-up state as software can see. However, C1 is in link-down
state in hardware. It means the link state is out of synchronization
between hardware and software, resulting in inappropriate channel (C1)
selected as active one.
This resolves the issue by sending separate GLS (Get Link Status)
commands to all channels in the package before trying to do failover.
The last link states of all channels in the package are retrieved.
With it, C0 (not C1) is selected as active one as expected.
Signed-off-by: Gavin Shan <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
There are several if/else statements in the state machine implemented
by switch/case in ncsi_suspend_channel() to avoid duplicated code. It
makes the code a bit hard to be understood.
This drops if/else statements in ncsi_suspend_channel() to improve the
code readability as Joel Stanley suggested. Also, it becomes easy to
add more states in the state machine without affecting current code.
No logical changes introduced by this.
Suggested-by: Joel Stanley <[email protected]>
Signed-off-by: Gavin Shan <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
This stops NCSI device when closing the network device so that the
NCSI device can be reenabled later.
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Joel Stanley <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
This introduces ncsi_stop_dev(), as counterpart to ncsi_start_dev(),
to stop the NCSI device so that it can be reenabled in future. This
API should be called when the network device driver is going to
shutdown the device. There are 3 things done in the function: Stop
the channel monitoring; Reset channels to inactive state; Report
NCSI link down.
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Joel Stanley <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
The original NCSI channel monitoring was implemented based on a
backoff algorithm: the GLS response should be received in the
specified interval. Otherwise, the channel is regarded as dead
and failover should be taken if current channel is an active one.
There are several problems in the implementation: (A) On BCM5718,
we found when the IID (Instance ID) in the GLS command packet
changes from 255 to 1, the response corresponding to IID#1 never
comes in. It means we cannot make the unfair judgement that the
channel is dead when one response is missed. (B) The code's
readability should be improved. (C) We should do failover when
current channel is active one and the channel monitoring should
be marked as disabled before doing failover.
This reworks the channel monitoring to address all above issues.
The fields for channel monitoring is put into separate struct
and the state of channel monitoring is predefined. The channel
is regarded alive if the network controller responses to one of
two GLS commands or both of them in 5 seconds.
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Joel Stanley <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
There is only one NCSI request property for now: the response for
the sent command need drive the workqueue or not. So we had one
field (@driven) for the purpose. We lost the flexibility to extend
NCSI request properties.
This replaces @driven with @flags and @req_flags in NCSI request
and NCSI command argument struct. Each bit of the newly introduced
field can be used for one property. No functional changes introduced.
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Joel Stanley <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
The NCSI request index (struct ncsi_request::id) is put into instance
ID (IID) field while sending NCSI command packet. It was designed the
available IDs are given in round-robin fashion. @ndp->request_id was
introduced to represent the next available ID, but it has been used
as number of successively allocated IDs. It breaks the round-robin
design. Besides, we shouldn't put 0 to NCSI command packet's IID
field, meaning ID#0 should be reserved according section 6.3.1.1
in NCSI spec (v1.1.0).
This fixes above two issues. With it applied, the available IDs will
be assigned in round-robin fashion and ID#0 won't be assigned.
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Joel Stanley <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
We needn't send CIS (Clear Initial State) command to the NCSI
reserved channel (0x1f) in the enumeration. We shouldn't receive
a valid response from CIS on NCSI channel 0x1f.
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Joel Stanley <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
This defines NCSI_RESERVED_CHANNEL as the reserved NCSI channel
ID (0x1f). No logical changes introduced.
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Joel Stanley <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
xchg() is used to set NCSI channel's state in order for consistent
access to the state. xchg()'s return value should be used. Otherwise,
one build warning will be raised (with -Wunused-value) as below message
indicates. It is reported by ia64-linux-gcc (GCC) 4.9.0.
net/ncsi/ncsi-manage.c: In function 'ncsi_channel_monitor':
arch/ia64/include/uapi/asm/cmpxchg.h:56:2: warning: value computed is \
not used [-Wunused-value]
((__typeof__(*(ptr))) __xchg((unsigned long) (x), (ptr), sizeof(*(ptr))))
^
net/ncsi/ncsi-manage.c:202:3: note: in expansion of macro 'xchg'
xchg(&nc->state, NCSI_CHANNEL_INACTIVE);
This removes the atomic access to NCSI channel's state avoid the above
build warning. We have to hold the channel's lock when its state is readed
or updated. No functional changes introduced.
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Joel Stanley <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
This fixes warning reported from sparse:
pci-ioda.c:451:49: warning: incorrect type in argument 2 (different base types)
Fixes: 262af557dd ("powerpc/powernv: Enable M64 aperatus for PHB3")
Signed-off-by: Gavin Shan <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
This fixes the warning reported from sparse:
eeh-powernv.c:875:23: warning: constant 0x8000000000000000 is so big it is unsigned long
Fixes: ebe2253127 ("powerpc/powernv: Support PCI slot ID")
Suggested-by: Michael Ellerman <[email protected]>
Signed-off-by: Gavin Shan <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
The hub diag-data type is filled with big-endian data by OPAL call
opal_pci_get_hub_diag_data(). We need convert it to CPU-endian value
before using it. The issue is reported by sparse as pointed by Michael
Ellerman:
eeh-powernv.c:1309:21: warning: restricted __be16 degrades to integer
This converts hub diag-data type to CPU-endian before using it in
pnv_eeh_get_and_dump_hub_diag().
Fixes: 2a485ad7c8 ("powerpc/powernv: Drop PHB operation next_error()")
Cc: [email protected] # v4.1+
Suggested-by: Michael Ellerman <[email protected]>
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Russell Currey <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
The PE number (@frozen_pe_no), filled by opal_pci_next_error() is in
big-endian format. It should be converted to CPU-endian before it is
passed to opal_pci_eeh_freeze_clear() when clearing the frozen state if
the PE is invalid one. As Michael Ellerman pointed out, the issue is
also detected by sparse:
eeh-powernv.c:1541:41: warning: incorrect type in argument 2 (different base types)
This passes CPU-endian PE number to opal_pci_eeh_freeze_clear() and it
should be part of commit <0f36db77643b> ("powerpc/eeh: Fix wrong printed
PE number"), which was merged to 4.3 kernel.
Fixes: 71b540adff ("powerpc/powernv: Don't escalate non-existing frozen PE")
Cc: [email protected] # v4.3+
Suggested-by: Paul Mackerras <[email protected]>
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Russell Currey <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
This replaces of_get_property() with of_property_read_u32() or
of_property_read_string() so that we needn't consider the endian
issue, the returned value always is in CPU-endian.
