test/common: Commonize get_block_dev*() functions

Adjust get_nvme_name_from_bdf() so it calls to
get_block_dev_from_nvme().

Change-Id: If5f0eb573878a42e0ad0ab0453c73d39cbdf6020
Signed-off-by: Micha

test/common: Commonize get_block_dev*() functions

Adjust get_nvme_name_from_bdf() so it calls to
get_block_dev_from_nvme().

Change-Id: If5f0eb573878a42e0ad0ab0453c73d39cbdf6020
Signed-off-by: Michal Berger <michal.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/23508
Reviewed-by: Jim Harris <jim.harris@samsung.com>
Reviewed-by: Marek Chomnicki <marek.chomnicki@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>

show more ...

scripts/setup: Make sure get_block_dev_from_nvme() doesn't trigger errexit

This function should just list block devices which belong to given
subsysnqn, anything else should be simply ignored.

Chan

scripts/setup: Make sure get_block_dev_from_nvme() doesn't trigger errexit

This function should just list block devices which belong to given
subsysnqn, anything else should be simply ignored.

Change-Id: I06804b8c29e20bc4bf359f5fd8499ec2edde45bf
Signed-off-by: Michal Berger <michal.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/23507
Reviewed-by: Marek Chomnicki <marek.chomnicki@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Jim Harris <jim.harris@samsung.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>

show more ...

scripts/setup.sh: Use HUGE_EVEN_ALLOC logic by default

To that end, remove it altogether and allow setup.sh to always split
requested amount of hugepages across all available nodes. Custom
setups pe

scripts/setup.sh: Use HUGE_EVEN_ALLOC logic by default

To that end, remove it altogether and allow setup.sh to always split
requested amount of hugepages across all available nodes. Custom
setups per node are still available through the HUGENODE var.

Adjust some of the hugepages.sh tests to adhere to the new default.
This change allows us to remove the per_node_1G_alloc() test since
its flow is now covered by the custom_alloc().

By default, autotest.sh sets HUGEMEM=4096 and by spreading it across
all the nodes, for the minimal scenario, we get only 2GB on a single
node. For vhost tests, the default alloc per VM via vm_setup() is set
to 1GB so sharing that 2GB between qemu and SPDK spreads it a bit too
thin.

Case in point, for vhost.vhost_blk_packed_ring_integrity test roughly
> 600 hugepages is used for vhost. Since it slurps everything from
node0, only < 512 remains. Since VMs are assigned per node, the most
basic setups keep VM_*_qemu_numa_node set to 0. This is then used by
qemu to bind to a specific node. With 1G, < 512 hugepages is simply
not enough.

With that in mind, for vhost tests, keep all allocations on a single
node instead of trying to figure out right amount of memory per node
to keep the old behavior in (this may change in the future when
SPDK/vhost becomes numa-aware).

Change-Id: I83e18bfa4cc6de0a777804b354de083ae6ae9d8c
Signed-off-by: Michal Berger <michal.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/24176
Community-CI: Mellanox Build Bot
Reviewed-by: Tomasz Zawadzki <tomasz@tzawadzki.com>
Reviewed-by: Jim Harris <jim.harris@samsung.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>

show more ...

scripts/setup: Try to gracefully handle unsupported nic_uio devices

Under FreeBSD, since we perform a full nic_uio|contigmem reload, we
may affect devices which are not directly supported by the set

scripts/setup: Try to gracefully handle unsupported nic_uio devices

Under FreeBSD, since we perform a full nic_uio|contigmem reload, we
may affect devices which are not directly supported by the setup.sh.

To mitigate it, try to lookup kernel environment to see if loader
already provides setup for "extra" devices. If so, bail and allow
to continue only when new $FORCE_NIC_UIO_REBIND var is passed.

When $FORCE_NIC_UIO_REBIND is passed, the list of devices to set
via hw.nic_uio.bdfs will include all the "extra" devices to make
sure they are reconfigured as well after the driver's reload -
this includes the 'reset' mode.

Change-Id: I9f1b7784e3cd03c5e7cb4867970e206b6e04bd13
Signed-off-by: Michal Berger <michal.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/24175
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Jim Harris <jim.harris@samsung.com>

show more ...

scripts/check_format: Adjust bash code to adhere to shfmt v3.8.0

This bump ensures some of the obscure formatting bugs are finally
squashed. All the adjustments introduced here adhere to what we
alr

scripts/check_format: Adjust bash code to adhere to shfmt v3.8.0

This bump ensures some of the obscure formatting bugs are finally
squashed. All the adjustments introduced here adhere to what we
already try to enforce across the entire repo, but due to said bugs
could not be implemented.

Remove SHFMT_NO_EDITORCONFIG since it's not supported in latest
shfmt releases.

