diff mbox series

[net-next,v4,7/8] netdev: add queue stats

Message ID 20240314085459.115933-8-xuanzhuo@linux.alibaba.com (mailing list archive)
State Superseded
Delegated to: Netdev Maintainers
Headers show
Series virtio-net: support device stats | expand

Checks

Context Check Description
netdev/series_format success Posting correctly formatted
netdev/tree_selection success Clearly marked for net-next
netdev/ynl fail Tree is dirty after regen; no warnings/errors; no diff in generated;
netdev/fixes_present success Fixes tag not required for -next series
netdev/header_inline success No static functions without inline keyword in header files
netdev/build_32bit success Errors and warnings before: 4886 this patch: 4886
netdev/build_tools success Errors and warnings before: 0 this patch: 0
netdev/cc_maintainers success CCed 8 of 8 maintainers
netdev/build_clang success Errors and warnings before: 1062 this patch: 1062
netdev/verify_signedoff success Signed-off-by tag matches author and committer
netdev/deprecated_api success None detected
netdev/check_selftest success No net selftest shell script
netdev/verify_fixes success No Fixes tag
netdev/build_allmodconfig_warn success Errors and warnings before: 5170 this patch: 5170
netdev/checkpatch warning WARNING: line length of 100 exceeds 80 columns WARNING: line length of 82 exceeds 80 columns WARNING: line length of 86 exceeds 80 columns WARNING: line length of 90 exceeds 80 columns WARNING: line length of 94 exceeds 80 columns WARNING: line length of 96 exceeds 80 columns
netdev/build_clang_rust success No Rust files in patch. Skipping build
netdev/kdoc success Errors and warnings before: 0 this patch: 0
netdev/source_inline success Was 0 now: 0

Commit Message

Xuan Zhuo March 14, 2024, 8:54 a.m. UTC
These stats are commonly. Support reporting those via netdev-genl queue
stats.

name: rx-hw-drops
name: rx-hw-drop-overruns
name: rx-csum-unnecessary
name: rx-csum-none
name: rx-csum-bad
name: rx-hw-gro-packets
name: rx-hw-gro-bytes
name: rx-hw-gro-wire-packets
name: rx-hw-gro-wire-bytes
name: rx-hw-drop-ratelimits
name: tx-hw-drops
name: tx-hw-drop-errors
name: tx-csum-none
name: tx-needs-csum
name: tx-hw-gso-packets
name: tx-hw-gso-bytes
name: tx-hw-gso-wire-packets
name: tx-hw-gso-wire-bytes
name: tx-hw-drop-ratelimits

Signed-off-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
---
 Documentation/netlink/specs/netdev.yaml | 105 ++++++++++++++++++++++++
 include/net/netdev_queues.h             |  27 ++++++
 include/uapi/linux/netdev.h             |  20 +++++
 net/core/netdev-genl.c                  |  23 +++++-
 tools/include/uapi/linux/netdev.h       |  19 +++++
 5 files changed, 192 insertions(+), 2 deletions(-)

Comments

Jakub Kicinski March 14, 2024, 11:21 p.m. UTC | #1
On Thu, 14 Mar 2024 16:54:58 +0800 Xuan Zhuo wrote:
> +      -
> +        name: rx-hw-drops
> +        doc: |
> +          Number of packets that arrived at the device but never left it,
> +          encompassing packets dropped for reasons such as insufficient buffer

s/encompassing/including/

> +          space, processing errors, as well as those affected by explicitly
> +          defined policies and packet filtering criteria.

I'm afraid the attempt to "correct my English" ended up decreasing the
clarity. Maybe use what I suggested more directly?

  Number of all packets which entered the device, but never left it,
  including but not limited to: packets dropped due to lack of buffer
  space, processing errors, explicit or implicit policies and packet filters.

> +        type: uint
> +      -
> +        name: rx-hw-drop-overruns
> +        doc: |
> +          Number of packets that arrived at the device but never left it due to
> +          no descriptors.

I put a paragraph in my previous email explaining how tricky this is to
state for HW devices and you just went with the virtio spec :|

  Number of packets dropped due to transient lack of resources, such as
  buffer space, host descriptors etc.

> +      -
> +        name: rx-csum-bad
> +        doc: |
> +          Number of packets that are dropped by device due to invalid checksum.

Quoting myself:

>> Maybe add a note in "bad" that packets with bad csum are not
>> discarded, but still delivered to the stack.