Signed-off-by: Gavin Shan <[email protected]>
[mpe: Fold in the change to the "ibm,slot-surprise-pluggable" case]
Signed-off-by: Michael Ellerman <[email protected]>
This supports PCI surprise hotplug. The design is highlighted as
below:
* The PCI slot's surprise hotplug capability is exposed through
device node property "ibm,slot-surprise-pluggable", meaning
PCI surprise hotplug will be disabled if skiboot doesn't support
it yet.
* The interrupt because of presence or link state change is raised
on surprise hotplug event. One event is allocated and queued to
the PCI slot for workqueue to pick it up and process in serialized
fashion. The code flow for surprise hotplug is same to that for
managed hotplug except: the affected PEs are put into frozen state
to avoid unexpected EEH error reporting in surprise hot remove path.
Signed-off-by: Gavin Shan <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
This removes likely() and unlikely() in pnv_php.c as the code isn't
running in hot path. Those macros to affect CPU's branch stream don't
help a lot for performance. I used them to identify the cases are
likely or unlikely to happen. No logical changes introduced.
Signed-off-by: Gavin Shan <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
This unfreezes PE when it's initialized because the PE might be put
into frozen state in the last hot remove path. It's not harmful to
do so if the PE is already in unfrozen state.
Signed-off-by: Gavin Shan <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
This exports eeh_pe_state_mark(). It will be used to mark the surprise
hot removed PE as isolated to avoid unexpected EEH error reporting in
surprise remove path.
Signed-off-by: Gavin Shan <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
This exports @confirm_error_lock so that eeh_serialize_{lock, unlock}()
can be used to freeze the affected PE in PCI surprise hot remove path.
Signed-off-by: Gavin Shan <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
Function eeh_pe_set_option() is used to apply the requested options
(enable, disable, unfreeze) in EEH virtualization path. The semantics
of this function isn't complete until freezing is supported.
This allows to freeze the indicated PE. The new semantics is going to
be used in PCI surprise hot remove path, to freeze removed PCI devices
(PE) to avoid unexpected EEH error reporting.
Signed-off-by: Gavin Shan <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
When issuing PHB reset, OPAL API opal_pci_poll() is called to drive
the state machine in OPAL forward. However, we needn't always call
the function under some circumstances like reset deassert.
This avoids calling opal_pci_poll() when OPAL_SUCCESS is returned
from opal_pci_reset(). Except the overhead introduced by additional
one unnecessary OPAL call, I didn't run into real issue because of
this.
Reported-by: Pridhiviraj Paidipeddi <[email protected]>
Signed-off-by: Gavin Shan <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
There is stale interrupt (PHYSTS_CHG in ISR, bit#6 in 0x0) from
the bootloader (uboot) when enabling the MAC. The stale interrupts
aren't part of kernel and should be cleared.
This clears the stale interrupts in ISR (0x0) when enabling the MAC.
Signed-off-by: Gavin Shan <[email protected]>
Signed-off-by: Joel Stanley <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
The PCI hotplug can be part of EEH error recovery. The @pdn and
the device's PE number aren't removed and added afterwords. The
PE number in @pdn should be set to an invalid one. Otherwise, the
PE's device count is decreased on removing devices while failing
to be increased on adding devices. It leads to unbalanced PE's
device count and make normal PCI hotplug path broken.
Fixes: c5f7700bbd ("powerpc/powernv: Dynamically release PE")
Signed-off-by: Gavin Shan <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
The PE for root bus (root PE) can be removed because of PCI hot
remove in EEH recovery path for fenced PHB error. We need update
@phb->root_pe_populated accordingly so that the root PE can be
populated again in forthcoming PCI hot add path. Also, the PE
shouldn't be destroyed as it's global and reserved resource.
Fixes: c5f7700bbd ("powerpc/powernv: Dynamically release PE")
Reported-by: Frederic Barrat <[email protected]>
Signed-off-by: Gavin Shan <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
In pnv_ioda_free_pe(), the PE object (including the associated PE
number) is cleared before resetting the corresponding bit in the
PE allocation bitmap. It means PE#0 is always released to the bitmap
wrongly.
This fixes above issue by caching the PE number before the PE object
is cleared.
Fixes: 1e9167726c ("powerpc/powernv: Use PE instead of number during setup and release"
Cc: [email protected] # v4.7+
Signed-off-by: Gavin Shan <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
The compound PE is created to accommodate the devices attached to
one specific PCI bus that consume multiple M64 segments. The compound
PE is made up of one master PE and possibly multiple slave PEs. The
slave PEs should be destroyed when releasing the master PE. A kernel
crash happens when derferencing @pe->pdev on releasing the slave PE
in pnv_ioda_deconfigure_pe().
# echo 0 > /sys/bus/pci/slots/C7/power
iommu: Removing device 0000:01:00.1 from group 0
iommu: Removing device 0000:01:00.0 from group 0
Unable to handle kernel paging request for data at address 0x00000010
Faulting instruction address: 0xc00000000005d898
cpu 0x1: Vector: 300 (Data Access) at [c000000fe8217620]
pc: c00000000005d898: pnv_ioda_release_pe+0x288/0x610
lr: c00000000005dbdc: pnv_ioda_release_pe+0x5cc/0x610
sp: c000000fe82178a0
msr: 9000000000009033
dar: 10
dsisr: 40000000
current = 0xc000000fe815ab80
paca = 0xc00000000ff00400 softe: 0 irq_happened: 0x01
pid = 2709, comm = sh
Linux version 4.8.0-rc5-gavin-00006-g745efdb (gwshan@gwshan) \
(gcc version 4.9.3 (Buildroot 2016.02-rc2-00093-g5ea3bce) ) #586 SMP \
Tue Sep 6 13:37:29 AEST 2016
enter ? for help
[c000000fe8217940] c00000000005d684 pnv_ioda_release_pe+0x74/0x610
[c000000fe82179e0] c000000000034460 pcibios_release_device+0x50/0x70
[c000000fe8217a10] c0000000004aba80 pci_release_dev+0x50/0xa0
[c000000fe8217a40] c000000000704898 device_release+0x58/0xf0
[c000000fe8217ac0] c000000000470510 kobject_release+0x80/0xf0
[c000000fe8217b00] c000000000704dd4 put_device+0x24/0x40
[c000000fe8217b20] c0000000004af94c pci_remove_bus_device+0x12c/0x150
[c000000fe8217b60] c000000000034244 pci_hp_remove_devices+0x94/0xd0
[c000000fe8217ba0] c0000000004ca444 pnv_php_disable_slot+0x64/0xb0
[c000000fe8217bd0] c0000000004c88c0 power_write_file+0xa0/0x190
[c000000fe8217c50] c0000000004c248c pci_slot_attr_store+0x3c/0x60
[c000000fe8217c70] c0000000002d6494 sysfs_kf_write+0x94/0xc0
[c000000fe8217cb0] c0000000002d50f0 kernfs_fop_write+0x180/0x260
[c000000fe8217d00] c0000000002334a0 __vfs_write+0x40/0x190
[c000000fe8217d90] c000000000234738 vfs_write+0xc8/0x240
[c000000fe8217de0] c000000000236250 SyS_write+0x60/0x110
[c000000fe8217e30] c000000000009524 system_call+0x38/0x108
It fixes the kernel crash by bypassing releasing resources (DMA,
IO and memory segments, PELTM) because there are no resources assigned
to the slave PE.
