From: Long Li
The deprecation notice has been sent in
7642e3b0e0 ("doc: announce vmbus API deprecation")
Remove those interfaces in this release.
Signed-off-by: Long Li
---
doc/guides/rel_notes/release_25_07.rst | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/doc/guides/
From: Long Li
The RDMA kernel driver may expose two MANA RDMA devices, for RC and RAW
QP types. The purpose is to support fast service mode at SOC. Depending
on kernel version, the probe may fail if DPDK picks up the wrong device.
Add check for vendor_part_id and RAW QP capability when probing t
From: Long Li
The RDMA kernel driver may expose two MANA RDMA devices, for RC and RAW
QP types. The purpose is to support fast service mode at SOC. Depending
on kernel version, the probe may fail if DPDK picks up the wrong device.
Add check for vendor_part_id and RAW QP capability when probing t
From: Long Li
The netvsc driver should add per-queue and rx_nombuf counters from VF.
Fixes: 4e9c73e96e83 ("net/netvsc: add Hyper-V network device")
Cc: sta...@dpdk.org
Signed-off-by: Long Li
---
Change log:
v2 is a resend, there is no change to v1
drivers/net/netvsc/hn_ethdev.c | 10 +
From: Long Li
Hyper-V uses 4k page size, regardless of the system page size used. Define
Hyper-V page size for use in drivers.
Fixes: 831dba47bd ("bus/vmbus: add Hyper-V virtual bus support")
Cc: sta...@dpdk.org
Signed-off-by: Long Li
---
drivers/bus/vmbus/bus_vmbus_driver.h | 4
1 file c
From: Long Li
The driver should always use Hyper-V page size for implementing RNDIS and
calculating PFN (Page Frame Number) for communicating with Hyper-V VSP.
It should not use system page size as it may be different to the Hyper-V
page size.
Fixes: 4e9c73e96e ("net/netvsc: add Hyper-V network
From: Long Li
The interrupt and monitor pages mapped from Hyper-V via kernel are always
4k in sizes. Use Hyper-V page size to map them.
Fixes: 831dba47bd ("bus/vmbus: add Hyper-V virtual bus support")
Cc: sta...@dpdk.org
Signed-off-by: Long Li
---
drivers/bus/vmbus/linux/vmbus_uio.c | 2 +-
d
From: Long Li
The ring buffer data region always starts at the system page boundary
after ring buffer head. The current code assumes the system page size is
4k. This is not always correct.
Fix this by using system page size for addressing ring buffer data.
Fixes: 831dba47bd ("bus/vmbus: add Hyp
From: Long Li
The Hyper-V uses 4k page size regardless of system page size.
This patchset fixes issues while running on systems with page sizes other than
4k.
Long Li (4):
bus/vmbus: Align ring buffer data region to system page boundary
bus/vmbus: Define Hyper-V page size
bus/vmbus: Use
From: Long Li
Introduce a function get the vmbus device from hn_data. For secondary
process, the vmbus device is in eth_dev's private region.
Signed-off-by: Long Li
---
drivers/net/netvsc/hn_nvs.c | 15 +++
drivers/net/netvsc/hn_nvs.h | 2 ++
2 files changed, 17 insertions(+)
dif
From: Long Li
Hyperv host may offer channels without monitor enabled. The max monitor
ID it supports is 128. Over those channels without monitor enabled,
Hyperv does not send or receive large amount of data traffic and almost all
the data traffic is going over the VF.
Change the code to not fail
From: Long Li
The secondary process can access its vmbus device through device private
region. Add and pass it on all call chains leading to vmbus code.
Signed-off-by: Long Li
---
Change log.
v5: rebased to the latest net-next
drivers/bus/vmbus/linux/vmbus_uio.c | 2 +-
drivers/bus/vmbus/priv
From: Long Li
Hyperv may expose VMBUS channels without monitoring enabled. In this case,
it programs almost all the data traffic to VF.
This patchset enabled vmbus/netvsc to use channels without monitoring
enabled.
This needs to change the APIs exposed by drivers/bus/vmbus. Because those
APIs a
From: Long Li
To prepare for supporting to set hyperv event from secondary process
when the channel has monitoring disable, introduce a private data
region for storing the vmbus device. The secondary process will get
access to its vmbus device in case it needs to signal the host.
Signed-off-by:
From: Long Li
For vmbus channels without monitoring support, use kernel UIO interface
to indicate packet through interrupt page and UIO file handle.
