Hi Harish,
> -Original Message-
> From: Naga Harish K, S V
> Sent: Thursday, December 1, 2022 11:38 AM
> To: Kundapura, Ganapati ; dev@dpdk.org;
> jer...@marvell.com; Gujjar, Abhinandan S
> Cc: Jayatheerthan, Jay
> Subject: RE: [PATCH v1 2/5] eventdev/crypto: wrong crypto enqueue count
When many ops are failed to enqueue to eventdev, crypto
adapter stores one event in buffer for later processing.
Add all failed ops to buffer for later processing.
Signed-off-by: Ganapati Kundapura
---
v2:
* Updated subject line commit message
diff --git a/lib/eventdev/rte_event_crypto_adapter.
Crypto adapter checks CPM backpressure once in enq_run()
This leads to buffer overflow if some ops failed to flush
to cryptodev.
Checked CPM backpressure for every iteration in enq_run()
Fixes: 7901eac3409a ("eventdev: add crypto adapter implementation")
Signed-off-by: Ganapati Kundapura
---
v2
Events enqueued to eventdev from the beginning of the circular buffer.
This leads to invalid or already freed events getting enqueued to eventdev
from the circular buffer.
Fixed by enqueuing the events to eventdev from the head pointer of
circular buffer.
Fixes: 7901eac3409a ("eventdev: add crypt
crypto_enq_count is updated on failure to enqueue ops to cryptodev.
Updated crypto_enq_count on successful enqueue of ops to cryptodev.
Fixes: 7901eac3409a ("eventdev: add crypto adapter implementation")
Signed-off-by: Ganapati Kundapura
---
v2:
* Updated subject line in commit message
diff --
In the current implementation adapter queries event device's capability for
implicit release support.
This information is used to decide whether events are enqueued back as
NEW or FWD events.
This patch updates the adapter to query the port caps for implicit release
to decide on events enqueuing
Hi Ganapti,
> -Original Message-
> From: Kundapura, Ganapati
> Sent: Wednesday, November 30, 2022 10:40 PM
> To: dev@dpdk.org; jer...@marvell.com; Naga Harish K, S V
> ; Gujjar, Abhinandan S
>
> Cc: Jayatheerthan, Jay
> Subject: [PATCH v1 2/5] eventdev/crypto: wrong crypto enqueue count
From: Satheesh Paul
Flow parsing was not handling pattern matching on the
fields of inner VLAN even though MKEX profile could be
extracting inner VLAN fields. Code has been modified
to handle matching fields on two VLAN tags.
Fixes: c34ea71b878 ("common/cnxk: add NPC parsing API")
Cc: sta...@dpd
From: Kiran Kumar K
For inline IPsec device, reserve number of rules specified by
``max_ipsec_rules`` and use them while installing rules with
action as security. Rule priority should be 0. If specified
number of rules not available, then only available number of
rules will be allocated and used.
From: Kiran Kumar K
Add a flag to check flow validate. In case of MSNS don't
need to configure the API to SA translation while validating
the flow. Skip this with a check.
Signed-off-by: Kiran Kumar K
Reviewed-by: Satheesh Paul
---
drivers/net/cnxk/cnxk_flow.c | 9 +
1 file changed, 5
From: Satheesh Paul
In case of IPsec, the inbound SPI can be random. HW supports mapping
SPI to an arbitrary SA index. SPI to SA index is done using a lookup
in NPC cam entry with key as SPI, MATCH_ID, LFID. Adding Mbox API
changes to configure the match table. And adding RTE FLow changes to
prog
From: Kiran Kumar K
While installing flow rule, if user provide item type
as RTE_FLOW_ITEM_TYPE_ETH, it should be applied to both
first and second pass. Adding changes to mask the ltype
to match both.
Signed-off-by: Kiran Kumar K
Reviewed-by: Satheesh Paul
---
drivers/common/cnxk/roc_npc_mcam
From: Kiran Kumar K
While configuring the RSS key, mask has been created to match multiple
ltypes for L4. Since ICMP is placed in between, it is also being
considered for RSS. So, moving the ICMP to end of enum.
Signed-off-by: Kiran Kumar K
Reviewed-by: Jerin Jacob
---
drivers/common/cnxk/hw/
https://bugs.dpdk.org/show_bug.cgi?id=1138
Bug ID: 1138
Summary: net/i40e all packets dropped after
rte_eth_dev_flow_ctrl_set
Product: DPDK
Version: 21.11
Hardware: All
OS: Linux
Status: UNCONFIR
The FIB library builds on Windows as long as sys/queue.h is defined.