Fixes: c5f7700bbd ("powerpc/powernv: Dynamically release PE")
Reported-by: Frederic Barrat <[email protected]>
Signed-off-by: Gavin Shan <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
@mynodes is set to NULL when __unflatten_device_tree() is called
to unflatten device sub-tree in PCI hot add scenario on PowerPC
PowerNV platform. Marking @mynodes detached unconditionally causes
kernel crash as below backtrace shows:
Unable to handle kernel paging request for data at address 0x00000000
Faulting instruction address: 0xc000000000b26f64
cpu 0x0: Vector: 300 (Data Access) at [c000003fcc7cf740]
pc: c000000000b26f64: __unflatten_device_tree+0xf4/0x190
lr: c000000000b26f40: __unflatten_device_tree+0xd0/0x190
sp: c000003fcc7cf9c0
msr: 900000000280b033
dar: 0
dsisr: 40000000
current = 0xc000003fcc281680
paca = 0xc00000000ff00000 softe: 0 irq_happened: 0x01
pid = 2724, comm = sh
Linux version 4.7.0-gavin-07754-g92a6836 (gwshan@gwshan) (gcc version \
4.9.3 (Buildroot 2016.02-rc2-00093-g5ea3bce) ) #539 SMP Mon Aug 1 \
12:40:29 AEST 2016
enter ? for help
[c000003fcc7cfa50] c000000000b27060 of_fdt_unflatten_tree+0x60/0x90
[c000003fcc7cfaa0] c0000000004c6288 pnv_php_set_slot_power_state+0x118/0x440
[c000003fcc7cfb80] c0000000004c6a10 pnv_php_enable+0xc0/0x170
[c000003fcc7cfbd0] c0000000004c4d80 power_write_file+0xa0/0x190
[c000003fcc7cfc50] c0000000004be93c pci_slot_attr_store+0x3c/0x60
[c000003fcc7cfc70] c0000000002d3fd4 sysfs_kf_write+0x94/0xc0
[c000003fcc7cfcb0] c0000000002d2c30 kernfs_fop_write+0x180/0x260
[c000003fcc7cfd00] c000000000230fe0 __vfs_write+0x40/0x190
[c000003fcc7cfd90] c000000000232278 vfs_write+0xc8/0x240
[c000003fcc7cfde0] c000000000233d90 SyS_write+0x60/0x110
[c000003fcc7cfe30] c000000000009524 system_call+0x38/0x108
This avoids the kernel crash by marking @mynodes detached only when
@mynodes is dereferencing valid device node in __unflatten_device_tree().
Fixes: 1d1bde550e ("of: fdt: mark unflattened tree as detached")
Reported-by: Meng Li <[email protected]>
Signed-off-by: Gavin Shan <[email protected]>
Signed-off-by: Rob Herring <[email protected]>
The initial MAC address is retrieved from hardware if it's not
provided by device-tree. The reserved MAC address from hardware
will be used if non-reserved MAC address is invalid. It will
cause mismatched MAC address seen by hardware and software.
This disallows using the reserved hardware MAC address to avoid
the mismatched MAC address seen by hardware and software.
Fixes: 113ce107af ("net/faraday: Read MAC address from chip")
Suggested-by: David Laight <[email protected]>
Suggested-by: Benjamin Herrenschmidt <[email protected]>
Signed-off-by: Gavin Shan <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Bogus PHY interrupts are observed. This masks the PHY interrupt
when the interface works in NCSI mode as there is no attached
PHY under the circumstance.
Signed-off-by: Gavin Shan <[email protected]>
Acked-by: Joel Stanley <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
This matches the driver with devices compatible with "faraday,ftgmac100"
declared in the device tree. Originally, device's name from device
tree for it.
Signed-off-by: Gavin Shan <[email protected]>
Acked-by: Joel Stanley <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
This makes ftgmac100 driver support NCSI mode. The NCSI is enabled
on the interface if property "use-nc-si" or "use-ncsi" is found from
the device node in device tree.
* No PHY device is used when NCSI mode is enabled.
* The NCSI device (struct ncsi_dev) is created when probing the
device while it's enabled/started when the interface is brought
up.
* Hardware IP checksum dosn't work when NCSI mode is enabled. It
is disabled on enabled NCSI.
Signed-off-by: Gavin Shan <[email protected]>
Acked-by: Joel Stanley <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
The device is assigned with random MAC address. It isn't reasonable.
An valid MAC address might have been provided by (uboot) firmware by
device-tree or in chip. It's reasonable to use it to maintain consistency.
This uses the MAC address from device-tree or that in the chip if it's
valid. Otherwise, a random MAC address is given as before.
Signed-off-by: Gavin Shan <[email protected]>
Acked-by: Joel Stanley <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
This introduces two helper functions to create or destroy MDIO
interface. No logical changes introduced except the proper MDIO
names are given when having more than one MDIO bus.
Signed-off-by: Gavin Shan <[email protected]>
Acked-by: Joel Stanley <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
This introduces NCSI AEN packet handlers that result in (A) the
currently active channel is reconfigured; (B) Currently active
channel is deconfigured and disabled, another channel is chosen
as active one and configured. Case (B) won't happen if hardware
arbitration has been enabled, the channel that was in active
state is suspended simply.
Signed-off-by: Gavin Shan <[email protected]>
Acked-by: Joel Stanley <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
This manages NCSI packages and channels:
* The available packages and channels are enumerated in the first
time of calling ncsi_start_dev(). The channels' capabilities are
probed in the meanwhile. The NCSI network topology won't change
until the NCSI device is destroyed.
* There in a queue in every NCSI device. The element in the queue,
channel, is waiting for configuration (bringup) or suspending
(teardown). The channel's state (inactive/active) indicates the
futher action (configuration or suspending) will be applied on the
channel. Another channel's state (invisible) means the requested
action is being applied.
* The hardware arbitration will be enabled if all available packages
and channels support it. All available channels try to provide
service when hardware arbitration is enabled. Otherwise, one channel
is selected as the active one at once.
* When channel is in active state, meaning it's providing service, a
timer started to retrieve the channe's link status. If the channel's
link status fails to be updated in the determined period, the channel
is going to be reconfigured. It's the error handling implementation
as defined in NCSI spec.
