Remove abandonned capi support for the Mellanox CX4.
This reverts commit a19bd79e31.
Signed-off-by: Frederic Barrat <fbarrat@linux.ibm.com>
Acked-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Remove abandonned capi support for the Mellanox CX4.
This reverts commit 317f5ef1b3.
Signed-off-by: Alastair D'Silva <alastair@d-silva.org>
Acked-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
We want to use the dma_direct_ namespace for a generic implementation,
so rename powerpc to the second best choice: dma_nommu_.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Commit 14a3ae34bf ("cxl: Prevent read/write to AFU config space while AFU
not configured") introduced a rwsem to fix an invalid memory access that
occurred when someone attempts to access the config space of an AFU on a
vPHB whilst the AFU is deconfigured, such as during EEH recovery.
It turns out that it's possible to run into a nested locking issue when EEH
recovery fails and a full device hotplug is required.
cxl_pci_error_detected() deconfigures the AFU, taking a writer lock on
configured_rwsem. When EEH recovery fails, the EEH code calls
pci_hp_remove_devices() to remove the device, which in turn calls
cxl_remove() -> cxl_pci_remove_afu() -> pci_deconfigure_afu(), which tries
to grab the writer lock that's already held.
Standard rwsem semantics don't express what we really want to do here and
don't allow for nested locking. Fix this by replacing the rwsem with an
atomic_t which we can control more finely. Allow the AFU to be locked
multiple times so long as there are no readers.
Fixes: 14a3ae34bf ("cxl: Prevent read/write to AFU config space while AFU not configured")
Cc: stable@vger.kernel.org # v4.9+
Signed-off-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com>
Acked-by: Frederic Barrat <fbarrat@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
During EEH recovery, we deconfigure all AFUs whilst leaving the
corresponding vPHB and virtual PCI device in place.
If something attempts to interact with the AFU's PCI config space (e.g.
running lspci) after the AFU has been deconfigured and before it's
reconfigured, cxl_pcie_{read,write}_config() will read invalid values from
the deconfigured struct cxl_afu and proceed to Oops when they try to
dereference pointers that have been set to NULL during deconfiguration.
Add a rwsem to struct cxl_afu so we can prevent interaction with config
space while the AFU is deconfigured.
Reported-by: Pradipta Ghosh <pradghos@in.ibm.com>
Suggested-by: Frederic Barrat <fbarrat@linux.vnet.ibm.com>
Cc: stable@vger.kernel.org # v4.9+
Signed-off-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com>
Signed-off-by: Vaibhav Jain <vaibhav@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
When cxl removes a vPHB, it's possible that the pci_controller may be freed
before all references to the devices on the vPHB have been released. This
in turn causes an invalid memory access when the devices are eventually
released, as pcibios_release_device() attempts to call the phb's
release_device hook.
In cxl_pci_vphb_remove(), remove the existing call to
pcibios_free_controller(). Instead, use
pcibios_free_controller_deferred() to free the pci_controller after all
devices have been released. Export pci_set_host_bridge_release() so we can
do this.
Cc: stable@vger.kernel.org
Signed-off-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com>
Reviewed-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Acked-by: Ian Munsie <imunsie@au1.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Make native_irq_wait() static and use NULL rather than 0 to initialise
phb->cfg_data in cxl_pci_vphb_add() to remove sparse warnings.
Signed-off-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com>
Reviewed-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Acked-by: Ian Munsie <imunsie@au1.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
This hooks up support for using the kernel API with a real PHB. After
the AFU initialisation has completed it calls into the PHB code to pass
it the AFU that will be used by other peer physical functions on the
adapter.
The cxl_pci_to_afu API is extended to work with peer PCI devices,
retrieving the peer AFU from the PHB. This API may also now return an
error if it is called on a PCI device that is not associated with either
a cxl vPHB or a peer PCI device to an AFU, and this error is propagated
down.
Signed-off-by: Ian Munsie <imunsie@au1.ibm.com>
Reviewed-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
The vPHB model of the cxl kernel API is a hierarchy where the AFU is
represented by the vPHB, and it's AFU configuration records are exposed
as functions under that vPHB. If there are no AFU configuration records
we will create a vPHB with nothing under it, which is a waste of
resources and will opt us into EEH handling despite not having anything
special to handle.
This also does not make sense for cards using the peer model of the cxl
kernel API, where the other functions of the device are exposed via
additional peer physical functions rather than AFU configuration
records. This model will also not work with the existing EEH handling in
the cxl driver, as that is designed around the vPHB model.
Skip creating the vPHB for AFUs without any AFU configuration records,
and opt out of EEH handling for them.
