RE: [PATCH 0/2] net/mlx5: add modify MPLS support
Hi, > -Original Message- > From: Michael Baum > Sent: Wednesday, June 14, 2023 8:53 AM > To: dev@dpdk.org > Cc: Matan Azrad ; Raslan Darawsheh > ; Slava Ovsiienko > Subject: [PATCH 0/2] net/mlx5: add modify MPLS support > > Add support for MPLS header modification. > > RFC: > https://patchwork.dpdk.org/project/dpdk/patch/20230420094347.523784 > -1-michae...@nvidia.com/ > > Michael Baum (2): > net/mlx5: align implementation with modify API > net/mlx5: add MPLS modify field support > > doc/guides/nics/mlx5.rst| 2 ++ > drivers/common/mlx5/mlx5_prm.h | 5 +++ > drivers/net/mlx5/mlx5_flow.h| 13 > drivers/net/mlx5/mlx5_flow_dv.c | 41 > drivers/net/mlx5/mlx5_flow_hw.c | 57 +-- > -- > 5 files changed, 105 insertions(+), 13 deletions(-) > > -- > 2.25.1 Series applied to next-net-mlx, Kindest regards, Raslan Darawsheh
Re: [PATCH v4 1/4] kni: move IOVA build check
On Wed, Jun 21, 2023 at 07:00:55PM +0200, David Marchand wrote: > kni dependency to IOVA configuration does not need to be expressed in > the top level lib/meson.build file. > > Signed-off-by: David Marchand > --- > lib/kni/meson.build | 5 + > lib/meson.build | 3 --- > 2 files changed, 5 insertions(+), 3 deletions(-) Acked-by: Bruce Richardson even if the rest of the set gets pushed to next release, I think this patch should go in this one anyway. It's a simple cleanup. Thanks.
Re: [PATCH v4 2/4] build: rename enabled libraries list
On Wed, Jun 21, 2023 at 07:00:56PM +0200, David Marchand wrote: > The enabled_libs variable is renamed and moved to the top level > meson.build file, for sake of consistency wrt to drivers and > applications similar variables. > > Signed-off-by: David Marchand > --- Acked-by: Bruce Richardson Again, good cleanup for consistency. thanks.
Re: [PATCH v4 3/4] build: select deprecated libraries
On Wed, Jun 21, 2023 at 07:00:57PM +0200, David Marchand wrote: > Rework deprecated libraries selection by introducing a new configuration > option. > > This breaks existing configurations that were relying on disable_libs='' > for enabling deprecated libraries. > On the other hand, it will make enabling optional libraries more > straightforward by taking the deprecated libraries out of the picture. > > Signed-off-by: David Marchand This gives us a single on/off value for the deprecated libs. So if you wants to build only a single deprecated lib, you need to turn on this option and then use "disable_libs/enable_libs" option to then selectively pick which of the deprecated libs you actually want. Is that the expected behaviour? Just checking that we don't want this to be a list too.
Re: [PATCH v4 4/4] build: select optional libraries
On Wed, Jun 21, 2023 at 07:00:58PM +0200, David Marchand wrote:
> There is currently no way to know which libraries are optional.
> Introduce a enable_libs option (close to what we have for drivers) so
> that packagers or projects consuming DPDK can more easily select the
> optional libraries that matter to them and disable other optional
> libraries.
>
> Signed-off-by: David Marchand
> ---
LGTM - one minor suggestion inline below.
Acked-by: Bruce Richardson
> Changes since v3:
> - split non related cleanup changes,
> - fixed false positive complaints about disabling mandatory libs by
> building the list of always enabled libs, appending this list to
> enable_list and checking the disable side,
> - updated enable/disable_libs descriptions,
>
> Changes since v2:
> - moved the IOVA check for kni build in lib/kni/meson.build,
> - reworked deprecated libraries handling by only considering them when
> no enable_libs option is set. With this, no need for a default value
> in meson_options.txt, yet configurations in the CI must be adjusted,
> - moved mandatory libraries check after enable/disable_libs evaluation,
>
> ---
> lib/meson.build | 33 +++--
> meson_options.txt | 4 +++-
> 2 files changed, 26 insertions(+), 11 deletions(-)
>
> diff --git a/lib/meson.build b/lib/meson.build
> index a6e9e19b42..407769b270 100644
> --- a/lib/meson.build
> +++ b/lib/meson.build
> @@ -92,17 +92,20 @@ dpdk_libs_deprecated += [
> 'kni',
> ]
>
> -disabled_libs = []
> -opt_disabled_libs = run_command(list_dir_globs, get_option('disable_libs'),
> -check: true).stdout().split()
> -foreach l:opt_disabled_libs
> +disable_libs = run_command(list_dir_globs, get_option('disable_libs'),
> check: true).stdout().split()
> +
> +enable_libs = run_command(list_dir_globs, get_option('enable_libs'), check:
> true).stdout().split()
> +if enable_libs.length() == 0
> +enable_libs += optional_libs
> +endif
> +
> +always_enable = []
> +foreach l:libraries
> if not optional_libs.contains(l)
> -warning('Cannot disable mandatory library "@0@"'.format(l))
> -continue
> +always_enable += l
> endif
> -disabled_libs += l
> endforeach
Minor nit, I'd actually put this block further up the file just after the
optional_libs list. It's more an array definition, and moving it up puts
the next line (below) to add to enable_libs immediately after the other
assignments to enable_libs.
> -
> +enable_libs += always_enable
>
> default_cflags = machine_args
> default_cflags += ['-DALLOW_EXPERIMENTAL_API']
> @@ -139,12 +142,22 @@ foreach l:libraries
> if dpdk_libs_deprecated.contains(l) and not
> get_option('enable_deprecated_libs')
> build = false
> reason = 'deprecated libraries disabled via build config'
> -elif disabled_libs.contains(l)
> +elif not enable_libs.contains(l)
> build = false
> -reason = 'explicitly disabled via build config'
> +reason = 'not in enabled libraries build config'
> if dpdk_libs_deprecated.contains(l)
> reason += ' (deprecated lib)'
> endif
> +elif disable_libs.contains(l)
> +if always_enable.contains(l)
> +warning('Cannot disable mandatory library "@0@"'.format(l))
> +else
> +build = false
> +reason = 'explicitly disabled via build config'
> +if dpdk_libs_deprecated.contains(l)
> +reason += ' (deprecated lib)'
> +endif
> +endif
> endif
>
> if build
> diff --git a/meson_options.txt b/meson_options.txt
> index f959015e46..9951563146 100644
> --- a/meson_options.txt
> +++ b/meson_options.txt
> @@ -11,7 +11,7 @@ option('disable_apps', type: 'string', value: '',
> description:
> option('disable_drivers', type: 'string', value: '', description:
> 'Comma-separated list of drivers to explicitly disable.')
> option('disable_libs', type: 'string', value: '', description:
> - 'Comma-separated list of libraries to explicitly disable. [NOTE: not
> all libs can be disabled]')
> + 'Comma-separated list of optional libraries to explicitly disable.
> [NOTE: mandatory libs cannot be disabled]')
> option('drivers_install_subdir', type: 'string', value:
> 'dpdk/pmds-', description:
> 'Subdirectory of libdir where to install PMDs. Defaults to using a
> versioned subdirectory.')
> option('enable_docs', type: 'boolean', value: false, description:
> @@ -26,6 +26,8 @@ option('enable_driver_sdk', type: 'boolean', value: false,
> description:
> 'Install headers to build drivers.')
> option('enable_kmods', type: 'boolean', value: false, description:
> 'build kernel modules')
> +option('enable_libs', type: 'string', value: '', description:
> + 'Comma-separated list of optional libraries to explicitly enable.
> [NOTE: mandatory libs are always enabled]')
> option('examples', type: 'string',
Re: [PATCH v4 3/4] build: select deprecated libraries
On Thu, Jun 22, 2023 at 09:43:18AM +0100, Bruce Richardson wrote: > On Wed, Jun 21, 2023 at 07:00:57PM +0200, David Marchand wrote: > > Rework deprecated libraries selection by introducing a new configuration > > option. > > > > This breaks existing configurations that were relying on disable_libs='' > > for enabling deprecated libraries. > > On the other hand, it will make enabling optional libraries more > > straightforward by taking the deprecated libraries out of the picture. > > > > Signed-off-by: David Marchand > > This gives us a single on/off value for the deprecated libs. So if you > wants to build only a single deprecated lib, you need to turn on this > option and then use "disable_libs/enable_libs" option to then selectively > pick which of the deprecated libs you actually want. Is that the expected > behaviour? Just checking that we don't want this to be a list too. > The patch is fine itself though, so: Acked-by: Bruce Richardson
Re: [PATCH] maintainers: update for PCI bus driver
On 6/15/2023 5:14 PM, Nipun Gupta wrote: > Add myself as co-maintainer of PCI bus driver > > Signed-off-by: Nipun Gupta > --- > This patch is based on top of: > http://patches.dpdk.org/project/dpdk/patch/20230613065738.42370-1-chenbo@intel.com/ > > MAINTAINERS | 1 + > 1 file changed, 1 insertion(+) > > diff --git a/MAINTAINERS b/MAINTAINERS > index fea84b8cb9..bd3fff326a 100644 > --- a/MAINTAINERS > +++ b/MAINTAINERS > @@ -587,6 +587,7 @@ F: drivers/bus/fslmc/ > > PCI bus driver > M: Chenbo Xia > +M: Nipun Gupta > F: drivers/bus/pci/ > > Platform bus driver Acked-by: Ferruh Yigit
Re: [PATCH v4 0/4] Select optional libraries
On Wed, Jun 21, 2023 at 07:00:54PM +0200, David Marchand wrote: > This series is one implementation to try and please users who want to > select more easily which parts of DPDK are built. > > It introduces a change in behavior for enabling deprecated libraries: > this series is aimed at the next release but sent early as a demo of > what changes are required. > > If no strong opposition is met, a deprecation notice will be sent for > v23.07 to announce this change in v23.11. > > > Changes since v3: > - split kni cleanup, > - split variable rename cleanup, > - introduced a new meson option to control deprecated libraries > activation, > - simplified the actual implementation of enable_libs to mimic > enable_drivers behavior, > > > -- > David Marchand > > David Marchand (4): > kni: move IOVA build check > build: rename enabled libraries list > build: select deprecated libraries > build: select optional libraries > I think the first 2 patches should definitely go into 23.07. The latter two cause a change in how one needs to build DPDK with libs enabled, so we may want to push that out, and flag the change in advance. Since it's a relatively minor change, and a build-time one only, I'm fine with it going in either 23.07 or 23.11, as maintainers decide. Series-reviewed-by: Bruce Richardson
[PATCH v3] common/sfc_efx/base: set return code in case of the error
If the prefix for the rss hash was not chosen the
ENOTSUP error should be returned. Before this patch
the zero code was returned in case of an error.
Fixes: f784cdc5cbb1 ("common/sfc_efx/base: provide control to deliver RSS hash")
Signed-off-by: Artemii Morozov
Reviewed-by: Andy Moreton
---
v3: update commit log as fix commit
v2: don't use capital letters in email
drivers/common/sfc_efx/base/efx_rx.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/drivers/common/sfc_efx/base/efx_rx.c
b/drivers/common/sfc_efx/base/efx_rx.c
index 68f42f5cac..61726a9f0b 100644
--- a/drivers/common/sfc_efx/base/efx_rx.c
+++ b/drivers/common/sfc_efx/base/efx_rx.c
@@ -937,8 +937,10 @@ efx_rx_qcreate_internal(
rss_hash_field =
&erplp->erpl_fields[EFX_RX_PREFIX_FIELD_RSS_HASH];
- if (rss_hash_field->erpfi_width_bits == 0)
+ if (rss_hash_field->erpfi_width_bits == 0) {
+ rc = ENOTSUP;
goto fail5;
+ }
}
enp->en_rx_qcount++;
--
2.34.1
Re: [PATCH v3] common/sfc_efx/base: set return code in case of the error
On 6/22/23 12:13, Artemii Morozov wrote:
If the prefix for the rss hash was not chosen the
rss -> RSS
ENOTSUP error should be returned. Before this patch
the zero code was returned in case of an error.
Fixes: f784cdc5cbb1 ("common/sfc_efx/base: provide control to deliver RSS hash")
I guess Cc: sta...@dpdk.org is missing
Signed-off-by: Artemii Morozov
Reviewed-by: Andy Moreton
Acked-by: Andrew Rybchenko
Re: [PATCH v3] build: select optional libraries
On Wed, Jun 21, 2023 at 11:54 AM David Marchand wrote: > On Tue, Jun 20, 2023 at 5:05 PM Bruce Richardson > wrote: > > > > On Tue, Jun 20, 2023 at 04:33:15PM +0200, Thomas Monjalon wrote: > > > 20/06/2023 11:03, Bruce Richardson: > > > > On Tue, Jun 20, 2023 at 10:48:50AM +0200, David Marchand wrote: > > > > > On Tue, Jun 20, 2023 at 10:45 AM Bruce Richardson > > > > > wrote: > > > > > > > > > > I notice the change in behaviour for enabling the > deprecated libs. Is there > > > > > > > > > > any other change in behaviour for current users? > > > > > > > > > > > > > > > > > > The only change I see, is that this implementation breaks > enabling > > > > > > > > > deprecated libs via disable_libs. > > > > > > > > > It may break existing developer build directory and maybe > some > > > > > > > > > packaging scripts, this is why I am a bit puzzled. > > > > > > > > > > > > > > > > > > Relooking at the disable_libs option current > implementation, it seems > > > > > > > > > backward to pass a disable_libs option when you want to > build some > > > > > > > > > deprecated library. > > > > > > > > > It is more straightforward to request building libraries > via > > > > > > > > > -Denable_libs= explicitly or > -Denable_libs=* > > > > > > > > > implicitly. > > > > > > > > > > > > > > > > > > But again, we may be breaking something for people who > relied on this behavior. > > > > > > > > > > > > > > > > > > > > > > > > > That's what I expected, and I think that is ok. I just > wanted to check that > > > > > > > > the change in behaviour was only for the deprecated libs > case. > > > > > > > > > > > > > > Thomas, wdyt? > > > > > > > It requires some release note, at least. > > > > > > > > > > > > > I am assuming this is not targetting this release though, right? > Assuming > > > > > > 23.11, we can put in a deprecation note informing of the change > ahead of > > > > > > time too. > > > > > > > > > > I was hoping to get it in this release. > > > > > But I am fine with postponing and announcing the change beforehand. > > > > > > > > > Given the fact that we are likely changing behaviour, and the fact > that the > > > > deprecated libs makes it more complicated than the drivers one > (since we > > > > have always on, default on and default off cases to consider), I > think it's > > > > best we don't rush this. > > > > > > I'm not sure what is the best behaviour. > > > I tend to think such options should be simple to understand > > > with only 3 cases: > > > - no option -> default > > > - enable option -> only core mandatory and listed libraries > > > - disable option -> all but the listed libraries > > > It looks simpler to forbid having both enable and disable libraries. > > > > > > Would you be open to change the behaviour of the drivers options? > > > > > > > [reducing CC list a bit] > > > > As a further follow-up, I really think we need to move slowly and more > > carefully on this. While I can see the simplicity in disallowing the two > > options to be specified, depending on how we go about choosing the > > enabling of the deprecated libs, we may want to keep support for allowing > > both. > > I agree, we need more time on this topic, sorry for being pushy earlier > :-). > > > > > > Specifically, my current concern/thinking is: > > * David points out that using the "disabled_libs" options may not make > the > > most logic sense for *enabling* deprecated libs. > > * While that is true, I think the usability of enabling them via > > "enabled_libs" could be pretty terrible - unless we start adding more > > complications. Specifically, if someone wants to just enable KNI in the > > build using "enable_libs", specifying "-Denable_libs=kni" will not do > > what they want - sure it will enable kni, but will disable dozens of > > other parts of DPDK. > > * Therefore, I think keeping the disabling of deprecated parts of DPDK > via > > disabled_libs is easier on users - though agreed it is slightly less > > logical. However, if we forbid using enabled and disabled options > > together, that would mean that to switch to enabled libs style, the > user > > has to set both enabled libs, *and* clear the default disabled libs > option > > of the deprecated ones. > > * Therefore, right now I'm tending more towards something like the status > > quo - disabling deprecated via disable option, but allowing both enable > > and disable options together. This hasn't caused us issues with drivers > > thus far, so I'd be hopeful for using it for libs. > > Mixing enable/disable_drivers has been possible since the introduction > of enable_drivers. > Looking at the code, it is unclear the intention was to make them > exclusive. > Is there no usecase for using both options in some soc configuration? > Juraj, do you have an opinion? > > I seem to remember that there wasn't any intention of using them together - I think I did it this way because it was essentially "free" (i.e. the ability to use them together was an
[PATCH v4] common/sfc_efx/base: set return code in case of the error
If the prefix for the RSS hash was not chosen the
ENOTSUP error should be returned. Before this patch
the zero code was returned in case of an error.
Fixes: f784cdc5cbb1 ("common/sfc_efx/base: provide control to deliver RSS hash")
Cc: sta...@dpdk.org
Signed-off-by: Artemii Morozov
Reviewed-by: Andy Moreton
Acked-by: Andrew Rybchenko
---
v4: add Cc: sta...@dpdk.org and transform rss to RSS
v3: update commit log as fix commit
v2: don't use capital letters in email
drivers/common/sfc_efx/base/efx_rx.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/drivers/common/sfc_efx/base/efx_rx.c
b/drivers/common/sfc_efx/base/efx_rx.c
index 68f42f5cac..61726a9f0b 100644
--- a/drivers/common/sfc_efx/base/efx_rx.c
+++ b/drivers/common/sfc_efx/base/efx_rx.c
@@ -937,8 +937,10 @@ efx_rx_qcreate_internal(
rss_hash_field =
&erplp->erpl_fields[EFX_RX_PREFIX_FIELD_RSS_HASH];
- if (rss_hash_field->erpfi_width_bits == 0)
+ if (rss_hash_field->erpfi_width_bits == 0) {
+ rc = ENOTSUP;
goto fail5;
+ }
}
enp->en_rx_qcount++;
--
2.34.1
Re: [PATCH v2 1/4] common/sfc_efx/base: discover NIC partitioning mode
On 6/22/23 06:47, Denis Pryazhennikov wrote:
From: Sandilya Bhagi
NIC Partitioning mode in SFC devices means multiple PFs
per network port. When NIC Partitioning is configured,
apart from the privileged adapter(s) the other
unprivileged adapter(s) will share the same physical port.
Determining NIC Partitioning mode is required to take
necessary action(s) for unprivileged adapter to work seamlessly.
BNIC Partitioning is determined using heuristic approach.
If the physical ports are shared between PFs then either
NIC Partitioning or SR-IOV is in use.
Signed-off-by: Sandilya Bhagi
Signed-off-by: Denis Pryazhennikov
Reviewed-by: Andy Moreton
Overall LGMT with few style notes:
Acked-by: Andrew Rybchenko
[snip]
+#define CAP_PFS_TO_PORTS(_n) \
+ (MC_CMD_GET_CAPABILITIES_V2_OUT_PFS_TO_PORTS_ASSIGNMENT_ ## _n)
+
+ encp->enc_port_usage = EFX_PORT_USAGE_UNKNOWN;
+
+ if (req.emr_out_length_used >= MC_CMD_GET_CAPABILITIES_V2_OUT_LEN) {
+ /* PFs to ports assignment */
+ uint8_t pfs_to_ports[CAP_PFS_TO_PORTS(NUM)];
+
+ EFX_STATIC_ASSERT((CAP_PFS_TO_PORTS(NUM) *
CAP_PFS_TO_PORTS(LEN)) ==
+ EFX_ARRAY_SIZE(pfs_to_ports));
As far as I remember libefx style, it should be 4 spaces indent above
relative to mail line.
+
+ memcpy(pfs_to_ports, MCDI_OUT(req, efx_byte_t,
CAP_PFS_TO_PORTS(OFST)),
+ EFX_ARRAY_SIZE(pfs_to_ports));
same here, 4 spaces indent
+
+ rc = ef10_nic_get_physical_port_usage(enp, pfs_to_ports,
+
EFX_ARRAY_SIZE(pfs_to_ports),
+ &encp->enc_port_usage);
same here, 4 spaces indent
+ if (rc != 0) {
+ /* PF to port mapping lookup failed */
+ encp->enc_port_usage = EFX_PORT_USAGE_UNKNOWN;
+ }
+ }
+#undef CAP_PFS_TO_PORTS
+
/*
* Check if firmware reports the VI window mode.
* Medford2 has a variable VI window size (8K, 16K or 64K).
[snip]
Re: [PATCH v2 4/4] net/sfc: add configurable Rx CRC stripping
On 6/22/23 06:47, Denis Pryazhennikov wrote: Configurable Rx CRC stripping is allowed only if running firmware variant supports it and if NIC is configured with single PF per port and without VFs. When KEEP_CRC is supported, CRC will be part of the packet payload. The packet length will also contain CRC length. At the same time, CRC length should be removed from stats. Signed-off-by: Denis Pryazhennikov Signed-off-by: Roman Zhukov Reviewed-by: Andy Moreton Acked-by: Andrew Rybchenko
[PATCH] net/octeon_ep: support backward compatibility
From: Sathesh Edara
Add backward compatibility support between VF
and PF mailbox messages.
Signed-off-by: Sathesh Edara
Signed-off-by: Vamsi Attunuru
---
drivers/net/octeon_ep/otx_ep_common.h | 3 +++
drivers/net/octeon_ep/otx_ep_ethdev.c | 6 +
drivers/net/octeon_ep/otx_ep_mbox.c | 38 ++-
drivers/net/octeon_ep/otx_ep_mbox.h | 11 ++--
4 files changed, 50 insertions(+), 8 deletions(-)
diff --git a/drivers/net/octeon_ep/otx_ep_common.h
b/drivers/net/octeon_ep/otx_ep_common.h
index 42aa065a3a..c150cbe619 100644
--- a/drivers/net/octeon_ep/otx_ep_common.h
+++ b/drivers/net/octeon_ep/otx_ep_common.h
@@ -538,6 +538,9 @@ struct otx_ep_device {
/* Mailbox receive message length */
int32_t mbox_rcv_message_len;
+
+ /* Negotiated Mbox version */
+ uint32_t mbox_neg_ver;
};
int otx_ep_setup_iqs(struct otx_ep_device *otx_ep, uint32_t iq_no,
diff --git a/drivers/net/octeon_ep/otx_ep_ethdev.c
b/drivers/net/octeon_ep/otx_ep_ethdev.c
index a9868909f8..57b965ad06 100644
--- a/drivers/net/octeon_ep/otx_ep_ethdev.c
+++ b/drivers/net/octeon_ep/otx_ep_ethdev.c
@@ -666,6 +666,12 @@ otx_ep_eth_dev_init(struct rte_eth_dev *eth_dev)
otx_epvf->port_id = eth_dev->data->port_id;
eth_dev->dev_ops = &otx_ep_eth_dev_ops;
rte_spinlock_init(&otx_epvf->mbox_lock);
+
+ /*
+* Initialize negotiated Mbox version to base version of VF Mbox
+* This will address working legacy PF with latest VF.
+*/
+ otx_epvf->mbox_neg_ver = OTX_EP_MBOX_VERSION_V1;
eth_dev->data->mac_addrs = rte_zmalloc("otx_ep", RTE_ETHER_ADDR_LEN, 0);
if (eth_dev->data->mac_addrs == NULL) {
otx_ep_err("MAC addresses memory allocation failed\n");
diff --git a/drivers/net/octeon_ep/otx_ep_mbox.c
b/drivers/net/octeon_ep/otx_ep_mbox.c
index 1ad36e14c8..4118645dc7 100644
--- a/drivers/net/octeon_ep/otx_ep_mbox.c
+++ b/drivers/net/octeon_ep/otx_ep_mbox.c
@@ -12,6 +12,14 @@
#include "cnxk_ep_vf.h"
#include "otx_ep_mbox.h"
+/*
+ * When a new command is implemented, the below table should be updated
+ * with new command and it's version info.
+ */
+static uint32_t otx_ep_cmd_versions[OTX_EP_MBOX_CMD_MAX] = {
+ [0 ... OTX_EP_MBOX_CMD_DEV_REMOVE] = OTX_EP_MBOX_VERSION_V1
+};
+
static int
__otx_ep_send_mbox_cmd(struct otx_ep_device *otx_ep,
union otx_ep_mbox_word cmd,
@@ -56,6 +64,12 @@ otx_ep_send_mbox_cmd(struct otx_ep_device *otx_ep,
int ret;
rte_spinlock_lock(&otx_ep->mbox_lock);
+ if (otx_ep_cmd_versions[cmd.s.opcode] > otx_ep->mbox_neg_ver) {
+ otx_ep_dbg("CMD:%d not supported in Version:%d\n", cmd.s.opcode,
+ otx_ep->mbox_neg_ver);
+ rte_spinlock_unlock(&otx_ep->mbox_lock);
+ return -EOPNOTSUPP;
+ }
ret = __otx_ep_send_mbox_cmd(otx_ep, cmd, rsp);
rte_spinlock_unlock(&otx_ep->mbox_lock);
return ret;
@@ -284,15 +298,27 @@ int otx_ep_mbox_version_check(struct rte_eth_dev *eth_dev)
cmd.u64 = 0;
cmd.s_version.opcode = OTX_EP_MBOX_CMD_VERSION;
- cmd.s_version.version = OTX_EP_MBOX_VERSION;
+ cmd.s_version.version = OTX_EP_MBOX_VERSION_CURRENT;
ret = otx_ep_send_mbox_cmd(otx_ep, cmd, &rsp);
- if (!ret)
- return 0;
- if (ret == OTX_EP_MBOX_CMD_STATUS_NACK) {
- otx_ep_err("VF Mbox version:%u is not compatible with PF\n",
+
+ /*
+* VF receives NACK or version info as zero
+* only if PF driver running old version of Mailbox
+* In this case VF mailbox version fallbacks to base
+* mailbox vesrion OTX_EP_MBOX_VERSION_V1.
+* Default VF mbox_neg_ver is set to OTX_EP_MBOX_VERSION_V1
+* during initialization of PMD driver.
+*/
+ if (ret == OTX_EP_MBOX_CMD_STATUS_NACK || rsp.s_version.version == 0) {
+ otx_ep_dbg("VF Mbox version fallback to base version from:%u\n",
(uint32_t)cmd.s_version.version);
+ return 0;
}
- return ret;
+ otx_ep->mbox_neg_ver = (uint32_t)rsp.s_version.version;
+ otx_ep_dbg("VF Mbox version:%u Negotiated VF version with PF:%u\n",
+ (uint32_t)cmd.s_version.version,
+ (uint32_t)rsp.s_version.version);
+ return 0;
}
int otx_ep_mbox_send_dev_exit(struct rte_eth_dev *eth_dev)
diff --git a/drivers/net/octeon_ep/otx_ep_mbox.h
b/drivers/net/octeon_ep/otx_ep_mbox.h
index 9df3c53edd..a3fc15cca7 100644
--- a/drivers/net/octeon_ep/otx_ep_mbox.h
+++ b/drivers/net/octeon_ep/otx_ep_mbox.h
@@ -5,8 +5,15 @@
#ifndef _OTX_EP_MBOX_H_
#define _OTX_EP_MBOX_H_
+/*
+ * When a new command is implemented, VF Mbox version should be bumped.
+ */
+enum octep_pfvf_mbox_version {
+ OTX_EP_MBOX_VERSION_V0,
+ OTX_EP_MBOX_VERSION_V1,
+};
-#define OTX_EP_MBOX_VERSION1
+#define OTX_EP_MBOX_VERSI
Re: [PATCH v4] common/sfc_efx/base: set return code in case of the error
On 6/22/2023 10:31 AM, Artemii Morozov wrote:
> If the prefix for the RSS hash was not chosen the
> ENOTSUP error should be returned. Before this patch
> the zero code was returned in case of an error.
>
What is the impact of the problem to user?
Previously 'efx_rx_qcreate_internal()' was returning success although it
fails, and 'erpp' was not set.
There are multiple callers of 'efx_rx_qcreate_internal()', those
functions will assume functions succeed when it failed, what kind of
problem is this cause?
Does it cause a crash, or failure in the Rx queue creation, or no impact
at all, can you please describe?
Impact information is not always easy to understand from code change or
low level details, but most of the times that is what users are interested.
Like fixing a return value can be a simple refactoring without any
functional impact at all, or it can be causing a crash and making driver
completely unusable for some cases, although code change can be similar
these two cases has different priority.
So commit message/title should highlight the impact, starting from 'fix'
keyword.
Assume it cause Rx queue creation failure, commit can be something like:
```
common/sfc_efx/base: fix Rx queue creation without RSS hash prefix
If the prefix for the RSS hash was not chosen the ENOTSUP error should
be returned.
Before this patch success was returned for this case causing Rx queue
creation to fail.
Fixing return value to indicate failure.
Fixes: ...
Cc: ...
...
```
> Fixes: f784cdc5cbb1 ("common/sfc_efx/base: provide control to deliver RSS
> hash")
> Cc: sta...@dpdk.org
>
> Signed-off-by: Artemii Morozov
> Reviewed-by: Andy Moreton
> Acked-by: Andrew Rybchenko
> ---
> v4: add Cc: sta...@dpdk.org and transform rss to RSS
>
> v3: update commit log as fix commit
>
> v2: don't use capital letters in email
>
> drivers/common/sfc_efx/base/efx_rx.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/common/sfc_efx/base/efx_rx.c
> b/drivers/common/sfc_efx/base/efx_rx.c
> index 68f42f5cac..61726a9f0b 100644
> --- a/drivers/common/sfc_efx/base/efx_rx.c
> +++ b/drivers/common/sfc_efx/base/efx_rx.c
> @@ -937,8 +937,10 @@ efx_rx_qcreate_internal(
>
> rss_hash_field =
> &erplp->erpl_fields[EFX_RX_PREFIX_FIELD_RSS_HASH];
> - if (rss_hash_field->erpfi_width_bits == 0)
> + if (rss_hash_field->erpfi_width_bits == 0) {
> + rc = ENOTSUP;
> goto fail5;
> + }
> }
>
> enp->en_rx_qcount++;
[PATCH v6 0/4] net/sfc: support VLAN stripping offload
This patch series adds VLAN stripping offload. Note that this offload are device level offload. v6: * highlight that efx_port_vlan_strip_set() must be called before any filter insertion * avoid an extra check if offload is not requested v5: * fixed problems with naming * fixed problems with abbreviations * fixed problems with isolated mode * fixed problems with consistency v4: * fix apply patch failure warning v3: * fix apply patch failure warning v2: * rebase patches on top of dpdk-next-net/main Artemii Morozov (4): common/sfc_efx/base: report VLAN stripping capability common/sfc_efx/base: add API to get installed filters count common/sfc_efx/base: add support to enable VLAN stripping net/sfc: support VLAN stripping offload doc/guides/nics/sfc_efx.rst | 4 +-- doc/guides/rel_notes/release_23_07.rst| 6 drivers/common/sfc_efx/base/ef10_filter.c | 26 +++ drivers/common/sfc_efx/base/ef10_impl.h | 6 drivers/common/sfc_efx/base/ef10_nic.c| 6 drivers/common/sfc_efx/base/efx.h | 13 drivers/common/sfc_efx/base/efx_filter.c | 27 drivers/common/sfc_efx/base/efx_impl.h| 8 + drivers/common/sfc_efx/base/efx_port.c| 39 +++ drivers/common/sfc_efx/base/efx_rx.c | 14 drivers/common/sfc_efx/base/rhead_rx.c| 3 ++ drivers/common/sfc_efx/base/siena_nic.c | 1 + drivers/common/sfc_efx/version.map| 1 + drivers/net/sfc/sfc.h | 1 + drivers/net/sfc/sfc_dp_rx.h | 1 + drivers/net/sfc/sfc_ef100_rx.c| 24 +- drivers/net/sfc/sfc_port.c| 12 +++ drivers/net/sfc/sfc_rx.c | 10 ++ 18 files changed, 199 insertions(+), 3 deletions(-) -- 2.34.1
[PATCH v6 1/4] common/sfc_efx/base: report VLAN stripping capability
These changes are necessary in order to add support for stripping
VLAN tags in the future.