Signed-off-by: Gavin Shan <[email protected]>
Acked-by: Joel Stanley <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
The NCSI response packets are sent to MC (Management Controller)
from the remote end. They are responses of NCSI command packets
for multiple purposes: completion status of NCSI command packets,
return NCSI channel's capability or configuration etc.
This defines struct to represent NCSI response packets and introduces
function ncsi_rcv_rsp() which will be used to receive NCSI response
packets and parse them.
Signed-off-by: Gavin Shan <[email protected]>
Acked-by: Joel Stanley <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
The NCSI command packets are sent from MC (Management Controller)
to remote end. They are used for multiple purposes: probe existing
NCSI package/channel, retrieve NCSI channel's capability, configure
NCSI channel etc.
This defines struct to represent NCSI command packets and introduces
function ncsi_xmit_cmd(), which will be used to transmit NCSI command
packet according to the request. The request is represented by struct
ncsi_cmd_arg.
Signed-off-by: Gavin Shan <[email protected]>
Acked-by: Joel Stanley <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
NCSI spec (DSP0222) defines several objects: package, channel, mode,
filter, version and statistics etc. This introduces the data structs
to represent those objects and implement functions to manage them.
Also, this introduces CONFIG_NET_NCSI for the newly implemented NCSI
stack.
* The user (e.g. netdev driver) dereference NCSI device by
"struct ncsi_dev", which is embedded to "struct ncsi_dev_priv".
The later one is used by NCSI stack internally.
* Every NCSI device can have multiple packages simultaneously, up
to 8 packages. It's represented by "struct ncsi_package" and
identified by 3-bits ID.
* Every NCSI package can have multiple channels, up to 32. It's
represented by "struct ncsi_channel" and identified by 5-bits ID.
* Every NCSI channel has version, statistics, various modes and
filters. They are represented by "struct ncsi_channel_version",
"struct ncsi_channel_stats", "struct ncsi_channel_mode" and
"struct ncsi_channel_filter" separately.
* Apart from AEN (Asynchronous Event Notification), the NCSI stack
works in terms of command and response. This introduces "struct
ncsi_req" to represent a complete NCSI transaction made of NCSI
request and response.
link: https://www.dmtf.org/sites/default/files/standards/documents/DSP0222_1.1.0.pdf
Signed-off-by: Gavin Shan <[email protected]>
Acked-by: Joel Stanley <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
When calling eeh_rmv_device() in eeh_reset_device() for partial hotplug
case, @rmv_data instead of its address is the proper argument.
Otherwise, the stack frame is corrupted when writing to
@rmv_data (actually its address) in eeh_rmv_device(). It results in
kernel crash as observed.
This fixes the issue by passing @rmv_data, not its address to
eeh_rmv_device() in eeh_reset_device().
Fixes: 67086e32b5 ("powerpc/eeh: powerpc/eeh: Support error recovery for VF PE")
Reported-by: Pridhiviraj Paidipeddi <[email protected]>
Signed-off-by: Gavin Shan <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
This adds standalone driver to support PCI hotplug for PowerPC PowerNV
platform that runs on top of skiboot firmware. The firmware identifies
hotpluggable slots and marked their device tree node with proper
"ibm,slot-pluggable" and "ibm,reset-by-firmware". The driver scans
device tree nodes to create/register PCI hotplug slot accordingly.
The PCI slots are organized in fashion of tree, which means one
PCI slot might have parent PCI slot and parent PCI slot possibly
contains multiple child PCI slots. At the plugging time, the parent
PCI slot is populated before its children. The child PCI slots are
removed before their parent PCI slot can be removed from the system.
If the skiboot firmware doesn't support slot status retrieval, the PCI
slot device node shouldn't have property "ibm,reset-by-firmware". In
that case, none of valid PCI slots will be detected from device tree.
The skiboot firmware doesn't export the capability to access attention
LEDs yet and it's something for TBD.
Signed-off-by: Gavin Shan <[email protected]>
Acked-by: Bjorn Helgaas <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
This exports 4 functions, which base on the corresponding OPAL
APIs to get/set PCI slot status. Those functions are going to
be used by PowerNV PCI hotplug driver:
pnv_pci_get_device_tree() opal_get_device_tree()
pnv_pci_get_presence_state() opal_pci_get_presence_state()
pnv_pci_get_power_state() opal_pci_get_power_state()
pnv_pci_set_power_state() opal_pci_set_power_state()
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Alexey Kardashevskiy <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
This introduces pnv_pci_get_slot_id() to get the hotpluggable PCI
slot ID from the corresponding device node. It will be used by
hotplug driver.
Requested-by: Andrew Donnellan <[email protected]>
Signed-off-by: Gavin Shan <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
The (OPAL) firmware might provide the PCI slot reset capability
which is identified by property "ibm,reset-by-firmware" on the
PCI slot associated device node.
This routes the reset request to firmware if "ibm,reset-by-firmware"
exists in the PCI slot device node. Otherwise, the reset is done
inside kernel as before.
Signed-off-by: Gavin Shan <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
The reset and poll functionality from (OPAL) firmware supports
PHB and PCI slot at same time. They are identified by ID. This
supports PCI slot ID by:
* Rename the argument name for opal_pci_reset() and opal_pci_poll()
accordingly
* Rename pnv_eeh_phb_poll() to pnv_eeh_poll() and adjust its argument
name.
* One macro is added to produce PCI slot ID.
Signed-off-by: Gavin Shan <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
The pdn (struct pci_dn) instances are allocated from memblock or
bootmem when creating PCI controller (hoses) in setup_arch(). PCI
hotplug, which will be supported by proceeding patches, releases
PCI device nodes and their corresponding pdn on unplugging event.
The memory chunks for pdn instances allocated from memblock or
bootmem are hard to reused after being released.
This delays creating pdn by pci_devs_phb_init() from setup_arch()
to core_initcall() so that they are allocated from slab. The memory
consumed by pdn can be released to system without problem during
PCI unplugging time. It indicates that pci_dn is unavailable in
setup_arch() and the the fixup on pdn (like AGP's) can't be carried
out that time. We have to do that in pcibios_root_bridge_prepare()
on maple/pasemi/powermac platforms where/when the pdn is available.
pcibios_root_bridge_prepare is called from subsys_initcall() which
is executed after core_initcall() so the code flow does not change.
At the mean while, the EEH device is created when pdn is populated,
meaning pdn and EEH device have same life cycle. In turn, we needn't
call eeh_dev_init() to create EEH device explicitly.
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Alexey Kardashevskiy <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
On the PCI plugging event, PCI slot's subordinate devices are
scanned and their (IO and MMIO) resources are assigned. Platform
dependent resources (PE#, IO/MMIO/DMA windows) are allocated or
created on updating windows of the slot's upstream bridge.
