* Re: [PATCH 2/6] move _body_io_syscall to the generic syscall.h
From: Jeff Moyer @ 2018-01-05 22:49 UTC (permalink / raw)
To: Benjamin LaHaise
Cc: Christoph Hellwig, linux-aio, Avi Kivity, linux-fsdevel, netdev,
linux-kernel
In-Reply-To: <20180105164031.GK27850@kvack.org>
Hi, Ben,
Thanks for the quick reply.
Benjamin LaHaise <ben@communityfibre.ca> writes:
> On Fri, Jan 05, 2018 at 11:25:17AM -0500, Jeff Moyer wrote:
>> Christoph Hellwig <hch@lst.de> writes:
>>
>> > This way it can be used for the fallback 6-argument version on
>> > all architectures.
>> >
>> > Signed-off-by: Christoph Hellwig <hch@lst.de>
>>
>> This is a strange way to do things. However, I was never really sold on
>> libaio having to implement its own system call wrappers. That decision
>> definitely resulted in some maintenance overhead.
>>
>> Ben, what was your reasoning for not just using syscall?
>
> The main issue was that glibc's pthreads implementation really sucked back
> during initial development and there was a use-case for having the io_XXX
> functions usable directly from clone()ed threads that didn't have all the
> glibc pthread state setup for per-cpu areas to handle per-thread errno.
> That made sense back then, but is rather silly today.
Thanks for the background info.
> Technically, I'm not sure the generic syscall wrapper is safe to use. The
> io_XXX arch wrappers don't modify errno, while it appears the generic one
> does. That said, nobody has ever noticed...
Good point. Common architectures don't use the generic syscall wrapper,
so I'm not sure we can conclude that it won't break anything. At the
same time, I'm not sure I want to write and test the io_syscall6
assembly for all of the supported arches. I could save and restore
errno. That sounds ugly, but less painful than the other options.
Does anyone have any strong preferences?
-Jeff
--
To unsubscribe, send a message with 'unsubscribe linux-aio' in
the body to majordomo@kvack.org. For more info on Linux AIO,
see: http://www.kvack.org/aio/
Don't email: <a href=mailto:"aart@kvack.org">aart@kvack.org</a>
^ permalink raw reply
* Re: [PATCH 6/6] add test for aio poll and io_pgetevents
From: Jeff Moyer @ 2018-01-05 22:55 UTC (permalink / raw)
To: Christoph Hellwig
Cc: linux-aio, Avi Kivity, linux-fsdevel, netdev, linux-kernel
In-Reply-To: <20180104080325.14716-7-hch@lst.de>
Christoph Hellwig <hch@lst.de> writes:
> + p = fork();
> + switch (p) {
[snip]
> + default:
> + close(pipe1[0]);
> + close(pipe2[1]);
> +
> + io_prep_poll(&iocb, pipe2[0], POLLIN);
> +
> + ret = io_setup(1, &ctx);
> + if (ret) {
> + printf("child: io_setup failed\n");
parent
> + return 1;
> + }
> +
> + ret = io_submit(ctx, 1, iocbs);
> + if (ret != 1) {
> + printf("child: io_submit failed\n");
parent
Other than that, looks ok to me. Thanks for writing a test!
I can fix this up, no need to repost.
-Jeff
--
To unsubscribe, send a message with 'unsubscribe linux-aio' in
the body to majordomo@kvack.org. For more info on Linux AIO,
see: http://www.kvack.org/aio/
Don't email: <a href=mailto:"aart@kvack.org">aart@kvack.org</a>
^ permalink raw reply
* Re: libaio: resurrect aio poll and add io_pgetevents support
From: Jeff Moyer @ 2018-01-05 22:57 UTC (permalink / raw)
To: Christoph Hellwig
Cc: linux-aio, Avi Kivity, linux-fsdevel, netdev, linux-kernel
In-Reply-To: <20180104080325.14716-1-hch@lst.de>
Christoph Hellwig <hch@lst.de> writes:
> Hi all,
>
> this series resurrects IOCB_CMD_POLL support and adds support for the
> new io_pgetevents system call, as well as adding a test case.
This looks good to me. There may be a couple of changes to the syscall
bits, but I can take care of that. I'll review the kernel bits more
thoroughly next week.
-Jeff
--
To unsubscribe, send a message with 'unsubscribe linux-aio' in
the body to majordomo@kvack.org. For more info on Linux AIO,
see: http://www.kvack.org/aio/
Don't email: <a href=mailto:"aart@kvack.org">aart@kvack.org</a>
^ permalink raw reply
* [PATCH net-next v2 00/10] net: create dynamic software irq moderation library
From: Andy Gospodarek @ 2018-01-05 22:58 UTC (permalink / raw)
To: netdev; +Cc: mchan, talgi, ogerlitz, Andy Gospodarek
From: Andy Gospodarek <gospo@broadcom.com>
This converts the dynamic interrupt moderation library from the mlx5e
driver into a library so it can be used by any driver. The penultimate
patch in this set adds support for thiw new dynamic interrupt moderation
library in the bnxt_en driver and the last patch creates an entry in the
MAINTAINERS file for this library.
The main purpose of this code is to allow an administrator to make sure
that default coalesce settings are optimized for low latency, but
quickly adapt to handle high throughput/bulk traffic by altering how
much time passes before popping an interrupt.
For any new driver the following changes would be needed to use this
library:
- add elements in ring struct to track items needed by this library
- create function that can be called to actually set coalesce settings
for the driver
Credit to Rob Rice and Lee Reed for doing some of the initial proof of
concept and testing for this patch and Tal Gilboa and Or Gerlitz for
their comments, etc on this set.
v2: Spelling fixes from Stephen Hemminger, bnxt_en suggestions from
Michael Chan, spelling and formatting fixes from Or Gerlitz, and
spelling and mlx5e changes suggested by Tal Gilboa.
Andy Gospodarek (10):
net/mlx5e: Move interrupt moderation structs to new file
net/mlx5e: Move interrupt moderation forward declarations
net/mlx5e: Remove rq references in mlx5e_rx_am
net/mlx5e: Move AM logic enums
net/mlx5e: Move generic functions to new file
net/mlx5e: Change Mellanox references in DIM code
net/mlx5e: Move dynamic interrupt coalescing code to include/linux
net/dim: use struct net_dim_sample as arg to net_dim
bnxt_en: add support for software dynamic interrupt moderation
MAINTAINERS: add entry for Dynamic Interrupt Moderation
MAINTAINERS | 5 +
drivers/net/ethernet/broadcom/bnxt/Makefile | 2 +-
drivers/net/ethernet/broadcom/bnxt/bnxt.c | 49 +++
drivers/net/ethernet/broadcom/bnxt/bnxt.h | 34 +-
drivers/net/ethernet/broadcom/bnxt/bnxt_dim.c | 32 ++
drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c | 12 +
drivers/net/ethernet/mellanox/mlx5/core/Makefile | 2 +-
drivers/net/ethernet/mellanox/mlx5/core/en.h | 49 +--
drivers/net/ethernet/mellanox/mlx5/core/en_dim.c | 49 +++
.../net/ethernet/mellanox/mlx5/core/en_ethtool.c | 12 +-
drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 52 ++-
drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 4 +-
drivers/net/ethernet/mellanox/mlx5/core/en_rx_am.c | 341 -------------------
drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c | 10 +-
include/linux/mlx5/mlx5_ifc.h | 6 -
include/linux/net_dim.h | 372 +++++++++++++++++++++
16 files changed, 604 insertions(+), 427 deletions(-)
create mode 100644 drivers/net/ethernet/broadcom/bnxt/bnxt_dim.c
create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/en_dim.c
delete mode 100644 drivers/net/ethernet/mellanox/mlx5/core/en_rx_am.c
create mode 100644 include/linux/net_dim.h
--
2.7.4
^ permalink raw reply
* [PATCH net-next v2 01/10] net/mlx5e: Move interrupt moderation structs to new file
From: Andy Gospodarek @ 2018-01-05 22:58 UTC (permalink / raw)
To: netdev; +Cc: mchan, talgi, ogerlitz, Andy Gospodarek
In-Reply-To: <1515193094-9554-1-git-send-email-andy@greyhouse.net>
From: Andy Gospodarek <gospo@broadcom.com>
Create new header file to prepare to move code that handles irq
moderation to a library that lives in a header file.
Signed-off-by: Andy Gospodarek <gospo@broadcom.com>
Acked-by: Tal Gilboa <talgi@mellanox.com>
Acked-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en.h | 33 +----------
drivers/net/ethernet/mellanox/mlx5/core/en_dim.h | 75 ++++++++++++++++++++++++
include/linux/mlx5/mlx5_ifc.h | 6 --
3 files changed, 76 insertions(+), 38 deletions(-)
create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/en_dim.h
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h
index 543060c..ddb5429 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h
@@ -49,6 +49,7 @@
#include "wq.h"
#include "mlx5_core.h"
#include "en_stats.h"
+#include "en_dim.h"
#define MLX5_SET_CFG(p, f, v) MLX5_SET(create_flow_group_in, p, f, v)
@@ -226,12 +227,6 @@ enum mlx5e_priv_flag {
#define MLX5E_MAX_BW_ALLOC 100 /* Max percentage of BW allocation */
#endif
-struct mlx5e_cq_moder {
- u16 usec;
- u16 pkts;
- u8 cq_period_mode;
-};
-
struct mlx5e_params {
u8 log_sq_size;
u8 rq_wq_type;
@@ -472,32 +467,6 @@ struct mlx5e_mpw_info {
u16 skbs_frags[MLX5_MPWRQ_PAGES_PER_WQE];
};
-struct mlx5e_rx_am_stats {
- int ppms; /* packets per msec */
- int bpms; /* bytes per msec */
- int epms; /* events per msec */
-};
-
-struct mlx5e_rx_am_sample {
- ktime_t time;
- u32 pkt_ctr;
- u32 byte_ctr;
- u16 event_ctr;
-};
-
-struct mlx5e_rx_am { /* Adaptive Moderation */
- u8 state;
- struct mlx5e_rx_am_stats prev_stats;
- struct mlx5e_rx_am_sample start_sample;
- struct work_struct work;
- u8 profile_ix;
- u8 mode;
- u8 tune_state;
- u8 steps_right;
- u8 steps_left;
- u8 tired;
-};
-
/* a single cache unit is capable to serve one napi call (for non-striding rq)
* or a MPWQE (for striding rq).
*/
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_dim.h b/drivers/net/ethernet/mellanox/mlx5/core/en_dim.h
new file mode 100644
index 0000000..84b8524
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_dim.h
@@ -0,0 +1,75 @@
+/*
+ * Copyright (c) 2013-2015, Mellanox Technologies, Ltd. All rights reserved.
+ * Copyright (c) 2017, Broadcom Limited
+ *
+ * This software is available to you under a choice of one of two
+ * licenses. You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ * Redistribution and use in source and binary forms, with or
+ * without modification, are permitted provided that the following
+ * conditions are met:
+ *
+ * - Redistributions of source code must retain the above
+ * copyright notice, this list of conditions and the following
+ * disclaimer.
+ *
+ * - Redistributions in binary form must reproduce the above
+ * copyright notice, this list of conditions and the following
+ * disclaimer in the documentation and/or other materials
+ * provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+*/
+
+#ifndef MLX5_AM_H
+#define MLX5_AM_H
+
+struct mlx5e_cq_moder {
+ u16 usec;
+ u16 pkts;
+ u8 cq_period_mode;
+};
+
+struct mlx5e_rx_am_sample {
+ ktime_t time;
+ u32 pkt_ctr;
+ u32 byte_ctr;
+ u16 event_ctr;
+};
+
+struct mlx5e_rx_am_stats {
+ int ppms; /* packets per msec */
+ int bpms; /* bytes per msec */
+ int epms; /* events per msec */
+};
+
+struct mlx5e_rx_am { /* Adaptive Moderation */
+ u8 state;
+ struct mlx5e_rx_am_stats prev_stats;
+ struct mlx5e_rx_am_sample start_sample;
+ struct work_struct work;
+ u8 profile_ix;
+ u8 mode;
+ u8 tune_state;
+ u8 steps_right;
+ u8 steps_left;
+ u8 tired;
+};
+
+enum {
+ MLX5_CQ_PERIOD_MODE_START_FROM_EQE = 0x0,
+ MLX5_CQ_PERIOD_MODE_START_FROM_CQE = 0x1,
+ MLX5_CQ_PERIOD_NUM_MODES
+};
+
+#endif /* MLX5_AM_H */
diff --git a/include/linux/mlx5/mlx5_ifc.h b/include/linux/mlx5/mlx5_ifc.h
index d44ec5f..60d17e5 100644
--- a/include/linux/mlx5/mlx5_ifc.h
+++ b/include/linux/mlx5/mlx5_ifc.h
@@ -2918,12 +2918,6 @@ enum {
MLX5_CQC_ST_FIRED = 0xa,
};
-enum {
- MLX5_CQ_PERIOD_MODE_START_FROM_EQE = 0x0,
- MLX5_CQ_PERIOD_MODE_START_FROM_CQE = 0x1,
- MLX5_CQ_PERIOD_NUM_MODES
-};
-
struct mlx5_ifc_cqc_bits {
u8 status[0x4];
u8 reserved_at_4[0x4];
--
2.7.4
^ permalink raw reply related
* [PATCH net-next v2 02/10] net/mlx5e: Move interrupt moderation forward declarations
From: Andy Gospodarek @ 2018-01-05 22:58 UTC (permalink / raw)
To: netdev; +Cc: mchan, talgi, ogerlitz, Andy Gospodarek
In-Reply-To: <1515193094-9554-1-git-send-email-andy@greyhouse.net>
From: Andy Gospodarek <gospo@broadcom.com>
Move these to newly created file to prepare to move these functions to a
library.
Signed-off-by: Andy Gospodarek <gospo@broadcom.com>
Acked-by: Tal Gilboa <talgi@mellanox.com>
Acked-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en.h | 4 ----
drivers/net/ethernet/mellanox/mlx5/core/en_dim.h | 5 +++++
2 files changed, 5 insertions(+), 4 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h
index ddb5429..2ccedf6 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h
@@ -829,10 +829,6 @@ void mlx5e_dealloc_rx_wqe(struct mlx5e_rq *rq, u16 ix);
void mlx5e_dealloc_rx_mpwqe(struct mlx5e_rq *rq, u16 ix);
void mlx5e_free_rx_mpwqe(struct mlx5e_rq *rq, struct mlx5e_mpw_info *wi);
-void mlx5e_rx_am(struct mlx5e_rq *rq);
-void mlx5e_rx_am_work(struct work_struct *work);
-struct mlx5e_cq_moder mlx5e_am_get_def_profile(u8 rx_cq_period_mode);
-
void mlx5e_update_stats(struct mlx5e_priv *priv, bool full);
int mlx5e_create_flow_steering(struct mlx5e_priv *priv);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_dim.h b/drivers/net/ethernet/mellanox/mlx5/core/en_dim.h
index 84b8524..f5f6535 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_dim.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_dim.h
@@ -72,4 +72,9 @@ enum {
MLX5_CQ_PERIOD_NUM_MODES
};
+struct mlx5e_rq;
+void mlx5e_rx_am(struct mlx5e_rq *rq);
+void mlx5e_rx_am_work(struct work_struct *work);
+struct mlx5e_cq_moder mlx5e_am_get_def_profile(u8 rx_cq_period_mode);
+
#endif /* MLX5_AM_H */
--
2.7.4
^ permalink raw reply related
* [PATCH net-next v2 03/10] net/mlx5e: Remove rq references in mlx5e_rx_am
From: Andy Gospodarek @ 2018-01-05 22:58 UTC (permalink / raw)
To: netdev; +Cc: mchan, talgi, ogerlitz, Andy Gospodarek
In-Reply-To: <1515193094-9554-1-git-send-email-andy@greyhouse.net>
From: Andy Gospodarek <gospo@broadcom.com>
This makes mlx5e_am_sample more generic so that it can be called easily
from a driver that does not use the same data structure to store these
values in a single structure.
Signed-off-by: Andy Gospodarek <gospo@broadcom.com>
Acked-by: Tal Gilboa <talgi@mellanox.com>
Acked-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en_dim.h | 6 ++++--
drivers/net/ethernet/mellanox/mlx5/core/en_rx_am.c | 22 +++++++++++++---------
drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c | 5 ++++-
3 files changed, 21 insertions(+), 12 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_dim.h b/drivers/net/ethernet/mellanox/mlx5/core/en_dim.h
index f5f6535..b676a057 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_dim.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_dim.h
@@ -72,8 +72,10 @@ enum {
MLX5_CQ_PERIOD_NUM_MODES
};
-struct mlx5e_rq;
-void mlx5e_rx_am(struct mlx5e_rq *rq);
+void mlx5e_rx_am(struct mlx5e_rx_am *am,
+ u16 event_ctr,
+ u64 packets,
+ u64 bytes);
void mlx5e_rx_am_work(struct work_struct *work);
struct mlx5e_cq_moder mlx5e_am_get_def_profile(u8 rx_cq_period_mode);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rx_am.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rx_am.c
index e401d9d..1630076 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rx_am.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rx_am.c
@@ -264,13 +264,15 @@ static bool mlx5e_am_decision(struct mlx5e_rx_am_stats *curr_stats,
return am->profile_ix != prev_ix;
}
-static void mlx5e_am_sample(struct mlx5e_rq *rq,
+static void mlx5e_am_sample(u16 event_ctr,
+ u64 packets,
+ u64 bytes,
struct mlx5e_rx_am_sample *s)
{
s->time = ktime_get();
- s->pkt_ctr = rq->stats.packets;
- s->byte_ctr = rq->stats.bytes;
- s->event_ctr = rq->cq.event_ctr;
+ s->pkt_ctr = packets;
+ s->byte_ctr = bytes;
+ s->event_ctr = event_ctr;
}
#define MLX5E_AM_NEVENTS 64
@@ -309,20 +311,22 @@ void mlx5e_rx_am_work(struct work_struct *work)
am->state = MLX5E_AM_START_MEASURE;
}
-void mlx5e_rx_am(struct mlx5e_rq *rq)
+void mlx5e_rx_am(struct mlx5e_rx_am *am,
+ u16 event_ctr,
+ u64 packets,
+ u64 bytes)
{
- struct mlx5e_rx_am *am = &rq->am;
struct mlx5e_rx_am_sample end_sample;
struct mlx5e_rx_am_stats curr_stats;
u16 nevents;
switch (am->state) {
case MLX5E_AM_MEASURE_IN_PROGRESS:
- nevents = BIT_GAP(BITS_PER_TYPE(u16), rq->cq.event_ctr,
+ nevents = BIT_GAP(BITS_PER_TYPE(u16), event_ctr,
am->start_sample.event_ctr);
if (nevents < MLX5E_AM_NEVENTS)
break;
- mlx5e_am_sample(rq, &end_sample);
+ mlx5e_am_sample(event_ctr, packets, bytes, &end_sample);
mlx5e_am_calc_stats(&am->start_sample, &end_sample,
&curr_stats);
if (mlx5e_am_decision(&curr_stats, am)) {
@@ -332,7 +336,7 @@ void mlx5e_rx_am(struct mlx5e_rq *rq)
}
/* fall through */
case MLX5E_AM_START_MEASURE:
- mlx5e_am_sample(rq, &am->start_sample);
+ mlx5e_am_sample(event_ctr, packets, bytes, &am->start_sample);
am->state = MLX5E_AM_MEASURE_IN_PROGRESS;
break;
case MLX5E_AM_APPLY_NEW_PROFILE:
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c
index ab92298..1849169 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c
@@ -79,7 +79,10 @@ int mlx5e_napi_poll(struct napi_struct *napi, int budget)
mlx5e_cq_arm(&c->sq[i].cq);
if (MLX5E_TEST_BIT(c->rq.state, MLX5E_RQ_STATE_AM))
- mlx5e_rx_am(&c->rq);
+ mlx5e_rx_am(&c->rq.am,
+ c->rq.cq.event_ctr,
+ c->rq.stats.packets,
+ c->rq.stats.bytes);
mlx5e_cq_arm(&c->rq.cq);
mlx5e_cq_arm(&c->icosq.cq);
--
2.7.4
^ permalink raw reply related
* [PATCH net-next v2 04/10] net/mlx5e: Move AM logic enums
From: Andy Gospodarek @ 2018-01-05 22:58 UTC (permalink / raw)
To: netdev; +Cc: mchan, talgi, ogerlitz, Andy Gospodarek
In-Reply-To: <1515193094-9554-1-git-send-email-andy@greyhouse.net>
From: Andy Gospodarek <gospo@broadcom.com>
More movement to help make this code more generic.
Signed-off-by: Andy Gospodarek <gospo@broadcom.com>
Acked-by: Tal Gilboa <talgi@mellanox.com>
Acked-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en_dim.h | 26 ++++++++++++++++++++++
drivers/net/ethernet/mellanox/mlx5/core/en_rx_am.c | 25 ---------------------
2 files changed, 26 insertions(+), 25 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_dim.h b/drivers/net/ethernet/mellanox/mlx5/core/en_dim.h
index b676a057..c9f0d05 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_dim.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_dim.h
@@ -72,6 +72,32 @@ enum {
MLX5_CQ_PERIOD_NUM_MODES
};
+/* Adaptive moderation logic */
+enum {
+ MLX5E_AM_START_MEASURE,
+ MLX5E_AM_MEASURE_IN_PROGRESS,
+ MLX5E_AM_APPLY_NEW_PROFILE,
+};
+
+enum {
+ MLX5E_AM_PARKING_ON_TOP,
+ MLX5E_AM_PARKING_TIRED,
+ MLX5E_AM_GOING_RIGHT,
+ MLX5E_AM_GOING_LEFT,
+};
+
+enum {
+ MLX5E_AM_STATS_WORSE,
+ MLX5E_AM_STATS_SAME,
+ MLX5E_AM_STATS_BETTER,
+};
+
+enum {
+ MLX5E_AM_STEPPED,
+ MLX5E_AM_TOO_TIRED,
+ MLX5E_AM_ON_EDGE,
+};
+
void mlx5e_rx_am(struct mlx5e_rx_am *am,
u16 event_ctr,
u64 packets,
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rx_am.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rx_am.c
index 1630076..337dd60 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rx_am.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rx_am.c
@@ -82,31 +82,6 @@ struct mlx5e_cq_moder mlx5e_am_get_def_profile(u8 rx_cq_period_mode)
return mlx5e_am_get_profile(rx_cq_period_mode, default_profile_ix);
}
-/* Adaptive moderation logic */
-enum {
- MLX5E_AM_START_MEASURE,
- MLX5E_AM_MEASURE_IN_PROGRESS,
- MLX5E_AM_APPLY_NEW_PROFILE,
-};
-
-enum {
- MLX5E_AM_PARKING_ON_TOP,
- MLX5E_AM_PARKING_TIRED,
- MLX5E_AM_GOING_RIGHT,
- MLX5E_AM_GOING_LEFT,
-};
-
-enum {
- MLX5E_AM_STATS_WORSE,
- MLX5E_AM_STATS_SAME,
- MLX5E_AM_STATS_BETTER,
-};
-
-enum {
- MLX5E_AM_STEPPED,
- MLX5E_AM_TOO_TIRED,
- MLX5E_AM_ON_EDGE,
-};
static bool mlx5e_am_on_top(struct mlx5e_rx_am *am)
{
--
2.7.4
^ permalink raw reply related
* [PATCH net-next v2 05/10] net/mlx5e: Move generic functions to new file
From: Andy Gospodarek @ 2018-01-05 22:58 UTC (permalink / raw)
To: netdev; +Cc: mchan, talgi, ogerlitz, Andy Gospodarek
In-Reply-To: <1515193094-9554-1-git-send-email-andy@greyhouse.net>
From: Andy Gospodarek <gospo@broadcom.com>
These functions were identified as ones that could be made generic and
used by multiple drivers. Most of the contents of en_rx_am.c are moved
to net_dim.c.
Signed-off-by: Andy Gospodarek <gospo@broadcom.com>
Acked-by: Tal Gilboa <talgi@mellanox.com>
Acked-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/Makefile | 4 +-
drivers/net/ethernet/mellanox/mlx5/core/en.h | 2 +-
drivers/net/ethernet/mellanox/mlx5/core/en_dim.c | 48 ++++
drivers/net/ethernet/mellanox/mlx5/core/en_dim.h | 108 -------
drivers/net/ethernet/mellanox/mlx5/core/en_rx_am.c | 320 ---------------------
drivers/net/ethernet/mellanox/mlx5/core/net_dim.c | 307 ++++++++++++++++++++
drivers/net/ethernet/mellanox/mlx5/core/net_dim.h | 109 +++++++
7 files changed, 467 insertions(+), 431 deletions(-)
create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/en_dim.c
delete mode 100644 drivers/net/ethernet/mellanox/mlx5/core/en_dim.h
delete mode 100644 drivers/net/ethernet/mellanox/mlx5/core/en_rx_am.c
create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/net_dim.c
create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/net_dim.h
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/Makefile b/drivers/net/ethernet/mellanox/mlx5/core/Makefile
index 19b21b4..b46b6de2 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/Makefile
+++ b/drivers/net/ethernet/mellanox/mlx5/core/Makefile
@@ -14,8 +14,8 @@ mlx5_core-$(CONFIG_MLX5_FPGA) += fpga/cmd.o fpga/core.o fpga/conn.o fpga/sdk.o \
fpga/ipsec.o
mlx5_core-$(CONFIG_MLX5_CORE_EN) += en_main.o en_common.o en_fs.o en_ethtool.o \
- en_tx.o en_rx.o en_rx_am.o en_txrx.o en_stats.o vxlan.o \
- en_arfs.o en_fs_ethtool.o en_selftest.o
+ en_tx.o en_rx.o en_dim.o en_txrx.o en_stats.o vxlan.o \
+ en_arfs.o en_fs_ethtool.o en_selftest.o net_dim.o
mlx5_core-$(CONFIG_MLX5_MPFS) += lib/mpfs.o
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h
index 2ccedf6..121f280 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h
@@ -49,7 +49,7 @@
#include "wq.h"
#include "mlx5_core.h"
#include "en_stats.h"
-#include "en_dim.h"
+#include "net_dim.h"
#define MLX5_SET_CFG(p, f, v) MLX5_SET(create_flow_group_in, p, f, v)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_dim.c b/drivers/net/ethernet/mellanox/mlx5/core/en_dim.c
new file mode 100644
index 0000000..b9b434b
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_dim.c
@@ -0,0 +1,48 @@
+/*
+ * Copyright (c) 2016, Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses. You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ * Redistribution and use in source and binary forms, with or
+ * without modification, are permitted provided that the following
+ * conditions are met:
+ *
+ * - Redistributions of source code must retain the above
+ * copyright notice, this list of conditions and the following
+ * disclaimer.
+ *
+ * - Redistributions in binary form must reproduce the above
+ * copyright notice, this list of conditions and the following
+ * disclaimer in the documentation and/or other materials
+ * provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include "en.h"
+
+void mlx5e_rx_am_work(struct work_struct *work)
+{
+ struct mlx5e_rx_am *am = container_of(work, struct mlx5e_rx_am,
+ work);
+ struct mlx5e_rq *rq = container_of(am, struct mlx5e_rq, am);
+ struct mlx5e_cq_moder cur_profile = mlx5e_am_get_profile(am->mode,
+ am->profile_ix);
+
+ mlx5_core_modify_cq_moderation(rq->mdev, &rq->cq.mcq,
+ cur_profile.usec, cur_profile.pkts);
+
+ am->state = MLX5E_AM_START_MEASURE;
+}
+
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_dim.h b/drivers/net/ethernet/mellanox/mlx5/core/en_dim.h
deleted file mode 100644
index c9f0d05..0000000
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_dim.h
+++ /dev/null
@@ -1,108 +0,0 @@
-/*
- * Copyright (c) 2013-2015, Mellanox Technologies, Ltd. All rights reserved.
- * Copyright (c) 2017, Broadcom Limited
- *
- * This software is available to you under a choice of one of two
- * licenses. You may choose to be licensed under the terms of the GNU
- * General Public License (GPL) Version 2, available from the file
- * COPYING in the main directory of this source tree, or the
- * OpenIB.org BSD license below:
- *
- * Redistribution and use in source and binary forms, with or
- * without modification, are permitted provided that the following
- * conditions are met:
- *
- * - Redistributions of source code must retain the above
- * copyright notice, this list of conditions and the following
- * disclaimer.