Signed-off-by: Ian Munsie <imunsie@au1.ibm.com>
Reviewed-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
The cxl kernel API has a concept of a default context associated with
each PCI device under the virtual PHB. The Mellanox CX4 will also use
the cxl kernel API, but it does not use a virtual PHB - rather, the AFU
appears as a physical function as a peer to the networking functions.
In order to allow the kernel API to work with those networking
functions, we will need to associate a default context with them as
well. To this end, refactor the corresponding code to do this in vphb.c
and export it so that it can be called from the PHB code.
Signed-off-by: Ian Munsie <imunsie@au1.ibm.com>
Reviewed-by: Frederic Barrat <fbarrat@linux.vnet.ibm.com>
Reviewed-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
On bare-metal, when a device is attached to the cxl card, lsvpd shows
a location code such as (with cxlflash):
# lsvpd -l sg22
...
*YL U78CB.001.WZS0073-P1-C33-B0-T0-L0
which makes it hard to easily identify the cxl adapter owning the
flash device, since in this example C33 refers to a P8 processor.
lsvpd looks in the parent devices until it finds a location code, so the
device node for the vPHB ends up being used.
By reusing the device node of the adapter for the vPHB, lsvpd shows:
# lsvpd -l sg16
...
*YL U78C9.001.WZS09XA-P1-C7-B1-T0-L3
where C7 is the PCI slot of the cxl adapter.
On powerVM, the vPHB was already using the adapter device node, so
there's no change there.
Tested by cxlflash on bare-metal and powerVM.
Signed-off-by: Frederic Barrat <fbarrat@linux.vnet.ibm.com>
Reviewed-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Acked-by: Ian Munsie <imunsie@au1.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Add a check at the beginning of cxl_probe function to ignore virtual pci
devices created for each afu registered. This fixes the the errors
messages logged about missing CXL vsec, when cxl probe is unable to
find necessary vsec entries in device pci config space. The error
message logged are of the form :
cxl-pci 0004:00:00.0: ABORTING: CXL VSEC not found!
cxl-pci 0004:00:00.0: cxl_init_adapter failed: -19
Cc: Ian Munsie <imunsie@au1.ibm.com>
Cc: Frederic Barrat <fbarrat@linux.vnet.ibm.com>
Signed-off-by: Vaibhav Jain <vaibhav@linux.vnet.ibm.com>
Reviewed-by: fbarrat@linux.vnet.ibm.com
Reviewed-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com>
Acked-by: Ian Munsie <imunsie@au1.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Check the AFU state whenever an API is called. The hypervisor may
issue a reset of the adapter when it detects a fault. When it happens,
it launches an error recovery which will either move the AFU to a
permanent failure state, or in the disabled state.
If the AFU is found to be disabled, detach all existing contexts from
it before issuing a AFU reset to re-enable it.
Before detaching contexts, notify any kernel driver through the EEH
callbacks of the AFU pci device.
Co-authored-by: Frederic Barrat <fbarrat@linux.vnet.ibm.com>
Signed-off-by: Frederic Barrat <fbarrat@linux.vnet.ibm.com>
Signed-off-by: Christophe Lombard <clombard@linux.vnet.ibm.com>
Reviewed-by: Manoj Kumar <manoj@linux.vnet.ibm.com>
Acked-by: Ian Munsie <imunsie@au1.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Like on bare-metal, the cxl driver creates a virtual PHB and a pci
device for the AFU. The configuration space of the device is mapped to
the configuration record of the AFU.
Reuse the code defined in afu_cr_read8|16|32() when reading the
configuration space of the AFU device.
Even though the (virtual) AFU device is a pci device, the adapter is
not. So a driver using the cxl kernel API cannot read the VPD of the
adapter through the usual PCI interface. Therefore, we add a call to
the cxl kernel API:
ssize_t cxl_read_adapter_vpd(struct pci_dev *dev, void *buf, size_t count);
Co-authored-by: Christophe Lombard <clombard@linux.vnet.ibm.com>
Signed-off-by: Frederic Barrat <fbarrat@linux.vnet.ibm.com>
Signed-off-by: Christophe Lombard <clombard@linux.vnet.ibm.com>
Reviewed-by: Manoj Kumar <manoj@linux.vnet.ibm.com>
Acked-by: Ian Munsie <imunsie@au1.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Introduce sub-structures containing the bare-metal specific fields in
the structures describing the adapter (struct cxl) and AFU (struct
cxl_afu).
Update all their references.