This updates the windows of the hot plugged slot's upstream bridge
in pcibios_finish_adding_to_bus() so that the platform resources
(PE#, IO/MMIO/DMA segments) are allocated or created accordingly.
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Alexey Kardashevskiy <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
This supports releasing PEs dynamically. A reference count is
introduced to PE representing number of PCI devices associated
with the PE. The reference count is increased when PCI device
joins the PE and decreased when PCI device leaves the PE in
pnv_pci_release_device(). When the count becomes zero, the PE
and its consumed resources are released. Note that the count
is accessed concurrently. So a counter with "int" type is enough
here.
In order to release the sources consumed by the PE, couple of
helper functions are introduced as below:
* pnv_pci_ioda1_unset_window() - Unset IODA1 DMA32 window
* pnv_pci_ioda1_release_dma_pe() - Release IODA1 DMA32 segments
* pnv_pci_ioda2_release_dma_pe() - Release IODA2 DMA resource
* pnv_ioda_release_pe_seg() - Unmap IO/M32/M64 segments
Signed-off-by: Gavin Shan <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
pnv_ioda_deconfigure_pe() is visible only when CONFIG_PCI_IOV is
enabled. The function will be used to tear down PE's associated
mapping in PCI hotplug path that doesn't depend on CONFIG_PCI_IOV.
This makes pnv_ioda_deconfigure_pe() visible and not depend on
CONFIG_PCI_IOV.
Signed-off-by: Gavin Shan <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
The PCI slots are associated with root port or downstream ports
of the PCIe switch connected to root port. When adapter is hot
added to the PCI slot, it usually requests more IO or memory
resource from the directly connected parent bridge (port) and
update the bridge's windows accordingly. The resource windows
of upstream bridges can't be updated automatically. It possibly
leads to unbalanced resource across the bridges: The window of
downstream bridge is overruning that of upstream bridge. The
IO or MMIO path won't work.
This resolves the above issue by extending bridge windows of
root port and upstream port of the PCIe switch connected to
the root port to PHB's windows.
The windows of root port and bridge behind that are extended to
the PHB's windows to accomodate the PCI hotplug happening in
future. The PHB's 64KB 32-bits MSI region is included in bridge's
M32 windows (in hardware) though it's excluded in the corresponding
resource, as the bridge's M32 windows have 1MB as their minimal
alignment. We observed EEH error during system boot when the MSI
region is included in bridge's M32 window.
This excludes top 1MB (including 64KB 32-bits MSI region) region
from bridge's M32 windows when extending them.
Signed-off-by: Gavin Shan <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
There is no parent bridge for root bus, meaning pcibios_setup_bridge()
isn't invoked for root bus. The PE for root bus is the ancestor of
other PEs in PELTV. It means we need PE for root bus populated before
all others.
This populates the PE for root bus in pcibios_setup_bridge() path
if it's not populated yet. The PE number next to the reserved one
is used as the PE# to avoid holes in continuous M64 space.
Signed-off-by: Gavin Shan <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
Currently, the PEs and their associated resources are assigned in
ppc_md.pcibios_fixup() except those used by SRIOV VFs. The function
is called for once after PCI probing and resources assignment is
completed. So it's obviously not hotplug friendly.
This creates PEs dynamically in pcibios_setup_bridge() that is
called for the event during system bootup and PCI hotplug: updating
PCI bridge's windows after resource assignment/reassignment are done.
In partial hotplug case, not all PCI devices included to one particular
PE are unplugged and plugged again, we just need unbinding/binding the
hot added PCI devices with the corresponding PE without creating new
one. The change is applied to IODA1 and IODA2 PHBs only. The behaviour
on NPU PHBs aren't changed. There are no PCI bridges on NPU PHBs,
meaning pcibios_setup_bridge() won't be invoked there. We have to use
old path (pnv_pci_ioda_fixup()) to setup PEs on NPU PHBs.
Signed-off-by: Gavin Shan <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
PE number for one particular PE can be allocated dynamically or
reserved according to the consumed M64 (64-bits prefetchable)
segments of the PE. The M64 segment can't be remapped to arbitrary
PE, meaning the PE number is determined according to the index
of the consumed M64 segment. As below figure shows, M64 resource
grows from low to high end, meaning the PE (number) reserved
according to M64 segment grows from low to high end as well,
so does the dynamically allocated PE number. It will lead to
conflict: PE number (M64 segment) reserved by dynamic allocation
is required by hot added PCI adapter at later point. It fails
the PCI hotplug because of the PE number can't be reserved
based on the index of the consumed M64 segment.
+---+---+---+---+---+--------------------------------+-----+
| 0 | 1 | 2 | 3 | 4 | ....... | 255 |
+---+---+---+---+---+--------------------------------+-----+
PE number for dynamic allocation ----------------->
PE number reserved for M64 segment ----------------->
To resolve above conflicts, this forces the PE number to be
allocated dynamically in reverse order. With this patch applied,
the PE numbers are reserved in ascending order, but allocated
dynamically in reverse order.
Signed-off-by: Gavin Shan <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
Each PHB maintains an array helping to translate 2-bytes Request
ID (RID) to PE# with the assumption that PE# takes one byte, meaning
that we can't have more than 256 PEs. However, pci_dn->pe_number
already had 4-bytes for the PE#.
This extends the PE# capacity for every PHB. After that, the PE number
is represented by 4-bytes value. Then we can reuse IODA_INVALID_PE to
check the PE# in phb->pe_rmap[] is valid or not.
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Daniel Axtens <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
pnv_pci_ioda_setup_opal_tce_kill() called by pnv_ioda_setup_dma()
to remap the TCE kill regiter. What's done in pnv_ioda_setup_dma()
will be covered in pcibios_setup_bridge() which is invoked on each
PCI bridge. It means we will possibly remap the TCE kill register
for multiple times and it's unnecessary.
This moves pnv_pci_ioda_setup_opal_tce_kill() to where the PHB is
initialized (pnv_pci_init_ioda_phb()) to avoid above issue.
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Alexey Kardashevskiy <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
The macro defined in arch/powerpc/platforms/powernv/pci.c isn't
used by anyone. Just remove it.
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Andrew Donnellan <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
This overrides pcibios_setup_bridge() that is called to update PCI
bridge windows when PCI resource assignment is completed, to assign
PE and setup various (resource) mapping for the PE in subsequent
patches.
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Alexey Kardashevskiy <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
Currently, PowerPC PowerNV platform utilizes ppc_md.pcibios_fixup(),
which is called for once after PCI probing and resource assignment
are completed, to allocate platform required resources for PCI devices:
PE#, IO and MMIO mapping, DMA address translation (TCE) table etc.
