Refill all consumed descriptors instead of refilling a set amount of
descriptors everytime. Also, lower the rx-free-thresh to refill more
aggressively. This yields in fewer packet drops for DQO queue format.
Signed-off-by: Rushil Gupta
Reviewed-by: Joshua Washington
---
drivers/net/gve
Acked-by: Rushil Gupta
Thanks!
On Fri, Aug 9, 2024 at 11:49 AM Tathagat Priyadarshi
wrote:
>
> The patch intends on adding support for TSO in DQO RDA format.
>
> Signed-off-by: Tathagat Priyadarshi
> Signed-off-by: Varun Lakkur Ambaji Rao
> ---
> drivers/net/
Makes sense. The virtual device only does L4 checksum.
Acked-by: Rushil Gupta
On Thu, Feb 16, 2023 at 10:58 AM Levend Sayar wrote:
> Google Virtual NIC is not doing IPv4 checksummimg.
> Removed that capability from PMD.
>
> Signed-off-by: Levend Sayar
> ---
> drivers/ne
acked-by: Rushil Gupta
On Thu, May 4, 2023 at 7:19 PM Junfeng Guo wrote:
>
> Add maintainers from Google for GVE.
>
> Signed-off-by: Junfeng Guo
> Signed-off-by: Jeroen de Borst
> Signed-off-by: Rushil Gupta
> Signed-off-by: Joshua Washington
> ---
> MAINTAINER
Thanks for reviewing Ferruh Yigit! I wanted to answer your queries
before sending the next patch.
Please find answers below:
On Thu, May 4, 2023 at 8:06 AM Ferruh Yigit wrote:
>
> On 4/26/2023 10:37 PM, Rushil Gupta wrote:
> > Change gve_driver_info fields to report DPDK as OS type
Change gve_driver_info fields to report DPDK as OS type and DPDK RTE
version as OS version, reserving driver_version fields for GVE driver
version based on features.
Signed-off-by: Rushil Gupta
Signed-off-by: Joshua Washington
Signed-off-by: Junfeng Guo
Signed-off-by: Jeroen de Borst
Change gve_driver_info fields to report DPDK as OS type and DPDK RTE
version as OS version, reserving driver_version fields for GVE driver
version based on features.
Signed-off-by: Rushil Gupta
Signed-off-by: Joshua Washington
Signed-off-by: Junfeng Guo
Signed-off-by: Jeroen de Borst
e.com/
On Mon, May 8, 2023 at 9:00 AM Ferruh Yigit wrote:
>
> On 5/8/2023 7:23 AM, Rushil Gupta wrote:
> >>> +#ifndef _GVE_VERSION_H_
> >>> +#define _GVE_VERSION_H_
> >>> +
> >>> +#include
> >>> +
> >>> +#define GVE_VERSION
tested-by: Rushil Gupta
On Mon, May 8, 2023 at 8:07 PM Junfeng Guo wrote:
> Add support for queue operations for GQI:
> - gve_rx_queue_start
> - gve_tx_queue_start
> - gve_rx_queue_stop
> - gve_tx_queue_stop
>
> Add support for queue operations for DQO:
>
Change gve_driver_info fields to report DPDK as OS type and DPDK RTE
version as OS version, reserving driver_version fields for GVE driver
version based on features.
Signed-off-by: Rushil Gupta
Signed-off-by: Joshua Washington
Signed-off-by: Junfeng Guo
Signed-off-by: Jeroen de Borst
Thanks for reviewing. Here is v3:
http://patchwork.dpdk.org/project/dpdk/patch/20230518174005.1377467-1-rush...@google.com/
On Wed, May 17, 2023 at 9:59 AM Ferruh Yigit wrote:
> On 5/8/2023 8:15 PM, Rushil Gupta wrote:
> > Change gve_driver_info fields to report DPDK as OS type and
Change gve_driver_info fields to report DPDK as OS type and DPDK RTE
version as OS version, reserving driver_version fields for GVE driver
version based on features.
Signed-off-by: Rushil Gupta
Signed-off-by: Joshua Washington
Signed-off-by: Junfeng Guo
Signed-off-by: Jeroen de Borst
Change gve_driver_info fields to report DPDK as OS type and DPDK RTE
version as OS version, reserving driver_version fields for GVE driver
version based on features.
Signed-off-by: Rushil Gupta
Signed-off-by: Joshua Washington
Signed-off-by: Junfeng Guo
Signed-off-by: Jeroen de Borst
Change gve_driver_info fields to report DPDK as OS type and DPDK RTE
version as OS version, reserving driver_version fields for GVE driver
version based on features.