Devices should not drop packets due to invalid L4 checksums.

> +        type: uint
> +      -
> +        name: rx-hw-gro-packets
> +        doc: |
> +          Number of packets that area coalesced from smaller packets by the device,
> +          that do not cover LRO packets.

s/area/were/ ?
s/that do not cover LRO packets/Counts only packets coalesced with the
HW-GRO netdevice feature, LRO-coalesced packets are not counted./

> +        type: uint
> +      -
> +        name: rx-hw-gro-bytes
> +        doc: see `rx-hw-gro-packets`.
> +        type: uint
> +      -
> +        name: rx-hw-gro-wire-packets
> +        doc: |
> +          Number of packets that area coalesced to bigger packets(no LRO packets)
> +          by the device.

s/area/were/
s/packets(no LRO packets)/packets by the HW-GRO feature of the device/

> +        type: uint
> +      -
> +        name: rx-hw-gro-wire-bytes
> +        doc: see `rx-hw-gro-wire-packets`.

Please make sure the "See `xyz`." references are capitalized (See)
and end with a full stop.

> +        type: uint
> +      -
> +        name: rx-hw-drop-ratelimits
> +        doc: |
> +          Number of the packets dropped by the device due to the received
> +          packets bitrate exceeding the device speed limit.

Maybe s/speed/rate/

> +        type: uint
> +      -
> +        name: tx-hw-drops
> +        doc: |
> +          Number of packets that arrived at the device but never left it,
> +          encompassing packets dropped for reasons such as processing errors, as
> +          well as those affected by explicitly defined policies and packet
> +          filtering criteria.
> +        type: uint
> +      -
> +        name: tx-hw-drop-errors
> +        doc: Number of packets dropped because they were invalid or malformed.
> +        type: uint
> +      -
> +        name: tx-csum-none
> +        doc: |
> +          Number of packets that do not require the device to calculate the
> +          checksum.

I think we should use past tense everywhere.

> +        type: uint
> +      -
> +        name: tx-needs-csum
> +        doc: |
> +          Number of packets that require the device to calculate the
> +          checksum.
> +        type: uint
> +      -
> +        name: tx-hw-gso-packets
> +        doc: |
> +          Number of packets that necessitated segmentation into smaller packets
> +          by the device.
> +        type: uint
> +      -
> +        name: tx-hw-gso-bytes
> +        doc: |
> +          Successfully segmented into smaller packets by the device, see
> +          `tx-hw-gso-packets`.

Maybe stick to "see `tx-hw-gso-packets`", none of the other counters
mention the send being "successful".

> +        type: uint
> +      -
> +        name: tx-hw-gso-wire-packets
> +        doc: |
> +          Number of the small packets that segmented from bigger packets by the
> +          device.

Maybe

  Number of wire-sized packets generated by processing
  `tx-hw-gso-packets`

?

> +        type: uint
> +      -
> +        name: tx-hw-gso-wire-bytes
> +        doc: See `tx-hw-gso-wire-packets`.
> +
> +        type: uint
> +      -
> +        name: tx-hw-drop-ratelimits
> +        doc: |
> +          Number of the packets dropped by the device due to the transmit
> +          packets bitrate exceeding the device speed limit.

s/speed/rate/

> +        type: uint
>  
>  operations:
>    list:
> diff --git a/include/net/netdev_queues.h b/include/net/netdev_queues.h
> index 1ec408585373..c7ac4539eafc 100644
> --- a/include/net/netdev_queues.h
> +++ b/include/net/netdev_queues.h
> @@ -9,11 +9,38 @@ struct netdev_queue_stats_rx {
>  	u64 bytes;
>  	u64 packets;
>  	u64 alloc_fail;
> +
> +	u64 hw_drops;
> +	u64 hw_drop_overruns;
> +
> +	u64 csum_unnecessary;
> +	u64 csum_none;
> +	u64 csum_bad;
> +
> +	u64 hw_gro_packets;
> +	u64 hw_gro_bytes;
> +	u64 hw_gro_wire_packets;
> +	u64 hw_gro_wire_bytes;
> +
> +	u64 hw_drop_ratelimits;
>  };
>  
>  struct netdev_queue_stats_tx {
>  	u64 bytes;
>  	u64 packets;
> +
> +	u64 hw_drops;
> +	u64 hw_drop_errors;
> +
> +	u64 csum_none;
> +	u64 needs_csum;
> +
> +	u64 hw_gso_packets;
> +	u64 hw_gso_bytes;
> +	u64 hw_gso_wire_packets;
> +	u64 hw_gso_wire_bytes;
> +
> +	u64 hw_drop_ratelimits;
>  };
>  
>  /**
> diff --git a/include/uapi/linux/netdev.h b/include/uapi/linux/netdev.h
> index bb65ee840cda..702d69b09f14 100644
> --- a/include/uapi/linux/netdev.h
> +++ b/include/uapi/linux/netdev.h
> @@ -147,6 +147,26 @@ enum {
>  	NETDEV_A_QSTATS_TX_BYTES,
>  	NETDEV_A_QSTATS_RX_ALLOC_FAIL,
>  