- *
- * - Redistributions in binary form must reproduce the above
- * copyright notice, this list of conditions and the following
- * disclaimer in the documentation and/or other materials
- * provided with the distribution.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
- * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
- * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
- * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
- * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
- * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
- * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
-*/
-
-#ifndef MLX5_AM_H
-#define MLX5_AM_H
-
-struct mlx5e_cq_moder {
- u16 usec;
- u16 pkts;
- u8 cq_period_mode;
-};
-
-struct mlx5e_rx_am_sample {
- ktime_t time;
- u32 pkt_ctr;
- u32 byte_ctr;
- u16 event_ctr;
-};
-
-struct mlx5e_rx_am_stats {
- int ppms; /* packets per msec */
- int bpms; /* bytes per msec */
- int epms; /* events per msec */
-};
-
-struct mlx5e_rx_am { /* Adaptive Moderation */
- u8 state;
- struct mlx5e_rx_am_stats prev_stats;
- struct mlx5e_rx_am_sample start_sample;
- struct work_struct work;
- u8 profile_ix;
- u8 mode;
- u8 tune_state;
- u8 steps_right;
- u8 steps_left;
- u8 tired;
-};
-
-enum {
- MLX5_CQ_PERIOD_MODE_START_FROM_EQE = 0x0,
- MLX5_CQ_PERIOD_MODE_START_FROM_CQE = 0x1,
- MLX5_CQ_PERIOD_NUM_MODES
-};
-
-/* Adaptive moderation logic */
-enum {
- MLX5E_AM_START_MEASURE,
- MLX5E_AM_MEASURE_IN_PROGRESS,
- MLX5E_AM_APPLY_NEW_PROFILE,
-};
-
-enum {
- MLX5E_AM_PARKING_ON_TOP,
- MLX5E_AM_PARKING_TIRED,
- MLX5E_AM_GOING_RIGHT,
- MLX5E_AM_GOING_LEFT,
-};
-
-enum {
- MLX5E_AM_STATS_WORSE,
- MLX5E_AM_STATS_SAME,
- MLX5E_AM_STATS_BETTER,
-};
-
-enum {
- MLX5E_AM_STEPPED,
- MLX5E_AM_TOO_TIRED,
- MLX5E_AM_ON_EDGE,
-};
-
-void mlx5e_rx_am(struct mlx5e_rx_am *am,
- u16 event_ctr,
- u64 packets,
- u64 bytes);
-void mlx5e_rx_am_work(struct work_struct *work);
-struct mlx5e_cq_moder mlx5e_am_get_def_profile(u8 rx_cq_period_mode);
-
-#endif /* MLX5_AM_H */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rx_am.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rx_am.c
deleted file mode 100644
index 337dd60..0000000
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rx_am.c
+++ /dev/null
@@ -1,320 +0,0 @@
-/*
- * Copyright (c) 2016, Mellanox Technologies. All rights reserved.
- *
- * This software is available to you under a choice of one of two
- * licenses. You may choose to be licensed under the terms of the GNU
- * General Public License (GPL) Version 2, available from the file
- * COPYING in the main directory of this source tree, or the
- * OpenIB.org BSD license below:
- *
- * Redistribution and use in source and binary forms, with or
- * without modification, are permitted provided that the following
- * conditions are met:
- *
- * - Redistributions of source code must retain the above
- * copyright notice, this list of conditions and the following
- * disclaimer.
- *
- * - Redistributions in binary form must reproduce the above
- * copyright notice, this list of conditions and the following
- * disclaimer in the documentation and/or other materials
- * provided with the distribution.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
- * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
- * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
- * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
- * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
- * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
- * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-
-#include "en.h"
-
-/* Adaptive moderation profiles */
-#define MLX5E_AM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE 256
-#define MLX5E_RX_AM_DEF_PROFILE_CQE 1
-#define MLX5E_RX_AM_DEF_PROFILE_EQE 1
-#define MLX5E_PARAMS_AM_NUM_PROFILES 5
-
-/* All profiles sizes must be MLX5E_PARAMS_AM_NUM_PROFILES */
-#define MLX5_AM_EQE_PROFILES { \
- {1, MLX5E_AM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE}, \
- {8, MLX5E_AM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE}, \
- {64, MLX5E_AM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE}, \
- {128, MLX5E_AM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE}, \
- {256, MLX5E_AM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE}, \
-}
-
-#define MLX5_AM_CQE_PROFILES { \
- {2, 256}, \
- {8, 128}, \
- {16, 64}, \
- {32, 64}, \
- {64, 64} \
-}
-
-static const struct mlx5e_cq_moder
-profile[MLX5_CQ_PERIOD_NUM_MODES][MLX5E_PARAMS_AM_NUM_PROFILES] = {
- MLX5_AM_EQE_PROFILES,
- MLX5_AM_CQE_PROFILES,
-};
-
-static inline struct mlx5e_cq_moder mlx5e_am_get_profile(u8 cq_period_mode, int ix)
-{
- struct mlx5e_cq_moder cq_moder;
-
- cq_moder = profile[cq_period_mode][ix];
- cq_moder.cq_period_mode = cq_period_mode;
- return cq_moder;
-}
-
-struct mlx5e_cq_moder mlx5e_am_get_def_profile(u8 rx_cq_period_mode)
-{
- int default_profile_ix;
-
- if (rx_cq_period_mode == MLX5_CQ_PERIOD_MODE_START_FROM_CQE)
- default_profile_ix = MLX5E_RX_AM_DEF_PROFILE_CQE;
- else /* MLX5_CQ_PERIOD_MODE_START_FROM_EQE */
- default_profile_ix = MLX5E_RX_AM_DEF_PROFILE_EQE;
-
- return mlx5e_am_get_profile(rx_cq_period_mode, default_profile_ix);
-}
-
-
-static bool mlx5e_am_on_top(struct mlx5e_rx_am *am)
-{
- switch (am->tune_state) {
- case MLX5E_AM_PARKING_ON_TOP:
- case MLX5E_AM_PARKING_TIRED:
- return true;
- case MLX5E_AM_GOING_RIGHT:
- return (am->steps_left > 1) && (am->steps_right == 1);
- default: /* MLX5E_AM_GOING_LEFT */
- return (am->steps_right > 1) && (am->steps_left == 1);
- }
-}
-
-static void mlx5e_am_turn(struct mlx5e_rx_am *am)
-{
- switch (am->tune_state) {
- case MLX5E_AM_PARKING_ON_TOP:
- case MLX5E_AM_PARKING_TIRED:
- break;
- case MLX5E_AM_GOING_RIGHT:
- am->tune_state = MLX5E_AM_GOING_LEFT;
- am->steps_left = 0;
- break;
- case MLX5E_AM_GOING_LEFT:
- am->tune_state = MLX5E_AM_GOING_RIGHT;
- am->steps_right = 0;
- break;
- }
-}
-
-static int mlx5e_am_step(struct mlx5e_rx_am *am)
-{
- if (am->tired == (MLX5E_PARAMS_AM_NUM_PROFILES * 2))
- return MLX5E_AM_TOO_TIRED;
-
- switch (am->tune_state) {
- case MLX5E_AM_PARKING_ON_TOP:
- case MLX5E_AM_PARKING_TIRED:
- break;
- case MLX5E_AM_GOING_RIGHT:
- if (am->profile_ix == (MLX5E_PARAMS_AM_NUM_PROFILES - 1))
- return MLX5E_AM_ON_EDGE;
- am->profile_ix++;
- am->steps_right++;
- break;
- case MLX5E_AM_GOING_LEFT:
- if (am->profile_ix == 0)
- return MLX5E_AM_ON_EDGE;
- am->profile_ix--;
- am->steps_left++;
- break;
- }
-
- am->tired++;
- return MLX5E_AM_STEPPED;
-}
-
-static void mlx5e_am_park_on_top(struct mlx5e_rx_am *am)
-{
- am->steps_right = 0;
- am->steps_left = 0;
- am->tired = 0;
- am->tune_state = MLX5E_AM_PARKING_ON_TOP;
-}
-
-static void mlx5e_am_park_tired(struct mlx5e_rx_am *am)
-{
- am->steps_right = 0;
- am->steps_left = 0;
- am->tune_state = MLX5E_AM_PARKING_TIRED;
-}
-
-static void mlx5e_am_exit_parking(struct mlx5e_rx_am *am)
-{
- am->tune_state = am->profile_ix ? MLX5E_AM_GOING_LEFT :
- MLX5E_AM_GOING_RIGHT;
- mlx5e_am_step(am);
-}
-
-#define IS_SIGNIFICANT_DIFF(val, ref) \
- (((100 * abs((val) - (ref))) / (ref)) > 10) /* more than 10% difference */
-
-static int mlx5e_am_stats_compare(struct mlx5e_rx_am_stats *curr,
- struct mlx5e_rx_am_stats *prev)
-{
- if (!prev->bpms)
- return curr->bpms ? MLX5E_AM_STATS_BETTER :
- MLX5E_AM_STATS_SAME;
-
- if (IS_SIGNIFICANT_DIFF(curr->bpms, prev->bpms))
- return (curr->bpms > prev->bpms) ? MLX5E_AM_STATS_BETTER :
- MLX5E_AM_STATS_WORSE;
-
- if (IS_SIGNIFICANT_DIFF(curr->ppms, prev->ppms))
- return (curr->ppms > prev->ppms) ? MLX5E_AM_STATS_BETTER :
- MLX5E_AM_STATS_WORSE;
-
- if (IS_SIGNIFICANT_DIFF(curr->epms, prev->epms))
- return (curr->epms < prev->epms) ? MLX5E_AM_STATS_BETTER :
- MLX5E_AM_STATS_WORSE;
-
- return MLX5E_AM_STATS_SAME;
-}
-
-static bool mlx5e_am_decision(struct mlx5e_rx_am_stats *curr_stats,
- struct mlx5e_rx_am *am)
-{
- int prev_state = am->tune_state;
- int prev_ix = am->profile_ix;
- int stats_res;
- int step_res;
-
- switch (am->tune_state) {
- case MLX5E_AM_PARKING_ON_TOP:
- stats_res = mlx5e_am_stats_compare(curr_stats, &am->prev_stats);
- if (stats_res != MLX5E_AM_STATS_SAME)
- mlx5e_am_exit_parking(am);
- break;
-
- case MLX5E_AM_PARKING_TIRED:
- am->tired--;
- if (!am->tired)
- mlx5e_am_exit_parking(am);
- break;
-
- case MLX5E_AM_GOING_RIGHT:
- case MLX5E_AM_GOING_LEFT:
- stats_res = mlx5e_am_stats_compare(curr_stats, &am->prev_stats);
- if (stats_res != MLX5E_AM_STATS_BETTER)
- mlx5e_am_turn(am);
-
- if (mlx5e_am_on_top(am)) {
- mlx5e_am_park_on_top(am);
- break;
- }
-
- step_res = mlx5e_am_step(am);
- switch (step_res) {
- case MLX5E_AM_ON_EDGE:
- mlx5e_am_park_on_top(am);
- break;
- case MLX5E_AM_TOO_TIRED:
- mlx5e_am_park_tired(am);
- break;
- }
-
- break;
- }
-
- if ((prev_state != MLX5E_AM_PARKING_ON_TOP) ||
- (am->tune_state != MLX5E_AM_PARKING_ON_TOP))
- am->prev_stats = *curr_stats;
-
- return am->profile_ix != prev_ix;
-}
-
-static void mlx5e_am_sample(u16 event_ctr,
- u64 packets,
- u64 bytes,
- struct mlx5e_rx_am_sample *s)
-{
- s->time = ktime_get();
- s->pkt_ctr = packets;
- s->byte_ctr = bytes;
- s->event_ctr = event_ctr;
-}
-
-#define MLX5E_AM_NEVENTS 64
-#define BITS_PER_TYPE(type) (sizeof(type) * BITS_PER_BYTE)
-#define BIT_GAP(bits, end, start) ((((end) - (start)) + BIT_ULL(bits)) & (BIT_ULL(bits) - 1))
-
-static void mlx5e_am_calc_stats(struct mlx5e_rx_am_sample *start,
- struct mlx5e_rx_am_sample *end,
- struct mlx5e_rx_am_stats *curr_stats)
-{
- /* u32 holds up to 71 minutes, should be enough */
- u32 delta_us = ktime_us_delta(end->time, start->time);
- u32 npkts = BIT_GAP(BITS_PER_TYPE(u32), end->pkt_ctr, start->pkt_ctr);
- u32 nbytes = BIT_GAP(BITS_PER_TYPE(u32), end->byte_ctr,
- start->byte_ctr);
-
- if (!delta_us)
- return;
-
- curr_stats->ppms = DIV_ROUND_UP(npkts * USEC_PER_MSEC, delta_us);
- curr_stats->bpms = DIV_ROUND_UP(nbytes * USEC_PER_MSEC, delta_us);
- curr_stats->epms = DIV_ROUND_UP(MLX5E_AM_NEVENTS * USEC_PER_MSEC,
- delta_us);
-}
-
-void mlx5e_rx_am_work(struct work_struct *work)
-{
- struct mlx5e_rx_am *am = container_of(work, struct mlx5e_rx_am,
- work);
- struct mlx5e_rq *rq = container_of(am, struct mlx5e_rq, am);
- struct mlx5e_cq_moder cur_profile = profile[am->mode][am->profile_ix];
-
- mlx5_core_modify_cq_moderation(rq->mdev, &rq->cq.mcq,
- cur_profile.usec, cur_profile.pkts);
-
- am->state = MLX5E_AM_START_MEASURE;
-}
-
-void mlx5e_rx_am(struct mlx5e_rx_am *am,
- u16 event_ctr,
- u64 packets,
- u64 bytes)
-{
- struct mlx5e_rx_am_sample end_sample;
- struct mlx5e_rx_am_stats curr_stats;
- u16 nevents;
-
- switch (am->state) {
- case MLX5E_AM_MEASURE_IN_PROGRESS:
- nevents = BIT_GAP(BITS_PER_TYPE(u16), event_ctr,
- am->start_sample.event_ctr);
- if (nevents < MLX5E_AM_NEVENTS)
- break;
- mlx5e_am_sample(event_ctr, packets, bytes, &end_sample);
- mlx5e_am_calc_stats(&am->start_sample, &end_sample,
- &curr_stats);
- if (mlx5e_am_decision(&curr_stats, am)) {
- am->state = MLX5E_AM_APPLY_NEW_PROFILE;
- schedule_work(&am->work);
- break;
- }
- /* fall through */
- case MLX5E_AM_START_MEASURE:
- mlx5e_am_sample(event_ctr, packets, bytes, &am->start_sample);
- am->state = MLX5E_AM_MEASURE_IN_PROGRESS;
- break;
- case MLX5E_AM_APPLY_NEW_PROFILE:
- break;
- }
-}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/net_dim.c b/drivers/net/ethernet/mellanox/mlx5/core/net_dim.c
new file mode 100644
index 0000000..ca05f4e
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/net_dim.c
@@ -0,0 +1,307 @@
+/*
+ * Copyright (c) 2016, Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2017, Broadcom Limiited. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses. You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ * Redistribution and use in source and binary forms, with or
+ * without modification, are permitted provided that the following
+ * conditions are met:
+ *
+ * - Redistributions of source code must retain the above
+ * copyright notice, this list of conditions and the following
+ * disclaimer.
+ *
+ * - Redistributions in binary form must reproduce the above
+ * copyright notice, this list of conditions and the following
+ * disclaimer in the documentation and/or other materials
+ * provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include "en.h"
+
+#define MLX5E_PARAMS_AM_NUM_PROFILES 5
+/* Adaptive moderation profiles */
+#define MLX5E_AM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE 256
+#define MLX5E_RX_AM_DEF_PROFILE_CQE 1
+#define MLX5E_RX_AM_DEF_PROFILE_EQE 1
+
+/* All profiles sizes must be MLX5E_PARAMS_AM_NUM_PROFILES */
+#define MLX5_AM_EQE_PROFILES { \
+ {1, MLX5E_AM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE}, \
+ {8, MLX5E_AM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE}, \
+ {64, MLX5E_AM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE}, \
+ {128, MLX5E_AM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE}, \
+ {256, MLX5E_AM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE}, \
+}
+
+#define MLX5_AM_CQE_PROFILES { \
+ {2, 256}, \
+ {8, 128}, \
+ {16, 64}, \
+ {32, 64}, \
+ {64, 64} \
+}
+
+static const struct mlx5e_cq_moder
+profile[MLX5_CQ_PERIOD_NUM_MODES][MLX5E_PARAMS_AM_NUM_PROFILES] = {
+ MLX5_AM_EQE_PROFILES,
+ MLX5_AM_CQE_PROFILES,
+};
+
+struct mlx5e_cq_moder mlx5e_am_get_profile(u8 cq_period_mode, int ix)
+{
+ struct mlx5e_cq_moder cq_moder;
+
+ cq_moder = profile[cq_period_mode][ix];
+ cq_moder.cq_period_mode = cq_period_mode;
+ return cq_moder;
+}
+
+struct mlx5e_cq_moder mlx5e_am_get_def_profile(u8 rx_cq_period_mode)
+{
+ int default_profile_ix;
+
+ if (rx_cq_period_mode == MLX5_CQ_PERIOD_MODE_START_FROM_CQE)
+ default_profile_ix = MLX5E_RX_AM_DEF_PROFILE_CQE;
+ else /* MLX5_CQ_PERIOD_MODE_START_FROM_EQE */
+ default_profile_ix = MLX5E_RX_AM_DEF_PROFILE_EQE;
+
+ return mlx5e_am_get_profile(rx_cq_period_mode, default_profile_ix);
+}
+
+static bool mlx5e_am_on_top(struct mlx5e_rx_am *am)
+{
+ switch (am->tune_state) {
+ case MLX5E_AM_PARKING_ON_TOP:
+ case MLX5E_AM_PARKING_TIRED:
+ return true;
+ case MLX5E_AM_GOING_RIGHT:
+ return (am->steps_left > 1) && (am->steps_right == 1);
+ default: /* MLX5E_AM_GOING_LEFT */
+ return (am->steps_right > 1) && (am->steps_left == 1);
+ }
+}
+
+static void mlx5e_am_turn(struct mlx5e_rx_am *am)
+{
+ switch (am->tune_state) {
+ case MLX5E_AM_PARKING_ON_TOP:
+ case MLX5E_AM_PARKING_TIRED:
+ break;
+ case MLX5E_AM_GOING_RIGHT:
+ am->tune_state = MLX5E_AM_GOING_LEFT;
+ am->steps_left = 0;
+ break;
+ case MLX5E_AM_GOING_LEFT:
+ am->tune_state = MLX5E_AM_GOING_RIGHT;
+ am->steps_right = 0;
+ break;
+ }
+}
+
+static int mlx5e_am_step(struct mlx5e_rx_am *am)
+{
+ if (am->tired == (MLX5E_PARAMS_AM_NUM_PROFILES * 2))
+ return MLX5E_AM_TOO_TIRED;
+
+ switch (am->tune_state) {
+ case MLX5E_AM_PARKING_ON_TOP:
+ case MLX5E_AM_PARKING_TIRED:
+ break;
+ case MLX5E_AM_GOING_RIGHT:
+ if (am->profile_ix == (MLX5E_PARAMS_AM_NUM_PROFILES - 1))
+ return MLX5E_AM_ON_EDGE;
+ am->profile_ix++;
+ am->steps_right++;
+ break;
+ case MLX5E_AM_GOING_LEFT:
+ if (am->profile_ix == 0)
+ return MLX5E_AM_ON_EDGE;
+ am->profile_ix--;
+ am->steps_left++;
+ break;
+ }
+
+ am->tired++;
+ return MLX5E_AM_STEPPED;
+}
+
+static void mlx5e_am_park_on_top(struct mlx5e_rx_am *am)
+{
+ am->steps_right = 0;
+ am->steps_left = 0;
+ am->tired = 0;
+ am->tune_state = MLX5E_AM_PARKING_ON_TOP;
+}
+
+static void mlx5e_am_park_tired(struct mlx5e_rx_am *am)
+{
+ am->steps_right = 0;
+ am->steps_left = 0;
+ am->tune_state = MLX5E_AM_PARKING_TIRED;
+}
+
+static void mlx5e_am_exit_parking(struct mlx5e_rx_am *am)
+{
+ am->tune_state = am->profile_ix ? MLX5E_AM_GOING_LEFT :
+ MLX5E_AM_GOING_RIGHT;
+ mlx5e_am_step(am);
+}
+
+#define IS_SIGNIFICANT_DIFF(val, ref) \
+ (((100 * abs((val) - (ref))) / (ref)) > 10) /* more than 10% difference */
+
+static int mlx5e_am_stats_compare(struct mlx5e_rx_am_stats *curr,
+ struct mlx5e_rx_am_stats *prev)
+{
+ if (!prev->bpms)
+ return curr->bpms ? MLX5E_AM_STATS_BETTER :
+ MLX5E_AM_STATS_SAME;
+
+ if (IS_SIGNIFICANT_DIFF(curr->bpms, prev->bpms))
+ return (curr->bpms > prev->bpms) ? MLX5E_AM_STATS_BETTER :
+ MLX5E_AM_STATS_WORSE;
+
+ if (IS_SIGNIFICANT_DIFF(curr->ppms, prev->ppms))
+ return (curr->ppms > prev->ppms) ? MLX5E_AM_STATS_BETTER :
+ MLX5E_AM_STATS_WORSE;
+
+ if (IS_SIGNIFICANT_DIFF(curr->epms, prev->epms))
+ return (curr->epms < prev->epms) ? MLX5E_AM_STATS_BETTER :
+ MLX5E_AM_STATS_WORSE;
+
+ return MLX5E_AM_STATS_SAME;
+}
+
+static bool mlx5e_am_decision(struct mlx5e_rx_am_stats *curr_stats,
+ struct mlx5e_rx_am *am)
+{
+ int prev_state = am->tune_state;
+ int prev_ix = am->profile_ix;
+ int stats_res;
+ int step_res;
+
+ switch (am->tune_state) {
+ case MLX5E_AM_PARKING_ON_TOP:
+ stats_res = mlx5e_am_stats_compare(curr_stats, &am->prev_stats);
+ if (stats_res != MLX5E_AM_STATS_SAME)
+ mlx5e_am_exit_parking(am);
+ break;
+
+ case MLX5E_AM_PARKING_TIRED:
+ am->tired--;
+ if (!am->tired)
+ mlx5e_am_exit_parking(am);
+ break;
+
+ case MLX5E_AM_GOING_RIGHT:
+ case MLX5E_AM_GOING_LEFT:
+ stats_res = mlx5e_am_stats_compare(curr_stats, &am->prev_stats);
+ if (stats_res != MLX5E_AM_STATS_BETTER)
+ mlx5e_am_turn(am);
+
+ if (mlx5e_am_on_top(am)) {
+ mlx5e_am_park_on_top(am);
+ break;
+ }
+
+ step_res = mlx5e_am_step(am);
+ switch (step_res) {
+ case MLX5E_AM_ON_EDGE:
+ mlx5e_am_park_on_top(am);
+ break;
+ case MLX5E_AM_TOO_TIRED:
+ mlx5e_am_park_tired(am);
+ break;
+ }
+
+ break;
+ }
+
+ if ((prev_state != MLX5E_AM_PARKING_ON_TOP) ||
+ (am->tune_state != MLX5E_AM_PARKING_ON_TOP))
+ am->prev_stats = *curr_stats;
+
+ return am->profile_ix != prev_ix;
+}
+
+static void mlx5e_am_sample(u16 event_ctr,
+ u64 packets,
+ u64 bytes,
+ struct mlx5e_rx_am_sample *s)
+{
+ s->time = ktime_get();
+ s->pkt_ctr = packets;
+ s->byte_ctr = bytes;
+ s->event_ctr = event_ctr;
+}
+
+#define MLX5E_AM_NEVENTS 64
+#define BITS_PER_TYPE(type) (sizeof(type) * BITS_PER_BYTE)
+#define BIT_GAP(bits, end, start) ((((end) - (start)) + BIT_ULL(bits)) & (BIT_ULL(bits) - 1))
+
+static void mlx5e_am_calc_stats(struct mlx5e_rx_am_sample *start,
+ struct mlx5e_rx_am_sample *end,
+ struct mlx5e_rx_am_stats *curr_stats)
+{
+ /* u32 holds up to 71 minutes, should be enough */
+ u32 delta_us = ktime_us_delta(end->time, start->time);
+ u32 npkts = BIT_GAP(BITS_PER_TYPE(u32), end->pkt_ctr, start->pkt_ctr);
+ u32 nbytes = BIT_GAP(BITS_PER_TYPE(u32), end->byte_ctr,
+ start->byte_ctr);
+
+ if (!delta_us)
+ return;
+
+ curr_stats->ppms = DIV_ROUND_UP(npkts * USEC_PER_MSEC, delta_us);
+ curr_stats->bpms = DIV_ROUND_UP(nbytes * USEC_PER_MSEC, delta_us);
+ curr_stats->epms = DIV_ROUND_UP(MLX5E_AM_NEVENTS * USEC_PER_MSEC,
+ delta_us);
+}
+
+void mlx5e_rx_am(struct mlx5e_rx_am *am,
+ u16 event_ctr,
+ u64 packets,
+ u64 bytes)
+{
+ struct mlx5e_rx_am_sample end_sample;
+ struct mlx5e_rx_am_stats curr_stats;
+ u16 nevents;
+
+ switch (am->state) {
+ case MLX5E_AM_MEASURE_IN_PROGRESS:
+ nevents = BIT_GAP(BITS_PER_TYPE(u16), event_ctr,
+ am->start_sample.event_ctr);
+ if (nevents < MLX5E_AM_NEVENTS)
+ break;
+ mlx5e_am_sample(event_ctr, packets, bytes, &end_sample);
+ mlx5e_am_calc_stats(&am->start_sample, &end_sample,
+ &curr_stats);
+ if (mlx5e_am_decision(&curr_stats, am)) {
+ am->state = MLX5E_AM_APPLY_NEW_PROFILE;
+ schedule_work(&am->work);
+ break;
+ }
+ /* fall through */
+ case MLX5E_AM_START_MEASURE:
+ mlx5e_am_sample(event_ctr, packets, bytes, &am->start_sample);
+ am->state = MLX5E_AM_MEASURE_IN_PROGRESS;
+ break;
+ case MLX5E_AM_APPLY_NEW_PROFILE:
+ break;
+ }
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/net_dim.h b/drivers/net/ethernet/mellanox/mlx5/core/net_dim.h
new file mode 100644
index 0000000..5ce8e54
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/net_dim.h
@@ -0,0 +1,109 @@
+/*
+ * Copyright (c) 2013-2015, Mellanox Technologies, Ltd. All rights reserved.
+ * Copyright (c) 2017, Broadcom Limited
+ *
+ * This software is available to you under a choice of one of two
+ * licenses. You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ * Redistribution and use in source and binary forms, with or
+ * without modification, are permitted provided that the following
+ * conditions are met:
+ *
+ * - Redistributions of source code must retain the above
+ * copyright notice, this list of conditions and the following
+ * disclaimer.
+ *
+ * - Redistributions in binary form must reproduce the above
+ * copyright notice, this list of conditions and the following
+ * disclaimer in the documentation and/or other materials
+ * provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+*/
+
+#ifndef MLX5_AM_H
+#define MLX5_AM_H
+
+struct mlx5e_cq_moder {
+ u16 usec;
+ u16 pkts;
+ u8 cq_period_mode;
+};
+
+struct mlx5e_rx_am_sample {
+ ktime_t time;
+ u32 pkt_ctr;
+ u32 byte_ctr;
+ u16 event_ctr;
+};
+
+struct mlx5e_rx_am_stats {
+ int ppms; /* packets per msec */
+ int bpms; /* bytes per msec */
+ int epms; /* events per msec */
+};
+
+struct mlx5e_rx_am { /* Adaptive Moderation */
+ u8 state;
+ struct mlx5e_rx_am_stats prev_stats;
+ struct mlx5e_rx_am_sample start_sample;
+ struct work_struct work;
+ u8 profile_ix;
+ u8 mode;
+ u8 tune_state;
+ u8 steps_right;
+ u8 steps_left;
+ u8 tired;
+};
+
+enum {
+ MLX5_CQ_PERIOD_MODE_START_FROM_EQE = 0x0,
+ MLX5_CQ_PERIOD_MODE_START_FROM_CQE = 0x1,
+ MLX5_CQ_PERIOD_NUM_MODES
+};
+
+/* Adaptive moderation logic */
+enum {
+ MLX5E_AM_START_MEASURE,
+ MLX5E_AM_MEASURE_IN_PROGRESS,
+ MLX5E_AM_APPLY_NEW_PROFILE,
+};
+
+enum {
+ MLX5E_AM_PARKING_ON_TOP,
+ MLX5E_AM_PARKING_TIRED,
+ MLX5E_AM_GOING_RIGHT,
+ MLX5E_AM_GOING_LEFT,
+};
+
+enum {
+ MLX5E_AM_STATS_WORSE,
+ MLX5E_AM_STATS_SAME,
+ MLX5E_AM_STATS_BETTER,
+};
+
+enum {
+ MLX5E_AM_STEPPED,
+ MLX5E_AM_TOO_TIRED,
+ MLX5E_AM_ON_EDGE,
+};
+
+void mlx5e_rx_am(struct mlx5e_rx_am *am,
+ u16 event_ctr,
+ u64 packets,
+ u64 bytes);
+void mlx5e_rx_am_work(struct work_struct *work);
+struct mlx5e_cq_moder mlx5e_am_get_def_profile(u8 rx_cq_period_mode);
+struct mlx5e_cq_moder mlx5e_am_get_profile(u8 cq_period_mode, int ix);
+
+#endif /* MLX5_AM_H */
--
2.7.4
^ permalink raw reply related
* [PATCH net-next v2 06/10] net/mlx5e: Change Mellanox references in DIM code
From: Andy Gospodarek @ 2018-01-05 22:58 UTC (permalink / raw)
To: netdev; +Cc: mchan, talgi, ogerlitz, Andy Gospodarek
In-Reply-To: <1515193094-9554-1-git-send-email-andy@greyhouse.net>
From: Andy Gospodarek <gospo@broadcom.com>
Change all appropriate mlx5_am* and MLX5_AM* references to net_dim and
NET_DIM, respectively, in code that handles dynamic interrupt
moderation. Also change all references from 'am' to 'dim' when used as
local variables.