Signed-off-by: Rushil Gupta
Signed-off-by: Joshua Washington
Signed-off-by: Junfeng Guo
Signed-off-by: Jeroen de Borst
resource for :00:04.0
...
EAL: Error - exiting with code: 1
This is probably because the adminq command is sharing memory to
report driver-info to the gvnic device and that needs to be in dpdk
memory.
On Fri, May 19, 2023 at 12:26 AM Rushil Gupta wrote:
>
> Change gve_driver_info fields to
I agree. Other adminq commands like gve_adminq_report_link_speed use
gve_alloc_dma_mem and gve_free_dma_mem (which use
rte_memzone_reserve_aligned under the hood)
so let's stick to that.
On Fri, May 19, 2023 at 3:04 AM Ferruh Yigit wrote:
>
> On 5/19/2023 8:41 AM, Rushil Gupta wr
Change gve_driver_info fields to report DPDK as OS type and DPDK RTE
version as OS version, reserving driver_version fields for GVE driver
version based on features.
Signed-off-by: Rushil Gupta
Signed-off-by: Joshua Washington
Signed-off-by: Junfeng Guo
Signed-off-by: Jeroen de Borst
Change gve_driver_info fields to report DPDK as OS type and DPDK RTE
version as OS version, reserving driver_version fields for GVE driver
version based on features.
Signed-off-by: Rushil Gupta
Signed-off-by: Joshua Washington
Signed-off-by: Junfeng Guo
Signed-off-by: Jeroen de Borst
We are not validating anything.
This is for our internal analysis and product requirements.
On Fri, May 19, 2023 at 1:56 PM Stephen Hemminger <
step...@networkplumber.org> wrote:
> On Fri, 19 May 2023 13:46:18 -0700
> Rushil Gupta wrote:
>
> > +#include
> >
&
tchwork.dpdk.org/project/dpdk/patch/20230519204618.1507956-1-rush...@google.com/
On Mon, May 22, 2023 at 1:52 AM Ferruh Yigit wrote:
> On 5/19/2023 9:46 PM, Rushil Gupta wrote:
> > +static int
> > +gve_verify_driver_compatibility(struct gve_priv *priv)
> > +{
> > +
Change gve_driver_info fields to report DPDK as OS type and DPDK RTE
version as OS version, reserving driver_version fields for GVE driver
version based on features.
Signed-off-by: Rushil Gupta
Signed-off-by: Joshua Washington
Signed-off-by: Junfeng Guo
Signed-off-by: Jeroen de Borst
the new patch:
http://patchwork.dpdk.org/project/dpdk/patch/20230524171324.2072742-1-rush...@google.com/
On Tue, May 23, 2023 at 3:22 AM Ferruh Yigit wrote:
>
> On 5/19/2023 9:46 PM, Rushil Gupta wrote:
> > diff --git a/drivers/net/gve/base/gve_osdep.h
> > b/drivers/net
http://patchwork.dpdk.org/project/dpdk/patch/20230524171324.2072742-1-rush...@google.com/
Does this look good to you?
On Wed, May 24, 2023 at 10:14 AM Rushil Gupta wrote:
>
> As noted from spec: "Some commands require an additional dma memory
> region to be passed to the device&quo
gVNIC requires physical address to be passed in the adminq command.
This was initially rightly pointed by ferruh.yigit@.
Fixed by passing 'driver_info_mem->iova'.
Signed-off-by: Rushil Gupta
---
drivers/net/gve/gve_ethdev.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
gVNIC requires physical address to be passed in the adminq command.
This was initially rightly pointed by ferruh.yigit@.
Fixed by passing 'driver_info_mem->iova'.
Signed-off-by: Rushil Gupta
---
drivers/net/gve/gve_ethdev.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
Thanks for the quick response!
Just for my own knowledge, what Junfeng described is the process to
fix the bug if a bug is present in the main dpdk repo?
On Thu, Jun 1, 2023 at 3:24 AM Ferruh Yigit wrote:
>
> On 6/1/2023 9:26 AM, Ferruh Yigit wrote:
> > On 6/1/2023 5:49 AM, Rushil
Thanks for the clarification and for squashing the bug into the main commit.
I just tested the net-next branch and it seems to log the right
message on the device.
On Fri, Jun 2, 2023 at 7:49 AM Ferruh Yigit wrote:
>
> On 6/1/2023 5:32 PM, Rushil Gupta wrote:
> > Thanks for the qu
net/gve: fix completion path for chained segments
Dqo Tx path had a bug where driver was overwriting mbufs in sw-ring.