Looks hand written. Once you update the yaml file, you should run:

./tools/net/ynl/ynl-regen.sh

This will generate the uAPI changes for you.

> +	NETDEV_A_QSTATS_RX_HW_DROPS,
> +	NETDEV_A_QSTATS_RX_HW_DROP_OVERRUNS,
> +	NETDEV_A_QSTATS_RX_CSUM_UNNECESSARY,
> +	NETDEV_A_QSTATS_RX_CSUM_NONE,
> +	NETDEV_A_QSTATS_RX_CSUM_BAD,
> +	NETDEV_A_QSTATS_RX_HW_GRO_PACKETS,
> +	NETDEV_A_QSTATS_RX_HW_GRO_BYTES,
> +	NETDEV_A_QSTATS_RX_HW_GRO_WIRE_PACKETS,
> +	NETDEV_A_QSTATS_RX_HW_GRO_WIRE_BYTES,
> +	NETDEV_A_QSTATS_RX_HW_DROP_RATELIMITS,
> +	NETDEV_A_QSTATS_TX_HW_DROPS,
> +	NETDEV_A_QSTATS_TX_HW_DROP_ERRORS,
> +	NETDEV_A_QSTATS_TX_CSUM_NONE,
> +	NETDEV_A_QSTATS_TX_NEEDS_CSUM,
> +	NETDEV_A_QSTATS_TX_HW_GSO_PACKETS,
> +	NETDEV_A_QSTATS_TX_HW_GSO_BYTES,
> +	NETDEV_A_QSTATS_TX_HW_GSO_WIRE_PACKETS,
> +	NETDEV_A_QSTATS_TX_HW_GSO_WIRE_BYTES,
> +	NETDEV_A_QSTATS_TX_HW_DROP_RATELIMITS,
> +
>  	__NETDEV_A_QSTATS_MAX,
>  	NETDEV_A_QSTATS_MAX = (__NETDEV_A_QSTATS_MAX - 1)
>  };
Xuan Zhuo March 15, 2024, 8:11 a.m. UTC | #2
On Thu, 14 Mar 2024 16:21:42 -0700, Jakub Kicinski <kuba@kernel.org> wrote:
> On Thu, 14 Mar 2024 16:54:58 +0800 Xuan Zhuo wrote:
> > +      -
> > +        name: rx-hw-drops
> > +        doc: |
> > +          Number of packets that arrived at the device but never left it,
> > +          encompassing packets dropped for reasons such as insufficient buffer
>
> s/encompassing/including/
>
> > +          space, processing errors, as well as those affected by explicitly
> > +          defined policies and packet filtering criteria.
>
> I'm afraid the attempt to "correct my English" ended up decreasing the
> clarity.

I apologize for the misunderstanding; that was not my intention. Your original
phrasing is indeed clear and comprehensive. Let's use your suggested
description.

> Maybe use what I suggested more directly?
>
>   Number of all packets which entered the device, but never left it,
>   including but not limited to: packets dropped due to lack of buffer
>   space, processing errors, explicit or implicit policies and packet filters.
>
> > +        type: uint
> > +      -
> > +        name: rx-hw-drop-overruns
> > +        doc: |
> > +          Number of packets that arrived at the device but never left it due to
> > +          no descriptors.
>
> I put a paragraph in my previous email explaining how tricky this is to
> state for HW devices and you just went with the virtio spec :|

Sorry, not native speaker, I misunderstood.

Other is ok. I will fix in next version.

Thanks.