Signed-off-by: Long Li
---
Changes:
v4: replace RTE_ATOMIC(uint32_t) * with RTE_ATOMIC(uint32_t *) for declaring
variable
v6: revert the v4 versi
From: Long Li
Secondary process will get access to vmbus device and this UIO fd for
signaling hyperv host on channels without monitoring support.
Signed-off-by: Long Li
---
drivers/bus/vmbus/vmbus_common_uio.c | 9 ++---
1 file changed, 6 insertions(+), 3 deletions(-)
diff --git a/drivers
From: Long Li
Secondary process will get access to vmbus device and this UIO fd for
signaling hyperv host on channels without monitoring support.
Signed-off-by: Long Li
---
drivers/bus/vmbus/vmbus_common_uio.c | 9 ++---
1 file changed, 6 insertions(+), 3 deletions(-)
diff --git a/drivers
From: Long Li
To prepare for supporting to set hyperv event from secondary process
when the channel has monitoring disable, introduce a private data
region for storing the vmbus device. The secondary process will get
access to its vmbus device in case it needs to signal the host.
Signed-off-by:
From: Long Li
For vmbus channels without monitoring support, use kernel UIO interface
to indicate packet through interrupt page and UIO file handle.
Signed-off-by: Long Li
---
Changes:
v4: replace RTE_ATOMIC(uint32_t) * with RTE_ATOMIC(uint32_t *) for declaring
variable
drivers/bus/vmbus/vmb
From: Long Li
Hyperv host may offer channels without monitor enabled. The max monitor
ID it supports is 128. Over those channels without monitor enabled,
Hyperv does not send or receive large amount of data traffic and almost all
the data traffic is going over the VF.
Change the code to not fail
From: Long Li
Introduce a function get the vmbus device from hn_data. For secondary
process, the vmbus device is in eth_dev's private region.
Signed-off-by: Long Li
---
drivers/net/netvsc/hn_nvs.c | 15 +++
drivers/net/netvsc/hn_nvs.h | 2 ++
2 files changed, 17 insertions(+)
dif
From: Long Li
The secondary process can access its vmbus device through device private
region. Add and pass it on all call chains leading to vmbus code.
Signed-off-by: Long Li
---
Change log.
v5: rebased to the latest net-next
drivers/bus/vmbus/linux/vmbus_uio.c | 2 +-
drivers/bus/vmbus/pri
From: Long Li
Secondary process will get access to vmbus device and this UIO fd for
signaling hyperv host on channels without monitoring support.
Signed-off-by: Long Li
---
drivers/bus/vmbus/vmbus_common_uio.c | 9 ++---
1 file changed, 6 insertions(+), 3 deletions(-)
diff --git a/drivers
From: Long Li
Hyperv may expose VMBUS channels without monitoring enabled. In this case,
it programs almost all the data traffic to VF.
This patchset enabled vmbus/netvsc to use channels without monitoring
enabled.
This needs to change the APIs exposed by drivers/bus/vmbus. Because those
APIs a
From: Long Li
For vmbus channels without monitoring support, use kernel UIO interface
to indicate packet through interrupt page and UIO file handle.
Signed-off-by: Long Li
---
Changes:
v4: replace RTE_ATOMIC(uint32_t) * with RTE_ATOMIC(uint32_t *) for declaring
variable
drivers/bus/vmbus/vmb
From: Long Li
The secondary process can access its vmbus device through device private
region. Add and pass it on all call chains leading to vmbus code.
Signed-off-by: Long Li
---
drivers/bus/vmbus/linux/vmbus_uio.c | 2 +-
drivers/bus/vmbus/private.h | 2 +-
drivers/bus/vmbus/rte_bu
From: Long Li
Hyperv host may offer channels without monitor enabled. The max monitor
ID it supports is 128. Over those channels without monitor enabled,
Hyperv does not send or receive large amount of data traffic and almost all
the data traffic is going over the VF.
Change the code to not fail
From: Long Li
Introduce a function get the vmbus device from hn_data. For secondary
process, the vmbus device is in eth_dev's private region.
Signed-off-by: Long Li
---
drivers/net/netvsc/hn_nvs.c | 15 +++
drivers/net/netvsc/hn_nvs.h | 2 ++
2 files changed, 17 insertions(+)
dif
From: Long Li
To prepare for supporting to set hyperv event from secondary process
when the channel has monitoring disable, introduce a private data
region for storing the vmbus device. The secondary process will get
access to its vmbus device in case it needs to signal the host.