We fixed this bug by cleaning slots for all segments.
Fixes: 4022f9 ("net/gve: support basic Tx data path for DQO")
Cc: sta...@dpdk.org
Signed-off-by: Rushil Gupta
R
Dqo Tx path had a bug where driver was overwriting mbufs in sw-ring.
We fixed this bug by cleaning slots for all segments.
Fixes: 4022f9 ("net/gve: support basic Tx data path for DQO")
Cc: sta...@dpdk.org
Signed-off-by: Rushil Gupta
Reviewed-by: Joshua Washington
---
drive
This was causing failure for testpmd runs (for queues >=15)
presumably due to flooding of logs due to descriptor ring being
overwritten.
Fixes: a01854 ("net/gve: fix dqo bug for chained descriptors")
Cc: sta...@dpdk.org
Signed-off-by: Rushil Gupta
Reviewed-by: Joshua Washington
> On Mon, 19 Feb 2024 02:44:35 +
> Rushil Gupta wrote:
>
> > This was causing failure for testpmd runs (for queues >=15)
> > presumably due to flooding of logs due to descriptor ring being
> > overwritten.
> >
> > Fixes: a01854 ("net/g
, 3:03 PM Ferruh Yigit wrote:
> On 2/19/2024 2:44 AM, Rushil Gupta wrote:
> > This was causing failure for testpmd runs (for queues >=15)
> > presumably due to flooding of logs due to descriptor ring being
> > overwritten.
> >
> > Fixes: a01854 ("net
Read from shared region to retrieve imissed statistics for GQ.
Change-Id: I8123d70ec5019b12f368e67b858c382bb7812e5c
---
dpdk/base/gve_adminq.h | 11 ++
dpdk/gve_ethdev.c | 84 ++
dpdk/gve_ethdev.h | 6 +++
3 files changed, 101 insertions(+)
Read from shared region to retrieve imissed statistics for GQ.
Tested using `show port xstats ` in interactive mode.
This metric can be triggered by using queues > cores.
Signed-off-by: Rushil Gupta
Reviewed-by: Joshua Washington
---
drivers/net/gve/base/gve_adminq.h | 11
drivers/net/
Read from shared region to retrieve imissed statistics for GQ.
Tested using `show port xstats ` in interactive mode.
This metric can be triggered by using queues > cores.
Signed-off-by: Rushil Gupta
Reviewed-by: Joshua Washington
---
drivers/net/gve/base/gve_adminq.h | 11
drivers/net/
Read from shared region to retrieve imissed statistics for GQ.
Tested using `show port xstats ` in interactive mode.
This metric can be triggered by using queues > cores.
Signed-off-by: Rushil Gupta
Reviewed-by: Joshua Washington
---
drivers/net/gve/base/gve_adminq.h | 11 +++
drivers/net/
Read from shared region to retrieve imissed statistics for GQ from device.
Tested using `show port xstats ` in interactive mode.
This metric can be triggered by using queues > cores.
---
drivers/net/gve/base/gve_adminq.h | 11 +++
drivers/net/gve/gve_ethdev.c | 117 ++
Read from shared region to retrieve imissed statistics for GQ from device.
Tested using `show port xstats ` in interactive mode.
This metric can be triggered by using queues > cores.
Signed-off-by: Rushil Gupta
Reviewed-by: Joshua Washington
---
drivers/net/gve/base/gve_adminq.h |
Read from shared region to retrieve imissed statistics for GQ from device.
Tested using `show port xstats ` in interactive mode.
This metric can be triggered by using queues > cores.
Signed-off-by: Rushil Gupta
Reviewed-by: Joshua Washington
---
drivers/net/gve/base/gve_adminq.h |
On Fri, Jan 12, 2024 at 8:36 PM Ferruh Yigit wrote:
> On 12/22/2023 3:39 PM, Rushil Gupta wrote:
> > Read from shared region to retrieve imissed statistics for GQ from
> device.
> > Tested using `show port xstats ` in interactive mode.
> > This metric can be triggered
Read from shared region to retrieve imissed statistics for GQ from device.
Tested using `show port xstats ` in interactive mode.
Signed-off-by: Rushil Gupta
Reviewed-by: Joshua Washington
---
drivers/net/gve/base/gve_adminq.h | 11
drivers/net/gve/gve_ethdev.c | 100
Those are fair points. I'll fix this by simply
calling gve_get_imissed_from_nic from gve_xstats_get in the v3 patch.
On Wed, Jan 17, 2024 at 3:10 PM Ferruh Yigit wrote:
> On 1/16/2024 6:18 AM, Rushil Gupta wrote:
> >
> >
> > On Fri, Jan 12, 2024 at 8:36 PM Ferruh
Read from shared region to retrieve imissed statistics for GQ from device.
Signed-off-by: Rushil Gupta
Reviewed-by: Joshua Washington
---
drivers/net/gve/base/gve_adminq.h | 11
drivers/net/gve/gve_ethdev.c | 103 ++
drivers/net/gve/gve_ethdev.h
I misinterpreted your comment earlier and prefixed memzone variable name
with "gve" instead of memzone name. Will fix it in v4. Thanks!
On Fri, Jan 19, 2024 at 8:55 PM Ferruh Yigit wrote:
> On 1/19/2024 2:26 PM, Rushil Gupta wrote:
> > Read from shared region to retrieve imi
Read from shared region to retrieve imissed statistics for GQ from device.
Signed-off-by: Rushil Gupta
Reviewed-by: Joshua Washington
---
drivers/net/gve/base/gve_adminq.h | 11
drivers/net/gve/gve_ethdev.c | 104 ++
drivers/net/gve/gve_ethdev.h
Hi Stephen
We wish to capture both rte version and linux kernel version.
The current implementation is needed to answer questions like how many
customers are on Dpdk 23.11 for Ubuntu 22 vs Debian 11.
Therefore, we need the uts library and the adminq is working as intended.
On Fri, Jan 12, 2024, 1
Bump up the dpdk dqo driver ring size to 4096.
The queue size is controlled by queue_setup method.
Signed-off-by: Rushil Gupta
Reviewed-by: Joshua Washington
---
drivers/net/gve/base/gve_adminq.c | 8
drivers/net/gve/gve_ethdev.c | 4 ++--
drivers/net/gve/gve_ethdev.h | 1
Gvnic's DQO format allows offloading IPv4 checksum.
Made changes to Tx and Rx path to translate DPDK flags
to descriptor for offloading (and vice-versa).
Added ptype adminq support to only add this flags for
supported L3/L4 packet-types.
---
drivers/net/gve/gve_ethdev.c | 29 ++
Gvnic's DQO format allows offloading IPv4 checksum.
Made changes to Tx and Rx path to translate DPDK flags
to descriptor for offloading (and vice-versa).
Add ptype adminq support to only add this flags for
supported L3/L4 packet-types.
Signed-off-by: Rushil Gupta
Reviewed-by: Joshua Washi
Gvnic's DQO format allows offloading IPv4 checksum.
Made changes to Tx and Rx path to translate DPDK flags
to descriptor for offloading (and vice-versa).
Add ptype adminq support to only add this flags for
supported L3/L4 packet-types.
Signed-off-by: Rushil Gupta
Reviewed-by: Joshua Washi
Gvnic's DQO format allows offloading IPv4 checksum.
Made changes to Tx and Rx path to translate DPDK flags
to descriptor for offloading (and vice-versa).
Add ptype adminq support to only add this flags for
supported L3/L4 packet-types.
Signed-off-by: Rushil Gupta
Reviewed-by: Joshua Washi
On Tue, Mar 28, 2023 at 3:37 AM Ferruh Yigit wrote:
> On 3/28/2023 10:45 AM, Junfeng Guo wrote:
> > Add Google LLC as one of the copyright holders for GVE.
> >
> > Signed-off-by: Rushil Gupta
> > Signed-off-by: Joshua Washington
> > Signed-off-by: Junfeng
les under BSD-3 License.
>
> The code under BSD is different of the MIT code?
> If it is the same with a new approved license, you can just change the
> license.
>
> > Please correct me if there are still misunderstanding in my statement.
> > Thanks Thomas for pointing out my mistake. I'll be careful to fix this.
> >
> > Copyright holder for the gve base code will stay unchanged. Google LLC
> > will be added as one of the copyright holders for the gve common code.
> > @Rushil Gupta Please also be more active and responsive for the code
> > review and contribution in the community. Thanks!
>
>
>
>
with BSD-3 License (approved by Google).
> > So I suppose that the correct way may be 1) first remove all these code
> > under MIT License and 2) then add the new files under BSD-3 License.
>
> The code under BSD is different of the MIT code?