Signed-off-by: Artemii Morozov
Reviewed-by: Ivan Malov
Reviewed-by: Andy Moreton
---
drivers/common/sfc_efx/base/ef10_nic.c | 6 ++
drivers/common/sfc_efx/base/efx.h | 1 +
drivers/common/sfc_efx/base/siena_nic.c | 1 +
3 files changed, 8 insertions(+)
diff --git a/drivers/common/sfc_efx/base/ef10_nic.c
b/drivers/common/sfc_efx/base/ef10_nic.c
index e1709d1200..070f4d08b5 100644
--- a/drivers/common/sfc_efx/base/ef10_nic.c
+++ b/drivers/common/sfc_efx/base/ef10_nic.c
@@ -1146,6 +1146,12 @@ ef10_get_datapath_caps(
else
encp->enc_hw_tx_insert_vlan_enabled = B_FALSE;
+ /* Check if firmware supports VLAN stripping. */
+ if (CAP_FLAGS1(req, RX_VLAN_STRIPPING))
+ encp->enc_rx_vlan_stripping_supported = B_TRUE;
+ else
+ encp->enc_rx_vlan_stripping_supported = B_FALSE;
+
/* Check if the firmware supports RX event batching */
if (CAP_FLAGS1(req, RX_BATCHING))
encp->enc_rx_batching_enabled = B_TRUE;
diff --git a/drivers/common/sfc_efx/base/efx.h
b/drivers/common/sfc_efx/base/efx.h
index 10c412bcd7..9a29583ecb 100644
--- a/drivers/common/sfc_efx/base/efx.h
+++ b/drivers/common/sfc_efx/base/efx.h
@@ -1616,6 +1616,7 @@ typedef struct efx_nic_cfg_s {
/* Number of TSO contexts on the NIC (FATSOv2) */
uint32_tenc_fw_assisted_tso_v2_n_contexts;
boolean_t enc_hw_tx_insert_vlan_enabled;
+ boolean_t enc_rx_vlan_stripping_supported;
/* Number of PFs on the NIC */
uint32_tenc_hw_pf_count;
/* Datapath firmware vadapter/vport/vswitch support */
diff --git a/drivers/common/sfc_efx/base/siena_nic.c
b/drivers/common/sfc_efx/base/siena_nic.c
index 9f14faf271..4b7f7cbb87 100644
--- a/drivers/common/sfc_efx/base/siena_nic.c
+++ b/drivers/common/sfc_efx/base/siena_nic.c
@@ -179,6 +179,7 @@ siena_board_cfg(
(encp->enc_rxq_limit * EFX_RXQ_DC_NDESCS(EFX_RXQ_DC_SIZE));
encp->enc_hw_tx_insert_vlan_enabled = B_FALSE;
+ encp->enc_rx_vlan_stripping_supported = B_FALSE;
encp->enc_fw_assisted_tso_enabled = B_FALSE;
encp->enc_fw_assisted_tso_v2_enabled = B_FALSE;
encp->enc_fw_assisted_tso_v2_n_contexts = 0;
--
2.34.1
[PATCH v6 2/4] common/sfc_efx/base: add API to get installed filters count
This API allows to get number of installed filters. This will
be used in the future patches.
Signed-off-by: Artemii Morozov
---
drivers/common/sfc_efx/base/ef10_filter.c | 20 +
drivers/common/sfc_efx/base/ef10_impl.h | 6 +
drivers/common/sfc_efx/base/efx_filter.c | 27 +++
drivers/common/sfc_efx/base/efx_impl.h| 7 ++
4 files changed, 60 insertions(+)
diff --git a/drivers/common/sfc_efx/base/ef10_filter.c
b/drivers/common/sfc_efx/base/ef10_filter.c
index d6940011c0..278502fb61 100644
--- a/drivers/common/sfc_efx/base/ef10_filter.c
+++ b/drivers/common/sfc_efx/base/ef10_filter.c
@@ -2113,6 +2113,26 @@ ef10_filter_reconfigure(
return (rc);
}
+ __checkReturn efx_rc_t
+ef10_filter_get_count(
+ __inefx_nic_t *enp,
+ __out uint32_t *count)
+{
+ ef10_filter_table_t *table = enp->en_filter.ef_ef10_filter_table;
+ uint32_t filters_count;
+
+ EFSYS_ASSERT(EFX_FAMILY_IS_EF100(enp) || EFX_FAMILY_IS_EF10(enp));
+ EFSYS_ASSERT(count != NULL);
+
+ filters_count = table->eft_unicst_filter_count +
+ table->eft_mulcst_filter_count +
+ table->eft_encap_filter_count;
+
+ *count = filters_count;
+
+ return (0);
+}
+
void
ef10_filter_get_default_rxq(
__inefx_nic_t *enp,
diff --git a/drivers/common/sfc_efx/base/ef10_impl.h
b/drivers/common/sfc_efx/base/ef10_impl.h
index 877aedad45..914bba10ce 100644
--- a/drivers/common/sfc_efx/base/ef10_impl.h
+++ b/drivers/common/sfc_efx/base/ef10_impl.h
@@ -1347,6 +1347,12 @@ ef10_filter_reconfigure(
__in_ecount(6*count)uint8_t const *addrs,
__inuint32_t count);
+LIBEFX_INTERNAL
+extern __checkReturn efx_rc_t
+ef10_filter_get_count(
+ __inefx_nic_t *enp,
+ __out uint32_t *count);
+
LIBEFX_INTERNAL
extern void
ef10_filter_get_default_rxq(
diff --git a/drivers/common/sfc_efx/base/efx_filter.c
b/drivers/common/sfc_efx/base/efx_filter.c
index 83c37ff859..a8b62ee88d 100644
--- a/drivers/common/sfc_efx/base/efx_filter.c
+++ b/drivers/common/sfc_efx/base/efx_filter.c
@@ -53,6 +53,7 @@ static const efx_filter_ops_t __efx_filter_siena_ops = {
siena_filter_delete,/* efo_delete */
siena_filter_supported_filters, /* efo_supported_filters */
NULL, /* efo_reconfigure */
+ NULL, /* efo_get_count */
};
#endif /* EFSYS_OPT_SIENA */
@@ -65,6 +66,7 @@ static const efx_filter_ops_t __efx_filter_ef10_ops = {
ef10_filter_delete, /* efo_delete */
ef10_filter_supported_filters, /* efo_supported_filters */
ef10_filter_reconfigure,/* efo_reconfigure */
+ ef10_filter_get_count, /* efo_get_count */
};
#endif /* EFX_OPTS_EF10() */
@@ -77,6 +79,7 @@ static const efx_filter_ops_t __efx_filter_rhead_ops = {
ef10_filter_delete, /* efo_delete */
ef10_filter_supported_filters, /* efo_supported_filters */
ef10_filter_reconfigure,/* efo_reconfigure */
+ ef10_filter_get_count, /* efo_get_count */
};
#endif /* EFSYS_OPT_RIVERHEAD */
@@ -306,6 +309,30 @@ efx_filter_reconfigure(
return (0);
+fail1:
+ EFSYS_PROBE1(fail1, efx_rc_t, rc);
+
+ return (rc);
+}
+
+ __checkReturn efx_rc_t
+efx_filter_get_count(
+ __inefx_nic_t *enp,
+ __out uint32_t *count)
+{
+ efx_rc_t rc;
+
+ EFSYS_ASSERT3U(enp->en_magic, ==, EFX_NIC_MAGIC);
+ EFSYS_ASSERT3U(enp->en_mod_flags, &, EFX_MOD_PROBE);
+ EFSYS_ASSERT3U(enp->en_mod_flags, &, EFX_MOD_FILTER);
+
+ if (enp->en_efop->efo_get_count != NULL) {
+ if ((rc = enp->en_efop->efo_get_count(enp, count)) != 0)
+ goto fail1;
+ }
+
+ return (0);
+
fail1:
EFSYS_PROBE1(fail1, efx_rc_t, rc);
diff --git a/drivers/common/sfc_efx/base/efx_impl.h
b/drivers/common/sfc_efx/base/efx_impl.h
index 45e99d01c5..d657734bc5 100644
--- a/drivers/common/sfc_efx/base/efx_impl.h
+++ b/drivers/common/sfc_efx/base/efx_impl.h
@@ -288,6 +288,7 @@ typedef struct efx_filter_ops_s {
efx_rc_t(*efo_reconfigure)(efx_nic_t *, uint8_t const *,
boolean_t,
boolean_t, boolean_t, boolean_t,
uint8_t const *, uint32_t);
+ efx_rc_t(*efo_get_count)(efx_nic_t *, uint32_t *);
} efx_filter_ops_t;
LIBEFX_INTERNAL
@@ -302,6 +303,12 @@ efx_filter_reconfigure(
__in_ecount(6*count)uint8_t const *addrs,
__inuint32_t count);
+LIBEFX_INTERNAL
+extern __checkReturn efx_rc_t
+efx_filter_get_count(
+ __inefx_nic_t *enp,
+ __out uint32_t *count);
+
#endif /* EFSYS_OPT_FILTER */
#if EFSYS_OPT_TUNNEL
--
2.34.1
[PATCH v6 3/4] common/sfc_efx/base: add support to enable VLAN stripping
To enable VLAN stripping, two conditions must be met:
the corresponding flag must be set and the appropriate
Rx prefix should be requested.
VLAN stripping is supported on EF100.
Signed-off-by: Artemii Morozov
Reviewed-by: Ivan Malov
Reviewed-by: Andy Moreton
---
drivers/common/sfc_efx/base/ef10_filter.c | 6
drivers/common/sfc_efx/base/efx.h | 12 +++
drivers/common/sfc_efx/base/efx_impl.h| 1 +
drivers/common/sfc_efx/base/efx_port.c| 39 +++
drivers/common/sfc_efx/base/efx_rx.c | 14
drivers/common/sfc_efx/base/rhead_rx.c| 3 ++
drivers/common/sfc_efx/version.map| 1 +
7 files changed, 76 insertions(+)
diff --git a/drivers/common/sfc_efx/base/ef10_filter.c
b/drivers/common/sfc_efx/base/ef10_filter.c
index 278502fb61..827b3e8f00 100644
--- a/drivers/common/sfc_efx/base/ef10_filter.c
+++ b/drivers/common/sfc_efx/base/ef10_filter.c
@@ -171,6 +171,7 @@ efx_mcdi_filter_op_add(
EFX_MCDI_DECLARE_BUF(payload, MC_CMD_FILTER_OP_V3_IN_LEN,
MC_CMD_FILTER_OP_EXT_OUT_LEN);
efx_filter_match_flags_t match_flags;
+ efx_port_t *epp = &(enp->en_port);
uint32_t port_id;
efx_rc_t rc;
@@ -338,6 +339,11 @@ efx_mcdi_filter_op_add(
FILTER_OP_V3_IN_MATCH_SET_FLAG, 1);
}
+ if (epp->ep_vlan_strip) {
+ MCDI_IN_SET_DWORD_FIELD(req, FILTER_OP_V3_IN_MATCH_ACTION_FLAGS,
+ FILTER_OP_V3_IN_MATCH_STRIP_VLAN, 1);
+ }
+
efx_mcdi_execute(enp, &req);
if (req.emr_rc != 0) {
diff --git a/drivers/common/sfc_efx/base/efx.h
b/drivers/common/sfc_efx/base/efx.h
index 9a29583ecb..016bbc8ec9 100644
--- a/drivers/common/sfc_efx/base/efx.h
+++ b/drivers/common/sfc_efx/base/efx.h
@@ -1147,6 +1147,12 @@ efx_port_poll(
__inefx_nic_t *enp,
__out_opt efx_link_mode_t *link_modep);
+LIBEFX_API
+extern __checkReturn efx_rc_t
+efx_port_vlan_strip_set(
+ __inefx_nic_t *enp,
+ __inboolean_t enabled);
+
LIBEFX_API
extern void
efx_port_fini(
@@ -3101,6 +3107,12 @@ typedef enum efx_rxq_type_e {
* Request user flag field in the Rx prefix of a queue.
*/
#defineEFX_RXQ_FLAG_USER_FLAG 0x20
+/*
+ * Request VLAN TCI field in the Rx prefix. The flag just
+ * controls delivery of the stripped VLAN TCI if VLAN stripping
+ * is enabled and done.
+ */
+#defineEFX_RXQ_FLAG_VLAN_STRIPPED_TCI 0x40
LIBEFX_API
extern __checkReturn efx_rc_t
diff --git a/drivers/common/sfc_efx/base/efx_impl.h
b/drivers/common/sfc_efx/base/efx_impl.h
index d657734bc5..de9d1dddc8 100644
--- a/drivers/common/sfc_efx/base/efx_impl.h
+++ b/drivers/common/sfc_efx/base/efx_impl.h
@@ -370,6 +370,7 @@ typedef struct efx_port_s {
uint32_tep_default_adv_cap_mask;
uint32_tep_phy_cap_mask;
boolean_t ep_mac_drain;
+ boolean_t ep_vlan_strip;
#if EFSYS_OPT_BIST
efx_bist_type_t ep_current_bist;
#endif
diff --git a/drivers/common/sfc_efx/base/efx_port.c
b/drivers/common/sfc_efx/base/efx_port.c
index a5f982e335..e5a9fa6c53 100644
--- a/drivers/common/sfc_efx/base/efx_port.c
+++ b/drivers/common/sfc_efx/base/efx_port.c
@@ -204,6 +204,45 @@ efx_loopback_type_name(
#endif /* EFSYS_OPT_LOOPBACK */
+ __checkReturn efx_rc_t
+efx_port_vlan_strip_set(
+ __inefx_nic_t *enp,
+ __inboolean_t enabled)
+{
+ efx_port_t *epp = &(enp->en_port);
+ efx_nic_cfg_t *encp = &(enp->en_nic_cfg);
+ uint32_t filters_count = 0;
+ efx_rc_t rc;
+
+ EFSYS_ASSERT3U(enp->en_magic, ==, EFX_NIC_MAGIC);
+
+ if (enabled && !encp->enc_rx_vlan_stripping_supported) {
+ rc = ENOTSUP;
+ goto fail1;
+ }
+
+ if ((rc = efx_filter_get_count(enp, &filters_count)) != 0)
+ goto fail2;
+
+ if (filters_count != 0) {
+ rc = EINVAL;
+ goto fail3;
+ }
+
+ epp->ep_vlan_strip = enabled;
+
+ return (0);
+
+fail3:
+ EFSYS_PROBE(fail3);
+fail2:
+ EFSYS_PROBE(fail2);
+fail1:
+ EFSYS_PROBE1(fail1, efx_rc_t, rc);
+
+ return (rc);
+}
+
void
efx_port_fini(
__inefx_nic_t *enp)
diff --git a/drivers/common/sfc_efx/base/efx_rx.c
b/drivers/common/sfc_efx/base/efx_rx.c
index 68f42f5cac..b3d9e14c67 100644
--- a/drivers/common/sfc_efx/base/efx_rx.c
+++ b/drivers/common/sfc_efx/base/efx_rx.c
@@ -941,11 +941,25 @@ efx_rx_qcreate_internal(
goto fail5;
}
+ if (flags & EFX_RXQ_FLAG_VLAN_STRIPPED_TCI) {
+ const efx_rx_prefix_layout_t *erplp = &erp->er_prefix_layout;
+ const efx_rx_prefix_field_info_t *vlan_tci_field;
+
+ vlan_tci_field =
+ &erplp->erpl_fields[EFX_RX_PREFIX_FIEL
[PATCH v6 4/4] net/sfc: support VLAN stripping offload
Extract VLAN TCI provided by the HW in the prefix and put it to mbuf.
VLAN stripping is supported for ef100 datapath only. This is device
level offload.
Signed-off-by: Artemii Morozov
Reviewed-by: Viacheslav Galaktionov
Reviewed-by: Ivan Malov
Reviewed-by: Andy Moreton
---
doc/guides/nics/sfc_efx.rst| 4 ++--
doc/guides/rel_notes/release_23_07.rst | 6 ++
drivers/net/sfc/sfc.h | 1 +
drivers/net/sfc/sfc_dp_rx.h| 1 +
drivers/net/sfc/sfc_ef100_rx.c | 24 +++-
drivers/net/sfc/sfc_port.c | 12
drivers/net/sfc/sfc_rx.c | 10 ++
7 files changed, 55 insertions(+), 3 deletions(-)
diff --git a/doc/guides/nics/sfc_efx.rst b/doc/guides/nics/sfc_efx.rst
index de0656876b..44fa24e1ba 100644
--- a/doc/guides/nics/sfc_efx.rst
+++ b/doc/guides/nics/sfc_efx.rst
@@ -118,6 +118,8 @@ SFC EFX PMD has support for:
- Port representors (see :ref: switch_representation)
+- VLAN stripping (if running firmware variant supports it)
+
Non-supported Features
--
@@ -132,8 +134,6 @@ The features not yet supported include:
- VLAN filtering
-- VLAN stripping
-
- LRO
diff --git a/doc/guides/rel_notes/release_23_07.rst
b/doc/guides/rel_notes/release_23_07.rst
index dc0d250e16..b961ef139f 100644
--- a/doc/guides/rel_notes/release_23_07.rst
+++ b/doc/guides/rel_notes/release_23_07.rst
@@ -177,6 +177,12 @@ New Features
See :doc:`../prog_guide/pdcp_lib` for more information.
+* **Updated Solarflare network PMD.**
+
+ Updated the Solarflare ``sfc_efx`` driver with changes including:
+
+ * Added VLAN stripping support on SN1000 SmartNICs
+
Removed Items
-
diff --git a/drivers/net/sfc/sfc.h b/drivers/net/sfc/sfc.h
index e42abe42cb..b68bcc7d4f 100644
--- a/drivers/net/sfc/sfc.h
+++ b/drivers/net/sfc/sfc.h
@@ -73,6 +73,7 @@ struct sfc_port {
unsigned intflow_ctrl;
boolean_t flow_ctrl_autoneg;
+ boolean_t vlan_strip;
size_t pdu;
/*
diff --git a/drivers/net/sfc/sfc_dp_rx.h b/drivers/net/sfc/sfc_dp_rx.h
index 8a504bdcf1..9f9bf28988 100644
--- a/drivers/net/sfc/sfc_dp_rx.h
+++ b/drivers/net/sfc/sfc_dp_rx.h
@@ -70,6 +70,7 @@ struct sfc_dp_rx_qcreate_info {
unsigned intflags;
#define SFC_RXQ_FLAG_RSS_HASH 0x1
#define SFC_RXQ_FLAG_INGRESS_MPORT 0x2
+#define SFC_RXQ_FLAG_VLAN_STRIPPED_TCI 0x4
/** Rx queue size */
unsigned intrxq_entries;
diff --git a/drivers/net/sfc/sfc_ef100_rx.c b/drivers/net/sfc/sfc_ef100_rx.c
index 37b754fa33..07381df5cf 100644
--- a/drivers/net/sfc/sfc_ef100_rx.c
+++ b/drivers/net/sfc/sfc_ef100_rx.c
@@ -68,6 +68,7 @@ struct sfc_ef100_rxq {
#define SFC_EF100_RXQ_INGRESS_MPORT0x80
#define SFC_EF100_RXQ_USER_FLAG0x100
#define SFC_EF100_RXQ_NIC_DMA_MAP 0x200
+#define SFC_EF100_RXQ_VLAN_STRIPPED_TCI0x400
unsigned intptr_mask;
unsigned intevq_phase_bit_shift;
unsigned intready_pkts;
@@ -392,6 +393,7 @@ static const efx_rx_prefix_layout_t
sfc_ef100_rx_prefix_layout = {
SFC_EF100_RX_PREFIX_FIELD(RSS_HASH, B_FALSE),
SFC_EF100_RX_PREFIX_FIELD(USER_FLAG, B_FALSE),
SFC_EF100_RX_PREFIX_FIELD(USER_MARK, B_FALSE),
+ SFC_EF100_RX_PREFIX_FIELD(VLAN_STRIP_TCI, B_FALSE),
#undef SFC_EF100_RX_PREFIX_FIELD
}
@@ -472,6 +474,14 @@ sfc_ef100_rx_prefix_to_offloads(const struct sfc_ef100_rxq
*rxq,
ESF_GZ_RX_PREFIX_INGRESS_MPORT);
}
+ if (rxq->flags & SFC_EF100_RXQ_VLAN_STRIPPED_TCI &&
+ EFX_TEST_XWORD_BIT(rx_prefix[0],
+ ESF_GZ_RX_PREFIX_VLAN_STRIPPED_LBN)) {
+ ol_flags |= RTE_MBUF_F_RX_VLAN | RTE_MBUF_F_RX_VLAN_STRIPPED;
+ m->vlan_tci = EFX_XWORD_FIELD(rx_prefix[0],
+
ESF_GZ_RX_PREFIX_VLAN_STRIP_TCI);
+ }
+
m->ol_flags = ol_flags;
return true;
}
@@ -813,6 +823,9 @@ sfc_ef100_rx_qcreate(uint16_t port_id, uint16_t queue_id,
if (info->flags & SFC_RXQ_FLAG_INGRESS_MPORT)
rxq->flags |= SFC_EF100_RXQ_INGRESS_MPORT;
+ if (info->flags & SFC_RXQ_FLAG_VLAN_STRIPPED_TCI)
+ rxq->flags |= SFC_EF100_RXQ_VLAN_STRIPPED_TCI;
+
sfc_ef100_rx_debug(rxq, "RxQ doorbell is %p", rxq->doorbell);
*dp_rxqp = &rxq->dp;
@@ -892,6 +905,15 @@ sfc_ef100_rx_qstart(struct sfc_dp_rxq *dp_rxq, unsigned
int evq_read_ptr,
(rxq->flags & SFC_EF100_RXQ_INGRESS_MPORT))
return ENOTSUP;
+ /*
+* Exclude the SFC_EF100_RXQ_VLAN_STRIPPED_TCI if offload was not
requested
+* or the prefix does not contain the corresponding fie
[PATCH v3 0/4] net/sfc: support KEEP_CRC offload
This patch series adds support for RTE_ETH_RX_OFFLOAD_KEEP_CRC offload in the SFC driver. Changes in v2: * Applied review notes; * Fixed stats handling; * Added missing release notes to [4/4]. Changes in v3: * Fixed wrong indents in [1/4]. Denis Pryazhennikov (3): common/sfc_efx/base: detect and report FCS include support common/sfc_efx/base: add support for configure MAC to keep FCS net/sfc: add configurable Rx CRC stripping Sandilya Bhagi (1): common/sfc_efx/base: discover NIC partitioning mode doc/guides/nics/features/sfc.ini| 1 + doc/guides/nics/sfc_efx.rst | 6 +- doc/guides/rel_notes/release_23_07.rst | 2 + drivers/common/sfc_efx/base/ef10_mac.c | 5 +- drivers/common/sfc_efx/base/ef10_nic.c | 107 drivers/common/sfc_efx/base/efx.h | 14 drivers/common/sfc_efx/base/efx_impl.h | 1 + drivers/common/sfc_efx/base/efx_mac.c | 48 +++ drivers/common/sfc_efx/base/siena_nic.c | 1 + drivers/common/sfc_efx/version.map | 1 + drivers/net/sfc/sfc.h | 1 + drivers/net/sfc/sfc_ef100_rx.c | 2 +- drivers/net/sfc/sfc_ef10_rx.c | 3 +- drivers/net/sfc/sfc_ethdev.c| 7 +- drivers/net/sfc/sfc_port.c | 12 +++ drivers/net/sfc/sfc_rx.c| 6 +- 16 files changed, 209 insertions(+), 8 deletions(-) -- 2.39.2 (Apple Git-143)
[PATCH v3 1/4] common/sfc_efx/base: discover NIC partitioning mode
From: Sandilya Bhagi
NIC Partitioning mode in SFC devices means multiple PFs
per network port. When NIC Partitioning is configured,
apart from the privileged adapter(s) the other
unprivileged adapter(s) will share the same physical port.
Determining NIC Partitioning mode is required to take
necessary action(s) for unprivileged adapter to work seamlessly.
NIC Partitioning is determined using heuristic approach.
If the physical ports are shared between PFs then either
NIC Partitioning or SR-IOV is in use.
Signed-off-by: Sandilya Bhagi
Signed-off-by: Denis Pryazhennikov
Reviewed-by: Andy Moreton
Acked-by: Andrew Rybchenko
---
drivers/common/sfc_efx/base/ef10_nic.c | 101 +
drivers/common/sfc_efx/base/efx.h | 8 ++
2 files changed, 109 insertions(+)
diff --git a/drivers/common/sfc_efx/base/ef10_nic.c
b/drivers/common/sfc_efx/base/ef10_nic.c
index e1709d120093..0c6a2eee4453 100644
--- a/drivers/common/sfc_efx/base/ef10_nic.c
+++ b/drivers/common/sfc_efx/base/ef10_nic.c
@@ -1044,6 +1044,83 @@ ef10_mcdi_get_pf_count(
return (rc);
}
+static __checkReturn efx_rc_t
+ef10_nic_get_physical_port_usage(
+ __inefx_nic_t *enp,
+ __in_ecount(pfs_to_ports_size) uint8_t *pfs_to_ports,
+ __insize_t pfs_to_ports_size,
+ __out efx_port_usage_t *port_usagep)
+{
+ efx_nic_cfg_t *encp = &(enp->en_nic_cfg);
+ efx_port_usage_t port_usage;
+ uint8_t phy_port;
+ efx_rc_t rc;
+ size_t pf;
+
+ /*
+* The sharing of physical ports between functions are determined
+* in the following way.
+* 1. If VFs are enabled then the physical port is shared.
+* 2. Retrieve PFs to ports assignment.
+* 3. If PF 0 assignment cannot be retrieved(ACCESS_DENIED), it
+*implies this is an unprivileged function. An unprivileged
+*function indicates the physical port must be shared with
+*another privileged function.
+* 4. If PF 0 assignment can be retrieved, it indicates this
+*function is privileged. Now, read all other PF's physical
+*port number assignment and check if the current PF's physical
+*port is shared with any other PF's physical port.
+* NOTE: PF 0 is always privileged function.
+*/
+
+ if (EFX_PCI_FUNCTION_IS_VF(encp)) {
+ port_usage = EFX_PORT_USAGE_SHARED;
+ goto out;
+ }
+
+ if (pfs_to_ports[0] ==
+ MC_CMD_GET_CAPABILITIES_V2_OUT_ACCESS_NOT_PERMITTED) {
+ /*
+* This is unprivileged function as it do not have sufficient
+* privileges to read the value, this implies the physical port
+* is shared between this function and another privileged
+* function
+*/
+ port_usage = EFX_PORT_USAGE_SHARED;
+ goto out;
+ }
+
+ if (encp->enc_pf >= pfs_to_ports_size) {
+ rc = EINVAL;
+ goto fail1;
+ }
+ phy_port = pfs_to_ports[encp->enc_pf];
+
+ /*
+* This is privileged function as it is able read the value of
+* PF 0. Now, check if any other function share the same physical
+* port number as this function.
+*/
+ for (pf = 0; pf < pfs_to_ports_size; pf++) {
+ if ((encp->enc_pf != pf) && (phy_port == pfs_to_ports[pf])) {
+ /* Found match, PFs share the same physical port */
+ port_usage = EFX_PORT_USAGE_SHARED;
+ goto out;
+ }
+ }
+
+ port_usage = EFX_PORT_USAGE_EXCLUSIVE;
+
+out:
+ *port_usagep = port_usage;
+ return (0);
+
+fail1:
+ EFSYS_PROBE1(fail1, efx_rc_t, rc);
+
+ return (rc);
+}
+
static __checkReturn efx_rc_t
ef10_get_datapath_caps(
__inefx_nic_t *enp)
@@ -1307,6 +1384,30 @@ ef10_get_datapath_caps(
encp->enc_tunnel_config_udp_entries_max = 0;
}
+#define CAP_PFS_TO_PORTS(_n) \
+ (MC_CMD_GET_CAPABILITIES_V2_OUT_PFS_TO_PORTS_ASSIGNMENT_ ## _n)
+
+ encp->enc_port_usage = EFX_PORT_USAGE_UNKNOWN;
+
+ if (req.emr_out_length_used >= MC_CMD_GET_CAPABILITIES_V2_OUT_LEN) {
+ /* PFs to ports assignment */
+ uint8_t pfs_to_ports[CAP_PFS_TO_PORTS(NUM)];
+
+ EFX_STATIC_ASSERT((CAP_PFS_TO_PORTS(NUM) *
CAP_PFS_TO_PORTS(LEN)) ==
+ EFX_ARRAY_SIZE(pfs_to_ports));
+
+ memcpy(pfs_to_ports, MCDI_OUT(req, efx_byte_t,
CAP_PFS_TO_PORTS(OFST)),
+ EFX_ARRAY_SIZE(pfs_to_ports));
+
+ rc = ef10_nic_get_physical_port_usage(enp, pfs_to_ports,
+ EFX_ARRAY_SIZE(pfs_to_ports), &encp->enc_port_usage);
+ if (rc != 0) {
+
[PATCH v3 2/4] common/sfc_efx/base: detect and report FCS include support
From: Roman Zhukov
A new variable was added to efx_nic_cfg_s to detect and
report if FCS is supported by FW.
