[v2,net-next,1/5] net: dsa: microchip: add rmon grouping for ethtool statistics
Commit Message
Add support for ethtool standard device statistics grouping. Support rmon
statistics grouping using rmon groups parameter in ethtool command. rmon
provides packet size based range grouping. Common mib parameters are used
across all KSZ series swtches for packet size statistics, except for
KSZ8830. KSZ series have mib counters for packets with size:
- less than 64 Bytes,
- 65 to 127 Bytes,
- 128 to 255 Bytes,
- 256 to 511 Bytes,
- 512 to 1023 Bytes,
- 1024 to 1522 Bytes,
- 1523 to 2000 Bytes and
- More than 2001 Bytes
KSZ8830 have mib counters upto 1024-1522 range only. Since no other change,
common range used across all KSZ series, but used upto only upto 1024-1522
for KSZ8830.
Co-developed-by: Thangaraj Samynathan <Thangaraj.S@microchip.com>
Signed-off-by: Rakesh Sankaranarayanan <rakesh.sankaranarayanan@microchip.com>
---
drivers/net/dsa/microchip/Makefile | 1 +
drivers/net/dsa/microchip/ksz_common.c | 15 ++
drivers/net/dsa/microchip/ksz_common.h | 3 +
drivers/net/dsa/microchip/ksz_ethtool.c | 180 ++++++++++++++++++++++++
drivers/net/dsa/microchip/ksz_ethtool.h | 18 +++
5 files changed, 217 insertions(+)
create mode 100644 drivers/net/dsa/microchip/ksz_ethtool.c
create mode 100644 drivers/net/dsa/microchip/ksz_ethtool.h
Comments
From: Rakesh Sankaranarayanan <rakesh.sankaranarayanan@microchip.com>
Date: Fri, 17 Feb 2023 16:32:07 +0530
> Add support for ethtool standard device statistics grouping. Support rmon
> statistics grouping using rmon groups parameter in ethtool command. rmon
> provides packet size based range grouping. Common mib parameters are used
> across all KSZ series swtches for packet size statistics, except for
> KSZ8830. KSZ series have mib counters for packets with size:
[...]
> +void ksz8_get_rmon_stats(struct ksz_device *dev, int port,
> + struct ethtool_rmon_stats *rmon_stats,
> + const struct ethtool_rmon_hist_range **ranges)
> +{
> + struct ksz_port_mib *mib;
> + u64 *cnt;
Nit: I guess it can be const since you only read it (in every such
callback)?
> + u8 i;
> +
> + mib = &dev->ports[port].mib;
> +
> + mutex_lock(&mib->cnt_mutex);
> +
> + cnt = &mib->counters[KSZ8_RX_UNDERSIZE];
> + dev->dev_ops->r_mib_pkt(dev, port, KSZ8_RX_UNDERSIZE, NULL, cnt);
> + rmon_stats->undersize_pkts = *cnt;
> +
> + cnt = &mib->counters[KSZ8_RX_OVERSIZE];
> + dev->dev_ops->r_mib_pkt(dev, port, KSZ8_RX_OVERSIZE, NULL, cnt);
> + rmon_stats->oversize_pkts = *cnt;
> +
> + cnt = &mib->counters[KSZ8_RX_FRAGMENTS];
> + dev->dev_ops->r_mib_pkt(dev, port, KSZ8_RX_FRAGMENTS, NULL, cnt);
> + rmon_stats->fragments = *cnt;
> +
> + cnt = &mib->counters[KSZ8_RX_JABBERS];
> + dev->dev_ops->r_mib_pkt(dev, port, KSZ8_RX_JABBERS, NULL, cnt);
> + rmon_stats->jabbers = *cnt;
> +
> + for (i = 0; i < KSZ8_HIST_LEN; i++) {
> + cnt = &mib->counters[KSZ8_RX_64_OR_LESS + i];
> + dev->dev_ops->r_mib_pkt(dev, port,
> + (KSZ8_RX_64_OR_LESS + i), NULL, cnt);
Weird linewrap. Please align the following lines with the opening brace
of the first one, e.g.
dev->dev_ops->r_mib_pkt(dev, port,
KSZ8_RX_64_OR_LESS + i, NULL,
cnt);
BUT I don't see why you need those braces around `macro + i` and without
them you can fit it into the previous line I believe.