>
>   Number of packets dropped due to transient lack of resources, such as
>   buffer space, host descriptors etc.
>
> > +      -
> > +        name: rx-csum-bad
> > +        doc: |
> > +          Number of packets that are dropped by device due to invalid checksum.
>
> Quoting myself:
>
> >> Maybe add a note in "bad" that packets with bad csum are not
> >> discarded, but still delivered to the stack.
>
> Devices should not drop packets due to invalid L4 checksums.
>
> > +        type: uint
> > +      -
> > +        name: rx-hw-gro-packets
> > +        doc: |
> > +          Number of packets that area coalesced from smaller packets by the device,
> > +          that do not cover LRO packets.
>
> s/area/were/ ?
> s/that do not cover LRO packets/Counts only packets coalesced with the
> HW-GRO netdevice feature, LRO-coalesced packets are not counted./
>
> > +        type: uint
> > +      -
> > +        name: rx-hw-gro-bytes
> > +        doc: see `rx-hw-gro-packets`.
> > +        type: uint
> > +      -
> > +        name: rx-hw-gro-wire-packets
> > +        doc: |
> > +          Number of packets that area coalesced to bigger packets(no LRO packets)
> > +          by the device.
>
> s/area/were/
> s/packets(no LRO packets)/packets by the HW-GRO feature of the device/
>
> > +        type: uint
> > +      -
> > +        name: rx-hw-gro-wire-bytes
> > +        doc: see `rx-hw-gro-wire-packets`.
>
> Please make sure the "See `xyz`." references are capitalized (See)
> and end with a full stop.
>
> > +        type: uint
> > +      -
> > +        name: rx-hw-drop-ratelimits
> > +        doc: |
> > +          Number of the packets dropped by the device due to the received
> > +          packets bitrate exceeding the device speed limit.
>
> Maybe s/speed/rate/
>
> > +        type: uint
> > +      -
> > +        name: tx-hw-drops
> > +        doc: |
> > +          Number of packets that arrived at the device but never left it,
> > +          encompassing packets dropped for reasons such as processing errors, as
> > +          well as those affected by explicitly defined policies and packet
> > +          filtering criteria.
> > +        type: uint
> > +      -
> > +        name: tx-hw-drop-errors
> > +        doc: Number of packets dropped because they were invalid or malformed.
> > +        type: uint
> > +      -
> > +        name: tx-csum-none
> > +        doc: |
> > +          Number of packets that do not require the device to calculate the
> > +          checksum.
>
> I think we should use past tense everywhere.
>
> > +        type: uint
> > +      -
> > +        name: tx-needs-csum
> > +        doc: |
> > +          Number of packets that require the device to calculate the
> > +          checksum.
> > +        type: uint
> > +      -
> > +        name: tx-hw-gso-packets
> > +        doc: |
> > +          Number of packets that necessitated segmentation into smaller packets
> > +          by the device.
> > +        type: uint
> > +      -
> > +        name: tx-hw-gso-bytes
> > +        doc: |
> > +          Successfully segmented into smaller packets by the device, see
> > +          `tx-hw-gso-packets`.
>
> Maybe stick to "see `tx-hw-gso-packets`", none of the other counters
> mention the send being "successful".
>
> > +        type: uint
> > +      -
> > +        name: tx-hw-gso-wire-packets
> > +        doc: |
> > +          Number of the small packets that segmented from bigger packets by the
> > +          device.
>
> Maybe
>
>   Number of wire-sized packets generated by processing
>   `tx-hw-gso-packets`
>
> ?
>
> > +        type: uint
> > +      -
> > +        name: tx-hw-gso-wire-bytes
> > +        doc: See `tx-hw-gso-wire-packets`.
> > +
> > +        type: uint
> > +      -
> > +        name: tx-hw-drop-ratelimits
> > +        doc: |
> > +          Number of the packets dropped by the device due to the transmit
> > +          packets bitrate exceeding the device speed limit.
>
> s/speed/rate/
>
> > +        type: uint
> >
> >  operations:
> >    list:
> > diff --git a/include/net/netdev_queues.h b/include/net/netdev_queues.h
> > index 1ec408585373..c7ac4539eafc 100644
> > --- a/include/net/netdev_queues.h
> > +++ b/include/net/netdev_queues.h
> > @@ -9,11 +9,38 @@ struct netdev_queue_stats_rx {
> >  	u64 bytes;
> >  	u64 packets;
> >  	u64 alloc_fail;
> > +
> > +	u64 hw_drops;
> > +	u64 hw_drop_overruns;
> > +
> > +	u64 csum_unnecessary;
> > +	u64 csum_none;
> > +	u64 csum_bad;
> > +
> > +	u64 hw_gro_packets;
> > +	u64 hw_gro_bytes;
> > +	u64 hw_gro_wire_packets;
> > +	u64 hw_gro_wire_bytes;
> > +
> > +	u64 hw_drop_ratelimits;
> >  };
> >
> >  struct netdev_queue_stats_tx {
> >  	u64 bytes;
> >  	u64 packets;
> > +
> > +	u64 hw_drops;
> > +	u64 hw_drop_errors;
> > +
> > +	u64 csum_none;
> > +	u64 needs_csum;
> > +
> > +	u64 hw_gso_packets;
> > +	u64 hw_gso_bytes;
> > +	u64 hw_gso_wire_packets;
> > +	u64 hw_gso_wire_bytes;
> > +
> > +	u64 hw_drop_ratelimits;
> >  };
> >
> >  /**
> > diff --git a/include/uapi/linux/netdev.h b/include/uapi/linux/netdev.h
> > index bb65ee840cda..702d69b09f14 100644
> > --- a/include/uapi/linux/netdev.h
> > +++ b/include/uapi/linux/netdev.h
> > @@ -147,6 +147,26 @@ enum {
> >  	NETDEV_A_QSTATS_TX_BYTES,
> >  	NETDEV_A_QSTATS_RX_ALLOC_FAIL,
> >
>
> Looks hand written. Once you update the yaml file, you should run:
>
> ./tools/net/ynl/ynl-regen.sh
>
> This will generate the uAPI changes for you.
>
> > +	NETDEV_A_QSTATS_RX_HW_DROPS,
> > +	NETDEV_A_QSTATS_RX_HW_DROP_OVERRUNS,
> > +	NETDEV_A_QSTATS_RX_CSUM_UNNECESSARY,
> > +	NETDEV_A_QSTATS_RX_CSUM_NONE,
> > +	NETDEV_A_QSTATS_RX_CSUM_BAD,
> > +	NETDEV_A_QSTATS_RX_HW_GRO_PACKETS,
> > +	NETDEV_A_QSTATS_RX_HW_GRO_BYTES,
> > +	NETDEV_A_QSTATS_RX_HW_GRO_WIRE_PACKETS,
> > +	NETDEV_A_QSTATS_RX_HW_GRO_WIRE_BYTES,
> > +	NETDEV_A_QSTATS_RX_HW_DROP_RATELIMITS,
> > +	NETDEV_A_QSTATS_TX_HW_DROPS,
> > +	NETDEV_A_QSTATS_TX_HW_DROP_ERRORS,
> > +	NETDEV_A_QSTATS_TX_CSUM_NONE,
> > +	NETDEV_A_QSTATS_TX_NEEDS_CSUM,
> > +	NETDEV_A_QSTATS_TX_HW_GSO_PACKETS,
> > +	NETDEV_A_QSTATS_TX_HW_GSO_BYTES,
> > +	NETDEV_A_QSTATS_TX_HW_GSO_WIRE_PACKETS,
> > +	NETDEV_A_QSTATS_TX_HW_GSO_WIRE_BYTES,
> > +	NETDEV_A_QSTATS_TX_HW_DROP_RATELIMITS,
> > +
> >  	__NETDEV_A_QSTATS_MAX,
> >  	NETDEV_A_QSTATS_MAX = (__NETDEV_A_QSTATS_MAX - 1)
> >  };
>
diff mbox series