Signed-off-by: Roman Zhukov
Signed-off-by: Denis Pryazhennikov
Reviewed-by: Viacheslav Galaktionov
Acked-by: Andrew Rybchenko
---
drivers/common/sfc_efx/base/ef10_nic.c | 6 ++
drivers/common/sfc_efx/base/efx.h | 1 +
drivers/common/sfc_efx/base/siena_nic.c | 1 +
3 files changed, 8 insertions(+)
diff --git a/drivers/common/sfc_efx/base/ef10_nic.c
b/drivers/common/sfc_efx/base/ef10_nic.c
index 0c6a2eee4453..bf9cb9d30990 100644
--- a/drivers/common/sfc_efx/base/ef10_nic.c
+++ b/drivers/common/sfc_efx/base/ef10_nic.c
@@ -1244,6 +1244,12 @@ ef10_get_datapath_caps(
/* No limit on maximum number of Rx scatter elements per packet. */
encp->enc_rx_scatter_max = -1;
+ /* Check if the firmware supports include FCS on RX */
+ if (CAP_FLAGS1(req, RX_INCLUDE_FCS))
+ encp->enc_rx_include_fcs_supported = B_TRUE;
+ else
+ encp->enc_rx_include_fcs_supported = B_FALSE;
+
/* Check if the firmware supports packed stream mode */
if (CAP_FLAGS1(req, RX_PACKED_STREAM))
encp->enc_rx_packed_stream_supported = B_TRUE;
diff --git a/drivers/common/sfc_efx/base/efx.h
b/drivers/common/sfc_efx/base/efx.h
index a63211612249..b5bd390169ae 100644
--- a/drivers/common/sfc_efx/base/efx.h
+++ b/drivers/common/sfc_efx/base/efx.h
@@ -1629,6 +1629,7 @@ typedef struct efx_nic_cfg_s {
/* Datapath firmware vport reconfigure support */
boolean_t enc_vport_reconfigure_supported;
boolean_t enc_rx_disable_scatter_supported;
+ boolean_t enc_rx_include_fcs_supported;
/* Maximum number of Rx scatter segments supported by HW */
uint32_tenc_rx_scatter_max;
boolean_t enc_allow_set_mac_with_installed_filters;
diff --git a/drivers/common/sfc_efx/base/siena_nic.c
b/drivers/common/sfc_efx/base/siena_nic.c
index 1f1fb7d2afd1..ca38eefa7e37 100644
--- a/drivers/common/sfc_efx/base/siena_nic.c
+++ b/drivers/common/sfc_efx/base/siena_nic.c
@@ -187,6 +187,7 @@ siena_board_cfg(
encp->enc_allow_set_mac_with_installed_filters = B_TRUE;
encp->enc_rx_packed_stream_supported = B_FALSE;
encp->enc_rx_var_packed_stream_supported = B_FALSE;
+ encp->enc_rx_include_fcs_supported = B_FALSE;
encp->enc_rx_es_super_buffer_supported = B_FALSE;
encp->enc_fw_subvariant_no_tx_csum_supported = B_FALSE;
--
2.39.2 (Apple Git-143)
[PATCH v3 3/4] common/sfc_efx/base: add support for configure MAC to keep FCS
From: Roman Zhukov
Signed-off-by: Denis Pryazhennikov
Reviewed-by: Andy Moreton
Acked-by: Andrew Rybchenko
---
drivers/common/sfc_efx/base/ef10_mac.c | 5 +--
drivers/common/sfc_efx/base/efx.h | 5 +++
drivers/common/sfc_efx/base/efx_impl.h | 1 +
drivers/common/sfc_efx/base/efx_mac.c | 48 ++
drivers/common/sfc_efx/version.map | 1 +
5 files changed, 58 insertions(+), 2 deletions(-)
diff --git a/drivers/common/sfc_efx/base/ef10_mac.c
b/drivers/common/sfc_efx/base/ef10_mac.c
index 28228a9fb784..bfc82b80c7e5 100644
--- a/drivers/common/sfc_efx/base/ef10_mac.c
+++ b/drivers/common/sfc_efx/base/ef10_mac.c
@@ -307,9 +307,10 @@ ef10_mac_reconfigure(
*/
MCDI_IN_SET_DWORD(req, SET_MAC_IN_FCNTL, MC_CMD_FCNTL_AUTO);
- /* Do not include the Ethernet frame checksum in RX packets */
+ /* Include the Ethernet frame checksum in RX packets if it's required */
MCDI_IN_POPULATE_DWORD_1(req, SET_MAC_IN_FLAGS,
- SET_MAC_IN_FLAG_INCLUDE_FCS, 0);
+ SET_MAC_IN_FLAG_INCLUDE_FCS,
+ epp->ep_include_fcs ? 1 : 0);
efx_mcdi_execute_quiet(enp, &req);
diff --git a/drivers/common/sfc_efx/base/efx.h
b/drivers/common/sfc_efx/base/efx.h
index b5bd390169ae..ef626cc55a7e 100644
--- a/drivers/common/sfc_efx/base/efx.h
+++ b/drivers/common/sfc_efx/base/efx.h
@@ -747,6 +747,11 @@ efx_mac_fcntl_get(
__out unsigned int *fcntl_wantedp,
__out unsigned int *fcntl_linkp);
+LIBEFX_API
+extern __checkReturn efx_rc_t
+efx_mac_include_fcs_set(
+ __in efx_nic_t *enp,
+ __in boolean_t enabled);
#if EFSYS_OPT_MAC_STATS
diff --git a/drivers/common/sfc_efx/base/efx_impl.h
b/drivers/common/sfc_efx/base/efx_impl.h
index 09b1e95c594c..92a30c34ae28 100644
--- a/drivers/common/sfc_efx/base/efx_impl.h
+++ b/drivers/common/sfc_efx/base/efx_impl.h
@@ -363,6 +363,7 @@ typedef struct efx_port_s {
uint32_tep_default_adv_cap_mask;
uint32_tep_phy_cap_mask;
boolean_t ep_mac_drain;
+ boolean_t ep_include_fcs;
#if EFSYS_OPT_BIST
efx_bist_type_t ep_current_bist;
#endif
diff --git a/drivers/common/sfc_efx/base/efx_mac.c
b/drivers/common/sfc_efx/base/efx_mac.c
index c51e86b52c29..13cac5a75130 100644
--- a/drivers/common/sfc_efx/base/efx_mac.c
+++ b/drivers/common/sfc_efx/base/efx_mac.c
@@ -527,6 +527,54 @@ efx_mac_filter_default_rxq_clear(
emop->emo_filter_default_rxq_clear(enp);
}
+ __checkReturn efx_rc_t
+efx_mac_include_fcs_set(
+ __inefx_nic_t *enp,
+ __inboolean_t enabled)
+{
+ efx_port_t *epp = &(enp->en_port);
+ efx_nic_cfg_t *encp = &(enp->en_nic_cfg);
+ const efx_mac_ops_t *emop = epp->ep_emop;
+ efx_rc_t rc;
+
+ EFSYS_ASSERT3U(enp->en_magic, ==, EFX_NIC_MAGIC);
+ EFSYS_ASSERT3U(enp->en_mod_flags, &, EFX_MOD_PORT);
+ EFSYS_ASSERT(emop != NULL);
+
+ if (enabled && !encp->enc_rx_include_fcs_supported) {
+ rc = ENOTSUP;
+ goto fail1;
+ }
+
+ /*
+* Enabling 'include FCS' changes link control state and affects
+* behaviour for all PCI functions on the port, so to avoid this it
+* can be enabled only if the PCI function is exclusive port user
+*/
+ if (enabled && encp->enc_port_usage != EFX_PORT_USAGE_EXCLUSIVE) {
+ rc = EACCES;
+ goto fail2;
+ }
+
+ if (epp->ep_include_fcs != enabled) {
+ epp->ep_include_fcs = enabled;
+
+ rc = emop->emo_reconfigure(enp);
+ if (rc != 0)
+ goto fail3;
+ }
+
+ return 0;
+
+fail3:
+ EFSYS_PROBE(fail3);
+fail2:
+ EFSYS_PROBE(fail2);
+fail1:
+ EFSYS_PROBE1(fail1, efx_rc_t, rc);
+
+ return rc;
+}
#if EFSYS_OPT_MAC_STATS
diff --git a/drivers/common/sfc_efx/version.map
b/drivers/common/sfc_efx/version.map
index e91bcbcad863..01113bffa7cb 100644
--- a/drivers/common/sfc_efx/version.map
+++ b/drivers/common/sfc_efx/version.map
@@ -71,6 +71,7 @@ INTERNAL {
efx_mac_drain;
efx_mac_fcntl_get;
efx_mac_fcntl_set;
+ efx_mac_include_fcs_set;
efx_mac_filter_default_rxq_clear;
efx_mac_filter_default_rxq_set;
efx_mac_filter_get_all_ucast_mcast;
--
2.39.2 (Apple Git-143)
[PATCH v3 4/4] net/sfc: add configurable Rx CRC stripping
Configurable Rx CRC stripping is allowed only if
running firmware variant supports it and if NIC is
configured with single PF per port and without VFs.
When KEEP_CRC is supported, CRC will be part of
the packet payload. The packet length will also
contain CRC length. At the same time, CRC length
should be removed from stats.
Signed-off-by: Denis Pryazhennikov
Signed-off-by: Roman Zhukov
Reviewed-by: Andy Moreton
Acked-by: Andrew Rybchenko
---
doc/guides/nics/features/sfc.ini | 1 +
doc/guides/nics/sfc_efx.rst| 6 --
doc/guides/rel_notes/release_23_07.rst | 2 ++
drivers/net/sfc/sfc.h | 1 +
drivers/net/sfc/sfc_ef100_rx.c | 2 +-
drivers/net/sfc/sfc_ef10_rx.c | 3 ++-
drivers/net/sfc/sfc_ethdev.c | 7 ++-
drivers/net/sfc/sfc_port.c | 12
drivers/net/sfc/sfc_rx.c | 6 +-
9 files changed, 34 insertions(+), 6 deletions(-)
diff --git a/doc/guides/nics/features/sfc.ini b/doc/guides/nics/features/sfc.ini
index a4b53c619c12..8a9198adcb5b 100644
--- a/doc/guides/nics/features/sfc.ini
+++ b/doc/guides/nics/features/sfc.ini
@@ -23,6 +23,7 @@ RSS key update = Y
RSS reta update = Y
SR-IOV = Y
Flow control = Y
+CRC offload = Y
VLAN offload = P
FEC = Y
L3 checksum offload = Y
diff --git a/doc/guides/nics/sfc_efx.rst b/doc/guides/nics/sfc_efx.rst
index ba82b020932b..24459da33ea8 100644
--- a/doc/guides/nics/sfc_efx.rst
+++ b/doc/guides/nics/sfc_efx.rst
@@ -114,6 +114,10 @@ SFC EFX PMD has support for:
- Loopback
+- Configurable Rx CRC stripping (if running firmware variant supports it and
+ if NIC is configured with single PF per port and without VFs, otherwise
+ always stripped)
+
- SR-IOV PF
- Port representors (see :ref: switch_representation)
@@ -126,8 +130,6 @@ The features not yet supported include:
- Priority-based flow control
-- Configurable RX CRC stripping (always stripped)
-
- Header split on receive
- VLAN filtering
diff --git a/doc/guides/rel_notes/release_23_07.rst
b/doc/guides/rel_notes/release_23_07.rst
index 3db443d16018..738d35d9f7c6 100644
--- a/doc/guides/rel_notes/release_23_07.rst
+++ b/doc/guides/rel_notes/release_23_07.rst
@@ -145,6 +145,8 @@ New Features
* Added support for transfer flow action INDIRECT with subtype COUNT,
for aggregated statistics.
+ * Added support for keeping CRC.
+
* **Added vmxnet3 version 7 support.**
Added support for vmxnet3 version 7 which includes support
diff --git a/drivers/net/sfc/sfc.h b/drivers/net/sfc/sfc.h
index 89e2c46fa974..25cdeaa5cd25 100644
--- a/drivers/net/sfc/sfc.h
+++ b/drivers/net/sfc/sfc.h
@@ -74,6 +74,7 @@ struct sfc_port {
unsigned intflow_ctrl;
boolean_t flow_ctrl_autoneg;
+ boolean_t include_fcs;
size_t pdu;
/*
diff --git a/drivers/net/sfc/sfc_ef100_rx.c b/drivers/net/sfc/sfc_ef100_rx.c
index 37b754fa3305..5563bd9a0bd1 100644
--- a/drivers/net/sfc/sfc_ef100_rx.c
+++ b/drivers/net/sfc/sfc_ef100_rx.c
@@ -1004,7 +1004,7 @@ struct sfc_dp_rx sfc_ef100_rx = {
SFC_DP_RX_FEAT_FLOW_MARK |
SFC_DP_RX_FEAT_INTR |
SFC_DP_RX_FEAT_STATS,
- .dev_offload_capa = 0,
+ .dev_offload_capa = RTE_ETH_RX_OFFLOAD_KEEP_CRC,
.queue_offload_capa = RTE_ETH_RX_OFFLOAD_CHECKSUM |
RTE_ETH_RX_OFFLOAD_OUTER_IPV4_CKSUM |
RTE_ETH_RX_OFFLOAD_OUTER_UDP_CKSUM |
diff --git a/drivers/net/sfc/sfc_ef10_rx.c b/drivers/net/sfc/sfc_ef10_rx.c
index 7be224c9c412..30a320d0791c 100644
--- a/drivers/net/sfc/sfc_ef10_rx.c
+++ b/drivers/net/sfc/sfc_ef10_rx.c
@@ -825,7 +825,8 @@ struct sfc_dp_rx sfc_ef10_rx = {
SFC_DP_RX_FEAT_INTR,
.dev_offload_capa = RTE_ETH_RX_OFFLOAD_CHECKSUM |
RTE_ETH_RX_OFFLOAD_OUTER_IPV4_CKSUM |
- RTE_ETH_RX_OFFLOAD_RSS_HASH,
+ RTE_ETH_RX_OFFLOAD_RSS_HASH |
+ RTE_ETH_RX_OFFLOAD_KEEP_CRC,
.queue_offload_capa = RTE_ETH_RX_OFFLOAD_SCATTER,
.get_dev_info = sfc_ef10_rx_get_dev_info,
.qsize_up_rings = sfc_ef10_rx_qsize_up_rings,
diff --git a/drivers/net/sfc/sfc_ethdev.c b/drivers/net/sfc/sfc_ethdev.c
index f120ddc5a8d8..1efe64a36a7f 100644
--- a/drivers/net/sfc/sfc_ethdev.c
+++ b/drivers/net/sfc/sfc_ethdev.c
@@ -689,8 +689,13 @@ sfc_stats_get(struct rte_eth_dev *dev, struct
rte_eth_stats *stats)
sfc_adapter_lock(sa);
- if (have_dp_rx_stats)
+ if (have_dp_rx_stats) {
sfc_stats_get_dp_rx(sa, &stats->ipackets, &stats->ibytes);
+ if (dev->data->dev_c
Re: [PATCH v6 1/4] common/sfc_efx/base: report VLAN stripping capability
On 6/22/23 14:31, Artemii Morozov wrote: These changes are necessary in order to add support for stripping VLAN tags in the future. Signed-off-by: Artemii Morozov Reviewed-by: Ivan Malov Reviewed-by: Andy Moreton Acked-by: Andrew Rybchenko
[PATCH v4 1/4] common/sfc_efx/base: discover NIC partitioning mode
From: Sandilya Bhagi
NIC Partitioning mode in SFC devices means multiple PFs
per network port. When NIC Partitioning is configured,
apart from the privileged adapter(s) the other
unprivileged adapter(s) will share the same physical port.
Determining NIC Partitioning mode is required to take
necessary action(s) for unprivileged adapter to work seamlessly.
NIC Partitioning is determined using heuristic approach.
If the physical ports are shared between PFs then either
NIC Partitioning or SR-IOV is in use.
Signed-off-by: Sandilya Bhagi
Signed-off-by: Denis Pryazhennikov
Reviewed-by: Andy Moreton
Acked-by: Andrew Rybchenko
---
drivers/common/sfc_efx/base/ef10_nic.c | 101 +
drivers/common/sfc_efx/base/efx.h | 8 ++
2 files changed, 109 insertions(+)
diff --git a/drivers/common/sfc_efx/base/ef10_nic.c
b/drivers/common/sfc_efx/base/ef10_nic.c
index e1709d120093..0c6a2eee4453 100644
--- a/drivers/common/sfc_efx/base/ef10_nic.c
+++ b/drivers/common/sfc_efx/base/ef10_nic.c
@@ -1044,6 +1044,83 @@ ef10_mcdi_get_pf_count(
return (rc);
}
+static __checkReturn efx_rc_t
+ef10_nic_get_physical_port_usage(
+ __inefx_nic_t *enp,
+ __in_ecount(pfs_to_ports_size) uint8_t *pfs_to_ports,
+ __insize_t pfs_to_ports_size,
+ __out efx_port_usage_t *port_usagep)
+{
+ efx_nic_cfg_t *encp = &(enp->en_nic_cfg);
+ efx_port_usage_t port_usage;
+ uint8_t phy_port;
+ efx_rc_t rc;
+ size_t pf;
+
+ /*
+* The sharing of physical ports between functions are determined
+* in the following way.
+* 1. If VFs are enabled then the physical port is shared.
+* 2. Retrieve PFs to ports assignment.
+* 3. If PF 0 assignment cannot be retrieved(ACCESS_DENIED), it
+*implies this is an unprivileged function. An unprivileged
+*function indicates the physical port must be shared with
+*another privileged function.
+* 4. If PF 0 assignment can be retrieved, it indicates this
+*function is privileged. Now, read all other PF's physical
+*port number assignment and check if the current PF's physical
+*port is shared with any other PF's physical port.
+* NOTE: PF 0 is always privileged function.
+*/
+
+ if (EFX_PCI_FUNCTION_IS_VF(encp)) {
+ port_usage = EFX_PORT_USAGE_SHARED;
+ goto out;
+ }
+
+ if (pfs_to_ports[0] ==
+ MC_CMD_GET_CAPABILITIES_V2_OUT_ACCESS_NOT_PERMITTED) {
+ /*
+* This is unprivileged function as it do not have sufficient
+* privileges to read the value, this implies the physical port
+* is shared between this function and another privileged
+* function
+*/
+ port_usage = EFX_PORT_USAGE_SHARED;
+ goto out;
+ }
+
+ if (encp->enc_pf >= pfs_to_ports_size) {
+ rc = EINVAL;
+ goto fail1;
+ }
+ phy_port = pfs_to_ports[encp->enc_pf];
+
+ /*
+* This is privileged function as it is able read the value of
+* PF 0. Now, check if any other function share the same physical
+* port number as this function.
+*/
+ for (pf = 0; pf < pfs_to_ports_size; pf++) {
+ if ((encp->enc_pf != pf) && (phy_port == pfs_to_ports[pf])) {
+ /* Found match, PFs share the same physical port */
+ port_usage = EFX_PORT_USAGE_SHARED;
+ goto out;
+ }
+ }
+
+ port_usage = EFX_PORT_USAGE_EXCLUSIVE;
+
+out:
+ *port_usagep = port_usage;
+ return (0);
+
+fail1:
+ EFSYS_PROBE1(fail1, efx_rc_t, rc);
+
+ return (rc);
+}
+
static __checkReturn efx_rc_t
ef10_get_datapath_caps(
__inefx_nic_t *enp)
@@ -1307,6 +1384,30 @@ ef10_get_datapath_caps(
encp->enc_tunnel_config_udp_entries_max = 0;
}
+#define CAP_PFS_TO_PORTS(_n) \
+ (MC_CMD_GET_CAPABILITIES_V2_OUT_PFS_TO_PORTS_ASSIGNMENT_ ## _n)
+
+ encp->enc_port_usage = EFX_PORT_USAGE_UNKNOWN;
+
+ if (req.emr_out_length_used >= MC_CMD_GET_CAPABILITIES_V2_OUT_LEN) {
+ /* PFs to ports assignment */
+ uint8_t pfs_to_ports[CAP_PFS_TO_PORTS(NUM)];
+
+ EFX_STATIC_ASSERT((CAP_PFS_TO_PORTS(NUM) *
CAP_PFS_TO_PORTS(LEN)) ==
+ EFX_ARRAY_SIZE(pfs_to_ports));
+
+ memcpy(pfs_to_ports, MCDI_OUT(req, efx_byte_t,
CAP_PFS_TO_PORTS(OFST)),
+ EFX_ARRAY_SIZE(pfs_to_ports));
+
+ rc = ef10_nic_get_physical_port_usage(enp, pfs_to_ports,
+ EFX_ARRAY_SIZE(pfs_to_ports), &encp->enc_port_usage);
+ if (rc != 0) {
+
[PATCH v4 0/4] net/sfc: support KEEP_CRC offload
This patch series adds support for RTE_ETH_RX_OFFLOAD_KEEP_CRC offload in the SFC driver. Changes in v2: * Applied review notes; * Fixed stats handling; * Added missing release notes to [4/4]. Changes in v3: * Fixed wrong indents in [1/4]. Changes in v4: * Fixed "From" section in [3/4]. Denis Pryazhennikov (3): common/sfc_efx/base: detect and report FCS include support common/sfc_efx/base: add support for configure MAC to keep FCS net/sfc: add configurable Rx CRC stripping Sandilya Bhagi (1): common/sfc_efx/base: discover NIC partitioning mode doc/guides/nics/features/sfc.ini| 1 + doc/guides/nics/sfc_efx.rst | 6 +- doc/guides/rel_notes/release_23_07.rst | 2 + drivers/common/sfc_efx/base/ef10_mac.c | 5 +- drivers/common/sfc_efx/base/ef10_nic.c | 107 drivers/common/sfc_efx/base/efx.h | 14 drivers/common/sfc_efx/base/efx_impl.h | 1 + drivers/common/sfc_efx/base/efx_mac.c | 48 +++ drivers/common/sfc_efx/base/siena_nic.c | 1 + drivers/common/sfc_efx/version.map | 1 + drivers/net/sfc/sfc.h | 1 + drivers/net/sfc/sfc_ef100_rx.c | 2 +- drivers/net/sfc/sfc_ef10_rx.c | 3 +- drivers/net/sfc/sfc_ethdev.c| 7 +- drivers/net/sfc/sfc_port.c | 12 +++ drivers/net/sfc/sfc_rx.c| 6 +- 16 files changed, 209 insertions(+), 8 deletions(-) -- 2.39.2 (Apple Git-143)
[PATCH v4 2/4] common/sfc_efx/base: detect and report FCS include support
From: Roman Zhukov
A new variable was added to efx_nic_cfg_s to detect and
report if FCS is supported by FW.
Signed-off-by: Roman Zhukov
Signed-off-by: Denis Pryazhennikov
Reviewed-by: Viacheslav Galaktionov
Acked-by: Andrew Rybchenko
---
drivers/common/sfc_efx/base/ef10_nic.c | 6 ++
drivers/common/sfc_efx/base/efx.h | 1 +
drivers/common/sfc_efx/base/siena_nic.c | 1 +
3 files changed, 8 insertions(+)
diff --git a/drivers/common/sfc_efx/base/ef10_nic.c
b/drivers/common/sfc_efx/base/ef10_nic.c
index 0c6a2eee4453..bf9cb9d30990 100644
--- a/drivers/common/sfc_efx/base/ef10_nic.c
+++ b/drivers/common/sfc_efx/base/ef10_nic.c
@@ -1244,6 +1244,12 @@ ef10_get_datapath_caps(
/* No limit on maximum number of Rx scatter elements per packet. */
encp->enc_rx_scatter_max = -1;
+ /* Check if the firmware supports include FCS on RX */
+ if (CAP_FLAGS1(req, RX_INCLUDE_FCS))
+ encp->enc_rx_include_fcs_supported = B_TRUE;
+ else
+ encp->enc_rx_include_fcs_supported = B_FALSE;
+
/* Check if the firmware supports packed stream mode */
if (CAP_FLAGS1(req, RX_PACKED_STREAM))
encp->enc_rx_packed_stream_supported = B_TRUE;
diff --git a/drivers/common/sfc_efx/base/efx.h
b/drivers/common/sfc_efx/base/efx.h
index a63211612249..b5bd390169ae 100644
--- a/drivers/common/sfc_efx/base/efx.h
+++ b/drivers/common/sfc_efx/base/efx.h
@@ -1629,6 +1629,7 @@ typedef struct efx_nic_cfg_s {
/* Datapath firmware vport reconfigure support */
boolean_t enc_vport_reconfigure_supported;
boolean_t enc_rx_disable_scatter_supported;
+ boolean_t enc_rx_include_fcs_supported;
/* Maximum number of Rx scatter segments supported by HW */
uint32_tenc_rx_scatter_max;
boolean_t enc_allow_set_mac_with_installed_filters;
diff --git a/drivers/common/sfc_efx/base/siena_nic.c
b/drivers/common/sfc_efx/base/siena_nic.c
index 1f1fb7d2afd1..ca38eefa7e37 100644
--- a/drivers/common/sfc_efx/base/siena_nic.c
+++ b/drivers/common/sfc_efx/base/siena_nic.c
@@ -187,6 +187,7 @@ siena_board_cfg(
encp->enc_allow_set_mac_with_installed_filters = B_TRUE;
encp->enc_rx_packed_stream_supported = B_FALSE;
encp->enc_rx_var_packed_stream_supported = B_FALSE;
+ encp->enc_rx_include_fcs_supported = B_FALSE;
encp->enc_rx_es_super_buffer_supported = B_FALSE;
encp->enc_fw_subvariant_no_tx_csum_supported = B_FALSE;
--
2.39.2 (Apple Git-143)
[PATCH v4 3/4] common/sfc_efx/base: add support for configure MAC to keep FCS
From: Roman Zhukov
Drivers cannot determine if received packet includes the FCS or not.
Only packets with an external port have the FCS included and functions
without link control privilege cannot determine the MAC configuration.
This patch is trying to make assumptions that: if PF is the only function
(there are no VFs or additional PFs); it can set the MAC configuration and
it never expects packets it sends to be looped back then it can assume that
changed the MAC configuration to include the FCS is safe and that all
received packets will include their FCS.
Signed-off-by: Roman Zhukov
Signed-off-by: Denis Pryazhennikov
Reviewed-by: Andy Moreton
Acked-by: Andrew Rybchenko
---
drivers/common/sfc_efx/base/ef10_mac.c | 5 +--
drivers/common/sfc_efx/base/efx.h | 5 +++
drivers/common/sfc_efx/base/efx_impl.h | 1 +
drivers/common/sfc_efx/base/efx_mac.c | 48 ++
drivers/common/sfc_efx/version.map | 1 +
5 files changed, 58 insertions(+), 2 deletions(-)
diff --git a/drivers/common/sfc_efx/base/ef10_mac.c
b/drivers/common/sfc_efx/base/ef10_mac.c
index 28228a9fb784..bfc82b80c7e5 100644
--- a/drivers/common/sfc_efx/base/ef10_mac.c
+++ b/drivers/common/sfc_efx/base/ef10_mac.c
@@ -307,9 +307,10 @@ ef10_mac_reconfigure(
*/
MCDI_IN_SET_DWORD(req, SET_MAC_IN_FCNTL, MC_CMD_FCNTL_AUTO);
- /* Do not include the Ethernet frame checksum in RX packets */
+ /* Include the Ethernet frame checksum in RX packets if it's required */
MCDI_IN_POPULATE_DWORD_1(req, SET_MAC_IN_FLAGS,
- SET_MAC_IN_FLAG_INCLUDE_FCS, 0);
+ SET_MAC_IN_FLAG_INCLUDE_FCS,
+ epp->ep_include_fcs ? 1 : 0);
efx_mcdi_execute_quiet(enp, &req);
diff --git a/drivers/common/sfc_efx/base/efx.h
b/drivers/common/sfc_efx/base/efx.h
index b5bd390169ae..ef626cc55a7e 100644
--- a/drivers/common/sfc_efx/base/efx.h
+++ b/drivers/common/sfc_efx/base/efx.h
@@ -747,6 +747,11 @@ efx_mac_fcntl_get(
__out unsigned int *fcntl_wantedp,
__out unsigned int *fcntl_linkp);
+LIBEFX_API
+extern __checkReturn efx_rc_t
+efx_mac_include_fcs_set(
+ __in efx_nic_t *enp,
+ __in boolean_t enabled);
#if EFSYS_OPT_MAC_STATS
diff --git a/drivers/common/sfc_efx/base/efx_impl.h
b/drivers/common/sfc_efx/base/efx_impl.h
index 09b1e95c594c..92a30c34ae28 100644
--- a/drivers/common/sfc_efx/base/efx_impl.h
+++ b/drivers/common/sfc_efx/base/efx_impl.h
@@ -363,6 +363,7 @@ typedef struct efx_port_s {
uint32_tep_default_adv_cap_mask;
uint32_tep_phy_cap_mask;
boolean_t ep_mac_drain;
+ boolean_t ep_include_fcs;
#if EFSYS_OPT_BIST
efx_bist_type_t ep_current_bist;
#endif
diff --git a/drivers/common/sfc_efx/base/efx_mac.c
b/drivers/common/sfc_efx/base/efx_mac.c
index c51e86b52c29..13cac5a75130 100644
--- a/drivers/common/sfc_efx/base/efx_mac.c
+++ b/drivers/common/sfc_efx/base/efx_mac.c
@@ -527,6 +527,54 @@ efx_mac_filter_default_rxq_clear(
emop->emo_filter_default_rxq_clear(enp);
}
+ __checkReturn efx_rc_t
+efx_mac_include_fcs_set(
+ __inefx_nic_t *enp,
+ __inboolean_t enabled)
+{
+ efx_port_t *epp = &(enp->en_port);
+ efx_nic_cfg_t *encp = &(enp->en_nic_cfg);
+ const efx_mac_ops_t *emop = epp->ep_emop;
+ efx_rc_t rc;
+
+ EFSYS_ASSERT3U(enp->en_magic, ==, EFX_NIC_MAGIC);
+ EFSYS_ASSERT3U(enp->en_mod_flags, &, EFX_MOD_PORT);
+ EFSYS_ASSERT(emop != NULL);
+
+ if (enabled && !encp->enc_rx_include_fcs_supported) {
+ rc = ENOTSUP;
+ goto fail1;
+ }
+
+ /*
+* Enabling 'include FCS' changes link control state and affects
+* behaviour for all PCI functions on the port, so to avoid this it
+* can be enabled only if the PCI function is exclusive port user
+*/
+ if (enabled && encp->enc_port_usage != EFX_PORT_USAGE_EXCLUSIVE) {
+ rc = EACCES;
+ goto fail2;
+ }
+
+ if (epp->ep_include_fcs != enabled) {
+ epp->ep_include_fcs = enabled;
+
+ rc = emop->emo_reconfigure(enp);
+ if (rc != 0)
+ goto fail3;
+ }
+
+ return 0;
+
+fail3:
+ EFSYS_PROBE(fail3);
+fail2:
+ EFSYS_PROBE(fail2);
+fail1:
+ EFSYS_PROBE1(fail1, efx_rc_t, rc);
+
+ return rc;
+}
#if EFSYS_OPT_MAC_STATS
diff --git a/drivers/common/sfc_efx/version.map
b/drivers/common/sfc_efx/version.map
index e91bcbcad863..01113bffa7cb 100644
--- a/drivers/common/sfc_efx/version.map
+++ b/drivers/common/sfc_efx/version.map
@@ -71,6 +71,7 @@ INTERNAL {
efx_mac_drain;
efx_mac_fcntl_get;
efx_mac_fcntl_set;
+ efx_mac_include_fcs_set;
efx_mac_filter_default_rx
[PATCH v4 4/4] net/sfc: add configurable Rx CRC stripping
Configurable Rx CRC stripping is allowed only if
running firmware variant supports it and if NIC is
configured with single PF per port and without VFs.
When KEEP_CRC is supported, CRC will be part of
the packet payload. The packet length will also
contain CRC length. At the same time, CRC length
should be removed from stats.
Signed-off-by: Denis Pryazhennikov
Signed-off-by: Roman Zhukov
Reviewed-by: Andy Moreton
Acked-by: Andrew Rybchenko
---
doc/guides/nics/features/sfc.ini | 1 +
doc/guides/nics/sfc_efx.rst| 6 --
doc/guides/rel_notes/release_23_07.rst | 2 ++
drivers/net/sfc/sfc.h | 1 +
drivers/net/sfc/sfc_ef100_rx.c | 2 +-
drivers/net/sfc/sfc_ef10_rx.c | 3 ++-
drivers/net/sfc/sfc_ethdev.c | 7 ++-
drivers/net/sfc/sfc_port.c | 12
drivers/net/sfc/sfc_rx.c | 6 +-
9 files changed, 34 insertions(+), 6 deletions(-)
diff --git a/doc/guides/nics/features/sfc.ini b/doc/guides/nics/features/sfc.ini
index a4b53c619c12..8a9198adcb5b 100644
--- a/doc/guides/nics/features/sfc.ini
+++ b/doc/guides/nics/features/sfc.ini
@@ -23,6 +23,7 @@ RSS key update = Y
RSS reta update = Y
SR-IOV = Y
Flow control = Y
+CRC offload = Y
VLAN offload = P
FEC = Y
L3 checksum offload = Y
diff --git a/doc/guides/nics/sfc_efx.rst b/doc/guides/nics/sfc_efx.rst
index ba82b020932b..24459da33ea8 100644
--- a/doc/guides/nics/sfc_efx.rst
+++ b/doc/guides/nics/sfc_efx.rst
@@ -114,6 +114,10 @@ SFC EFX PMD has support for:
- Loopback
+- Configurable Rx CRC stripping (if running firmware variant supports it and
+ if NIC is configured with single PF per port and without VFs, otherwise
+ always stripped)
+
- SR-IOV PF
- Port representors (see :ref: switch_representation)
@@ -126,8 +130,6 @@ The features not yet supported include:
- Priority-based flow control
-- Configurable RX CRC stripping (always stripped)
-
- Header split on receive
- VLAN filtering
diff --git a/doc/guides/rel_notes/release_23_07.rst
b/doc/guides/rel_notes/release_23_07.rst
index 3db443d16018..738d35d9f7c6 100644
--- a/doc/guides/rel_notes/release_23_07.rst
+++ b/doc/guides/rel_notes/release_23_07.rst
@@ -145,6 +145,8 @@ New Features
* Added support for transfer flow action INDIRECT with subtype COUNT,
for aggregated statistics.