Obviously, it's not hotplug friendly.
This adds weak function pcibios_setup_bridge(), which is called by
pci_setup_bridge(). PowerPC PowerNV platform will reuse the function
to assign above platform required resources to newly plugged PCI devices
during PCI hotplug in subsequent patches.
Signed-off-by: Gavin Shan <[email protected]>
Acked-by: Bjorn Helgaas <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
The PE primary bus cannot be got from its child devices when having
full hotplug in error recovery. The PE primary bus is cached, which
is done in commit <05ba75f84864> ("powerpc/eeh: Fix stale cached primary
bus"). In eeh_reset_device(), the flag (EEH_PE_PRI_BUS) is cleared
before the PCI hot remove. eeh_pe_bus_get() then returns NULL as the
PE primary bus in pnv_eeh_reset() and it crashes the kernel eventually.
This fixes the issue by clearing the flag (EEH_PE_PRI_BUS) before the
PCI hot add. With it, the PowerNV EEH reset backend (pnv_eeh_reset())
can get valid PE primary bus through eeh_pe_bus_get().
Fixes: 67086e32b5 ("powerpc/eeh: powerpc/eeh: Support error recovery for VF PE")
Reported-by: Pridhiviraj Paidipeddi <[email protected]>
Signed-off-by: Gavin Shan <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
The function is unflattening device sub-tree blob if @dad passed to
the function is valid. Currently, this functionality is used by PPC
PowerNV PCI hotplug driver only. There are possibly multiple nodes
in the first level of depth, fdt_next_node() bails immediately when
@depth becomes negative before the second device node can be probed
successfully. It leads to the device nodes except the first one won't
be unflattened successfully.
This fixes the issue by setting the initial depth (@inital_depth) to
1 when this function is called to unflatten device sub-tree blob. No
logic changes when this function is used to unflatten non-sub-tree
blob.
Cc: Rhyland Klein <[email protected]>
Fixes: 78c44d910 ("drivers/of: Fix depth when unflattening devicetree")
Signed-off-by: Gavin Shan <[email protected]>
Tested-by: Rhyland Klein <[email protected]>
Tested-by: Andrew Donnellan <[email protected]>
Signed-off-by: Rob Herring <[email protected]>
In commit 8445a87f70 "powerpc/iommu: Remove the dependency on EEH
struct in DDW mechanism", the PE address was replaced with the PCI
config address in order to remove dependency on EEH. According to PAPR
spec, firmware (pHyp or QEMU) should accept "xxBBSSxx" format PCI config
address, not "xxxxBBSS" provided by the patch. Note that "BB" is PCI bus
number and "SS" is the combination of slot and function number.
This fixes the PCI address passed to DDW RTAS calls.
Fixes: 8445a87f70 ("powerpc/iommu: Remove the dependency on EEH struct in DDW mechanism")
Cc: [email protected] # v3.4+
Reported-by: Guilherme G. Piccoli <[email protected]>
Signed-off-by: Gavin Shan <[email protected]>
Tested-by: Guilherme G. Piccoli <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
When we have !NO_BOOTMEM, the deferred page struct initialization
doesn't work well because the pages reserved in bootmem are released to
the page allocator uncoditionally. It causes memory corruption and
system crash eventually.
As Mel suggested, the bootmem is retiring slowly. We fix the issue by
simply hiding DEFERRED_STRUCT_PAGE_INIT when bootmem is enabled.
Link: http://lkml.kernel.org/r/[email protected]
Signed-off-by: Gavin Shan <[email protected]>
Acked-by: Mel Gorman <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
One timer, whose handler keeps reading on MMIO register for EEH
core to detect error in time, is started when the PCI device driver
is loaded. MMIO register can't be accessed during PE reset in EEH
recovery. Otherwise, the unexpected recursive error is triggered.
The timer isn't closed that time if the interface isn't brought
up. So the unexpected recursive error is seen during EEH recovery
when the interface is down.
This avoids the unexpected recursive EEH error by closing the timer
in qlge_io_error_detected() before EEH PE reset unconditionally. The
timer is started unconditionally after EEH PE reset in qlge_io_resume().
Also, the timer should be closed unconditionally when the device is
removed from the system permanently in qlge_io_error_detected().
Reported-by: Shriya R. Kulkarni <[email protected]>
Signed-off-by: Gavin Shan <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
Function populate_node() is used to unflatten FDT blob to device
tree. It supports maximal 64 level of device nodes. There is one
array @fpsizes[64] tracking the full name length of last unflattened
device node in the corresponding level (index of element in the
array - 1). Build warning is seen with CONFIG_FRAME_WARN=1024 like
below on ARM64 as Geert reported. The issue can be reproduced on
PPC64 as well.
$ make drivers/of/fdt.o
drivers/of/fdt.c:443:1: warning: the frame size of 1136 bytes is \
larger than 1024 bytes [-Wframe-larger-than=]
This changes the data type of @fpsizes[i] from "unsigned long" to
"unsigned int" to avoid the build warning. The return value type
of populate_node() and its @fpsize argument is adjusted accordingly.
With this applied, 256 bytes saved from the stack frame on ARM64 and
PPC64 platforms and the above warning isn't seen.
Fixes: 50800082f1 ("drivers/of: Avoid recursively calling unflatten_dt_node()")
Reported-by: Geert Uytterhoeven <[email protected]>
Signed-off-by: Gavin Shan <[email protected]>
Signed-off-by: Rob Herring <[email protected]>
This returns the allocate memory chunk, storing the unflattened device
tree, from of_fdt_unflatten_tree() so that memory chunk can be released
on demand in PowerNV PCI hotplug driver.
Signed-off-by: Gavin Shan <[email protected]>
Acked-by: Rob Herring <[email protected]>
Signed-off-by: Rob Herring <[email protected]>
This adds one more argument to of_fdt_unflatten_tree() to specify
the parent node of the FDT blob that is going to be unflattened.
In the result, the function can be used to unflatten FDT blob that
represents device sub-tree in PowerNV PCI hotplug driver.
Cc: Jyri Sarha <[email protected]>
Signed-off-by: Gavin Shan <[email protected]>
Acked-by: Rob Herring <[email protected]>
Acked-by: Jyri Sarha <[email protected]>
Signed-off-by: Rob Herring <[email protected]>
This renames unflatten_dt_node() to unflatten_dt_nodes() as it
populates multiple device nodes from FDT blob. No logical changes
introduced.
Signed-off-by: Gavin Shan <[email protected]>
Acked-by: Rob Herring <[email protected]>
Signed-off-by: Rob Herring <[email protected]>
In current implementation, unflatten_dt_node() is called recursively
to unflatten device nodes in FDT blob. It's stress to limited stack
capacity, especially to adopt the function to unflatten device sub-tree
that possibly has multiple root nodes. In that case, we runs out of
stack and the system can't boot up successfully.