Patch

diff --git a/Documentation/netlink/specs/netdev.yaml b/Documentation/netlink/specs/netdev.yaml
index 76352dbd2be4..41ba22cf1123 100644
--- a/Documentation/netlink/specs/netdev.yaml
+++ b/Documentation/netlink/specs/netdev.yaml
@@ -335,6 +335,111 @@  attribute-sets:
           Allocation failure may, or may not result in a packet drop, depending
           on driver implementation and whether system recovers quickly.
         type: uint
+      -
+        name: rx-hw-drops
+        doc: |
+          Number of packets that arrived at the device but never left it,
+          encompassing packets dropped for reasons such as insufficient buffer
+          space, processing errors, as well as those affected by explicitly
+          defined policies and packet filtering criteria.
+        type: uint
+      -
+        name: rx-hw-drop-overruns
+        doc: |
+          Number of packets that arrived at the device but never left it due to
+          no descriptors.
+        type: uint
+      -
+        name: rx-csum-unnecessary
+        doc: Number of packets that are marked as CHECKSUM_UNNECESSARY.
+        type: uint
+      -
+        name: rx-csum-none
+        doc: Number of packets that are not checksummed by device.
+        type: uint
+      -
+        name: rx-csum-bad
+        doc: |
+          Number of packets that are dropped by device due to invalid checksum.
+        type: uint
+      -
+        name: rx-hw-gro-packets
+        doc: |
+          Number of packets that area coalesced from smaller packets by the device,
+          that do not cover LRO packets.
+        type: uint
+      -
+        name: rx-hw-gro-bytes
+        doc: see `rx-hw-gro-packets`.
+        type: uint
+      -
+        name: rx-hw-gro-wire-packets
+        doc: |
+          Number of packets that area coalesced to bigger packets(no LRO packets)
+          by the device.
+        type: uint
+      -
+        name: rx-hw-gro-wire-bytes
+        doc: see `rx-hw-gro-wire-packets`.
+        type: uint
+      -
+        name: rx-hw-drop-ratelimits
+        doc: |
+          Number of the packets dropped by the device due to the received
+          packets bitrate exceeding the device speed limit.
+        type: uint
+      -
+        name: tx-hw-drops
+        doc: |
+          Number of packets that arrived at the device but never left it,
+          encompassing packets dropped for reasons such as processing errors, as
+          well as those affected by explicitly defined policies and packet
+          filtering criteria.
+        type: uint
+      -
+        name: tx-hw-drop-errors
+        doc: Number of packets dropped because they were invalid or malformed.
+        type: uint
+      -
+        name: tx-csum-none
+        doc: |
+          Number of packets that do not require the device to calculate the
+          checksum.
+        type: uint
+      -
+        name: tx-needs-csum
+        doc: |
+          Number of packets that require the device to calculate the
+          checksum.
+        type: uint
+      -
+        name: tx-hw-gso-packets
+        doc: |
+          Number of packets that necessitated segmentation into smaller packets
+          by the device.
+        type: uint
+      -
+        name: tx-hw-gso-bytes
+        doc: |
+          Successfully segmented into smaller packets by the device, see
+          `tx-hw-gso-packets`.
+        type: uint
+      -
+        name: tx-hw-gso-wire-packets
+        doc: |
+          Number of the small packets that segmented from bigger packets by the
+          device.
+        type: uint
+      -
+        name: tx-hw-gso-wire-bytes
+        doc: See `tx-hw-gso-wire-packets`.
+        type: uint
+      -
+        name: tx-hw-drop-ratelimits
+        doc: |
+          Number of the packets dropped by the device due to the transmit
+          packets bitrate exceeding the device speed limit.
+        type: uint
 
 operations:
   list:
diff --git a/include/net/netdev_queues.h b/include/net/netdev_queues.h
index 1ec408585373..c7ac4539eafc 100644
--- a/include/net/netdev_queues.h
+++ b/include/net/netdev_queues.h
@@ -9,11 +9,38 @@  struct netdev_queue_stats_rx {
 	u64 bytes;
 	u64 packets;
 	u64 alloc_fail;
+
+	u64 hw_drops;
+	u64 hw_drop_overruns;
+
+	u64 csum_unnecessary;
+	u64 csum_none;
+	u64 csum_bad;
+
+	u64 hw_gro_packets;
+	u64 hw_gro_bytes;
+	u64 hw_gro_wire_packets;
+	u64 hw_gro_wire_bytes;
+
+	u64 hw_drop_ratelimits;
 };
 
 struct netdev_queue_stats_tx {
 	u64 bytes;
 	u64 packets;
+
+	u64 hw_drops;
+	u64 hw_drop_errors;
+
+	u64 csum_none;
+	u64 needs_csum;
+
+	u64 hw_gso_packets;
+	u64 hw_gso_bytes;
+	u64 hw_gso_wire_packets;
+	u64 hw_gso_wire_bytes;
+
+	u64 hw_drop_ratelimits;
 };
 
 /**
diff --git a/include/uapi/linux/netdev.h b/include/uapi/linux/netdev.h
index bb65ee840cda..702d69b09f14 100644
--- a/include/uapi/linux/netdev.h
+++ b/include/uapi/linux/netdev.h
@@ -147,6 +147,26 @@  enum {
 	NETDEV_A_QSTATS_TX_BYTES,
 	NETDEV_A_QSTATS_RX_ALLOC_FAIL,
 