This patch doesn't introduce any functional change in the updated
files.

Change-Id: Ia2b011b86c8fe12376c4a8455ee638473a9f7179
Signed-off-by: Michal Berger <michal.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/23088
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Community-CI: Mellanox Build Bot

show more ...

scripts: Disable DSA and IAA default vfio-pci binding

The DSA and IAA device IDs have been added to the vfio denylist,
due to an erratum (see INTEL-SA-01084 security advisory).
This patch disables t

scripts: Disable DSA and IAA default vfio-pci binding

The DSA and IAA device IDs have been added to the vfio denylist,
due to an erratum (see INTEL-SA-01084 security advisory).
This patch disables the default behavior of vfio-pci binding
for these devices.

To force the binding, it is possible to use the PCI_ALLOWED variable.

Change-Id: I25be317156f19f400f76961c5928d602043125dc
Signed-off-by: Slawomir Ptak <slawomir.ptak@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/23196
Reviewed-by: Sebastian Brzezinka <sebastian.brzezinka@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Community-CI: Mellanox Build Bot

show more ...

setup.sh: emit command line to /dev/kmsg on Linux

This can help delineate kernel log messages as result of running
setup.sh vs. other reasons.

Signed-off-by: Jim Harris <jim.harris@samsung.com>
Cha

setup.sh: emit command line to /dev/kmsg on Linux

This can help delineate kernel log messages as result of running
setup.sh vs. other reasons.

Signed-off-by: Jim Harris <jim.harris@samsung.com>
Change-Id: I9b523a0dc54975beca121696ab33c264a430a1d2
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/22673
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Michal Berger <michal.berger@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>

show more ...

scripts/setup: Refactor get_block_dev_from_bdf()

Split it into get_block_dev_from_nvme() and fetch block devices
based on their subsysnqn association rather than their direct
link to the ctrl - setu

scripts/setup: Refactor get_block_dev_from_bdf()

Split it into get_block_dev_from_nvme() and fetch block devices
based on their subsysnqn association rather than their direct
link to the ctrl - setups where block device points at a virtual
subsystem, even though the underlying target is pcie, seem to
be possible.

Change-Id: I86b1b2db3a7f69176dcdf9e978ff3abd6b0caaf0
Signed-off-by: Michal Berger <michal.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/21492
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Jim Harris <jim.harris@samsung.com>

show more ...

scripts/setup: Add support for $DEV_TYPE

This var should hold a pattern matching any of the supported types.
In case of a match, only given type of devices would be acted on.
extglob patterns are al

scripts/setup: Add support for $DEV_TYPE

This var should hold a pattern matching any of the supported types.
In case of a match, only given type of devices would be acted on.
extglob patterns are allowed as well for more flexibility, e.g.:

$ DEV_TYPE=@(nvme|virtio) ./scripts/setup.sh status

Change-Id: Ia526fdafcd6456555b5dde5ddce82021426c1400
Signed-off-by: Michal Berger <michal.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/22583
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <jim.harris@samsung.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>

show more ...

scripts/setup: Fix driver printout

In case given device was not bound to any driver it was returning
just an empty string. So changing something like this:

0000:5d:05.5 (8086 201d): WARNING: 10000:

scripts/setup: Fix driver printout

In case given device was not bound to any driver it was returning
just an empty string. So changing something like this:

0000:5d:05.5 (8086 201d): WARNING: 10000:00:02.0 not bound to vfio-pci ()

into:

0000:5d:05.5 (8086 201d): WARNING: 10000:00:02.0 not bound to vfio-pci (no driver)

Change-Id: I8cb7795c39350bbe1bb22a778b777cacefc7fd7a
Signed-off-by: Michal Berger <michal.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/21874
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <jim.harris@samsung.com>

show more ...

scripts: Commonize routines for creating map of supported devices

It may be useful to have these outside of setup.sh for various
use-cases, e.g. VMD handling.

This comes with several enhancements:

scripts: Commonize routines for creating map of supported devices

It may be useful to have these outside of setup.sh for various
use-cases, e.g. VMD handling.

This comes with several enhancements:

- drop of the redundant driver lookup; using solely pci_bus_driver[]
- more detailed message regarding the VMD setup and nvme drives that
may be behind it

Change-Id: I5d5289fa8af435d036b75ac1fdcf9fa5ad7e53bf
Signed-off-by: Michal Berger <michal.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/21063
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <jim.harris@samsung.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>

show more ...

scripts/setup: Fix iommug[@] lookup

b44bbd240f forgot to update sysfs paths used while checking the
driver link, fix it.

Change-Id: I3256dcc6f8e7924be2994af07c62043e8ec646b7
Signed-off-by: Michal B

scripts/setup: Fix iommug[@] lookup

b44bbd240f forgot to update sysfs paths used while checking the
driver link, fix it.