Signed-off-by:
From: Long Li
Hyperv may expose VMBUS channels without monitoring enabled. In this case,
it programs almost all the data traffic to VF.
This patchset enabled vmbus/netvsc to use channels without monitoring
enabled.
This needs to change the APIs exposed by drivers/bus/vmbus. Because those
APIs a
From: Long Li
Introduce a function get the vmbus device from hn_data. For secondary
process, the vmbus device is in eth_dev's private region.
Signed-off-by: Long Li
---
drivers/net/netvsc/hn_nvs.c | 15 +++
drivers/net/netvsc/hn_nvs.h | 2 ++
2 files changed, 17 insertions(+)
dif
From: Long Li
The secondary process can access its vmbus device through device private
region. Add and pass it on all call chains leading to vmbus code.
Signed-off-by: Long Li
---
drivers/bus/vmbus/linux/vmbus_uio.c | 2 +-
drivers/bus/vmbus/private.h | 2 +-
drivers/bus/vmbus/rte_bu
From: Long Li
Secondary process will get access to vmbus device and this UIO fd for
signaling hyperv host on channels without monitoring support.
Signed-off-by: Long Li
---
drivers/bus/vmbus/vmbus_common_uio.c | 9 ++---
1 file changed, 6 insertions(+), 3 deletions(-)
diff --git a/drivers
From: Long Li
Hyperv host may offer channels without monitor enabled. The max monitor
ID it supports is 128. Over those channels without monitor enabled,
Hyperv does not send or receive large amount of data traffic and almost all
the data traffic is going over the VF.
Change the code to not fail
From: Long Li
For vmbus channels without monitoring support, use kernel UIO interface
to indicate packet through interrupt page and UIO file handle.
Signed-off-by: Long Li
---
Change log:
v2: fixed compilation warnings
drivers/bus/vmbus/vmbus_channel.c | 21 ++---
1 file chang
From: Long Li
To prepare for supporting to set hyperv event from secondary process
when the channel has monitoring disable, introduce a private data
region for storing the vmbus device. The secondary process will get
access to its vmbus device in case it needs to signal the host.
Signed-off-by:
From: root
Hyperv may expose VMBUS channels without monitoring enabled. In this case,
it programs almost all the data traffic to VF.
This patchset enabled vmbus/netvsc to use channels without monitoring
enabled.
This needs to change the APIs exposed by drivers/bus/vmbus. Becuase those
APIs are
From: Long Li
All vmbus APIs are used internally by DPDK core and net/netvsc PMD.
It's not feasible or practical to use those APIs by the application.
Those APIs will become internal in DPDK 25.11.
Signed-off-by: Long Li
---
Change in
v2: add details on what APIs will be changed
v3: remove the
From: Long Li
All vmbus APIs are used internally by DPDK core and net/netvsc PMD.
It's not feasible or practical to use those APIs by the application.
Those APIs will be moved from "DPDK" to "Internal" in DPDK 25.11.
Signed-off-by: Long Li
---
Change in v2:
add details on what APIs will be chan
From: Long Li
All vmbus APIs are used internally by DPDK core and net/netvsc PMD.
It's not feasible or practical to use those APIs by the application.
Those APIs will be moved from "DPDK" to "Internal" in DPDK 25.11.
Signed-off-by: Long Li
---
doc/guides/rel_notes/deprecation.rst | 4
1 f
From: Long Li
For vmbus channels without monitoring support, use kernel UIO interface
to indicate packet through interrupt page and UIO file handle.
Signed-off-by: Long Li
---
Change log:
v2: fixed compilation warnings
drivers/bus/vmbus/vmbus_channel.c | 21 ++---
1 file chang
From: Long Li
Hyperv may expose VMBUS channels without monitoring enabled. In this case,
it programs almost all the data traffic to VF.
This patchset enabled vmbus/netvsc to use channels without monitoring
enabled.
Long Li (6):
net/netvsc: introduce private data for storing vmbus device for
From: Long Li
Hyperv host may offer channels without monitor enabled. The max monitor
ID it supports is 128. Over those channels without monitor enabled,
Hyperv does not send or receive large amount of data traffic and almost all
the data traffic is going over the VF.