+ * Added support for keeping CRC.
+
* **Added vmxnet3 version 7 support.**
Added support for vmxnet3 version 7 which includes support
diff --git a/drivers/net/sfc/sfc.h b/drivers/net/sfc/sfc.h
index 89e2c46fa974..25cdeaa5cd25 100644
--- a/drivers/net/sfc/sfc.h
+++ b/drivers/net/sfc/sfc.h
@@ -74,6 +74,7 @@ struct sfc_port {
unsigned intflow_ctrl;
boolean_t flow_ctrl_autoneg;
+ boolean_t include_fcs;
size_t pdu;
/*
diff --git a/drivers/net/sfc/sfc_ef100_rx.c b/drivers/net/sfc/sfc_ef100_rx.c
index 37b754fa3305..5563bd9a0bd1 100644
--- a/drivers/net/sfc/sfc_ef100_rx.c
+++ b/drivers/net/sfc/sfc_ef100_rx.c
@@ -1004,7 +1004,7 @@ struct sfc_dp_rx sfc_ef100_rx = {
SFC_DP_RX_FEAT_FLOW_MARK |
SFC_DP_RX_FEAT_INTR |
SFC_DP_RX_FEAT_STATS,
- .dev_offload_capa = 0,
+ .dev_offload_capa = RTE_ETH_RX_OFFLOAD_KEEP_CRC,
.queue_offload_capa = RTE_ETH_RX_OFFLOAD_CHECKSUM |
RTE_ETH_RX_OFFLOAD_OUTER_IPV4_CKSUM |
RTE_ETH_RX_OFFLOAD_OUTER_UDP_CKSUM |
diff --git a/drivers/net/sfc/sfc_ef10_rx.c b/drivers/net/sfc/sfc_ef10_rx.c
index 7be224c9c412..30a320d0791c 100644
--- a/drivers/net/sfc/sfc_ef10_rx.c
+++ b/drivers/net/sfc/sfc_ef10_rx.c
@@ -825,7 +825,8 @@ struct sfc_dp_rx sfc_ef10_rx = {
SFC_DP_RX_FEAT_INTR,
.dev_offload_capa = RTE_ETH_RX_OFFLOAD_CHECKSUM |
RTE_ETH_RX_OFFLOAD_OUTER_IPV4_CKSUM |
- RTE_ETH_RX_OFFLOAD_RSS_HASH,
+ RTE_ETH_RX_OFFLOAD_RSS_HASH |
+ RTE_ETH_RX_OFFLOAD_KEEP_CRC,
.queue_offload_capa = RTE_ETH_RX_OFFLOAD_SCATTER,
.get_dev_info = sfc_ef10_rx_get_dev_info,
.qsize_up_rings = sfc_ef10_rx_qsize_up_rings,
diff --git a/drivers/net/sfc/sfc_ethdev.c b/drivers/net/sfc/sfc_ethdev.c
index f120ddc5a8d8..1efe64a36a7f 100644
--- a/drivers/net/sfc/sfc_ethdev.c
+++ b/drivers/net/sfc/sfc_ethdev.c
@@ -689,8 +689,13 @@ sfc_stats_get(struct rte_eth_dev *dev, struct
rte_eth_stats *stats)
sfc_adapter_lock(sa);
- if (have_dp_rx_stats)
+ if (have_dp_rx_stats) {
sfc_stats_get_dp_rx(sa, &stats->ipackets, &stats->ibytes);
+ if (dev->data->dev_c
Re: [PATCH v6 2/4] common/sfc_efx/base: add API to get installed filters count
On 6/22/23 14:31, Artemii Morozov wrote:
This API allows to get number of installed filters. This will
be used in the future patches.
Signed-off-by: Artemii Morozov
---
drivers/common/sfc_efx/base/ef10_filter.c | 20 +
drivers/common/sfc_efx/base/ef10_impl.h | 6 +
drivers/common/sfc_efx/base/efx_filter.c | 27 +++
drivers/common/sfc_efx/base/efx_impl.h| 7 ++
4 files changed, 60 insertions(+)
diff --git a/drivers/common/sfc_efx/base/ef10_filter.c
b/drivers/common/sfc_efx/base/ef10_filter.c
index d6940011c0..278502fb61 100644
--- a/drivers/common/sfc_efx/base/ef10_filter.c
+++ b/drivers/common/sfc_efx/base/ef10_filter.c
@@ -2113,6 +2113,26 @@ ef10_filter_reconfigure(
return (rc);
}
+ __checkReturn efx_rc_t
+ef10_filter_get_count(
+ __inefx_nic_t *enp,
+ __out uint32_t *count)
+{
+ ef10_filter_table_t *table = enp->en_filter.ef_ef10_filter_table;
+ uint32_t filters_count;
+
+ EFSYS_ASSERT(EFX_FAMILY_IS_EF100(enp) || EFX_FAMILY_IS_EF10(enp));
+ EFSYS_ASSERT(count != NULL);
+
+ filters_count = table->eft_unicst_filter_count +
+ table->eft_mulcst_filter_count +
+ table->eft_encap_filter_count;
I guess encap filters are orthogonal to Rx VLAN stripping. But may be it
is not a problem.
+
+ *count = filters_count;
+
+ return (0);
+}
+
void
ef10_filter_get_default_rxq(
__inefx_nic_t *enp,
diff --git a/drivers/common/sfc_efx/base/ef10_impl.h
b/drivers/common/sfc_efx/base/ef10_impl.h
index 877aedad45..914bba10ce 100644
--- a/drivers/common/sfc_efx/base/ef10_impl.h
+++ b/drivers/common/sfc_efx/base/ef10_impl.h
@@ -1347,6 +1347,12 @@ ef10_filter_reconfigure(
__in_ecount(6*count)uint8_t const *addrs,
__inuint32_t count);
+LIBEFX_INTERNAL
+extern __checkReturn efx_rc_t
+ef10_filter_get_count(
+ __inefx_nic_t *enp,
+ __out uint32_t *count);
+
LIBEFX_INTERNAL
externvoid
ef10_filter_get_default_rxq(
diff --git a/drivers/common/sfc_efx/base/efx_filter.c
b/drivers/common/sfc_efx/base/efx_filter.c
index 83c37ff859..a8b62ee88d 100644
--- a/drivers/common/sfc_efx/base/efx_filter.c
+++ b/drivers/common/sfc_efx/base/efx_filter.c
@@ -53,6 +53,7 @@ static const efx_filter_ops_t __efx_filter_siena_ops = {
siena_filter_delete,/* efo_delete */
siena_filter_supported_filters, /* efo_supported_filters */
NULL, /* efo_reconfigure */
+ NULL, /* efo_get_count */
};
#endif /* EFSYS_OPT_SIENA */
@@ -65,6 +66,7 @@ static const efx_filter_ops_t __efx_filter_ef10_ops = {
ef10_filter_delete, /* efo_delete */
ef10_filter_supported_filters, /* efo_supported_filters */
ef10_filter_reconfigure,/* efo_reconfigure */
+ ef10_filter_get_count, /* efo_get_count */
};
#endif /* EFX_OPTS_EF10() */
@@ -77,6 +79,7 @@ static const efx_filter_ops_t __efx_filter_rhead_ops = {
ef10_filter_delete, /* efo_delete */
ef10_filter_supported_filters, /* efo_supported_filters */
ef10_filter_reconfigure,/* efo_reconfigure */
+ ef10_filter_get_count, /* efo_get_count */
};
#endif /* EFSYS_OPT_RIVERHEAD */
@@ -306,6 +309,30 @@ efx_filter_reconfigure(
return (0);
+fail1:
+ EFSYS_PROBE1(fail1, efx_rc_t, rc);
+
+ return (rc);
+}
+
+ __checkReturn efx_rc_t
+efx_filter_get_count(
+ __inefx_nic_t *enp,
+ __out uint32_t *count)
+{
+ efx_rc_t rc;
+
+ EFSYS_ASSERT3U(enp->en_magic, ==, EFX_NIC_MAGIC);
+ EFSYS_ASSERT3U(enp->en_mod_flags, &, EFX_MOD_PROBE);
+ EFSYS_ASSERT3U(enp->en_mod_flags, &, EFX_MOD_FILTER);
+
+ if (enp->en_efop->efo_get_count != NULL) {
It looks bad to return 0, but do not fill in count.
IMHO, it should be an error if callback is not provided.
+ if ((rc = enp->en_efop->efo_get_count(enp, count)) != 0)
+ goto fail1;
+ }
+
+ return (0);
+
fail1:
EFSYS_PROBE1(fail1, efx_rc_t, rc);
diff --git a/drivers/common/sfc_efx/base/efx_impl.h b/drivers/common/sfc_efx/base/efx_impl.h
index 45e99d01c5..d657734bc5 100644
--- a/drivers/common/sfc_efx/base/efx_impl.h
+++ b/drivers/common/sfc_efx/base/efx_impl.h
@@ -288,6 +288,7 @@ typedef struct efx_filter_ops_s {
efx_rc_t(*efo_reconfigure)(efx_nic_t *, uint8_t const *,
boolean_t,
boolean_t, boolean_t, boolean_t,
uint8_t const *, uint32_t);
+ efx_rc_t(*efo_get_count)(efx_nic_t *, uint32_t *);
} efx_filter_ops_t;
LIBEFX_INTERNAL
@@ -302,6 +303,12 @@ efx_filter_reconfigure(
__in_ecount(6*count)uint8_t const *add
Re: [PATCH v6 3/4] common/sfc_efx/base: add support to enable VLAN stripping
On 6/22/23 14:31, Artemii Morozov wrote:
To enable VLAN stripping, two conditions must be met:
the corresponding flag must be set and the appropriate
Rx prefix should be requested.
VLAN stripping is supported on EF100.
Signed-off-by: Artemii Morozov
Reviewed-by: Ivan Malov
Reviewed-by: Andy Moreton
with minor style fix, see below:
Acked-by: Andrew Rybchenko
diff --git a/drivers/common/sfc_efx/base/ef10_filter.c
b/drivers/common/sfc_efx/base/ef10_filter.c
index 278502fb61..827b3e8f00 100644
--- a/drivers/common/sfc_efx/base/ef10_filter.c
+++ b/drivers/common/sfc_efx/base/ef10_filter.c
@@ -171,6 +171,7 @@ efx_mcdi_filter_op_add(
EFX_MCDI_DECLARE_BUF(payload, MC_CMD_FILTER_OP_V3_IN_LEN,
MC_CMD_FILTER_OP_EXT_OUT_LEN);
efx_filter_match_flags_t match_flags;
+ efx_port_t *epp = &(enp->en_port);
uint32_t port_id;
efx_rc_t rc;
@@ -338,6 +339,11 @@ efx_mcdi_filter_op_add(
FILTER_OP_V3_IN_MATCH_SET_FLAG, 1);
}
+ if (epp->ep_vlan_strip) {
+ MCDI_IN_SET_DWORD_FIELD(req, FILTER_OP_V3_IN_MATCH_ACTION_FLAGS,
+ FILTER_OP_V3_IN_MATCH_STRIP_VLAN, 1);
4 spaces alignment, please, on line continuation in libefx
+ }
+
efx_mcdi_execute(enp, &req);
if (req.emr_rc != 0) {
[snip]
Re: [PATCH v6 4/4] net/sfc: support VLAN stripping offload
On 6/22/23 14:31, Artemii Morozov wrote: Extract VLAN TCI provided by the HW in the prefix and put it to mbuf. VLAN stripping is supported for ef100 datapath only. This is device level offload. Signed-off-by: Artemii Morozov Reviewed-by: Viacheslav Galaktionov Reviewed-by: Ivan Malov Reviewed-by: Andy Moreton One, but important question below: diff --git a/drivers/net/sfc/sfc_ef100_rx.c b/drivers/net/sfc/sfc_ef100_rx.c index 37b754fa33..07381df5cf 100644 --- a/drivers/net/sfc/sfc_ef100_rx.c +++ b/drivers/net/sfc/sfc_ef100_rx.c [snip] @@ -892,6 +905,15 @@ sfc_ef100_rx_qstart(struct sfc_dp_rxq *dp_rxq, unsigned int evq_read_ptr, (rxq->flags & SFC_EF100_RXQ_INGRESS_MPORT)) return ENOTSUP; + /* +* Exclude the SFC_EF100_RXQ_VLAN_STRIPPED_TCI if offload was not requested +* or the prefix does not contain the corresponding field. +*/ + if (!((rxq->flags & SFC_EF100_RXQ_VLAN_STRIPPED_TCI) && Sorry, it is complicated to understand the condition. Please, avoid leading not. + ((unsup_rx_prefix_fields & + (1U << EFX_RX_PREFIX_FIELD_VLAN_STRIP_TCI)) == 0))) + rxq->flags &= ~SFC_EF100_RXQ_VLAN_STRIPPED_TCI; + I don't understand why do we need to check Rx prefix field presence here. We should not allow to enable VLAN stripping if Rx prefix field is not available. rxq->prefix_size = pinfo->erpl_length; rxq->rearm_data = sfc_ef100_mk_mbuf_rearm_data(rxq->dp.dpq.port_id, rxq->prefix_size); [snip]
RE: [PATCH] common/mlx5: adjust fork call with the new kernel API
Hi,
> -Original Message-
> From: Erez Ferber
> Sent: Wednesday, May 24, 2023 3:02 PM
> To: dev@dpdk.org
> Cc: Slava Ovsiienko ; Matan Azrad
> ; Raslan Darawsheh ; Erez Ferber
> ; sta...@dpdk.org
> Subject: [PATCH] common/mlx5: adjust fork call with the new kernel API
>
> From: Erez Ferber
>
> While doing process fork() the operating system remaps all the parent
> process's memory to the address space of the child process and activates
> the Copy-on-Write mechanics - it duplicates physical pages once memory
> writing happens in the child process. Sometimes memory duplication is
> not allowed - for example, if the page contains hardware queue
> descriptors. To handle similar issues the rdma-core library should be
> prepared for forking.
>
> The ibv_fork_init() prepares the library to track all the related memory
> and prevent it from forking using madvise() system API. This approach
> allows fork, but not all the memory is forked to the child process and,
> application should care not to touch pages where the parent application
> allocated the rdma-core objects.
>
> The newer kernels propose an option of copy-on-fork for DMA pages and
> tracking all the memory and disabling it for the forking is no longer
> needed. The new API routine ibv_is_fork_initialized() should be involved
> to decide if library initialization for forking is required.
>
> Fixes: 0e83b8e536 ("net/mlx5: move rdma-core calls to separate file")
> Cc: sta...@dpdk.org
> Signed-off-by: Erez Ferber
Patch applied to next-net-mlx,
Kindest regards,
Raslan Darawsheh
[PATCH v5] common/sfc_efx/base: fix Rx queue creation without RSS hash prefix
If the prefix for the RSS hash was not chosen the ENOTSUP error should
be returned.
Before this patch success was returned for this case causing Rx queue
creation to fail.
Fixing return value to indicate failure.
Fixes: f784cdc5cbb1 ("common/sfc_efx/base: provide control to deliver RSS hash")
Cc: sta...@dpdk.org
Signed-off-by: Artemii Morozov
Reviewed-by: Andy Moreton
Acked-by: Andrew Rybchenko
---
v5: update commit message
v4: add Cc: sta...@dpdk.org and transform rss to RSS
v3: update commit log as fix commit
v2: don't use capital letters in email
drivers/common/sfc_efx/base/efx_rx.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/drivers/common/sfc_efx/base/efx_rx.c
b/drivers/common/sfc_efx/base/efx_rx.c
index 68f42f5cac..61726a9f0b 100644
--- a/drivers/common/sfc_efx/base/efx_rx.c
+++ b/drivers/common/sfc_efx/base/efx_rx.c
@@ -937,8 +937,10 @@ efx_rx_qcreate_internal(
rss_hash_field =
&erplp->erpl_fields[EFX_RX_PREFIX_FIELD_RSS_HASH];
- if (rss_hash_field->erpfi_width_bits == 0)
+ if (rss_hash_field->erpfi_width_bits == 0) {
+ rc = ENOTSUP;
goto fail5;
+ }
}
enp->en_rx_qcount++;
--
2.34.1
Re: [PATCH v4 0/4] net/sfc: support KEEP_CRC offload
On 6/22/2023 12:47 PM, Denis Pryazhennikov wrote: > This patch series adds support for RTE_ETH_RX_OFFLOAD_KEEP_CRC > offload in the SFC driver. > > Changes in v2: > * Applied review notes; > * Fixed stats handling; > * Added missing release notes to [4/4]. > > Changes in v3: > * Fixed wrong indents in [1/4]. > > Changes in v4: > * Fixed "From" section in [3/4]. > > Denis Pryazhennikov (3): > common/sfc_efx/base: detect and report FCS include support > common/sfc_efx/base: add support for configure MAC to keep FCS > net/sfc: add configurable Rx CRC stripping > > Sandilya Bhagi (1): > common/sfc_efx/base: discover NIC partitioning mode > Series applied to dpdk-next-net/main, thanks.
Re: [PATCH v5] common/sfc_efx/base: fix Rx queue creation without RSS hash prefix
On 6/22/2023 1:31 PM, Artemii Morozov wrote:
> If the prefix for the RSS hash was not chosen the ENOTSUP error should
> be returned.
>
> Before this patch success was returned for this case causing Rx queue
> creation to fail.
>
> Fixing return value to indicate failure.
>
It looks like you used sample commit log as it is, so can you please
confirm this is the actual case (because this was my assumption without
really knowing the actual problem), problem is not crash or undefined
behavior because of 'erpp' was not set etc..?
> Fixes: f784cdc5cbb1 ("common/sfc_efx/base: provide control to deliver RSS
> hash")
> Cc: sta...@dpdk.org
>
> Signed-off-by: Artemii Morozov
> Reviewed-by: Andy Moreton
> Acked-by: Andrew Rybchenko
> ---
> v5: update commit message
>
> v4: add Cc: sta...@dpdk.org and transform rss to RSS
>
> v3: update commit log as fix commit
>
> v2: don't use capital letters in email
>
> drivers/common/sfc_efx/base/efx_rx.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/common/sfc_efx/base/efx_rx.c
> b/drivers/common/sfc_efx/base/efx_rx.c
> index 68f42f5cac..61726a9f0b 100644
> --- a/drivers/common/sfc_efx/base/efx_rx.c
> +++ b/drivers/common/sfc_efx/base/efx_rx.c
> @@ -937,8 +937,10 @@ efx_rx_qcreate_internal(
>
> rss_hash_field =
> &erplp->erpl_fields[EFX_RX_PREFIX_FIELD_RSS_HASH];
> - if (rss_hash_field->erpfi_width_bits == 0)
> + if (rss_hash_field->erpfi_width_bits == 0) {
> + rc = ENOTSUP;
> goto fail5;
> + }
> }
>
> enp->en_rx_qcount++;
[PATCH] examples/l2fwd-cat: fix external build
From: David Marchand
cpu_set_t definition requires _GNU_SOURCE.
Fixes: e0473c6d5b18 ("eal: fix build with musl")
Cc: sta...@dpdk.org
Signed-off-by: David Marchand
---
This patch was missing in the patchset to make all examples compile.
It will be merged with others from:
"Test examples compilation externally"
---
examples/l2fwd-cat/Makefile | 1 +
1 file changed, 1 insertion(+)
diff --git a/examples/l2fwd-cat/Makefile b/examples/l2fwd-cat/Makefile
index 23a09550a4..d06053451a 100644
--- a/examples/l2fwd-cat/Makefile
+++ b/examples/l2fwd-cat/Makefile
@@ -35,6 +35,7 @@ endif
endif
CFLAGS += -DALLOW_EXPERIMENTAL_API
+CFLAGS += -D_GNU_SOURCE
LDFLAGS += -lpqos
--
2.41.0
Re: [PATCH v5] common/sfc_efx/base: fix Rx queue creation without RSS hash prefix
On 6/22/23 17:13, Ferruh Yigit wrote:
On 6/22/2023 1:31 PM, Artemii Morozov wrote:
If the prefix for the RSS hash was not chosen the ENOTSUP error should
be returned.
Before this patch success was returned for this case causing Rx queue
creation to fail.
Fixing return value to indicate failure.
It looks like you used sample commit log as it is, so can you please
confirm this is the actual case (because this was my assumption without
really knowing the actual problem), problem is not crash or undefined
behavior because of 'erpp' was not set etc..?
The main problem is that if the RSS prefix was requested and an error
occurred when creating the Rx queue, then this is ignored. Now the
commit message reflects this.
Fixes: f784cdc5cbb1 ("common/sfc_efx/base: provide control to deliver RSS hash")
Cc: sta...@dpdk.org
Signed-off-by: Artemii Morozov
Reviewed-by: Andy Moreton
Acked-by: Andrew Rybchenko
---
v5: update commit message
v4: add Cc: sta...@dpdk.org and transform rss to RSS
v3: update commit log as fix commit
v2: don't use capital letters in email
drivers/common/sfc_efx/base/efx_rx.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/drivers/common/sfc_efx/base/efx_rx.c
b/drivers/common/sfc_efx/base/efx_rx.c
index 68f42f5cac..61726a9f0b 100644
--- a/drivers/common/sfc_efx/base/efx_rx.c
+++ b/drivers/common/sfc_efx/base/efx_rx.c
@@ -937,8 +937,10 @@ efx_rx_qcreate_internal(
rss_hash_field =
&erplp->erpl_fields[EFX_RX_PREFIX_FIELD_RSS_HASH];
- if (rss_hash_field->erpfi_width_bits == 0)
+ if (rss_hash_field->erpfi_width_bits == 0) {
+ rc = ENOTSUP;
goto fail5;
+ }
}
enp->en_rx_qcount++;
Re: [PATCH v3 0/4] Test examples compilation externally
20/06/2023 16:07, David Marchand: > As DPDK provides examples compiled with makefiles, we need some tests in > the CI. So far, a few maintainers have been testing them but a simple > issue has been missed for some time and there was no way to try to build > all examples that were built with meson. > > Additionnally, this series can help in identify issues in public headers > that the current headers check can not catch as it relies on the meson > framework for finding headers include path (which points at sources). > > Changes since v2: > - tweaked patch 2 commitlog, > - dropped -e sed option in patch 3/4, > > Changes since v1: > - reworked built examples discovery, Applied with the additional patch for l2fwd-cat, thanks.
Re: [PATCH] examples/l2fwd-cat: fix external build
22/06/2023 15:48, Thomas Monjalon:
> From: David Marchand
>
> cpu_set_t definition requires _GNU_SOURCE.
>
> Fixes: e0473c6d5b18 ("eal: fix build with musl")
> Cc: sta...@dpdk.org
>
> Signed-off-by: David Marchand
> ---
> This patch was missing in the patchset to make all examples compile.
> It will be merged with others from:
> "Test examples compilation externally"
Tested-by: Thomas Monjalon
and applied.
[PATCH 0/7] expand list of optional libraries
DPDK still has many libraries which cannot be disabled as part of a build. With the ongoing work to make it easier to only build a subset of the libraries in DPDK, we can also work to expand the list of libraries which can be enabled/disabled as desired. This patch addresses a number of the "low-hanging fruit" libraries, where only the unit test builds need minor changes to support making the library optional. The rest of the build system is already well set up for selective disabling of libraries. For better support of enabling components, especially those more integrated into DPDK unit tests, rework of the test meson.build file is likely needed. For example, it could probably be better rewritten to use a dictionary of files and the dependencies of each file, and the unit test commands each provides. However, such rework is a significant effort, and outside the scope of this patchset. Bruce Richardson (7): build: make most device classes optional build: make membership library optional build: make bpf library optional build: make efd library optional build: make distributor library optional build: make fragmentation libary optional build: make reorder library optional app/test/meson.build | 81 +++- lib/meson.build | 13 +++ 2 files changed, 63 insertions(+), 31 deletions(-) -- 2.39.2
[PATCH 1/7] build: make most device classes optional
Apart from ethdev and cryptodev, which have lots of components and tests
which depend on them, we can make the device class libraries optional
without too much work.
This patch marks:
* bbdev,
* compressdev,
* dmadev,
* eventdev,
* mldev,
* rawdev,
* regexdev
optional, and ensures that DPDK - including tests - can be built with
these components disabled.