Signed-off-by: Andy Gospodarek <gospo@broadcom.com>
Acked-by: Tal Gilboa <talgi@mellanox.com>
Acked-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en.h | 12 +-
drivers/net/ethernet/mellanox/mlx5/core/en_dim.c | 14 +-
.../net/ethernet/mellanox/mlx5/core/en_ethtool.c | 12 +-
drivers/net/ethernet/mellanox/mlx5/core/en_main.c | 52 ++--
drivers/net/ethernet/mellanox/mlx5/core/en_rep.c | 4 +-
drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c | 2 +-
drivers/net/ethernet/mellanox/mlx5/core/net_dim.c | 284 ++++++++++-----------
drivers/net/ethernet/mellanox/mlx5/core/net_dim.h | 63 +++--
8 files changed, 232 insertions(+), 211 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h
index 121f280..732f275 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h
@@ -115,6 +115,9 @@
#define MLX5E_PARAMS_DEFAULT_MIN_RX_WQES 0x80
#define MLX5E_PARAMS_DEFAULT_MIN_RX_WQES_MPW 0x2
+#define MLX5E_CQ_PERIOD_MODE_START_FROM_EQE 0x0
+#define MLX5E_CQ_PERIOD_MODE_START_FROM_CQE 0x1
+
#define MLX5E_LOG_INDIR_RQT_SIZE 0x7
#define MLX5E_INDIR_RQT_SIZE BIT(MLX5E_LOG_INDIR_RQT_SIZE)
#define MLX5E_MIN_NUM_CHANNELS 0x1
@@ -237,8 +240,8 @@ struct mlx5e_params {
u16 num_channels;
u8 num_tc;
bool rx_cqe_compress_def;
- struct mlx5e_cq_moder rx_cq_moderation;
- struct mlx5e_cq_moder tx_cq_moderation;
+ struct net_dim_cq_moder rx_cq_moderation;
+ struct net_dim_cq_moder tx_cq_moderation;
bool lro_en;
u32 lro_wqe_sz;
u16 tx_max_inline;
@@ -248,7 +251,7 @@ struct mlx5e_params {
u32 indirection_rqt[MLX5E_INDIR_RQT_SIZE];
bool vlan_strip_disable;
bool scatter_fcs_en;
- bool rx_am_enabled;
+ bool rx_dim_enabled;
u32 lro_timeout;
u32 pflags;
struct bpf_prog *xdp_prog;
@@ -527,7 +530,7 @@ struct mlx5e_rq {
unsigned long state;
int ix;
- struct mlx5e_rx_am am; /* Adaptive Moderation */
+ struct net_dim dim; /* Dynamic Interrupt Moderation */
/* XDP */
struct bpf_prog *xdp_prog;
@@ -1075,4 +1078,5 @@ void mlx5e_build_nic_params(struct mlx5_core_dev *mdev,
struct mlx5e_params *params,
u16 max_channels);
u8 mlx5e_params_calculate_tx_min_inline(struct mlx5_core_dev *mdev);
+void mlx5e_rx_dim_work(struct work_struct *work);
#endif /* __MLX5_EN_H__ */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_dim.c b/drivers/net/ethernet/mellanox/mlx5/core/en_dim.c
index b9b434b..f620325 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_dim.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_dim.c
@@ -32,17 +32,17 @@
#include "en.h"
-void mlx5e_rx_am_work(struct work_struct *work)
+void mlx5e_rx_dim_work(struct work_struct *work)
{
- struct mlx5e_rx_am *am = container_of(work, struct mlx5e_rx_am,
- work);
- struct mlx5e_rq *rq = container_of(am, struct mlx5e_rq, am);
- struct mlx5e_cq_moder cur_profile = mlx5e_am_get_profile(am->mode,
- am->profile_ix);
+ struct net_dim *dim = container_of(work, struct net_dim,
+ work);
+ struct mlx5e_rq *rq = container_of(dim, struct mlx5e_rq, dim);
+ struct net_dim_cq_moder cur_profile = net_dim_get_profile(dim->mode,
+ dim->profile_ix);
mlx5_core_modify_cq_moderation(rq->mdev, &rq->cq.mcq,
cur_profile.usec, cur_profile.pkts);
- am->state = MLX5E_AM_START_MEASURE;
+ dim->state = NET_DIM_START_MEASURE;
}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
index 8f05efa..62ac4c9 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c
@@ -480,7 +480,7 @@ int mlx5e_ethtool_get_coalesce(struct mlx5e_priv *priv,
coal->rx_max_coalesced_frames = priv->channels.params.rx_cq_moderation.pkts;
coal->tx_coalesce_usecs = priv->channels.params.tx_cq_moderation.usec;
coal->tx_max_coalesced_frames = priv->channels.params.tx_cq_moderation.pkts;
- coal->use_adaptive_rx_coalesce = priv->channels.params.rx_am_enabled;
+ coal->use_adaptive_rx_coalesce = priv->channels.params.rx_dim_enabled;
return 0;
}
@@ -534,7 +534,7 @@ int mlx5e_ethtool_set_coalesce(struct mlx5e_priv *priv,
new_channels.params.tx_cq_moderation.pkts = coal->tx_max_coalesced_frames;
new_channels.params.rx_cq_moderation.usec = coal->rx_coalesce_usecs;
new_channels.params.rx_cq_moderation.pkts = coal->rx_max_coalesced_frames;
- new_channels.params.rx_am_enabled = !!coal->use_adaptive_rx_coalesce;
+ new_channels.params.rx_dim_enabled = !!coal->use_adaptive_rx_coalesce;
if (!test_bit(MLX5E_STATE_OPENED, &priv->state)) {
priv->channels.params = new_channels.params;
@@ -542,7 +542,7 @@ int mlx5e_ethtool_set_coalesce(struct mlx5e_priv *priv,
}
/* we are opened */
- reset = !!coal->use_adaptive_rx_coalesce != priv->channels.params.rx_am_enabled;
+ reset = !!coal->use_adaptive_rx_coalesce != priv->channels.params.rx_dim_enabled;
if (!reset) {
mlx5e_set_priv_channels_coalesce(priv, coal);
priv->channels.params = new_channels.params;
@@ -1465,14 +1465,14 @@ static int set_pflag_cqe_based_moder(struct net_device *netdev, bool enable,
int err = 0;
cq_period_mode = enable ?
- MLX5_CQ_PERIOD_MODE_START_FROM_CQE :
- MLX5_CQ_PERIOD_MODE_START_FROM_EQE;
+ MLX5E_CQ_PERIOD_MODE_START_FROM_CQE :
+ MLX5E_CQ_PERIOD_MODE_START_FROM_EQE;
current_cq_period_mode = is_rx_cq ?
priv->channels.params.rx_cq_moderation.cq_period_mode :
priv->channels.params.tx_cq_moderation.cq_period_mode;
mode_changed = cq_period_mode != current_cq_period_mode;
- if (cq_period_mode == MLX5_CQ_PERIOD_MODE_START_FROM_CQE &&
+ if (cq_period_mode == MLX5E_CQ_PERIOD_MODE_START_FROM_CQE &&
!MLX5_CAP_GEN(mdev, cq_period_start_from_cqe))
return -EOPNOTSUPP;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
index 3aa1c90..fb78d2d 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -674,8 +674,17 @@ static int mlx5e_alloc_rq(struct mlx5e_channel *c,
wqe->data.lkey = rq->mkey_be;
}
- INIT_WORK(&rq->am.work, mlx5e_rx_am_work);
- rq->am.mode = params->rx_cq_moderation.cq_period_mode;
+ INIT_WORK(&rq->dim.work, mlx5e_rx_dim_work);
+
+ switch (params->rx_cq_moderation.cq_period_mode) {
+ case MLX5E_CQ_PERIOD_MODE_START_FROM_CQE:
+ rq->dim.mode = NET_DIM_CQ_PERIOD_MODE_START_FROM_CQE;
+ break;
+ case MLX5E_CQ_PERIOD_MODE_START_FROM_EQE:
+ default:
+ rq->dim.mode = NET_DIM_CQ_PERIOD_MODE_START_FROM_EQE;
+ }
+
rq->page_cache.head = 0;
rq->page_cache.tail = 0;
@@ -919,7 +928,7 @@ static int mlx5e_open_rq(struct mlx5e_channel *c,
if (err)
goto err_destroy_rq;
- if (params->rx_am_enabled)
+ if (params->rx_dim_enabled)
c->rq.state |= BIT(MLX5E_RQ_STATE_AM);
return 0;
@@ -952,7 +961,7 @@ static void mlx5e_deactivate_rq(struct mlx5e_rq *rq)
static void mlx5e_close_rq(struct mlx5e_rq *rq)
{
- cancel_work_sync(&rq->am.work);
+ cancel_work_sync(&rq->dim.work);
mlx5e_destroy_rq(rq);
mlx5e_free_rx_descs(rq);
mlx5e_free_rq(rq);
@@ -1565,7 +1574,7 @@ static void mlx5e_destroy_cq(struct mlx5e_cq *cq)
}
static int mlx5e_open_cq(struct mlx5e_channel *c,
- struct mlx5e_cq_moder moder,
+ struct net_dim_cq_moder moder,
struct mlx5e_cq_param *param,
struct mlx5e_cq *cq)
{
@@ -1747,7 +1756,7 @@ static int mlx5e_open_channel(struct mlx5e_priv *priv, int ix,
struct mlx5e_channel_param *cparam,
struct mlx5e_channel **cp)
{
- struct mlx5e_cq_moder icocq_moder = {0, 0};
+ struct net_dim_cq_moder icocq_moder = {0, 0};
struct net_device *netdev = priv->netdev;
int cpu = mlx5e_get_cpu(priv, ix);
struct mlx5e_channel *c;
@@ -1999,7 +2008,7 @@ static void mlx5e_build_ico_cq_param(struct mlx5e_priv *priv,
mlx5e_build_common_cq_param(priv, param);
- param->cq_period_mode = MLX5_CQ_PERIOD_MODE_START_FROM_EQE;
+ param->cq_period_mode = NET_DIM_CQ_PERIOD_MODE_START_FROM_EQE;
}
static void mlx5e_build_icosq_param(struct mlx5e_priv *priv,
@@ -4016,13 +4025,13 @@ void mlx5e_set_tx_cq_mode_params(struct mlx5e_params *params, u8 cq_period_mode)
params->tx_cq_moderation.usec =
MLX5E_PARAMS_DEFAULT_TX_CQ_MODERATION_USEC;
- if (cq_period_mode == MLX5_CQ_PERIOD_MODE_START_FROM_CQE)
+ if (cq_period_mode == MLX5E_CQ_PERIOD_MODE_START_FROM_CQE)
params->tx_cq_moderation.usec =
MLX5E_PARAMS_DEFAULT_TX_CQ_MODERATION_USEC_FROM_CQE;
MLX5E_SET_PFLAG(params, MLX5E_PFLAG_TX_CQE_BASED_MODER,
params->tx_cq_moderation.cq_period_mode ==
- MLX5_CQ_PERIOD_MODE_START_FROM_CQE);
+ MLX5E_CQ_PERIOD_MODE_START_FROM_CQE);
}
void mlx5e_set_rx_cq_mode_params(struct mlx5e_params *params, u8 cq_period_mode)
@@ -4034,17 +4043,26 @@ void mlx5e_set_rx_cq_mode_params(struct mlx5e_params *params, u8 cq_period_mode)
params->rx_cq_moderation.usec =
MLX5E_PARAMS_DEFAULT_RX_CQ_MODERATION_USEC;
- if (cq_period_mode == MLX5_CQ_PERIOD_MODE_START_FROM_CQE)
+ if (cq_period_mode == MLX5E_CQ_PERIOD_MODE_START_FROM_CQE)
params->rx_cq_moderation.usec =
MLX5E_PARAMS_DEFAULT_RX_CQ_MODERATION_USEC_FROM_CQE;
- if (params->rx_am_enabled)
- params->rx_cq_moderation =
- mlx5e_am_get_def_profile(cq_period_mode);
+ if (params->rx_dim_enabled) {
+ switch (cq_period_mode) {
+ case MLX5E_CQ_PERIOD_MODE_START_FROM_CQE:
+ params->rx_cq_moderation =
+ net_dim_get_def_profile(NET_DIM_CQ_PERIOD_MODE_START_FROM_CQE);
+ break;
+ case MLX5E_CQ_PERIOD_MODE_START_FROM_EQE:
+ default:
+ params->rx_cq_moderation =
+ net_dim_get_def_profile(NET_DIM_CQ_PERIOD_MODE_START_FROM_EQE);
+ }
+ }
MLX5E_SET_PFLAG(params, MLX5E_PFLAG_RX_CQE_BASED_MODER,
params->rx_cq_moderation.cq_period_mode ==
- MLX5_CQ_PERIOD_MODE_START_FROM_CQE);
+ MLX5E_CQ_PERIOD_MODE_START_FROM_CQE);
}
u32 mlx5e_choose_lro_timeout(struct mlx5_core_dev *mdev, u32 wanted_timeout)
@@ -4100,9 +4118,9 @@ void mlx5e_build_nic_params(struct mlx5_core_dev *mdev,
/* CQ moderation params */
cq_period_mode = MLX5_CAP_GEN(mdev, cq_period_start_from_cqe) ?
- MLX5_CQ_PERIOD_MODE_START_FROM_CQE :
- MLX5_CQ_PERIOD_MODE_START_FROM_EQE;
- params->rx_am_enabled = MLX5_CAP_GEN(mdev, cq_moderation);
+ MLX5E_CQ_PERIOD_MODE_START_FROM_CQE :
+ MLX5E_CQ_PERIOD_MODE_START_FROM_EQE;
+ params->rx_dim_enabled = MLX5_CAP_GEN(mdev, cq_moderation);
mlx5e_set_rx_cq_mode_params(params, cq_period_mode);
mlx5e_set_tx_cq_mode_params(params, cq_period_mode);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
index c6a77f8..05771fb 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c
@@ -877,8 +877,8 @@ static void mlx5e_build_rep_params(struct mlx5_core_dev *mdev,
struct mlx5e_params *params)
{
u8 cq_period_mode = MLX5_CAP_GEN(mdev, cq_period_start_from_cqe) ?
- MLX5_CQ_PERIOD_MODE_START_FROM_CQE :
- MLX5_CQ_PERIOD_MODE_START_FROM_EQE;
+ MLX5E_CQ_PERIOD_MODE_START_FROM_CQE :
+ MLX5E_CQ_PERIOD_MODE_START_FROM_EQE;
params->log_sq_size = MLX5E_PARAMS_MINIMUM_LOG_SQ_SIZE;
params->rq_wq_type = MLX5_WQ_TYPE_LINKED_LIST;
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c
index 1849169..dae77a9 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c
@@ -79,7 +79,7 @@ int mlx5e_napi_poll(struct napi_struct *napi, int budget)
mlx5e_cq_arm(&c->sq[i].cq);
if (MLX5E_TEST_BIT(c->rq.state, MLX5E_RQ_STATE_AM))
- mlx5e_rx_am(&c->rq.am,
+ net_dim(&c->rq.dim,
c->rq.cq.event_ctr,
c->rq.stats.packets,
c->rq.stats.bytes);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/net_dim.c b/drivers/net/ethernet/mellanox/mlx5/core/net_dim.c
index ca05f4e..00b9ae3 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/net_dim.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/net_dim.c
@@ -33,22 +33,22 @@
#include "en.h"
-#define MLX5E_PARAMS_AM_NUM_PROFILES 5
+#define NET_DIM_PARAMS_NUM_PROFILES 5
/* Adaptive moderation profiles */
-#define MLX5E_AM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE 256
-#define MLX5E_RX_AM_DEF_PROFILE_CQE 1
-#define MLX5E_RX_AM_DEF_PROFILE_EQE 1
-
-/* All profiles sizes must be MLX5E_PARAMS_AM_NUM_PROFILES */
-#define MLX5_AM_EQE_PROFILES { \
- {1, MLX5E_AM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE}, \
- {8, MLX5E_AM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE}, \
- {64, MLX5E_AM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE}, \
- {128, MLX5E_AM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE}, \
- {256, MLX5E_AM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE}, \
+#define NET_DIM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE 256
+#define NET_DIM_DEF_PROFILE_CQE 1
+#define NET_DIM_DEF_PROFILE_EQE 1
+
+/* All profiles sizes must be NET_PARAMS_DIM_NUM_PROFILES */
+#define NET_DIM_EQE_PROFILES { \
+ {1, NET_DIM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE}, \
+ {8, NET_DIM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE}, \
+ {64, NET_DIM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE}, \
+ {128, NET_DIM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE}, \
+ {256, NET_DIM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE}, \
}
-#define MLX5_AM_CQE_PROFILES { \
+#define NET_DIM_CQE_PROFILES { \
{2, 256}, \
{8, 128}, \
{16, 64}, \
@@ -56,193 +56,193 @@
{64, 64} \
}
-static const struct mlx5e_cq_moder
-profile[MLX5_CQ_PERIOD_NUM_MODES][MLX5E_PARAMS_AM_NUM_PROFILES] = {
- MLX5_AM_EQE_PROFILES,
- MLX5_AM_CQE_PROFILES,
+static const struct net_dim_cq_moder
+profile[NET_DIM_CQ_PERIOD_NUM_MODES][NET_DIM_PARAMS_NUM_PROFILES] = {
+ NET_DIM_EQE_PROFILES,
+ NET_DIM_CQE_PROFILES,
};
-struct mlx5e_cq_moder mlx5e_am_get_profile(u8 cq_period_mode, int ix)
+struct net_dim_cq_moder net_dim_get_profile(u8 cq_period_mode, int ix)
{
- struct mlx5e_cq_moder cq_moder;
+ struct net_dim_cq_moder cq_moder;
cq_moder = profile[cq_period_mode][ix];
cq_moder.cq_period_mode = cq_period_mode;
return cq_moder;
}
-struct mlx5e_cq_moder mlx5e_am_get_def_profile(u8 rx_cq_period_mode)
+struct net_dim_cq_moder net_dim_get_def_profile(u8 rx_cq_period_mode)
{
int default_profile_ix;
- if (rx_cq_period_mode == MLX5_CQ_PERIOD_MODE_START_FROM_CQE)
- default_profile_ix = MLX5E_RX_AM_DEF_PROFILE_CQE;
- else /* MLX5_CQ_PERIOD_MODE_START_FROM_EQE */
- default_profile_ix = MLX5E_RX_AM_DEF_PROFILE_EQE;
+ if (rx_cq_period_mode == NET_DIM_CQ_PERIOD_MODE_START_FROM_CQE)
+ default_profile_ix = NET_DIM_DEF_PROFILE_CQE;
+ else /* NET_DIM_CQ_PERIOD_MODE_START_FROM_EQE */
+ default_profile_ix = NET_DIM_DEF_PROFILE_EQE;
- return mlx5e_am_get_profile(rx_cq_period_mode, default_profile_ix);
+ return net_dim_get_profile(rx_cq_period_mode, default_profile_ix);
}
-static bool mlx5e_am_on_top(struct mlx5e_rx_am *am)
+static bool net_dim_on_top(struct net_dim *dim)
{
- switch (am->tune_state) {
- case MLX5E_AM_PARKING_ON_TOP:
- case MLX5E_AM_PARKING_TIRED:
+ switch (dim->tune_state) {
+ case NET_DIM_PARKING_ON_TOP:
+ case NET_DIM_PARKING_TIRED:
return true;
- case MLX5E_AM_GOING_RIGHT:
- return (am->steps_left > 1) && (am->steps_right == 1);
- default: /* MLX5E_AM_GOING_LEFT */
- return (am->steps_right > 1) && (am->steps_left == 1);
+ case NET_DIM_GOING_RIGHT:
+ return (dim->steps_left > 1) && (dim->steps_right == 1);
+ default: /* NET_DIM_GOING_LEFT */
+ return (dim->steps_right > 1) && (dim->steps_left == 1);
}
}
-static void mlx5e_am_turn(struct mlx5e_rx_am *am)
+static void net_dim_turn(struct net_dim *dim)
{
- switch (am->tune_state) {
- case MLX5E_AM_PARKING_ON_TOP:
- case MLX5E_AM_PARKING_TIRED:
+ switch (dim->tune_state) {
+ case NET_DIM_PARKING_ON_TOP:
+ case NET_DIM_PARKING_TIRED:
break;
- case MLX5E_AM_GOING_RIGHT:
- am->tune_state = MLX5E_AM_GOING_LEFT;
- am->steps_left = 0;
+ case NET_DIM_GOING_RIGHT:
+ dim->tune_state = NET_DIM_GOING_LEFT;
+ dim->steps_left = 0;
break;
- case MLX5E_AM_GOING_LEFT:
- am->tune_state = MLX5E_AM_GOING_RIGHT;
- am->steps_right = 0;
+ case NET_DIM_GOING_LEFT:
+ dim->tune_state = NET_DIM_GOING_RIGHT;
+ dim->steps_right = 0;
break;
}
}
-static int mlx5e_am_step(struct mlx5e_rx_am *am)
+static int net_dim_step(struct net_dim *dim)
{
- if (am->tired == (MLX5E_PARAMS_AM_NUM_PROFILES * 2))
- return MLX5E_AM_TOO_TIRED;
+ if (dim->tired == (NET_DIM_PARAMS_NUM_PROFILES * 2))
+ return NET_DIM_TOO_TIRED;
- switch (am->tune_state) {
- case MLX5E_AM_PARKING_ON_TOP:
- case MLX5E_AM_PARKING_TIRED:
+ switch (dim->tune_state) {
+ case NET_DIM_PARKING_ON_TOP:
+ case NET_DIM_PARKING_TIRED:
break;
- case MLX5E_AM_GOING_RIGHT:
- if (am->profile_ix == (MLX5E_PARAMS_AM_NUM_PROFILES - 1))
- return MLX5E_AM_ON_EDGE;
- am->profile_ix++;
- am->steps_right++;
+ case NET_DIM_GOING_RIGHT:
+ if (dim->profile_ix == (NET_DIM_PARAMS_NUM_PROFILES - 1))
+ return NET_DIM_ON_EDGE;
+ dim->profile_ix++;
+ dim->steps_right++;
break;
- case MLX5E_AM_GOING_LEFT:
- if (am->profile_ix == 0)
- return MLX5E_AM_ON_EDGE;
- am->profile_ix--;
- am->steps_left++;
+ case NET_DIM_GOING_LEFT:
+ if (dim->profile_ix == 0)
+ return NET_DIM_ON_EDGE;
+ dim->profile_ix--;
+ dim->steps_left++;
break;
}
- am->tired++;
- return MLX5E_AM_STEPPED;
+ dim->tired++;
+ return NET_DIM_STEPPED;
}
-static void mlx5e_am_park_on_top(struct mlx5e_rx_am *am)
+static void net_dim_park_on_top(struct net_dim *dim)
{
- am->steps_right = 0;
- am->steps_left = 0;
- am->tired = 0;
- am->tune_state = MLX5E_AM_PARKING_ON_TOP;
+ dim->steps_right = 0;
+ dim->steps_left = 0;
+ dim->tired = 0;
+ dim->tune_state = NET_DIM_PARKING_ON_TOP;
}
-static void mlx5e_am_park_tired(struct mlx5e_rx_am *am)
+static void net_dim_park_tired(struct net_dim *dim)
{
- am->steps_right = 0;
- am->steps_left = 0;
- am->tune_state = MLX5E_AM_PARKING_TIRED;
+ dim->steps_right = 0;
+ dim->steps_left = 0;
+ dim->tune_state = NET_DIM_PARKING_TIRED;
}
-static void mlx5e_am_exit_parking(struct mlx5e_rx_am *am)
+static void net_dim_exit_parking(struct net_dim *dim)
{
- am->tune_state = am->profile_ix ? MLX5E_AM_GOING_LEFT :
- MLX5E_AM_GOING_RIGHT;
- mlx5e_am_step(am);
+ dim->tune_state = dim->profile_ix ? NET_DIM_GOING_LEFT :
+ NET_DIM_GOING_RIGHT;
+ net_dim_step(dim);
}
#define IS_SIGNIFICANT_DIFF(val, ref) \
(((100 * abs((val) - (ref))) / (ref)) > 10) /* more than 10% difference */
-static int mlx5e_am_stats_compare(struct mlx5e_rx_am_stats *curr,
- struct mlx5e_rx_am_stats *prev)
+static int net_dim_stats_compare(struct net_dim_stats *curr,
+ struct net_dim_stats *prev)
{
if (!prev->bpms)
- return curr->bpms ? MLX5E_AM_STATS_BETTER :
- MLX5E_AM_STATS_SAME;
+ return curr->bpms ? NET_DIM_STATS_BETTER :
+ NET_DIM_STATS_SAME;
if (IS_SIGNIFICANT_DIFF(curr->bpms, prev->bpms))
- return (curr->bpms > prev->bpms) ? MLX5E_AM_STATS_BETTER :
- MLX5E_AM_STATS_WORSE;
+ return (curr->bpms > prev->bpms) ? NET_DIM_STATS_BETTER :
+ NET_DIM_STATS_WORSE;
if (IS_SIGNIFICANT_DIFF(curr->ppms, prev->ppms))
- return (curr->ppms > prev->ppms) ? MLX5E_AM_STATS_BETTER :
- MLX5E_AM_STATS_WORSE;
+ return (curr->ppms > prev->ppms) ? NET_DIM_STATS_BETTER :
+ NET_DIM_STATS_WORSE;
if (IS_SIGNIFICANT_DIFF(curr->epms, prev->epms))
- return (curr->epms < prev->epms) ? MLX5E_AM_STATS_BETTER :
- MLX5E_AM_STATS_WORSE;
+ return (curr->epms < prev->epms) ? NET_DIM_STATS_BETTER :
+ NET_DIM_STATS_WORSE;
- return MLX5E_AM_STATS_SAME;
+ return NET_DIM_STATS_SAME;
}
-static bool mlx5e_am_decision(struct mlx5e_rx_am_stats *curr_stats,
- struct mlx5e_rx_am *am)
+static bool net_dim_decision(struct net_dim_stats *curr_stats,
+ struct net_dim *dim)
{
- int prev_state = am->tune_state;
- int prev_ix = am->profile_ix;
+ int prev_state = dim->tune_state;
+ int prev_ix = dim->profile_ix;
int stats_res;
int step_res;
- switch (am->tune_state) {
- case MLX5E_AM_PARKING_ON_TOP:
- stats_res = mlx5e_am_stats_compare(curr_stats, &am->prev_stats);
- if (stats_res != MLX5E_AM_STATS_SAME)
- mlx5e_am_exit_parking(am);
+ switch (dim->tune_state) {
+ case NET_DIM_PARKING_ON_TOP:
+ stats_res = net_dim_stats_compare(curr_stats, &dim->prev_stats);
+ if (stats_res != NET_DIM_STATS_SAME)
+ net_dim_exit_parking(dim);
break;
- case MLX5E_AM_PARKING_TIRED:
- am->tired--;
- if (!am->tired)
- mlx5e_am_exit_parking(am);
+ case NET_DIM_PARKING_TIRED:
+ dim->tired--;
+ if (!dim->tired)
+ net_dim_exit_parking(dim);
break;
- case MLX5E_AM_GOING_RIGHT:
- case MLX5E_AM_GOING_LEFT:
- stats_res = mlx5e_am_stats_compare(curr_stats, &am->prev_stats);
- if (stats_res != MLX5E_AM_STATS_BETTER)
- mlx5e_am_turn(am);
+ case NET_DIM_GOING_RIGHT:
+ case NET_DIM_GOING_LEFT:
+ stats_res = net_dim_stats_compare(curr_stats, &dim->prev_stats);
+ if (stats_res != NET_DIM_STATS_BETTER)
+ net_dim_turn(dim);
- if (mlx5e_am_on_top(am)) {
- mlx5e_am_park_on_top(am);
+ if (net_dim_on_top(dim)) {
+ net_dim_park_on_top(dim);
break;
}
- step_res = mlx5e_am_step(am);
+ step_res = net_dim_step(dim);
switch (step_res) {
- case MLX5E_AM_ON_EDGE:
- mlx5e_am_park_on_top(am);
+ case NET_DIM_ON_EDGE:
+ net_dim_park_on_top(dim);
break;
- case MLX5E_AM_TOO_TIRED:
- mlx5e_am_park_tired(am);
+ case NET_DIM_TOO_TIRED:
+ net_dim_park_tired(dim);
break;
}
break;
}
- if ((prev_state != MLX5E_AM_PARKING_ON_TOP) ||
- (am->tune_state != MLX5E_AM_PARKING_ON_TOP))
- am->prev_stats = *curr_stats;
+ if ((prev_state != NET_DIM_PARKING_ON_TOP) ||
+ (dim->tune_state != NET_DIM_PARKING_ON_TOP))
+ dim->prev_stats = *curr_stats;
- return am->profile_ix != prev_ix;
+ return dim->profile_ix != prev_ix;
}
-static void mlx5e_am_sample(u16 event_ctr,
- u64 packets,
- u64 bytes,
- struct mlx5e_rx_am_sample *s)
+static void net_dim_sample(u16 event_ctr,
+ u64 packets,
+ u64 bytes,
+ struct net_dim_sample *s)
{
s->time = ktime_get();
s->pkt_ctr = packets;
@@ -250,13 +250,13 @@ static void mlx5e_am_sample(u16 event_ctr,
s->event_ctr = event_ctr;
}
-#define MLX5E_AM_NEVENTS 64
+#define NET_DIM_NEVENTS 64
#define BITS_PER_TYPE(type) (sizeof(type) * BITS_PER_BYTE)
#define BIT_GAP(bits, end, start) ((((end) - (start)) + BIT_ULL(bits)) & (BIT_ULL(bits) - 1))
-static void mlx5e_am_calc_stats(struct mlx5e_rx_am_sample *start,
- struct mlx5e_rx_am_sample *end,
- struct mlx5e_rx_am_stats *curr_stats)
+static void net_dim_calc_stats(struct net_dim_sample *start,
+ struct net_dim_sample *end,
+ struct net_dim_stats *curr_stats)
{
/* u32 holds up to 71 minutes, should be enough */
u32 delta_us = ktime_us_delta(end->time, start->time);
@@ -269,39 +269,39 @@ static void mlx5e_am_calc_stats(struct mlx5e_rx_am_sample *start,
curr_stats->ppms = DIV_ROUND_UP(npkts * USEC_PER_MSEC, delta_us);
curr_stats->bpms = DIV_ROUND_UP(nbytes * USEC_PER_MSEC, delta_us);
- curr_stats->epms = DIV_ROUND_UP(MLX5E_AM_NEVENTS * USEC_PER_MSEC,
+ curr_stats->epms = DIV_ROUND_UP(NET_DIM_NEVENTS * USEC_PER_MSEC,
delta_us);
}
-void mlx5e_rx_am(struct mlx5e_rx_am *am,
- u16 event_ctr,
- u64 packets,
- u64 bytes)
+void net_dim(struct net_dim *dim,
+ u16 event_ctr,
+ u64 packets,
+ u64 bytes)
{
- struct mlx5e_rx_am_sample end_sample;
- struct mlx5e_rx_am_stats curr_stats;
+ struct net_dim_sample end_sample;
+ struct net_dim_stats curr_stats;
u16 nevents;
- switch (am->state) {
- case MLX5E_AM_MEASURE_IN_PROGRESS:
+ switch (dim->state) {
+ case NET_DIM_MEASURE_IN_PROGRESS:
nevents = BIT_GAP(BITS_PER_TYPE(u16), event_ctr,
- am->start_sample.event_ctr);
- if (nevents < MLX5E_AM_NEVENTS)
+ dim->start_sample.event_ctr);
+ if (nevents < NET_DIM_NEVENTS)
break;
- mlx5e_am_sample(event_ctr, packets, bytes, &end_sample);
- mlx5e_am_calc_stats(&am->start_sample, &end_sample,
+ net_dim_sample(event_ctr, packets, bytes, &end_sample);
+ net_dim_calc_stats(&dim->start_sample, &end_sample,
&curr_stats);
- if (mlx5e_am_decision(&curr_stats, am)) {
- am->state = MLX5E_AM_APPLY_NEW_PROFILE;
- schedule_work(&am->work);
+ if (net_dim_decision(&curr_stats, dim)) {
+ dim->state = NET_DIM_APPLY_NEW_PROFILE;
+ schedule_work(&dim->work);
break;
}
/* fall through */
- case MLX5E_AM_START_MEASURE:
- mlx5e_am_sample(event_ctr, packets, bytes, &am->start_sample);
- am->state = MLX5E_AM_MEASURE_IN_PROGRESS;
+ case NET_DIM_START_MEASURE:
+ net_dim_sample(event_ctr, packets, bytes, &dim->start_sample);
+ dim->state = NET_DIM_MEASURE_IN_PROGRESS;
break;
- case MLX5E_AM_APPLY_NEW_PROFILE:
+ case NET_DIM_APPLY_NEW_PROFILE:
break;
}
}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/net_dim.h b/drivers/net/ethernet/mellanox/mlx5/core/net_dim.h
index 5ce8e54..a775c12 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/net_dim.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/net_dim.h
@@ -31,32 +31,32 @@
* SOFTWARE.