Change the code to not fail
From: Long Li
The secondary process can access its vmbus device through device private
region. Add and pass it on all call chains leading to vmbus code.
Signed-off-by: Long Li
---
drivers/bus/vmbus/linux/vmbus_uio.c | 2 +-
drivers/bus/vmbus/private.h | 2 +-
drivers/bus/vmbus/rte_bu
From: Long Li
Introduce a function get the vmbus device from hn_data. For secondary
process, the vmbus device is in eth_dev's private region.
Signed-off-by: Long Li
---
drivers/net/netvsc/hn_nvs.c | 15 +++
drivers/net/netvsc/hn_nvs.h | 2 ++
2 files changed, 17 insertions(+)
dif
From: Long Li
Secondary process will get access to vmbus device and this UIO fd for
signaling hyperv host on channels without monitoring support.
Signed-off-by: Long Li
---
drivers/bus/vmbus/vmbus_common_uio.c | 9 ++---
1 file changed, 6 insertions(+), 3 deletions(-)
diff --git a/drivers
From: Long Li
To prepare for supporting to set hyperv event from secondary process
when the channel has monitoring disable, introduce a private data
region for storing the vmbus device. The secondary process will get
access to its vmbus device in case it needs to signal the host.
Signed-off-by:
From: Long Li
The pathname can be defined as name[MAX_PATH]. This makes the driver
compilable using MSVC.
Signed-off-by: Long Li
---
drivers/net/mana/mana.c | 7 ++-
1 file changed, 2 insertions(+), 5 deletions(-)
diff --git a/drivers/net/mana/mana.c b/drivers/net/mana/mana.c
index c37c4e
From: Long Li
Introduce a function get the vmbus device from hn_data. For secondary
process, the vmbus device is in eth_dev's private region.
Signed-off-by: Long Li
---
drivers/net/netvsc/hn_nvs.c | 15 +++
drivers/net/netvsc/hn_nvs.h | 2 ++
2 files changed, 17 insertions(+)
dif
From: Long Li
Hyperv may expose VMBUS channels without monitoring enabled. In this case,
it programs almost all the data traffic to VF.
This patchset enabled vmbus/netvsc to use channels without monitoring
enabled.
Long Li (6):
net/netvsc: introduce private data for storing vmbus device for
From: Long Li
Hyperv host may offer channels without monitor enabled. The max monitor
ID it supports is 128. Over those channels without monitor enabled,
Hyperv does not send or receive large amount of data traffic and almost all
the data traffic is going over the VF.
Change the code to not fail
From: Long Li
The secondary process can access its vmbus device through device private
region. Add and pass it on all call chains leading to vmbus code.
Signed-off-by: Long Li
---
drivers/bus/vmbus/linux/vmbus_uio.c | 2 +-
drivers/bus/vmbus/private.h | 2 +-
drivers/bus/vmbus/rte_bu
From: Long Li
For vmbus channels without monitoring support, use kernel UIO interface
to indicate packet through interrupt page and UIO file handle.
Signed-off-by: Long Li
---
drivers/bus/vmbus/vmbus_channel.c | 21 ++---
1 file changed, 18 insertions(+), 3 deletions(-)
diff -
From: Long Li
To prepare for supporting to set hyperv event from secondary process
when the channel has monitoring disable, introduce a private data
region for storing the vmbus device. The secondary process will get
access to its vmbus device in case it needs to signal the host.
Signed-off-by:
From: Long Li
Secondary process will get access to vmbus device and this UIO fd for
signaling hyperv host on channels without monitoring support.
Signed-off-by: Long Li
---
drivers/bus/vmbus/vmbus_common_uio.c | 9 ++---
1 file changed, 6 insertions(+), 3 deletions(-)
diff --git a/drivers
From: Long Li
The driver uses mana_shared_data for tracking usage count for primary
process. This is not correct as the mana_shared_data is allocated
by the primary and is meant to track usage of secondary process by the
primary process. And it creates a race condition when the device is
removed
From: Long Li
The netvsc driver should add per-queue and rx_nombuf counters from VF.
Fixes: 4e9c73e96e83 ("net/netvsc: add Hyper-V network device")
Cc: sta...@dpdk.org
Signed-off-by: Long Li
---
drivers/net/netvsc/hn_ethdev.c | 10 +-
1 file changed, 5 insertions(+), 5 deletions(-)
di
From: Long Li
The driver uses mana_shared_data for tracking usage count for primary
process. This is not correct as the mana_shared_data is allocated
by the primary and is meant to track usage of secondary process by the
primary process. And it creates a race condition when the device is
removed
From: Long Li
The current code has the wrong assumption that a PCI device can have only
one Ethernet device. This is not correct as a PCI device can be multi
functional and have multiple Ethernet devices.