Signed-off-by: Bruce Richardson
---
app/test/meson.build | 33 +
lib/meson.build | 7 +++
2 files changed, 28 insertions(+), 12 deletions(-)
diff --git a/app/test/meson.build b/app/test/meson.build
index d0fabcbb8b..780005f320 100644
--- a/app/test/meson.build
+++ b/app/test/meson.build
@@ -41,19 +41,12 @@ test_sources = files(
'test_devargs.c',
'test_distributor.c',
'test_distributor_perf.c',
-'test_dmadev.c',
-'test_dmadev_api.c',
'test_eal_flags.c',
'test_eal_fs.c',
'test_efd.c',
'test_efd_perf.c',
'test_errno.c',
'test_ethdev_link.c',
-'test_event_crypto_adapter.c',
-'test_event_eth_rx_adapter.c',
-'test_event_ring.c',
-'test_event_timer_adapter.c',
-'test_eventdev.c',
'test_external_mem.c',
'test_fbarray.c',
'test_fib.c',
@@ -105,7 +98,6 @@ test_sources = files(
'test_power_intel_uncore.c',
'test_prefetch.c',
'test_rand_perf.c',
-'test_rawdev.c',
'test_rcu_qsbr.c',
'test_rcu_qsbr_perf.c',
'test_reciprocal_division.c',
@@ -129,7 +121,6 @@ test_sources = files(
'test_sched.c',
'test_security.c',
'test_security_inline_macsec.c',
-'test_security_inline_proto.c',
'test_seqlock.c',
'test_service_cores.c',
'test_spinlock.c',
@@ -188,7 +179,6 @@ fast_tests = [
['eal_fs_autotest', true, true],
['errno_autotest', true, true],
['ethdev_link_status', true, true],
-['event_ring_autotest', true, true],
['fib_autotest', true, true],
['fib6_autotest', true, true],
['func_reentrancy_autotest', false, true],
@@ -234,7 +224,6 @@ fast_tests = [
['version_autotest', true, true],
['crc_autotest', true, true],
['distributor_autotest', false, true],
-['eventdev_common_autotest', true, true],
['fbarray_autotest', true, true],
['hash_readwrite_func_autotest', false, true],
['ipsec_autotest', true, true],
@@ -321,7 +310,6 @@ driver_test_names = [
'cryptodev_sw_snow3g_autotest',
'cryptodev_sw_zuc_autotest',
'cryptodev_uadk_autotest',
-'dmadev_autotest',
]
dump_test_names = []
@@ -359,6 +347,25 @@ if dpdk_conf.has('RTE_EVENT_SKELETON')
test_deps += 'event_skeleton'
endif
+if dpdk_conf.has('RTE_LIB_DMADEV')
+test_sources += ['test_dmadev.c', 'test_dmadev_api.c']
+driver_test_names += 'dmadev_autotest'
+endif
+if dpdk_conf.has('RTE_LIB_EVENTDEV')
+test_sources += [
+'test_event_eth_rx_adapter.c',
+'test_event_ring.c',
+'test_event_timer_adapter.c',
+'test_eventdev.c',
+]
+fast_tests += [
+['event_ring_autotest', true, true],
+['eventdev_common_autotest', true, true],
+]
+if dpdk_conf.has('RTE_LIB_CRYPTODEV')
+test_sources += 'test_event_crypto_adapter.c'
+endif
+endif
if dpdk_conf.has('RTE_LIB_FLOW_CLASSIFY')
test_sources += 'test_flow_classify.c'
fast_tests += [['flow_classify_autotest', false, true]]
@@ -403,6 +410,7 @@ if dpdk_conf.has('RTE_LIB_EVENTDEV') and
dpdk_conf.has('RTE_NET_RING')
test_sources += 'test_pmd_ring.c'
test_sources += 'test_event_eth_tx_adapter.c'
test_sources += 'sample_packet_forward.c'
+test_sources += 'test_security_inline_proto.c'
fast_tests += [['ring_pmd_autotest', true, true]]
perf_test_names += 'ring_pmd_perf_autotest'
fast_tests += [['event_eth_tx_adapter_autotest', false, true]]
@@ -425,6 +433,7 @@ if dpdk_conf.has('RTE_NET_NULL')
fast_tests += [['vdev_autotest', true, true]]
endif
if dpdk_conf.has('RTE_RAW_SKELETON')
+test_sources += 'test_rawdev.c'
test_deps += 'raw_skeleton'
fast_tests += [['rawdev_autotest', true, true]]
endif
diff --git a/lib/meson.build b/lib/meson.build
index 9677239236..679a81f62f 100644
--- a/lib/meson.build
+++ b/lib/meson.build
@@ -68,8 +68,12 @@ libraries = [
]
optional_libs = [
+'bbdev',
'bitratestats',
'cfgfile',
+'compressdev',
+'dmadev',
+'eventdev',
'flow_classify',
'gpudev',
'gro',
@@ -78,11 +82,14 @@ optional_libs = [
'jobstats',
'latencystats',
'metrics',
+'mldev',
'node',
'pdump',
'pipeline',
'port',
'power',
+'rawdev',
+'regexdev',
'table',
'vhost',
]
--
2.39.2
[PATCH 2/7] build: make membership library optional
Signed-off-by: Bruce Richardson
---
app/test/meson.build | 9 +
lib/meson.build | 1 +
2 files changed, 6 insertions(+), 4 deletions(-)
diff --git a/app/test/meson.build b/app/test/meson.build
index 780005f320..bed6acfdaf 100644
--- a/app/test/meson.build
+++ b/app/test/meson.build
@@ -78,8 +78,6 @@ test_sources = files(
'test_malloc.c',
'test_malloc_perf.c',
'test_mbuf.c',
-'test_member.c',
-'test_member_perf.c',
'test_memcpy.c',
'test_memcpy_perf.c',
'test_memory.c',
@@ -229,7 +227,6 @@ fast_tests = [
['ipsec_autotest', true, true],
['kni_autotest', false, true],
['kvargs_autotest', true, true],
-['member_autotest', true, true],
['power_cpufreq_autotest', false, true],
['power_autotest', true, true],
['power_kvm_vm_autotest', false, true],
@@ -267,7 +264,6 @@ perf_test_names = [
'timer_racecond_autotest',
'efd_autotest',
'hash_functions_autotest',
-'member_perf_autotest',
'efd_perf_autotest',
'lpm6_perf_autotest',
'rib6_slow_autotest',
@@ -374,6 +370,11 @@ if dpdk_conf.has('RTE_LIB_METRICS')
test_sources += ['test_metrics.c']
fast_tests += [['metrics_autotest', true, true]]
endif
+if dpdk_conf.has('RTE_LIB_MEMBER')
+test_sources += ['test_member.c', 'test_member_perf.c']
+fast_tests += [['member_autotest', true, true]]
+perf_test_names += 'member_perf_autotest'
+endif
if not is_windows and dpdk_conf.has('RTE_LIB_TELEMETRY')
test_sources += ['test_telemetry_json.c', 'test_telemetry_data.c']
fast_tests += [['telemetry_json_autotest', true, true]]
diff --git a/lib/meson.build b/lib/meson.build
index 679a81f62f..e46b141657 100644
--- a/lib/meson.build
+++ b/lib/meson.build
@@ -81,6 +81,7 @@ optional_libs = [
'kni',
'jobstats',
'latencystats',
+'member',
'metrics',
'mldev',
'node',
--
2.39.2
[PATCH 3/7] build: make bpf library optional
Signed-off-by: Bruce Richardson
---
app/test/meson.build | 10 +++---
lib/meson.build | 1 +
2 files changed, 8 insertions(+), 3 deletions(-)
diff --git a/app/test/meson.build b/app/test/meson.build
index bed6acfdaf..d8ad8b1d97 100644
--- a/app/test/meson.build
+++ b/app/test/meson.build
@@ -15,7 +15,6 @@ test_sources = files(
'test_barrier.c',
'test_bitops.c',
'test_bitmap.c',
-'test_bpf.c',
'test_byteorder.c',
'test_cksum.c',
'test_cksum_perf.c',
@@ -152,8 +151,6 @@ fast_tests = [
['acl_autotest', true, true],
['atomic_autotest', false, true],
['bitmap_autotest', true, true],
-['bpf_autotest', true, true],
-['bpf_convert_autotest', true, true],
['bitops_autotest', true, true],
['byteorder_autotest', true, true],
['cksum_autotest', true, true],
@@ -343,6 +340,13 @@ if dpdk_conf.has('RTE_EVENT_SKELETON')
test_deps += 'event_skeleton'
endif
+if dpdk_conf.has('RTE_LIB_BPF')
+test_sources += 'test_bpf.c'
+fast_tests += [
+['bpf_autotest', true, true],
+['bpf_convert_autotest', true, true],
+]
+endif
if dpdk_conf.has('RTE_LIB_DMADEV')
test_sources += ['test_dmadev.c', 'test_dmadev_api.c']
driver_test_names += 'dmadev_autotest'
diff --git a/lib/meson.build b/lib/meson.build
index e46b141657..105f167d45 100644
--- a/lib/meson.build
+++ b/lib/meson.build
@@ -70,6 +70,7 @@ libraries = [
optional_libs = [
'bbdev',
'bitratestats',
+'bpf',
'cfgfile',
'compressdev',
'dmadev',
--
2.39.2
[PATCH 4/7] build: make efd library optional
Signed-off-by: Bruce Richardson
---
app/test/meson.build | 8
lib/meson.build | 1 +
2 files changed, 5 insertions(+), 4 deletions(-)
diff --git a/app/test/meson.build b/app/test/meson.build
index d8ad8b1d97..ab5bd370bf 100644
--- a/app/test/meson.build
+++ b/app/test/meson.build
@@ -42,8 +42,6 @@ test_sources = files(
'test_distributor_perf.c',
'test_eal_flags.c',
'test_eal_fs.c',
-'test_efd.c',
-'test_efd_perf.c',
'test_errno.c',
'test_ethdev_link.c',
'test_external_mem.c',
@@ -259,9 +257,7 @@ perf_test_names = [
'barrier_autotest',
'hash_multiwriter_autotest',
'timer_racecond_autotest',
-'efd_autotest',
'hash_functions_autotest',
-'efd_perf_autotest',
'lpm6_perf_autotest',
'rib6_slow_autotest',
'fib6_slow_autotest',
@@ -370,6 +366,10 @@ if dpdk_conf.has('RTE_LIB_FLOW_CLASSIFY')
test_sources += 'test_flow_classify.c'
fast_tests += [['flow_classify_autotest', false, true]]
endif
+if dpdk_conf.has('RTE_LIB_EFD')
+test_sources += ['test_efd.c', 'test_efd_perf.c']
+perf_test_names += ['efd_autotest', 'efd_perf_autotest']
+endif
if dpdk_conf.has('RTE_LIB_METRICS')
test_sources += ['test_metrics.c']
fast_tests += [['metrics_autotest', true, true]]
diff --git a/lib/meson.build b/lib/meson.build
index 105f167d45..7de3b0d32c 100644
--- a/lib/meson.build
+++ b/lib/meson.build
@@ -74,6 +74,7 @@ optional_libs = [
'cfgfile',
'compressdev',
'dmadev',
+'efd',
'eventdev',
'flow_classify',
'gpudev',
--
2.39.2
[PATCH 5/7] build: make distributor library optional
Signed-off-by: Bruce Richardson
---
app/test/meson.build | 9 +
lib/meson.build | 1 +
2 files changed, 6 insertions(+), 4 deletions(-)
diff --git a/app/test/meson.build b/app/test/meson.build
index ab5bd370bf..749e929945 100644
--- a/app/test/meson.build
+++ b/app/test/meson.build
@@ -38,8 +38,6 @@ test_sources = files(
'test_cycles.c',
'test_debug.c',
'test_devargs.c',
-'test_distributor.c',
-'test_distributor_perf.c',
'test_eal_flags.c',
'test_eal_fs.c',
'test_errno.c',
@@ -216,7 +214,6 @@ fast_tests = [
['user_delay_us', true, true],
['version_autotest', true, true],
['crc_autotest', true, true],
-['distributor_autotest', false, true],
['fbarray_autotest', true, true],
['hash_readwrite_func_autotest', false, true],
['ipsec_autotest', true, true],
@@ -265,7 +262,6 @@ perf_test_names = [
'rcu_qsbr_perf_autotest',
'red_perf',
'pie_perf',
-'distributor_perf_autotest',
'pmd_perf_autotest',
'service_perf_autotest',
'stack_perf_autotest',
@@ -366,6 +362,11 @@ if dpdk_conf.has('RTE_LIB_FLOW_CLASSIFY')
test_sources += 'test_flow_classify.c'
fast_tests += [['flow_classify_autotest', false, true]]
endif
+if dpdk_conf.has('RTE_LIB_DISTRIBUTOR')
+test_sources += ['test_distributor.c', 'test_distributor_perf.c']
+fast_tests += [['distributor_autotest', false, true]]
+perf_test_names += 'distributor_perf_autotest'
+endif
if dpdk_conf.has('RTE_LIB_EFD')
test_sources += ['test_efd.c', 'test_efd_perf.c']
perf_test_names += ['efd_autotest', 'efd_perf_autotest']
diff --git a/lib/meson.build b/lib/meson.build
index 7de3b0d32c..ce147a0766 100644
--- a/lib/meson.build
+++ b/lib/meson.build
@@ -73,6 +73,7 @@ optional_libs = [
'bpf',
'cfgfile',
'compressdev',
+'distributor',
'dmadev',
'efd',
'eventdev',
--
2.39.2
[PATCH 6/7] build: make fragmentation libary optional
Signed-off-by: Bruce Richardson
---
app/test/meson.build | 6 --
lib/meson.build | 1 +
2 files changed, 5 insertions(+), 2 deletions(-)
diff --git a/app/test/meson.build b/app/test/meson.build
index 749e929945..e8dbbadd5e 100644
--- a/app/test/meson.build
+++ b/app/test/meson.build
@@ -58,7 +58,6 @@ test_sources = files(
'test_hash_perf.c',
'test_hash_readwrite_lf_perf.c',
'test_interrupts.c',
-'test_ipfrag.c',
'test_ipsec.c',
'test_ipsec_sad.c',
'test_ipsec_perf.c',
@@ -175,7 +174,6 @@ fast_tests = [
['func_reentrancy_autotest', false, true],
['hash_autotest', true, true],
['interrupt_autotest', true, true],
-['ipfrag_autotest', false, true],
['lcores_autotest', true, true],
['logs_autotest', true, true],
['lpm_autotest', true, true],
@@ -371,6 +369,10 @@ if dpdk_conf.has('RTE_LIB_EFD')
test_sources += ['test_efd.c', 'test_efd_perf.c']
perf_test_names += ['efd_autotest', 'efd_perf_autotest']
endif
+if dpdk_conf.has('RTE_IP_FRAG')
+test_sources += 'test_ipfrag.c'
+fast_tests += [['ipfrag_autotest', false, true]]
+endif
if dpdk_conf.has('RTE_LIB_METRICS')
test_sources += ['test_metrics.c']
fast_tests += [['metrics_autotest', true, true]]
diff --git a/lib/meson.build b/lib/meson.build
index ce147a0766..7fb0cf4c4e 100644
--- a/lib/meson.build
+++ b/lib/meson.build
@@ -82,6 +82,7 @@ optional_libs = [
'gro',
'gso',
'kni',
+'ip_frag',
'jobstats',
'latencystats',
'member',
--
2.39.2
[PATCH 7/7] build: make reorder library optional
Signed-off-by: Bruce Richardson
---
app/test/meson.build | 6 --
lib/meson.build | 1 +
2 files changed, 5 insertions(+), 2 deletions(-)
diff --git a/app/test/meson.build b/app/test/meson.build
index e8dbbadd5e..13bd752002 100644
--- a/app/test/meson.build
+++ b/app/test/meson.build
@@ -96,7 +96,6 @@ test_sources = files(
'test_reciprocal_division_perf.c',
'test_red.c',
'test_pie.c',
-'test_reorder.c',
'test_rib.c',
'test_rib6.c',
'test_ring.c',
@@ -221,7 +220,6 @@ fast_tests = [
['power_autotest', true, true],
['power_kvm_vm_autotest', false, true],
['power_intel_uncore_autotest', true, true],
-['reorder_autotest', true, true],
['service_autotest', true, true],
['thash_autotest', true, true],
['threads_autotest', true, true],
@@ -382,6 +380,10 @@ if dpdk_conf.has('RTE_LIB_MEMBER')
fast_tests += [['member_autotest', true, true]]
perf_test_names += 'member_perf_autotest'
endif
+if dpdk_conf.has('RTE_LIB_REORDER')
+test_sources += 'test_reorder.c'
+fast_tests += [['reorder_autotest', true, true]]
+endif
if not is_windows and dpdk_conf.has('RTE_LIB_TELEMETRY')
test_sources += ['test_telemetry_json.c', 'test_telemetry_data.c']
fast_tests += [['telemetry_json_autotest', true, true]]
diff --git a/lib/meson.build b/lib/meson.build
index 7fb0cf4c4e..be4c6113fe 100644
--- a/lib/meson.build
+++ b/lib/meson.build
@@ -95,6 +95,7 @@ optional_libs = [
'power',
'rawdev',
'regexdev',
+'reorder',
'table',
'vhost',
]
--
2.39.2
RE: [PATCH 0/7] expand list of optional libraries
> From: Bruce Richardson [mailto:bruce.richard...@intel.com] > Sent: Thursday, 22 June 2023 15.49 > > DPDK still has many libraries which cannot be disabled as part of a > build. With the ongoing work to make it easier to only build a subset > of the libraries in DPDK, we can also work to expand the list of > libraries which can be enabled/disabled as desired. > > This patch addresses a number of the "low-hanging fruit" libraries, > where only the unit test builds need minor changes to support > making the library optional. The rest of the build system is already > well set up for selective disabling of libraries. > > For better support of enabling components, especially those more > integrated into DPDK unit tests, rework of the test meson.build file > is likely needed. For example, it could probably be better rewritten > to use a dictionary of files and the dependencies of each file, and > the unit test commands each provides. However, such rework is a > significant effort, and outside the scope of this patchset. Thank you, Bruce. Series-acked-by: Morten Brørup
[PATCH v2 0/7] expand list of optional libraries
DPDK still has many libraries which cannot be disabled as part of a build. With the ongoing work to make it easier to only build a subset of the libraries in DPDK, we can also work to expand the list of libraries which can be enabled/disabled as desired. This patch addresses a number of the "low-hanging fruit" libraries, where only the unit test builds need minor changes to support making the library optional. The rest of the build system is already well set up for selective disabling of libraries. For better support of enabling components, especially those more integrated into DPDK unit tests, rework of the test meson.build file is likely needed. For example, it could probably be better rewritten to use a dictionary of files and the dependencies of each file, and the unit test commands each provides. However, such rework is a significant effort, and outside the scope of this patchset. V2: fix checkpatch issues, since checkpatch doesn't like empty commit messages (even if the title is pretty self-explanatory!) Bruce Richardson (7): build: make most device classes optional build: make membership library optional build: make bpf library optional build: make efd library optional build: make distributor library optional build: make fragmentation library optional build: make reorder library optional app/test/meson.build | 81 +++- lib/meson.build | 13 +++ 2 files changed, 63 insertions(+), 31 deletions(-) -- 2.39.2
[PATCH v2 1/7] build: make most device classes optional
Apart from ethdev and cryptodev, which have lots of components and tests
which depend on them, we can make the device class libraries optional
without too much work.
This patch marks:
* bbdev,
* compressdev,
* dmadev,
* eventdev,
* mldev,
* rawdev,
* regexdev
optional, and ensures that DPDK - including tests - can be built with
these components disabled.
Signed-off-by: Bruce Richardson
Acked-by: Morten Brørup
---
app/test/meson.build | 33 +
lib/meson.build | 7 +++
2 files changed, 28 insertions(+), 12 deletions(-)
diff --git a/app/test/meson.build b/app/test/meson.build
index d0fabcbb8b..780005f320 100644
--- a/app/test/meson.build
+++ b/app/test/meson.build
@@ -41,19 +41,12 @@ test_sources = files(
'test_devargs.c',
'test_distributor.c',
'test_distributor_perf.c',
-'test_dmadev.c',
-'test_dmadev_api.c',
'test_eal_flags.c',
'test_eal_fs.c',
'test_efd.c',
'test_efd_perf.c',
'test_errno.c',
'test_ethdev_link.c',
-'test_event_crypto_adapter.c',
-'test_event_eth_rx_adapter.c',
-'test_event_ring.c',
-'test_event_timer_adapter.c',
-'test_eventdev.c',
'test_external_mem.c',
'test_fbarray.c',
'test_fib.c',
@@ -105,7 +98,6 @@ test_sources = files(
'test_power_intel_uncore.c',
'test_prefetch.c',
'test_rand_perf.c',
-'test_rawdev.c',
'test_rcu_qsbr.c',
'test_rcu_qsbr_perf.c',
'test_reciprocal_division.c',
@@ -129,7 +121,6 @@ test_sources = files(
'test_sched.c',
'test_security.c',
'test_security_inline_macsec.c',
-'test_security_inline_proto.c',
'test_seqlock.c',
'test_service_cores.c',
'test_spinlock.c',
@@ -188,7 +179,6 @@ fast_tests = [
['eal_fs_autotest', true, true],
['errno_autotest', true, true],
['ethdev_link_status', true, true],
-['event_ring_autotest', true, true],
['fib_autotest', true, true],
['fib6_autotest', true, true],
['func_reentrancy_autotest', false, true],
@@ -234,7 +224,6 @@ fast_tests = [
['version_autotest', true, true],
['crc_autotest', true, true],
['distributor_autotest', false, true],
-['eventdev_common_autotest', true, true],
['fbarray_autotest', true, true],
['hash_readwrite_func_autotest', false, true],
['ipsec_autotest', true, true],
@@ -321,7 +310,6 @@ driver_test_names = [
'cryptodev_sw_snow3g_autotest',
'cryptodev_sw_zuc_autotest',
'cryptodev_uadk_autotest',
-'dmadev_autotest',
]
dump_test_names = []
@@ -359,6 +347,25 @@ if dpdk_conf.has('RTE_EVENT_SKELETON')
test_deps += 'event_skeleton'
endif
+if dpdk_conf.has('RTE_LIB_DMADEV')
+test_sources += ['test_dmadev.c', 'test_dmadev_api.c']
+driver_test_names += 'dmadev_autotest'
+endif
+if dpdk_conf.has('RTE_LIB_EVENTDEV')
+test_sources += [
+'test_event_eth_rx_adapter.c',
+'test_event_ring.c',
+'test_event_timer_adapter.c',
+'test_eventdev.c',
+]
+fast_tests += [
+['event_ring_autotest', true, true],
+['eventdev_common_autotest', true, true],
+]
+if dpdk_conf.has('RTE_LIB_CRYPTODEV')
+test_sources += 'test_event_crypto_adapter.c'
+endif
+endif
if dpdk_conf.has('RTE_LIB_FLOW_CLASSIFY')
test_sources += 'test_flow_classify.c'
fast_tests += [['flow_classify_autotest', false, true]]
@@ -403,6 +410,7 @@ if dpdk_conf.has('RTE_LIB_EVENTDEV') and
dpdk_conf.has('RTE_NET_RING')
test_sources += 'test_pmd_ring.c'
test_sources += 'test_event_eth_tx_adapter.c'
test_sources += 'sample_packet_forward.c'
+test_sources += 'test_security_inline_proto.c'
fast_tests += [['ring_pmd_autotest', true, true]]
perf_test_names += 'ring_pmd_perf_autotest'
fast_tests += [['event_eth_tx_adapter_autotest', false, true]]
@@ -425,6 +433,7 @@ if dpdk_conf.has('RTE_NET_NULL')
fast_tests += [['vdev_autotest', true, true]]
endif
if dpdk_conf.has('RTE_RAW_SKELETON')
+test_sources += 'test_rawdev.c'
test_deps += 'raw_skeleton'
fast_tests += [['rawdev_autotest', true, true]]
endif
diff --git a/lib/meson.build b/lib/meson.build
index 9677239236..679a81f62f 100644
--- a/lib/meson.build
+++ b/lib/meson.build
@@ -68,8 +68,12 @@ libraries = [
]
optional_libs = [
+'bbdev',
'bitratestats',
'cfgfile',
+'compressdev',
+'dmadev',
+'eventdev',
'flow_classify',
'gpudev',
'gro',
@@ -78,11 +82,14 @@ optional_libs = [
'jobstats',
'latencystats',
'metrics',
+'mldev',
'node',
'pdump',
'pipeline',
'port',
'power',
+'rawdev',
+'regexdev',
'table',
'vho
[PATCH v2 2/7] build: make membership library optional
This library is not essential for most DPDK uses, so mark it as optional
in the build config.
Signed-off-by: Bruce Richardson
Acked-by: Morten Brørup
---
app/test/meson.build | 9 +
lib/meson.build | 1 +
2 files changed, 6 insertions(+), 4 deletions(-)
diff --git a/app/test/meson.build b/app/test/meson.build
index 780005f320..bed6acfdaf 100644
--- a/app/test/meson.build
+++ b/app/test/meson.build
@@ -78,8 +78,6 @@ test_sources = files(
'test_malloc.c',
'test_malloc_perf.c',
'test_mbuf.c',
-'test_member.c',
-'test_member_perf.c',
'test_memcpy.c',
'test_memcpy_perf.c',
'test_memory.c',
@@ -229,7 +227,6 @@ fast_tests = [
['ipsec_autotest', true, true],
['kni_autotest', false, true],
['kvargs_autotest', true, true],
-['member_autotest', true, true],
['power_cpufreq_autotest', false, true],
['power_autotest', true, true],
['power_kvm_vm_autotest', false, true],
@@ -267,7 +264,6 @@ perf_test_names = [
'timer_racecond_autotest',
'efd_autotest',
'hash_functions_autotest',
-'member_perf_autotest',
'efd_perf_autotest',
'lpm6_perf_autotest',
'rib6_slow_autotest',
@@ -374,6 +370,11 @@ if dpdk_conf.has('RTE_LIB_METRICS')
test_sources += ['test_metrics.c']
fast_tests += [['metrics_autotest', true, true]]
endif
+if dpdk_conf.has('RTE_LIB_MEMBER')
+test_sources += ['test_member.c', 'test_member_perf.c']
+fast_tests += [['member_autotest', true, true]]
+perf_test_names += 'member_perf_autotest'
+endif
if not is_windows and dpdk_conf.has('RTE_LIB_TELEMETRY')
test_sources += ['test_telemetry_json.c', 'test_telemetry_data.c']
fast_tests += [['telemetry_json_autotest', true, true]]
diff --git a/lib/meson.build b/lib/meson.build
index 679a81f62f..e46b141657 100644
--- a/lib/meson.build
+++ b/lib/meson.build
@@ -81,6 +81,7 @@ optional_libs = [
'kni',
'jobstats',
'latencystats',
+'member',
'metrics',
'mldev',
'node',
--
2.39.2
[PATCH v2 3/7] build: make bpf library optional
This library is not essential for most DPDK uses, so mark it as optional
in the build config.
Signed-off-by: Bruce Richardson
Acked-by: Morten Brørup
---
app/test/meson.build | 10 +++---
lib/meson.build | 1 +
2 files changed, 8 insertions(+), 3 deletions(-)
diff --git a/app/test/meson.build b/app/test/meson.build
index bed6acfdaf..d8ad8b1d97 100644
--- a/app/test/meson.build
+++ b/app/test/meson.build
@@ -15,7 +15,6 @@ test_sources = files(
'test_barrier.c',
'test_bitops.c',
'test_bitmap.c',
-'test_bpf.c',
'test_byteorder.c',
'test_cksum.c',
'test_cksum_perf.c',
@@ -152,8 +151,6 @@ fast_tests = [
['acl_autotest', true, true],
['atomic_autotest', false, true],
['bitmap_autotest', true, true],
-['bpf_autotest', true, true],
-['bpf_convert_autotest', true, true],
['bitops_autotest', true, true],
['byteorder_autotest', true, true],
['cksum_autotest', true, true],
@@ -343,6 +340,13 @@ if dpdk_conf.has('RTE_EVENT_SKELETON')
test_deps += 'event_skeleton'
endif
+if dpdk_conf.has('RTE_LIB_BPF')
+test_sources += 'test_bpf.c'
+fast_tests += [
+['bpf_autotest', true, true],
+['bpf_convert_autotest', true, true],
+]
+endif
if dpdk_conf.has('RTE_LIB_DMADEV')
test_sources += ['test_dmadev.c', 'test_dmadev_api.c']
driver_test_names += 'dmadev_autotest'
diff --git a/lib/meson.build b/lib/meson.build
index e46b141657..105f167d45 100644
--- a/lib/meson.build
+++ b/lib/meson.build
@@ -70,6 +70,7 @@ libraries = [
optional_libs = [
'bbdev',
'bitratestats',
+'bpf',
'cfgfile',
'compressdev',
'dmadev',
--
2.39.2
[PATCH v2 4/7] build: make efd library optional
This library is not essential for most DPDK uses, so mark it as optional
in the build config.
Signed-off-by: Bruce Richardson
Acked-by: Morten Brørup
---
app/test/meson.build | 8
lib/meson.build | 1 +
2 files changed, 5 insertions(+), 4 deletions(-)
diff --git a/app/test/meson.build b/app/test/meson.build
index d8ad8b1d97..ab5bd370bf 100644
--- a/app/test/meson.build
+++ b/app/test/meson.build
@@ -42,8 +42,6 @@ test_sources = files(
'test_distributor_perf.c',
'test_eal_flags.c',
'test_eal_fs.c',
-'test_efd.c',
-'test_efd_perf.c',
'test_errno.c',
'test_ethdev_link.c',
'test_external_mem.c',
@@ -259,9 +257,7 @@ perf_test_names = [
'barrier_autotest',
'hash_multiwriter_autotest',
'timer_racecond_autotest',
-'efd_autotest',
'hash_functions_autotest',
-'efd_perf_autotest',
'lpm6_perf_autotest',
'rib6_slow_autotest',
'fib6_slow_autotest',
@@ -370,6 +366,10 @@ if dpdk_conf.has('RTE_LIB_FLOW_CLASSIFY')
test_sources += 'test_flow_classify.c'
fast_tests += [['flow_classify_autotest', false, true]]
endif
+if dpdk_conf.has('RTE_LIB_EFD')
+test_sources += ['test_efd.c', 'test_efd_perf.c']
+perf_test_names += ['efd_autotest', 'efd_perf_autotest']
+endif
if dpdk_conf.has('RTE_LIB_METRICS')
test_sources += ['test_metrics.c']
fast_tests += [['metrics_autotest', true, true]]
diff --git a/lib/meson.build b/lib/meson.build
index 105f167d45..7de3b0d32c 100644
--- a/lib/meson.build
+++ b/lib/meson.build
@@ -74,6 +74,7 @@ optional_libs = [
'cfgfile',
'compressdev',
'dmadev',
+'efd',
'eventdev',
'flow_classify',
'gpudev',
--
2.39.2
[PATCH v2 5/7] build: make distributor library optional
This library is not essential for most DPDK uses, so mark it as optional
in the build config.
Signed-off-by: Bruce Richardson
Acked-by: Morten Brørup
---
app/test/meson.build | 9 +
lib/meson.build | 1 +
2 files changed, 6 insertions(+), 4 deletions(-)
diff --git a/app/test/meson.build b/app/test/meson.build
index ab5bd370bf..749e929945 100644
--- a/app/test/meson.build
+++ b/app/test/meson.build
@@ -38,8 +38,6 @@ test_sources = files(
'test_cycles.c',
'test_debug.c',
'test_devargs.c',
-'test_distributor.c',
-'test_distributor_perf.c',
'test_eal_flags.c',
'test_eal_fs.c',
'test_errno.c',
@@ -216,7 +214,6 @@ fast_tests = [
['user_delay_us', true, true],
['version_autotest', true, true],
['crc_autotest', true, true],
-['distributor_autotest', false, true],
['fbarray_autotest', true, true],
['hash_readwrite_func_autotest', false, true],
['ipsec_autotest', true, true],
@@ -265,7 +262,6 @@ perf_test_names = [
'rcu_qsbr_perf_autotest',
'red_perf',
'pie_perf',
-'distributor_perf_autotest',
'pmd_perf_autotest',
'service_perf_autotest',
'stack_perf_autotest',
@@ -366,6 +362,11 @@ if dpdk_conf.has('RTE_LIB_FLOW_CLASSIFY')
test_sources += 'test_flow_classify.c'
fast_tests += [['flow_classify_autotest', false, true]]
endif
+if dpdk_conf.has('RTE_LIB_DISTRIBUTOR')
+test_sources += ['test_distributor.c', 'test_distributor_perf.c']
+fast_tests += [['distributor_autotest', false, true]]
+perf_test_names += 'distributor_perf_autotest'
+endif
if dpdk_conf.has('RTE_LIB_EFD')
test_sources += ['test_efd.c', 'test_efd_perf.c']
perf_test_names += ['efd_autotest', 'efd_perf_autotest']
diff --git a/lib/meson.build b/lib/meson.build
index 7de3b0d32c..ce147a0766 100644
--- a/lib/meson.build
+++ b/lib/meson.build
@@ -73,6 +73,7 @@ optional_libs = [
'bpf',
'cfgfile',
'compressdev',
+'distributor',
'dmadev',
'efd',
'eventdev',
--
2.39.2
[PATCH v2 6/7] build: make fragmentation library optional
This library is not essential for most DPDK uses, so mark it as optional
in the build config.
Signed-off-by: Bruce Richardson
Acked-by: Morten Brørup
---
app/test/meson.build | 6 --
lib/meson.build | 1 +
2 files changed, 5 insertions(+), 2 deletions(-)
diff --git a/app/test/meson.build b/app/test/meson.build
index 749e929945..e8dbbadd5e 100644
--- a/app/test/meson.build
+++ b/app/test/meson.build
@@ -58,7 +58,6 @@ test_sources = files(
'test_hash_perf.c',
'test_hash_readwrite_lf_perf.c',
'test_interrupts.c',
-'test_ipfrag.c',
'test_ipsec.c',
'test_ipsec_sad.c',
'test_ipsec_perf.c',
@@ -175,7 +174,6 @@ fast_tests = [
['func_reentrancy_autotest', false, true],
['hash_autotest', true, true],
['interrupt_autotest', true, true],
-['ipfrag_autotest', false, true],
['lcores_autotest', true, true],
['logs_autotest', true, true],
['lpm_autotest', true, true],
@@ -371,6 +369,10 @@ if dpdk_conf.has('RTE_LIB_EFD')
test_sources += ['test_efd.c', 'test_efd_perf.c']
perf_test_names += ['efd_autotest', 'efd_perf_autotest']
endif
+if dpdk_conf.has('RTE_IP_FRAG')
+test_sources += 'test_ipfrag.c'
+fast_tests += [['ipfrag_autotest', false, true]]
+endif
if dpdk_conf.has('RTE_LIB_METRICS')
test_sources += ['test_metrics.c']
fast_tests += [['metrics_autotest', true, true]]
diff --git a/lib/meson.build b/lib/meson.build
index ce147a0766..7fb0cf4c4e 100644
--- a/lib/meson.build
+++ b/lib/meson.build
@@ -82,6 +82,7 @@ optional_libs = [
'gro',
'gso',
'kni',
+'ip_frag',
'jobstats',
'latencystats',
'member',
--
2.39.2
[PATCH v2 7/7] build: make reorder library optional
This library is not essential for most DPDK uses, so mark it as optional
in the build config.
Signed-off-by: Bruce Richardson
Acked-by: Morten Brørup
---
app/test/meson.build | 6 --
lib/meson.build | 1 +
2 files changed, 5 insertions(+), 2 deletions(-)
diff --git a/app/test/meson.build b/app/test/meson.build
index e8dbbadd5e..13bd752002 100644
--- a/app/test/meson.build
+++ b/app/test/meson.build
@@ -96,7 +96,6 @@ test_sources = files(
'test_reciprocal_division_perf.c',
'test_red.c',
'test_pie.c',
-'test_reorder.c',
'test_rib.c',
'test_rib6.c',
'test_ring.c',
@@ -221,7 +220,6 @@ fast_tests = [
['power_autotest', true, true],
['power_kvm_vm_autotest', false, true],
['power_intel_uncore_autotest', true, true],
-['reorder_autotest', true, true],
['service_autotest', true, true],
['thash_autotest', true, true],
['threads_autotest', true, true],
@@ -382,6 +380,10 @@ if dpdk_conf.has('RTE_LIB_MEMBER')
fast_tests += [['member_autotest', true, true]]
perf_test_names += 'member_perf_autotest'
endif
+if dpdk_conf.has('RTE_LIB_REORDER')
+test_sources += 'test_reorder.c'
+fast_tests += [['reorder_autotest', true, true]]
+endif
if not is_windows and dpdk_conf.has('RTE_LIB_TELEMETRY')
test_sources += ['test_telemetry_json.c', 'test_telemetry_data.c']
fast_tests += [['telemetry_json_autotest', true, true]]
diff --git a/lib/meson.build b/lib/meson.build
index 7fb0cf4c4e..be4c6113fe 100644
--- a/lib/meson.build
+++ b/lib/meson.build
@@ -95,6 +95,7 @@ optional_libs = [
'power',
'rawdev',
'regexdev',
+'reorder',
'table',
'vhost',
]
--
2.39.2
Re: [PATCH v5] common/sfc_efx/base: fix Rx queue creation without RSS hash prefix
On 6/22/2023 1:31 PM, Artemii Morozov wrote:
> If the prefix for the RSS hash was not chosen the ENOTSUP error should
> be returned.
>
> Before this patch success was returned for this case causing Rx queue
> creation to fail.
>
> Fixing return value to indicate failure.
>
> Fixes: f784cdc5cbb1 ("common/sfc_efx/base: provide control to deliver RSS
> hash")
> Cc: sta...@dpdk.org
>
> Signed-off-by: Artemii Morozov
> Reviewed-by: Andy Moreton
> Acked-by: Andrew Rybchenko
>
Applied to dpdk-next-net/main, thanks.
[PATCH v7 0/4] net/sfc: support VLAN stripping offload
This patch series adds VLAN stripping offload. Note that this offload are device level offload. v7: * raise an error if there is no callback for efo_get_count * fix alignment * remove the extra check v6: * highlight that efx_port_vlan_strip_set() must be called before any filter insertion * avoid an extra check if offload is not requested v5: * fixed problems with naming * fixed problems with abbreviations * fixed problems with isolated mode * fixed problems with consistency v4: * fix apply patch failure warning v3: * fix apply patch failure warning v2: * rebase patches on top of dpdk-next-net/main Artemii Morozov (4): common/sfc_efx/base: report VLAN stripping capability common/sfc_efx/base: add API to get installed filters count common/sfc_efx/base: add support to enable VLAN stripping net/sfc: support VLAN stripping offload doc/guides/nics/sfc_efx.rst | 4 +-- doc/guides/rel_notes/release_23_07.rst| 6 drivers/common/sfc_efx/base/ef10_filter.c | 26 +++ drivers/common/sfc_efx/base/ef10_impl.h | 6 drivers/common/sfc_efx/base/ef10_nic.c| 6 drivers/common/sfc_efx/base/efx.h | 13 drivers/common/sfc_efx/base/efx_filter.c | 32 +++ drivers/common/sfc_efx/base/efx_impl.h| 8 + drivers/common/sfc_efx/base/efx_port.c| 39 +++ drivers/common/sfc_efx/base/efx_rx.c | 14 drivers/common/sfc_efx/base/rhead_rx.c| 3 ++ drivers/common/sfc_efx/base/siena_nic.c | 1 + drivers/common/sfc_efx/version.map| 1 + drivers/net/sfc/sfc.h | 1 + drivers/net/sfc/sfc_dp_rx.h | 1 + drivers/net/sfc/sfc_ef100_rx.c| 16 +- drivers/net/sfc/sfc_port.c| 11 +++ drivers/net/sfc/sfc_rx.c | 10 ++ 18 files changed, 195 insertions(+), 3 deletions(-) -- 2.34.1
[PATCH v7 1/4] common/sfc_efx/base: report VLAN stripping capability
These changes are necessary in order to add support for stripping
VLAN tags in the future.