In order to reuse the function to unflatten device sub-tree, this avoids
calling the function recursively, meaning the device nodes are unflattened
in one call on unflatten_dt_node(): two arrays are introduced to track the
parent path size and the device node of current level of depth, which will
be used by the device node on next level of depth to be unflattened. All
device nodes in more than 64 level of depth are dropped and hopefully,
the system can boot up successfully with the partial device-tree.
Also, the parameter "poffset" and "fpsize" are unused and dropped and the
parameter "dryrun" is figured out from "mem == NULL". Besides, the return
value of the function is changed to indicate the size of memory consumed by
the unflatten device tree or error code.
Signed-off-by: Gavin Shan <[email protected]>
Acked-by: Rob Herring <[email protected]>
Signed-off-by: Rob Herring <[email protected]>
The function unflatten_dt_node() is called recursively to unflatten
device nodes and properties in the FDT blob. It looks complicated
and hard to be understood.
This splits the function into 3 functions: populate_properties(),
populate_node() and unflatten_dt_node(). populate_properties(),
which is called by populate_node(), creates properties for the
indicated device node. The later one creates the device nodes
from FDT blob. populate_node() gets the offset in FDT blob for
next device nodes and then calls populate_node(). No logical
changes introduced.
Signed-off-by: Gavin Shan <[email protected]>
Acked-by: Rob Herring <[email protected]>
Signed-off-by: Rob Herring <[email protected]>
The label "reset" in eeh_pe_change_owner() is used only for once.
No need to keep it and just drop it. No logical changes introduced.
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: David Gibson <[email protected]>
Reviewed-by: Russell Currey <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
The function eeh_pe_reset_and_recover() is used to recover EEH
error when the passthrough device are transferred to guest and
backwards, meaning the device's driver is vfio-pci or none. In
both cases, the handlers triggered by eeh_report_reset() and
eeh_report_resume() shouldn't be called.
This ignores the error handlers from eeh_report_reset() and
eeh_report_resume().
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Russell Currey <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
The function eeh_pe_reset_and_recover() is used to recover EEH
error when the passthrou device are transferred to guest and
backwards. The content in the device's config space will be lost
on PE reset issued in the middle of the recovery. The function
saves/restores it before/after the reset. However, config access
to some adapters like Broadcom BCM5719 at this point will causes
fenced PHB. The config space is always blocked and we save 0xFF's
that are restored at late point. The memory BARs are totally
corrupted, causing another EEH error upon access to one of the
memory BARs.
This restores the config space on those adapters like BCM5719
from the content saved to the EEH device when it's populated,
to resolve above issue.
Fixes: 5cfb20b9 ("powerpc/eeh: Emulate EEH recovery for VFIO devices")
Cc: [email protected] #v3.18+
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Russell Currey <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
The function eeh_pe_reset_and_recover() is used to recover EEH
error when the passthrough device are transferred to guest and
backwards, meaning the device's driver is vfio-pci or none.
When the driver is vfio-pci that provides error_detected() error
handler only, the handler simply stops the guest and it's not
expected behaviour. On the other hand, no error handlers will
be called if we don't have a bound driver.
This ignores the error handler in eeh_pe_reset_and_recover()
that reports the error to device driver to avoid the exceptional
behaviour.
Fixes: 5cfb20b9 ("powerpc/eeh: Emulate EEH recovery for VFIO devices")
Cc: [email protected] #v3.18+
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Russell Currey <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
The function pnv_pci_reset_secondary_bus() is called like below.
It's impossible for call the function on root bus. So it's safe
to remove the root bus case in the function. No functional changes
introduced.
pci_parent_bus_reset() / pci_bus_reset() / pci_try_reset_bus()
pci_reset_bridge_secondary_bus()
pcibios_reset_secondary_bus()
pnv_pci_reset_secondary_bus()
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Daniel Axtens <[email protected]>
Reviewed-by: Alexey Kardashevskiy <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
This drops unnecessary nested if statements in pnv_eeh_reset() to
improve the code readability. After the changes, the unused local
variable "ret" is dropped as well. No logical changes introduced.
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Andrew Donnellan <[email protected]>
Reviewed-by: Alexey Kardashevskiy <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
In hotplug case, function pci_add_pci_devices() is called to rescan
the specified PCI bus, which might not have any child devices. Access
to the PCI bus's child device node will cause kernel crash without
exception.
This adds one more check to skip scanning PCI bus that doesn't have
any subordinate devices from device-tree, in order to avoid kernel
crash.
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Alexey Kardashevskiy <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
This renames traverse_pci_devices() to pci_traverse_device_nodes().
The function traverses all subordinate device nodes of the specified
one. Also, below cleanup applied to the function. No logical changes
introduced.
* Rename "pre" to "fn".
* Avoid assignment in if condition reported from checkpatch.pl.
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Alexey Kardashevskiy <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
This implements and exports pci_remove_device_node_info(). It's
used to remove the pdn (struct pci_dn) for the indicated device
node. The function is going to be used by PowerNV PCI hotplug
driver.
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Alexey Kardashevskiy <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
This renames update_dn_pci_info() to pci_add_device_node_info()
with corresponding adjustment on the parameter type and exports it.
The function is used to create pdn (struct pci_dn) for the indicated
device node. Another function add_pdn(), almost wrapper of
pci_add_device_node_info(), to be used in traverse_pci_devices(). No
logical changes introduced.
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Alexey Kardashevskiy <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
This moves pci_find_bus_by_node() from arch/powerpc/platforms/
pseries/pci_dlpar.c to arch/powerpc/kernel/pci-hotplug.c so that
the function can be used by pSeries and PowerNV platform at the
same time. Also, below cleanup applied. No functional changes
introduced.
* Remove variable "busdn" in find_bus_among_children()
* Use PCI_DN() to convert device node to pci_dn
Signed-off-by: Gavin Shan <[email protected]>
Acked-by: Benjamin Herrenschmidt <[email protected]>
Reviewed-by: Andrew Donnellan <[email protected]>
Reviewed-by: Alexey Kardashevskiy <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
This renames pcibios_find_pci_bus() to pci_find_bus_by_node() to
avoid conflicts with those PCI subsystem weak function names, which
have prefix "pcibios". No logical changes introduced.
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Alexey Kardashevskiy <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
This renames pcibios_{add,remove}_pci_devices() to avoid conflicts
with names of the weak functions in PCI subsystem, which have the
prefix "pcibios". No logical changes introduced.