Fix this by scanning all the devices under a PCI device.
Fixes: a2a23a794b3a ("net/netvsc:
From: Long Li
There is no recovery code path if netvsc failed to switch data path.
Log an error in this case for troubleshooting.
Signed-off-by: Long Li
---
drivers/net/netvsc/hn_vf.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/drivers/net/netvsc/hn_vf.c b/drivers
From: Long Li
An RTE device can have multiple Ethernet devices. On hot plug events, it
can't be removed until all its Ethernet devices have been removed.
Fixes: a2a23a794b3a ("net/netvsc: support VF device hot add/remove")
Cc: sta...@dpdk.org
Signed-off-by: Long Li
---
drivers/net/netvsc/hn_vf
From: Long Li
On some systems, one PCI device may have muiltple NET devices. This
patchset fixes issues on those systems when such PCI device is doing hot
add/remove.
Change in v2: added cover letter, drop the 4th patch that will be implemented
in EAL instead for device args caching
Long Li (
From: Long Li
The driver uses mana_shared_data for tracking usage count for primary
process. This is not correct as the mana_shared_data is allocated
by the primary and is meant to track usage of secondary process by the
primary process. And it creates a race condition when the device is
removed
From: Long Li
The current code has the wrong assumption that a PCI device can have only
one Ethernet device. This is not correct as a PCI device can be multi
functional and have multiple Ethernet devices.
Fix this by scanning all the devices under a PCI device.
Fixes: a2a23a794b3a ("net/netvsc:
From: Long Li
There is no recovery code path if netvsc failed to switch data path.
Log an error in this case for troubleshooting.
Signed-off-by: Long Li
---
drivers/net/netvsc/hn_vf.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/drivers/net/netvsc/hn_vf.c b/drivers
From: Long Li
If a device is hot removed and hot plugged, it needs the same driver
parameters that are passed to EAL. However, during device removal, all
EAL driver parameters are freed as part of the cleanup.
Cache those driver parameters for future hot plug events. Because we don't
know which
From: Long Li
An RTE device can have multiple Ethernet devices. On hot plug events, it
can't be removed until all its Ethernet devices have been removed.
Fixes: a2a23a794b3a ("net/netvsc: support VF device hot add/remove")
Cc: sta...@dpdk.org
Signed-off-by: Long Li
---
drivers/net/netvsc/hn_vf
From: Long Li
The VSP assumes the packet doesn't have VLAN tags. When VLAN tag is
present in a TX packet, always strip it and use PPI to send VLAN info
through VSP packet.
Fixes: 4e9c73e96e83 ("net/netvsc: add Hyper-V network device")
Cc: Hui (Hui) Ling
Cc: Stephen Hemminger
Cc: Wei Hu
Cc: st
From: Alan Elder
The previous code allowed the number of Tx queues to be set higher than
the number of Rx queues. If a packet was sent on a Tx queue with index
>= number Rx queues there was a segfault due to accessing beyond the end
of the dev->data->rx_queues[] array.
This commit fixes the iss
From: Long Li
The driver has been verified on Linux arm64. Enable this build option and
add a missing header file for arm64.
Signed-off-by: Long Li
---
drivers/net/mana/meson.build | 4 ++--
drivers/net/mana/mp.c| 1 +
2 files changed, 3 insertions(+), 2 deletions(-)
diff --git a/driv
From: Malcolm Bumgardner
In the device event unregister code, it unconditionally remove all
callbacks which are registered with device_name set to NULL. This results
in many callbacks uncorrectly removed.
Fix this by only removing callbacks with matching cb_fn and cb_arg.
Signed-off-by: Malcolm
From: Stephen Hemminger
The current code uses unnecessary locking to set VF MTU, resulting in
deadlock on hot add/remove path. Fix this by using rte_eth_dev_set_mtu()
to set VF MTU.
Signed-off-by: Stephen Hemminger
Fixes: 45c83603087e ("net/netvsc: support MTU set")
Cc: sta...@dpdk.org
Signed-
From: Long Li
On MR cache expension failure, the request should fail as there is no path
to get a new MR into the tree. Attempting to insert a new MR to the cache
tree will result in memory violation.