Signed-off-by: Stephen Hemminger
---
app/test/test_fib.c | 22 +-
app/test/test_fib6.c | 24 ++--
app/test/test_fib6_perf.c | 15 +--
app/test/test_fib_perf.c | 19
Allow building on Windows, need to provide some compatability
wrappers for writev() and if_nametoindex.
Signed-off-by: Stephen Hemminger
---
lib/pcapng/meson.build | 6 -
lib/pcapng/rte_pcapng.c | 59 +++--
2 files changed, 57 insertions(+), 8 deletions(
The RIB library builds on Windows as long as sys/queue.h is included
Signed-off-by: Stephen Hemminger
---
app/test/test_rib.c | 19 ---
app/test/test_rib6.c | 24 ++--
lib/rib/meson.build | 6 --
lib/rib/rte_rib.c| 1 +
lib/rib/rte_rib6.c | 1 +
This build works on Windows if sys/queue.h is included.
Signed-off-by: Stephen Hemminger
---
app/test/test_ipfrag.c | 12
lib/ip_frag/ip_frag_common.h | 2 ++
lib/ip_frag/meson.build | 6 --
3 files changed, 2 insertions(+), 18 deletions(-)
diff --git a/app/test/te
This builds on Windows if sys/queue.h is included.
Signed-off-by: Stephen Hemminger
---
app/test/test_reorder.c | 11 ---
lib/reorder/meson.build | 6 --
lib/reorder/rte_reorder.c | 1 +
3 files changed, 1 insertion(+), 17 deletions(-)
diff --git a/app/test/test_reorder.c b/ap
This builds on Windows without changes.
Signed-off-by: Stephen Hemminger
---
app/test/test_lpm.c | 15 +--
app/test/test_lpm6.c | 12
app/test/test_lpm6_data.h | 3 ++-
app/test/test_lpm6_perf.c | 14 +-
app/test/test_lpm_perf.c | 39
This driver builds ok on windows, but need to modify the
test meson build to skip the eventdev test on Windows.
Signed-off-by: Stephen Hemminger
---
app/test/meson.build | 2 +-
drivers/net/ring/meson.build | 6 --
2 files changed, 1 insertion(+), 7 deletions(-)
diff --git a/app/tes
Builds fine with current code, no changes needed.
Signed-off-by: Stephen Hemminger
---
drivers/net/null/meson.build | 6 --
1 file changed, 6 deletions(-)
diff --git a/drivers/net/null/meson.build b/drivers/net/null/meson.build
index 4a483955a7a9..a51f8f5211b0 100644
--- a/drivers/net/null/
While diagnosing some Windows cross build errors;
noticed that lots of important DPDK libraries are not
being built on Windows.
This patchset enables build them.
Tested with MingW cross build.
Stephen Hemminger (8):
net/null: build null PMD on Windows
net/ring: build on Windows
lpm: enable
Add a new section 'Supported APIs', inculding the supported
APIs/items/actions of rte_flow.
Signed-off-by: Chaoyong He
Reviewed-by: Niklas Söderlund
---
doc/guides/nics/nfp.rst | 71 +
1 file changed, 71 insertions(+)
diff --git a/doc/guides/nics/nfp.rst
hi folks,
i'd like to continue work moving to our platform abstracted rte_thread
but ran into a hiccup. for some recent and not so recent apis it appears
they managed to slip in without ever being __experimental.
as a function of the dpdk project api/abi policy this means we can't
change or remo
On Wed, 30 Nov 2022 18:41:16 +
bugzi...@dpdk.org wrote:
> https://bugs.dpdk.org/show_bug.cgi?id=1137
>
> Bug ID: 1137
>Summary: CPU affinity set incorrectly when lcore_id 0 is not
> the master-lcore
>Product: DPDK
>Version: 2
https://bugs.dpdk.org/show_bug.cgi?id=1137
Bug ID: 1137
Summary: CPU affinity set incorrectly when lcore_id 0 is not
the master-lcore
Product: DPDK
Version: 22.11
Hardware: All
OS: Linux
Status:
When many ops are failed to enqueue to eventdev, crypto
adapter stores one event in buffer for later processing.
Add all failed ops to buffer for later processing.
Signed-off-by: Ganapati Kundapura
diff --git a/lib/eventdev/rte_event_crypto_adapter.c
b/lib/eventdev/rte_event_crypto_adapter.c
i
Crypto adapter checks CPM backpressure once in enq_run()
This leads to buffer overflow if some ops failed to flush
to cryptodev.