Co-authored-by: Frederic Barrat <fbarrat@linux.vnet.ibm.com>
Signed-off-by: Frederic Barrat <fbarrat@linux.vnet.ibm.com>
Signed-off-by: Christophe Lombard <clombard@linux.vnet.ibm.com>
Reviewed-by: Manoj Kumar <manoj@linux.vnet.ibm.com>
Acked-by: Ian Munsie <imunsie@au1.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
The backend API (in cxl.h) lists some low-level functions whose
implementation is different on bare-metal and in a guest. Each
environment implements its own functions, and the common code uses
them through function pointers, defined in cxl_backend_ops
Co-authored-by: Christophe Lombard <clombard@linux.vnet.ibm.com>
Signed-off-by: Frederic Barrat <fbarrat@linux.vnet.ibm.com>
Signed-off-by: Christophe Lombard <clombard@linux.vnet.ibm.com>
Reviewed-by: Manoj Kumar <manoj@linux.vnet.ibm.com>
Acked-by: Ian Munsie <imunsie@au1.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
When writing a value to config space, cxl_pcie_write_config() calls
cxl_pcie_config_info() to obtain a mask and shift value, shifts the new
value accordingly, then uses the mask to combine the shifted value with the
existing value at the address as part of a read-modify-write pattern.
Currently, we use a logical OR operator rather than a bitwise OR operator,
which means any use of this function results in an incorrect value being
written. Replace the logical OR operator with a bitwise OR operator so the
value is written correctly.
Reported-by: Michael Ellerman <mpe@ellerman.id.au>
Cc: stable@vger.kernel.org
Fixes: 6f7f0b3df6 ("cxl: Add AFU virtual PHB and kernel API")
Signed-off-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com>
Acked-by: Ian Munsie <imunsie@au1.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
When adding a vPHB in cxl_pci_vphb_add(), we allocate a pci_controller
struct using pcibios_alloc_controller(). However, we don't free it in
cxl_pci_vphb_remove(), causing a leak.
Call pcibios_free_controller() in cxl_pci_vphb_remove() to free the vPHB
data structure correctly.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
cxl_pci_enable_device_hook() is called when attempting to enable an AFU
sitting on a vPHB. At present, the state of the underlying CXL card's PCI
channel is only checked when it calls cxl_afu_check_and_enable() at the
very end, after it has already set DMA options and initialised a default
context.
Check the CXL card's link status before setting DMA options or initialising
a default context. If the link is down, print a warning and return
immediately.
Signed-off-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com>
Acked-by: Ian Munsie <imunsie@au1.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
EEH (Enhanced Error Handling) allows a driver to recover from the
temporary failure of an attached PCI card. Enable basic CXL support
for EEH.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
If the PCI channel has gone down, don't attempt to poke the hardware.
We need to guard every time cxl_whatever_(read|write) is called. This
is because a call to those functions will dereference an offset into an
mmio register, and the mmio mappings get invalidated in the EEH
teardown.
Check in the read/write functions in the header.
We give them the same semantics as usual PCI operations:
- a write to a channel that is down is ignored.
- a read from a channel that is down returns all fs.
Also, we try to access the MMIO space of a vPHB device as part of the
PCI disable path. Because that's a read that bypasses most of our usual
checks, we handle it explicitly.
As far as user visible warnings go:
- Check link state in file ops, return -EIO if down.
- Be reasonably quiet if there's an error in a teardown path,
or when we already know the hardware is going down.
- Throw a big WARN if someone tries to start a CXL operation
while the card is down. This gives a useful stacktrace for
debugging whatever is doing that.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
static Anlaysis detected below error:-
(error) Possible null pointer dereference: phb
So, Use phb after NULL check.
Signed-off-by: Maninder Singh <maninder1.s@samsung.com>
Acked-by: Ian Munsie <imunsie@au1.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
When we release the device, we should also invalidate the default context.
With this cxl_get_context() will return null after removal.
Signed-off-by: Michael Neuling <mikey@neuling.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
This patch does two things.
Firstly it presents the Accelerator Function Unit (AFUs) behind the POWER
Service Layer (PSL) as PCI devices on a virtual PCI Host Bridge (vPHB). This
in in addition to the PSL being a PCI device itself.
As part of the Coherent Accelerator Interface Architecture (CAIA) AFUs can
provide an AFU configuration. This AFU configuration recored is architected to
be the same as a PCI config space.
This patch sets discovers the AFU configuration records, provides AFU config
space read/write functions to these configuration records. It then enumerates
the PCI bus. It also hooks in PCI ops where appropriate. It also destroys the
vPHB when the physical card is removed.
Secondly, it add an in kernel API for AFU to use CXL. AFUs must present a
driver that firstly binds as a PCI device. This PCI device can then be using
to do CXL specific operations (that can't sit in the PCI ops) using this API.
Signed-off-by: Michael Neuling <mikey@neuling.org>
Acked-by: Ian Munsie <imunsie@au1.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