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-By: Alistair Popple <[email protected]>
Reviewed-by: Andrew Donnellan <[email protected]>
Reviewed-by: Alexey Kardashevskiy <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
In current implementation, the PEs that are allocated or picked
from the reserved list are identified by PE number. The PE instance
has to be picked according to the PE number eventually. We have
same issue when PE is released.
For pnv_ioda_pick_m64_pe() and pnv_ioda_alloc_pe(), this returns
PE instance so that pnv_ioda_setup_bus_PE() can use the allocated
or reserved PE instance directly. Also, pnv_ioda_setup_bus_PE()
returns the reserved/allocated PE instance to be used in subsequent
patches. On the other hand, pnv_ioda_free_pe() uses PE instance
(not number) as its argument. No logical changes introduced.
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Alexey Kardashevskiy <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
In current implementation, the DMA32 segments required by one specific
PE isn't calculated with the information hold in the PE independently.
It conflicts with the PCI hotplug design: PE centralized, meaning the
PE's DMA32 segments should be calculated from the information hold in
the PE independently.
This introduces an array (@dma32_segmap) for every PHB to track the
DMA32 segmeng usage. Besides, this moves the logic calculating PE's
consumed DMA32 segments to pnv_pci_ioda1_setup_dma_pe() so that PE's
DMA32 segments are calculated/allocated from the information hold in
the PE (DMA32 weight). Also the logic is improved: we try to allocate
as much DMA32 segments as we can. It's acceptable that number of DMA32
segments less than the expected number are allocated.
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Alexey Kardashevskiy <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
PEs are put into PHB DMA32 list (phb->ioda.pe_dma_list) according
to their DMA32 weight. The PEs on the list are iterated to setup
their TCE32 tables at system booting time. The list is used for
once at boot time and no need to keep it.
This moves the logic calculating DMA32 weight of PHB and PE to
pnv_ioda_setup_dma() to drop PHB's DMA32 list. Also, every PE
traces the consumed DMA32 segment by @tce32_seg and @tce32_segcount
are useless and they're removed.
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Alexey Kardashevskiy <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
Currently, there is one macro (TCE32_TABLE_SIZE) representing the
TCE table size for one DMA32 segment. The constant representing
the DMA32 segment size (1 << 28) is still used in the code.
This defines PNV_IODA1_DMA32_SEGSIZE representing one DMA32
segment size. the TCE table size can be calcualted when the page
has fixed 4KB size. So all the related calculation depends on one
macro (PNV_IODA1_DMA32_SEGSIZE). No logical changes introduced.
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-By: Alistair Popple <[email protected]>
Reviewed-by: Alexey Kardashevskiy <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
This renames pnv_pci_ioda_setup_dma_pe() to pnv_pci_ioda1_setup_dma_pe()
as it's the counter-part of IODA2's pnv_pci_ioda2_setup_dma_pe().
No logical changes introduced.
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Alexey Kardashevskiy <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
This enables M64 window on P7IOC, which has been enabled on PHB3.
Different from PHB3 where 16 M64 BARs are supported and each of
them can be owned by one particular PE# exclusively or divided
evenly to 256 segments, every P7IOC PHB has 16 M64 BARs and each
of them are divided to 8 segments. So every P7IOC PHB supports
128 M64 segments in total. P7IOC has M64DT, which helps mapping
one particular M64 segment# to arbitrary PE#. PHB3 doesn't have
M64DT, indicating that one M64 segment can only be pinned to the
fixed PE#.
In order to unified M64 support M64 on P7IOC and PHB3, we just
provide 128 M64 segments on every P7IOC PHB and each of them is
pinned to the fixed PE# by bypassing the function of M64DT. In
turn, we just need different phb->init_m64() for P7IOC and PHB3
and maps M64 segment in pnv_ioda_reserve_m64_pe() for P7IOC, most
of the code are shared by them.
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Alistair Popple <[email protected]>
Reviewed-by: Alexey Kardashevskiy <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
This renames those functions picking PE number based on consumed
M64 segments, mapping M64 segments to PEs as those functions are
going to be shared by IODA1/IODA2 in next patch. No logical changes
introduced.
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Alexey Kardashevskiy <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
When unplugging PCI devices, their parent PEs might be offline.
The consumed M64 resource by the PEs should be released at that
time. As we track M32 segment consumption, this introduces an
array to the PHB to track the mapping between M64 segment and
PE number.
Note: M64 mapping isn't covered by pnv_ioda_setup_pe_seg() as
IODA2 doesn't support the mapping explicitly while it's supported
on IODA1. Until now, no M64 is supported on IODA1 in software.
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Alexey Kardashevskiy <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
Currently, the IO and M32 segments are mapped to the corresponding
PE based on the windows of the parent bridge of PE's primary bus.
It's not going to work when the windows of root port or upstream
port of the PCIe switch behind root port are extended to PHB's
apertures in order to support hotplug in subsequent patch.
This fixes the issue by mapping IO and M32 segments based on the
resources of the PCI devices included in the PE, instead of the
windows of the parent bridge of the PE's primary bus.
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Alexey Kardashevskiy <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
pnv_ioda_setup_pe_seg() associates the IO and M32 segments with the
owner PE. The code mapping segments should be fixed and immune from
logic changes introduced to pnv_ioda_setup_pe_seg().
This moves the code mapping segments to helper pnv_ioda_setup_pe_res().
The data type for @rc is changed to "int64_t". Also, argument @hose is
removed from pnv_ioda_setup_pe() as it can be got from @pe. No functional
changes introduced.
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-By: Alistair Popple <[email protected]>
Reviewed-by: Alexey Kardashevskiy <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
There are two arrays for IO and M32 segment maps on every PHB.
The index of the arrays are segment number and the value stored
in the corresponding element is PE number, indicating the segment
is assigned to the PE. Initially, all elements in those two arrays
are zeroes, meaning all segments are assigned to PE#0. It's wrong.
This fixes the initial values in the elements of those two arrays
to IODA_INVALID_PE, meaning all segments aren't assigned to any
PE.
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Alexey Kardashevskiy <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
This changes the data type of PE number from "int" to "unsigned int"
in order to match the fact PE number is never negative:
* The number of PE to which the specified PCI device is attached.
* The PE number map for SRIOV VFs.
* The returned PE number from pnv_ioda_alloc_pe().
* The returned PE number from pnv_ioda2_pick_m64_pe().
Suggested-by: Alexey Kardashevskiy <[email protected]>
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-By: Alistair Popple <[email protected]>
Reviewed-by: Alexey Kardashevskiy <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
This renames the fields related to PE number in "struct pnv_phb"
for better reflecting of their usages as Alexey suggested. No
logical changes introduced.
Signed-off-by: Gavin Shan <[email protected]>
Reviewed-by: Alexey Kardashevskiy <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
Gavin Shan