Signed-off-by: Artemii Morozov
Reviewed-by: Ivan Malov
Reviewed-by: Andy Moreton
Acked-by: Andrew Rybchenko
---
drivers/common/sfc_efx/base/ef10_nic.c | 6 ++
drivers/common/sfc_efx/base/efx.h | 1 +
drivers/common/sfc_efx/base/siena_nic.c | 1 +
3 files changed, 8 insertions(+)
diff --git a/drivers/common/sfc_efx/base/ef10_nic.c
b/drivers/common/sfc_efx/base/ef10_nic.c
index bf9cb9d309..79d596b5ef 100644
--- a/drivers/common/sfc_efx/base/ef10_nic.c
+++ b/drivers/common/sfc_efx/base/ef10_nic.c
@@ -1223,6 +1223,12 @@ ef10_get_datapath_caps(
else
encp->enc_hw_tx_insert_vlan_enabled = B_FALSE;
+ /* Check if firmware supports VLAN stripping. */
+ if (CAP_FLAGS1(req, RX_VLAN_STRIPPING))
+ encp->enc_rx_vlan_stripping_supported = B_TRUE;
+ else
+ encp->enc_rx_vlan_stripping_supported = B_FALSE;
+
/* Check if the firmware supports RX event batching */
if (CAP_FLAGS1(req, RX_BATCHING))
encp->enc_rx_batching_enabled = B_TRUE;
diff --git a/drivers/common/sfc_efx/base/efx.h
b/drivers/common/sfc_efx/base/efx.h
index ef626cc55a..7fcf48f5e1 100644
--- a/drivers/common/sfc_efx/base/efx.h
+++ b/drivers/common/sfc_efx/base/efx.h
@@ -1627,6 +1627,7 @@ typedef struct efx_nic_cfg_s {
/* Number of TSO contexts on the NIC (FATSOv2) */
uint32_tenc_fw_assisted_tso_v2_n_contexts;
boolean_t enc_hw_tx_insert_vlan_enabled;
+ boolean_t enc_rx_vlan_stripping_supported;
/* Number of PFs on the NIC */
uint32_tenc_hw_pf_count;
/* Datapath firmware vadapter/vport/vswitch support */
diff --git a/drivers/common/sfc_efx/base/siena_nic.c
b/drivers/common/sfc_efx/base/siena_nic.c
index ca38eefa7e..e334f01b03 100644
--- a/drivers/common/sfc_efx/base/siena_nic.c
+++ b/drivers/common/sfc_efx/base/siena_nic.c
@@ -179,6 +179,7 @@ siena_board_cfg(
(encp->enc_rxq_limit * EFX_RXQ_DC_NDESCS(EFX_RXQ_DC_SIZE));
encp->enc_hw_tx_insert_vlan_enabled = B_FALSE;
+ encp->enc_rx_vlan_stripping_supported = B_FALSE;
encp->enc_fw_assisted_tso_enabled = B_FALSE;
encp->enc_fw_assisted_tso_v2_enabled = B_FALSE;
encp->enc_fw_assisted_tso_v2_n_contexts = 0;
--
2.34.1
[PATCH v7 2/4] common/sfc_efx/base: add API to get installed filters count
This API allows to get number of installed filters. This will
be used in the future patches.
Signed-off-by: Artemii Morozov
---
drivers/common/sfc_efx/base/ef10_filter.c | 20 ++
drivers/common/sfc_efx/base/ef10_impl.h | 6 +
drivers/common/sfc_efx/base/efx_filter.c | 32 +++
drivers/common/sfc_efx/base/efx_impl.h| 7 +
4 files changed, 65 insertions(+)
diff --git a/drivers/common/sfc_efx/base/ef10_filter.c
b/drivers/common/sfc_efx/base/ef10_filter.c
index d6940011c0..278502fb61 100644
--- a/drivers/common/sfc_efx/base/ef10_filter.c
+++ b/drivers/common/sfc_efx/base/ef10_filter.c
@@ -2113,6 +2113,26 @@ ef10_filter_reconfigure(
return (rc);
}
+ __checkReturn efx_rc_t
+ef10_filter_get_count(
+ __inefx_nic_t *enp,
+ __out uint32_t *count)
+{
+ ef10_filter_table_t *table = enp->en_filter.ef_ef10_filter_table;
+ uint32_t filters_count;
+
+ EFSYS_ASSERT(EFX_FAMILY_IS_EF100(enp) || EFX_FAMILY_IS_EF10(enp));
+ EFSYS_ASSERT(count != NULL);
+
+ filters_count = table->eft_unicst_filter_count +
+ table->eft_mulcst_filter_count +
+ table->eft_encap_filter_count;
+
+ *count = filters_count;
+
+ return (0);
+}
+
void
ef10_filter_get_default_rxq(
__inefx_nic_t *enp,
diff --git a/drivers/common/sfc_efx/base/ef10_impl.h
b/drivers/common/sfc_efx/base/ef10_impl.h
index 877aedad45..914bba10ce 100644
--- a/drivers/common/sfc_efx/base/ef10_impl.h
+++ b/drivers/common/sfc_efx/base/ef10_impl.h
@@ -1347,6 +1347,12 @@ ef10_filter_reconfigure(
__in_ecount(6*count)uint8_t const *addrs,
__inuint32_t count);
+LIBEFX_INTERNAL
+extern __checkReturn efx_rc_t
+ef10_filter_get_count(
+ __inefx_nic_t *enp,
+ __out uint32_t *count);
+
LIBEFX_INTERNAL
extern void
ef10_filter_get_default_rxq(
diff --git a/drivers/common/sfc_efx/base/efx_filter.c
b/drivers/common/sfc_efx/base/efx_filter.c
index 83c37ff859..064630a66b 100644
--- a/drivers/common/sfc_efx/base/efx_filter.c
+++ b/drivers/common/sfc_efx/base/efx_filter.c
@@ -53,6 +53,7 @@ static const efx_filter_ops_t __efx_filter_siena_ops = {
siena_filter_delete,/* efo_delete */
siena_filter_supported_filters, /* efo_supported_filters */
NULL, /* efo_reconfigure */
+ NULL, /* efo_get_count */
};
#endif /* EFSYS_OPT_SIENA */
@@ -65,6 +66,7 @@ static const efx_filter_ops_t __efx_filter_ef10_ops = {
ef10_filter_delete, /* efo_delete */
ef10_filter_supported_filters, /* efo_supported_filters */
ef10_filter_reconfigure,/* efo_reconfigure */
+ ef10_filter_get_count, /* efo_get_count */
};
#endif /* EFX_OPTS_EF10() */
@@ -77,6 +79,7 @@ static const efx_filter_ops_t __efx_filter_rhead_ops = {
ef10_filter_delete, /* efo_delete */
ef10_filter_supported_filters, /* efo_supported_filters */
ef10_filter_reconfigure,/* efo_reconfigure */
+ ef10_filter_get_count, /* efo_get_count */
};
#endif /* EFSYS_OPT_RIVERHEAD */
@@ -309,6 +312,35 @@ efx_filter_reconfigure(
fail1:
EFSYS_PROBE1(fail1, efx_rc_t, rc);
+ return (rc);
+}
+
+ __checkReturn efx_rc_t
+efx_filter_get_count(
+ __inefx_nic_t *enp,
+ __out uint32_t *count)
+{
+ efx_rc_t rc;
+
+ EFSYS_ASSERT3U(enp->en_magic, ==, EFX_NIC_MAGIC);
+ EFSYS_ASSERT3U(enp->en_mod_flags, &, EFX_MOD_PROBE);
+ EFSYS_ASSERT3U(enp->en_mod_flags, &, EFX_MOD_FILTER);
+
+ if (enp->en_efop->efo_get_count == NULL) {
+ rc = ENOTSUP;
+ goto fail1;
+ }
+
+ if ((rc = enp->en_efop->efo_get_count(enp, count)) != 0)
+ goto fail2;
+
+ return (0);
+
+fail2:
+ EFSYS_PROBE(fail2);
+fail1:
+ EFSYS_PROBE1(fail1, efx_rc_t, rc);
+
return (rc);
}
diff --git a/drivers/common/sfc_efx/base/efx_impl.h
b/drivers/common/sfc_efx/base/efx_impl.h
index 92a30c34ae..91ba187c73 100644
--- a/drivers/common/sfc_efx/base/efx_impl.h
+++ b/drivers/common/sfc_efx/base/efx_impl.h
@@ -288,6 +288,7 @@ typedef struct efx_filter_ops_s {
efx_rc_t(*efo_reconfigure)(efx_nic_t *, uint8_t const *,
boolean_t,
boolean_t, boolean_t, boolean_t,
uint8_t const *, uint32_t);
+ efx_rc_t(*efo_get_count)(efx_nic_t *, uint32_t *);
} efx_filter_ops_t;
LIBEFX_INTERNAL
@@ -302,6 +303,12 @@ efx_filter_reconfigure(
__in_ecount(6*count)uint8_t const *addrs,
__inuint32_t count);
+LIBEFX_INTERNAL
+extern __checkReturn efx_rc_t
+efx_filter_get_count(
+ __inefx_nic_t *enp,
+ __out uint32_t *count);
[PATCH v7 3/4] common/sfc_efx/base: add support to enable VLAN stripping
To enable VLAN stripping, two conditions must be met:
the corresponding flag must be set and the appropriate
Rx prefix should be requested.
VLAN stripping is supported on EF100.
Signed-off-by: Artemii Morozov
Reviewed-by: Ivan Malov
Reviewed-by: Andy Moreton
---
drivers/common/sfc_efx/base/ef10_filter.c | 6
drivers/common/sfc_efx/base/efx.h | 12 +++
drivers/common/sfc_efx/base/efx_impl.h| 1 +
drivers/common/sfc_efx/base/efx_port.c| 39 +++
drivers/common/sfc_efx/base/efx_rx.c | 14
drivers/common/sfc_efx/base/rhead_rx.c| 3 ++
drivers/common/sfc_efx/version.map| 1 +
7 files changed, 76 insertions(+)
diff --git a/drivers/common/sfc_efx/base/ef10_filter.c
b/drivers/common/sfc_efx/base/ef10_filter.c
index 278502fb61..3e6dd2e35c 100644
--- a/drivers/common/sfc_efx/base/ef10_filter.c
+++ b/drivers/common/sfc_efx/base/ef10_filter.c
@@ -171,6 +171,7 @@ efx_mcdi_filter_op_add(
EFX_MCDI_DECLARE_BUF(payload, MC_CMD_FILTER_OP_V3_IN_LEN,
MC_CMD_FILTER_OP_EXT_OUT_LEN);
efx_filter_match_flags_t match_flags;
+ efx_port_t *epp = &(enp->en_port);
uint32_t port_id;
efx_rc_t rc;
@@ -338,6 +339,11 @@ efx_mcdi_filter_op_add(
FILTER_OP_V3_IN_MATCH_SET_FLAG, 1);
}
+ if (epp->ep_vlan_strip) {
+ MCDI_IN_SET_DWORD_FIELD(req, FILTER_OP_V3_IN_MATCH_ACTION_FLAGS,
+ FILTER_OP_V3_IN_MATCH_STRIP_VLAN, 1);
+ }
+
efx_mcdi_execute(enp, &req);
if (req.emr_rc != 0) {
diff --git a/drivers/common/sfc_efx/base/efx.h
b/drivers/common/sfc_efx/base/efx.h
index 7fcf48f5e1..77f855bfb0 100644
--- a/drivers/common/sfc_efx/base/efx.h
+++ b/drivers/common/sfc_efx/base/efx.h
@@ -1158,6 +1158,12 @@ efx_port_poll(
__inefx_nic_t *enp,
__out_opt efx_link_mode_t *link_modep);
+LIBEFX_API
+extern __checkReturn efx_rc_t
+efx_port_vlan_strip_set(
+ __inefx_nic_t *enp,
+ __inboolean_t enabled);
+
LIBEFX_API
extern void
efx_port_fini(
@@ -3117,6 +3123,12 @@ typedef enum efx_rxq_type_e {
* Request user flag field in the Rx prefix of a queue.
*/
#defineEFX_RXQ_FLAG_USER_FLAG 0x20
+/*
+ * Request VLAN TCI field in the Rx prefix. The flag just
+ * controls delivery of the stripped VLAN TCI if VLAN stripping
+ * is enabled and done.
+ */
+#defineEFX_RXQ_FLAG_VLAN_STRIPPED_TCI 0x40
LIBEFX_API
extern __checkReturn efx_rc_t
diff --git a/drivers/common/sfc_efx/base/efx_impl.h
b/drivers/common/sfc_efx/base/efx_impl.h
index 91ba187c73..78cc056576 100644
--- a/drivers/common/sfc_efx/base/efx_impl.h
+++ b/drivers/common/sfc_efx/base/efx_impl.h
@@ -371,6 +371,7 @@ typedef struct efx_port_s {
uint32_tep_phy_cap_mask;
boolean_t ep_mac_drain;
boolean_t ep_include_fcs;
+ boolean_t ep_vlan_strip;
#if EFSYS_OPT_BIST
efx_bist_type_t ep_current_bist;
#endif
diff --git a/drivers/common/sfc_efx/base/efx_port.c
b/drivers/common/sfc_efx/base/efx_port.c
index a5f982e335..e5a9fa6c53 100644
--- a/drivers/common/sfc_efx/base/efx_port.c
+++ b/drivers/common/sfc_efx/base/efx_port.c
@@ -204,6 +204,45 @@ efx_loopback_type_name(
#endif /* EFSYS_OPT_LOOPBACK */
+ __checkReturn efx_rc_t
+efx_port_vlan_strip_set(
+ __inefx_nic_t *enp,
+ __inboolean_t enabled)
+{
+ efx_port_t *epp = &(enp->en_port);
+ efx_nic_cfg_t *encp = &(enp->en_nic_cfg);
+ uint32_t filters_count = 0;
+ efx_rc_t rc;
+
+ EFSYS_ASSERT3U(enp->en_magic, ==, EFX_NIC_MAGIC);
+
+ if (enabled && !encp->enc_rx_vlan_stripping_supported) {
+ rc = ENOTSUP;
+ goto fail1;
+ }
+
+ if ((rc = efx_filter_get_count(enp, &filters_count)) != 0)
+ goto fail2;
+
+ if (filters_count != 0) {
+ rc = EINVAL;
+ goto fail3;
+ }
+
+ epp->ep_vlan_strip = enabled;
+
+ return (0);
+
+fail3:
+ EFSYS_PROBE(fail3);
+fail2:
+ EFSYS_PROBE(fail2);
+fail1:
+ EFSYS_PROBE1(fail1, efx_rc_t, rc);
+
+ return (rc);
+}
+
void
efx_port_fini(
__inefx_nic_t *enp)
diff --git a/drivers/common/sfc_efx/base/efx_rx.c
b/drivers/common/sfc_efx/base/efx_rx.c
index 68f42f5cac..b3d9e14c67 100644
--- a/drivers/common/sfc_efx/base/efx_rx.c
+++ b/drivers/common/sfc_efx/base/efx_rx.c
@@ -941,11 +941,25 @@ efx_rx_qcreate_internal(
goto fail5;
}
+ if (flags & EFX_RXQ_FLAG_VLAN_STRIPPED_TCI) {
+ const efx_rx_prefix_layout_t *erplp = &erp->er_prefix_layout;
+ const efx_rx_prefix_field_info_t *vlan_tci_field;
+
+ vlan_tci_field =
+ &erplp->erpl_fields[EFX_RX_PREFIX_FIELD_VLAN_STRIP_
[PATCH v7 4/4] net/sfc: support VLAN stripping offload
Extract VLAN TCI provided by the HW in the prefix and put it to mbuf.
VLAN stripping is supported for ef100 datapath only. This is device
level offload.
Signed-off-by: Artemii Morozov
Reviewed-by: Viacheslav Galaktionov
Reviewed-by: Ivan Malov
Reviewed-by: Andy Moreton
---
doc/guides/nics/sfc_efx.rst| 4 ++--
doc/guides/rel_notes/release_23_07.rst | 6 ++
drivers/net/sfc/sfc.h | 1 +
drivers/net/sfc/sfc_dp_rx.h| 1 +
drivers/net/sfc/sfc_ef100_rx.c | 16 +++-
drivers/net/sfc/sfc_port.c | 11 +++
drivers/net/sfc/sfc_rx.c | 10 ++
7 files changed, 46 insertions(+), 3 deletions(-)
diff --git a/doc/guides/nics/sfc_efx.rst b/doc/guides/nics/sfc_efx.rst
index 24459da33e..eafb88191a 100644
--- a/doc/guides/nics/sfc_efx.rst
+++ b/doc/guides/nics/sfc_efx.rst
@@ -122,6 +122,8 @@ SFC EFX PMD has support for:
- Port representors (see :ref: switch_representation)
+- VLAN stripping (if running firmware variant supports it)
+
Non-supported Features
--
@@ -134,8 +136,6 @@ The features not yet supported include:
- VLAN filtering
-- VLAN stripping
-
- LRO
diff --git a/doc/guides/rel_notes/release_23_07.rst
b/doc/guides/rel_notes/release_23_07.rst
index 738d35d9f7..ce523800c7 100644
--- a/doc/guides/rel_notes/release_23_07.rst
+++ b/doc/guides/rel_notes/release_23_07.rst
@@ -185,6 +185,12 @@ New Features
See :doc:`../prog_guide/pdcp_lib` for more information.
+* **Updated Solarflare network PMD.**
+
+ Updated the Solarflare ``sfc_efx`` driver with changes including:
+
+ * Added VLAN stripping support on SN1000 SmartNICs
+
Removed Items
-
diff --git a/drivers/net/sfc/sfc.h b/drivers/net/sfc/sfc.h
index 25cdeaa5cd..2432a2307e 100644
--- a/drivers/net/sfc/sfc.h
+++ b/drivers/net/sfc/sfc.h
@@ -75,6 +75,7 @@ struct sfc_port {
unsigned intflow_ctrl;
boolean_t flow_ctrl_autoneg;
boolean_t include_fcs;
+ boolean_t vlan_strip;
size_t pdu;
/*
diff --git a/drivers/net/sfc/sfc_dp_rx.h b/drivers/net/sfc/sfc_dp_rx.h
index 8a504bdcf1..9f9bf28988 100644
--- a/drivers/net/sfc/sfc_dp_rx.h
+++ b/drivers/net/sfc/sfc_dp_rx.h
@@ -70,6 +70,7 @@ struct sfc_dp_rx_qcreate_info {
unsigned intflags;
#define SFC_RXQ_FLAG_RSS_HASH 0x1
#define SFC_RXQ_FLAG_INGRESS_MPORT 0x2
+#define SFC_RXQ_FLAG_VLAN_STRIPPED_TCI 0x4
/** Rx queue size */
unsigned intrxq_entries;
diff --git a/drivers/net/sfc/sfc_ef100_rx.c b/drivers/net/sfc/sfc_ef100_rx.c
index 5563bd9a0b..2677003da3 100644
--- a/drivers/net/sfc/sfc_ef100_rx.c
+++ b/drivers/net/sfc/sfc_ef100_rx.c
@@ -68,6 +68,7 @@ struct sfc_ef100_rxq {
#define SFC_EF100_RXQ_INGRESS_MPORT0x80
#define SFC_EF100_RXQ_USER_FLAG0x100
#define SFC_EF100_RXQ_NIC_DMA_MAP 0x200
+#define SFC_EF100_RXQ_VLAN_STRIPPED_TCI0x400
unsigned intptr_mask;
unsigned intevq_phase_bit_shift;
unsigned intready_pkts;
@@ -392,6 +393,7 @@ static const efx_rx_prefix_layout_t
sfc_ef100_rx_prefix_layout = {
SFC_EF100_RX_PREFIX_FIELD(RSS_HASH, B_FALSE),
SFC_EF100_RX_PREFIX_FIELD(USER_FLAG, B_FALSE),
SFC_EF100_RX_PREFIX_FIELD(USER_MARK, B_FALSE),
+ SFC_EF100_RX_PREFIX_FIELD(VLAN_STRIP_TCI, B_FALSE),
#undef SFC_EF100_RX_PREFIX_FIELD
}
@@ -472,6 +474,14 @@ sfc_ef100_rx_prefix_to_offloads(const struct sfc_ef100_rxq
*rxq,
ESF_GZ_RX_PREFIX_INGRESS_MPORT);
}
+ if (rxq->flags & SFC_EF100_RXQ_VLAN_STRIPPED_TCI &&
+ EFX_TEST_XWORD_BIT(rx_prefix[0],
+ ESF_GZ_RX_PREFIX_VLAN_STRIPPED_LBN)) {
+ ol_flags |= RTE_MBUF_F_RX_VLAN | RTE_MBUF_F_RX_VLAN_STRIPPED;
+ m->vlan_tci = EFX_XWORD_FIELD(rx_prefix[0],
+
ESF_GZ_RX_PREFIX_VLAN_STRIP_TCI);
+ }
+
m->ol_flags = ol_flags;
return true;
}
@@ -813,6 +823,9 @@ sfc_ef100_rx_qcreate(uint16_t port_id, uint16_t queue_id,
if (info->flags & SFC_RXQ_FLAG_INGRESS_MPORT)
rxq->flags |= SFC_EF100_RXQ_INGRESS_MPORT;
+ if (info->flags & SFC_RXQ_FLAG_VLAN_STRIPPED_TCI)
+ rxq->flags |= SFC_EF100_RXQ_VLAN_STRIPPED_TCI;
+
sfc_ef100_rx_debug(rxq, "RxQ doorbell is %p", rxq->doorbell);
*dp_rxqp = &rxq->dp;
@@ -1004,7 +1017,8 @@ struct sfc_dp_rx sfc_ef100_rx = {
SFC_DP_RX_FEAT_FLOW_MARK |
SFC_DP_RX_FEAT_INTR |
SFC_DP_RX_FEAT_STATS,
- .dev_offload_capa = RTE_ETH_RX_OFFLOAD_KEEP_CRC,
+ .de
Re: [PATCH v14 3/6] memarea: support alloc and free API
On 2/9/2023 6:36 AM, Chengwen Feng wrote:
This patch supports rte_memarea_alloc() and rte_memarea_free() API.
Signed-off-by: Chengwen Feng
Reviewed-by: Dongdong Liu
Acked-by: Morten Brørup
---
General note: this patchset could benefit from a bit more comments. I
don't suggest commenting every line, but at least more comments denoting
various logical steps (like you have in `rte_memarea_free`) would be
nice to have.
#include
#include
@@ -88,6 +90,8 @@ memarea_alloc_area(const struct rte_memarea_param *init)
init->numa_socket);
else if (init->source == RTE_MEMAREA_SOURCE_LIBC)
ptr = memarea_alloc_from_libc(init->total_sz);
+ else if (init->source == RTE_MEMAREA_SOURCE_MEMAREA)
+ ptr = rte_memarea_alloc(init->src_ma, init->total_sz);
Why `if` not `switch`?
return ptr;
}
@@ -99,6 +103,8 @@ memarea_free_area(const struct rte_memarea_param *init, void
*ptr)
rte_free(ptr);
else if (init->source == RTE_MEMAREA_SOURCE_LIBC)
free(ptr);
+ else if (init->source == RTE_MEMAREA_SOURCE_MEMAREA)
+ rte_memarea_free(init->src_ma, ptr);
Similarly here: why `if` not `switch`?
+static inline void
+memarea_add_node(struct rte_memarea *ma, struct memarea_objhdr *hdr, size_t
alloc_sz)
+{
+#ifdef RTE_LIBRTE_MEMAREA_DEBUG
+ struct memarea_objtlr *cur_tlr;
+#endif
+ struct memarea_objhdr *new_hdr;
+
+#ifdef RTE_LIBRTE_MEMAREA_DEBUG
+ cur_tlr = RTE_PTR_ADD(hdr, sizeof(struct memarea_objhdr) + alloc_sz);
+ cur_tlr->cookie = MEMAREA_OBJECT_TRAILER_COOKIE;
+ new_hdr = RTE_PTR_ADD(cur_tlr, sizeof(struct memarea_objtlr));
+ new_hdr->cookie = MEMAREA_OBJECT_HEADER_AVAILABLE_COOKIE;
+#else
+ new_hdr = RTE_PTR_ADD(hdr, sizeof(struct memarea_objhdr) + alloc_sz);
+#endif
+ TAILQ_INSERT_AFTER(&ma->obj_list, hdr, new_hdr, obj_next);
+ TAILQ_INSERT_AFTER(&ma->avail_list, hdr, new_hdr, avail_next);
+}
It seems to me that this function isn't "adding" node but rather is
splitting the `hdr` into two nodes. This is nitpicking, but I feel like
this part could be clearer semantically (splitting the function, adding
comments, some other way...).
+
+void *
+rte_memarea_alloc(struct rte_memarea *ma, size_t size)
+{
+ size_t align_sz = RTE_ALIGN(size, MEMAREA_OBJECT_SIZE_ALIGN);
+ struct memarea_objhdr *hdr;
+ size_t avail_sz;
+ void *ptr = NULL;
+
+ if (unlikely(ma == NULL || size == 0 || align_sz < size))
+ return ptr;
It would be nice if API also set rte_errno to indicate what kind of
error has happened.
+
+ memarea_lock(ma);
+ TAILQ_FOREACH(hdr, &ma->avail_list, avail_next) {
+ memarea_check_cookie(ma, hdr, 0);
+ avail_sz = MEMAREA_OBJECT_GET_SIZE(hdr);
+ if (avail_sz < align_sz)
+ continue;
+ if (memarea_whether_add_node(avail_sz, align_sz))
+ memarea_add_node(ma, hdr, align_sz);
I didn't get this at first, which means it needs comments :)
Specifically, it seems to me that we're only "adding" a node when we can
comfortably split it. So, in addition to comments documenting the above,
perhaps the above functions should also be called differently? Like
`memarea_can_split()` and `memarea_split`? IMO it'd communicate the
intent better (unless I misunderstood the intent, that is!).
+ TAILQ_REMOVE(&ma->avail_list, hdr, avail_next);
+ MEMAREA_OBJECT_MARK_ALLOCATED(hdr);
+#ifdef RTE_LIBRTE_MEMAREA_DEBUG
+ hdr->cookie = MEMAREA_OBJECT_HEADER_ALLOCATED_COOKIE;
+#endif
+ ptr = RTE_PTR_ADD(hdr, sizeof(struct memarea_objhdr));
+ break;
+ }
+ memarea_unlock(ma);
+
+ return ptr;
It seems that it's possible to reach the end of the loop and return NULL
as `ptr`, without any error. I would suggest setting rte_errno to
`ENOMEM` initially, and clearing it when we find a suitable element.
+}
+
+static inline void
+memarea_merge_node(struct rte_memarea *ma, struct memarea_objhdr *curr,
+ struct memarea_objhdr *next)
+{
+#ifdef RTE_LIBRTE_MEMAREA_DEBUG
+ struct memarea_objtlr *tlr;
+#endif
+ RTE_SET_USED(curr);
+ TAILQ_REMOVE(&ma->obj_list, next, obj_next);
+ TAILQ_REMOVE(&ma->avail_list, next, avail_next);
+#ifdef RTE_LIBRTE_MEMAREA_DEBUG
+ next->cookie = 0;
+ tlr = RTE_PTR_SUB(next, sizeof(struct memarea_objtlr));
+ tlr->cookie = 0;
+#endif
+}
+
+void
+rte_memarea_free(struct rte_memarea *ma, void *ptr)
+{
+ struct memarea_objhdr *hdr, *prev, *next;
+
+ if (unlikely(ma == NULL || ptr == NULL))
+ return;
+
+ hdr = RTE_PTR_SUB(ptr, sizeof(struct memarea_objhdr));
+ if (unlikely(!MEMAREA_OBJECT_IS_ALLOCATED(hdr))) {
Here and above - I question the value of using `unlikely` here
Re: [PATCH v7 3/4] common/sfc_efx/base: add support to enable VLAN stripping
On 6/22/23 18:11, Artemii Morozov wrote: To enable VLAN stripping, two conditions must be met: the corresponding flag must be set and the appropriate Rx prefix should be requested. VLAN stripping is supported on EF100. Signed-off-by: Artemii Morozov Reviewed-by: Ivan Malov Reviewed-by: Andy Moreton Acked-by: Andrew Rybchenko
Re: [PATCH v7 4/4] net/sfc: support VLAN stripping offload
On 6/22/23 18:11, Artemii Morozov wrote: Extract VLAN TCI provided by the HW in the prefix and put it to mbuf. VLAN stripping is supported for ef100 datapath only. This is device level offload. Signed-off-by: Artemii Morozov Reviewed-by: Viacheslav Galaktionov Reviewed-by: Ivan Malov Reviewed-by: Andy Moreton Acked-by: Andrew Rybchenko
Re: [PATCH v7 2/4] common/sfc_efx/base: add API to get installed filters count
On 6/22/23 18:11, Artemii Morozov wrote: This API allows to get number of installed filters. This will be used in the future patches. Signed-off-by: Artemii Morozov Acked-by: Andrew Rybchenko However, it would be good if it is reviewed as well by Andy and Ivan.
Re: [PATCH v14 5/6] memarea: support dump API
On 2/9/2023 6:36 AM, Chengwen Feng wrote:
This patch supports rte_memarea_dump() API which could be used for
debug.
Signed-off-by: Chengwen Feng
Reviewed-by: Dongdong Liu
Acked-by: Morten Brørup
---
Provisionally,
Acked-by: Anatoly Burakov
As long as below is addressed.
+static void
+memarea_dump_objects_detail(struct rte_memarea *ma, FILE *f)
+{
+ struct memarea_objhdr *hdr;
+ size_t offset;
+ void *ptr;
+
+ fprintf(f, " objects:\n");
+ TAILQ_FOREACH(hdr, &ma->obj_list, obj_next) {
+ if (hdr == ma->guard_hdr)
+ break;
+ memarea_check_cookie(ma, hdr, 2);
+ ptr = RTE_PTR_ADD(hdr, sizeof(struct memarea_objhdr));
+ offset = RTE_PTR_DIFF(ptr, ma->area_base);
+#ifdef RTE_LIBRTE_MEMAREA_DEBUG
+ fprintf(f, "%p off: 0x%zx size: 0x%zx %s\n",
+ ptr, offset, MEMAREA_OBJECT_GET_SIZE(hdr),
+ MEMAREA_OBJECT_IS_ALLOCATED(hdr) ? "allocated" : "");
+#else
+ fprintf(f, "off: 0x%zx size: 0x%zx %s\n",
+ offset, MEMAREA_OBJECT_GET_SIZE(hdr),
+ MEMAREA_OBJECT_IS_ALLOCATED(hdr) ? "allocated" : "");
+#endif
+ }.
+}
+
+int
+rte_memarea_dump(struct rte_memarea *ma, FILE *f, bool dump_all)
+{
+ if (ma == NULL || f == NULL)
+ return -EINVAL;
I feel like the API is inconsistent in this way. I would suggest picking
a method of error reporting, and sticking with it. I would suggest
returning 0/-1 or ptr/NULL with rte_errno set to indicate error, as that
is how most libraries in DPDK behave.