*/
-#ifndef MLX5_AM_H
-#define MLX5_AM_H
+#ifndef NET_DIM_H
+#define NET_DIM_H
-struct mlx5e_cq_moder {
+struct net_dim_cq_moder {
u16 usec;
u16 pkts;
u8 cq_period_mode;
};
-struct mlx5e_rx_am_sample {
+struct net_dim_sample {
ktime_t time;
u32 pkt_ctr;
u32 byte_ctr;
u16 event_ctr;
};
-struct mlx5e_rx_am_stats {
+struct net_dim_stats {
int ppms; /* packets per msec */
int bpms; /* bytes per msec */
int epms; /* events per msec */
};
-struct mlx5e_rx_am { /* Adaptive Moderation */
+struct net_dim { /* Adaptive Moderation */
u8 state;
- struct mlx5e_rx_am_stats prev_stats;
- struct mlx5e_rx_am_sample start_sample;
+ struct net_dim_stats prev_stats;
+ struct net_dim_sample start_sample;
struct work_struct work;
u8 profile_ix;
u8 mode;
@@ -67,43 +67,42 @@ struct mlx5e_rx_am { /* Adaptive Moderation */
};
enum {
- MLX5_CQ_PERIOD_MODE_START_FROM_EQE = 0x0,
- MLX5_CQ_PERIOD_MODE_START_FROM_CQE = 0x1,
- MLX5_CQ_PERIOD_NUM_MODES
+ NET_DIM_CQ_PERIOD_MODE_START_FROM_EQE = 0x0,
+ NET_DIM_CQ_PERIOD_MODE_START_FROM_CQE = 0x1,
+ NET_DIM_CQ_PERIOD_NUM_MODES
};
/* Adaptive moderation logic */
enum {
- MLX5E_AM_START_MEASURE,
- MLX5E_AM_MEASURE_IN_PROGRESS,
- MLX5E_AM_APPLY_NEW_PROFILE,
+ NET_DIM_START_MEASURE,
+ NET_DIM_MEASURE_IN_PROGRESS,
+ NET_DIM_APPLY_NEW_PROFILE,
};
enum {
- MLX5E_AM_PARKING_ON_TOP,
- MLX5E_AM_PARKING_TIRED,
- MLX5E_AM_GOING_RIGHT,
- MLX5E_AM_GOING_LEFT,
+ NET_DIM_PARKING_ON_TOP,
+ NET_DIM_PARKING_TIRED,
+ NET_DIM_GOING_RIGHT,
+ NET_DIM_GOING_LEFT,
};
enum {
- MLX5E_AM_STATS_WORSE,
- MLX5E_AM_STATS_SAME,
- MLX5E_AM_STATS_BETTER,
+ NET_DIM_STATS_WORSE,
+ NET_DIM_STATS_SAME,
+ NET_DIM_STATS_BETTER,
};
enum {
- MLX5E_AM_STEPPED,
- MLX5E_AM_TOO_TIRED,
- MLX5E_AM_ON_EDGE,
+ NET_DIM_STEPPED,
+ NET_DIM_TOO_TIRED,
+ NET_DIM_ON_EDGE,
};
-void mlx5e_rx_am(struct mlx5e_rx_am *am,
- u16 event_ctr,
- u64 packets,
- u64 bytes);
-void mlx5e_rx_am_work(struct work_struct *work);
-struct mlx5e_cq_moder mlx5e_am_get_def_profile(u8 rx_cq_period_mode);
-struct mlx5e_cq_moder mlx5e_am_get_profile(u8 cq_period_mode, int ix);
+void net_dim(struct net_dim *dim,
+ u16 event_ctr,
+ u64 packets,
+ u64 bytes);
+struct net_dim_cq_moder net_dim_get_def_profile(u8 rx_cq_period_mode);
+struct net_dim_cq_moder net_dim_get_profile(u8 cq_period_mode, int ix);
-#endif /* MLX5_AM_H */
+#endif /* NET_DIM_H */
--
2.7.4
^ permalink raw reply related
* [PATCH net-next v2 07/10] net/mlx5e: Move dynamic interrupt coalescing code to include/linux
From: Andy Gospodarek @ 2018-01-05 22:58 UTC (permalink / raw)
To: netdev; +Cc: mchan, talgi, ogerlitz, Andy Gospodarek
In-Reply-To: <1515193094-9554-1-git-send-email-andy@greyhouse.net>
From: Andy Gospodarek <gospo@broadcom.com>
This move allows drivers to add private structure elements to track the
number of packets, bytes, and interrupts events per ring. A driver
also defines a workqueue handler to act on this collected data once per
poll and modify the coalescing parameters per ring.
Signed-off-by: Andy Gospodarek <gospo@broadcom.com>
Acked-by: Tal Gilboa <talgi@mellanox.com>
Acked-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/Makefile | 2 +-
drivers/net/ethernet/mellanox/mlx5/core/en.h | 2 +-
drivers/net/ethernet/mellanox/mlx5/core/en_dim.c | 1 +
drivers/net/ethernet/mellanox/mlx5/core/net_dim.c | 307 ------------------
drivers/net/ethernet/mellanox/mlx5/core/net_dim.h | 108 -------
include/linux/net_dim.h | 376 ++++++++++++++++++++++
6 files changed, 379 insertions(+), 417 deletions(-)
delete mode 100644 drivers/net/ethernet/mellanox/mlx5/core/net_dim.c
delete mode 100644 drivers/net/ethernet/mellanox/mlx5/core/net_dim.h
create mode 100644 include/linux/net_dim.h
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/Makefile b/drivers/net/ethernet/mellanox/mlx5/core/Makefile
index b46b6de2..c805769 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/Makefile
+++ b/drivers/net/ethernet/mellanox/mlx5/core/Makefile
@@ -15,7 +15,7 @@ mlx5_core-$(CONFIG_MLX5_FPGA) += fpga/cmd.o fpga/core.o fpga/conn.o fpga/sdk.o \
mlx5_core-$(CONFIG_MLX5_CORE_EN) += en_main.o en_common.o en_fs.o en_ethtool.o \
en_tx.o en_rx.o en_dim.o en_txrx.o en_stats.o vxlan.o \
- en_arfs.o en_fs_ethtool.o en_selftest.o net_dim.o
+ en_arfs.o en_fs_ethtool.o en_selftest.o
mlx5_core-$(CONFIG_MLX5_MPFS) += lib/mpfs.o
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h
index 732f275..cb9abc9 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h
@@ -46,10 +46,10 @@
#include <linux/mlx5/transobj.h>
#include <linux/rhashtable.h>
#include <net/switchdev.h>
+#include <linux/net_dim.h>
#include "wq.h"
#include "mlx5_core.h"
#include "en_stats.h"
-#include "net_dim.h"
#define MLX5_SET_CFG(p, f, v) MLX5_SET(create_flow_group_in, p, f, v)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_dim.c b/drivers/net/ethernet/mellanox/mlx5/core/en_dim.c
index f620325..2b89951 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_dim.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_dim.c
@@ -30,6 +30,7 @@
* SOFTWARE.
*/
+#include <linux/net_dim.h>
#include "en.h"
void mlx5e_rx_dim_work(struct work_struct *work)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/net_dim.c b/drivers/net/ethernet/mellanox/mlx5/core/net_dim.c
deleted file mode 100644
index 00b9ae3..0000000
--- a/drivers/net/ethernet/mellanox/mlx5/core/net_dim.c
+++ /dev/null
@@ -1,307 +0,0 @@
-/*
- * Copyright (c) 2016, Mellanox Technologies. All rights reserved.
- * Copyright (c) 2017, Broadcom Limiited. All rights reserved.
- *
- * This software is available to you under a choice of one of two
- * licenses. You may choose to be licensed under the terms of the GNU
- * General Public License (GPL) Version 2, available from the file
- * COPYING in the main directory of this source tree, or the
- * OpenIB.org BSD license below:
- *
- * Redistribution and use in source and binary forms, with or
- * without modification, are permitted provided that the following
- * conditions are met:
- *
- * - Redistributions of source code must retain the above
- * copyright notice, this list of conditions and the following
- * disclaimer.
- *
- * - Redistributions in binary form must reproduce the above
- * copyright notice, this list of conditions and the following
- * disclaimer in the documentation and/or other materials
- * provided with the distribution.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
- * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
- * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
- * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
- * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
- * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
- * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- */
-
-#include "en.h"
-
-#define NET_DIM_PARAMS_NUM_PROFILES 5
-/* Adaptive moderation profiles */
-#define NET_DIM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE 256
-#define NET_DIM_DEF_PROFILE_CQE 1
-#define NET_DIM_DEF_PROFILE_EQE 1
-
-/* All profiles sizes must be NET_PARAMS_DIM_NUM_PROFILES */
-#define NET_DIM_EQE_PROFILES { \
- {1, NET_DIM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE}, \
- {8, NET_DIM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE}, \
- {64, NET_DIM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE}, \
- {128, NET_DIM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE}, \
- {256, NET_DIM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE}, \
-}
-
-#define NET_DIM_CQE_PROFILES { \
- {2, 256}, \
- {8, 128}, \
- {16, 64}, \
- {32, 64}, \
- {64, 64} \
-}
-
-static const struct net_dim_cq_moder
-profile[NET_DIM_CQ_PERIOD_NUM_MODES][NET_DIM_PARAMS_NUM_PROFILES] = {
- NET_DIM_EQE_PROFILES,
- NET_DIM_CQE_PROFILES,
-};
-
-struct net_dim_cq_moder net_dim_get_profile(u8 cq_period_mode, int ix)
-{
- struct net_dim_cq_moder cq_moder;
-
- cq_moder = profile[cq_period_mode][ix];
- cq_moder.cq_period_mode = cq_period_mode;
- return cq_moder;
-}
-
-struct net_dim_cq_moder net_dim_get_def_profile(u8 rx_cq_period_mode)
-{
- int default_profile_ix;
-
- if (rx_cq_period_mode == NET_DIM_CQ_PERIOD_MODE_START_FROM_CQE)
- default_profile_ix = NET_DIM_DEF_PROFILE_CQE;
- else /* NET_DIM_CQ_PERIOD_MODE_START_FROM_EQE */
- default_profile_ix = NET_DIM_DEF_PROFILE_EQE;
-
- return net_dim_get_profile(rx_cq_period_mode, default_profile_ix);
-}
-
-static bool net_dim_on_top(struct net_dim *dim)
-{
- switch (dim->tune_state) {
- case NET_DIM_PARKING_ON_TOP:
- case NET_DIM_PARKING_TIRED:
- return true;
- case NET_DIM_GOING_RIGHT:
- return (dim->steps_left > 1) && (dim->steps_right == 1);
- default: /* NET_DIM_GOING_LEFT */
- return (dim->steps_right > 1) && (dim->steps_left == 1);
- }
-}
-
-static void net_dim_turn(struct net_dim *dim)
-{
- switch (dim->tune_state) {
- case NET_DIM_PARKING_ON_TOP:
- case NET_DIM_PARKING_TIRED:
- break;
- case NET_DIM_GOING_RIGHT:
- dim->tune_state = NET_DIM_GOING_LEFT;
- dim->steps_left = 0;
- break;
- case NET_DIM_GOING_LEFT:
- dim->tune_state = NET_DIM_GOING_RIGHT;
- dim->steps_right = 0;
- break;
- }
-}
-
-static int net_dim_step(struct net_dim *dim)
-{
- if (dim->tired == (NET_DIM_PARAMS_NUM_PROFILES * 2))
- return NET_DIM_TOO_TIRED;
-
- switch (dim->tune_state) {
- case NET_DIM_PARKING_ON_TOP:
- case NET_DIM_PARKING_TIRED:
- break;
- case NET_DIM_GOING_RIGHT:
- if (dim->profile_ix == (NET_DIM_PARAMS_NUM_PROFILES - 1))
- return NET_DIM_ON_EDGE;
- dim->profile_ix++;
- dim->steps_right++;
- break;
- case NET_DIM_GOING_LEFT:
- if (dim->profile_ix == 0)
- return NET_DIM_ON_EDGE;
- dim->profile_ix--;
- dim->steps_left++;
- break;
- }
-
- dim->tired++;
- return NET_DIM_STEPPED;
-}
-
-static void net_dim_park_on_top(struct net_dim *dim)
-{
- dim->steps_right = 0;
- dim->steps_left = 0;
- dim->tired = 0;
- dim->tune_state = NET_DIM_PARKING_ON_TOP;
-}
-
-static void net_dim_park_tired(struct net_dim *dim)
-{
- dim->steps_right = 0;
- dim->steps_left = 0;
- dim->tune_state = NET_DIM_PARKING_TIRED;
-}
-
-static void net_dim_exit_parking(struct net_dim *dim)
-{
- dim->tune_state = dim->profile_ix ? NET_DIM_GOING_LEFT :
- NET_DIM_GOING_RIGHT;
- net_dim_step(dim);
-}
-
-#define IS_SIGNIFICANT_DIFF(val, ref) \
- (((100 * abs((val) - (ref))) / (ref)) > 10) /* more than 10% difference */
-
-static int net_dim_stats_compare(struct net_dim_stats *curr,
- struct net_dim_stats *prev)
-{
- if (!prev->bpms)
- return curr->bpms ? NET_DIM_STATS_BETTER :
- NET_DIM_STATS_SAME;
-
- if (IS_SIGNIFICANT_DIFF(curr->bpms, prev->bpms))
- return (curr->bpms > prev->bpms) ? NET_DIM_STATS_BETTER :
- NET_DIM_STATS_WORSE;
-
- if (IS_SIGNIFICANT_DIFF(curr->ppms, prev->ppms))
- return (curr->ppms > prev->ppms) ? NET_DIM_STATS_BETTER :
- NET_DIM_STATS_WORSE;
-
- if (IS_SIGNIFICANT_DIFF(curr->epms, prev->epms))
- return (curr->epms < prev->epms) ? NET_DIM_STATS_BETTER :
- NET_DIM_STATS_WORSE;
-
- return NET_DIM_STATS_SAME;
-}
-
-static bool net_dim_decision(struct net_dim_stats *curr_stats,
- struct net_dim *dim)
-{
- int prev_state = dim->tune_state;
- int prev_ix = dim->profile_ix;
- int stats_res;
- int step_res;
-
- switch (dim->tune_state) {
- case NET_DIM_PARKING_ON_TOP:
- stats_res = net_dim_stats_compare(curr_stats, &dim->prev_stats);
- if (stats_res != NET_DIM_STATS_SAME)
- net_dim_exit_parking(dim);
- break;
-
- case NET_DIM_PARKING_TIRED:
- dim->tired--;
- if (!dim->tired)
- net_dim_exit_parking(dim);
- break;
-
- case NET_DIM_GOING_RIGHT:
- case NET_DIM_GOING_LEFT:
- stats_res = net_dim_stats_compare(curr_stats, &dim->prev_stats);
- if (stats_res != NET_DIM_STATS_BETTER)
- net_dim_turn(dim);
-
- if (net_dim_on_top(dim)) {
- net_dim_park_on_top(dim);
- break;
- }
-
- step_res = net_dim_step(dim);
- switch (step_res) {
- case NET_DIM_ON_EDGE:
- net_dim_park_on_top(dim);
- break;
- case NET_DIM_TOO_TIRED:
- net_dim_park_tired(dim);
- break;
- }
-
- break;
- }
-
- if ((prev_state != NET_DIM_PARKING_ON_TOP) ||
- (dim->tune_state != NET_DIM_PARKING_ON_TOP))
- dim->prev_stats = *curr_stats;
-
- return dim->profile_ix != prev_ix;
-}
-
-static void net_dim_sample(u16 event_ctr,
- u64 packets,
- u64 bytes,
- struct net_dim_sample *s)
-{
- s->time = ktime_get();
- s->pkt_ctr = packets;
- s->byte_ctr = bytes;
- s->event_ctr = event_ctr;
-}
-
-#define NET_DIM_NEVENTS 64
-#define BITS_PER_TYPE(type) (sizeof(type) * BITS_PER_BYTE)
-#define BIT_GAP(bits, end, start) ((((end) - (start)) + BIT_ULL(bits)) & (BIT_ULL(bits) - 1))
-
-static void net_dim_calc_stats(struct net_dim_sample *start,
- struct net_dim_sample *end,
- struct net_dim_stats *curr_stats)
-{
- /* u32 holds up to 71 minutes, should be enough */
- u32 delta_us = ktime_us_delta(end->time, start->time);
- u32 npkts = BIT_GAP(BITS_PER_TYPE(u32), end->pkt_ctr, start->pkt_ctr);
- u32 nbytes = BIT_GAP(BITS_PER_TYPE(u32), end->byte_ctr,
- start->byte_ctr);
-
- if (!delta_us)
- return;
-
- curr_stats->ppms = DIV_ROUND_UP(npkts * USEC_PER_MSEC, delta_us);
- curr_stats->bpms = DIV_ROUND_UP(nbytes * USEC_PER_MSEC, delta_us);
- curr_stats->epms = DIV_ROUND_UP(NET_DIM_NEVENTS * USEC_PER_MSEC,
- delta_us);
-}
-
-void net_dim(struct net_dim *dim,
- u16 event_ctr,
- u64 packets,
- u64 bytes)
-{
- struct net_dim_sample end_sample;
- struct net_dim_stats curr_stats;
- u16 nevents;
-
- switch (dim->state) {
- case NET_DIM_MEASURE_IN_PROGRESS:
- nevents = BIT_GAP(BITS_PER_TYPE(u16), event_ctr,
- dim->start_sample.event_ctr);
- if (nevents < NET_DIM_NEVENTS)
- break;
- net_dim_sample(event_ctr, packets, bytes, &end_sample);
- net_dim_calc_stats(&dim->start_sample, &end_sample,
- &curr_stats);
- if (net_dim_decision(&curr_stats, dim)) {
- dim->state = NET_DIM_APPLY_NEW_PROFILE;
- schedule_work(&dim->work);
- break;
- }
- /* fall through */
- case NET_DIM_START_MEASURE:
- net_dim_sample(event_ctr, packets, bytes, &dim->start_sample);
- dim->state = NET_DIM_MEASURE_IN_PROGRESS;
- break;
- case NET_DIM_APPLY_NEW_PROFILE:
- break;
- }
-}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/net_dim.h b/drivers/net/ethernet/mellanox/mlx5/core/net_dim.h
deleted file mode 100644
index a775c12..0000000
--- a/drivers/net/ethernet/mellanox/mlx5/core/net_dim.h
+++ /dev/null
@@ -1,108 +0,0 @@
-/*
- * Copyright (c) 2013-2015, Mellanox Technologies, Ltd. All rights reserved.
- * Copyright (c) 2017, Broadcom Limited
- *
- * This software is available to you under a choice of one of two
- * licenses. You may choose to be licensed under the terms of the GNU
- * General Public License (GPL) Version 2, available from the file
- * COPYING in the main directory of this source tree, or the
- * OpenIB.org BSD license below:
- *
- * Redistribution and use in source and binary forms, with or
- * without modification, are permitted provided that the following
- * conditions are met:
- *
- * - Redistributions of source code must retain the above
- * copyright notice, this list of conditions and the following
- * disclaimer.
- *
- * - Redistributions in binary form must reproduce the above
- * copyright notice, this list of conditions and the following
- * disclaimer in the documentation and/or other materials
- * provided with the distribution.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
- * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
- * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
- * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
- * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
- * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
- * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
-*/
-
-#ifndef NET_DIM_H
-#define NET_DIM_H
-
-struct net_dim_cq_moder {
- u16 usec;
- u16 pkts;
- u8 cq_period_mode;
-};
-
-struct net_dim_sample {
- ktime_t time;
- u32 pkt_ctr;
- u32 byte_ctr;
- u16 event_ctr;
-};
-
-struct net_dim_stats {
- int ppms; /* packets per msec */
- int bpms; /* bytes per msec */
- int epms; /* events per msec */
-};
-
-struct net_dim { /* Adaptive Moderation */
- u8 state;
- struct net_dim_stats prev_stats;
- struct net_dim_sample start_sample;
- struct work_struct work;
- u8 profile_ix;
- u8 mode;
- u8 tune_state;
- u8 steps_right;
- u8 steps_left;
- u8 tired;
-};
-
-enum {
- NET_DIM_CQ_PERIOD_MODE_START_FROM_EQE = 0x0,
- NET_DIM_CQ_PERIOD_MODE_START_FROM_CQE = 0x1,
- NET_DIM_CQ_PERIOD_NUM_MODES
-};
-
-/* Adaptive moderation logic */
-enum {
- NET_DIM_START_MEASURE,
- NET_DIM_MEASURE_IN_PROGRESS,
- NET_DIM_APPLY_NEW_PROFILE,
-};
-
-enum {
- NET_DIM_PARKING_ON_TOP,
- NET_DIM_PARKING_TIRED,
- NET_DIM_GOING_RIGHT,
- NET_DIM_GOING_LEFT,
-};
-
-enum {
- NET_DIM_STATS_WORSE,
- NET_DIM_STATS_SAME,
- NET_DIM_STATS_BETTER,
-};
-
-enum {
- NET_DIM_STEPPED,
- NET_DIM_TOO_TIRED,
- NET_DIM_ON_EDGE,
-};
-
-void net_dim(struct net_dim *dim,
- u16 event_ctr,
- u64 packets,
- u64 bytes);
-struct net_dim_cq_moder net_dim_get_def_profile(u8 rx_cq_period_mode);
-struct net_dim_cq_moder net_dim_get_profile(u8 cq_period_mode, int ix);
-
-#endif /* NET_DIM_H */
diff --git a/include/linux/net_dim.h b/include/linux/net_dim.h
new file mode 100644
index 0000000..bb99073
--- /dev/null
+++ b/include/linux/net_dim.h
@@ -0,0 +1,376 @@
+/*
+ * Copyright (c) 2013-2015, Mellanox Technologies, Ltd. All rights reserved.
+ * Copyright (c) 2017, Broadcom Limited
+ *
+ * This software is available to you under a choice of one of two
+ * licenses. You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ * Redistribution and use in source and binary forms, with or
+ * without modification, are permitted provided that the following
+ * conditions are met:
+ *
+ * - Redistributions of source code must retain the above
+ * copyright notice, this list of conditions and the following
+ * disclaimer.