> + rmon_stats->hist[i] = *cnt;
> + }
> +
> + mutex_unlock(&mib->cnt_mutex);
> +
> + *ranges = ksz_rmon_ranges;
> +}
> +
> +void ksz9477_get_rmon_stats(struct ksz_device *dev, int port,
> + struct ethtool_rmon_stats *rmon_stats,
> + const struct ethtool_rmon_hist_range **ranges)
> +{
> + struct ksz_port_mib *mib;
> + u64 *cnt;
> + u8 i;
> +
> + mib = &dev->ports[port].mib;
> +
> + mutex_lock(&mib->cnt_mutex);
> +
> + cnt = &mib->counters[KSZ9477_RX_UNDERSIZE];
> + dev->dev_ops->r_mib_pkt(dev, port, KSZ9477_RX_UNDERSIZE, NULL, cnt);
> + rmon_stats->undersize_pkts = *cnt;
> +
> + cnt = &mib->counters[KSZ9477_RX_OVERSIZE];
> + dev->dev_ops->r_mib_pkt(dev, port, KSZ9477_RX_OVERSIZE, NULL, cnt);
> + rmon_stats->oversize_pkts = *cnt;
> +
> + cnt = &mib->counters[KSZ9477_RX_FRAGMENTS];
> + dev->dev_ops->r_mib_pkt(dev, port, KSZ9477_RX_FRAGMENTS, NULL, cnt);
> + rmon_stats->fragments = *cnt;
> +
> + cnt = &mib->counters[KSZ9477_RX_JABBERS];
> + dev->dev_ops->r_mib_pkt(dev, port, KSZ9477_RX_JABBERS, NULL, cnt);
> + rmon_stats->jabbers = *cnt;
> +
> + for (i = 0; i < KSZ9477_HIST_LEN; i++) {
> + cnt = &mib->counters[KSZ9477_RX_64_OR_LESS + i];
> + dev->dev_ops->r_mib_pkt(dev, port,
> + (KSZ9477_RX_64_OR_LESS + i), NULL, cnt);
(same, and please check all other places in the series)
> + rmon_stats->hist[i] = *cnt;
> + }
> +
> + mutex_unlock(&mib->cnt_mutex);
> +
> + *ranges = ksz_rmon_ranges;
> +}
[...]
Thanks,
Olek
On Fri, Feb 17, 2023 at 04:32:07PM +0530, Rakesh Sankaranarayanan wrote:
> Add support for ethtool standard device statistics grouping. Support rmon
> statistics grouping using rmon groups parameter in ethtool command. rmon
> provides packet size based range grouping. Common mib parameters are used
> across all KSZ series swtches for packet size statistics, except for
> KSZ8830. KSZ series have mib counters for packets with size:
> - less than 64 Bytes,
> - 65 to 127 Bytes,
> - 128 to 255 Bytes,
> - 256 to 511 Bytes,
> - 512 to 1023 Bytes,
> - 1024 to 1522 Bytes,
> - 1523 to 2000 Bytes and
> - More than 2001 Bytes
> KSZ8830 have mib counters upto 1024-1522 range only. Since no other change,
> common range used across all KSZ series, but used upto only upto 1024-1522
> for KSZ8830.
Why are all commit messages indented in this way? Please keep the
default text indentation at 0 characters. I have never seen this style
in "git log".
>
> Co-developed-by: Thangaraj Samynathan <Thangaraj.S@microchip.com>
Documentation/process/submitting-patches.rst:
Co-developed-by: states that the patch was co-created by multiple developers;
it is used to give attribution to co-authors (in addition to the author
attributed by the From: tag) when several people work on a single patch. Since
Co-developed-by: denotes authorship, every Co-developed-by: must be immediately
followed by a Signed-off-by: of the associated co-author. Standard sign-off
procedure applies, i.e. the ordering of Signed-off-by: tags should reflect the
chronological history of the patch insofar as possible, regardless of whether
the author is attributed via From: or Co-developed-by:. Notably, the last
Signed-off-by: must always be that of the developer submitting the patch.
On Fri, Feb 17, 2023 at 06:53:46PM +0200, Vladimir Oltean wrote:
> On Fri, Feb 17, 2023 at 04:32:07PM +0530, Rakesh Sankaranarayanan wrote:
> > Add support for ethtool standard device statistics grouping. Support rmon
> > statistics grouping using rmon groups parameter in ethtool command. rmon
> > provides packet size based range grouping. Common mib parameters are used
> > across all KSZ series swtches for packet size statistics, except for
> > KSZ8830. KSZ series have mib counters for packets with size:
> > - less than 64 Bytes,
> > - 65 to 127 Bytes,
> > - 128 to 255 Bytes,
> > - 256 to 511 Bytes,
> > - 512 to 1023 Bytes,
> > - 1024 to 1522 Bytes,
> > - 1523 to 2000 Bytes and
> > - More than 2001 Bytes
> > KSZ8830 have mib counters upto 1024-1522 range only. Since no other change,
> > common range used across all KSZ series, but used upto only upto 1024-1522
> > for KSZ8830.
>
> Why are all commit messages indented in this way? Please keep the
> default text indentation at 0 characters. I have never seen this style
> in "git log".
I can make a guess
git show HEAD
notice how the commit message is indented. So if you where to
cut/paste that, you get the extra indent. It is better to do:
git commit -C 8e757b50555f3ae ; git commit --am
to copy a commit message from 8e757b50555f3ae, and then edit it to
suite.
Andrew
On Fri, 2023-02-17 at 15:59 +0100, Alexander Lobakin wrote:
> [Some people who received this message don't often get email from
> aleksander.lobakin@intel.com. Learn why this is important at
> https://aka.ms/LearnAboutSenderIdentification ]
>
> EXTERNAL EMAIL: Do not click links or open attachments unless you
> know the content is safe
>
> From: Rakesh Sankaranarayanan <rakesh.sankaranarayanan@microchip.com>
> Date: Fri, 17 Feb 2023 16:32:07 +0530
>
> > Add support for ethtool standard device statistics grouping.
> > Support rmon
> > statistics grouping using rmon groups parameter in ethtool
> > command. rmon
> > provides packet size based range grouping. Common mib
> > parameters are used
> > across all KSZ series swtches for packet size statistics,
> > except for
> > KSZ8830. KSZ series have mib counters for packets with size:
>
> [...]
>
> > +void ksz8_get_rmon_stats(struct ksz_device *dev, int port,
> > + struct ethtool_rmon_stats *rmon_stats,
> > + const struct ethtool_rmon_hist_range
> > **ranges)
> > +{
> > + struct ksz_port_mib *mib;
> > + u64 *cnt;
>
> Nit: I guess it can be const since you only read it (in every such
> callback)?
This variable is getting updated on every call back after reading from
registers, so we cannot have it as constant.
>
> > + u8 i;
> > +
> > + mib = &dev->ports[port].mib;
> > +
> > + mutex_lock(&mib->cnt_mutex);
> > +
> > + cnt = &mib->counters[KSZ8_RX_UNDERSIZE];
> > + dev->dev_ops->r_mib_pkt(dev, port, KSZ8_RX_UNDERSIZE, NULL,
> > cnt);
> > + rmon_stats->undersize_pkts = *cnt;
> > +
> > + cnt = &mib->counters[KSZ8_RX_OVERSIZE];
> > + dev->dev_ops->r_mib_pkt(dev, port, KSZ8_RX_OVERSIZE, NULL,
> > cnt);
> > + rmon_stats->oversize_pkts = *cnt;
> > +
> > + cnt = &mib->counters[KSZ8_RX_FRAGMENTS];
> > + dev->dev_ops->r_mib_pkt(dev, port, KSZ8_RX_FRAGMENTS, NULL,
> > cnt);
> > + rmon_stats->fragments = *cnt;
> > +
> > + cnt = &mib->counters[KSZ8_RX_JABBERS];
> > + dev->dev_ops->r_mib_pkt(dev, port, KSZ8_RX_JABBERS, NULL,
> > cnt);
> > + rmon_stats->jabbers = *cnt;
> > +
> > + for (i = 0; i < KSZ8_HIST_LEN; i++) {
> > + cnt = &mib->counters[KSZ8_RX_64_OR_LESS + i];
> > + dev->dev_ops->r_mib_pkt(dev, port,
> > + (KSZ8_RX_64_OR_LESS + i), NULL, cnt);
>
> Weird linewrap. Please align the following lines with the opening
> brace
> of the first one, e.g.
>
> dev->dev_ops->r_mib_pkt(dev, port,
> KSZ8_RX_64_OR_LESS + i, NULL,
> cnt);
>
> BUT I don't see why you need those braces around `macro + i` and
> without
> them you can fit it into the previous line I believe.
Will update it in next revision
>
> > + rmon_stats->hist[i] = *cnt;
> > + }
> > +
> > + mutex_unlock(&mib->cnt_mutex);
> > +
> > + *ranges = ksz_rmon_ranges;
> > +}
> > +
> > +void ksz9477_get_rmon_stats(struct ksz_device *dev, int port,
> > + struct ethtool_rmon_stats *rmon_stats,
> > + const struct ethtool_rmon_hist_range
> > **ranges)
> > +{
> > + struct ksz_port_mib *mib;
> > + u64 *cnt;
> > + u8 i;
> > +
> > + mib = &dev->ports[port].mib;
> > +
> > + mutex_lock(&mib->cnt_mutex);
> > +
> > + cnt = &mib->counters[KSZ9477_RX_UNDERSIZE];
> > + dev->dev_ops->r_mib_pkt(dev, port, KSZ9477_RX_UNDERSIZE,
> > NULL, cnt);
> > + rmon_stats->undersize_pkts = *cnt;
> > +
> > + cnt = &mib->counters[KSZ9477_RX_OVERSIZE];
> > + dev->dev_ops->r_mib_pkt(dev, port, KSZ9477_RX_OVERSIZE, NULL,
> > cnt);
> > + rmon_stats->oversize_pkts = *cnt;
> > +
> > + cnt = &mib->counters[KSZ9477_RX_FRAGMENTS];
> > + dev->dev_ops->r_mib_pkt(dev, port, KSZ9477_RX_FRAGMENTS,
> > NULL, cnt);
> > + rmon_stats->fragments = *cnt;
> > +
> > + cnt = &mib->counters[KSZ9477_RX_JABBERS];
> > + dev->dev_ops->r_mib_pkt(dev, port, KSZ9477_RX_JABBERS, NULL,
> > cnt);
> > + rmon_stats->jabbers = *cnt;
> > +
> > + for (i = 0; i < KSZ9477_HIST_LEN; i++) {
> > + cnt = &mib->counters[KSZ9477_RX_64_OR_LESS + i];
> > + dev->dev_ops->r_mib_pkt(dev, port,
> > + (KSZ9477_RX_64_OR_LESS + i), NULL,
> > cnt);
>
> (same, and please check all other places in the series)
sure, will update in next version.
>
> > + rmon_stats->hist[i] = *cnt;
> > + }
> > +
> > + mutex_unlock(&mib->cnt_mutex);
> > +
> > + *ranges = ksz_rmon_ranges;
> > +}
> [...]
>
> Thanks,
> Olek
Hi Vlad,
On Fri, 2023-02-17 at 18:53 +0200, Vladimir Oltean wrote:
> EXTERNAL EMAIL: Do not click links or open attachments unless you
> know the content is safe
>
> On Fri, Feb 17, 2023 at 04:32:07PM +0530, Rakesh Sankaranarayanan
> wrote:
> > Add support for ethtool standard device statistics grouping.
> > Support rmon
> > statistics grouping using rmon groups parameter in ethtool
> > command. rmon
> > provides packet size based range grouping. Common mib
> > parameters are used
> > across all KSZ series swtches for packet size statistics,
> > except for
> > KSZ8830. KSZ series have mib counters for packets with size:
> > - less than 64 Bytes,
> > - 65 to 127 Bytes,
> > - 128 to 255 Bytes,
> > - 256 to 511 Bytes,
> > - 512 to 1023 Bytes,
> > - 1024 to 1522 Bytes,
> > - 1523 to 2000 Bytes and
> > - More than 2001 Bytes
> > KSZ8830 have mib counters upto 1024-1522 range only. Since no
> > other change,
> > common range used across all KSZ series, but used upto only
> > upto 1024-1522
> > for KSZ8830.
>
> Why are all commit messages indented in this way? Please keep the
> default text indentation at 0 characters. I have never seen this
> style
> in "git log".
Sure, will update the indentation in next version
>
> >
> > Co-developed-by: Thangaraj Samynathan <Thangaraj.S@microchip.com>
>
> Documentation/process/submitting-patches.rst:
>
> Co-developed-by: states that the patch was co-created by multiple
> developers;
> it is used to give attribution to co-authors (in addition to the
> author
> attributed by the From: tag) when several people work on a single
> patch. Since
> Co-developed-by: denotes authorship, every Co-developed-by: must be
> immediately
> followed by a Signed-off-by: of the associated co-author. Standard
> sign-off
> procedure applies, i.e. the ordering of Signed-off-by: tags should
> reflect the
> chronological history of the patch insofar as possible, regardless of
> whether
> the author is attributed via From: or Co-developed-by:. Notably, the
> last
> Signed-off-by: must always be that of the developer submitting the
> patch.
Will add Signed-off-by information in next revision
@@ -1,6 +1,7 @@
# SPDX-License-Identifier: GPL-2.0-only
obj-$(CONFIG_NET_DSA_MICROCHIP_KSZ_COMMON) += ksz_switch.o
ksz_switch-objs := ksz_common.o
+ksz_switch-objs += ksz_ethtool.o
ksz_switch-objs += ksz9477.o
ksz_switch-objs += ksz8795.o
ksz_switch-objs += lan937x_main.o
@@ -27,6 +27,7 @@
#include <net/switchdev.h>
#include "ksz_common.h"
+#include "ksz_ethtool.h"
#include "ksz_ptp.h"
#include "ksz8.h"
#include "ksz9477.h"
@@ -204,6 +205,7 @@ static const struct ksz_dev_ops ksz8_dev_ops = {
.freeze_mib = ksz8_freeze_mib,
.port_init_cnt = ksz8_port_init_cnt,
.fdb_dump = ksz8_fdb_dump,
+ .get_rmon_stats = ksz8_get_rmon_stats,
.mdb_add = ksz8_mdb_add,
.mdb_del = ksz8_mdb_del,
.vlan_filtering = ksz8_port_vlan_filtering,
@@ -241,6 +243,7 @@ static const struct ksz_dev_ops ksz9477_dev_ops = {
.r_mib_stat64 = ksz_r_mib_stats64,
.freeze_mib = ksz9477_freeze_mib,
.port_init_cnt = ksz9477_port_init_cnt,
+ .get_rmon_stats = ksz9477_get_rmon_stats,
.vlan_filtering = ksz9477_port_vlan_filtering,
.vlan_add = ksz9477_port_vlan_add,
.vlan_del = ksz9477_port_vlan_del,
@@ -277,6 +280,7 @@ static const struct ksz_dev_ops lan937x_dev_ops = {
.r_mib_stat64 = ksz_r_mib_stats64,
.freeze_mib = ksz9477_freeze_mib,
.port_init_cnt = ksz9477_port_init_cnt,
+ .get_rmon_stats = ksz9477_get_rmon_stats,
.vlan_filtering = ksz9477_port_vlan_filtering,
.vlan_add = ksz9477_port_vlan_add,
.vlan_del = ksz9477_port_vlan_del,
@@ -1730,6 +1734,16 @@ static void ksz_get_pause_stats(struct dsa_switch *ds, int port,
spin_unlock(&mib->stats64_lock);
}
+static void ksz_get_rmon_stats(struct dsa_switch *ds, int port,
+ struct ethtool_rmon_stats *rmon_stats,
+ const struct ethtool_rmon_hist_range **ranges)
+{
+ struct ksz_device *dev = ds->priv;
+
+ if (dev->dev_ops->get_rmon_stats)
+ dev->dev_ops->get_rmon_stats(dev, port, rmon_stats, ranges);
+}
+
static void ksz_get_strings(struct dsa_switch *ds, int port,
u32 stringset, uint8_t *buf)
{
@@ -3186,6 +3200,7 @@ static const struct dsa_switch_ops ksz_switch_ops = {
.port_mirror_del = ksz_port_mirror_del,
.get_stats64 = ksz_get_stats64,
.get_pause_stats = ksz_get_pause_stats,
+ .get_rmon_stats = ksz_get_rmon_stats,
.port_change_mtu = ksz_change_mtu,
.port_max_mtu = ksz_max_mtu,
.get_ts_info = ksz_get_ts_info,
@@ -361,6 +361,9 @@ struct ksz_dev_ops {
int (*reset)(struct ksz_device *dev);
int (*init)(struct ksz_device *dev);
void (*exit)(struct ksz_device *dev);
+ void (*get_rmon_stats)(struct ksz_device *dev, int port,
+ struct ethtool_rmon_stats *rmon_stats,
+ const struct ethtool_rmon_hist_range **ranges);
};
struct ksz_device *ksz_switch_alloc(struct device *base, void *priv);
new file mode 100644
@@ -0,0 +1,180 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Microchip KSZ series ethtool statistics
+ *
+ * Copyright (C) 2022 Microchip Technology Inc.
+ */
+
+#include "ksz_common.h"
+#include "ksz_ethtool.h"
+
+enum ksz8_mib_entry {
+ KSZ8_RX,
+ KSZ8_RX_HI,
+ KSZ8_RX_UNDERSIZE,
+ KSZ8_RX_FRAGMENTS,
+ KSZ8_RX_OVERSIZE,
+ KSZ8_RX_JABBERS,
+ KSZ8_RX_SYMBOL_ERR,
+ KSZ8_RX_CRC_ERR,
+ KSZ8_RX_ALIGN_ERR,
+ KSZ8_RX_MAC_CTL,
+ KSZ8_RX_PAUSE,
+ KSZ8_RX_BCAST,
+ KSZ8_RX_MCAST,
+ KSZ8_RX_UCAST,
+ KSZ8_RX_64_OR_LESS,
+ KSZ8_RX_65_127,
+ KSZ8_RX_128_255,
+ KSZ8_RX_256_511,
+ KSZ8_RX_512_1023,
+ KSZ8_RX_1024_1522,
+ KSZ8_TX,
+ KSZ8_TX_HI,
+ KSZ8_TX_LATE_COL,
+ KSZ8_TX_PAUSE,
+ KSZ8_TX_BCAST,
+ KSZ8_TX_MCAST,
+ KSZ8_TX_UCAST,
+ KSZ8_TX_DEFERRED,
+ KSZ8_TX_TOTAL_COL,
+ KSZ8_TX_EXC_COL,
+ KSZ8_TX_SINGLE_COL,
+ KSZ8_TX_MULT_COL,
+ KSZ8_RX_DISCARDS,
+ KSZ8_TX_DISCARDS,
+};
+
+enum ksz9477_mib_entry {
+ KSZ9477_RX_HI,
+ KSZ9477_RX_UNDERSIZE,
+ KSZ9477_RX_FRAGMENTS,
+ KSZ9477_RX_OVERSIZE,
+ KSZ9477_RX_JABBERS,
+ KSZ9477_RX_SYMBOL_ERR,
+ KSZ9477_RX_CRC_ERR,
+ KSZ9477_RX_ALIGN_ERR,
+ KSZ9477_RX_MAC_CTL,
+ KSZ9477_RX_PAUSE,
+ KSZ9477_RX_BCAST,
+ KSZ9477_RX_MCAST,
+ KSZ9477_RX_UCAST,
+ KSZ9477_RX_64_OR_LESS,
+ KSZ9477_RX_65_127,
+ KSZ9477_RX_128_255,
+ KSZ9477_RX_256_511,
+ KSZ9477_RX_512_1023,
+ KSZ9477_RX_1024_1522,
+ KSZ9477_RX_1523_2000,
+ KSZ9477_RX_2001,
+ KSZ9477_TX_HI,
+ KSZ9477_TX_LATE_COL,
+ KSZ9477_TX_PAUSE,
+ KSZ9477_TX_BCAST,
+ KSZ9477_TX_MCAST,
+ KSZ9477_TX_UCAST,
+ KSZ9477_TX_DEFERRED,
+ KSZ9477_TX_TOTAL_COL,
+ KSZ9477_TX_EXC_COL,
+ KSZ9477_TX_SINGLE_COL,
+ KSZ9477_TX_MULT_COL,
+ KSZ9477_RX_TOTAL,
+ KSZ9477_TX_TOTAL,
+ KSZ9477_RX_DISCARDS,
+ KSZ9477_TX_DISCARDS,
+};
+
+static const struct ethtool_rmon_hist_range ksz_rmon_ranges[] = {
+ { 0, 64 },
+ { 65, 127 },
+ { 128, 255 },
+ { 256, 511 },
+ { 512, 1023 },
+ { 1024, 1522 },
+ { 1523, 2000 },
+ { 2001, 9000 },
+ {}
+};
+
+#define KSZ8_HIST_LEN 6
+#define KSZ9477_HIST_LEN 8
+
+void ksz8_get_rmon_stats(struct ksz_device *dev, int port,
+ struct ethtool_rmon_stats *rmon_stats,
+ const struct ethtool_rmon_hist_range **ranges)
+{
+ struct ksz_port_mib *mib;
+ u64 *cnt;
+ u8 i;
+
+ mib = &dev->ports[port].mib;
+
+ mutex_lock(&mib->cnt_mutex);
+
+ cnt = &mib->counters[KSZ8_RX_UNDERSIZE];
+ dev->dev_ops->r_mib_pkt(dev, port, KSZ8_RX_UNDERSIZE, NULL, cnt);
+ rmon_stats->undersize_pkts = *cnt;
+
+ cnt = &mib->counters[KSZ8_RX_OVERSIZE];
+ dev->dev_ops->r_mib_pkt(dev, port, KSZ8_RX_OVERSIZE, NULL, cnt);
+ rmon_stats->oversize_pkts = *cnt;
+
+ cnt = &mib->counters[KSZ8_RX_FRAGMENTS];
+ dev->dev_ops->r_mib_pkt(dev, port, KSZ8_RX_FRAGMENTS, NULL, cnt);
+ rmon_stats->fragments = *cnt;
+
+ cnt = &mib->counters[KSZ8_RX_JABBERS];
+ dev->dev_ops->r_mib_pkt(dev, port, KSZ8_RX_JABBERS, NULL, cnt);
+ rmon_stats->jabbers = *cnt;
+
+ for (i = 0; i < KSZ8_HIST_LEN; i++) {
+ cnt = &mib->counters[KSZ8_RX_64_OR_LESS + i];
+ dev->dev_ops->r_mib_pkt(dev, port,
+ (KSZ8_RX_64_OR_LESS + i), NULL, cnt);
+ rmon_stats->hist[i] = *cnt;
+ }
+
+ mutex_unlock(&mib->cnt_mutex);
+
+ *ranges = ksz_rmon_ranges;
+}
+
+void ksz9477_get_rmon_stats(struct ksz_device *dev, int port,
+ struct ethtool_rmon_stats *rmon_stats,
+ const struct ethtool_rmon_hist_range **ranges)
+{
+ struct ksz_port_mib *mib;
+ u64 *cnt;
+ u8 i;
+
+ mib = &dev->ports[port].mib;
+
+ mutex_lock(&mib->cnt_mutex);
+
+ cnt = &mib->counters[KSZ9477_RX_UNDERSIZE];
+ dev->dev_ops->r_mib_pkt(dev, port, KSZ9477_RX_UNDERSIZE, NULL, cnt);
+ rmon_stats->undersize_pkts = *cnt;
+
+ cnt = &mib->counters[KSZ9477_RX_OVERSIZE];
+ dev->dev_ops->r_mib_pkt(dev, port, KSZ9477_RX_OVERSIZE, NULL, cnt);
+ rmon_stats->oversize_pkts = *cnt;
+
+ cnt = &mib->counters[KSZ9477_RX_FRAGMENTS];
+ dev->dev_ops->r_mib_pkt(dev, port, KSZ9477_RX_FRAGMENTS, NULL, cnt);
+ rmon_stats->fragments = *cnt;
+
+ cnt = &mib->counters[KSZ9477_RX_JABBERS];
+ dev->dev_ops->r_mib_pkt(dev, port, KSZ9477_RX_JABBERS, NULL, cnt);
+ rmon_stats->jabbers = *cnt;
+
+ for (i = 0; i < KSZ9477_HIST_LEN; i++) {
+ cnt = &mib->counters[KSZ9477_RX_64_OR_LESS + i];
+ dev->dev_ops->r_mib_pkt(dev, port,
+ (KSZ9477_RX_64_OR_LESS + i), NULL, cnt);
+ rmon_stats->hist[i] = *cnt;
+ }
+
+ mutex_unlock(&mib->cnt_mutex);
+
+ *ranges = ksz_rmon_ranges;
+}
new file mode 100644
@@ -0,0 +1,18 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Microchip KSZ series ethtool statistics
+ *
+ * Copyright (C) 2022 Microchip Technology Inc.
+ */
+
+#ifndef __KSZ_ETHTOOL_H
+#define __KSZ_ETHTOOL_H
+
+void ksz8_get_rmon_stats(struct ksz_device *dev, int port,
+ struct ethtool_rmon_stats *rmon_stats,
+ const struct ethtool_rmon_hist_range **ranges);
+
+void ksz9477_get_rmon_stats(struct ksz_device *dev, int port,
+ struct ethtool_rmon_stats *rmon_stats,
+ const struct ethtool_rmon_hist_range **ranges);
+#endif