From: Kalesh AP
This series contains bnxt PMD bug fixes. Please apply.
Ajit Khaparde (4):
net/bnxt: fix ring teardown
net/bnxt: fix PAM4 mask setting
net/bnxt: fix pointer access
net/bnxt: check VF rep pointer before access
Kalesh AP (13):
net/bnxt: fix bnxt_dev_set_mc_addr_list_op
From: Kalesh AP
Fix to cache the multicast mac addresses added to the port
to the driver private structure memory. Use this cached mc list
to program the FW.
This fixes an issue where multicast packets reception is
successful only if the multicast mac address of the packets
is the first one in t
From: Kalesh AP
During reset recovery, driver is not restoring the multicast
mac addresses. Added code to restore them during reset recovery.
Fixes: b02f1573cd07 ("net/bnxt: restore MAC filters during reset recovery")
Cc: sta...@dpdk.org
Signed-off-by: Kalesh AP
Reviewed-by: Ajit Khaparde
Rev
From: Kalesh AP
When there are no active Rx queues(i.e when all queues have been stopped),
clear the RSS redirection table of the VNIC on Thor.
Fixes: 9b63c6fd70e3 ("net/bnxt: support Rx/Tx queue start/stop")
Cc: sta...@dpdk.org
Signed-off-by: Kalesh AP
Reviewed-by: Ajit Khaparde
Reviewed-by:
From: Kalesh AP
During reset recovery, driver is not restoring the VNIC rss hash key.
It's generating a new random hash key which results in unexpected
RSS behavior after recovery. Fixed this by storing the VNIC rss
configuration to a local struct and then applying the cached value
during the rec
From: Kalesh AP
During port start, driver calls bnxt_vlan_offload_set_op()
to program VLAN Filter/VLAN Strip setting to the HW. This
in turns add the vlan filters.
This results in a failure when bnxt_restore_filters() invokes
bnxt_restore_vlan_filters() during the recovery as the vlans are
alrea
From: Kalesh AP
The Maximum number of receive mac addr is hard coded to 128
in the ethdev library(RTE_ETH_NUM_RECEIVE_MAC_ADDR).
But the bnxt devices support more than 128 unicast MAC filters
which could result in a segfault while user tries to add more
than 128 unicast MAC addresses to the port.
From: Kalesh AP
During reset recovery, fixed to set the fast-path pointers
only if recovery succeeds.
Fixes: 720b55ad278e ("net/bnxt: fix crash caused by error recovery")
Cc: sta...@dpdk.org
Signed-off-by: Kalesh AP
Reviewed-by: Ajit Khaparde
Reviewed-by: Somnath Kotur
---
drivers/net/bnxt/
From: Kalesh AP
Make these messages more specific.
Signed-off-by: Kalesh AP
Reviewed-by: Ajit Khaparde
Reviewed-by: Somnath Kotur
---
drivers/net/bnxt/bnxt_ethdev.c | 6 --
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/drivers/net/bnxt/bnxt_ethdev.c b/drivers/net/bnxt/bnx
From: Kalesh AP
During the port start, driver ignores mark table allocation failure.
This could cause a segfault due to NULL pointer dereference in
bnxt_set_mark_in_mbuf(). Fix this by checking the pointer validity
before accessing it.
Fixes: b87abb2e55cb ("net/bnxt: support marking packet")
Cc:
From: Kalesh AP
This patch reverts the commit "5d47d06b2c83".
Revert this commit as it caused a regression on legacy chips.
On newer chips we use TruFlow based flow creation instead of
HWRM based flow creation.
Fixes: 5d47d06b2c83 ("net/bnxt: modify VNIC accounting")
Cc: sta...@dpdk.org
Signed-
From: Kalesh AP
The HWRM_FUNC_QCAPS response indicates the maximum number
of multicast filters that can be supported by this function
on the RX side.
Fixed to use this value instead of the hard coded value 16.
Fixes: d69851df12b2 ("net/bnxt: support multicast filter and set MAC addr")
Cc: sta..
From: Somnath Kotur
By accepting rte_eth_dev as input, bnxt_stop_rxtx() can be used
in multiple scenarios such as representor devices.
Signed-off-by: Somnath Kotur
Reviewed-by: Ajit Khaparde
Reviewed-by: Kalesh AP
---
drivers/net/bnxt/bnxt_cpr.c| 16
drivers/net/bnxt/bnx
From: Kalesh AP
When there is a change in the default VLAN of the VF,
FW sends the VF_CFG_CHANGE async event to the driver.