+ *
+ * - Redistributions in binary form must reproduce the above
+ * copyright notice, this list of conditions and the following
+ * disclaimer in the documentation and/or other materials
+ * provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+*/
+
+#ifndef NET_DIM_H
+#define NET_DIM_H
+
+#include <linux/module.h>
+
+struct net_dim_cq_moder {
+ u16 usec;
+ u16 pkts;
+ u8 cq_period_mode;
+};
+
+struct net_dim_sample {
+ ktime_t time;
+ u32 pkt_ctr;
+ u32 byte_ctr;
+ u16 event_ctr;
+};
+
+struct net_dim_stats {
+ int ppms; /* packets per msec */
+ int bpms; /* bytes per msec */
+ int epms; /* events per msec */
+};
+
+struct net_dim { /* Adaptive Moderation */
+ u8 state;
+ struct net_dim_stats prev_stats;
+ struct net_dim_sample start_sample;
+ struct work_struct work;
+ u8 profile_ix;
+ u8 mode;
+ u8 tune_state;
+ u8 steps_right;
+ u8 steps_left;
+ u8 tired;
+};
+
+enum {
+ NET_DIM_CQ_PERIOD_MODE_START_FROM_EQE = 0x0,
+ NET_DIM_CQ_PERIOD_MODE_START_FROM_CQE = 0x1,
+ NET_DIM_CQ_PERIOD_NUM_MODES
+};
+
+/* Adaptive moderation logic */
+enum {
+ NET_DIM_START_MEASURE,
+ NET_DIM_MEASURE_IN_PROGRESS,
+ NET_DIM_APPLY_NEW_PROFILE,
+};
+
+enum {
+ NET_DIM_PARKING_ON_TOP,
+ NET_DIM_PARKING_TIRED,
+ NET_DIM_GOING_RIGHT,
+ NET_DIM_GOING_LEFT,
+};
+
+enum {
+ NET_DIM_STATS_WORSE,
+ NET_DIM_STATS_SAME,
+ NET_DIM_STATS_BETTER,
+};
+
+enum {
+ NET_DIM_STEPPED,
+ NET_DIM_TOO_TIRED,
+ NET_DIM_ON_EDGE,
+};
+
+#define NET_DIM_PARAMS_NUM_PROFILES 5
+/* Adaptive moderation profiles */
+#define NET_DIM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE 256
+#define NET_DIM_DEF_PROFILE_CQE 1
+#define NET_DIM_DEF_PROFILE_EQE 1
+
+/* All profiles sizes must be NET_PARAMS_DIM_NUM_PROFILES */
+#define NET_DIM_EQE_PROFILES { \
+ {1, NET_DIM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE}, \
+ {8, NET_DIM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE}, \
+ {64, NET_DIM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE}, \
+ {128, NET_DIM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE}, \
+ {256, NET_DIM_DEFAULT_RX_CQ_MODERATION_PKTS_FROM_EQE}, \
+}
+
+#define NET_DIM_CQE_PROFILES { \
+ {2, 256}, \
+ {8, 128}, \
+ {16, 64}, \
+ {32, 64}, \
+ {64, 64} \
+}
+
+static const struct net_dim_cq_moder
+profile[NET_DIM_CQ_PERIOD_NUM_MODES][NET_DIM_PARAMS_NUM_PROFILES] = {
+ NET_DIM_EQE_PROFILES,
+ NET_DIM_CQE_PROFILES,
+};
+
+static inline struct net_dim_cq_moder net_dim_get_profile(u8 cq_period_mode, int ix)
+{
+ struct net_dim_cq_moder cq_moder;
+
+ cq_moder = profile[cq_period_mode][ix];
+ cq_moder.cq_period_mode = cq_period_mode;
+ return cq_moder;
+}
+
+static inline struct net_dim_cq_moder net_dim_get_def_profile(u8 rx_cq_period_mode)
+{
+ int default_profile_ix;
+
+ if (rx_cq_period_mode == NET_DIM_CQ_PERIOD_MODE_START_FROM_CQE)
+ default_profile_ix = NET_DIM_DEF_PROFILE_CQE;
+ else /* NET_DIM_CQ_PERIOD_MODE_START_FROM_EQE */
+ default_profile_ix = NET_DIM_DEF_PROFILE_EQE;
+
+ return net_dim_get_profile(rx_cq_period_mode, default_profile_ix);
+}
+
+static inline bool net_dim_on_top(struct net_dim *dim)
+{
+ switch (dim->tune_state) {
+ case NET_DIM_PARKING_ON_TOP:
+ case NET_DIM_PARKING_TIRED:
+ return true;
+ case NET_DIM_GOING_RIGHT:
+ return (dim->steps_left > 1) && (dim->steps_right == 1);
+ default: /* NET_DIM_GOING_LEFT */
+ return (dim->steps_right > 1) && (dim->steps_left == 1);
+ }
+}
+
+static inline void net_dim_turn(struct net_dim *dim)
+{
+ switch (dim->tune_state) {
+ case NET_DIM_PARKING_ON_TOP:
+ case NET_DIM_PARKING_TIRED:
+ break;
+ case NET_DIM_GOING_RIGHT:
+ dim->tune_state = NET_DIM_GOING_LEFT;
+ dim->steps_left = 0;
+ break;
+ case NET_DIM_GOING_LEFT:
+ dim->tune_state = NET_DIM_GOING_RIGHT;
+ dim->steps_right = 0;
+ break;
+ }
+}
+
+static inline int net_dim_step(struct net_dim *dim)
+{
+ if (dim->tired == (NET_DIM_PARAMS_NUM_PROFILES * 2))
+ return NET_DIM_TOO_TIRED;
+
+ switch (dim->tune_state) {
+ case NET_DIM_PARKING_ON_TOP:
+ case NET_DIM_PARKING_TIRED:
+ break;
+ case NET_DIM_GOING_RIGHT:
+ if (dim->profile_ix == (NET_DIM_PARAMS_NUM_PROFILES - 1))
+ return NET_DIM_ON_EDGE;
+ dim->profile_ix++;
+ dim->steps_right++;
+ break;
+ case NET_DIM_GOING_LEFT:
+ if (dim->profile_ix == 0)
+ return NET_DIM_ON_EDGE;
+ dim->profile_ix--;
+ dim->steps_left++;
+ break;
+ }
+
+ dim->tired++;
+ return NET_DIM_STEPPED;
+}
+
+static inline void net_dim_park_on_top(struct net_dim *dim)
+{
+ dim->steps_right = 0;
+ dim->steps_left = 0;
+ dim->tired = 0;
+ dim->tune_state = NET_DIM_PARKING_ON_TOP;
+}
+
+static inline void net_dim_park_tired(struct net_dim *dim)
+{
+ dim->steps_right = 0;
+ dim->steps_left = 0;
+ dim->tune_state = NET_DIM_PARKING_TIRED;
+}
+
+static inline void net_dim_exit_parking(struct net_dim *dim)
+{
+ dim->tune_state = dim->profile_ix ? NET_DIM_GOING_LEFT :
+ NET_DIM_GOING_RIGHT;
+ net_dim_step(dim);
+}
+
+#define IS_SIGNIFICANT_DIFF(val, ref) \
+ (((100 * abs((val) - (ref))) / (ref)) > 10) /* more than 10% difference */
+
+static inline int net_dim_stats_compare(struct net_dim_stats *curr,
+ struct net_dim_stats *prev)
+{
+ if (!prev->bpms)
+ return curr->bpms ? NET_DIM_STATS_BETTER :
+ NET_DIM_STATS_SAME;
+
+ if (IS_SIGNIFICANT_DIFF(curr->bpms, prev->bpms))
+ return (curr->bpms > prev->bpms) ? NET_DIM_STATS_BETTER :
+ NET_DIM_STATS_WORSE;
+
+ if (IS_SIGNIFICANT_DIFF(curr->ppms, prev->ppms))
+ return (curr->ppms > prev->ppms) ? NET_DIM_STATS_BETTER :
+ NET_DIM_STATS_WORSE;
+
+ if (IS_SIGNIFICANT_DIFF(curr->epms, prev->epms))
+ return (curr->epms < prev->epms) ? NET_DIM_STATS_BETTER :
+ NET_DIM_STATS_WORSE;
+
+ return NET_DIM_STATS_SAME;
+}
+
+static inline bool net_dim_decision(struct net_dim_stats *curr_stats,
+ struct net_dim *dim)
+{
+ int prev_state = dim->tune_state;
+ int prev_ix = dim->profile_ix;
+ int stats_res;
+ int step_res;
+
+ switch (dim->tune_state) {
+ case NET_DIM_PARKING_ON_TOP:
+ stats_res = net_dim_stats_compare(curr_stats, &dim->prev_stats);
+ if (stats_res != NET_DIM_STATS_SAME)
+ net_dim_exit_parking(dim);
+ break;
+
+ case NET_DIM_PARKING_TIRED:
+ dim->tired--;
+ if (!dim->tired)
+ net_dim_exit_parking(dim);
+ break;
+
+ case NET_DIM_GOING_RIGHT:
+ case NET_DIM_GOING_LEFT:
+ stats_res = net_dim_stats_compare(curr_stats, &dim->prev_stats);
+ if (stats_res != NET_DIM_STATS_BETTER)
+ net_dim_turn(dim);
+
+ if (net_dim_on_top(dim)) {
+ net_dim_park_on_top(dim);
+ break;
+ }
+
+ step_res = net_dim_step(dim);
+ switch (step_res) {
+ case NET_DIM_ON_EDGE:
+ net_dim_park_on_top(dim);
+ break;
+ case NET_DIM_TOO_TIRED:
+ net_dim_park_tired(dim);
+ break;
+ }
+
+ break;
+ }
+
+ if ((prev_state != NET_DIM_PARKING_ON_TOP) ||
+ (dim->tune_state != NET_DIM_PARKING_ON_TOP))
+ dim->prev_stats = *curr_stats;
+
+ return dim->profile_ix != prev_ix;
+}
+
+static inline void net_dim_sample(u16 event_ctr,
+ u64 packets,
+ u64 bytes,
+ struct net_dim_sample *s)
+{
+ s->time = ktime_get();
+ s->pkt_ctr = packets;
+ s->byte_ctr = bytes;
+ s->event_ctr = event_ctr;
+}
+
+#define NET_DIM_NEVENTS 64
+#define BITS_PER_TYPE(type) (sizeof(type) * BITS_PER_BYTE)
+#define BIT_GAP(bits, end, start) ((((end) - (start)) + BIT_ULL(bits)) & (BIT_ULL(bits) - 1))
+
+static inline void net_dim_calc_stats(struct net_dim_sample *start,
+ struct net_dim_sample *end,
+ struct net_dim_stats *curr_stats)
+{
+ /* u32 holds up to 71 minutes, should be enough */
+ u32 delta_us = ktime_us_delta(end->time, start->time);
+ u32 npkts = BIT_GAP(BITS_PER_TYPE(u32), end->pkt_ctr, start->pkt_ctr);
+ u32 nbytes = BIT_GAP(BITS_PER_TYPE(u32), end->byte_ctr,
+ start->byte_ctr);
+
+ if (!delta_us)
+ return;
+
+ curr_stats->ppms = DIV_ROUND_UP(npkts * USEC_PER_MSEC, delta_us);
+ curr_stats->bpms = DIV_ROUND_UP(nbytes * USEC_PER_MSEC, delta_us);
+ curr_stats->epms = DIV_ROUND_UP(NET_DIM_NEVENTS * USEC_PER_MSEC,
+ delta_us);
+}
+
+static inline void net_dim(struct net_dim *dim,
+ u16 event_ctr,
+ u64 packets,
+ u64 bytes)
+{
+ struct net_dim_sample end_sample;
+ struct net_dim_stats curr_stats;
+ u16 nevents;
+
+ switch (dim->state) {
+ case NET_DIM_MEASURE_IN_PROGRESS:
+ nevents = BIT_GAP(BITS_PER_TYPE(u16), event_ctr,
+ dim->start_sample.event_ctr);
+ if (nevents < NET_DIM_NEVENTS)
+ break;
+ net_dim_sample(event_ctr, packets, bytes, &end_sample);
+ net_dim_calc_stats(&dim->start_sample, &end_sample,
+ &curr_stats);
+ if (net_dim_decision(&curr_stats, dim)) {
+ dim->state = NET_DIM_APPLY_NEW_PROFILE;
+ schedule_work(&dim->work);
+ break;
+ }
+ /* fall through */
+ case NET_DIM_START_MEASURE:
+ net_dim_sample(event_ctr, packets, bytes, &dim->start_sample);
+ dim->state = NET_DIM_MEASURE_IN_PROGRESS;
+ break;
+ case NET_DIM_APPLY_NEW_PROFILE:
+ break;
+ }
+}
+
+#endif /* NET_DIM_H */
--
2.7.4
^ permalink raw reply related
* [PATCH net-next v2 08/10] net/dim: use struct net_dim_sample as arg to net_dim
From: Andy Gospodarek @ 2018-01-05 22:58 UTC (permalink / raw)
To: netdev; +Cc: mchan, talgi, ogerlitz, Andy Gospodarek
In-Reply-To: <1515193094-9554-1-git-send-email-andy@greyhouse.net>
From: Andy Gospodarek <gospo@broadcom.com>
Simplify the arguments net_dim() by formatting them into a struct
net_dim_sample before calling the function.
Signed-off-by: Andy Gospodarek <gospo@broadcom.com>
Suggested-by: Tal Gilboa <talgi@mellanox.com>
Acked-by: Tal Gilboa <talgi@mellanox.com>
Acked-by: Saeed Mahameed <saeedm@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c | 13 ++++++++-----
include/linux/net_dim.h | 10 +++-------
2 files changed, 11 insertions(+), 12 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c
index dae77a9..f292bb3 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c
@@ -78,11 +78,14 @@ int mlx5e_napi_poll(struct napi_struct *napi, int budget)
for (i = 0; i < c->num_tc; i++)
mlx5e_cq_arm(&c->sq[i].cq);
- if (MLX5E_TEST_BIT(c->rq.state, MLX5E_RQ_STATE_AM))
- net_dim(&c->rq.dim,
- c->rq.cq.event_ctr,
- c->rq.stats.packets,
- c->rq.stats.bytes);
+ if (MLX5E_TEST_BIT(c->rq.state, MLX5E_RQ_STATE_AM)) {
+ struct net_dim_sample dim_sample;
+ net_dim_sample(c->rq.cq.event_ctr,
+ c->rq.stats.packets,
+ c->rq.stats.bytes,
+ &dim_sample);
+ net_dim(&c->rq.dim, dim_sample);
+ }
mlx5e_cq_arm(&c->rq.cq);
mlx5e_cq_arm(&c->icosq.cq);
diff --git a/include/linux/net_dim.h b/include/linux/net_dim.h
index bb99073..2cceefa 100644
--- a/include/linux/net_dim.h
+++ b/include/linux/net_dim.h
@@ -341,21 +341,18 @@ static inline void net_dim_calc_stats(struct net_dim_sample *start,
}
static inline void net_dim(struct net_dim *dim,
- u16 event_ctr,
- u64 packets,
- u64 bytes)
+ struct net_dim_sample end_sample)
{
- struct net_dim_sample end_sample;
struct net_dim_stats curr_stats;
u16 nevents;
switch (dim->state) {
case NET_DIM_MEASURE_IN_PROGRESS:
- nevents = BIT_GAP(BITS_PER_TYPE(u16), event_ctr,
+ nevents = BIT_GAP(BITS_PER_TYPE(u16),
+ end_sample.event_ctr,
dim->start_sample.event_ctr);
if (nevents < NET_DIM_NEVENTS)
break;
- net_dim_sample(event_ctr, packets, bytes, &end_sample);
net_dim_calc_stats(&dim->start_sample, &end_sample,
&curr_stats);
if (net_dim_decision(&curr_stats, dim)) {
@@ -365,7 +362,6 @@ static inline void net_dim(struct net_dim *dim,
}
/* fall through */
case NET_DIM_START_MEASURE:
- net_dim_sample(event_ctr, packets, bytes, &dim->start_sample);
dim->state = NET_DIM_MEASURE_IN_PROGRESS;
break;
case NET_DIM_APPLY_NEW_PROFILE:
--
2.7.4
^ permalink raw reply related
* [PATCH net-next v2 09/10] bnxt_en: add support for software dynamic interrupt moderation
From: Andy Gospodarek @ 2018-01-05 22:58 UTC (permalink / raw)
To: netdev; +Cc: mchan, talgi, ogerlitz, Andy Gospodarek
In-Reply-To: <1515193094-9554-1-git-send-email-andy@greyhouse.net>
From: Andy Gospodarek <gospo@broadcom.com>
This implements the changes needed for the bnxt_en driver to add support
for dynamic interrupt moderation per ring.
This does add additional counters in the receive path, but testing shows
that any additional instructions are offset by throughput gain when the
default configuration is for low latency.
Signed-off-by: Andy Gospodarek <gospo@broadcom.com>
Cc: Michael Chan <mchan@broadcom.com>
---
drivers/net/ethernet/broadcom/bnxt/Makefile | 2 +-
drivers/net/ethernet/broadcom/bnxt/bnxt.c | 49 +++++++++++++++++++++++
drivers/net/ethernet/broadcom/bnxt/bnxt.h | 34 +++++++++++-----
drivers/net/ethernet/broadcom/bnxt/bnxt_dim.c | 32 +++++++++++++++
drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c | 12 ++++++
5 files changed, 117 insertions(+), 12 deletions(-)
create mode 100644 drivers/net/ethernet/broadcom/bnxt/bnxt_dim.c
diff --git a/drivers/net/ethernet/broadcom/bnxt/Makefile b/drivers/net/ethernet/broadcom/bnxt/Makefile
index 59c8ec9..7c560d5 100644
--- a/drivers/net/ethernet/broadcom/bnxt/Makefile
+++ b/drivers/net/ethernet/broadcom/bnxt/Makefile
@@ -1,4 +1,4 @@
obj-$(CONFIG_BNXT) += bnxt_en.o
-bnxt_en-y := bnxt.o bnxt_sriov.o bnxt_ethtool.o bnxt_dcb.o bnxt_ulp.o bnxt_xdp.o bnxt_vfr.o bnxt_devlink.o
+bnxt_en-y := bnxt.o bnxt_sriov.o bnxt_ethtool.o bnxt_dcb.o bnxt_ulp.o bnxt_xdp.o bnxt_vfr.o bnxt_devlink.o bnxt_dim.o
bnxt_en-$(CONFIG_BNXT_FLOWER_OFFLOAD) += bnxt_tc.o
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
index 9efbdc6..b9d4c61 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
@@ -1645,6 +1645,8 @@ static int bnxt_rx_pkt(struct bnxt *bp, struct bnxt_napi *bnapi, u32 *raw_cons,
rxr->rx_next_cons = NEXT_RX(cons);
next_rx_no_prod:
+ cpr->rx_packets += 1;
+ cpr->rx_bytes += len;
*raw_cons = tmp_raw_cons;
return rc;
@@ -1802,6 +1804,7 @@ static irqreturn_t bnxt_msix(int irq, void *dev_instance)
struct bnxt_cp_ring_info *cpr = &bnapi->cp_ring;
u32 cons = RING_CMP(cpr->cp_raw_cons);
+ cpr->event_ctr++;
prefetch(&cpr->cp_desc_ring[CP_RING(cons)][CP_IDX(cons)]);
napi_schedule(&bnapi->napi);
return IRQ_HANDLED;
@@ -2025,6 +2028,14 @@ static int bnxt_poll(struct napi_struct *napi, int budget)
break;
}
}
+ if (bp->flags & BNXT_FLAG_DIM) {
+ struct net_dim_sample dim_sample;
+ net_dim_sample(cpr->event_ctr,
+ cpr->rx_packets,
+ cpr->rx_bytes,
+ &dim_sample);
+ net_dim(&cpr->dim, dim_sample);
+ }
mmiowb();
return work_done;
}
@@ -2610,6 +2621,8 @@ static void bnxt_init_cp_rings(struct bnxt *bp)
struct bnxt_ring_struct *ring = &cpr->cp_ring_struct;
ring->fw_ring_id = INVALID_HW_RING_ID;
+ cpr->rx_ring_coal.coal_ticks = bp->rx_coal.coal_ticks;
+ cpr->rx_ring_coal.coal_bufs = bp->rx_coal.coal_bufs;
}
}
@@ -4583,6 +4596,36 @@ static void bnxt_hwrm_set_coal_params(struct bnxt_coal *hw_coal,
req->flags = cpu_to_le16(flags);
}
+int bnxt_hwrm_set_ring_coal(struct bnxt *bp, struct bnxt_napi *bnapi)
+{
+ struct hwrm_ring_cmpl_ring_cfg_aggint_params_input req_rx = {0};
+ struct bnxt_cp_ring_info *cpr = &bnapi->cp_ring;
+ struct bnxt_coal coal;
+ unsigned int grp_idx;
+
+ /* Tick values in micro seconds.
+ * 1 coal_buf x bufs_per_record = 1 completion record.
+ */
+ memcpy(&coal, &bp->rx_coal, sizeof(struct bnxt_coal));
+
+ coal.coal_ticks = cpr->rx_ring_coal.coal_ticks;
+ coal.coal_bufs = cpr->rx_ring_coal.coal_bufs;
+
+ if (!bnapi->rx_ring)
+ return -ENODEV;
+
+ bnxt_hwrm_cmd_hdr_init(bp, &req_rx,
+ HWRM_RING_CMPL_RING_CFG_AGGINT_PARAMS, -1, -1);
+
+ bnxt_hwrm_set_coal_params(&coal, &req_rx);
+
+ grp_idx = bnapi->index;
+ req_rx.ring_id = cpu_to_le16(bp->grp_info[grp_idx].cp_fw_ring_id);
+
+ return hwrm_send_message(bp, &req_rx, sizeof(req_rx),
+ HWRM_CMD_TIMEOUT);
+}
+
int bnxt_hwrm_set_coal(struct bnxt *bp)
{
int i, rc = 0;
@@ -5705,7 +5748,13 @@ static void bnxt_enable_napi(struct bnxt *bp)
int i;
for (i = 0; i < bp->cp_nr_rings; i++) {
+ struct bnxt_cp_ring_info *cpr = &bp->bnapi[i]->cp_ring;
bp->bnapi[i]->in_reset = false;
+
+ if (!(bp->bnapi[i]->flags & BNXT_NAPI_FLAG_XDP)) {
+ INIT_WORK(&cpr->dim.work, bnxt_dim_work);
+ cpr->dim.mode = NET_DIM_CQ_PERIOD_MODE_START_FROM_EQE;
+ }
napi_enable(&bp->bnapi[i]->napi);
}
}
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.h b/drivers/net/ethernet/broadcom/bnxt/bnxt.h
index 5359a1f..a2d9efd 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.h
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.h
@@ -23,6 +23,7 @@
#include <net/devlink.h>
#include <net/dst_metadata.h>
#include <net/switchdev.h>
+#include <linux/net_dim.h>
struct tx_bd {
__le32 tx_bd_len_flags_type;
@@ -607,6 +608,17 @@ struct bnxt_tx_ring_info {
struct bnxt_ring_struct tx_ring_struct;
};
+struct bnxt_coal {
+ u16 coal_ticks;
+ u16 coal_ticks_irq;
+ u16 coal_bufs;
+ u16 coal_bufs_irq;
+ /* RING_IDLE enabled when coal ticks < idle_thresh */
+ u16 idle_thresh;
+ u8 bufs_per_record;
+ u8 budget;
+};
+
struct bnxt_tpa_info {
void *data;
u8 *data_ptr;
@@ -670,6 +682,13 @@ struct bnxt_cp_ring_info {
u32 cp_raw_cons;
void __iomem *cp_doorbell;
+ struct bnxt_coal rx_ring_coal;
+ u64 rx_packets;
+ u64 rx_bytes;
+ u64 event_ctr;
+
+ struct net_dim dim;
+
struct tx_cmp *cp_desc_ring[MAX_CP_PAGES];
dma_addr_t cp_desc_mapping[MAX_CP_PAGES];
@@ -944,17 +963,6 @@ struct bnxt_test_info {
#define BNXT_CAG_REG_LEGACY_INT_STATUS 0x4014
#define BNXT_CAG_REG_BASE 0x300000
-struct bnxt_coal {
- u16 coal_ticks;
- u16 coal_ticks_irq;
- u16 coal_bufs;
- u16 coal_bufs_irq;
- /* RING_IDLE enabled when coal ticks < idle_thresh */
- u16 idle_thresh;
- u8 bufs_per_record;
- u8 budget;
-};
-
struct bnxt_tc_flow_stats {
u64 packets;
u64 bytes;
@@ -1126,6 +1134,7 @@ struct bnxt {
#define BNXT_FLAG_DOUBLE_DB 0x400000
#define BNXT_FLAG_FW_DCBX_AGENT 0x800000
#define BNXT_FLAG_CHIP_NITRO_A0 0x1000000
+ #define BNXT_FLAG_DIM 0x2000000
#define BNXT_FLAG_ALL_CONFIG_FEATS (BNXT_FLAG_TPA | \
BNXT_FLAG_RFS | \
@@ -1423,4 +1432,7 @@ int bnxt_setup_mq_tc(struct net_device *dev, u8 tc);
int bnxt_get_max_rings(struct bnxt *, int *, int *, bool);
void bnxt_restore_pf_fw_resources(struct bnxt *bp);
int bnxt_port_attr_get(struct bnxt *bp, struct switchdev_attr *attr);
+void bnxt_dim_work(struct work_struct *work);
+int bnxt_hwrm_set_ring_coal(struct bnxt *bp, struct bnxt_napi *bnapi);
+
#endif
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_dim.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_dim.c
new file mode 100644
index 0000000..156e025
--- /dev/null
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_dim.c
@@ -0,0 +1,32 @@
+/*
+ * Copyright (c) 2017 Broadcom Limited
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation.
+ */
+
+#include <linux/net_dim.h>
+#include "bnxt_hsi.h"
+#include "bnxt.h"
+
+void bnxt_dim_work(struct work_struct *work)
+{
+ struct net_dim *dim = container_of(work, struct net_dim,
+ work);
+ struct bnxt_cp_ring_info *cpr = container_of(dim,
+ struct bnxt_cp_ring_info,
+ dim);
+ struct bnxt_napi *bnapi = container_of(cpr,
+ struct bnxt_napi,
+ cp_ring);
+ struct net_dim_cq_moder cur_profile = net_dim_get_profile(dim->mode,
+ dim->profile_ix);
+
+ cpr->rx_ring_coal.coal_ticks = cur_profile.usec;
+ cpr->rx_ring_coal.coal_bufs = cur_profile.pkts;
+
+ bnxt_hwrm_set_ring_coal(bnapi->bp, bnapi);
+ dim->state = NET_DIM_START_MEASURE;
+}
+
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c
index fe7599f..632e03b 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c
@@ -49,6 +49,8 @@ static int bnxt_get_coalesce(struct net_device *dev,
memset(coal, 0, sizeof(*coal));
+ coal->use_adaptive_rx_coalesce = bp->flags & BNXT_FLAG_DIM;
+
hw_coal = &bp->rx_coal;
mult = hw_coal->bufs_per_record;
coal->rx_coalesce_usecs = hw_coal->coal_ticks;
@@ -77,6 +79,15 @@ static int bnxt_set_coalesce(struct net_device *dev,
int rc = 0;
u16 mult;
+ if (coal->use_adaptive_rx_coalesce)
+ bp->flags |= BNXT_FLAG_DIM;
+ else {
+ if (bp->flags & BNXT_FLAG_DIM) {
+ bp->flags &= ~(BNXT_FLAG_DIM);
+ goto reset_coalesce;
+ }
+ }
+
hw_coal = &bp->rx_coal;
mult = hw_coal->bufs_per_record;
hw_coal->coal_ticks = coal->rx_coalesce_usecs;
@@ -104,6 +115,7 @@ static int bnxt_set_coalesce(struct net_device *dev,
update_stats = true;
}
+reset_coalesce:
if (netif_running(dev)) {
if (update_stats) {
rc = bnxt_close_nic(bp, true, false);
--
2.7.4
^ permalink raw reply related
* [PATCH net-next v2 10/10] MAINTAINERS: add entry for Dynamic Interrupt Moderation
From: Andy Gospodarek @ 2018-01-05 22:58 UTC (permalink / raw)
To: netdev; +Cc: mchan, talgi, ogerlitz, Andy Gospodarek
In-Reply-To: <1515193094-9554-1-git-send-email-andy@greyhouse.net>
From: Andy Gospodarek <gospo@broadcom.com>
Signed-off-by: Andy Gospodarek <gospo@broadcom.com>
Signed-off-by: Tal Gilboa <talgi@mellanox.com>
Acked-by: Saeed Mahameed <saeedm@mellanox.com>
---
MAINTAINERS | 5 +++++
1 file changed, 5 insertions(+)
diff --git a/MAINTAINERS b/MAINTAINERS
index 753799d..178239dc 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -4944,6 +4944,11 @@ S: Maintained
F: lib/dynamic_debug.c
F: include/linux/dynamic_debug.h
+DYNAMIC INTERRUPT MODERATION
+M: Tal Gilboa <talgi@mellanox.com>
+S: Maintained
+F: include/linux/net_dim.h
+
DZ DECSTATION DZ11 SERIAL DRIVER
M: "Maciej W. Rozycki" <macro@linux-mips.org>
S: Maintained
--
2.7.4
^ permalink raw reply related
* [PATCH bpf] selftests/bpf: fix test_align
From: Alexei Starovoitov @ 2018-01-05 23:02 UTC (permalink / raw)
To: David S . Miller; +Cc: Daniel Borkmann, netdev
since commit 82abbf8d2fc4 the verifier rejects the bit-wise
arithmetic on pointers earlier.
The test 'dubious pointer arithmetic' now has less output to match on.
Adjust it.
Fixes: 82abbf8d2fc4 ("bpf: do not allow root to mangle valid pointers")
Reported-by: kernel test robot <xiaolong.ye@intel.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
---
tools/testing/selftests/bpf/test_align.c | 22 +---------------------
1 file changed, 1 insertion(+), 21 deletions(-)
diff --git a/tools/testing/selftests/bpf/test_align.c b/tools/testing/selftests/bpf/test_align.c
index 8591c89c0828..471bbbdb94db 100644
--- a/tools/testing/selftests/bpf/test_align.c
+++ b/tools/testing/selftests/bpf/test_align.c
@@ -474,27 +474,7 @@ static struct bpf_align_test tests[] = {
.result = REJECT,
.matches = {
{4, "R5=pkt(id=0,off=0,r=0,imm=0)"},
- /* ptr & 0x40 == either 0 or 0x40 */
- {5, "R5=inv(id=0,umax_value=64,var_off=(0x0; 0x40))"},
- /* ptr << 2 == unknown, (4n) */
- {7, "R5=inv(id=0,smax_value=9223372036854775804,umax_value=18446744073709551612,var_off=(0x0; 0xfffffffffffffffc))"},
- /* (4n) + 14 == (4n+2). We blow our bounds, because
- * the add could overflow.
- */
- {8, "R5=inv(id=0,var_off=(0x2; 0xfffffffffffffffc))"},
- /* Checked s>=0 */
- {10, "R5=inv(id=0,umin_value=2,umax_value=9223372036854775806,var_off=(0x2; 0x7ffffffffffffffc))"},
- /* packet pointer + nonnegative (4n+2) */
- {12, "R6=pkt(id=1,off=0,r=0,umin_value=2,umax_value=9223372036854775806,var_off=(0x2; 0x7ffffffffffffffc))"},
- {14, "R4=pkt(id=1,off=4,r=0,umin_value=2,umax_value=9223372036854775806,var_off=(0x2; 0x7ffffffffffffffc))"},
- /* NET_IP_ALIGN + (4n+2) == (4n), alignment is fine.
- * We checked the bounds, but it might have been able
- * to overflow if the packet pointer started in the
- * upper half of the address space.
- * So we did not get a 'range' on R6, and the access
- * attempt will fail.