> If it is the same with a new appr
This patch is dependent on
https://patchwork.dpdk.org/project/dpdk/patch/20230410064724.2094392-1-junfeng@intel.com/
Rushil Gupta (1):
net/gve: update base code for DQO
drivers/net/gve/base/gve.h | 1 +
drivers/net/gve/base/gve_adminq.c | 10 +-
drivers/net/gve/base
Update gve base code to support DQO.
This patch is based on this:
https://patchwork.dpdk.org/project/dpdk/list/?series=27647&state=*
Signed-off-by: Rushil Gupta
Signed-off-by: Junfeng Guo
---
drivers/net/gve/base/gve.h | 1 +
drivers/net/gve/base/gve_adminq.c
On Wed, Apr 12, 2023 at 2:41 AM Guo, Junfeng wrote:
>
>
> > -Original Message-
> > From: Ferruh Yigit
> > Sent: Wednesday, April 12, 2023 17:35
> > To: Guo, Junfeng ; Richardson, Bruce
> >
> > Cc: dev@dpdk.org; Zhang, Qi Z ; Rushil Gupta
Sorry for the confusion. I was talking about the same patch (titled
net/gve: update copyright holders); however, I am not able to find it on
patchwork.
On Wed, Apr 12, 2023 at 9:03 AM Ferruh Yigit wrote:
> On 4/12/2023 4:42 PM, Rushil Gupta wrote:
> >
> >
> > On Wed, Ap
Change gve_driver_info fields to report DPDK as OS type and DPDK RTE
version as OS version, reserving driver_version fields for GVE driver
version based on features.
This patch is dependent on
https://patchwork.dpdk.org/project/dpdk/list/?series=27687&state=*
Signed-off-by: Rushil G
Change gve_driver_info fields to report DPDK as OS type and DPDK RTE
version as OS version, reserving driver_version fields for GVE driver
version based on features.
This patch is dependent on
https://patchwork.dpdk.org/project/dpdk/list/?series=27687&state=*
Signed-off-by: Rushil G
I want to highlight that we wish to keep license changes separate from this
patch (probably for 23.11). This patch is to simply support basic
structures for the DQO data path.
On Wed, Apr 12, 2023 at 11:04 AM Rushil Gupta wrote:
> Sorry for the confusion. I was talking about the same pa
Change gve_driver_info fields to report DPDK as OS type and DPDK RTE
version as OS version, reserving driver_version fields for GVE driver
version based on features.
Depends-on: series-27687
Signed-off-by: Rushil Gupta
Signed-off-by: Joshua Washington
Signed-off-by: Junfeng Guo
Signed-off-by
> @@ -200,6 +200,9 @@ New Features
>
>Enhanced the GRO library to support TCP packets over IPv6 network.
>
> +* **Added GVE PMD enabling for DQO.**
> +
> + Added GVE PMD enabling for DQO queue format.
>
> Removed Items
> -
> --
> 2.34.1
>
Acked-by: Rushil Gupta
RS b/MAINTAINERS
> index 1c9922123e..d8c1d5272b 100644
> --- a/MAINTAINERS
> +++ b/MAINTAINERS
> @@ -698,6 +698,9 @@ F: doc/guides/nics/features/enic.ini
>
> Google Virtual Ethernet
> M: Junfeng Guo
> +M: Jeroen de Borst
> +M: Rushil Gupta
> +M: Jordan Kimbrough
>
ing
> > ; ferruh.yi...@xilinx.com; Xing, Beilei
> > ; dev@dpdk.org; jeroe...@google.com;
> > jr...@google.com; Zhang, Helin ; Rushil Gupta
> > ; Jeroen de Borst ;
> > Jordan Kimbrough
> > Subject: Re: [PATCH] maintainers: update for gve
> >
> > 09/11/2022 20
Makes sense.
On Thu, Nov 24, 2022 at 9:26 AM Ferruh Yigit wrote:
>
> On 11/24/2022 4:59 PM, Stephen Hemminger wrote:
> > On Thu, 24 Nov 2022 15:33:35 +0800
> > Junfeng Guo wrote:
> >
> >> +static int
> >> +gve_dev_stats_get(struct rte_eth_dev *dev, struct rte_eth_stats *stats)
> >> +{
> >> +
Hi all
Josh just found out some inconsistencies in the Tx/Rx statistics sum
for all ports. Not sure if we can screenshot here but it goes like
this:
Tx-dropped for port0: 447034
Tx-dropped for port1: 0
Accumulated forward statistics for all ports: 807630
Please note that this issue is only with Tx
67 matches
Mail list logo