Change-Id: I3256dcc6f8e7924be2994af07c62043e8ec646b7
Signed-off-by: Michal Berger <michal.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/21190
Reviewed-by: Jim Harris <jim.harris@samsung.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Karol Latecki <karol.latecki@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Community-CI: Mellanox Build Bot

show more ...

scripts/setup: Fix collect_driver() fallbacks for DSA and IAA devices

Cosmetic in nature, since the main modalias check would stil point at
proper driver, but nonetheless erroneous. Fix it and point

scripts/setup: Fix collect_driver() fallbacks for DSA and IAA devices

Cosmetic in nature, since the main modalias check would stil point at
proper driver, but nonetheless erroneous. Fix it and point at the idxd
as that's the right driver responsible for handling these supported
by SPDK devices.

Change-Id: Ibd486027d5eb9b09c6fd37b549aa02676ab82f09
Signed-off-by: Michal Berger <michal.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/21139
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Jim Harris <jim.harris@samsung.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Karol Latecki <karol.latecki@intel.com>

show more ...

scripts/setup: Skip nvme drives behind VMD sharing the same iommu_group

Also, keep track of the nvme-to-VMD relation in the status output to
easily indicate which nvme belongs to which VMD.

Change-

scripts/setup: Skip nvme drives behind VMD sharing the same iommu_group

Also, keep track of the nvme-to-VMD relation in the status output to
easily indicate which nvme belongs to which VMD.

Change-Id: I9295685be6e83e3ada1ef72b7e03d2e350678047
Signed-off-by: Michal Berger <michal.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/20958
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>

show more ...

scripts/setup: Use cached info about the iommu_groups

Change-Id: Ia64c374aa16aed9c22c8ca1619deb674e951d4c6
Signed-off-by: Michal Berger <michal.berger@intel.com>
Reviewed-on: https://review.spdk.io/

scripts/setup: Use cached info about the iommu_groups

Change-Id: Ia64c374aa16aed9c22c8ca1619deb674e951d4c6
Signed-off-by: Michal Berger <michal.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/20957
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Jim Harris <jim.harris@samsung.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot

show more ...

scripts/common: Wrap iommu check in is_iommu_enabled()

Change-Id: Id04820450aa3478621f01bbdbd18937da28d9bd8
Signed-off-by: Michal Berger <michal.berger@intel.com>
Reviewed-on: https://review.spdk.io

scripts/common: Wrap iommu check in is_iommu_enabled()

Change-Id: Id04820450aa3478621f01bbdbd18937da28d9bd8
Signed-off-by: Michal Berger <michal.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/20956
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Jim Harris <jim.harris@samsung.com>
Community-CI: Mellanox Build Bot

show more ...

scripts/setup: Change how nvme ctrl is picked up

Don't rely on block device's name since scenario where, for
instance, nvme0n1 is attached to nvme2 is plausible:

# readlink -f /sys/block/nvme0n1/d

scripts/setup: Change how nvme ctrl is picked up

Don't rely on block device's name since scenario where, for
instance, nvme0n1 is attached to nvme2 is plausible:

# readlink -f /sys/block/nvme0n1/device
/sys/devices/pci0000:80/0000:80:03.0/0000:85:00.0/nvme/nvme2

Change-Id: I21fcb5abb31bc424041d33750d4249224e0144a4
Signed-off-by: Michal Berger <michal.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/19656
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Jim Harris <jim.harris@gmail.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>

show more ...

scripts/setup: Add some extra details to pdevices()

To allow user identify given devices easier, especially when they
are bound to userspace drivers.

Change-Id: I66d4de6204aa62b23a0ddec4ff851fdc614

scripts/setup: Add some extra details to pdevices()

To allow user identify given devices easier, especially when they
are bound to userspace drivers.

Change-Id: I66d4de6204aa62b23a0ddec4ff851fdc61490fc3
Signed-off-by: Michal Berger <michal.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/19443
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Jim Harris <jim.harris@gmail.com>

show more ...

scripts/setup: Add quick interactive mode

Change-Id: I08dde2baba4967e1bf5c8944dd7513fc010d6d11
Signed-off-by: Michal Berger <michal.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spd

scripts/setup: Add quick interactive mode

Change-Id: I08dde2baba4967e1bf5c8944dd7513fc010d6d11
Signed-off-by: Michal Berger <michal.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/19442
Reviewed-by: Jim Harris <jim.harris@gmail.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Community-CI: Mellanox Build Bot

show more ...

scripts/setup: Interactive mode

Present a "hook" which allows setup.sh to enter interactive mode.
In this mode user can inspect setups.sh's view of the system and
adjust it if needed before falling

scripts/setup: Interactive mode

Present a "hook" which allows setup.sh to enter interactive mode.
In this mode user can inspect setups.sh's view of the system and
adjust it if needed before falling back to one of the main modes
(config, reset).