--
Thanks,
Anatoly
Re: [PATCH v7 4/4] net/sfc: support VLAN stripping offload
On 6/22/2023 4:11 PM, Artemii Morozov wrote: > Extract VLAN TCI provided by the HW in the prefix and put it to mbuf. > VLAN stripping is supported for ef100 datapath only. This is device > level offload. > > Signed-off-by: Artemii Morozov > Reviewed-by: Viacheslav Galaktionov > Reviewed-by: Ivan Malov > Reviewed-by: Andy Moreton <...> > diff --git a/doc/guides/rel_notes/release_23_07.rst > b/doc/guides/rel_notes/release_23_07.rst > index 738d35d9f7..ce523800c7 100644 > --- a/doc/guides/rel_notes/release_23_07.rst > +++ b/doc/guides/rel_notes/release_23_07.rst > @@ -185,6 +185,12 @@ New Features > >See :doc:`../prog_guide/pdcp_lib` for more information. > > +* **Updated Solarflare network PMD.** > + > + Updated the Solarflare ``sfc_efx`` driver with changes including: > + > + * Added VLAN stripping support on SN1000 SmartNICs > + > Can you please move release notes update to existing Solarflace block above?
DPDK22 issue: Unable to set more than 4 queues in Azure
Hi All,
We are observing the following issue with DPDK22.11. We didn’t find any
upstream patches for this issue on the DPDK github. Is there any known
issue, please let us know.
*Issue:*
On Azure platform, we are unable to configure more than 4 queues. When we
try to configure more than 4 queues its failing with “EAL: Cannot send more
than 8 FDs” error.
Here I am pasting the working and failing testpmd logs.
Please note that this issue is not observed in DPDK 21.11.
*DPDK Testpmd*:
*Working case with 4 rx & 4 tx queues:*
./build/app/dpdk-testpmd -l 0-3 -a 38d9:00:02.0 -- --nb-cores=3 --rxq=4
--txq=4 --burst=64 --mbcache=512 --max-pkt-len=128 --forward-mode=txonly -a
--stats-period 1
*Failing case with 8 rx & 8 tx queues:*
./build/app/dpdk-testpmd -l 0-7 -a 38d9:00:02.0 -- --nb-cores=4 --rxq=8
--txq=8 --burst=64 --mbcache=512 --max-pkt-len=128 --forward-mode=txonly -a
--stats-period 1
*DPDK log:*
Configuring Port 0 (socket 0)
EAL: Cannot send more than 8 FDs
tap_mp_req_on_rxtx(): Failed to send start req to secondary 7
*** stack smashing detected ***: terminated
Aborted
=
When we checked DPDK 22.11 code, tap_mp_req_on_rxtx() pasted below is newly
added in rte_eth_tap.c file which is causing the issue.
static int
tap_mp_req_on_rxtx(struct rte_eth_dev *dev)
{
struct rte_mp_msg msg;
struct ipc_queues *request_param = (struct ipc_queues
*)msg.param;
int err;
int fd_iterator = 0;
struct pmd_process_private *process_private =
dev->process_private;
int i;
memset(&msg, 0, sizeof(msg));
strlcpy(msg.name, TAP_MP_REQ_START_RXTX, sizeof(msg.name));
strlcpy(request_param->port_name, dev->data->name,
sizeof(request_param->port_name));
msg.len_param = sizeof(*request_param);
for (i = 0; i < dev->data->nb_tx_queues; i++) {
msg.fds[fd_iterator++] =
process_private->txq_fds[i];
msg.num_fds++;
request_param->txq_count++;
}
for (i = 0; i < dev->data->nb_rx_queues; i++) {
msg.fds[fd_iterator++] =
process_private->rxq_fds[i];
msg.num_fds++;
request_param->rxq_count++;
}
err = rte_mp_sendmsg(&msg); è This
function is sending msg with msg.num_fds > RTE_PMD_TAP_MAX_QUEUES (i.e. 8),
because of that its failing.
if (err < 0) {
TAP_LOG(ERR, "Failed to send start req to
secondary %d",
rte_errno);
return -1;
}
return 0;
}
Thanks
Nagesh
RE: [PATCH v4] app/testpmd: fix primary process not polling all queues
> -Original Message-
> From: Jie Hai
> Sent: Friday, June 9, 2023 12:04 PM
> To: Aman Singh ; Yuying Zhang
> ; Anatoly Burakov ;
> Matan Azrad ; Dmitry Kozlyuk
>
> Cc: dev@dpdk.org; liudongdo...@huawei.com; shiyangx...@intel.com;
> ferruh.yi...@amd.com
> Subject: [PATCH v4] app/testpmd: fix primary process not polling all queues
>
> Here's how the problem arises.
> step1: Start the app.
> dpdk-testpmd -a :35:00.0 -l 0-3 -- -i --rxq=10 --txq=10
>
> step2: Perform the following steps and send traffic. As expected,
> queue 7 does not send or receive packets, and other queues do.
> port 0 rxq 7 stop
> port 0 txq 7 stop
> set fwd mac
> start
>
> step3: Perform the following steps and send traffic. All queues
> are expected to send and receive packets normally, but that's not
> the case for queue 7.
> stop
> port stop all
> port start all
> start
> show port xstats all
>
> In fact, only the value of rx_q7_packets for queue 7 is not zero,
> which means queue 7 is enabled for the driver but is not involved
> in packet receiving and forwarding by software. If we check queue
> state by command 'show rxq info 0 7' and 'show txq info 0 7',
> we see queue 7 is started as other queues are.
> Rx queue state: started
> Tx queue state: started
> The queue 7 is started but cannot forward. That's the problem.
>
> We know that each stream has a read-only "disabled" field that
> control if this stream should be used to forward. This field
> depends on testpmd local queue state, please see
> commit 3c4426db54fc ("app/testpmd: do not poll stopped queues").
> DPDK framework maintains ethdev queue state that drivers reported,
> which indicates the real state of queues.
>
> There are commands that update these two kind queue state such as
> 'port X rxq|txq start|stop'. But these operations take effect only
> in one stop-start round. In the following stop-start round, the
> preceding operations do not take effect anymore. However, only
> the ethdev queue state is updated, causing the testpmd and ethdev
> state information to diverge and causing unexpected side effects
> as above problem.
>
> There was a similar problem for the secondary process, please see
> commit 5028f207a4fa ("app/testpmd: fix secondary process packet
> forwarding").
>
> This patch applies its workaround with some difference to the
> primary process. Not all PMDs implement rte_eth_rx_queue_info_get and
> rte_eth_tx_queue_info_get, however they may support deferred_start
> with primary process. To not break their behavior, retain the original
> testpmd local queue state for those PMDs.
>
> Fixes: 3c4426db54fc ("app/testpmd: do not poll stopped queues")
> Cc: sta...@dpdk.org
>
> Signed-off-by: Jie Hai
> ---
Hi Jie,
I see the error below when starting a representor port after reattaching it
with this patch, is it expected?
$ sudo ./build /app/dpdk-testpmd -n 4 -a
:08:00.0,dv_esw_en=1,representor=vf0-1 -a auxiliary: -a 00:00.0
--iova-mode="va" -- -i
[..]
testpmd> port stop all
testpmd> port close 0
testpmd> device detach :08:00.0
testpmd> port attach :08:00.0,dv_esw_en=1,representor=0-1
testpmd> port start 1
Configuring Port 1 (socket 0)
Port 1: FA:9E:D8:5F:D7:D8
Invalid Rx queue_id=0
testpmd: Failed to get rx queue info
Invalid Tx queue_id=0
testpmd: Failed to get tx queue info
Regards,
Ali
Re: [PATCH v4 0/4] Select optional libraries
22/06/2023 11:09, Bruce Richardson: > On Wed, Jun 21, 2023 at 07:00:54PM +0200, David Marchand wrote: > > This series is one implementation to try and please users who want to > > select more easily which parts of DPDK are built. > > > > It introduces a change in behavior for enabling deprecated libraries: > > this series is aimed at the next release but sent early as a demo of > > what changes are required. > > > > If no strong opposition is met, a deprecation notice will be sent for > > v23.07 to announce this change in v23.11. > > > > > > Changes since v3: > > - split kni cleanup, > > - split variable rename cleanup, > > - introduced a new meson option to control deprecated libraries > > activation, > > - simplified the actual implementation of enable_libs to mimic > > enable_drivers behavior, > > > > > I think the first 2 patches should definitely go into 23.07. The latter two > cause a change in how one needs to build DPDK with libs enabled, so we may > want to push that out, and flag the change in advance. Since it's a > relatively minor change, and a build-time one only, I'm fine with it going > in either 23.07 or 23.11, as maintainers decide. > > Series-reviewed-by: Bruce Richardson Applied first 2 patches, thanks.
Re: [PATCH 2/2] lib: make graph optional
20/06/2023 10:20, Jerin Jacob: > On Tue, Jun 20, 2023 at 1:40 PM Bruce Richardson > wrote: > > > > On Mon, Jun 19, 2023 at 10:46:50PM +0200, David Marchand wrote: > > > Allow disabling of the graph library in builds. > > > > > > Signed-off-by: David Marchand > > > --- > > > app/test/meson.build | 12 +++- > > > lib/meson.build | 1 + > > > 2 files changed, 8 insertions(+), 5 deletions(-) > > > > > Acked-by: Bruce Richardson > > Acked-by: Jerin Jacob Series applied, thanks.
Re: [PATCH] mbuf: fix doxygen for distributor metadata
19/06/2023 16:19, Bruce Richardson:
> On Mon, Jun 19, 2023 at 09:37:09AM +0200, David Marchand wrote:
> > On Thu, Jun 15, 2023 at 5:15 PM Bruce Richardson
> > wrote:
> > >
> > > On Thu, Jun 15, 2023 at 03:49:42PM +0200, David Marchand wrote:
> > > > /**< is for post annotations.
> > > >
> > > > Fixes: 839b20be0e9b ("ethdev: support metadata as flow rule criteria")
> > > > Cc: sta...@dpdk.org
> > > >
> > > > Signed-off-by: David Marchand
> > > > ---
> > > > lib/mbuf/rte_mbuf_core.h | 2 +-
> > > > 1 file changed, 1 insertion(+), 1 deletion(-)
> > > >
> > > > diff --git a/lib/mbuf/rte_mbuf_core.h b/lib/mbuf/rte_mbuf_core.h
> > > > index c692c33ec4..2030b3bef9 100644
> > > > --- a/lib/mbuf/rte_mbuf_core.h
> > > > +++ b/lib/mbuf/rte_mbuf_core.h
> > > > @@ -584,8 +584,8 @@ struct rte_mbuf {
> > > >* @see rte_event_eth_tx_adapter_txq_set()
> > > >*/
> > > > } txadapter; /**< Eventdev ethdev Tx adapter */
> > > > - /**< User defined tags. See
> > > > rte_distributor_process() */
> > > > uint32_t usr;
> > > > + /**< User defined tags. See
> > > > rte_distributor_process() */
> > >
> > > Yes, this fixes the issue, but...
> > > I dislike having the comment on the line under the item in question. I
> > > think the post-annotation comments should only be used for comments on the
> > > same line. If the comment is to be on a different line, I think it should
> > > go on the previous line to the item. I also think this tends to be the
> > > style used throughout DPDK generally.
> >
> > I dislike post annotations (regardless of being on the same line or
> > not) as it is easy to mix the description of fields in a structure.
> > But this file has many other usages of this form and, in doubt, I
> > aligned to them.
> >
> > I can send a followup cleanup if you want, is that ok for you?
> >
>
> If this is consistent with what is already done in the file, then I'm ok
> with it as a fix. Follow-up cleanup is obviously welcome, but doesn't have
> to be you doing it.
>
> Acked-by: Bruce Richardson
Applied, thanks.
Re: [PATCH] ci: fix libabigail cache in GHA
20/06/2023 16:21, Aaron Conole:
> David Marchand writes:
>
> > In repositories where multiple branches run the ABI checks using
> > different versions of libabigail (for example, a 22.11 branch using
> > libabigail-1.8 and a main branch using libabigail-2.1), a collision
> > happens on the libabigail binary cache entry.
> > As a single cache entry is used, the content of the cache (let's say the
> > cache was built for libabigail 2.1) won't match what the branch wants to
> > use (in this example running the check for 22.11 branch requires
> > libabigail 1.8).
> > .ci/linux-build.sh then tries to recompile libabigail but it fails as
> > the packages used for building libabigail are missing.
> >
> > Add the version to the cache entry name to avoid this collision.
> >
> > Fixes: 443267090edc ("ci: enable v21 ABI checks")
> > Cc: sta...@dpdk.org
> >
> > Signed-off-by: David Marchand
>
> Acked-by: Aaron Conole
Applied, thanks.
Re: [PATCH v3] bitmap: add scan from offset function
21/06/2023 12:01, Volodymyr Fialko: > Currently, in the case when we search for a bit set after a particular > value, the bitmap has to be scanned from the beginning and > rte_bitmap_scan() has to be called multiple times until we hit the value. > > Add a new rte_bitmap_scan_from_offset() function to initialize scan > state at the given offset and perform scan, this will allow getting > the next set bit after certain offset within one scan call. > > Signed-off-by: Volodymyr Fialko > --- > v2: > - added rte_bitmap_scan_from_offset > v3 > - added note for internal use only for init_at function [...] > +/** > + * @warning > + * @b EXPERIMENTAL: this API may change without prior notice. > + * > + * Bitmap initialize internal scan pointers at the given position for the > scan function. > + * > + * Note: for private/internal use, for public: > + * @see rte_bitmap_scan_from_offset() > + * > + * @param bmp > + * Handle to bitmap instance > + * @param pos > + * Bit position to start scan > + */ > +__rte_experimental > +static inline void > +__rte_bitmap_scan_init_at(struct rte_bitmap *bmp, uint32_t pos) I think it should marked with __rte_internal instead of experimental.
Re: [PATCH] eal: add notes to SMP memory barrier APIs
On 2023-06-21 08:44, Ruifeng Wang wrote: The rte_smp_xx() APIs are deprecated. But it is not mentioned in the function header. Added notes in function header for clarification. Signed-off-by: Ruifeng Wang --- lib/eal/include/generic/rte_atomic.h | 15 +++ 1 file changed, 15 insertions(+) diff --git a/lib/eal/include/generic/rte_atomic.h b/lib/eal/include/generic/rte_atomic.h index 58df843c54..542a2c16ff 100644 --- a/lib/eal/include/generic/rte_atomic.h +++ b/lib/eal/include/generic/rte_atomic.h @@ -55,6 +55,11 @@ static inline void rte_rmb(void); * Guarantees that the LOAD and STORE operations that precede the * rte_smp_mb() call are globally visible across the lcores * before the LOAD and STORE operations that follows it. + * + * @note + * This function is deprecated. It adds complexity to the memory model + * used by this project. C11 memory model should always be used. + * rte_atomic_thread_fence() should be used instead. It's somewhat confusing to learn I should use the C11 memory model, and then in the next sentence that I should call a function which is not in C11. I think it would be helpful to say which memory_model parameters should be used to replace the rte_smp_*mb() calls, and if there are any difference in semantics between the Linux kernel-style barriers and their C11 (near-)equivalents. Is there some particular reason these functions aren't marked __rte_deprecated? Too many warnings? */ static inline void rte_smp_mb(void); @@ -64,6 +69,11 @@ static inline void rte_smp_mb(void); * Guarantees that the STORE operations that precede the * rte_smp_wmb() call are globally visible across the lcores * before the STORE operations that follows it. + * + * @note + * This function is deprecated. It adds complexity to the memory model + * used by this project. C11 memory model should always be used. + * rte_atomic_thread_fence() should be used instead. */ static inline void rte_smp_wmb(void); @@ -73,6 +83,11 @@ static inline void rte_smp_wmb(void); * Guarantees that the LOAD operations that precede the * rte_smp_rmb() call are globally visible across the lcores * before the LOAD operations that follows it. + * + * @note + * This function is deprecated. It adds complexity to the memory model + * used by this project. C11 memory model should always be used. + * rte_atomic_thread_fence() should be used instead. */ static inline void rte_smp_rmb(void); ///@}
Re: [PATCH v3] doc: prefer installing using meson rather than ninja
09/06/2023 16:51, Bruce Richardson: > After doing a build, to install DPDK system-wide our documentation > recommended using the "ninja install" command. However, for anyone > building as a non-root user and only installing as root, the "meson > install" command is a better alternative, as it provides for > automatically dropping or elevating privileges as necessary in more > recent meson releases [1]. > > [1] https://mesonbuild.com/Installing.html#installing-as-the-superuser > > Signed-off-by: Bruce Richardson > > --- > V3: > * correct order of arguments to meson in CI scripts. The "-C" option > must follow the meson "install" command. [This is consistent with > other uses e.g. meson compile -C ..., meson test -C ...] > > V2: > * Fix one missed reference to "ninja install" in Linux GSG > * Changed CI scripts to use "meson install" to ensure step is properly > tested. > --- > .ci/linux-build.sh | 4 ++-- > doc/guides/contributing/coding_style.rst | 2 +- > doc/guides/cryptodevs/uadk.rst | 2 +- > doc/guides/freebsd_gsg/build_dpdk.rst| 2 +- > doc/guides/freebsd_gsg/build_sample_apps.rst | 2 +- > doc/guides/linux_gsg/build_dpdk.rst | 4 ++-- > doc/guides/prog_guide/build-sdk-meson.rst| 4 ++-- > 7 files changed, 10 insertions(+), 10 deletions(-) I see other occurences which could be replaced: .ci/linux-build.sh:[ -d install ] || DESTDIR=$(pwd)/install ninja -C build install devtools/test-meson-builds.sh: echo "DESTDIR=$2 $ninja_cmd -C $1 install" >&$verbose devtools/test-meson-builds.sh: DESTDIR=$2 $ninja_cmd -C $1 install >&$veryverbose doc/guides/nics/mlx4.rst:ninja install doc/guides/platform/mlx5.rst:ninja install
Re: [PATCH v5] lib/net: add MPLS insert and strip functionality
Hi, @olivier.m...@6wind.com
On Sun, 11 Jun 2023, 1:17 am Tanzeel-inline,
wrote:
> None of the foundational NICs currently supports MPLS insertion and
> stripping, this functionality can help users who rely on MPLS in their
> network application.
>
> Signed-off-by: Tanzeel Ahmed
>
> ---
>
> This patch is new version of [PATCH] lib/net: added push MPLS header API.
> I have also added the MPLS strip functionality to address the question
> asked in last patch.
>
> > To be honest, I have some doubts about the usefulness of the patch,
> > especially the function that strips all the MPLS headers.
>
> I believe it serves a practical purpose, in scenarios involving tapped
> traffic where MPLS headers
> have not been stripped.
> While some headers in the lib/net folder have well-defined functions,
> others are limited to their
> structure alone. It would be advantageous to have basic functions
> available for all headers.
>
> > I think the function should only strip the first MPLS header, it is
> > symmetric with the previous function, and more flexible.
>
> You are right, stripping one header is more flexible.
> I updated the function to return 1, in case of stripping last MPLS header.
>
> v5:
> * Updated the MPLS strip function to strip one header at a time.
> * Added the unit test cases.
>
> v4:
> * Removed extra void cast.
> * rte_pktmbuf_append/mtod now return void*.
> The memmove result is casted to rte_ether_hdr*.
>
> v3:
> * fixed patch check failure issue
>
> v2:
> * marked experimental
> * coding style fixed
> * changed rte_memcpy to memcpy
> * mpls header marked as const in parameter
> * added MPLS stripping functionality
> ---
> app/test/meson.build | 2 +
> app/test/test_mpls.c | 180 +++
> lib/net/rte_mpls.h | 106 +
> 3 files changed, 288 insertions(+)
> create mode 100644 app/test/test_mpls.c
>
> diff --git a/app/test/meson.build b/app/test/meson.build
> index f34d19e3c3..548349399f 100644
> --- a/app/test/meson.build
> +++ b/app/test/meson.build
> @@ -95,6 +95,7 @@ test_sources = files(
> 'test_meter.c',
> 'test_mcslock.c',
> 'test_mp_secondary.c',
> +'test_mpls.c',
> 'test_per_lcore.c',
> 'test_pflock.c',
> 'test_pmd_perf.c',
> @@ -205,6 +206,7 @@ fast_tests = [
> ['mempool_autotest', false, true],
> ['memzone_autotest', false, true],
> ['meter_autotest', true, true],
> +['mpls_autotest', false, true],
> ['multiprocess_autotest', false, false],
> ['per_lcore_autotest', true, true],
> ['pflock_autotest', true, true],
> diff --git a/app/test/test_mpls.c b/app/test/test_mpls.c
> new file mode 100644
> index 00..8ff701f6e0
> --- /dev/null
> +++ b/app/test/test_mpls.c
> @@ -0,0 +1,180 @@
> +/* SPDX-License-Identifier: BSD-3-Clause
> + * Copyright(c) 2010-2014 Intel Corporation
> + */
> +
> +#include "test.h"
> +
> +#include
> +#include
> +#include
> +#include
> +
> +#include
> +#include
> +#include
> +#include
> +#include
> +#include
> +#include
> +
> +#define MEMPOOL_CACHE_SIZE 32
> +#define MBUF_DATA_SIZE 2048
> +#define NB_MBUF 128
> +
> +static int
> +test_mpls_fail_push(struct rte_mbuf *m)
> +{
> + struct rte_mpls_hdr mpls;
> +
> + /* create dummy MPLS header */
> + mpls.tag_msb = 1;
> + mpls.tag_lsb = 2;
> + mpls.bs = 1;
> + mpls.tc = 1;
> + mpls.ttl = 255;
> +
> + /* push first MPLS header */
> + if (rte_mpls_push_over_l2(m, &mpls) != 0)
> + return 0;
> + return -1;
> +}
> +
> +static int
> +test_mpls_push(struct rte_mbuf *m)
> +{
> + struct rte_mpls_hdr mpls;
> +
> + /* create dummy MPLS header */
> + mpls.tag_msb = 1;
> + mpls.tag_lsb = 2;
> + mpls.bs = 1;
> + mpls.tc = 1;
> + mpls.ttl = 255;
> +
> + /* push first MPLS header */
> + if (rte_mpls_push_over_l2(m, &mpls) != 0) {
> + printf("Failed to insert mpls 1\n");
> + return -1;
> + }
> + if (rte_pktmbuf_pkt_len(m) != RTE_ETHER_HDR_LEN + RTE_MPLS_HLEN) {
> + printf("Bad pkt length after inserting first mpls
> header\n");
> + return -1;
> + }
> +
> + /* push second MPLS header*/
> + if (rte_mpls_push_over_l2(m, &mpls) != 0) {
> + printf("failed to insert mpls 1\n");
> + return -1;
> + }
> + if (rte_pktmbuf_pkt_len(m) != RTE_ETHER_HDR_LEN + RTE_MPLS_HLEN *
> 2) {
> + printf("bad pkt length after inserting second mpls
> header\n");
> + return -1;
> + }
> + return 0;
> +}
> +
> +static int
> +test_mpls_fail_strip(struct rte_mbuf *m)
> +{
> + /* strip MPLS headers */
> + if (rte_mpls_strip_over_l2(m) != 0)
> + return 0;
> + return -1;
> +}
> +
> +static int
> +test_mpls_strip(struct
Re: [PATCH v5 0/6] replace rte atomics with GCC builtin atomics
I want to report a possible regression from this patch series seen from CI testing on our Intel 82599ES 10G NIC, which we failed to report to patchwork when this initially went under CI due to a bug in our Jenkins reporting scripts. Use of the ixgbe driver appears to be affected. Tyler I apologize for the issues seen with reporting. We've made some temporary changes to avoid this happening again, and are currently reworking our reporting process entirely to provide greater reliability. Here is a DTS snippet showing the issue, and the full log for the failing virtio_smoke test can be downloaded here: https://dpdkdashboard.iol.unh.edu/results/dashboard/patchsets/26560/ 06/06/2023 18:22:58TestVirtioSmoke: Start send packets and verify 06/06/2023 18:22:58 tester: ifconfig enp134s0f0 mtu 9000 06/06/2023 18:22:58 tester: 06/06/2023 18:42:59TestVirtioSmoke: Test Case test_virtio_pvp Result FAILED: TIMEOUT on port start 0 06/06/2023 18:42:59TestVirtioSmoke: port start 0 ixgbe_dev_wait_setup_link_complete(): IXGBE link thread not complete too long time! ixgbe_dev_wait_setup_link_complete(): IXGBE link thread not complete too long time! ixgbe_dev_wait_setup_link_complete(): IXGBE link thread not complete too long time! We initially took this Intel10G testing offline to investigate as we thought it was a lab infra failure. Obviously that wasn't the case, so ideally we will bring this back online when appropriate. But, I don't want to do so right now and start failing everyone's patchseries which are obviously unrelated to this. Comments on this are welcome, otherwise of course I will just return this test coverage to our CI when the state of the git tree allows for it. Apologies for the missing report and the timeline on this. We are taking action to deliver results more reliably going forward. On Fri, Jun 9, 2023 at 11:13 AM Tyler Retzlaff wrote: > On Fri, Jun 09, 2023 at 05:01:53PM +0200, David Marchand wrote: > > On Tue, Jun 6, 2023 at 11:45 PM Tyler Retzlaff > > wrote: > > > > > > Replace the use of rte_atomic.h types and functions, instead use GCC > > > supplied C++11 memory model builtins. > > > > > > This series covers the libraries and drivers that are built on Windows. > > > > > > The code has be converted to use the __atomic builtins but there are > > > additional during conversion I notice that there may be some issues > > > that need to be addressed. > > > > > > I'll comment in the patches where my concerns are so the maintainers > > > may comment. > > > > > > v5: > > > * use relaxed ordering for counter increments in net/ring patch > > > * remove note comments from net/ring patch > > > > > > v4: > > > > > > * drop patch for lib/ring it will be provided by ARM / Honnappa > > > * rebase for changes in dma/idxd merge > > > * adapt __atomic_fetch_sub(...) - 1 == 0 to be > (__atomic_fetch_sub(...) == 1) > > > as per feedback. > > > * drop one /* NOTE: review for potential ordering optimization */ > since > > > the note reference non-critical to perf control path. > > > > > > note: > > > > > > Remainder of the NOTE comments have been retained since there > > > seems to be no consensus but stronger opinion/argument to keep > > > expressed. while I generally agree that changes should not > > > include ``TODO'' style comments I also agree that without these > > > comments in your face people are very unlikely to feel compelled > > > to make the review they are trying to solicit without them. if > > > it is absolute that the series won't be merged with them then I > > > will remove them, but please be explicit soon. > > > > > > v3: > > > * style, don't use c99 comments > > > > > > v2: > > > * comment code where optimizations may be possible now that memory > > > order can be specified. > > > * comment code where operations should potentially be atomic so that > > > maintainers can review. > > > * change a couple of variables labeled as counters to be unsigned. > > > > > > Tyler Retzlaff (6): > > > stack: replace rte atomics with GCC builtin atomics > > > dma/idxd: replace rte atomics with GCC builtin atomics > > > net/ice: replace rte atomics with GCC builtin atomics > > > net/ixgbe: replace rte atomics with GCC builtin atomics > > > net/null: replace rte atomics with GCC builtin atomics > > > net/ring: replace rte atomics with GCC builtin atomics > > > > > > drivers/dma/idxd/idxd_internal.h | 3 +-- > > > drivers/dma/idxd/idxd_pci.c | 11 ++- > > > drivers/net/ice/ice_dcf.c| 1 - > > > drivers/net/ice/ice_dcf_ethdev.c | 1 - > > > drivers/net/ice/ice_ethdev.c | 12 > > > drivers/net/ixgbe/ixgbe_bypass.c | 1 - > > > drivers/net/ixgbe/ixgbe_ethdev.c | 18 -- > > > drivers/net/ixgbe/ixgbe_ethdev.h | 3 ++- > > > drivers/net/ixgbe/ixgbe_flow.c | 1 - > > > drivers/net/ixgbe/ixgbe_rxtx.
Re: [PATCH] net/octeon_ep: support backward compatibility
On Thu, Jun 22, 2023 at 4:48 PM Vamsi Attunuru wrote:
>
> From: Sathesh Edara
>
> Add backward compatibility support between VF
> and PF mailbox messages.
>
> Signed-off-by: Sathesh Edara
> Signed-off-by: Vamsi Attunuru
Applied to dpdk-next-net-mrvl/for-next-net. Thanks
> ---
> drivers/net/octeon_ep/otx_ep_common.h | 3 +++
> drivers/net/octeon_ep/otx_ep_ethdev.c | 6 +
> drivers/net/octeon_ep/otx_ep_mbox.c | 38 ++-
> drivers/net/octeon_ep/otx_ep_mbox.h | 11 ++--
> 4 files changed, 50 insertions(+), 8 deletions(-)
>
> diff --git a/drivers/net/octeon_ep/otx_ep_common.h
> b/drivers/net/octeon_ep/otx_ep_common.h
> index 42aa065a3a..c150cbe619 100644
> --- a/drivers/net/octeon_ep/otx_ep_common.h
> +++ b/drivers/net/octeon_ep/otx_ep_common.h
> @@ -538,6 +538,9 @@ struct otx_ep_device {
>
> /* Mailbox receive message length */
> int32_t mbox_rcv_message_len;
> +
> + /* Negotiated Mbox version */
> + uint32_t mbox_neg_ver;
> };
>
> int otx_ep_setup_iqs(struct otx_ep_device *otx_ep, uint32_t iq_no,
> diff --git a/drivers/net/octeon_ep/otx_ep_ethdev.c
> b/drivers/net/octeon_ep/otx_ep_ethdev.c
> index a9868909f8..57b965ad06 100644
> --- a/drivers/net/octeon_ep/otx_ep_ethdev.c
> +++ b/drivers/net/octeon_ep/otx_ep_ethdev.c
> @@ -666,6 +666,12 @@ otx_ep_eth_dev_init(struct rte_eth_dev *eth_dev)
> otx_epvf->port_id = eth_dev->data->port_id;
> eth_dev->dev_ops = &otx_ep_eth_dev_ops;
> rte_spinlock_init(&otx_epvf->mbox_lock);
> +
> + /*
> +* Initialize negotiated Mbox version to base version of VF Mbox
> +* This will address working legacy PF with latest VF.
> +*/
> + otx_epvf->mbox_neg_ver = OTX_EP_MBOX_VERSION_V1;
> eth_dev->data->mac_addrs = rte_zmalloc("otx_ep", RTE_ETHER_ADDR_LEN,
> 0);
> if (eth_dev->data->mac_addrs == NULL) {
> otx_ep_err("MAC addresses memory allocation failed\n");
> diff --git a/drivers/net/octeon_ep/otx_ep_mbox.c
> b/drivers/net/octeon_ep/otx_ep_mbox.c
> index 1ad36e14c8..4118645dc7 100644
> --- a/drivers/net/octeon_ep/otx_ep_mbox.c
> +++ b/drivers/net/octeon_ep/otx_ep_mbox.c
> @@ -12,6 +12,14 @@
> #include "cnxk_ep_vf.h"
> #include "otx_ep_mbox.h"
>
> +/*
> + * When a new command is implemented, the below table should be updated
> + * with new command and it's version info.
> + */
> +static uint32_t otx_ep_cmd_versions[OTX_EP_MBOX_CMD_MAX] = {
> + [0 ... OTX_EP_MBOX_CMD_DEV_REMOVE] = OTX_EP_MBOX_VERSION_V1
> +};
> +
> static int
> __otx_ep_send_mbox_cmd(struct otx_ep_device *otx_ep,
>union otx_ep_mbox_word cmd,
> @@ -56,6 +64,12 @@ otx_ep_send_mbox_cmd(struct otx_ep_device *otx_ep,
> int ret;
>
> rte_spinlock_lock(&otx_ep->mbox_lock);
> + if (otx_ep_cmd_versions[cmd.s.opcode] > otx_ep->mbox_neg_ver) {
> + otx_ep_dbg("CMD:%d not supported in Version:%d\n",
> cmd.s.opcode,
> + otx_ep->mbox_neg_ver);
> + rte_spinlock_unlock(&otx_ep->mbox_lock);
> + return -EOPNOTSUPP;
> + }
> ret = __otx_ep_send_mbox_cmd(otx_ep, cmd, rsp);
> rte_spinlock_unlock(&otx_ep->mbox_lock);
> return ret;
> @@ -284,15 +298,27 @@ int otx_ep_mbox_version_check(struct rte_eth_dev
> *eth_dev)
>
> cmd.u64 = 0;
> cmd.s_version.opcode = OTX_EP_MBOX_CMD_VERSION;
> - cmd.s_version.version = OTX_EP_MBOX_VERSION;
> + cmd.s_version.version = OTX_EP_MBOX_VERSION_CURRENT;
> ret = otx_ep_send_mbox_cmd(otx_ep, cmd, &rsp);
> - if (!ret)
> - return 0;
> - if (ret == OTX_EP_MBOX_CMD_STATUS_NACK) {
> - otx_ep_err("VF Mbox version:%u is not compatible with PF\n",
> +
> + /*
> +* VF receives NACK or version info as zero
> +* only if PF driver running old version of Mailbox
> +* In this case VF mailbox version fallbacks to base
> +* mailbox vesrion OTX_EP_MBOX_VERSION_V1.
> +* Default VF mbox_neg_ver is set to OTX_EP_MBOX_VERSION_V1
> +* during initialization of PMD driver.
> +*/
> + if (ret == OTX_EP_MBOX_CMD_STATUS_NACK || rsp.s_version.version == 0)
> {
> + otx_ep_dbg("VF Mbox version fallback to base version
> from:%u\n",
> (uint32_t)cmd.s_version.version);
> + return 0;
> }
> - return ret;
> + otx_ep->mbox_neg_ver = (uint32_t)rsp.s_version.version;
> + otx_ep_dbg("VF Mbox version:%u Negotiated VF version with PF:%u\n",
> + (uint32_t)cmd.s_version.version,
> + (uint32_t)rsp.s_version.version);
> + return 0;
> }
>
> int otx_ep_mbox_send_dev_exit(struct rte_eth_dev *eth_dev)
> diff --git a/drivers/net/octeon_ep/otx_ep_mbox.h
> b/drivers/net/octeon_ep/otx_ep_mbox.h
> index 9df3c53edd..a3fc15cca7 100644
> --- a/drivers/net/octeon_ep/otx_ep_mbox.h
> +++ b/driv
[PATCH v8 0/4] net/sfc: support VLAN stripping offload
This patch series adds VLAN stripping offload. Note that this offload are device level offload. v8: * fix minor issues * update the release notes v7: * raise an error if there is no callback for efo_get_count * fix alignment * remove the extra check v6: * highlight that efx_port_vlan_strip_set() must be called before any filter insertion * avoid an extra check if offload is not requested v5: * fixed problems with naming * fixed problems with abbreviations * fixed problems with isolated mode * fixed problems with consistency v4: * fix apply patch failure warning v3: * fix apply patch failure warning v2: * rebase patches on top of dpdk-next-net/main Artemii Morozov (4): common/sfc_efx/base: report VLAN stripping capability common/sfc_efx/base: add API to get installed filters count common/sfc_efx/base: add support to enable VLAN stripping net/sfc: support VLAN stripping offload doc/guides/nics/sfc_efx.rst | 4 +-- doc/guides/rel_notes/release_23_07.rst| 2 ++ drivers/common/sfc_efx/base/ef10_filter.c | 26 +++ drivers/common/sfc_efx/base/ef10_impl.h | 6 drivers/common/sfc_efx/base/ef10_nic.c| 6 drivers/common/sfc_efx/base/efx.h | 13 drivers/common/sfc_efx/base/efx_filter.c | 32 +++ drivers/common/sfc_efx/base/efx_impl.h| 8 + drivers/common/sfc_efx/base/efx_port.c| 39 +++ drivers/common/sfc_efx/base/efx_rx.c | 14 drivers/common/sfc_efx/base/rhead_rx.c| 3 ++ drivers/common/sfc_efx/base/siena_nic.c | 1 + drivers/common/sfc_efx/version.map| 1 + drivers/net/sfc/sfc.h | 1 + drivers/net/sfc/sfc_dp_rx.h | 1 + drivers/net/sfc/sfc_ef100_rx.c| 16 +- drivers/net/sfc/sfc_port.c| 11 +++ drivers/net/sfc/sfc_rx.c | 10 ++ 18 files changed, 191 insertions(+), 3 deletions(-) -- 2.34.1
[PATCH v8 1/4] common/sfc_efx/base: report VLAN stripping capability
These changes are necessary in order to add support for stripping
VLAN tags in the future.
Signed-off-by: Artemii Morozov
Reviewed-by: Ivan Malov
Reviewed-by: Andy Moreton
Acked-by: Andrew Rybchenko
---
drivers/common/sfc_efx/base/ef10_nic.c | 6 ++
drivers/common/sfc_efx/base/efx.h | 1 +
drivers/common/sfc_efx/base/siena_nic.c | 1 +
3 files changed, 8 insertions(+)
diff --git a/drivers/common/sfc_efx/base/ef10_nic.c
b/drivers/common/sfc_efx/base/ef10_nic.c
index bf9cb9d309..79d596b5ef 100644
--- a/drivers/common/sfc_efx/base/ef10_nic.c
+++ b/drivers/common/sfc_efx/base/ef10_nic.c
@@ -1223,6 +1223,12 @@ ef10_get_datapath_caps(
else
encp->enc_hw_tx_insert_vlan_enabled = B_FALSE;
+ /* Check if firmware supports VLAN stripping. */
+ if (CAP_FLAGS1(req, RX_VLAN_STRIPPING))
+ encp->enc_rx_vlan_stripping_supported = B_TRUE;
+ else
+ encp->enc_rx_vlan_stripping_supported = B_FALSE;
+
/* Check if the firmware supports RX event batching */
if (CAP_FLAGS1(req, RX_BATCHING))
encp->enc_rx_batching_enabled = B_TRUE;
diff --git a/drivers/common/sfc_efx/base/efx.h
b/drivers/common/sfc_efx/base/efx.h
index ef626cc55a..7fcf48f5e1 100644
--- a/drivers/common/sfc_efx/base/efx.h
+++ b/drivers/common/sfc_efx/base/efx.h
@@ -1627,6 +1627,7 @@ typedef struct efx_nic_cfg_s {
/* Number of TSO contexts on the NIC (FATSOv2) */
uint32_tenc_fw_assisted_tso_v2_n_contexts;
boolean_t enc_hw_tx_insert_vlan_enabled;
+ boolean_t enc_rx_vlan_stripping_supported;
/* Number of PFs on the NIC */
uint32_tenc_hw_pf_count;
/* Datapath firmware vadapter/vport/vswitch support */
diff --git a/drivers/common/sfc_efx/base/siena_nic.c
b/drivers/common/sfc_efx/base/siena_nic.c
index ca38eefa7e..e334f01b03 100644
--- a/drivers/common/sfc_efx/base/siena_nic.c
+++ b/drivers/common/sfc_efx/base/siena_nic.c
@@ -179,6 +179,7 @@ siena_board_cfg(
(encp->enc_rxq_limit * EFX_RXQ_DC_NDESCS(EFX_RXQ_DC_SIZE));
encp->enc_hw_tx_insert_vlan_enabled = B_FALSE;
+ encp->enc_rx_vlan_stripping_supported = B_FALSE;
encp->enc_fw_assisted_tso_enabled = B_FALSE;
encp->enc_fw_assisted_tso_v2_enabled = B_FALSE;
encp->enc_fw_assisted_tso_v2_n_contexts = 0;
--
2.34.1
[PATCH v8 2/4] common/sfc_efx/base: add API to get installed filters count
This API allows to get number of installed filters. This will
be used in the future patches.
Signed-off-by: Artemii Morozov
Acked-by: Andrew Rybchenko
---
drivers/common/sfc_efx/base/ef10_filter.c | 20 ++
drivers/common/sfc_efx/base/ef10_impl.h | 6 +
drivers/common/sfc_efx/base/efx_filter.c | 32 +++
drivers/common/sfc_efx/base/efx_impl.h| 7 +
4 files changed, 65 insertions(+)
diff --git a/drivers/common/sfc_efx/base/ef10_filter.c
b/drivers/common/sfc_efx/base/ef10_filter.c
index d6940011c0..8ceaf98a5f 100644
--- a/drivers/common/sfc_efx/base/ef10_filter.c
+++ b/drivers/common/sfc_efx/base/ef10_filter.c
@@ -2113,6 +2113,26 @@ ef10_filter_reconfigure(
return (rc);
}
+ __checkReturn efx_rc_t
+ef10_filter_get_count(
+ __inefx_nic_t *enp,
+ __out uint32_t *countp)
+{
+ ef10_filter_table_t *table = enp->en_filter.ef_ef10_filter_table;
+ uint32_t filter_count;
+
+ EFSYS_ASSERT(EFX_FAMILY_IS_EF100(enp) || EFX_FAMILY_IS_EF10(enp));
+ EFSYS_ASSERT(countp != NULL);
+
+ filter_count = table->eft_unicst_filter_count +
+ table->eft_mulcst_filter_count +
+ table->eft_encap_filter_count;
+
+ *countp = filter_count;
+
+ return (0);
+}
+
void
ef10_filter_get_default_rxq(
__inefx_nic_t *enp,
diff --git a/drivers/common/sfc_efx/base/ef10_impl.h
b/drivers/common/sfc_efx/base/ef10_impl.h
index 877aedad45..3476f274ce 100644
--- a/drivers/common/sfc_efx/base/ef10_impl.h
+++ b/drivers/common/sfc_efx/base/ef10_impl.h
@@ -1347,6 +1347,12 @@ ef10_filter_reconfigure(
__in_ecount(6*count)uint8_t const *addrs,
__inuint32_t count);
+LIBEFX_INTERNAL
+extern __checkReturn efx_rc_t
+ef10_filter_get_count(
+ __inefx_nic_t *enp,
+ __out uint32_t *countp);
+
LIBEFX_INTERNAL
extern void
ef10_filter_get_default_rxq(
diff --git a/drivers/common/sfc_efx/base/efx_filter.c
b/drivers/common/sfc_efx/base/efx_filter.c
index 83c37ff859..2c8c7bdc33 100644
--- a/drivers/common/sfc_efx/base/efx_filter.c
+++ b/drivers/common/sfc_efx/base/efx_filter.c
@@ -53,6 +53,7 @@ static const efx_filter_ops_t __efx_filter_siena_ops = {
siena_filter_delete,/* efo_delete */
siena_filter_supported_filters, /* efo_supported_filters */
NULL, /* efo_reconfigure */
+ NULL, /* efo_get_count */
};
#endif /* EFSYS_OPT_SIENA */
@@ -65,6 +66,7 @@ static const efx_filter_ops_t __efx_filter_ef10_ops = {
ef10_filter_delete, /* efo_delete */
ef10_filter_supported_filters, /* efo_supported_filters */
ef10_filter_reconfigure,/* efo_reconfigure */
+ ef10_filter_get_count, /* efo_get_count */
};
#endif /* EFX_OPTS_EF10() */
@@ -77,6 +79,7 @@ static const efx_filter_ops_t __efx_filter_rhead_ops = {
ef10_filter_delete, /* efo_delete */
ef10_filter_supported_filters, /* efo_supported_filters */
ef10_filter_reconfigure,/* efo_reconfigure */
+ ef10_filter_get_count, /* efo_get_count */
};
#endif /* EFSYS_OPT_RIVERHEAD */
@@ -309,6 +312,35 @@ efx_filter_reconfigure(
fail1:
EFSYS_PROBE1(fail1, efx_rc_t, rc);
+ return (rc);
+}
+
+ __checkReturn efx_rc_t
+efx_filter_get_count(
+ __inefx_nic_t *enp,
+ __out uint32_t *countp)
+{
+ efx_rc_t rc;
+
+ EFSYS_ASSERT3U(enp->en_magic, ==, EFX_NIC_MAGIC);
+ EFSYS_ASSERT3U(enp->en_mod_flags, &, EFX_MOD_PROBE);
+ EFSYS_ASSERT3U(enp->en_mod_flags, &, EFX_MOD_FILTER);
+
+ if (enp->en_efop->efo_get_count == NULL) {
+ rc = ENOTSUP;
+ goto fail1;
+ }
+
+ if ((rc = enp->en_efop->efo_get_count(enp, countp)) != 0)
+ goto fail2;
+
+ return (0);
+
+fail2:
+ EFSYS_PROBE(fail2);
+fail1:
+ EFSYS_PROBE1(fail1, efx_rc_t, rc);
+
return (rc);
}
diff --git a/drivers/common/sfc_efx/base/efx_impl.h
b/drivers/common/sfc_efx/base/efx_impl.h
index 92a30c34ae..9a387a7efc 100644
--- a/drivers/common/sfc_efx/base/efx_impl.h
+++ b/drivers/common/sfc_efx/base/efx_impl.h
@@ -288,6 +288,7 @@ typedef struct efx_filter_ops_s {
efx_rc_t(*efo_reconfigure)(efx_nic_t *, uint8_t const *,
boolean_t,
boolean_t, boolean_t, boolean_t,
uint8_t const *, uint32_t);
+ efx_rc_t(*efo_get_count)(efx_nic_t *, uint32_t *);
} efx_filter_ops_t;
LIBEFX_INTERNAL
@@ -302,6 +303,12 @@ efx_filter_reconfigure(
__in_ecount(6*count)uint8_t const *addrs,
__inuint32_t count);
+LIBEFX_INTERNAL
+extern __checkReturn efx_rc_t
+efx_filter_get_count(
+ __inefx_nic_t *enp,
+
[PATCH v8 3/4] common/sfc_efx/base: add support to enable VLAN stripping
To enable VLAN stripping, two conditions must be met:
the corresponding flag must be set and the appropriate
Rx prefix should be requested.
VLAN stripping is supported on EF100.
Signed-off-by: Artemii Morozov
Reviewed-by: Ivan Malov
Reviewed-by: Andy Moreton
Acked-by: Andrew Rybchenko
---
drivers/common/sfc_efx/base/ef10_filter.c | 6
drivers/common/sfc_efx/base/efx.h | 12 +++
drivers/common/sfc_efx/base/efx_impl.h| 1 +
drivers/common/sfc_efx/base/efx_port.c| 39 +++
drivers/common/sfc_efx/base/efx_rx.c | 14
drivers/common/sfc_efx/base/rhead_rx.c| 3 ++
drivers/common/sfc_efx/version.map| 1 +
7 files changed, 76 insertions(+)
diff --git a/drivers/common/sfc_efx/base/ef10_filter.c
b/drivers/common/sfc_efx/base/ef10_filter.c
index 8ceaf98a5f..2a10720122 100644
--- a/drivers/common/sfc_efx/base/ef10_filter.c
+++ b/drivers/common/sfc_efx/base/ef10_filter.c
@@ -171,6 +171,7 @@ efx_mcdi_filter_op_add(
EFX_MCDI_DECLARE_BUF(payload, MC_CMD_FILTER_OP_V3_IN_LEN,
MC_CMD_FILTER_OP_EXT_OUT_LEN);
efx_filter_match_flags_t match_flags;
+ efx_port_t *epp = &(enp->en_port);
uint32_t port_id;
efx_rc_t rc;
@@ -338,6 +339,11 @@ efx_mcdi_filter_op_add(
FILTER_OP_V3_IN_MATCH_SET_FLAG, 1);
}
+ if (epp->ep_vlan_strip) {
+ MCDI_IN_SET_DWORD_FIELD(req, FILTER_OP_V3_IN_MATCH_ACTION_FLAGS,
+ FILTER_OP_V3_IN_MATCH_STRIP_VLAN, 1);
+ }
+
efx_mcdi_execute(enp, &req);
if (req.emr_rc != 0) {
diff --git a/drivers/common/sfc_efx/base/efx.h
b/drivers/common/sfc_efx/base/efx.h
index 7fcf48f5e1..77f855bfb0 100644
--- a/drivers/common/sfc_efx/base/efx.h
+++ b/drivers/common/sfc_efx/base/efx.h
@@ -1158,6 +1158,12 @@ efx_port_poll(
__inefx_nic_t *enp,
__out_opt efx_link_mode_t *link_modep);
+LIBEFX_API
+extern __checkReturn efx_rc_t
+efx_port_vlan_strip_set(
+ __inefx_nic_t *enp,
+ __inboolean_t enabled);
+
LIBEFX_API
extern void
efx_port_fini(
@@ -3117,6 +3123,12 @@ typedef enum efx_rxq_type_e {
* Request user flag field in the Rx prefix of a queue.
*/
#defineEFX_RXQ_FLAG_USER_FLAG 0x20
+/*
+ * Request VLAN TCI field in the Rx prefix. The flag just
+ * controls delivery of the stripped VLAN TCI if VLAN stripping
+ * is enabled and done.
+ */
+#defineEFX_RXQ_FLAG_VLAN_STRIPPED_TCI 0x40
LIBEFX_API
extern __checkReturn efx_rc_t
diff --git a/drivers/common/sfc_efx/base/efx_impl.h
b/drivers/common/sfc_efx/base/efx_impl.h
index 9a387a7efc..662a21e90c 100644
--- a/drivers/common/sfc_efx/base/efx_impl.h
+++ b/drivers/common/sfc_efx/base/efx_impl.h
@@ -371,6 +371,7 @@ typedef struct efx_port_s {
uint32_tep_phy_cap_mask;
boolean_t ep_mac_drain;
boolean_t ep_include_fcs;
+ boolean_t ep_vlan_strip;
#if EFSYS_OPT_BIST
efx_bist_type_t ep_current_bist;
#endif
diff --git a/drivers/common/sfc_efx/base/efx_port.c
b/drivers/common/sfc_efx/base/efx_port.c
index a5f982e335..389efb2fe9 100644
--- a/drivers/common/sfc_efx/base/efx_port.c
+++ b/drivers/common/sfc_efx/base/efx_port.c
@@ -204,6 +204,45 @@ efx_loopback_type_name(
#endif /* EFSYS_OPT_LOOPBACK */
+ __checkReturn efx_rc_t
+efx_port_vlan_strip_set(
+ __inefx_nic_t *enp,
+ __inboolean_t enabled)
+{
+ efx_port_t *epp = &(enp->en_port);
+ efx_nic_cfg_t *encp = &(enp->en_nic_cfg);
+ uint32_t filter_count = 0;
+ efx_rc_t rc;
+
+ EFSYS_ASSERT3U(enp->en_magic, ==, EFX_NIC_MAGIC);
+
+ if (enabled && !encp->enc_rx_vlan_stripping_supported) {
+ rc = ENOTSUP;
+ goto fail1;
+ }
+
+ if ((rc = efx_filter_get_count(enp, &filter_count)) != 0)
+ goto fail2;
+
+ if (filter_count != 0) {
+ rc = EINVAL;
+ goto fail3;
+ }
+
+ epp->ep_vlan_strip = enabled;
+
+ return (0);
+
+fail3:
+ EFSYS_PROBE(fail3);
+fail2:
+ EFSYS_PROBE(fail2);
+fail1:
+ EFSYS_PROBE1(fail1, efx_rc_t, rc);
+
+ return (rc);
+}
+
void
efx_port_fini(
__inefx_nic_t *enp)
diff --git a/drivers/common/sfc_efx/base/efx_rx.c
b/drivers/common/sfc_efx/base/efx_rx.c
index 68f42f5cac..b3d9e14c67 100644
--- a/drivers/common/sfc_efx/base/efx_rx.c
+++ b/drivers/common/sfc_efx/base/efx_rx.c
@@ -941,11 +941,25 @@ efx_rx_qcreate_internal(
goto fail5;
}
+ if (flags & EFX_RXQ_FLAG_VLAN_STRIPPED_TCI) {
+ const efx_rx_prefix_layout_t *erplp = &erp->er_prefix_layout;
+ const efx_rx_prefix_field_info_t *vlan_tci_field;
+
+ vlan_tci_field =
+ &erplp->erpl_fields[EFX_RX
[PATCH v8 4/4] net/sfc: support VLAN stripping offload
Extract VLAN TCI provided by the HW in the prefix and put it to mbuf.
VLAN stripping is supported for ef100 datapath only. This is device
level offload.
Signed-off-by: Artemii Morozov
Reviewed-by: Viacheslav Galaktionov
Reviewed-by: Ivan Malov
Reviewed-by: Andy Moreton
Acked-by: Andrew Rybchenko
---
doc/guides/nics/sfc_efx.rst| 4 ++--
doc/guides/rel_notes/release_23_07.rst | 2 ++
drivers/net/sfc/sfc.h | 1 +
drivers/net/sfc/sfc_dp_rx.h| 1 +
drivers/net/sfc/sfc_ef100_rx.c | 16 +++-
drivers/net/sfc/sfc_port.c | 11 +++
drivers/net/sfc/sfc_rx.c | 10 ++
7 files changed, 42 insertions(+), 3 deletions(-)
diff --git a/doc/guides/nics/sfc_efx.rst b/doc/guides/nics/sfc_efx.rst
index 24459da33e..eafb88191a 100644
--- a/doc/guides/nics/sfc_efx.rst
+++ b/doc/guides/nics/sfc_efx.rst
@@ -122,6 +122,8 @@ SFC EFX PMD has support for:
- Port representors (see :ref: switch_representation)
+- VLAN stripping (if running firmware variant supports it)
+
Non-supported Features
--
@@ -134,8 +136,6 @@ The features not yet supported include:
- VLAN filtering
-- VLAN stripping
-
- LRO
diff --git a/doc/guides/rel_notes/release_23_07.rst
b/doc/guides/rel_notes/release_23_07.rst
index 738d35d9f7..1d5e58eea2 100644
--- a/doc/guides/rel_notes/release_23_07.rst
+++ b/doc/guides/rel_notes/release_23_07.rst
@@ -147,6 +147,8 @@ New Features
* Added support for keeping CRC.
+ * Added VLAN stripping support on SN1000 SmartNICs.
+
* **Added vmxnet3 version 7 support.**
Added support for vmxnet3 version 7 which includes support
diff --git a/drivers/net/sfc/sfc.h b/drivers/net/sfc/sfc.h
index 25cdeaa5cd..2432a2307e 100644
--- a/drivers/net/sfc/sfc.h
+++ b/drivers/net/sfc/sfc.h
@@ -75,6 +75,7 @@ struct sfc_port {
unsigned intflow_ctrl;
boolean_t flow_ctrl_autoneg;
boolean_t include_fcs;
+ boolean_t vlan_strip;
size_t pdu;
/*
diff --git a/drivers/net/sfc/sfc_dp_rx.h b/drivers/net/sfc/sfc_dp_rx.h
index 8a504bdcf1..9f9bf28988 100644
--- a/drivers/net/sfc/sfc_dp_rx.h
+++ b/drivers/net/sfc/sfc_dp_rx.h
@@ -70,6 +70,7 @@ struct sfc_dp_rx_qcreate_info {
unsigned intflags;
#define SFC_RXQ_FLAG_RSS_HASH 0x1
#define SFC_RXQ_FLAG_INGRESS_MPORT 0x2
+#define SFC_RXQ_FLAG_VLAN_STRIPPED_TCI 0x4
/** Rx queue size */
unsigned intrxq_entries;
diff --git a/drivers/net/sfc/sfc_ef100_rx.c b/drivers/net/sfc/sfc_ef100_rx.c
index 5563bd9a0b..2677003da3 100644
--- a/drivers/net/sfc/sfc_ef100_rx.c
+++ b/drivers/net/sfc/sfc_ef100_rx.c
@@ -68,6 +68,7 @@ struct sfc_ef100_rxq {
#define SFC_EF100_RXQ_INGRESS_MPORT0x80
#define SFC_EF100_RXQ_USER_FLAG0x100
#define SFC_EF100_RXQ_NIC_DMA_MAP 0x200
+#define SFC_EF100_RXQ_VLAN_STRIPPED_TCI0x400
unsigned intptr_mask;
unsigned intevq_phase_bit_shift;
unsigned intready_pkts;
@@ -392,6 +393,7 @@ static const efx_rx_prefix_layout_t
sfc_ef100_rx_prefix_layout = {
SFC_EF100_RX_PREFIX_FIELD(RSS_HASH, B_FALSE),
SFC_EF100_RX_PREFIX_FIELD(USER_FLAG, B_FALSE),
SFC_EF100_RX_PREFIX_FIELD(USER_MARK, B_FALSE),
+ SFC_EF100_RX_PREFIX_FIELD(VLAN_STRIP_TCI, B_FALSE),
#undef SFC_EF100_RX_PREFIX_FIELD
}
@@ -472,6 +474,14 @@ sfc_ef100_rx_prefix_to_offloads(const struct sfc_ef100_rxq
*rxq,
ESF_GZ_RX_PREFIX_INGRESS_MPORT);
}
+ if (rxq->flags & SFC_EF100_RXQ_VLAN_STRIPPED_TCI &&
+ EFX_TEST_XWORD_BIT(rx_prefix[0],
+ ESF_GZ_RX_PREFIX_VLAN_STRIPPED_LBN)) {
+ ol_flags |= RTE_MBUF_F_RX_VLAN | RTE_MBUF_F_RX_VLAN_STRIPPED;
+ m->vlan_tci = EFX_XWORD_FIELD(rx_prefix[0],
+
ESF_GZ_RX_PREFIX_VLAN_STRIP_TCI);
+ }
+
m->ol_flags = ol_flags;
return true;
}
@@ -813,6 +823,9 @@ sfc_ef100_rx_qcreate(uint16_t port_id, uint16_t queue_id,
if (info->flags & SFC_RXQ_FLAG_INGRESS_MPORT)
rxq->flags |= SFC_EF100_RXQ_INGRESS_MPORT;
+ if (info->flags & SFC_RXQ_FLAG_VLAN_STRIPPED_TCI)
+ rxq->flags |= SFC_EF100_RXQ_VLAN_STRIPPED_TCI;
+
sfc_ef100_rx_debug(rxq, "RxQ doorbell is %p", rxq->doorbell);
*dp_rxqp = &rxq->dp;
@@ -1004,7 +1017,8 @@ struct sfc_dp_rx sfc_ef100_rx = {
SFC_DP_RX_FEAT_FLOW_MARK |
SFC_DP_RX_FEAT_INTR |
SFC_DP_RX_FEAT_STATS,
- .dev_offload_capa = RTE_ETH_RX_OFFLOAD_KEEP_CRC,
+ .dev_offload_capa = RTE_ETH_RX_OFFLO
RE: [EXT] [PATCH v8] app/dma-perf: introduce dma-perf application
Hi Cheng,
Thanks for the new version. Please see inline.
Thanks,
Anoob
> -Original Message-
> From: Cheng Jiang
> Sent: Tuesday, June 20, 2023 12:24 PM
> To: tho...@monjalon.net; bruce.richard...@intel.com;
> m...@smartsharesystems.com; chenbo@intel.com; Amit Prakash Shukla
> ; Anoob Joseph ;
> huangdeng...@huawei.com; kevin.la...@intel.com;
> fengcheng...@huawei.com; Jerin Jacob Kollanukkaran
>
> Cc: dev@dpdk.org; jiayu...@intel.com; xuan.d...@intel.com;
> wenwux...@intel.com; yuanx.w...@intel.com; xingguang...@intel.com;
> weix.l...@intel.com; Cheng Jiang
> Subject: [EXT] [PATCH v8] app/dma-perf: introduce dma-perf application
>
> External Email
>
> --
> There are many high-performance DMA devices supported in DPDK now,
> and these DMA devices can also be integrated into other modules of DPDK as
> accelerators, such as Vhost. Before integrating DMA into applications,
> developers need to know the performance of these DMA devices in various
> scenarios and the performance of CPUs in the same scenario, such as
> different buffer lengths. Only in this way can we know the target
> performance of the application accelerated by using them. This patch
> introduces a high-performance testing tool, which supports comparing the
> performance of CPU and DMA in different scenarios automatically with a pre-
> set config file. Memory Copy performance test are supported for now.
>
> Signed-off-by: Cheng Jiang
> Signed-off-by: Jiayu Hu
> Signed-off-by: Yuan Wang
> Acked-by: Morten Brørup
> Acked-by: Chenbo Xia
> ---
> v8:
> fixed string copy issue in parse_lcore();
> improved some data display format;
> added doc in doc/guides/tools;
> updated release notes;
>
> v7:
> fixed some strcpy issues;
> removed cache setup in calling rte_pktmbuf_pool_create();
> fixed some typos;
> added some memory free and null set operations;
> improved result calculation;
> v6:
> improved code based on Anoob's comments;
> fixed some code structure issues;
> v5:
> fixed some LONG_LINE warnings;
> v4:
> fixed inaccuracy of the memory footprint display;
> v3:
> fixed some typos;
> v2:
> added lcore/dmadev designation;
> added error case process;
> removed worker_threads parameter from config.ini;
> improved the logs;
> improved config file;
>
> app/meson.build| 1 +
> app/test-dma-perf/benchmark.c | 498 +
> app/test-dma-perf/config.ini | 61 +++
> app/test-dma-perf/main.c | 594 +
> app/test-dma-perf/main.h | 69 +++
> app/test-dma-perf/meson.build | 17 +
> doc/guides/rel_notes/release_23_07.rst | 6 +
> doc/guides/tools/dmaperf.rst | 103 +
> doc/guides/tools/index.rst | 1 +
> 9 files changed, 1350 insertions(+)
> create mode 100644 app/test-dma-perf/benchmark.c create mode 100644
> app/test-dma-perf/config.ini create mode 100644 app/test-dma-
> perf/main.c create mode 100644 app/test-dma-perf/main.h create mode
> 100644 app/test-dma-perf/meson.build create mode 100644
> doc/guides/tools/dmaperf.rst
>
> +/* Configuration of device. */
> +static void
> +configure_dmadev_queue(uint32_t dev_id, uint32_t ring_size) {
> + uint16_t vchan = 0;
> + struct rte_dma_info info;
> + struct rte_dma_conf dev_config = { .nb_vchans = 1 };
> + struct rte_dma_vchan_conf qconf = {
> + .direction = RTE_DMA_DIR_MEM_TO_MEM,
> + .nb_desc = ring_size
> + };
> +
> + if (rte_dma_configure(dev_id, &dev_config) != 0)
> + rte_exit(EXIT_FAILURE, "Error with dma configure.\n");
> +
> + if (rte_dma_vchan_setup(dev_id, vchan, &qconf) != 0)
> + rte_exit(EXIT_FAILURE, "Error with queue configuration.\n");
> +
> + rte_dma_info_get(dev_id, &info);
[Anoob] This API can return errors. Better to add handling.
> + if (info.nb_vchans != 1)
> + rte_exit(EXIT_FAILURE, "Error, no configured queues
> reported on device id. %u\n",
> + dev_id);
> +
> + if (rte_dma_start(dev_id) != 0)
> + rte_exit(EXIT_FAILURE, "Error with dma start.\n"); }
> +
> +static int
> +config_dmadevs(struct test_configure *cfg) {
> + uint32_t ring_size = cfg->ring_size.cur;
> + struct lcore_dma_map_t *ldm = &cfg->lcore_dma_map;
> + uint32_t nb_workers = ldm->cnt;
> + uint32_t i;
> + int dev_id;
> + uint16_t nb_dmadevs = 0;
> + char *dma_name;
> +
> + for (i = 0; i < ldm->cnt; i++) {
> + dma_name = ldm->dma_names[i];
> + dev_id = rte_dma_get_dev_id_by_name(dma_name);
> + if (dev_id == -1) {
[Anoob] Can you check the above API definition? I think it returns not just -1
in case of errors.
> + fprintf(stderr, "Error: Fail to find DMA %s.\n",
> dma_name);
> + goto end;
>