Checked CPM backpressure for every iteration in enq_run()
Signed-off-by: Ganapati Kundapura
diff --git a/lib/eventdev/rte_event_crypto_adapter.c
b/lib/eventdev/rte_e
Events enqueued to eventdev from the beginning of the circular buffer.
This leads to invalid or already freed events getting enqueued to
circular buffer.
Fixed by enqueuing the events from the head pointer of circular buffer.
Signed-off-by: Ganapati Kundapura
diff --git a/lib/eventdev/rte_event
crypto_enq_count is updated on failure to enqueue ops to cryptodev.
Updated crypto_enq_count on successful enqueue of ops to cryptodev.
Signed-off-by: Ganapati Kundapura
diff --git a/lib/eventdev/rte_event_crypto_adapter.c
b/lib/eventdev/rte_event_crypto_adapter.c
index 1c0a22b..ef3dbe9 100644
In the current implementation adapter queries event device's capability for
implicit release support.
This information is used to decide whether events are enqueued back as
NEW or FWD events.
This patch updates the adapter to query the port caps for implicit release
to decide on events enqueuing
On Wed, 30 Nov 2022 16:56:37 +0100
Maxime Coquelin wrote:
> + if (status != 0 || !dev->scvq)
> + goto out;
> +
> + status = virtio_send_command(&dev->scvq->cq, (struct virtio_pmd_ctrl
> *)hdr, dlen, nb_dlen);
> +out:
Maybe I am looking at the diff only, and not seeing someth
This patch removes the limitation of 8 queue pairs by
dynamically allocating vring metadata once we know the
maximum number of queue pairs supported by the backend.
This is especially useful for Vhost-vDPA with physical
devices, where the maximum queues supported may be much
more than 8 pairs.
Si
This patch advertizes control virtqueue support by the vDPA
backend if it supports VIRTIO_NET_F_CTRL_VQ.
Signed-off-by: Maxime Coquelin
---
drivers/net/virtio/virtio_user/vhost_vdpa.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/net/virtio/virtio_user/vhost_vdp
This patch implements control messages forwarding from the
regular control queue to the shadow control queue.
Signed-off-by: Maxime Coquelin
---
.../net/virtio/virtio_user/virtio_user_dev.c | 37 ---
.../net/virtio/virtio_user/virtio_user_dev.h | 3 --
drivers/net/virtio/virti
If the device fails to start, read the status from the
device and return early.
Fixes: 57912824615f ("net/virtio-user: support vhost status setting")
Cc: sta...@dpdk.org
Signed-off-by: Maxime Coquelin
---
drivers/net/virtio/virtio_user_ethdev.c | 11 ---
1 file changed, 8 insertions(+),
This patch introduces a new callback that is to be called
when the backend supports control virtqueue.
Implementation for Vhost-vDPA backend is added in this patch.
Signed-off-by: Maxime Coquelin
---
drivers/net/virtio/virtio_user/vhost.h | 1 +
drivers/net/virtio/virtio_user/vhost_v
When supported by the backend (only vDPA for now), this
patch gets the maximum number of queue pairs supported by
the device by querying it in its config space.
This is required for adding backend control queue support,
as is index equals the maximum number of queues supported
by the device as des
This patch adds sending the shadow control queue info
to the backend.
Signed-off-by: Maxime Coquelin
---
.../net/virtio/virtio_user/virtio_user_dev.c | 28 ---
1 file changed, 24 insertions(+), 4 deletions(-)
diff --git a/drivers/net/virtio/virtio_user/virtio_user_dev.c
b/driv
If the backends supports control virtqueue, allocate a
shadow control virtqueue, and implement the notify callback
that writes into the kickfd.
Signed-off-by: Maxime Coquelin
---
.../net/virtio/virtio_user/virtio_user_dev.c | 47 ++-
.../net/virtio/virtio_user/virtio_user_dev.h
The number of queue pairs is specified as a 16 bits
unsigned int in the Virtio specification.
Signed-off-by: Maxime Coquelin
---
drivers/net/virtio/virtio_user/virtio_user_dev.c | 2 +-
drivers/net/virtio/virtio_user/virtio_user_dev.h | 6 +++---
drivers/net/virtio/virtio_user_ethdev.c
The only reason two loops were needed to iterate over
queues at setup time was to be able to print whether it
was a Tx or Rx queue.
This patch changes queues iteration to a single loop.
Signed-off-by: Maxime Coquelin
---
drivers/net/virtio/virtio_user/virtio_user_dev.c | 16
1
This patch extracts the virtqueue initialization out of
the Virtio ethdev queue initialization, as preliminary
work to provide a way for Virtio-user to allocate its
shadow control virtqueue.
Signed-off-by: Maxime Coquelin
---
drivers/net/virtio/virtio_ethdev.c | 261 ++--
This patch only allocates the SW ring when vectorized
datapath is used. It also moves the SW ring and fake mbuf
in the virtnet_rx struct since this is Rx-only.
Signed-off-by: Maxime Coquelin
---
drivers/net/virtio/virtio_ethdev.c| 88 ---
drivers/net/virtio/virtio_rxt
Whatever its type (Rx, Tx or Ctl), all the virtqueue
require a memzone for the vrings. This patch moves its
pointer to the virtqueue struct, simplifying the code.
Signed-off-by: Maxime Coquelin
---
drivers/net/virtio/virtio_cvq.h| 1 -
drivers/net/virtio/virtio_ethdev.c | 11 ++-
dr
This patch refactors the indirect descriptors headers
initialization in a dedicated function, and makes it used
by both queue init and reset functions.
Signed-off-by: Maxime Coquelin
---
drivers/net/virtio/virtio_ethdev.c | 30 +
drivers/net/virtio/virtqueue.c | 68 ++
This patch removes the queue ID field in virtnet_rx struct.
Signed-off-by: Maxime Coquelin
---
drivers/net/virtio/virtio_rxtx.c | 1 -
drivers/net/virtio/virtio_rxtx.h | 2 --
2 files changed, 3 deletions(-)
diff --git a/drivers/net/virtio/virtio_rxtx.c b/drivers/net/virtio/virtio_rxtx.c
index
This patch removes the unused port ID information from
virtnet_ctl struct.
Signed-off-by: Maxime Coquelin
---
drivers/net/virtio/virtio_cvq.h | 1 -
1 file changed, 1 deletion(-)
diff --git a/drivers/net/virtio/virtio_cvq.h b/drivers/net/virtio/virtio_cvq.h
index 226561e6b8..0ff326b063 100644
-
The port and queue IDs are not used in virtnet_tx struct,
this patch removes them.
Signed-off-by: Maxime Coquelin
---
drivers/net/virtio/virtio_ethdev.c | 1 -
drivers/net/virtio/virtio_rxtx.c | 1 -
drivers/net/virtio/virtio_rxtx.h | 3 ---
3 files changed, 5 deletions(-)
diff --git a/driv
This patch refactors virtqueue initialization by moving
its headers allocation and deallocation in dedicated
function.
While at it, it renames the memzone metadata and address
pointers in the virtnet_tx and virtnet_ctl structures to
remove redundant virtio_net_ prefix.
Signed-off-by: Maxime Coque
The port ID information is duplicated in several places.
This patch removes it from the virtnet_rx struct as it can
be found in virtio_hw struct.
Signed-off-by: Maxime Coquelin
---
drivers/net/virtio/virtio_ethdev.c | 1 -
drivers/net/virtio/virtio_rxtx.c| 25 ++
This patch moves Virtio control queue code into a dedicated
file, as preliminary rework to support shadow control queue
in Virtio-user.
Signed-off-by: Maxime Coquelin
---
drivers/net/virtio/meson.build | 1 +
drivers/net/virtio/virtio_cvq.c| 230 +
drivers/n
This patch introduces a notification callback for the control
virtqueue as preliminary work to add shadow control virtqueue
support.
This new callback is required so that the shadow control queue
implemented in Virtio-user does not call the notifciation op
implemented for the driver layer.
Signed
This series introduces control queue support for Vhost-vDPA
backend. This is a requirement to support multiqueue, but
be usefull for other features like RSS for example.
Since the Virtio-user layer of the Virtio PMD must handle
some control messages, like the number of queue pairs to
be used by th
When many ops are failed to enqueue to eventdev, crypto
adapter stores one event in buffer for later processing.
Add all failed ops to buffer for later processing.
Signed-off-by: Ganapati Kundapura
diff --git a/lib/eventdev/rte_event_crypto_adapter.c
b/lib/eventdev/rte_event_crypto_adapter.c
i
Crypto adapter checks CPM backpressure once in enq_run()
This leads to buffer overflow if some ops failed to flush
to cryptodev.
Checked CPM backpressure for every iteration in enq_run()
Signed-off-by: Ganapati Kundapura
diff --git a/lib/eventdev/rte_event_crypto_adapter.c
b/lib/eventdev/rte_e
On Wed, Nov 30, 2022 at 04:02:26PM +0100, David Marchand wrote:
> On Wed, Nov 30, 2022 at 11:44 AM Ferruh Yigit wrote:
> >
> > On 11/30/2022 10:02 AM, David Marchand wrote:
> > > ld exports any global symbol by default if no version script is passed.