Fixes: 0f5db3c68ba7 ("net/mana: implement memory registration")
Cc: sta...@dpdk.org
Signed-off-
From: Long Li
Use a MR on the stack instead of allocating it. This fixes the memory
leak in the code where a MR is allocated but never freed.
Fixes: 0f5db3c68ba7 ("net/mana: implement memory registration")
Cc: sta...@dpdk.org
Signed-off-by: Long Li
---
Change in v2:
change commit message to in
From: Long Li
Instead of allocating mbufs one by one during RX, use
rte_pktmbuf_alloc_bulk() to allocate them in a batch.
With this patch, there are no measurable performance improvements in
benchmarks. However, this patch should improve CPU cycles and reduce
potential locking conflicts in real-
From: Long Li
Instead of allocating mbufs one by one during RX, use
rte_pktmbuf_alloc_bulk() to allocate them in a batch.
With this patch, there are no measurable performance improvements in
benchmarks. However, this patch should improve CPU cycles and reduce
potential locking conflicts in real-
From: Long Li
Instead of allocating mbufs one by one during RX, use
rte_pktmbuf_alloc_bulk() to allocate them in a batch.
There are no measurable performance improvements in benchmarks. However,
this patch should improve CPU cycles and reduce potential locking
conflicts in real-world application
From: Long Li
Secondary processes are started after primary, and in most cases with
the device already started. Make them being able to process packets as
soon as they start.
This also works with the case where the primary process decides to start
the device at a later time after secondary proce
From: Long Li
On MR cache expension failure, the request should fail as there is no path
to get a new MR into the tree. Attempting to insert a new MR to the cache
tree will result in memory violation.
Signed-off-by: Long Li
---
drivers/net/mana/mana.h | 6 +++---
drivers/net/mana/mr.c | 45
From: Long Li
The content of the MR is copied to the cache trees, it's not necessary to
allocate a MR to do this. Use a variable on the stack instead.
This also fixes the memory leak in the code where a MR is allocated but
never freed.
Signed-off-by: Long Li
---
drivers/net/mana/mr.c | 15 +++
From: Long Li
Instead of allocating mbufs one by one during RX, use rte_pktmbuf_alloc_bulk()
to allocate them in a batch.
Signed-off-by: Long Li
---
Change in v2:
use rte_calloc_socket() in place of rte_calloc()
drivers/net/mana/rx.c | 68 ---
1 file ch
From: Long Li
Instead of allocating mbufs one by one during RX, use
rte_pktmbuf_alloc_bulk() to allocate them in a batch.
Signed-off-by: Long Li
---
drivers/net/mana/rx.c | 67 +++
1 file changed, 43 insertions(+), 24 deletions(-)
diff --git a/drivers/n
From: Long Li
The function name mana_find_pmd_mr() is misleading as there might be
allocations to get a MR. Change function name to mana_alloc_pmd_mr().
Signed-off-by: Long Li
---
drivers/net/mana/mana.h | 6 +++---
drivers/net/mana/mr.c | 7 ---
drivers/net/mana/rx.c | 2 +-
drivers/n
From: Long Li
The device capabilities reported from RDMA layer are in int. Those values can
overflow with the data types defined in dev_info_get().
Fix this by doing a upper bound before returning those values.
Fixes: 517ed6e2d590 ("net/mana: add basic driver with build environment")
Cc: sta...
From: Long Li
Add a missing "\n" to DP logs. This makes the logs correctly formatted.
Signed-off-by: Long Li
---
drivers/net/mana/mana.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/mana/mana.h b/drivers/net/mana/mana.h
index 9bff90ab77..6836872dc2 100644
---
From: Long Li
Set the queue state when queue is started/stopped.
Signed-off-by: Long Li
Acked-by: Stephen Hemminger
---
Change since v1:
Fixed the following building error
error: 'for' loop initial declarations are only allowed in C99 mode
drivers/net/netvsc/hn_ethdev.c | 17 +++-
From: Long Li
The current timeout is set to 5 seconds. In Azure, tests show that it may
take up to 15 seconds for VSP to respond on busy nodes. The VSP schedules
unbounded work to process VSC resquest, there is no upper limit on how long
it takes to send response back to VSC.