+	NETDEV_A_QSTATS_RX_HW_DROPS,
+	NETDEV_A_QSTATS_RX_HW_DROP_OVERRUNS,
+	NETDEV_A_QSTATS_RX_CSUM_UNNECESSARY,
+	NETDEV_A_QSTATS_RX_CSUM_NONE,
+	NETDEV_A_QSTATS_RX_CSUM_BAD,
+	NETDEV_A_QSTATS_RX_HW_GRO_PACKETS,
+	NETDEV_A_QSTATS_RX_HW_GRO_BYTES,
+	NETDEV_A_QSTATS_RX_HW_GRO_WIRE_PACKETS,
+	NETDEV_A_QSTATS_RX_HW_GRO_WIRE_BYTES,
+	NETDEV_A_QSTATS_RX_HW_DROP_RATELIMITS,
+	NETDEV_A_QSTATS_TX_HW_DROPS,
+	NETDEV_A_QSTATS_TX_HW_DROP_ERRORS,
+	NETDEV_A_QSTATS_TX_CSUM_NONE,
+	NETDEV_A_QSTATS_TX_NEEDS_CSUM,
+	NETDEV_A_QSTATS_TX_HW_GSO_PACKETS,
+	NETDEV_A_QSTATS_TX_HW_GSO_BYTES,
+	NETDEV_A_QSTATS_TX_HW_GSO_WIRE_PACKETS,
+	NETDEV_A_QSTATS_TX_HW_GSO_WIRE_BYTES,
+	NETDEV_A_QSTATS_TX_HW_DROP_RATELIMITS,
+
 	__NETDEV_A_QSTATS_MAX,
 	NETDEV_A_QSTATS_MAX = (__NETDEV_A_QSTATS_MAX - 1)
 };
diff --git a/net/core/netdev-genl.c b/net/core/netdev-genl.c
index 7004b3399c2b..a2bf9af2dcf6 100644
--- a/net/core/netdev-genl.c
+++ b/net/core/netdev-genl.c
@@ -489,7 +489,17 @@  netdev_nl_stats_write_rx(struct sk_buff *rsp, struct netdev_queue_stats_rx *rx)
 {
 	if (netdev_stat_put(rsp, NETDEV_A_QSTATS_RX_PACKETS, rx->packets) ||
 	    netdev_stat_put(rsp, NETDEV_A_QSTATS_RX_BYTES, rx->bytes) ||
-	    netdev_stat_put(rsp, NETDEV_A_QSTATS_RX_ALLOC_FAIL, rx->alloc_fail))
+	    netdev_stat_put(rsp, NETDEV_A_QSTATS_RX_ALLOC_FAIL, rx->alloc_fail) ||
+	    netdev_stat_put(rsp, NETDEV_A_QSTATS_RX_HW_DROPS, rx->hw_drops) ||
+	    netdev_stat_put(rsp, NETDEV_A_QSTATS_RX_HW_DROP_OVERRUNS, rx->hw_drop_overruns) ||
+	    netdev_stat_put(rsp, NETDEV_A_QSTATS_RX_CSUM_UNNECESSARY, rx->csum_unnecessary) ||
+	    netdev_stat_put(rsp, NETDEV_A_QSTATS_RX_CSUM_NONE, rx->csum_none) ||
+	    netdev_stat_put(rsp, NETDEV_A_QSTATS_RX_CSUM_BAD, rx->csum_bad) ||
+	    netdev_stat_put(rsp, NETDEV_A_QSTATS_RX_HW_GRO_PACKETS, rx->hw_gro_packets) ||
+	    netdev_stat_put(rsp, NETDEV_A_QSTATS_RX_HW_GRO_BYTES, rx->hw_gro_bytes) ||
+	    netdev_stat_put(rsp, NETDEV_A_QSTATS_RX_HW_GRO_WIRE_PACKETS, rx->hw_gro_wire_packets) ||
+	    netdev_stat_put(rsp, NETDEV_A_QSTATS_RX_HW_GRO_WIRE_BYTES, rx->hw_gro_wire_bytes) ||
+	    netdev_stat_put(rsp, NETDEV_A_QSTATS_RX_HW_DROP_RATELIMITS, rx->hw_drop_ratelimits))
 		return -EMSGSIZE;
 	return 0;
 }
@@ -498,7 +508,16 @@  static int
 netdev_nl_stats_write_tx(struct sk_buff *rsp, struct netdev_queue_stats_tx *tx)
 {
 	if (netdev_stat_put(rsp, NETDEV_A_QSTATS_TX_PACKETS, tx->packets) ||
-	    netdev_stat_put(rsp, NETDEV_A_QSTATS_TX_BYTES, tx->bytes))
+	    netdev_stat_put(rsp, NETDEV_A_QSTATS_TX_BYTES, tx->bytes) ||
+	    netdev_stat_put(rsp, NETDEV_A_QSTATS_TX_HW_DROPS, tx->hw_drops) ||
+	    netdev_stat_put(rsp, NETDEV_A_QSTATS_TX_HW_DROP_ERRORS, tx->hw_drop_errors) ||
+	    netdev_stat_put(rsp, NETDEV_A_QSTATS_TX_CSUM_NONE, tx->csum_none) ||
+	    netdev_stat_put(rsp, NETDEV_A_QSTATS_TX_NEEDS_CSUM, tx->needs_csum) ||
+	    netdev_stat_put(rsp, NETDEV_A_QSTATS_TX_HW_GSO_PACKETS, tx->hw_gso_packets) ||
+	    netdev_stat_put(rsp, NETDEV_A_QSTATS_TX_HW_GSO_BYTES, tx->hw_gso_bytes) ||
+	    netdev_stat_put(rsp, NETDEV_A_QSTATS_TX_HW_GSO_WIRE_PACKETS, tx->hw_gso_wire_packets) ||
+	    netdev_stat_put(rsp, NETDEV_A_QSTATS_TX_HW_GSO_WIRE_BYTES, tx->hw_gso_wire_bytes) ||
+	    netdev_stat_put(rsp, NETDEV_A_QSTATS_TX_HW_DROP_RATELIMITS, tx->hw_drop_ratelimits))
 		return -EMSGSIZE;
 	return 0;
 }
diff --git a/tools/include/uapi/linux/netdev.h b/tools/include/uapi/linux/netdev.h
index bb65ee840cda..cf24f1d9adf8 100644
--- a/tools/include/uapi/linux/netdev.h
+++ b/tools/include/uapi/linux/netdev.h
@@ -146,6 +146,25 @@  enum {
 	NETDEV_A_QSTATS_TX_PACKETS,
 	NETDEV_A_QSTATS_TX_BYTES,
 	NETDEV_A_QSTATS_RX_ALLOC_FAIL,
+	NETDEV_A_QSTATS_RX_HW_DROPS,
+	NETDEV_A_QSTATS_RX_HW_DROP_OVERRUNS,
+	NETDEV_A_QSTATS_RX_CSUM_UNNECESSARY,
+	NETDEV_A_QSTATS_RX_CSUM_NONE,
+	NETDEV_A_QSTATS_RX_CSUM_BAD,
+	NETDEV_A_QSTATS_RX_HW_GRO_PACKETS,
+	NETDEV_A_QSTATS_RX_HW_GRO_BYTES,
+	NETDEV_A_QSTATS_RX_HW_GRO_WIRE_PACKETS,
+	NETDEV_A_QSTATS_RX_HW_GRO_WIRE_BYTES,
+	NETDEV_A_QSTATS_RX_HW_DROP_RATELIMITS,
+	NETDEV_A_QSTATS_TX_HW_DROPS,
+	NETDEV_A_QSTATS_TX_HW_DROP_ERRORS,
+	NETDEV_A_QSTATS_TX_CSUM_NONE,
+	NETDEV_A_QSTATS_TX_NEEDS_CSUM,
+	NETDEV_A_QSTATS_TX_HW_GSO_PACKETS,
+	NETDEV_A_QSTATS_TX_HW_GSO_BYTES,
+	NETDEV_A_QSTATS_TX_HW_GSO_WIRE_PACKETS,
+	NETDEV_A_QSTATS_TX_HW_GSO_WIRE_BYTES,
+	NETDEV_A_QSTATS_TX_HW_DROP_RATELIMITS,
 
 	__NETDEV_A_QSTATS_MAX,
 	NETDEV_A_QSTATS_MAX = (__NETDEV_A_QSTATS_MAX - 1)