> > > As a consequence, the incriminated change
Events enqueued to eventdev from the beginning of the circular buffer.
This leads to invalid or already freed events getting enqueued to
circular buffer.
Fixed by enqueuing the events from the head pointer of circular buffer.
Signed-off-by: Ganapati Kundapura
diff --git a/lib/eventdev/rte_event
On Wed, Nov 30, 2022 at 4:02 PM David Marchand
wrote:
> On Wed, Nov 30, 2022 at 11:44 AM Ferruh Yigit wrote:
> >
> > On 11/30/2022 10:02 AM, David Marchand wrote:
> > > ld exports any global symbol by default if no version script is passed.
> > > As a consequence, the incriminated change let any
crypto_enq_count is updated on failure to enqueue ops to cryptodev.
Updated crypto_enq_count on successful enqueue of ops to cryptodev.
Signed-off-by: Ganapati Kundapura
diff --git a/lib/eventdev/rte_event_crypto_adapter.c
b/lib/eventdev/rte_event_crypto_adapter.c
index c293a62..e1a0367 100644
On Wed, Nov 30, 2022 at 11:44 AM Ferruh Yigit wrote:
>
> On 11/30/2022 10:02 AM, David Marchand wrote:
> > ld exports any global symbol by default if no version script is passed.
> > As a consequence, the incriminated change let any public symbol leak
> > out of the driver shared libraries.
> >
>
When many ops are failed to enqueue to eventdev, crypto
adapter stores one event in buffer for later processing.
Add all failed ops to buffer for later processing.
Signed-off-by: Ganapati Kundapura
Signed-off-by: Stania Bartosz
diff --git a/lib/eventdev/rte_event_crypto_adapter.c
b/lib/eventd
In the current implementation adapter queries event device's capability for
implicit release support.
This information is used to decide whether events are enqueued back as
NEW or FWD events.
This patch updates the adapter to query the port caps for implicit release
to decide on events enqueuing
On 9/1/2022 3:20 PM, Cristian Dumitrescu wrote:
> Remove the Ethernet device flow API support.
>
> Signed-off-by: Cristian Dumitrescu
> Signed-off-by: Yogesh Jangra
> ---
> drivers/net/softnic/meson.build|1 -
> drivers/net/softnic/rte_eth_softnic.c |9 -
> drivers/net/
On 11/30/2022 10:02 AM, David Marchand wrote:
> ld exports any global symbol by default if no version script is passed.
> As a consequence, the incriminated change let any public symbol leak
> out of the driver shared libraries.
>
> Hide again those symbols by providing a default map file which
>
> From: Kamalakshitha Aligeri [mailto:kamalakshitha.alig...@arm.com]
> Sent: Tuesday, 29 November 2022 21.54
>
> > From: Morten Brørup
> > Sent: Wednesday, November 16, 2022 12:04 PM
> >
> > Zero-copy access to mempool caches is beneficial for PMD performance,
> and
> > must be provided by the me
Add test cases to verify scatter-gather (SGL) buffers, with multiple
segments.
Signed-off-by: Volodymyr Fialko
Signed-off-by: Archana Muniganti
---
app/test/test_cryptodev.c| 49 ++---
app/test/test_cryptodev_security_ipsec.c | 93 +---
app/test/test_
ld exports any global symbol by default if no version script is passed.
As a consequence, the incriminated change let any public symbol leak
out of the driver shared libraries.
Hide again those symbols by providing a default map file which
unexports any global symbol using a local: * catch-all sta
On 11/30/2022 8:27 AM, David Marchand wrote:
> On Wed, Nov 30, 2022 at 8:13 AM David Marchand
> wrote:
>>> I assume this will cause warnings for ABI check scripts, how can we
>>> prevent the warnings?
>>
>> Indeed.
>>
>> Some options:
>> - add exhaustive suppression rules in devtools/libabigail.ab
Hi Tomasz,
On Tue, Nov 29, 2022 at 5:28 PM, Tomasz Duszynski wrote:
Add support for programming PMU counters and reading their values
in runtime bypassing kernel completely.
This is especially useful in cases where CPU cores are isolated
(nohz_full) i.e run dedicated tasks. In such cases one ca
On Wed, Nov 30, 2022 at 8:13 AM David Marchand
wrote:
> > I assume this will cause warnings for ABI check scripts, how can we
> > prevent the warnings?
>
> Indeed.
>
> Some options:
> - add exhaustive suppression rules in devtools/libabigail.abignore,
> - retag the v22.11 release with this fix, bu
72 matches
Mail list logo