- */
- {16, "R6=pkt(id=1,off=0,r=0,umin_value=2,umax_value=9223372036854775806,var_off=(0x2; 0x7ffffffffffffffc))"},
+ /* R5 bitwise operator &= on pointer prohibited */
}
},
{
--
2.9.5
^ permalink raw reply related
* [patch net-next v6 00/11] net: sched: allow qdiscs to share filter block instances
From: Jiri Pirko @ 2018-01-05 23:09 UTC (permalink / raw)
To: netdev
Cc: davem, jhs, xiyou.wangcong, mlxsw, andrew, vivien.didelot,
f.fainelli, michael.chan, ganeshgr, saeedm, matanb, leonro,
idosch, jakub.kicinski, simon.horman, pieter.jansenvanvuuren,
john.hurley, alexander.h.duyck, ogerlitz, john.fastabend, daniel,
dsahern
From: Jiri Pirko <jiri@mellanox.com>
Currently the filters added to qdiscs are independent. So for example if you
have 2 netdevices and you create ingress qdisc on both and you want to add
identical filter rules both, you need to add them twice. This patchset
makes this easier and mainly saves resources allowing to share all filters
within a qdisc - I call it a "filter block". Also this helps to save
resources when we do offload to hw for example to expensive TCAM.
So back to the example. First, we create 2 qdiscs. Both will share
block number 22. "22" is just an identification. If we don't pass any
block number, a new one will be generated by kernel:
$ tc qdisc add dev ens7 ingress block 22
^^^^^^^^
$ tc qdisc add dev ens8 ingress block 22
^^^^^^^^
Now if we list the qdiscs, we will see the block index in the output:
$ tc qdisc
qdisc ingress ffff: dev ens7 parent ffff:fff1 block 22
qdisc ingress ffff: dev ens8 parent ffff:fff1 block 22
To make is more visual, the situation looks like this:
ens7 ingress qdisc ens7 ingress qdisc
| |
| |
+----------> block 22 <----------+
Unlimited number of qdiscs may share the same block.
Now we can add filter using the block index:
$ tc filter add block 22 protocol ip pref 25 flower dst_ip 192.168.0.0/16 action drop
Note we cannot use the qdisc for filter manipulations for shared blocks:
$ tc filter add dev ens8 ingress protocol ip pref 1 flower dst_ip 192.168.100.2 action drop
Error: Cannot work with shared block, please use block index.
We will see the same output if we list filters for ingress qdisc of
ens7 and ens8, also for the block 22:
$ tc filter show block 22
filter block 22 protocol ip pref 25 flower chain 0
filter block 22 protocol ip pref 25 flower chain 0 handle 0x1
...
$ tc filter show dev ens7 ingress
filter block 22 protocol ip pref 25 flower chain 0
filter block 22 protocol ip pref 25 flower chain 0 handle 0x1
...
$ tc filter show dev ens8 ingress
filter block 22 protocol ip pref 25 flower chain 0
filter block 22 protocol ip pref 25 flower chain 0 handle 0x1
...
---
v5->v6:
- added patch 6 that introduces block handle
v4->v5:
- patch 5:
- add tracking of binding of devs that are unable to offload and check
that before block cbs call.
v3->v4:
- patch 1:
- rebased on top of the current net-next
- added some extack strings
- patch 3:
- rebased on top of the current net-next
- patch 5:
- propagate netdev_ops->ndo_setup_tc error up to tcf_block_offload_bind
caller
- patch 7:
- rebased on top of the current net-next
v2->v3:
- removed original patch 1, removing tp->q cls_bpf dependency. Fixed by
Jakub in the meantime.
- patch 1:
- rebased on top of the current net-next
- patch 5:
- new patch
- patch 8:
- removed "p_" prefix from block index function args
- patch 10:
- add tc offload feature handling
Jiri Pirko (11):
net: sched: introduce support for multiple filter chain pointers
registration
net: sched: avoid usage of tp->q in tcf_classify
net: sched: introduce block mechanism to handle netif_keep_dst calls
net: sched: remove classid and q fields from tcf_proto
net: sched: keep track of offloaded filters and check tc offload
feature
net: sched: use block index as a handle instead of qdisc when block is
shared
net: sched: allow ingress and clsact qdiscs to share filter blocks
mlxsw: spectrum_acl: Reshuffle code around
mlxsw_sp_acl_ruleset_create/destroy
mlxsw: spectrum_acl: Don't store netdev and ingress for ruleset unbind
mlxsw: spectrum_acl: Implement TC block sharing
mlxsw: spectrum_acl: Pass mlxsw_sp_port down to ruleset bind/unbind
ops
drivers/net/ethernet/mellanox/mlxsw/spectrum.c | 182 ++++++-
drivers/net/ethernet/mellanox/mlxsw/spectrum.h | 44 +-
drivers/net/ethernet/mellanox/mlxsw/spectrum_acl.c | 302 ++++++++---
.../ethernet/mellanox/mlxsw/spectrum_acl_tcam.c | 44 +-
.../net/ethernet/mellanox/mlxsw/spectrum_flower.c | 41 +-
include/net/pkt_cls.h | 9 +
include/net/sch_generic.h | 27 +-
include/uapi/linux/pkt_sched.h | 11 +
net/sched/cls_api.c | 604 ++++++++++++++++-----
net/sched/cls_bpf.c | 9 +-
net/sched/cls_flow.c | 2 +-
net/sched/cls_flower.c | 3 +-
net/sched/cls_matchall.c | 3 +-
net/sched/cls_route.c | 2 +-
net/sched/cls_u32.c | 13 +-
net/sched/sch_ingress.c | 89 ++-
16 files changed, 1079 insertions(+), 306 deletions(-)
--
2.9.5
^ permalink raw reply
* [patch net-next v6 01/11] net: sched: introduce support for multiple filter chain pointers registration
From: Jiri Pirko @ 2018-01-05 23:09 UTC (permalink / raw)
To: netdev
Cc: davem, jhs, xiyou.wangcong, mlxsw, andrew, vivien.didelot,
f.fainelli, michael.chan, ganeshgr, saeedm, matanb, leonro,
idosch, jakub.kicinski, simon.horman, pieter.jansenvanvuuren,
john.hurley, alexander.h.duyck, ogerlitz, john.fastabend, daniel,
dsahern
In-Reply-To: <20180105230929.5645-1-jiri@resnulli.us>
From: Jiri Pirko <jiri@mellanox.com>
So far, there was possible only to register a single filter chain
pointer to block->chain[0]. However, when the blocks will get shareable,
we need to allow multiple filter chain pointers registration.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
v3->v4:
- rebased on top of the current net-next
- added some extack strings
v2->v3:
- rebased on top of the current net-next
---
include/net/pkt_cls.h | 3 +
include/net/sch_generic.h | 5 +-
net/sched/cls_api.c | 236 +++++++++++++++++++++++++++++++++++++++++-----
3 files changed, 216 insertions(+), 28 deletions(-)
diff --git a/include/net/pkt_cls.h b/include/net/pkt_cls.h
index 5cd3cf5..95c90da 100644
--- a/include/net/pkt_cls.h
+++ b/include/net/pkt_cls.h
@@ -29,6 +29,8 @@ struct tcf_block_ext_info {
enum tcf_block_binder_type binder_type;
tcf_chain_head_change_t *chain_head_change;
void *chain_head_change_priv;
+ bool shareable;
+ u32 block_index;
};
struct tcf_block_cb;
@@ -50,6 +52,7 @@ void tcf_block_put_ext(struct tcf_block *block, struct Qdisc *q,
static inline struct Qdisc *tcf_block_q(struct tcf_block *block)
{
+ WARN_ON(block->refcnt != 1);
return block->q;
}
diff --git a/include/net/sch_generic.h b/include/net/sch_generic.h
index ac029d5..5cc4d71 100644
--- a/include/net/sch_generic.h
+++ b/include/net/sch_generic.h
@@ -275,8 +275,7 @@ typedef void tcf_chain_head_change_t(struct tcf_proto *tp_head, void *priv);
struct tcf_chain {
struct tcf_proto __rcu *filter_chain;
- tcf_chain_head_change_t *chain_head_change;
- void *chain_head_change_priv;
+ struct list_head filter_chain_list;
struct list_head list;
struct tcf_block *block;
u32 index; /* chain index */
@@ -285,6 +284,8 @@ struct tcf_chain {
struct tcf_block {
struct list_head chain_list;
+ u32 index; /* block index for shared blocks */
+ unsigned int refcnt;
struct net *net;
struct Qdisc *q;
struct list_head cb_list;
diff --git a/net/sched/cls_api.c b/net/sched/cls_api.c
index 6708b69..2dd584a 100644
--- a/net/sched/cls_api.c
+++ b/net/sched/cls_api.c
@@ -24,6 +24,7 @@
#include <linux/init.h>
#include <linux/kmod.h>
#include <linux/slab.h>
+#include <linux/idr.h>
#include <net/net_namespace.h>
#include <net/sock.h>
#include <net/netlink.h>
@@ -179,6 +180,12 @@ static void tcf_proto_destroy(struct tcf_proto *tp)
kfree_rcu(tp, rcu);
}
+struct tcf_filter_chain_list_item {
+ struct list_head list;
+ tcf_chain_head_change_t *chain_head_change;
+ void *chain_head_change_priv;
+};
+
static struct tcf_chain *tcf_chain_create(struct tcf_block *block,
u32 chain_index)
{
@@ -187,6 +194,7 @@ static struct tcf_chain *tcf_chain_create(struct tcf_block *block,
chain = kzalloc(sizeof(*chain), GFP_KERNEL);
if (!chain)
return NULL;
+ INIT_LIST_HEAD(&chain->filter_chain_list);
list_add_tail(&chain->list, &block->chain_list);
chain->block = block;
chain->index = chain_index;
@@ -194,12 +202,19 @@ static struct tcf_chain *tcf_chain_create(struct tcf_block *block,
return chain;
}
+static void tcf_chain_head_change_item(struct tcf_filter_chain_list_item *item,
+ struct tcf_proto *tp_head)
+{
+ if (item->chain_head_change)
+ item->chain_head_change(tp_head, item->chain_head_change_priv);
+}
static void tcf_chain_head_change(struct tcf_chain *chain,
struct tcf_proto *tp_head)
{
- if (chain->chain_head_change)
- chain->chain_head_change(tp_head,
- chain->chain_head_change_priv);
+ struct tcf_filter_chain_list_item *item;
+
+ list_for_each_entry(item, &chain->filter_chain_list, list)
+ tcf_chain_head_change_item(item, tp_head);
}
static void tcf_chain_flush(struct tcf_chain *chain)
@@ -280,17 +295,91 @@ static void tcf_block_offload_unbind(struct tcf_block *block, struct Qdisc *q,
tcf_block_offload_cmd(block, q, ei, TC_BLOCK_UNBIND);
}
-int tcf_block_get_ext(struct tcf_block **p_block, struct Qdisc *q,
- struct tcf_block_ext_info *ei,
- struct netlink_ext_ack *extack)
+static int
+tcf_chain_head_change_cb_add(struct tcf_chain *chain,
+ struct tcf_block_ext_info *ei,
+ struct netlink_ext_ack *extack)
{
- struct tcf_block *block = kzalloc(sizeof(*block), GFP_KERNEL);
+ struct tcf_filter_chain_list_item *item;
+
+ item = kmalloc(sizeof(*item), GFP_KERNEL);
+ if (!item) {
+ NL_SET_ERR_MSG(extack, "Memory allocation for head change callback item failed");
+ return -ENOMEM;
+ }
+ item->chain_head_change = ei->chain_head_change;
+ item->chain_head_change_priv = ei->chain_head_change_priv;
+ if (chain->filter_chain)
+ tcf_chain_head_change_item(item, chain->filter_chain);
+ list_add(&item->list, &chain->filter_chain_list);
+ return 0;
+}
+
+static void
+tcf_chain_head_change_cb_del(struct tcf_chain *chain,
+ struct tcf_block_ext_info *ei)
+{
+ struct tcf_filter_chain_list_item *item;
+
+ list_for_each_entry(item, &chain->filter_chain_list, list) {
+ if ((!ei->chain_head_change && !ei->chain_head_change_priv) ||
+ (item->chain_head_change == ei->chain_head_change &&
+ item->chain_head_change_priv == ei->chain_head_change_priv)) {
+ tcf_chain_head_change_item(item, NULL);
+ list_del(&item->list);
+ kfree(item);
+ return;
+ }
+ }
+ WARN_ON(1);
+}
+
+struct tcf_net {
+ struct idr idr;
+};
+
+static unsigned int tcf_net_id;
+
+static int tcf_block_insert(struct tcf_block *block, struct net *net,
+ u32 block_index, struct netlink_ext_ack *extack)
+{
+ struct tcf_net *tn = net_generic(net, tcf_net_id);
+ int idr_start;
+ int idr_end;
+ int index;
+
+ if (block_index >= INT_MAX) {
+ NL_SET_ERR_MSG(extack, "Invalid block index value (>= INT_MAX)");
+ return -EINVAL;
+ }
+ idr_start = block_index ? block_index : 1;
+ idr_end = block_index ? block_index + 1 : INT_MAX;
+
+ index = idr_alloc(&tn->idr, block, idr_start, idr_end, GFP_KERNEL);
+ if (index < 0)
+ return index;
+ block->index = index;
+ return 0;
+}
+
+static void tcf_block_remove(struct tcf_block *block, struct net *net)
+{
+ struct tcf_net *tn = net_generic(net, tcf_net_id);
+
+ idr_remove(&tn->idr, block->index);
+}
+
+static struct tcf_block *tcf_block_create(struct net *net, struct Qdisc *q,
+ struct netlink_ext_ack *extack)
+{
+ struct tcf_block *block;
struct tcf_chain *chain;
int err;
+ block = kzalloc(sizeof(*block), GFP_KERNEL);
if (!block) {
NL_SET_ERR_MSG(extack, "Memory allocation for block failed");
- return -ENOMEM;
+ return ERR_PTR(-ENOMEM);
}
INIT_LIST_HEAD(&block->chain_list);
INIT_LIST_HEAD(&block->cb_list);
@@ -302,17 +391,75 @@ int tcf_block_get_ext(struct tcf_block **p_block, struct Qdisc *q,
err = -ENOMEM;
goto err_chain_create;
}
- WARN_ON(!ei->chain_head_change);
- chain->chain_head_change = ei->chain_head_change;
- chain->chain_head_change_priv = ei->chain_head_change_priv;
block->net = qdisc_net(q);
+ block->refcnt = 1;
+ block->net = net;
block->q = q;
+ return block;
+
+err_chain_create:
+ kfree(block);
+ return ERR_PTR(err);
+}
+
+static struct tcf_block *tcf_block_lookup(struct net *net, u32 block_index)
+{
+ struct tcf_net *tn = net_generic(net, tcf_net_id);
+
+ return idr_find(&tn->idr, block_index);
+}
+
+static struct tcf_chain *tcf_block_chain_zero(struct tcf_block *block)
+{
+ return list_first_entry(&block->chain_list, struct tcf_chain, list);
+}
+
+int tcf_block_get_ext(struct tcf_block **p_block, struct Qdisc *q,
+ struct tcf_block_ext_info *ei,
+ struct netlink_ext_ack *extack)
+{
+ struct net *net = qdisc_net(q);
+ struct tcf_block *block = NULL;
+ bool created = false;
+ int err;
+
+ if (ei->shareable) {
+ block = tcf_block_lookup(net, ei->block_index);
+ if (block)
+ block->refcnt++;
+ }
+
+ if (!block) {
+ block = tcf_block_create(net, q, extack);
+ if (IS_ERR(block))
+ return PTR_ERR(block);
+ created = true;
+ if (ei->shareable) {
+ err = tcf_block_insert(block, net,
+ ei->block_index, extack);
+ if (err)
+ goto err_block_insert;
+ }
+ }
+
+ err = tcf_chain_head_change_cb_add(tcf_block_chain_zero(block),
+ ei, extack);
+ if (err)
+ goto err_chain_head_change_cb_add;
+
tcf_block_offload_bind(block, q, ei);
*p_block = block;
return 0;
-err_chain_create:
- kfree(block);
+err_chain_head_change_cb_add:
+ if (created) {
+ if (ei->shareable)
+ tcf_block_remove(block, net);
+err_block_insert:
+ kfree(block);
+ } else {
+ block->refcnt--;
+ }
return err;
}
EXPORT_SYMBOL(tcf_block_get_ext);
@@ -346,26 +493,35 @@ void tcf_block_put_ext(struct tcf_block *block, struct Qdisc *q,
{
struct tcf_chain *chain, *tmp;
- /* Hold a refcnt for all chains, so that they don't disappear
- * while we are iterating.
- */
if (!block)
return;
- list_for_each_entry(chain, &block->chain_list, list)
- tcf_chain_hold(chain);
- list_for_each_entry(chain, &block->chain_list, list)
- tcf_chain_flush(chain);
+ tcf_chain_head_change_cb_del(tcf_block_chain_zero(block), ei);
+
+ if (--block->refcnt == 0) {
+ if (ei->shareable)
+ tcf_block_remove(block, block->net);
+
+ /* Hold a refcnt for all chains, so that they don't disappear
+ * while we are iterating.
+ */
+ list_for_each_entry(chain, &block->chain_list, list)
+ tcf_chain_hold(chain);
+
+ list_for_each_entry(chain, &block->chain_list, list)
+ tcf_chain_flush(chain);
+ }
tcf_block_offload_unbind(block, q, ei);
- /* At this point, all the chains should have refcnt >= 1. */
- list_for_each_entry_safe(chain, tmp, &block->chain_list, list)
- tcf_chain_put(chain);
+ if (block->refcnt == 0) {
+ /* At this point, all the chains should have refcnt >= 1. */
+ list_for_each_entry_safe(chain, tmp, &block->chain_list, list)
+ tcf_chain_put(chain);
- /* Finally, put chain 0 and allow block to be freed. */
- chain = list_first_entry(&block->chain_list, struct tcf_chain, list);
- tcf_chain_put(chain);
+ /* Finally, put chain 0 and allow block to be freed. */
+ tcf_chain_put(tcf_block_chain_zero(block));
+ }
}
EXPORT_SYMBOL(tcf_block_put_ext);
@@ -1252,12 +1408,40 @@ int tc_setup_cb_call(struct tcf_block *block, struct tcf_exts *exts,
}
EXPORT_SYMBOL(tc_setup_cb_call);
+static __net_init int tcf_net_init(struct net *net)
+{
+ struct tcf_net *tn = net_generic(net, tcf_net_id);
+
+ idr_init(&tn->idr);
+ return 0;
+}
+
+static void __net_exit tcf_net_exit(struct net *net)
+{
+ struct tcf_net *tn = net_generic(net, tcf_net_id);
+
+ idr_destroy(&tn->idr);
+}
+
+static struct pernet_operations tcf_net_ops = {
+ .init = tcf_net_init,
+ .exit = tcf_net_exit,
+ .id = &tcf_net_id,
+ .size = sizeof(struct tcf_net),
+};
+
static int __init tc_filter_init(void)
{
+ int err;
+
tc_filter_wq = alloc_ordered_workqueue("tc_filter_workqueue", 0);
if (!tc_filter_wq)
return -ENOMEM;
+ err = register_pernet_subsys(&tcf_net_ops);
+ if (err)
+ return err;
+
rtnl_register(PF_UNSPEC, RTM_NEWTFILTER, tc_ctl_tfilter, NULL, 0);
rtnl_register(PF_UNSPEC, RTM_DELTFILTER, tc_ctl_tfilter, NULL, 0);
rtnl_register(PF_UNSPEC, RTM_GETTFILTER, tc_ctl_tfilter,
--
2.9.5
^ permalink raw reply related
* [patch net-next v6 02/11] net: sched: avoid usage of tp->q in tcf_classify
From: Jiri Pirko @ 2018-01-05 23:09 UTC (permalink / raw)
To: netdev
Cc: davem, jhs, xiyou.wangcong, mlxsw, andrew, vivien.didelot,
f.fainelli, michael.chan, ganeshgr, saeedm, matanb, leonro,
idosch, jakub.kicinski, simon.horman, pieter.jansenvanvuuren,
john.hurley, alexander.h.duyck, ogerlitz, john.fastabend, daniel,
dsahern
In-Reply-To: <20180105230929.5645-1-jiri@resnulli.us>
From: Jiri Pirko <jiri@mellanox.com>
Use block index in the messages instead.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
net/sched/cls_api.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/net/sched/cls_api.c b/net/sched/cls_api.c
index 2dd584a..f50d203 100644
--- a/net/sched/cls_api.c
+++ b/net/sched/cls_api.c
@@ -680,8 +680,9 @@ int tcf_classify(struct sk_buff *skb, const struct tcf_proto *tp,
#ifdef CONFIG_NET_CLS_ACT
reset:
if (unlikely(limit++ >= max_reclassify_loop)) {
- net_notice_ratelimited("%s: reclassify loop, rule prio %u, protocol %02x\n",
- tp->q->ops->id, tp->prio & 0xffff,
+ net_notice_ratelimited("%u: reclassify loop, rule prio %u, protocol %02x\n",
+ tp->chain->block->index,
+ tp->prio & 0xffff,
ntohs(tp->protocol));
return TC_ACT_SHOT;
}
--
2.9.5
^ permalink raw reply related
* [patch net-next v6 04/11] net: sched: remove classid and q fields from tcf_proto
From: Jiri Pirko @ 2018-01-05 23:09 UTC (permalink / raw)
To: netdev
Cc: davem, jhs, xiyou.wangcong, mlxsw, andrew, vivien.didelot,
f.fainelli, michael.chan, ganeshgr, saeedm, matanb, leonro,
idosch, jakub.kicinski, simon.horman, pieter.jansenvanvuuren,
john.hurley, alexander.h.duyck, ogerlitz, john.fastabend, daniel,
dsahern
In-Reply-To: <20180105230929.5645-1-jiri@resnulli.us>
From: Jiri Pirko <jiri@mellanox.com>
Both are no longer used, so remove them.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
include/net/sch_generic.h | 2 --
net/sched/cls_api.c | 7 ++-----
2 files changed, 2 insertions(+), 7 deletions(-)
diff --git a/include/net/sch_generic.h b/include/net/sch_generic.h
index df97c3e..dba2214 100644
--- a/include/net/sch_generic.h
+++ b/include/net/sch_generic.h
@@ -255,8 +255,6 @@ struct tcf_proto {
/* All the rest */
u32 prio;
- u32 classid;
- struct Qdisc *q;
void *data;
const struct tcf_proto_ops *ops;
struct tcf_chain *chain;
diff --git a/net/sched/cls_api.c b/net/sched/cls_api.c
index 3e04d88..3e12ef9 100644
--- a/net/sched/cls_api.c
+++ b/net/sched/cls_api.c
@@ -122,8 +122,7 @@ static inline u32 tcf_auto_prio(struct tcf_proto *tp)
}
static struct tcf_proto *tcf_proto_create(const char *kind, u32 protocol,
- u32 prio, u32 parent, struct Qdisc *q,
- struct tcf_chain *chain)
+ u32 prio, struct tcf_chain *chain)
{
struct tcf_proto *tp;
int err;
@@ -157,8 +156,6 @@ static struct tcf_proto *tcf_proto_create(const char *kind, u32 protocol,
tp->classify = tp->ops->classify;
tp->protocol = protocol;
tp->prio = prio;
- tp->classid = parent;
- tp->q = q;
tp->chain = chain;
err = tp->ops->init(tp);
@@ -1077,7 +1074,7 @@ static int tc_ctl_tfilter(struct sk_buff *skb, struct nlmsghdr *n,
prio = tcf_auto_prio(tcf_chain_tp_prev(&chain_info));
tp = tcf_proto_create(nla_data(tca[TCA_KIND]),
- protocol, prio, parent, q, chain);
+ protocol, prio, chain);
if (IS_ERR(tp)) {
err = PTR_ERR(tp);
goto errout;
--
2.9.5
^ permalink raw reply related
* [patch net-next v6 03/11] net: sched: introduce block mechanism to handle netif_keep_dst calls
From: Jiri Pirko @ 2018-01-05 23:09 UTC (permalink / raw)
To: netdev
Cc: davem, jhs, xiyou.wangcong, mlxsw, andrew, vivien.didelot,
f.fainelli, michael.chan, ganeshgr, saeedm, matanb, leonro,
idosch, jakub.kicinski, simon.horman, pieter.jansenvanvuuren,
john.hurley, alexander.h.duyck, ogerlitz, john.fastabend, daniel,
dsahern
In-Reply-To: <20180105230929.5645-1-jiri@resnulli.us>
From: Jiri Pirko <jiri@mellanox.com>
Couple of classifiers call netif_keep_dst directly on q->dev. That is
not possible to do directly for shared blocke where multiple qdiscs are
owning the block. So introduce a infrastructure to keep track of the
block owners in list and use this list to implement block variant of
netif_keep_dst.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
v3->v4:
- rebased on top of the current net-next
v1->v2:
- fix binder_type to check "egress" as well as pointed out by Daniel
---
include/net/pkt_cls.h | 1 +
include/net/sch_generic.h | 2 ++
net/sched/cls_api.c | 69 +++++++++++++++++++++++++++++++++++++++++++++++
net/sched/cls_bpf.c | 4 +--
net/sched/cls_flow.c | 2 +-
net/sched/cls_route.c | 2 +-
6 files changed, 76 insertions(+), 4 deletions(-)
diff --git a/include/net/pkt_cls.h b/include/net/pkt_cls.h
index 95c90da..5b53fda 100644
--- a/include/net/pkt_cls.h
+++ b/include/net/pkt_cls.h
@@ -40,6 +40,7 @@ bool tcf_queue_work(struct work_struct *work);
struct tcf_chain *tcf_chain_get(struct tcf_block *block, u32 chain_index,
bool create);
void tcf_chain_put(struct tcf_chain *chain);
+void tcf_block_netif_keep_dst(struct tcf_block *block);
int tcf_block_get(struct tcf_block **p_block,
struct tcf_proto __rcu **p_filter_chain, struct Qdisc *q,
struct netlink_ext_ack *extack);
diff --git a/include/net/sch_generic.h b/include/net/sch_generic.h
index 5cc4d71..df97c3e 100644
--- a/include/net/sch_generic.h
+++ b/include/net/sch_generic.h
@@ -289,6 +289,8 @@ struct tcf_block {
struct net *net;
struct Qdisc *q;
struct list_head cb_list;
+ struct list_head owner_list;
+ bool keep_dst;
};
static inline void qdisc_cb_private_validate(const struct sk_buff *skb, int sz)
diff --git a/net/sched/cls_api.c b/net/sched/cls_api.c
index f50d203..3e04d88 100644
--- a/net/sched/cls_api.c
+++ b/net/sched/cls_api.c
@@ -383,6 +383,7 @@ static struct tcf_block *tcf_block_create(struct net *net, struct Qdisc *q,
}
INIT_LIST_HEAD(&block->chain_list);
INIT_LIST_HEAD(&block->cb_list);
+ INIT_LIST_HEAD(&block->owner_list);
/* Create chain 0 by default, it has to be always present. */
chain = tcf_chain_create(block, 0);
@@ -414,6 +415,65 @@ static struct tcf_chain *tcf_block_chain_zero(struct tcf_block *block)
return list_first_entry(&block->chain_list, struct tcf_chain, list);
}
+struct tcf_block_owner_item {
+ struct list_head list;
+ struct Qdisc *q;
+ enum tcf_block_binder_type binder_type;
+};
+
+static void
+tcf_block_owner_netif_keep_dst(struct tcf_block *block,
+ struct Qdisc *q,
+ enum tcf_block_binder_type binder_type)
+{
+ if (block->keep_dst &&
+ binder_type != TCF_BLOCK_BINDER_TYPE_CLSACT_INGRESS &&
+ binder_type != TCF_BLOCK_BINDER_TYPE_CLSACT_EGRESS)
+ netif_keep_dst(qdisc_dev(q));
+}
+
+void tcf_block_netif_keep_dst(struct tcf_block *block)
+{
+ struct tcf_block_owner_item *item;
+
+ block->keep_dst = true;
+ list_for_each_entry(item, &block->owner_list, list)
+ tcf_block_owner_netif_keep_dst(block, item->q,
+ item->binder_type);
+}
+EXPORT_SYMBOL(tcf_block_netif_keep_dst);
+
+static int tcf_block_owner_add(struct tcf_block *block,
+ struct Qdisc *q,
+ enum tcf_block_binder_type binder_type)
+{
+ struct tcf_block_owner_item *item;
+
+ item = kmalloc(sizeof(*item), GFP_KERNEL);
+ if (!item)
+ return -ENOMEM;
+ item->q = q;
+ item->binder_type = binder_type;
+ list_add(&item->list, &block->owner_list);
+ return 0;
+}
+
+static void tcf_block_owner_del(struct tcf_block *block,
+ struct Qdisc *q,
+ enum tcf_block_binder_type binder_type)
+{
+ struct tcf_block_owner_item *item;
+
+ list_for_each_entry(item, &block->owner_list, list) {
+ if (item->q == q && item->binder_type == binder_type) {
+ list_del(&item->list);
+ kfree(item);
+ return;
+ }
+ }
+ WARN_ON(1);
+}
+
int tcf_block_get_ext(struct tcf_block **p_block, struct Qdisc *q,
struct tcf_block_ext_info *ei,
struct netlink_ext_ack *extack)
@@ -442,6 +502,12 @@ int tcf_block_get_ext(struct tcf_block **p_block, struct Qdisc *q,
}
}
+ err = tcf_block_owner_add(block, q, ei->binder_type);
+ if (err)
+ goto err_block_owner_add;
+
+ tcf_block_owner_netif_keep_dst(block, q, ei->binder_type);
+
err = tcf_chain_head_change_cb_add(tcf_block_chain_zero(block),
ei, extack);
if (err)
@@ -452,6 +518,8 @@ int tcf_block_get_ext(struct tcf_block **p_block, struct Qdisc *q,
return 0;
err_chain_head_change_cb_add:
+ tcf_block_owner_del(block, q, ei->binder_type);
+err_block_owner_add:
if (created) {
if (ei->shareable)
tcf_block_remove(block, net);
@@ -497,6 +565,7 @@ void tcf_block_put_ext(struct tcf_block *block, struct Qdisc *q,
return;
tcf_chain_head_change_cb_del(tcf_block_chain_zero(block), ei);
+ tcf_block_owner_del(block, q, ei->binder_type);
if (--block->refcnt == 0) {
if (ei->shareable)
diff --git a/net/sched/cls_bpf.c b/net/sched/cls_bpf.c
index 8d78e7f..d79cc50 100644
--- a/net/sched/cls_bpf.c
+++ b/net/sched/cls_bpf.c
@@ -392,8 +392,8 @@ static int cls_bpf_prog_from_efd(struct nlattr **tb, struct cls_bpf_prog *prog,
prog->bpf_name = name;
prog->filter = fp;
- if (fp->dst_needed && !(tp->q->flags & TCQ_F_INGRESS))
- netif_keep_dst(qdisc_dev(tp->q));
+ if (fp->dst_needed)
+ tcf_block_netif_keep_dst(tp->chain->block);
return 0;
}
diff --git a/net/sched/cls_flow.c b/net/sched/cls_flow.c
index 25c2a88..28cd6fb 100644
--- a/net/sched/cls_flow.c
+++ b/net/sched/cls_flow.c
@@ -526,7 +526,7 @@ static int flow_change(struct net *net, struct sk_buff *in_skb,
timer_setup(&fnew->perturb_timer, flow_perturbation, TIMER_DEFERRABLE);
- netif_keep_dst(qdisc_dev(tp->q));
+ tcf_block_netif_keep_dst(tp->chain->block);
if (tb[TCA_FLOW_KEYS]) {
fnew->keymask = keymask;
diff --git a/net/sched/cls_route.c b/net/sched/cls_route.c
index ac9a5b8..a1f2b1b 100644
--- a/net/sched/cls_route.c
+++ b/net/sched/cls_route.c
@@ -527,7 +527,7 @@ static int route4_change(struct net *net, struct sk_buff *in_skb,
if (f->handle < f1->handle)
break;
- netif_keep_dst(qdisc_dev(tp->q));
+ tcf_block_netif_keep_dst(tp->chain->block);
rcu_assign_pointer(f->next, f1);
rcu_assign_pointer(*fp, f);
--
2.9.5
^ permalink raw reply related
* [patch net-next v6 06/11] net: sched: use block index as a handle instead of qdisc when block is shared
From: Jiri Pirko @ 2018-01-05 23:09 UTC (permalink / raw)
To: netdev
Cc: davem, jhs, xiyou.wangcong, mlxsw, andrew, vivien.didelot,
f.fainelli, michael.chan, ganeshgr, saeedm, matanb, leonro,
idosch, jakub.kicinski, simon.horman, pieter.jansenvanvuuren,
john.hurley, alexander.h.duyck, ogerlitz, john.fastabend, daniel,
dsahern
In-Reply-To: <20180105230929.5645-1-jiri@resnulli.us>
From: Jiri Pirko <jiri@mellanox.com>
As the tcm_ifindex 0 is invalid ifindex, reuse it to indicate that we
work with block, instead of qdisc. So if tcm_ifindex is 0, tcm_parent is
used to carry block_index.