Upon receiving this async event, driver currently only queries
the FW using HWRM_FUNC_QCFG. But this is not enough.
Driver has to clean up the existing filter and recreate
From: Ajit Khaparde
Check for valid bit while scanning for hwrm_done completion
during ring teardown. Not checking the valid bit could cause a
ring overflow when we ring the doorbell.
Fixes: 4fb6ab3f866d ("net/bnxt: check flush status during ring free")
Cc: sta...@dpdk.org
Signed-off-by: Ajit K
From: Ajit Khaparde
Set the PAM4 mask setting only when PAM4 signaling is enabled.
Setting it otherwise seems to result in link negotiation.
Also auto_pam4_link_speeds has been renamed to
auto_pam4_link_speed_mask to reflect its real usage.
Fixes: c23f9ded0391 ("net/bnxt: support 200G PAM4 link"
From: Ajit Khaparde
Validate the pointers in rxq before accessing them.
Otherwise it can cause a segfault.
Fixes: 657c2a7f1dd4 ("net/bnxt: create aggregation rings when needed")
Cc: sta...@dpdk.org
Signed-off-by: Ajit Khaparde
Reviewed-by: Kalesh AP
Reviewed-by: Somnath Kotur
---
drivers/ne
From: Kalesh AP
In bnxt_dev_xstats_get_op(), the PMD is not zeroing the whole
buffer supplied by the application. This can end up passing
junk statistics values to the application when the FW does not
support extended stats on a function.
Fixed to call memset() with corrrect size.
Fixes: f55e12f
From: Ajit Khaparde
The PF or trusted VF Rx handler could invoke the VF representor's
Rx function without knowledge of the application cleaning up the
representor ports. Check if the vfr_bp pointer is valid before
accessing it.
Fixes: 6dc83230b43b ("net/bnxt: support port representor data path")
> -Original Message-
> From: Zhang, Qi Z
> Sent: 2022年1月2日 16:46
> To: Liu, KevinX ; dev@dpdk.org
> Cc: Zhang, RobinX ; Wang, Jie1X
> ; Liu, KevinX
> Subject: RE: [PATCH] net/ice: fix error forwarding IPv6 VXLAN packet
>
>
>
> > -Original Message-
> > From: Kevin Liu
> > Se
The tunnel packets is missing some information after Tx forwarding.
In ice_txd_enable_offload, when set tunnel packet Tx checksum
offload enable, td_offset should be set with outer l2/l3 len instead
of inner l2/l3 len.
In ice_txd_enable_checksum, td_offset should also be set with outer
l3 len.
T
> -Original Message-
> From: pbhagavat...@marvell.com
> Sent: Friday, November 5, 2021 6:11 PM
> To: Ruifeng Wang ; david.march...@redhat.com;
> jer...@marvell.com; Yipeng Wang ; Sameh
> Gobriel ; Bruce Richardson
> ; Vladimir Medvedkin
>
> Cc: dev@dpdk.org; Pavan Nikhilesh
> Subject: [d
From: Abhimanyu Saini
Buffer for MCDI channel is allocated using rte_memzone_reserve_aligned
with zone name 'mcdi'. Since multiple MCDI channels are needed to
support multiple VF(s) and rte_memzone_reserve_aligned expects unique
zone names, append PCI address to zone name to make it unique.
Sign
In crypto producer mode, producer core enqueues cryptodev with software
generated crypto ops and worker core dequeues crypto completion events
from the eventdev. Event crypto metadata used for above processing is
pre-populated in each crypto session.
Parameter --prod_type_cryptodev can be used to
Done my own testing as well, for the logs;
Canonical(R) Testing
* Build tests of DPDK & OVS 2.13.3 on Ubuntu 20.04 (meson based)
* Functional and performance tests based on OVS-DPDK on x86_64
* Autopkgtests for DPDK and OpenvSwitch
The former other use case of 19.11 was Ubuntu 20.10 but
https://bugs.dpdk.org/show_bug.cgi?id=900
Christian Ehrhardt (christian.ehrha...@canonical.com) changed:
What|Removed |Added
CC||christian.e
Move crypto ops to circular buffer to retain crypto
ops when cryptodev/eventdev are temporarily full
---
v2:
* reset cryptp adapter next cdev id before dequeueing from the
next cdev
---
Signed-off-by: Ganapati Kundapura
diff --git a/lib/eventdev/rte_event_crypto_adapter.c
b/lib/eventdev/rte_
update rte_event_crypto_adapter_caps_get() to return
SW_CAP if PMD callback is not registered.
Signed-off-by: Ganapati Kundapura
diff --git a/lib/eventdev/rte_eventdev.c b/lib/eventdev/rte_eventdev.c
index 79b9ea3..6988bf1 100644
--- a/lib/eventdev/rte_eventdev.c
+++ b/lib/eventdev/rte_eventdev.
update rte_event_crypto_adapter_caps_get() to return
SW_CAP if PMD callback is not registered.
Signed-off-by: Ganapati Kundapura
diff --git a/lib/eventdev/rte_eventdev.c b/lib/eventdev/rte_eventdev.c
index 79b9ea3..6988bf1 100644
--- a/lib/eventdev/rte_eventdev.c
+++ b/lib/eventdev/rte_eventdev.
Move crypto ops to circular buffer to retain crypto
ops when cryptodev/eventdev are temporarily full
---
v2:
* reset cryptp adapter next cdev id before dequeueing from the
next cdev
---
Signed-off-by: Ganapati Kundapura
diff --git a/lib/eventdev/rte_event_crypto_adapter.c
b/lib/eventdev/rte_
+Cc Ori Kam, rte_flow maintainer
29/12/2021 15:34, Ivan Malov:
> Hi all,
>
> In 'rte_flow.h', there is 'struct rte_flow_action_rss'. In it, 'queue' is
> to provide "Queue indices to use". But it is unclear whether the order of
> elements is meaningful or not. Does that matter? Can queue indices
04/01/2022 03:41, eagost...@nvidia.com:
> From: Elena Agostini
>
> Enable the possibility to make a GPU memory area accessible from
> the CPU.
>
> GPU memory has to be allocated via rte_gpu_mem_alloc().
>
> This patch allows the gpudev library to pin, through the GPU driver,
> a chunk of GPU me
> 04/01/2022 03:41, eagost...@nvidia.com:
> > From: Elena Agostini
> >
> > Enable the possibility to make a GPU memory area accessible from
> > the CPU.
> >
> > GPU memory has to be allocated via rte_gpu_mem_alloc().
> >
> > This patch allows the gpudev library to pin, through the GPU driver,
> >
https://bugs.dpdk.org/show_bug.cgi?id=901
Christian Ehrhardt (christian.ehrha...@canonical.com) changed:
What|Removed |Added
Resolution|--- |FIXED
https://bugs.dpdk.org/show_bug.cgi?id=902
Christian Ehrhardt (christian.ehrha...@canonical.com) changed:
What|Removed |Added
Status|IN_PROGRESS |RESOLVED
https://bugs.dpdk.org/show_bug.cgi?id=903
Christian Ehrhardt (christian.ehrha...@canonical.com) changed:
What|Removed |Added
Status|IN_PROGRESS |RESOLVED
https://bugs.dpdk.org/show_bug.cgi?id=904
Christian Ehrhardt (christian.ehrha...@canonical.com) changed:
What|Removed |Added
Resolution|--- |FIXED
https://bugs.dpdk.org/show_bug.cgi?id=905
Christian Ehrhardt (christian.ehrha...@canonical.com) changed:
What|Removed |Added
Status|IN_PROGRESS |RESOLVED
https://bugs.dpdk.org/show_bug.cgi?id=908
Christian Ehrhardt (christian.ehrha...@canonical.com) changed:
What|Removed |Added
Status|IN_PROGRESS |RESOLVED
On Thu, Dec 23, 2021 at 9:37 AM Maxime Coquelin
wrote:
>
> This patch adds IOTLB mempool name when logging debug
> or error messages, and also prepends the socket path.
> to all the logs.
>
> Signed-off-by: Maxime Coquelin
> ---
> lib/vhost/iotlb.c | 26 +++---
> lib/vho
On Thu, Dec 23, 2021 at 9:37 AM Maxime Coquelin
wrote:
>
> This patch adds the Vhost socket path whenever possible in
> order to make debugging possible when multiple Vhost
> devices are in use.
This patch could go after patch 5 so that we have a dev->ifname
available instead of fd.
>
> Signed-
On Thu, Dec 23, 2021 at 9:37 AM Maxime Coquelin
wrote:
>
> This patch prepends Vhost logs with the Vhost-user socket
> patch when available to ease filtering logs for a given port.
>
> Signed-off-by: Maxime Coquelin
> ---
> lib/vhost/vhost.c | 104 +++---
>
On Thu, Dec 23, 2021 at 9:37 AM Maxime Coquelin
wrote:
> @@ -471,16 +468,14 @@ vhost_user_client_reconnect(void *arg __rte_unused)
> sizeof(reconn->un));
> if (ret == -2) {
> close(reconn->fd);
On Thu, Dec 23, 2021 at 9:37 AM Maxime Coquelin
wrote:
>
> This series aims at easing Vhost logs analysis, by
> prepending the Vhost-user socket path to all logs and to
> remove multi-line comments. Doing so, filtering Vhost-user
> ports logs is much easier.
>
> Maxime Coquelin (7):
> vhost: imp
Hi
> -Original Message-
> From: Singh, Aman Deep
> Sent: Wednesday, December 15, 2021 11:34
> To: Li, Xiaoyun ; Yigit, Ferruh
> ; olivier.m...@6wind.com;
> m...@smartsharesystems.com; Ananyev, Konstantin
> ; step...@networkplumber.org;
> Medvedkin, Vladimir
> Cc: dev@dpdk.org
> Subject:
Shijith Thotton writes:
>>> >>
>>> >> In crypto producer mode, producer core enqueues cryptodev with
>>> >> software generated crypto ops and worker core dequeues crypto
>>> >> completion events from the eventdev. Event crypto metadata used for
>>> >> above processing is pre- populated in each cr
Hi
> -Original Message-
> From: Li, Xiaoyun
> Sent: Tuesday, January 4, 2022 15:19
> To: Singh, Aman Deep ; Yigit, Ferruh
> ; olivier.m...@6wind.com;
> m...@smartsharesystems.com; Ananyev, Konstantin
> ; step...@networkplumber.org;
> Medvedkin, Vladimir
> Cc: dev@dpdk.org
> Subject: RE:
>>
>> In crypto producer mode, producer core enqueues cryptodev with
>> software generated crypto ops and worker core dequeues crypto
>> completion events from the eventdev. Event crypto metadata used for
>> above processing is pre- populated in each crypto session.
>>
On Tue, 04 Jan 2022 13:41:55 +0100
Thomas Monjalon wrote:
> +Cc Ori Kam, rte_flow maintainer
>
> 29/12/2021 15:34, Ivan Malov:
> > Hi all,
> >
> > In 'rte_flow.h', there is 'struct rte_flow_action_rss'. In it, 'queue' is
> > to provide "Queue indices to use". But it is unclear whether the order
On 20/12/2021 17:05, Bruce Richardson wrote:
When the maximum burst size supported by HW is less than the available
ring space, incorrect capacity was returned when there was already some
jobs queued up for submission. This was because the capacity calculation
failed to subtract the number of a
What happens when the Nvidia GPU driver kernel callback occurs to invalidate
the pinned GPU memory region? Doesn't the NIC need to cease all DMA transfers
to/from that region before the kernel callback can complete?
From: Elena Agostini
Sent: 04 January 2022 13:55
To: NBU-Contact-Thomas Monjal
This specific feature is not about NIC activity, it’s about make GPU memory
visible and accessible to the CPU so there isn’t an asynchronous hidden
invalidation procedure.
That’s the reason for my comment to change the name from rte_gpu_mem_pin() to
rte_gpu_mem_to_cpu() (or similar).
The scenar
Hi Stephen,
On Tue, 4 Jan 2022, Stephen Hemminger wrote:
On Tue, 04 Jan 2022 13:41:55 +0100
Thomas Monjalon wrote:
+Cc Ori Kam, rte_flow maintainer
29/12/2021 15:34, Ivan Malov:
Hi all,
In 'rte_flow.h', there is 'struct rte_flow_action_rss'. In it, 'queue' is
to provide "Queue indices to
On Tue, 4 Jan 2022 21:29:14 +0300 (MSK)
Ivan Malov wrote:
> Hi Stephen,
>
> On Tue, 4 Jan 2022, Stephen Hemminger wrote:
>
> > On Tue, 04 Jan 2022 13:41:55 +0100
> > Thomas Monjalon wrote:
> >
> >> +Cc Ori Kam, rte_flow maintainer
> >>
> >> 29/12/2021 15:34, Ivan Malov:
> >>> Hi all,
> >>>
Tested-by: Jiayu Hu
> -Original Message-
> From: Kevin Laatz
> Sent: Wednesday, January 5, 2022 1:17 AM
> To: Richardson, Bruce ; dev@dpdk.org
> Cc: sta...@dpdk.org
> Subject: Re: [PATCH] dma/idxd: fix burst capacity calculation
>
>
> On 20/12/2021 17:05, Bruce Richardson wrote:
> > Wh
55 matches
Mail list logo