In the NETVSC kerne
From: Long Li
Use %m through fscanf to allocate mac dynamically is safer than using
mac[20], this guarantees there is no overflow on mac[].
Signed-off-by: Long Li
---
drivers/net/mana/mana.c | 6 --
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/drivers/net/mana/mana.c b/dri
From: Long Li
Add support for updating MTU for MANA. MTU is updated in kernel through
socket interface.
Signed-off-by: Long Li
---
doc/guides/nics/features/mana.ini | 1 +
drivers/net/mana/mana.c | 96 ++-
drivers/net/mana/mana.h | 2 +-
3 file
From: Long Li
Bytes should be calculated using the packet length, not the mbuf segment
data length.
Fixes: 517ed6e2d590 ("net/mana: add basic driver with build environment")
Cc: sta...@dpdk.org
Signed-off-by: Long Li
---
drivers/net/mana/tx.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(
From: Long Li
hn_vf_reset() is defined but not used. Remove it.
Signed-off-by: Long Li
---
drivers/net/netvsc/hn_var.h | 1 -
drivers/net/netvsc/hn_vf.c | 5 -
2 files changed, 6 deletions(-)
diff --git a/drivers/net/netvsc/hn_var.h b/drivers/net/netvsc/hn_var.h
index 98a3b83033..e1f8e69
From: Long Li
The hardware specification specifies that WQE_COUNT should set to 0 for
the Receive Queue. Although currently the hardware doesn't enforce the
check, in the future releases it may check on this value.
Signed-off-by: Long Li
---
Change log:
v2: added "Signed-off-by"
drivers/net/m
From: Long Li
This counter should set to uint16_t, the same type as pkt_received.
Otherwise, it may overflow when pkt_received goes over 256.
Thanks Xinhao Kong for debugging this.
Fixes: 517ed6e2d590 ("net/mana: add basic driver with build environment")
Cc: sta...@dpdk.org
Signed-off-by: Long
From: Long Li
The hardware specification specifies that WQE_COUNT should set to 0 for
the Receive Queue. Although currently the hardware doesn't enforce the
check, in the future releases it may check on this value.
---
drivers/net/mana/mana.h | 2 +-
drivers/net/mana/rx.c | 11 +++
2
From: Long Li
This counter should set to uint16_t, the same type as pkt_received.
Otherwise, it may overflow when pkt_received goes over 256.
Thanks Xinhao Kong for debugging.
Fixes: 517ed6e2d590 ("net/mana: add basic driver with build environment")
Cc: sta...@dpdk.org
---
drivers/net/mana/rx
From: Long Li
Set the queue state when queue is started/stopped.
Signed-off-by: Long Li
---
drivers/net/netvsc/hn_ethdev.c | 15 ++-
1 file changed, 14 insertions(+), 1 deletion(-)
diff --git a/drivers/net/netvsc/hn_ethdev.c b/drivers/net/netvsc/hn_ethdev.c
index d0bbc0a4c0..0c15f
From: Long Li
Set the queue state when queue is started/stopped
Signed-off-by: Long Li
---
drivers/net/mana/rx.c | 15 +++
drivers/net/mana/tx.c | 13 +
2 files changed, 28 insertions(+)
diff --git a/drivers/net/mana/rx.c b/drivers/net/mana/rx.c
index 220b372b15..1047a
From: Long Li
On a fatal CQE error when coalescing is used, update the correct index
and allow proceeding to the next CQE.
Fixes: 3409e0f172f6 ("net/mana: implement Rx CQE coalescing")
Signed-off-by: Long Li
---
drivers/net/mana/rx.c | 18 --
1 file changed, 8 insertions(+), 10
From: Stephen Hemminger
Found by cppcheck.
drivers/net/netvsc/hn_rndis.c:332:21: warning: Found calculation inside
sizeof(). [sizeofCalculation] if (len < sizeof(3 * sizeof(uint32_t))) {
Fixes: 4e9c73e96e83 ("net/netvsc: add Hyper-V network device")
Signed-off-by: Stephen Hemminger
Signed-of
From: Long Li
With RX coalescing, one CQE entry can be used to indicate up to 4 packets
on the receive queue. This saves processing time and PCI bandwidth over
the CQ.
Cc: sta...@dpdk.org
Signed-off-by: Long Li
---
drivers/net/mana/mana.h | 4 +++
drivers/net/mana/rx.c | 71
1 - 100 of 320 matches
Mail list logo