If the block is set to be shared between at least 2 qdiscs, it is
forbidden to use the qdisc handle to add/delete filters. In that case,
userspace has to pass block_index.
Also, for dump of the filters, in case the block is shared in between at
least 2 qdiscs, the each filter is dumped with tcm_ifindex 0 and
tcm_parent set to block_index. That gives the user clear indication,
that the filter belongs to a shared block and not only to one qdisc
under which it is dumped.
Suggested-by: David Ahern <dsahern@gmail.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
v5->v6:
- new patch
---
include/net/pkt_cls.h | 7 +-
net/sched/cls_api.c | 208 ++++++++++++++++++++++++++++++--------------------
2 files changed, 132 insertions(+), 83 deletions(-)
diff --git a/include/net/pkt_cls.h b/include/net/pkt_cls.h
index 5b53fda..6f9999c 100644
--- a/include/net/pkt_cls.h
+++ b/include/net/pkt_cls.h
@@ -51,9 +51,14 @@ void tcf_block_put(struct tcf_block *block);
void tcf_block_put_ext(struct tcf_block *block, struct Qdisc *q,
struct tcf_block_ext_info *ei);
+static inline bool tcf_block_shared(struct tcf_block *block)
+{
+ return block->refcnt != 1;
+}
+
static inline struct Qdisc *tcf_block_q(struct tcf_block *block)
{
- WARN_ON(block->refcnt != 1);
+ WARN_ON(tcf_block_shared(block));
return block->q;
}
diff --git a/net/sched/cls_api.c b/net/sched/cls_api.c
index ae60fce..cf4fc7a 100644
--- a/net/sched/cls_api.c
+++ b/net/sched/cls_api.c
@@ -872,8 +872,9 @@ static struct tcf_proto *tcf_chain_tp_find(struct tcf_chain *chain,
}
static int tcf_fill_node(struct net *net, struct sk_buff *skb,
- struct tcf_proto *tp, struct Qdisc *q, u32 parent,
- void *fh, u32 portid, u32 seq, u16 flags, int event)
+ struct tcf_proto *tp, struct tcf_block *block,
+ struct Qdisc *q, u32 parent, void *fh,
+ u32 portid, u32 seq, u16 flags, int event)
{
struct tcmsg *tcm;
struct nlmsghdr *nlh;
@@ -886,8 +887,13 @@ static int tcf_fill_node(struct net *net, struct sk_buff *skb,
tcm->tcm_family = AF_UNSPEC;
tcm->tcm__pad1 = 0;
tcm->tcm__pad2 = 0;
- tcm->tcm_ifindex = qdisc_dev(q)->ifindex;
- tcm->tcm_parent = parent;
+ if (q) {
+ tcm->tcm_ifindex = qdisc_dev(q)->ifindex;
+ tcm->tcm_parent = parent;
+ } else {
+ tcm->tcm_ifindex = 0; /* block index is stored in parent */
+ tcm->tcm_parent = block->index;
+ }
tcm->tcm_info = TC_H_MAKE(tp->prio, tp->protocol);
if (nla_put_string(skb, TCA_KIND, tp->ops->kind))
goto nla_put_failure;
@@ -910,8 +916,8 @@ static int tcf_fill_node(struct net *net, struct sk_buff *skb,
static int tfilter_notify(struct net *net, struct sk_buff *oskb,
struct nlmsghdr *n, struct tcf_proto *tp,
- struct Qdisc *q, u32 parent,
- void *fh, int event, bool unicast)
+ struct tcf_block *block, struct Qdisc *q,
+ u32 parent, void *fh, int event, bool unicast)
{
struct sk_buff *skb;
u32 portid = oskb ? NETLINK_CB(oskb).portid : 0;
@@ -920,8 +926,8 @@ static int tfilter_notify(struct net *net, struct sk_buff *oskb,
if (!skb)
return -ENOBUFS;
- if (tcf_fill_node(net, skb, tp, q, parent, fh, portid, n->nlmsg_seq,
- n->nlmsg_flags, event) <= 0) {
+ if (tcf_fill_node(net, skb, tp, block, q, parent, fh, portid,
+ n->nlmsg_seq, n->nlmsg_flags, event) <= 0) {
kfree_skb(skb);
return -EINVAL;
}
@@ -935,8 +941,8 @@ static int tfilter_notify(struct net *net, struct sk_buff *oskb,
static int tfilter_del_notify(struct net *net, struct sk_buff *oskb,
struct nlmsghdr *n, struct tcf_proto *tp,
- struct Qdisc *q, u32 parent,
- void *fh, bool unicast, bool *last)
+ struct tcf_block *block, struct Qdisc *q,
+ u32 parent, void *fh, bool unicast, bool *last)
{
struct sk_buff *skb;
u32 portid = oskb ? NETLINK_CB(oskb).portid : 0;
@@ -946,8 +952,8 @@ static int tfilter_del_notify(struct net *net, struct sk_buff *oskb,
if (!skb)
return -ENOBUFS;
- if (tcf_fill_node(net, skb, tp, q, parent, fh, portid, n->nlmsg_seq,
- n->nlmsg_flags, RTM_DELTFILTER) <= 0) {
+ if (tcf_fill_node(net, skb, tp, block, q, parent, fh, portid,
+ n->nlmsg_seq, n->nlmsg_flags, RTM_DELTFILTER) <= 0) {
kfree_skb(skb);
return -EINVAL;
}
@@ -966,15 +972,16 @@ static int tfilter_del_notify(struct net *net, struct sk_buff *oskb,
}
static void tfilter_notify_chain(struct net *net, struct sk_buff *oskb,
- struct Qdisc *q, u32 parent,
- struct nlmsghdr *n,
+ struct tcf_block *block, struct Qdisc *q,
+ u32 parent, struct nlmsghdr *n,
struct tcf_chain *chain, int event)
{
struct tcf_proto *tp;
for (tp = rtnl_dereference(chain->filter_chain);
tp; tp = rtnl_dereference(tp->next))
- tfilter_notify(net, oskb, n, tp, q, parent, 0, event, false);
+ tfilter_notify(net, oskb, n, tp, block,
+ q, parent, 0, event, false);
}
/* Add/change/delete/get a filter node */
@@ -990,13 +997,11 @@ static int tc_ctl_tfilter(struct sk_buff *skb, struct nlmsghdr *n,
bool prio_allocate;
u32 parent;
u32 chain_index;
- struct net_device *dev;
- struct Qdisc *q;
+ struct Qdisc *q = NULL;
struct tcf_chain_info chain_info;
struct tcf_chain *chain = NULL;
struct tcf_block *block;
struct tcf_proto *tp;
- const struct Qdisc_class_ops *cops;
unsigned long cl;
void *fh;
int err;
@@ -1043,41 +1048,63 @@ static int tc_ctl_tfilter(struct sk_buff *skb, struct nlmsghdr *n,
/* Find head of filter chain. */
- /* Find link */
- dev = __dev_get_by_index(net, t->tcm_ifindex);
- if (dev == NULL)
- return -ENODEV;
- /* Find qdisc */
- if (!parent) {
- q = dev->qdisc;
- parent = q->handle;
- } else {
- q = qdisc_lookup(dev, TC_H_MAJ(t->tcm_parent));
- if (q == NULL)
+ if (t->tcm_ifindex) {
+ const struct Qdisc_class_ops *cops;
+ struct net_device *dev;
+
+ /* Find link */
+ dev = __dev_get_by_index(net, t->tcm_ifindex);
+ if (!dev)
+ return -ENODEV;
+
+ /* Find qdisc */
+ if (!parent) {
+ q = dev->qdisc;
+ parent = q->handle;
+ } else {
+ q = qdisc_lookup(dev, TC_H_MAJ(t->tcm_parent));
+ if (!q)
+ return -EINVAL;
+ }
+
+ /* Is it classful? */
+ cops = q->ops->cl_ops;
+ if (!cops)
return -EINVAL;
- }
- /* Is it classful? */
- cops = q->ops->cl_ops;
- if (!cops)
- return -EINVAL;
+ if (!cops->tcf_block)
+ return -EOPNOTSUPP;
- if (!cops->tcf_block)
- return -EOPNOTSUPP;
+ /* Do we search for filter, attached to class? */
+ if (TC_H_MIN(parent)) {
+ cl = cops->find(q, parent);
+ if (cl == 0)
+ return -ENOENT;
+ }
- /* Do we search for filter, attached to class? */
- if (TC_H_MIN(parent)) {
- cl = cops->find(q, parent);
- if (cl == 0)
- return -ENOENT;
- }
+ /* And the last stroke */
+ block = cops->tcf_block(q, cl, extack);
+ if (!block) {
+ err = -EINVAL;
+ goto errout;
+ }
+ if (tcf_block_shared(block)) {
+ NL_SET_ERR_MSG(extack, "Cannot work with shared block, please use block index");
+ err = -EOPNOTSUPP;
+ goto errout;
+ }
+ } else {
+ /* tcm_ifindex == means that the user wants to work with block.
+ * The block index is stored in tcm_parent.
+ */
+ u32 block_index = t->tcm_parent;
- /* And the last stroke */
- block = cops->tcf_block(q, cl, extack);
- if (!block) {
- err = -EINVAL;
- goto errout;
+ block = tcf_block_lookup(net, block_index);
+ if (!block) {
+ err = -EINVAL;
+ goto errout;
+ }
}
chain_index = tca[TCA_CHAIN] ? nla_get_u32(tca[TCA_CHAIN]) : 0;
@@ -1093,7 +1120,7 @@ static int tc_ctl_tfilter(struct sk_buff *skb, struct nlmsghdr *n,
}
if (n->nlmsg_type == RTM_DELTFILTER && prio == 0) {
- tfilter_notify_chain(net, skb, q, parent, n,
+ tfilter_notify_chain(net, skb, block, q, parent, n,
chain, RTM_DELTFILTER);
tcf_chain_flush(chain);
err = 0;
@@ -1141,7 +1168,7 @@ static int tc_ctl_tfilter(struct sk_buff *skb, struct nlmsghdr *n,
if (!fh) {
if (n->nlmsg_type == RTM_DELTFILTER && t->tcm_handle == 0) {
tcf_chain_tp_remove(chain, &chain_info, tp);
- tfilter_notify(net, skb, n, tp, q, parent, fh,
+ tfilter_notify(net, skb, n, tp, block, q, parent, fh,
RTM_DELTFILTER, false);
tcf_proto_destroy(tp);
err = 0;
@@ -1166,8 +1193,8 @@ static int tc_ctl_tfilter(struct sk_buff *skb, struct nlmsghdr *n,
}
break;
case RTM_DELTFILTER:
- err = tfilter_del_notify(net, skb, n, tp, q, parent,
- fh, false, &last);
+ err = tfilter_del_notify(net, skb, n, tp, block,
+ q, parent, fh, false, &last);
if (err)
goto errout;
if (last) {
@@ -1176,8 +1203,8 @@ static int tc_ctl_tfilter(struct sk_buff *skb, struct nlmsghdr *n,
}
goto errout;
case RTM_GETTFILTER:
- err = tfilter_notify(net, skb, n, tp, q, parent, fh,
- RTM_NEWTFILTER, true);
+ err = tfilter_notify(net, skb, n, tp, block, q, parent,
+ fh, RTM_NEWTFILTER, true);
goto errout;
default:
err = -EINVAL;
@@ -1190,7 +1217,7 @@ static int tc_ctl_tfilter(struct sk_buff *skb, struct nlmsghdr *n,
if (err == 0) {
if (tp_created)
tcf_chain_tp_insert(chain, &chain_info, tp);
- tfilter_notify(net, skb, n, tp, q, parent, fh,
+ tfilter_notify(net, skb, n, tp, block, q, parent, fh,
RTM_NEWTFILTER, false);
} else {
if (tp_created)
@@ -1210,6 +1237,7 @@ struct tcf_dump_args {
struct tcf_walker w;
struct sk_buff *skb;
struct netlink_callback *cb;
+ struct tcf_block *block;
struct Qdisc *q;
u32 parent;
};
@@ -1219,7 +1247,7 @@ static int tcf_node_dump(struct tcf_proto *tp, void *n, struct tcf_walker *arg)
struct tcf_dump_args *a = (void *)arg;
struct net *net = sock_net(a->skb->sk);
- return tcf_fill_node(net, a->skb, tp, a->q, a->parent,
+ return tcf_fill_node(net, a->skb, tp, a->block, a->q, a->parent,
n, NETLINK_CB(a->cb->skb).portid,
a->cb->nlh->nlmsg_seq, NLM_F_MULTI,
RTM_NEWTFILTER);
@@ -1230,6 +1258,7 @@ static bool tcf_chain_dump(struct tcf_chain *chain, struct Qdisc *q, u32 parent,
long index_start, long *p_index)
{
struct net *net = sock_net(skb->sk);
+ struct tcf_block *block = chain->block;
struct tcmsg *tcm = nlmsg_data(cb->nlh);
struct tcf_dump_args arg;
struct tcf_proto *tp;
@@ -1248,7 +1277,7 @@ static bool tcf_chain_dump(struct tcf_chain *chain, struct Qdisc *q, u32 parent,
memset(&cb->args[1], 0,
sizeof(cb->args) - sizeof(cb->args[0]));
if (cb->args[1] == 0) {
- if (tcf_fill_node(net, skb, tp, q, parent, 0,
+ if (tcf_fill_node(net, skb, tp, block, q, parent, 0,
NETLINK_CB(cb->skb).portid,
cb->nlh->nlmsg_seq, NLM_F_MULTI,
RTM_NEWTFILTER) <= 0)
@@ -1261,6 +1290,7 @@ static bool tcf_chain_dump(struct tcf_chain *chain, struct Qdisc *q, u32 parent,
arg.w.fn = tcf_node_dump;
arg.skb = skb;
arg.cb = cb;
+ arg.block = block;
arg.q = q;
arg.parent = parent;
arg.w.stop = 0;
@@ -1279,13 +1309,10 @@ static int tc_dump_tfilter(struct sk_buff *skb, struct netlink_callback *cb)
{
struct net *net = sock_net(skb->sk);
struct nlattr *tca[TCA_MAX + 1];
- struct net_device *dev;
- struct Qdisc *q;
+ struct Qdisc *q = NULL;
struct tcf_block *block;
struct tcf_chain *chain;
struct tcmsg *tcm = nlmsg_data(cb->nlh);
- unsigned long cl = 0;
- const struct Qdisc_class_ops *cops;
long index_start;
long index;
u32 parent;
@@ -1298,32 +1325,49 @@ static int tc_dump_tfilter(struct sk_buff *skb, struct netlink_callback *cb)
if (err)
return err;
- dev = __dev_get_by_index(net, tcm->tcm_ifindex);
- if (!dev)
- return skb->len;
+ if (tcm->tcm_ifindex) {
+ const struct Qdisc_class_ops *cops;
+ struct net_device *dev;
+ unsigned long cl = 0;
+
+ dev = __dev_get_by_index(net, tcm->tcm_ifindex);
+ if (!dev)
+ return skb->len;
- parent = tcm->tcm_parent;
- if (!parent) {
- q = dev->qdisc;
- parent = q->handle;
+ parent = tcm->tcm_parent;
+ if (!parent) {
+ q = dev->qdisc;
+ parent = q->handle;
+ } else {
+ q = qdisc_lookup(dev, TC_H_MAJ(tcm->tcm_parent));
+ }
+ if (!q)
+ goto out;
+ cops = q->ops->cl_ops;
+ if (!cops)
+ goto out;
+ if (!cops->tcf_block)
+ goto out;
+ if (TC_H_MIN(tcm->tcm_parent)) {
+ cl = cops->find(q, tcm->tcm_parent);
+ if (cl == 0)
+ goto out;
+ }
+ block = cops->tcf_block(q, cl, NULL);
+ if (!block)
+ goto out;
+ if (tcf_block_shared(block))
+ q = NULL;
} else {
- q = qdisc_lookup(dev, TC_H_MAJ(tcm->tcm_parent));
- }
- if (!q)
- goto out;
- cops = q->ops->cl_ops;
- if (!cops)
- goto out;
- if (!cops->tcf_block)
- goto out;
- if (TC_H_MIN(tcm->tcm_parent)) {
- cl = cops->find(q, tcm->tcm_parent);
- if (cl == 0)
+ /* tcm_ifindex == means that the user wants to work with block.
+ * The block index is stored in tcm_parent.
+ */
+ u32 block_index = tcm->tcm_parent;
+
+ block = tcf_block_lookup(net, block_index);
+ if (!block)
goto out;
}
- block = cops->tcf_block(q, cl, NULL);
- if (!block)
- goto out;
index_start = cb->args[0];
index = 0;
--
2.9.5
^ permalink raw reply related
* [patch net-next v6 05/11] net: sched: keep track of offloaded filters and check tc offload feature
From: Jiri Pirko @ 2018-01-05 23:09 UTC (permalink / raw)
To: netdev
Cc: davem, jhs, xiyou.wangcong, mlxsw, andrew, vivien.didelot,
f.fainelli, michael.chan, ganeshgr, saeedm, matanb, leonro,
idosch, jakub.kicinski, simon.horman, pieter.jansenvanvuuren,
john.hurley, alexander.h.duyck, ogerlitz, john.fastabend, daniel,
dsahern
In-Reply-To: <20180105230929.5645-1-jiri@resnulli.us>
From: Jiri Pirko <jiri@mellanox.com>
During block bind, we need to check tc offload feature. If it is
disabled yet still the block contains offloaded filters, forbid the
bind. Also forbid to register callback for a block that already
contains offloaded filters, as the play back is not supported now.
For keeping track of offloaded filters there is a new counter
introduced, alongside with couple of helpers called from cls_* code.
These helpers set and clear TCA_CLS_FLAGS_IN_HW flag.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
v4->v5:
- add tracking of binding of devs that are unable to offload and check
that before block cbs call.
v3->v4:
- propagate netdev_ops->ndo_setup_tc error up to tcf_block_offload_bind
caller
v2->v3:
- new patch
---
include/net/sch_generic.h | 18 +++++++++++
net/sched/cls_api.c | 79 ++++++++++++++++++++++++++++++++++++++---------
net/sched/cls_bpf.c | 5 ++-
net/sched/cls_flower.c | 3 +-
net/sched/cls_matchall.c | 3 +-
net/sched/cls_u32.c | 13 ++++----
6 files changed, 98 insertions(+), 23 deletions(-)
diff --git a/include/net/sch_generic.h b/include/net/sch_generic.h
index dba2214..ab86b64 100644
--- a/include/net/sch_generic.h
+++ b/include/net/sch_generic.h
@@ -289,8 +289,26 @@ struct tcf_block {
struct list_head cb_list;
struct list_head owner_list;
bool keep_dst;
+ unsigned int offloadcnt; /* Number of oddloaded filters */
+ unsigned int nooffloaddevcnt; /* Number of devs unable to do offload */
};
+static inline void tcf_block_offload_inc(struct tcf_block *block, u32 *flags)
+{
+ if (*flags & TCA_CLS_FLAGS_IN_HW)
+ return;
+ *flags |= TCA_CLS_FLAGS_IN_HW;
+ block->offloadcnt++;
+}
+
+static inline void tcf_block_offload_dec(struct tcf_block *block, u32 *flags)
+{
+ if (!(*flags & TCA_CLS_FLAGS_IN_HW))
+ return;
+ *flags &= ~TCA_CLS_FLAGS_IN_HW;
+ block->offloadcnt--;
+}
+
static inline void qdisc_cb_private_validate(const struct sk_buff *skb, int sz)
{
struct qdisc_skb_cb *qcb;
diff --git a/net/sched/cls_api.c b/net/sched/cls_api.c
index 3e12ef9..ae60fce 100644
--- a/net/sched/cls_api.c
+++ b/net/sched/cls_api.c
@@ -265,31 +265,66 @@ void tcf_chain_put(struct tcf_chain *chain)
}
EXPORT_SYMBOL(tcf_chain_put);
-static void tcf_block_offload_cmd(struct tcf_block *block, struct Qdisc *q,
- struct tcf_block_ext_info *ei,
- enum tc_block_command command)
+static bool tcf_block_offload_in_use(struct tcf_block *block)
+{
+ return block->offloadcnt;
+}
+
+static int tcf_block_offload_cmd(struct tcf_block *block,
+ struct net_device *dev,
+ struct tcf_block_ext_info *ei,
+ enum tc_block_command command)
{
- struct net_device *dev = q->dev_queue->dev;
struct tc_block_offload bo = {};
- if (!dev->netdev_ops->ndo_setup_tc)
- return;
bo.command = command;
bo.binder_type = ei->binder_type;
bo.block = block;
- dev->netdev_ops->ndo_setup_tc(dev, TC_SETUP_BLOCK, &bo);
+ return dev->netdev_ops->ndo_setup_tc(dev, TC_SETUP_BLOCK, &bo);
}
-static void tcf_block_offload_bind(struct tcf_block *block, struct Qdisc *q,
- struct tcf_block_ext_info *ei)
+static int tcf_block_offload_bind(struct tcf_block *block, struct Qdisc *q,
+ struct tcf_block_ext_info *ei)
{
- tcf_block_offload_cmd(block, q, ei, TC_BLOCK_BIND);
+ struct net_device *dev = q->dev_queue->dev;
+ int err;
+
+ if (!dev->netdev_ops->ndo_setup_tc)
+ goto no_offload_dev_inc;
+
+ /* If tc offload feature is disabled and the block we try to bind
+ * to already has some offloaded filters, forbid to bind.
+ */
+ if (!tc_can_offload(dev) && tcf_block_offload_in_use(block))
+ return -EOPNOTSUPP;
+
+ err = tcf_block_offload_cmd(block, dev, ei, TC_BLOCK_BIND);
+ if (err == -EOPNOTSUPP)
+ goto no_offload_dev_inc;
+ return err;
+
+no_offload_dev_inc:
+ if (tcf_block_offload_in_use(block))
+ return -EOPNOTSUPP;
+ block->nooffloaddevcnt++;
+ return 0;
}
static void tcf_block_offload_unbind(struct tcf_block *block, struct Qdisc *q,
struct tcf_block_ext_info *ei)
{
- tcf_block_offload_cmd(block, q, ei, TC_BLOCK_UNBIND);
+ struct net_device *dev = q->dev_queue->dev;
+ int err;
+
+ if (!dev->netdev_ops->ndo_setup_tc)
+ goto no_offload_dev_dec;
+ err = tcf_block_offload_cmd(block, dev, ei, TC_BLOCK_UNBIND);
+ if (err == -EOPNOTSUPP)
+ goto no_offload_dev_dec;
+ return;
+
+no_offload_dev_dec:
+ WARN_ON(block->nooffloaddevcnt-- == 0);
}
static int
@@ -510,10 +545,15 @@ int tcf_block_get_ext(struct tcf_block **p_block, struct Qdisc *q,
if (err)
goto err_chain_head_change_cb_add;
- tcf_block_offload_bind(block, q, ei);
+ err = tcf_block_offload_bind(block, q, ei);
+ if (err)
+ goto err_block_offload_bind;
+
*p_block = block;
return 0;
+err_block_offload_bind:
+ tcf_chain_head_change_cb_del(tcf_block_chain_zero(block), ei);
err_chain_head_change_cb_add:
tcf_block_owner_del(block, q, ei->binder_type);
err_block_owner_add:
@@ -645,9 +685,16 @@ struct tcf_block_cb *__tcf_block_cb_register(struct tcf_block *block,
{
struct tcf_block_cb *block_cb;
+ /* At this point, playback of previous block cb calls is not supported,
+ * so forbid to register to block which already has some offloaded
+ * filters present.