Change-Id: Ia945172fc7d50ffebc2da80310411c85b3957a99
Signed-off-by: Michal Berger <michal.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/18356
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benwalker@nvidia.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Community-CI: Mellanox Build Bot

show more ...

scripts/setup: Attempt to unbind all devices from target iommu group

Depending on the pci topology, there may be setups where target nvme
is part of iommu group which consists of more devices bound

scripts/setup: Attempt to unbind all devices from target iommu group

Depending on the pci topology, there may be setups where target nvme
is part of iommu group which consists of more devices bound to
various different drivers. In such a case, DPDK will refuse to
configure the device and mark given iommu group as "not viable".

To workaround this, we may try to unbind all remaining devices from
their drivers (binding to vfio or uio may fail, depending on what
kind of device we are dealing with). However, since this is an
invasive operation it should be done only on demand.

Successfully tested under WFP platform where nvme was connected to
fanout pcie switches, however, results may vary across the systems.

Change-Id: I4f315cf2e4178841859ecbba5ca09cae5fcbd497
Signed-off-by: Michal Berger <michal.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/18357
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Karol Latecki <karol.latecki@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>

show more ...

tests: Skip block devices marked as hidden

These devices don't come with their major:minor dev, hence they won't
pop up under /dev, i.e. are not really usable.

Change-Id: I49b39ccbedcdd1bfe37964819

tests: Skip block devices marked as hidden

These devices don't come with their major:minor dev, hence they won't
pop up under /dev, i.e. are not really usable.

Change-Id: I49b39ccbedcdd1bfe37964819e15b769af22cab6
Signed-off-by: Michal Berger <michal.berger@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/17774
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <smatsumoto@nvidia.com>

show more ...

scripts/setup: Stop checking for broken uio_pci_generic driver

This check was put in place as a workaround for some specific kernel
builds which were mainly shipped under centos8. Since we already
d

scripts/setup: Stop checking for broken uio_pci_generic driver

This check was put in place as a workaround for some specific kernel
builds which were mainly shipped under centos8. Since we already
dropped centos8 in the CI and replaced it with rocky8 (where kernel
is actually stable) there's no value in maintaining that anymore.

Signed-off-by: Michal Berger <michal.berger@intel.com>
Change-Id: Idd39b5bae888a991d7d628a19bc18acac6570f81
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/15819
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot

show more ...

test/nvmf: Make sure ctrl is fully released

There may be some timing issue where attempt to rebind the ctrl from
vfio_pci driver back to nvme, right after issuing the
bdev_nvme_detach_controller() c

test/nvmf: Make sure ctrl is fully released

There may be some timing issue where attempt to rebind the ctrl from
vfio_pci driver back to nvme, right after issuing the
bdev_nvme_detach_controller() call, fails as the vfio_pci might have
not fully released the device yet.

To mitigate, simply kill the application (as it's not needed anymore
at at point) before starting the kernel_target test - this should give
enough time for the device to be properly released.

As a precaution, make setup.sh to retry the probe attempt in case it
fails.

Signed-off-by: Michal Berger <michal.berger@intel.com>
Change-Id: Ifc4f4c18a90605154bf33b078575c8b41129f1f3
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/15767
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Kamil Godzwon <kamilx.godzwon@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot

show more ...

scripts/setup: Use driver_override to bind devices

Using new_id attribute is global in scope, meaning that depending
on the kernel's setup seen prior running setup.sh, single write
to it may re-bind

scripts/setup: Use driver_override to bind devices

Using new_id attribute is global in scope, meaning that depending
on the kernel's setup seen prior running setup.sh, single write
to it may re-bind ALL matching devices. This doesn't play well
with our PCI_{ALLOWED,BLOCKED} options as they can't be enforced
in such a case. Consider the following example:

> modprobe -r nvme # all nvme ctrls are detached from the kernel
> echo 0xdead 0xbeef >/sys/bus/pci/drivers/uio_pci_generic/new_id

# setup.sh-wise

> modprobe -r nvme
> PCI_BLOCKED=some:dead:beef.bdf setup.sh
# PCI_BLOCKED device still ends up bound to userspace driver.

After this single write, ALL matching devices will end up bound to
uio_pci_generic. To avoid this, we should override preferred driver
on per-bdf basis.

Signed-off-by: Michal Berger <michal.berger@intel.com>
Change-Id: Ic4613e33321303b92b47ce3f4d7e1f29ecca3036
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/13813
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Kamil Godzwon <kamilx.godzwon@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>

show more ...