+ */
+ if (tcf_block_offload_in_use(block))
+ return ERR_PTR(-EOPNOTSUPP);
+
block_cb = kzalloc(sizeof(*block_cb), GFP_KERNEL);
if (!block_cb)
- return NULL;
+ return ERR_PTR(-ENOMEM);
block_cb->cb = cb;
block_cb->cb_ident = cb_ident;
block_cb->cb_priv = cb_priv;
@@ -663,7 +710,7 @@ int tcf_block_cb_register(struct tcf_block *block,
struct tcf_block_cb *block_cb;
block_cb = __tcf_block_cb_register(block, cb, cb_ident, cb_priv);
- return block_cb ? 0 : -ENOMEM;
+ return IS_ERR(block_cb) ? PTR_ERR(block_cb) : 0;
}
EXPORT_SYMBOL(tcf_block_cb_register);
@@ -693,6 +740,10 @@ static int tcf_block_cb_call(struct tcf_block *block, enum tc_setup_type type,
int ok_count = 0;
int err;
+ /* Make sure all netdevs sharing this block are offload-capable. */
+ if (block->nooffloaddevcnt && err_stop)
+ return -EOPNOTSUPP;
+
list_for_each_entry(block_cb, &block->cb_list, list) {
err = block_cb->cb(type, type_data, block_cb->cb_priv);
if (err) {
diff --git a/net/sched/cls_bpf.c b/net/sched/cls_bpf.c
index d79cc50..cf72aef 100644
--- a/net/sched/cls_bpf.c
+++ b/net/sched/cls_bpf.c
@@ -167,13 +167,16 @@ static int cls_bpf_offload_cmd(struct tcf_proto *tp, struct cls_bpf_prog *prog,
cls_bpf.exts_integrated = obj->exts_integrated;
cls_bpf.gen_flags = obj->gen_flags;
+ if (oldprog)
+ tcf_block_offload_dec(block, &oldprog->gen_flags);
+
err = tc_setup_cb_call(block, NULL, TC_SETUP_CLSBPF, &cls_bpf, skip_sw);
if (prog) {
if (err < 0) {
cls_bpf_offload_cmd(tp, oldprog, prog);
return err;
} else if (err > 0) {
- prog->gen_flags |= TCA_CLS_FLAGS_IN_HW;
+ tcf_block_offload_inc(block, &prog->gen_flags);
}
}
diff --git a/net/sched/cls_flower.c b/net/sched/cls_flower.c
index 6132a73..f61df19 100644
--- a/net/sched/cls_flower.c
+++ b/net/sched/cls_flower.c
@@ -229,6 +229,7 @@ static void fl_hw_destroy_filter(struct tcf_proto *tp, struct cls_fl_filter *f)
tc_setup_cb_call(block, &f->exts, TC_SETUP_CLSFLOWER,
&cls_flower, false);
+ tcf_block_offload_dec(block, &f->flags);
}
static int fl_hw_replace_filter(struct tcf_proto *tp,
@@ -256,7 +257,7 @@ static int fl_hw_replace_filter(struct tcf_proto *tp,
fl_hw_destroy_filter(tp, f);
return err;
} else if (err > 0) {
- f->flags |= TCA_CLS_FLAGS_IN_HW;
+ tcf_block_offload_inc(block, &f->flags);
}
if (skip_sw && !(f->flags & TCA_CLS_FLAGS_IN_HW))
diff --git a/net/sched/cls_matchall.c b/net/sched/cls_matchall.c
index 66d4e00..d0e57c8 100644
--- a/net/sched/cls_matchall.c
+++ b/net/sched/cls_matchall.c
@@ -81,6 +81,7 @@ static void mall_destroy_hw_filter(struct tcf_proto *tp,
cls_mall.cookie = cookie;
tc_setup_cb_call(block, NULL, TC_SETUP_CLSMATCHALL, &cls_mall, false);
+ tcf_block_offload_dec(block, &head->flags);
}
static int mall_replace_hw_filter(struct tcf_proto *tp,
@@ -103,7 +104,7 @@ static int mall_replace_hw_filter(struct tcf_proto *tp,
mall_destroy_hw_filter(tp, head, cookie);
return err;
} else if (err > 0) {
- head->flags |= TCA_CLS_FLAGS_IN_HW;
+ tcf_block_offload_inc(block, &head->flags);
}
if (skip_sw && !(head->flags & TCA_CLS_FLAGS_IN_HW))
diff --git a/net/sched/cls_u32.c b/net/sched/cls_u32.c
index 507859c..020d328 100644
--- a/net/sched/cls_u32.c
+++ b/net/sched/cls_u32.c
@@ -529,16 +529,17 @@ static int u32_replace_hw_hnode(struct tcf_proto *tp, struct tc_u_hnode *h,
return 0;
}
-static void u32_remove_hw_knode(struct tcf_proto *tp, u32 handle)
+static void u32_remove_hw_knode(struct tcf_proto *tp, struct tc_u_knode *n)
{
struct tcf_block *block = tp->chain->block;
struct tc_cls_u32_offload cls_u32 = {};
tc_cls_common_offload_init(&cls_u32.common, tp);
cls_u32.command = TC_CLSU32_DELETE_KNODE;
- cls_u32.knode.handle = handle;
+ cls_u32.knode.handle = n->handle;
tc_setup_cb_call(block, NULL, TC_SETUP_CLSU32, &cls_u32, false);
+ tcf_block_offload_dec(block, &n->flags);
}
static int u32_replace_hw_knode(struct tcf_proto *tp, struct tc_u_knode *n,
@@ -567,10 +568,10 @@ static int u32_replace_hw_knode(struct tcf_proto *tp, struct tc_u_knode *n,
err = tc_setup_cb_call(block, NULL, TC_SETUP_CLSU32, &cls_u32, skip_sw);
if (err < 0) {
- u32_remove_hw_knode(tp, n->handle);
+ u32_remove_hw_knode(tp, n);
return err;
} else if (err > 0) {
- n->flags |= TCA_CLS_FLAGS_IN_HW;
+ tcf_block_offload_inc(block, &n->flags);
}
if (skip_sw && !(n->flags & TCA_CLS_FLAGS_IN_HW))
@@ -589,7 +590,7 @@ static void u32_clear_hnode(struct tcf_proto *tp, struct tc_u_hnode *ht)
RCU_INIT_POINTER(ht->ht[h],
rtnl_dereference(n->next));
tcf_unbind_filter(tp, &n->res);
- u32_remove_hw_knode(tp, n->handle);
+ u32_remove_hw_knode(tp, n);
idr_remove_ext(&ht->handle_idr, n->handle);
if (tcf_exts_get_net(&n->exts))
call_rcu(&n->rcu, u32_delete_key_freepf_rcu);
@@ -682,7 +683,7 @@ static int u32_delete(struct tcf_proto *tp, void *arg, bool *last)
goto out;
if (TC_U32_KEY(ht->handle)) {
- u32_remove_hw_knode(tp, ht->handle);
+ u32_remove_hw_knode(tp, (struct tc_u_knode *)ht);
ret = u32_delete_key(tp, (struct tc_u_knode *)ht);
goto out;
}
--
2.9.5
^ permalink raw reply related
* [patch net-next v6 07/11] net: sched: allow ingress and clsact qdiscs to share filter blocks
From: Jiri Pirko @ 2018-01-05 23:09 UTC (permalink / raw)
To: netdev
Cc: davem, jhs, xiyou.wangcong, mlxsw, andrew, vivien.didelot,
f.fainelli, michael.chan, ganeshgr, saeedm, matanb, leonro,
idosch, jakub.kicinski, simon.horman, pieter.jansenvanvuuren,
john.hurley, alexander.h.duyck, ogerlitz, john.fastabend, daniel,
dsahern
In-Reply-To: <20180105230929.5645-1-jiri@resnulli.us>
From: Jiri Pirko <jiri@mellanox.com>
Benefit from the previously introduced shared filter blocks
infrastructure and allow ingress and clsact qdisc instances to share
filter blocks. The block index is coming from userspace as qdisc option.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
v3->v4:
- rebased on top of the current net-next
v2->v3:
- removed "p_" prefix from block index function args
---
include/uapi/linux/pkt_sched.h | 11 ++++++
net/sched/sch_ingress.c | 89 +++++++++++++++++++++++++++++++++++++++++-
2 files changed, 99 insertions(+), 1 deletion(-)
diff --git a/include/uapi/linux/pkt_sched.h b/include/uapi/linux/pkt_sched.h
index 37b5096..8cc554a 100644
--- a/include/uapi/linux/pkt_sched.h
+++ b/include/uapi/linux/pkt_sched.h
@@ -934,4 +934,15 @@ enum {
#define TCA_CBS_MAX (__TCA_CBS_MAX - 1)
+/* Ingress/clsact */
+
+enum {
+ TCA_CLSACT_UNSPEC,
+ TCA_CLSACT_INGRESS_BLOCK,
+ TCA_CLSACT_EGRESS_BLOCK,
+ __TCA_CLSACT_MAX
+};
+
+#define TCA_CLSACT_MAX (__TCA_CLSACT_MAX - 1)
+
#endif
diff --git a/net/sched/sch_ingress.c b/net/sched/sch_ingress.c
index 7ca2be2..debcb8f 100644
--- a/net/sched/sch_ingress.c
+++ b/net/sched/sch_ingress.c
@@ -61,6 +61,29 @@ static void clsact_chain_head_change(struct tcf_proto *tp_head, void *priv)
struct mini_Qdisc_pair *miniqp = priv;
mini_qdisc_pair_swap(miniqp, tp_head);
+};
+
+static const struct nla_policy ingress_policy[TCA_CLSACT_MAX + 1] = {
+ [TCA_CLSACT_INGRESS_BLOCK] = { .type = NLA_U32 },
+};
+
+static int ingress_parse_opt(struct nlattr *opt, u32 *ingress_block_index)
+{
+ struct nlattr *tb[TCA_CLSACT_MAX + 1];
+ int err;
+
+ *ingress_block_index = 0;
+
+ if (!opt)
+ return 0;
+ err = nla_parse_nested(tb, TCA_CLSACT_MAX, opt, ingress_policy, NULL);
+ if (err)
+ return err;
+
+ if (tb[TCA_CLSACT_INGRESS_BLOCK])
+ *ingress_block_index =
+ nla_get_u32(tb[TCA_CLSACT_INGRESS_BLOCK]);
+ return 0;
}
static int ingress_init(struct Qdisc *sch, struct nlattr *opt,
@@ -74,6 +97,11 @@ static int ingress_init(struct Qdisc *sch, struct nlattr *opt,
mini_qdisc_pair_init(&q->miniqp, sch, &dev->miniq_ingress);
+ err = ingress_parse_opt(opt, &q->block_info.block_index);
+ if (err)
+ return err;
+
+ q->block_info.shareable = true;
q->block_info.binder_type = TCF_BLOCK_BINDER_TYPE_CLSACT_INGRESS;
q->block_info.chain_head_change = clsact_chain_head_change;
q->block_info.chain_head_change_priv = &q->miniqp;
@@ -97,11 +125,14 @@ static void ingress_destroy(struct Qdisc *sch)
static int ingress_dump(struct Qdisc *sch, struct sk_buff *skb)
{
+ struct ingress_sched_data *q = qdisc_priv(sch);
struct nlattr *nest;
nest = nla_nest_start(skb, TCA_OPTIONS);
if (nest == NULL)
goto nla_put_failure;
+ if (nla_put_u32(skb, TCA_CLSACT_INGRESS_BLOCK, q->block->index))
+ goto nla_put_failure;
return nla_nest_end(skb, nest);
@@ -170,6 +201,35 @@ static struct tcf_block *clsact_tcf_block(struct Qdisc *sch, unsigned long cl,
}
}
+static const struct nla_policy clsact_policy[TCA_CLSACT_MAX + 1] = {
+ [TCA_CLSACT_INGRESS_BLOCK] = { .type = NLA_U32 },
+ [TCA_CLSACT_EGRESS_BLOCK] = { .type = NLA_U32 },
+};
+
+static int clsact_parse_opt(struct nlattr *opt, u32 *ingress_block_index,
+ u32 *egress_block_index)
+{
+ struct nlattr *tb[TCA_CLSACT_MAX + 1];
+ int err;
+
+ *ingress_block_index = 0;
+ *egress_block_index = 0;
+
+ if (!opt)
+ return 0;
+ err = nla_parse_nested(tb, TCA_CLSACT_MAX, opt, clsact_policy, NULL);
+ if (err)
+ return err;
+
+ if (tb[TCA_CLSACT_INGRESS_BLOCK])
+ *ingress_block_index =
+ nla_get_u32(tb[TCA_CLSACT_INGRESS_BLOCK]);
+ if (tb[TCA_CLSACT_EGRESS_BLOCK])
+ *egress_block_index =
+ nla_get_u32(tb[TCA_CLSACT_EGRESS_BLOCK]);
+ return 0;
+}
+
static int clsact_init(struct Qdisc *sch, struct nlattr *opt,
struct netlink_ext_ack *extack)
{
@@ -182,6 +242,12 @@ static int clsact_init(struct Qdisc *sch, struct nlattr *opt,
mini_qdisc_pair_init(&q->miniqp_ingress, sch, &dev->miniq_ingress);
+ err = clsact_parse_opt(opt, &q->ingress_block_info.block_index,
+ &q->egress_block_info.block_index);
+ if (err)
+ return err;
+
+ q->ingress_block_info.shareable = true;
q->ingress_block_info.binder_type = TCF_BLOCK_BINDER_TYPE_CLSACT_INGRESS;
q->ingress_block_info.chain_head_change = clsact_chain_head_change;
q->ingress_block_info.chain_head_change_priv = &q->miniqp_ingress;
@@ -193,6 +259,7 @@ static int clsact_init(struct Qdisc *sch, struct nlattr *opt,
mini_qdisc_pair_init(&q->miniqp_egress, sch, &dev->miniq_egress);
+ q->egress_block_info.shareable = true;
q->egress_block_info.binder_type = TCF_BLOCK_BINDER_TYPE_CLSACT_EGRESS;
q->egress_block_info.chain_head_change = clsact_chain_head_change;
q->egress_block_info.chain_head_change_priv = &q->miniqp_egress;
@@ -218,6 +285,26 @@ static void clsact_destroy(struct Qdisc *sch)
net_dec_egress_queue();
}
+static int clsact_dump(struct Qdisc *sch, struct sk_buff *skb)
+{
+ struct clsact_sched_data *q = qdisc_priv(sch);
+ struct nlattr *nest;
+
+ nest = nla_nest_start(skb, TCA_OPTIONS);
+ if (!nest)
+ goto nla_put_failure;
+ if (nla_put_u32(skb, TCA_CLSACT_INGRESS_BLOCK, q->ingress_block->index))
+ goto nla_put_failure;
+ if (nla_put_u32(skb, TCA_CLSACT_EGRESS_BLOCK, q->egress_block->index))
+ goto nla_put_failure;
+
+ return nla_nest_end(skb, nest);
+
+nla_put_failure:
+ nla_nest_cancel(skb, nest);
+ return -1;
+}
+
static const struct Qdisc_class_ops clsact_class_ops = {
.leaf = ingress_leaf,
.find = clsact_find,
@@ -233,7 +320,7 @@ static struct Qdisc_ops clsact_qdisc_ops __read_mostly = {
.priv_size = sizeof(struct clsact_sched_data),
.init = clsact_init,
.destroy = clsact_destroy,
- .dump = ingress_dump,
+ .dump = clsact_dump,
.owner = THIS_MODULE,
};
--
2.9.5
^ permalink raw reply related
* [patch net-next v6 08/11] mlxsw: spectrum_acl: Reshuffle code around mlxsw_sp_acl_ruleset_create/destroy
From: Jiri Pirko @ 2018-01-05 23:09 UTC (permalink / raw)
To: netdev
Cc: davem, jhs, xiyou.wangcong, mlxsw, andrew, vivien.didelot,
f.fainelli, michael.chan, ganeshgr, saeedm, matanb, leonro,
idosch, jakub.kicinski, simon.horman, pieter.jansenvanvuuren,
john.hurley, alexander.h.duyck, ogerlitz, john.fastabend, daniel,
dsahern
In-Reply-To: <20180105230929.5645-1-jiri@resnulli.us>
From: Jiri Pirko <jiri@mellanox.com>
In order to prepare for follow-up changes, make the bind/unbind helpers
very simple. That required move of ht insertion/removal and bind/unbind
calls into mlxsw_sp_acl_ruleset_create/destroy.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
drivers/net/ethernet/mellanox/mlxsw/spectrum_acl.c | 102 ++++++++++-----------
1 file changed, 46 insertions(+), 56 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl.c
index 93dcd31..ead4cb8 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl.c
@@ -118,8 +118,26 @@ struct mlxsw_sp_fid *mlxsw_sp_acl_dummy_fid(struct mlxsw_sp *mlxsw_sp)
return mlxsw_sp->acl->dummy_fid;
}
+static int mlxsw_sp_acl_ruleset_bind(struct mlxsw_sp *mlxsw_sp,
+ struct mlxsw_sp_acl_ruleset *ruleset,
+ struct net_device *dev, bool ingress)
+{
+ const struct mlxsw_sp_acl_profile_ops *ops = ruleset->ht_key.ops;
+
+ return ops->ruleset_bind(mlxsw_sp, ruleset->priv, dev, ingress);
+}
+
+static void mlxsw_sp_acl_ruleset_unbind(struct mlxsw_sp *mlxsw_sp,
+ struct mlxsw_sp_acl_ruleset *ruleset)
+{
+ const struct mlxsw_sp_acl_profile_ops *ops = ruleset->ht_key.ops;
+
+ ops->ruleset_unbind(mlxsw_sp, ruleset->priv);
+}
+
static struct mlxsw_sp_acl_ruleset *
-mlxsw_sp_acl_ruleset_create(struct mlxsw_sp *mlxsw_sp,
+mlxsw_sp_acl_ruleset_create(struct mlxsw_sp *mlxsw_sp, struct net_device *dev,
+ bool ingress, u32 chain_index,
const struct mlxsw_sp_acl_profile_ops *ops)
{
struct mlxsw_sp_acl *acl = mlxsw_sp->acl;
@@ -132,6 +150,9 @@ mlxsw_sp_acl_ruleset_create(struct mlxsw_sp *mlxsw_sp,
if (!ruleset)
return ERR_PTR(-ENOMEM);
ruleset->ref_count = 1;
+ ruleset->ht_key.dev = dev;
+ ruleset->ht_key.ingress = ingress;
+ ruleset->ht_key.chain_index = chain_index;
ruleset->ht_key.ops = ops;
err = rhashtable_init(&ruleset->rule_ht, &mlxsw_sp_acl_rule_ht_params);
@@ -142,68 +163,49 @@ mlxsw_sp_acl_ruleset_create(struct mlxsw_sp *mlxsw_sp,
if (err)
goto err_ops_ruleset_add;
- return ruleset;
-
-err_ops_ruleset_add:
- rhashtable_destroy(&ruleset->rule_ht);
-err_rhashtable_init:
- kfree(ruleset);
- return ERR_PTR(err);
-}
-
-static void mlxsw_sp_acl_ruleset_destroy(struct mlxsw_sp *mlxsw_sp,
- struct mlxsw_sp_acl_ruleset *ruleset)
-{
- const struct mlxsw_sp_acl_profile_ops *ops = ruleset->ht_key.ops;
-
- ops->ruleset_del(mlxsw_sp, ruleset->priv);
- rhashtable_destroy(&ruleset->rule_ht);
- kfree(ruleset);
-}
-
-static int mlxsw_sp_acl_ruleset_bind(struct mlxsw_sp *mlxsw_sp,
- struct mlxsw_sp_acl_ruleset *ruleset,
- struct net_device *dev, bool ingress,
- u32 chain_index)
-{
- const struct mlxsw_sp_acl_profile_ops *ops = ruleset->ht_key.ops;
- struct mlxsw_sp_acl *acl = mlxsw_sp->acl;
- int err;
-
- ruleset->ht_key.dev = dev;
- ruleset->ht_key.ingress = ingress;
- ruleset->ht_key.chain_index = chain_index;
err = rhashtable_insert_fast(&acl->ruleset_ht, &ruleset->ht_node,
mlxsw_sp_acl_ruleset_ht_params);
if (err)
- return err;
- if (!ruleset->ht_key.chain_index) {
+ goto err_ht_insert;
+
+ if (!chain_index) {
/* We only need ruleset with chain index 0, the implicit one,
* to be directly bound to device. The rest of the rulesets
* are bound by "Goto action set".
*/
- err = ops->ruleset_bind(mlxsw_sp, ruleset->priv, dev, ingress);
+ err = mlxsw_sp_acl_ruleset_bind(mlxsw_sp, ruleset,
+ dev, ingress);
if (err)
- goto err_ops_ruleset_bind;
+ goto err_ruleset_bind;
}
- return 0;
-err_ops_ruleset_bind:
+ return ruleset;
+
+err_ruleset_bind:
rhashtable_remove_fast(&acl->ruleset_ht, &ruleset->ht_node,
mlxsw_sp_acl_ruleset_ht_params);
- return err;
+err_ht_insert:
+ ops->ruleset_del(mlxsw_sp, ruleset->priv);
+err_ops_ruleset_add:
+ rhashtable_destroy(&ruleset->rule_ht);
+err_rhashtable_init:
+ kfree(ruleset);
+ return ERR_PTR(err);
}
-static void mlxsw_sp_acl_ruleset_unbind(struct mlxsw_sp *mlxsw_sp,
- struct mlxsw_sp_acl_ruleset *ruleset)
+static void mlxsw_sp_acl_ruleset_destroy(struct mlxsw_sp *mlxsw_sp,
+ struct mlxsw_sp_acl_ruleset *ruleset)
{
const struct mlxsw_sp_acl_profile_ops *ops = ruleset->ht_key.ops;
struct mlxsw_sp_acl *acl = mlxsw_sp->acl;
if (!ruleset->ht_key.chain_index)
- ops->ruleset_unbind(mlxsw_sp, ruleset->priv);
+ mlxsw_sp_acl_ruleset_unbind(mlxsw_sp, ruleset);
rhashtable_remove_fast(&acl->ruleset_ht, &ruleset->ht_node,
mlxsw_sp_acl_ruleset_ht_params);
+ ops->ruleset_del(mlxsw_sp, ruleset->priv);
+ rhashtable_destroy(&ruleset->rule_ht);
+ kfree(ruleset);
}
static void mlxsw_sp_acl_ruleset_ref_inc(struct mlxsw_sp_acl_ruleset *ruleset)
@@ -216,7 +218,6 @@ static void mlxsw_sp_acl_ruleset_ref_dec(struct mlxsw_sp *mlxsw_sp,
{
if (--ruleset->ref_count)
return;
- mlxsw_sp_acl_ruleset_unbind(mlxsw_sp, ruleset);
mlxsw_sp_acl_ruleset_destroy(mlxsw_sp, ruleset);
}
@@ -263,7 +264,6 @@ mlxsw_sp_acl_ruleset_get(struct mlxsw_sp *mlxsw_sp, struct net_device *dev,
const struct mlxsw_sp_acl_profile_ops *ops;
struct mlxsw_sp_acl *acl = mlxsw_sp->acl;
struct mlxsw_sp_acl_ruleset *ruleset;
- int err;
ops = acl->ops->profile_ops(mlxsw_sp, profile);
if (!ops)
@@ -275,18 +275,8 @@ mlxsw_sp_acl_ruleset_get(struct mlxsw_sp *mlxsw_sp, struct net_device *dev,
mlxsw_sp_acl_ruleset_ref_inc(ruleset);
return ruleset;
}
- ruleset = mlxsw_sp_acl_ruleset_create(mlxsw_sp, ops);
- if (IS_ERR(ruleset))
- return ruleset;
- err = mlxsw_sp_acl_ruleset_bind(mlxsw_sp, ruleset, dev,
- ingress, chain_index);
- if (err)
- goto err_ruleset_bind;
- return ruleset;
-
-err_ruleset_bind:
- mlxsw_sp_acl_ruleset_destroy(mlxsw_sp, ruleset);
- return ERR_PTR(err);
+ return mlxsw_sp_acl_ruleset_create(mlxsw_sp, dev, ingress,
+ chain_index, ops);
}
void mlxsw_sp_acl_ruleset_put(struct mlxsw_sp *mlxsw_sp,
--
2.9.5
^ permalink raw reply related
* [patch net-next v6 09/11] mlxsw: spectrum_acl: Don't store netdev and ingress for ruleset unbind
From: Jiri Pirko @ 2018-01-05 23:09 UTC (permalink / raw)
To: netdev
Cc: davem, jhs, xiyou.wangcong, mlxsw, andrew, vivien.didelot,
f.fainelli, michael.chan, ganeshgr, saeedm, matanb, leonro,
idosch, jakub.kicinski, simon.horman, pieter.jansenvanvuuren,
john.hurley, alexander.h.duyck, ogerlitz, john.fastabend, daniel,
dsahern
In-Reply-To: <20180105230929.5645-1-jiri@resnulli.us>
From: Jiri Pirko <jiri@mellanox.com>
Instead, pass netdev and ingress flag to ruleset unbind op.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
---
drivers/net/ethernet/mellanox/mlxsw/spectrum.h | 3 +-
drivers/net/ethernet/mellanox/mlxsw/spectrum_acl.c | 9 ++++--
.../ethernet/mellanox/mlxsw/spectrum_acl_tcam.c | 33 +++++++++++-----------
3 files changed, 24 insertions(+), 21 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum.h b/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
index a0adcd8..523e64e 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum.h
@@ -477,7 +477,8 @@ struct mlxsw_sp_acl_profile_ops {
void (*ruleset_del)(struct mlxsw_sp *mlxsw_sp, void *ruleset_priv);
int (*ruleset_bind)(struct mlxsw_sp *mlxsw_sp, void *ruleset_priv,
struct net_device *dev, bool ingress);
- void (*ruleset_unbind)(struct mlxsw_sp *mlxsw_sp, void *ruleset_priv);
+ void (*ruleset_unbind)(struct mlxsw_sp *mlxsw_sp, void *ruleset_priv,
+ struct net_device *dev, bool ingress);
u16 (*ruleset_group_id)(void *ruleset_priv);
size_t rule_priv_size;
int (*rule_add)(struct mlxsw_sp *mlxsw_sp,
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl.c
index ead4cb8..7fb41a4 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl.c
@@ -128,11 +128,12 @@ static int mlxsw_sp_acl_ruleset_bind(struct mlxsw_sp *mlxsw_sp,
}
static void mlxsw_sp_acl_ruleset_unbind(struct mlxsw_sp *mlxsw_sp,
- struct mlxsw_sp_acl_ruleset *ruleset)
+ struct mlxsw_sp_acl_ruleset *ruleset,
+ struct net_device *dev, bool ingress)
{
const struct mlxsw_sp_acl_profile_ops *ops = ruleset->ht_key.ops;
- ops->ruleset_unbind(mlxsw_sp, ruleset->priv);
+ ops->ruleset_unbind(mlxsw_sp, ruleset->priv, dev, ingress);
}
static struct mlxsw_sp_acl_ruleset *
@@ -200,7 +201,9 @@ static void mlxsw_sp_acl_ruleset_destroy(struct mlxsw_sp *mlxsw_sp,
struct mlxsw_sp_acl *acl = mlxsw_sp->acl;
if (!ruleset->ht_key.chain_index)
- mlxsw_sp_acl_ruleset_unbind(mlxsw_sp, ruleset);
+ mlxsw_sp_acl_ruleset_unbind(mlxsw_sp, ruleset,
+ ruleset->ht_key.dev,
+ ruleset->ht_key.ingress);
rhashtable_remove_fast(&acl->ruleset_ht, &ruleset->ht_node,
mlxsw_sp_acl_ruleset_ht_params);
ops->ruleset_del(mlxsw_sp, ruleset->priv);
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_tcam.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_tcam.c
index 7e8284b..50b2f9a 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_tcam.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_acl_tcam.c
@@ -154,10 +154,6 @@ struct mlxsw_sp_acl_tcam_group {
struct list_head region_list;
unsigned int region_count;
struct rhashtable chunk_ht;
- struct {
- u16 local_port;
- bool ingress;
- } bound;
struct mlxsw_sp_acl_tcam_group_ops *ops;
const struct mlxsw_sp_acl_tcam_pattern *patterns;
unsigned int patterns_count;
@@ -271,26 +267,28 @@ mlxsw_sp_acl_tcam_group_bind(struct mlxsw_sp *mlxsw_sp,
return -EINVAL;
mlxsw_sp_port = netdev_priv(dev);
- group->bound.local_port = mlxsw_sp_port->local_port;
- group->bound.ingress = ingress;
- mlxsw_reg_ppbt_pack(ppbt_pl,
- group->bound.ingress ? MLXSW_REG_PXBT_E_IACL :
- MLXSW_REG_PXBT_E_EACL,
- MLXSW_REG_PXBT_OP_BIND, group->bound.local_port,
+ mlxsw_reg_ppbt_pack(ppbt_pl, ingress ? MLXSW_REG_PXBT_E_IACL :
+ MLXSW_REG_PXBT_E_EACL,
+ MLXSW_REG_PXBT_OP_BIND, mlxsw_sp_port->local_port,
group->id);
return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(ppbt), ppbt_pl);
}
static void
mlxsw_sp_acl_tcam_group_unbind(struct mlxsw_sp *mlxsw_sp,
- struct mlxsw_sp_acl_tcam_group *group)
+ struct mlxsw_sp_acl_tcam_group *group,
+ struct net_device *dev, bool ingress)
{
+ struct mlxsw_sp_port *mlxsw_sp_port;
char ppbt_pl[MLXSW_REG_PPBT_LEN];
- mlxsw_reg_ppbt_pack(ppbt_pl,
- group->bound.ingress ? MLXSW_REG_PXBT_E_IACL :
- MLXSW_REG_PXBT_E_EACL,
- MLXSW_REG_PXBT_OP_UNBIND, group->bound.local_port,
+ if (WARN_ON(!mlxsw_sp_port_dev_check(dev)))
+ return;
+
+ mlxsw_sp_port = netdev_priv(dev);
+ mlxsw_reg_ppbt_pack(ppbt_pl, ingress ? MLXSW_REG_PXBT_E_IACL :
+ MLXSW_REG_PXBT_E_EACL,
+ MLXSW_REG_PXBT_OP_UNBIND, mlxsw_sp_port->local_port,
group->id);
mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(ppbt), ppbt_pl);
}
@@ -1066,11 +1064,12 @@ mlxsw_sp_acl_tcam_flower_ruleset_bind(struct mlxsw_sp *mlxsw_sp,
static void
mlxsw_sp_acl_tcam_flower_ruleset_unbind(struct mlxsw_sp *mlxsw_sp,
- void *ruleset_priv)
+ void *ruleset_priv,
+ struct net_device *dev, bool ingress)
{
struct mlxsw_sp_acl_tcam_flower_ruleset *ruleset = ruleset_priv;
- mlxsw_sp_acl_tcam_group_unbind(mlxsw_sp, &ruleset->group);
+ mlxsw_sp_acl_tcam_group_unbind(mlxsw_sp, &ruleset->group, dev, ingress);
}
static u16
--
2.9.5
^ permalink raw reply related
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox