* Pull request: sfc-next 2013-12-16
@ 2013-12-16 21:29 Ben Hutchings
2013-12-16 21:31 ` [PATCH net-next 01/11] sfc: Add more information to many warnings using WARN() and netdev_WARN() Ben Hutchings
` (11 more replies)
0 siblings, 12 replies; 13+ messages in thread
From: Ben Hutchings @ 2013-12-16 21:29 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linux-net-drivers
[-- Attachment #1: Type: text/plain, Size: 2710 bytes --]
The following changes since commit bbbe7149bf0bd14432373de39315505e7a133acc:
sfc: Remove unnecessary condition for processing the TX timestamp queue (2013-12-12 22:07:18 +0000)
are available in the git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/bwh/sfc-next.git for-davem
for you to fetch changes up to a0bc3487e431eaea381c2c882a41274e9a02f028:
sfc: Fix RX drop filters for EF10 (2013-12-16 19:01:27 +0000)
Miscellaneous changes for 3.14:
1. Add more information to some WARN messages.
2. Refactor pushing of RSS configuration, from Andrew Rybchenko.
3. Refactor handling of automatic (device address list) vs manual (RX
NFC) MAC filters.
4. Implement clearing of manual RX filters on EF10 when ntuple offload
is disabled.
5. Remove definitions that are unused since the RX buffer allocation
changes, from Andrew Rybchenko.
6. Improve naming of some statistics, from Shradha Shah.
7. Add statistics for PTP support code.
8. Fix insertion of RX drop filters on EF10.
Ben.
----------------------------------------------------------------
Andrew Rybchenko (3):
sfc: Change efx_nic_type::rx_push_indir_table to push hash key as well
sfc: remove unused 'refcnt' from efx_rx_page_state
sfc: remove unused 'enum efx_rx_alloc_method'
Ben Hutchings (7):
sfc: Add more information to many warnings using WARN() and netdev_WARN()
sfc: Change priority and flags for automatic MAC filters
sfc: Don't refer to 'stack' in filter implementation
sfc: Allow filter removal only with exactly matching priority
sfc: Implement efx_nic_type::filter_clear_rx operation for EF10
sfc: Add PTP counters to ethtool stats
sfc: Fix RX drop filters for EF10
Shradha Shah (1):
sfc: Changed the statistic name emerg_{fetch,wait} to hlb_{fetch,wait}
drivers/net/ethernet/sfc/ef10.c | 251 ++++++++++++++++++++--------------
drivers/net/ethernet/sfc/efx.c | 2 +-
drivers/net/ethernet/sfc/efx.h | 11 --
drivers/net/ethernet/sfc/ethtool.c | 10 +-
drivers/net/ethernet/sfc/falcon.c | 26 +++-
drivers/net/ethernet/sfc/farch.c | 46 +++----
drivers/net/ethernet/sfc/filter.h | 17 ++-
drivers/net/ethernet/sfc/net_driver.h | 20 +--
drivers/net/ethernet/sfc/nic.h | 14 +-
drivers/net/ethernet/sfc/ptp.c | 126 +++++++++++++++--
drivers/net/ethernet/sfc/siena.c | 45 +++---
11 files changed, 362 insertions(+), 206 deletions(-)
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 490 bytes --]
^ permalink raw reply [flat|nested] 13+ messages in thread
* [PATCH net-next 01/11] sfc: Add more information to many warnings using WARN() and netdev_WARN()
2013-12-16 21:29 Pull request: sfc-next 2013-12-16 Ben Hutchings
@ 2013-12-16 21:31 ` Ben Hutchings
2013-12-16 21:31 ` [PATCH net-next 02/11] sfc: Change efx_nic_type::rx_push_indir_table to push hash key as well Ben Hutchings
` (10 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: Ben Hutchings @ 2013-12-16 21:31 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linux-net-drivers
In case of certain hardware and firmware errors it can be useful to
have more context than just the file and line number.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
---
drivers/net/ethernet/sfc/ef10.c | 30 ++++++++++++++++++++----------
1 file changed, 20 insertions(+), 10 deletions(-)
diff --git a/drivers/net/ethernet/sfc/ef10.c b/drivers/net/ethernet/sfc/ef10.c
index dfff8e7a55e2..9ec6072896c2 100644
--- a/drivers/net/ethernet/sfc/ef10.c
+++ b/drivers/net/ethernet/sfc/ef10.c
@@ -1252,7 +1252,8 @@ static void efx_ef10_tx_init(struct efx_tx_queue *tx_queue)
return;
fail:
- WARN_ON(true);
+ netdev_WARN(efx->net_dev, "failed to initialise TXQ %d\n",
+ tx_queue->queue);
}
static void efx_ef10_tx_fini(struct efx_tx_queue *tx_queue)
@@ -1492,9 +1493,9 @@ static void efx_ef10_rx_init(struct efx_rx_queue *rx_queue)
rc = efx_mcdi_rpc(efx, MC_CMD_INIT_RXQ, inbuf, inlen,
outbuf, sizeof(outbuf), &outlen);
- WARN_ON(rc);
-
- return;
+ if (rc)
+ netdev_WARN(efx->net_dev, "failed to initialise RXQ %d\n",
+ efx_rx_queue_index(rx_queue));
}
static void efx_ef10_rx_fini(struct efx_rx_queue *rx_queue)
@@ -1718,8 +1719,6 @@ static void efx_ef10_handle_rx_abort(struct efx_rx_queue *rx_queue)
{
unsigned int rx_desc_ptr;
- WARN_ON(rx_queue->scatter_n == 0);
-
netif_dbg(rx_queue->efx, hw, rx_queue->efx->net_dev,
"scattered RX aborted (dropping %u buffers)\n",
rx_queue->scatter_n);
@@ -1755,7 +1754,10 @@ static int efx_ef10_handle_rx_event(struct efx_channel *channel,
rx_l4_class = EFX_QWORD_FIELD(*event, ESF_DZ_RX_L4_CLASS);
rx_cont = EFX_QWORD_FIELD(*event, ESF_DZ_RX_CONT);
- WARN_ON(EFX_QWORD_FIELD(*event, ESF_DZ_RX_DROP_EVENT));
+ if (EFX_QWORD_FIELD(*event, ESF_DZ_RX_DROP_EVENT))
+ netdev_WARN(efx->net_dev, "saw RX_DROP_EVENT: event="
+ EFX_QWORD_FMT "\n",
+ EFX_QWORD_VAL(*event));
rx_queue = efx_channel_get_rx_queue(channel);
@@ -1770,7 +1772,12 @@ static int efx_ef10_handle_rx_event(struct efx_channel *channel,
/* detect rx abort */
if (unlikely(n_descs == rx_queue->scatter_n)) {
- WARN_ON(rx_bytes != 0);
+ if (rx_queue->scatter_n == 0 || rx_bytes != 0)
+ netdev_WARN(efx->net_dev,
+ "invalid RX abort: scatter_n=%u event="
+ EFX_QWORD_FMT "\n",
+ rx_queue->scatter_n,
+ EFX_QWORD_VAL(*event));
efx_ef10_handle_rx_abort(rx_queue);
return 0;
}
@@ -3035,8 +3042,11 @@ static void efx_ef10_filter_table_remove(struct efx_nic *efx)
table->entry[filter_idx].handle);
rc = efx_mcdi_rpc(efx, MC_CMD_FILTER_OP, inbuf, sizeof(inbuf),
NULL, 0, NULL);
-
- WARN_ON(rc != 0);
+ if (rc)
+ netdev_WARN(efx->net_dev,
+ "filter_idx=%#x handle=%#llx\n",
+ filter_idx,
+ table->entry[filter_idx].handle);
kfree(spec);
}
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply related [flat|nested] 13+ messages in thread
* [PATCH net-next 02/11] sfc: Change efx_nic_type::rx_push_indir_table to push hash key as well
2013-12-16 21:29 Pull request: sfc-next 2013-12-16 Ben Hutchings
2013-12-16 21:31 ` [PATCH net-next 01/11] sfc: Add more information to many warnings using WARN() and netdev_WARN() Ben Hutchings
@ 2013-12-16 21:31 ` Ben Hutchings
2013-12-16 21:32 ` [PATCH net-next 03/11] sfc: Change priority and flags for automatic MAC filters Ben Hutchings
` (9 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: Ben Hutchings @ 2013-12-16 21:31 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linux-net-drivers, Andrew Rybchenko
From: Andrew Rybchenko <Andrew.Rybchenko@oktetlabs.ru>
The EF10 implementation already does this, and it makes more logical
sense to group the RSS hash key and indirection table together.
Rename the operation to rx_push_rss_config.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
---
drivers/net/ethernet/sfc/ef10.c | 10 ++++----
drivers/net/ethernet/sfc/ethtool.c | 2 +-
drivers/net/ethernet/sfc/falcon.c | 26 ++++++++++++++++----
drivers/net/ethernet/sfc/farch.c | 5 +---
drivers/net/ethernet/sfc/net_driver.h | 4 ++--
drivers/net/ethernet/sfc/nic.h | 4 ----
drivers/net/ethernet/sfc/siena.c | 45 +++++++++++++++++++++--------------
7 files changed, 57 insertions(+), 39 deletions(-)
diff --git a/drivers/net/ethernet/sfc/ef10.c b/drivers/net/ethernet/sfc/ef10.c
index 9ec6072896c2..137c46b4b405 100644
--- a/drivers/net/ethernet/sfc/ef10.c
+++ b/drivers/net/ethernet/sfc/ef10.c
@@ -77,7 +77,7 @@ struct efx_ef10_filter_table {
/* An arbitrary search limit for the software hash table */
#define EFX_EF10_FILTER_SEARCH_LIMIT 200
-static void efx_ef10_rx_push_indir_table(struct efx_nic *efx);
+static void efx_ef10_rx_push_rss_config(struct efx_nic *efx);
static void efx_ef10_rx_free_indir_table(struct efx_nic *efx);
static void efx_ef10_filter_table_remove(struct efx_nic *efx);
@@ -679,7 +679,7 @@ static int efx_ef10_init_nic(struct efx_nic *efx)
nic_data->must_restore_piobufs = false;
}
- efx_ef10_rx_push_indir_table(efx);
+ efx_ef10_rx_push_rss_config(efx);
return 0;
}
@@ -1420,12 +1420,12 @@ static void efx_ef10_rx_free_indir_table(struct efx_nic *efx)
nic_data->rx_rss_context = EFX_EF10_RSS_CONTEXT_INVALID;
}
-static void efx_ef10_rx_push_indir_table(struct efx_nic *efx)
+static void efx_ef10_rx_push_rss_config(struct efx_nic *efx)
{
struct efx_ef10_nic_data *nic_data = efx->nic_data;
int rc;
- netif_dbg(efx, drv, efx->net_dev, "pushing RX indirection table\n");
+ netif_dbg(efx, drv, efx->net_dev, "pushing RSS config\n");
if (nic_data->rx_rss_context == EFX_EF10_RSS_CONTEXT_INVALID) {
rc = efx_ef10_alloc_rss_context(efx, &nic_data->rx_rss_context);
@@ -3574,7 +3574,7 @@ const struct efx_nic_type efx_hunt_a0_nic_type = {
.tx_init = efx_ef10_tx_init,
.tx_remove = efx_ef10_tx_remove,
.tx_write = efx_ef10_tx_write,
- .rx_push_indir_table = efx_ef10_rx_push_indir_table,
+ .rx_push_rss_config = efx_ef10_rx_push_rss_config,
.rx_probe = efx_ef10_rx_probe,
.rx_init = efx_ef10_rx_init,
.rx_remove = efx_ef10_rx_remove,
diff --git a/drivers/net/ethernet/sfc/ethtool.c b/drivers/net/ethernet/sfc/ethtool.c
index fb8993806167..bb2af8045281 100644
--- a/drivers/net/ethernet/sfc/ethtool.c
+++ b/drivers/net/ethernet/sfc/ethtool.c
@@ -1034,7 +1034,7 @@ static int efx_ethtool_set_rxfh_indir(struct net_device *net_dev,
struct efx_nic *efx = netdev_priv(net_dev);
memcpy(efx->rx_indir_table, indir, sizeof(efx->rx_indir_table));
- efx_nic_push_rx_indir_table(efx);
+ efx->type->rx_push_rss_config(efx);
return 0;
}
diff --git a/drivers/net/ethernet/sfc/falcon.c b/drivers/net/ethernet/sfc/falcon.c
index 76699f4e6e04..5c6d63cfc394 100644
--- a/drivers/net/ethernet/sfc/falcon.c
+++ b/drivers/net/ethernet/sfc/falcon.c
@@ -469,6 +469,24 @@ static irqreturn_t falcon_legacy_interrupt_a1(int irq, void *dev_id)
}
/**************************************************************************
*
+ * RSS
+ *
+ **************************************************************************
+ */
+
+static void falcon_b0_rx_push_rss_config(struct efx_nic *efx)
+{
+ efx_oword_t temp;
+
+ /* Set hash key for IPv4 */
+ memcpy(&temp, efx->rx_hash_key, sizeof(temp));
+ efx_writeo(efx, &temp, FR_BZ_RX_RSS_TKEY);
+
+ efx_farch_rx_push_indir_table(efx);
+}
+
+/**************************************************************************
+ *
* EEPROM/flash
*
**************************************************************************
@@ -2484,9 +2502,7 @@ static int falcon_init_nic(struct efx_nic *efx)
falcon_init_rx_cfg(efx);
if (efx_nic_rev(efx) >= EFX_REV_FALCON_B0) {
- /* Set hash key for IPv4 */
- memcpy(&temp, efx->rx_hash_key, sizeof(temp));
- efx_writeo(efx, &temp, FR_BZ_RX_RSS_TKEY);
+ falcon_b0_rx_push_rss_config(efx);
/* Set destination of both TX and RX Flush events */
EFX_POPULATE_OWORD_1(temp, FRF_BZ_FLS_EVQ_ID, 0);
@@ -2703,7 +2719,7 @@ const struct efx_nic_type falcon_a1_nic_type = {
.tx_init = efx_farch_tx_init,
.tx_remove = efx_farch_tx_remove,
.tx_write = efx_farch_tx_write,
- .rx_push_indir_table = efx_farch_rx_push_indir_table,
+ .rx_push_rss_config = efx_port_dummy_op_void,
.rx_probe = efx_farch_rx_probe,
.rx_init = efx_farch_rx_init,
.rx_remove = efx_farch_rx_remove,
@@ -2798,7 +2814,7 @@ const struct efx_nic_type falcon_b0_nic_type = {
.tx_init = efx_farch_tx_init,
.tx_remove = efx_farch_tx_remove,
.tx_write = efx_farch_tx_write,
- .rx_push_indir_table = efx_farch_rx_push_indir_table,
+ .rx_push_rss_config = falcon_b0_rx_push_rss_config,
.rx_probe = efx_farch_rx_probe,
.rx_init = efx_farch_rx_init,
.rx_remove = efx_farch_rx_remove,
diff --git a/drivers/net/ethernet/sfc/farch.c b/drivers/net/ethernet/sfc/farch.c
index 984e85ee76f6..4c64ad7c9200 100644
--- a/drivers/net/ethernet/sfc/farch.c
+++ b/drivers/net/ethernet/sfc/farch.c
@@ -1618,8 +1618,7 @@ void efx_farch_rx_push_indir_table(struct efx_nic *efx)
size_t i = 0;
efx_dword_t dword;
- if (efx_nic_rev(efx) < EFX_REV_FALCON_B0)
- return;
+ BUG_ON(efx_nic_rev(efx) < EFX_REV_FALCON_B0);
BUILD_BUG_ON(ARRAY_SIZE(efx->rx_indir_table) !=
FR_BZ_RX_INDIRECTION_TBL_ROWS);
@@ -1745,8 +1744,6 @@ void efx_farch_init_common(struct efx_nic *efx)
EFX_INVERT_OWORD(temp);
efx_writeo(efx, &temp, FR_AZ_FATAL_INTR_KER);
- efx_farch_rx_push_indir_table(efx);
-
/* Disable the ugly timer-based TX DMA backoff and allow TX DMA to be
* controlled by the RX FIFO fill level. Set arbitration to one pkt/Q.
*/
diff --git a/drivers/net/ethernet/sfc/net_driver.h b/drivers/net/ethernet/sfc/net_driver.h
index 653b8782c956..1b3c4e5207c5 100644
--- a/drivers/net/ethernet/sfc/net_driver.h
+++ b/drivers/net/ethernet/sfc/net_driver.h
@@ -1024,7 +1024,7 @@ struct efx_mtd_partition {
* @tx_init: Initialise TX queue on the NIC
* @tx_remove: Free resources for TX queue
* @tx_write: Write TX descriptors and doorbell
- * @rx_push_indir_table: Write RSS indirection table to the NIC
+ * @rx_push_rss_config: Write RSS hash key and indirection table to the NIC
* @rx_probe: Allocate resources for RX queue
* @rx_init: Initialise RX queue on the NIC
* @rx_remove: Free resources for RX queue
@@ -1141,7 +1141,7 @@ struct efx_nic_type {
void (*tx_init)(struct efx_tx_queue *tx_queue);
void (*tx_remove)(struct efx_tx_queue *tx_queue);
void (*tx_write)(struct efx_tx_queue *tx_queue);
- void (*rx_push_indir_table)(struct efx_nic *efx);
+ void (*rx_push_rss_config)(struct efx_nic *efx);
int (*rx_probe)(struct efx_rx_queue *rx_queue);
void (*rx_init)(struct efx_rx_queue *rx_queue);
void (*rx_remove)(struct efx_rx_queue *rx_queue);
diff --git a/drivers/net/ethernet/sfc/nic.h b/drivers/net/ethernet/sfc/nic.h
index 5d9e2dc121f7..c665ff6c8012 100644
--- a/drivers/net/ethernet/sfc/nic.h
+++ b/drivers/net/ethernet/sfc/nic.h
@@ -762,10 +762,6 @@ int falcon_reset_xaui(struct efx_nic *efx);
void efx_farch_dimension_resources(struct efx_nic *efx, unsigned sram_lim_qw);
void efx_farch_init_common(struct efx_nic *efx);
void efx_ef10_handle_drain_event(struct efx_nic *efx);
-static inline void efx_nic_push_rx_indir_table(struct efx_nic *efx)
-{
- efx->type->rx_push_indir_table(efx);
-}
void efx_farch_rx_push_indir_table(struct efx_nic *efx);
int efx_nic_alloc_buffer(struct efx_nic *efx, struct efx_buffer *buffer,
diff --git a/drivers/net/ethernet/sfc/siena.c b/drivers/net/ethernet/sfc/siena.c
index f65db356fe09..23f3a6f7737a 100644
--- a/drivers/net/ethernet/sfc/siena.c
+++ b/drivers/net/ethernet/sfc/siena.c
@@ -321,6 +321,31 @@ fail1:
return rc;
}
+static void siena_rx_push_rss_config(struct efx_nic *efx)
+{
+ efx_oword_t temp;
+
+ /* Set hash key for IPv4 */
+ memcpy(&temp, efx->rx_hash_key, sizeof(temp));
+ efx_writeo(efx, &temp, FR_BZ_RX_RSS_TKEY);
+
+ /* Enable IPv6 RSS */
+ BUILD_BUG_ON(sizeof(efx->rx_hash_key) <
+ 2 * sizeof(temp) + FRF_CZ_RX_RSS_IPV6_TKEY_HI_WIDTH / 8 ||
+ FRF_CZ_RX_RSS_IPV6_TKEY_HI_LBN != 0);
+ memcpy(&temp, efx->rx_hash_key, sizeof(temp));
+ efx_writeo(efx, &temp, FR_CZ_RX_RSS_IPV6_REG1);
+ memcpy(&temp, efx->rx_hash_key + sizeof(temp), sizeof(temp));
+ efx_writeo(efx, &temp, FR_CZ_RX_RSS_IPV6_REG2);
+ EFX_POPULATE_OWORD_2(temp, FRF_CZ_RX_RSS_IPV6_THASH_ENABLE, 1,
+ FRF_CZ_RX_RSS_IPV6_IP_THASH_ENABLE, 1);
+ memcpy(&temp, efx->rx_hash_key + 2 * sizeof(temp),
+ FRF_CZ_RX_RSS_IPV6_TKEY_HI_WIDTH / 8);
+ efx_writeo(efx, &temp, FR_CZ_RX_RSS_IPV6_REG3);
+
+ efx_farch_rx_push_indir_table(efx);
+}
+
/* This call performs hardware-specific global initialisation, such as
* defining the descriptor cache sizes and number of RSS channels.
* It does not set up any buffers, descriptor rings or event queues.
@@ -361,23 +386,7 @@ static int siena_init_nic(struct efx_nic *efx)
EFX_RX_USR_BUF_SIZE >> 5);
efx_writeo(efx, &temp, FR_AZ_RX_CFG);
- /* Set hash key for IPv4 */
- memcpy(&temp, efx->rx_hash_key, sizeof(temp));
- efx_writeo(efx, &temp, FR_BZ_RX_RSS_TKEY);
-
- /* Enable IPv6 RSS */
- BUILD_BUG_ON(sizeof(efx->rx_hash_key) <
- 2 * sizeof(temp) + FRF_CZ_RX_RSS_IPV6_TKEY_HI_WIDTH / 8 ||
- FRF_CZ_RX_RSS_IPV6_TKEY_HI_LBN != 0);
- memcpy(&temp, efx->rx_hash_key, sizeof(temp));
- efx_writeo(efx, &temp, FR_CZ_RX_RSS_IPV6_REG1);
- memcpy(&temp, efx->rx_hash_key + sizeof(temp), sizeof(temp));
- efx_writeo(efx, &temp, FR_CZ_RX_RSS_IPV6_REG2);
- EFX_POPULATE_OWORD_2(temp, FRF_CZ_RX_RSS_IPV6_THASH_ENABLE, 1,
- FRF_CZ_RX_RSS_IPV6_IP_THASH_ENABLE, 1);
- memcpy(&temp, efx->rx_hash_key + 2 * sizeof(temp),
- FRF_CZ_RX_RSS_IPV6_TKEY_HI_WIDTH / 8);
- efx_writeo(efx, &temp, FR_CZ_RX_RSS_IPV6_REG3);
+ siena_rx_push_rss_config(efx);
/* Enable event logging */
rc = efx_mcdi_log_ctrl(efx, true, false, 0);
@@ -940,7 +949,7 @@ const struct efx_nic_type siena_a0_nic_type = {
.tx_init = efx_farch_tx_init,
.tx_remove = efx_farch_tx_remove,
.tx_write = efx_farch_tx_write,
- .rx_push_indir_table = efx_farch_rx_push_indir_table,
+ .rx_push_rss_config = siena_rx_push_rss_config,
.rx_probe = efx_farch_rx_probe,
.rx_init = efx_farch_rx_init,
.rx_remove = efx_farch_rx_remove,
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply related [flat|nested] 13+ messages in thread
* [PATCH net-next 03/11] sfc: Change priority and flags for automatic MAC filters
2013-12-16 21:29 Pull request: sfc-next 2013-12-16 Ben Hutchings
2013-12-16 21:31 ` [PATCH net-next 01/11] sfc: Add more information to many warnings using WARN() and netdev_WARN() Ben Hutchings
2013-12-16 21:31 ` [PATCH net-next 02/11] sfc: Change efx_nic_type::rx_push_indir_table to push hash key as well Ben Hutchings
@ 2013-12-16 21:32 ` Ben Hutchings
2013-12-16 21:32 ` [PATCH net-next 04/11] sfc: Don't refer to 'stack' in filter implementation Ben Hutchings
` (8 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: Ben Hutchings @ 2013-12-16 21:32 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linux-net-drivers
MAC filters inserted automatically by the driver, based on the device
address list (EF10) or no-match filters (Siena), should be overridable
at MANUAL or REQUIRED priority. Currently they themselves have
REQUIRED priority and this requires some odd special-casing.
We also can't reliably tell whether such a MAC filter has or has
not been overridden. We just remember that it is wanted by the
stack (RX_STACK flag).
Add another priority level, AUTO, between HINT and MANUAL, and
use this for the automatic filters while they have not been
overridden. Remove the RX_STACK flag. Add an RX_OVER_AUTO
flag which is set only when an AUTO filter has been overridden
(or was requested to be inserted while a higher-priority filter
existed).
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
---
drivers/net/ethernet/sfc/ef10.c | 73 ++++++++++++++++++++++-----------------
drivers/net/ethernet/sfc/efx.h | 3 ++
drivers/net/ethernet/sfc/farch.c | 28 +++++++--------
drivers/net/ethernet/sfc/filter.h | 17 +++++----
4 files changed, 67 insertions(+), 54 deletions(-)
diff --git a/drivers/net/ethernet/sfc/ef10.c b/drivers/net/ethernet/sfc/ef10.c
index 137c46b4b405..e5eeac9063d2 100644
--- a/drivers/net/ethernet/sfc/ef10.c
+++ b/drivers/net/ethernet/sfc/ef10.c
@@ -2341,10 +2341,7 @@ static s32 efx_ef10_filter_insert(struct efx_nic *efx,
EFX_EF10_FILTER_FLAG_BUSY)
break;
if (spec->priority < saved_spec->priority &&
- !(saved_spec->priority ==
- EFX_FILTER_PRI_REQUIRED &&
- saved_spec->flags &
- EFX_FILTER_FLAG_RX_STACK)) {
+ spec->priority != EFX_FILTER_PRI_AUTO) {
rc = -EPERM;
goto out_unlock;
}
@@ -2398,9 +2395,11 @@ found:
*/
saved_spec = efx_ef10_filter_entry_spec(table, ins_index);
if (saved_spec) {
- if (spec->flags & EFX_FILTER_FLAG_RX_STACK) {
+ if (spec->priority == EFX_FILTER_PRI_AUTO &&
+ saved_spec->priority >= EFX_FILTER_PRI_AUTO) {
/* Just make sure it won't be removed */
- saved_spec->flags |= EFX_FILTER_FLAG_RX_STACK;
+ if (saved_spec->priority > EFX_FILTER_PRI_AUTO)
+ saved_spec->flags |= EFX_FILTER_FLAG_RX_OVER_AUTO;
table->entry[ins_index].spec &=
~EFX_EF10_FILTER_FLAG_STACK_OLD;
rc = ins_index;
@@ -2442,8 +2441,11 @@ found:
if (rc == 0) {
if (replacing) {
/* Update the fields that may differ */
+ if (saved_spec->priority == EFX_FILTER_PRI_AUTO)
+ saved_spec->flags |=
+ EFX_FILTER_FLAG_RX_OVER_AUTO;
saved_spec->priority = spec->priority;
- saved_spec->flags &= EFX_FILTER_FLAG_RX_STACK;
+ saved_spec->flags &= EFX_FILTER_FLAG_RX_OVER_AUTO;
saved_spec->flags |= spec->flags;
saved_spec->rss_context = spec->rss_context;
saved_spec->dmaq_id = spec->dmaq_id;
@@ -2542,26 +2544,41 @@ static int efx_ef10_filter_remove_internal(struct efx_nic *efx,
spin_unlock_bh(&efx->filter_lock);
schedule();
}
+
spec = efx_ef10_filter_entry_spec(table, filter_idx);
- if (!spec || spec->priority > priority ||
+ if (!spec ||
(!stack_requested &&
efx_ef10_filter_rx_match_pri(table, spec->match_flags) !=
filter_id / HUNT_FILTER_TBL_ROWS)) {
rc = -ENOENT;
goto out_unlock;
}
+
+ if (spec->flags & EFX_FILTER_FLAG_RX_OVER_AUTO &&
+ priority == EFX_FILTER_PRI_AUTO) {
+ /* Just remove flags */
+ spec->flags &= ~EFX_FILTER_FLAG_RX_OVER_AUTO;
+ table->entry[filter_idx].spec &= ~EFX_EF10_FILTER_FLAG_STACK_OLD;
+ rc = 0;
+ goto out_unlock;
+ }
+
+ if (spec->priority > priority) {
+ rc = -ENOENT;
+ goto out_unlock;
+ }
+
table->entry[filter_idx].spec |= EFX_EF10_FILTER_FLAG_BUSY;
spin_unlock_bh(&efx->filter_lock);
- if (spec->flags & EFX_FILTER_FLAG_RX_STACK && !stack_requested) {
+ if (spec->flags & EFX_FILTER_FLAG_RX_OVER_AUTO) {
/* Reset steering of a stack-owned filter */
struct efx_filter_spec new_spec = *spec;
- new_spec.priority = EFX_FILTER_PRI_REQUIRED;
+ new_spec.priority = EFX_FILTER_PRI_AUTO;
new_spec.flags = (EFX_FILTER_FLAG_RX |
- EFX_FILTER_FLAG_RX_RSS |
- EFX_FILTER_FLAG_RX_STACK);
+ EFX_FILTER_FLAG_RX_RSS);
new_spec.dmaq_id = 0;
new_spec.rss_context = EFX_FILTER_RSS_CONTEXT_DEFAULT;
rc = efx_ef10_filter_push(efx, &new_spec,
@@ -2589,6 +2606,7 @@ static int efx_ef10_filter_remove_internal(struct efx_nic *efx,
efx_ef10_filter_set_entry(table, filter_idx, NULL, 0);
}
}
+
table->entry[filter_idx].spec &= ~EFX_EF10_FILTER_FLAG_BUSY;
wake_up_all(&table->waitq);
out_unlock:
@@ -2731,8 +2749,6 @@ static s32 efx_ef10_filter_rfs_insert(struct efx_nic *efx,
rc = -EBUSY;
goto fail_unlock;
}
- EFX_WARN_ON_PARANOID(saved_spec->flags &
- EFX_FILTER_FLAG_RX_STACK);
if (spec->priority < saved_spec->priority) {
rc = -EPERM;
goto fail_unlock;
@@ -3118,9 +3134,8 @@ static void efx_ef10_filter_sync_rx_mode(struct efx_nic *efx)
/* Insert/renew unicast filters */
if (table->stack_uc_count >= 0) {
for (i = 0; i < table->stack_uc_count; i++) {
- efx_filter_init_rx(&spec, EFX_FILTER_PRI_REQUIRED,
- EFX_FILTER_FLAG_RX_RSS |
- EFX_FILTER_FLAG_RX_STACK,
+ efx_filter_init_rx(&spec, EFX_FILTER_PRI_AUTO,
+ EFX_FILTER_FLAG_RX_RSS,
0);
efx_filter_set_eth_local(&spec, EFX_FILTER_VID_UNSPEC,
table->stack_uc_list[i].addr);
@@ -3129,7 +3144,7 @@ static void efx_ef10_filter_sync_rx_mode(struct efx_nic *efx)
/* Fall back to unicast-promisc */
while (i--)
efx_ef10_filter_remove_safe(
- efx, EFX_FILTER_PRI_REQUIRED,
+ efx, EFX_FILTER_PRI_AUTO,
table->stack_uc_list[i].id);
table->stack_uc_count = -1;
break;
@@ -3138,9 +3153,8 @@ static void efx_ef10_filter_sync_rx_mode(struct efx_nic *efx)
}
}
if (table->stack_uc_count < 0) {
- efx_filter_init_rx(&spec, EFX_FILTER_PRI_REQUIRED,
- EFX_FILTER_FLAG_RX_RSS |
- EFX_FILTER_FLAG_RX_STACK,
+ efx_filter_init_rx(&spec, EFX_FILTER_PRI_AUTO,
+ EFX_FILTER_FLAG_RX_RSS,
0);
efx_filter_set_uc_def(&spec);
rc = efx_ef10_filter_insert(efx, &spec, true);
@@ -3155,9 +3169,8 @@ static void efx_ef10_filter_sync_rx_mode(struct efx_nic *efx)
/* Insert/renew multicast filters */
if (table->stack_mc_count >= 0) {
for (i = 0; i < table->stack_mc_count; i++) {
- efx_filter_init_rx(&spec, EFX_FILTER_PRI_REQUIRED,
- EFX_FILTER_FLAG_RX_RSS |
- EFX_FILTER_FLAG_RX_STACK,
+ efx_filter_init_rx(&spec, EFX_FILTER_PRI_AUTO,
+ EFX_FILTER_FLAG_RX_RSS,
0);
efx_filter_set_eth_local(&spec, EFX_FILTER_VID_UNSPEC,
table->stack_mc_list[i].addr);
@@ -3166,7 +3179,7 @@ static void efx_ef10_filter_sync_rx_mode(struct efx_nic *efx)
/* Fall back to multicast-promisc */
while (i--)
efx_ef10_filter_remove_safe(
- efx, EFX_FILTER_PRI_REQUIRED,
+ efx, EFX_FILTER_PRI_AUTO,
table->stack_mc_list[i].id);
table->stack_mc_count = -1;
break;
@@ -3175,9 +3188,8 @@ static void efx_ef10_filter_sync_rx_mode(struct efx_nic *efx)
}
}
if (table->stack_mc_count < 0) {
- efx_filter_init_rx(&spec, EFX_FILTER_PRI_REQUIRED,
- EFX_FILTER_FLAG_RX_RSS |
- EFX_FILTER_FLAG_RX_STACK,
+ efx_filter_init_rx(&spec, EFX_FILTER_PRI_AUTO,
+ EFX_FILTER_FLAG_RX_RSS,
0);
efx_filter_set_mc_def(&spec);
rc = efx_ef10_filter_insert(efx, &spec, true);
@@ -3197,9 +3209,8 @@ static void efx_ef10_filter_sync_rx_mode(struct efx_nic *efx)
for (i = 0; i < HUNT_FILTER_TBL_ROWS; i++) {
if (ACCESS_ONCE(table->entry[i].spec) &
EFX_EF10_FILTER_FLAG_STACK_OLD) {
- if (efx_ef10_filter_remove_internal(efx,
- EFX_FILTER_PRI_REQUIRED,
- i, true) < 0)
+ if (efx_ef10_filter_remove_internal(
+ efx, EFX_FILTER_PRI_AUTO, i, true) < 0)
remove_failed = true;
}
}
diff --git a/drivers/net/ethernet/sfc/efx.h b/drivers/net/ethernet/sfc/efx.h
index a653786fbbe7..3a67030c73ea 100644
--- a/drivers/net/ethernet/sfc/efx.h
+++ b/drivers/net/ethernet/sfc/efx.h
@@ -138,6 +138,9 @@ efx_filter_get_filter_safe(struct efx_nic *efx,
* efx_farch_filter_clear_rx - remove RX filters by priority
* @efx: NIC from which to remove the filters
* @priority: Maximum priority to remove
+ *
+ * Remove all RX filters whose priority is less than or equal to the
+ * given @priority and is not %EFX_FILTER_PRI_AUTO.
*/
static inline void efx_filter_clear_rx(struct efx_nic *efx,
enum efx_filter_priority priority)
diff --git a/drivers/net/ethernet/sfc/farch.c b/drivers/net/ethernet/sfc/farch.c
index 4c64ad7c9200..fbd923ddf546 100644
--- a/drivers/net/ethernet/sfc/farch.c
+++ b/drivers/net/ethernet/sfc/farch.c
@@ -2190,8 +2190,8 @@ efx_farch_filter_init_rx_for_stack(struct efx_nic *efx,
/* If there's only one channel then disable RSS for non VF
* traffic, thereby allowing VFs to use RSS when the PF can't.
*/
- spec->priority = EFX_FILTER_PRI_REQUIRED;
- spec->flags = (EFX_FILTER_FLAG_RX | EFX_FILTER_FLAG_RX_STACK |
+ spec->priority = EFX_FILTER_PRI_AUTO;
+ spec->flags = (EFX_FILTER_FLAG_RX |
(efx->n_rx_channels > 1 ? EFX_FILTER_FLAG_RX_RSS : 0) |
(efx->rx_scatter ? EFX_FILTER_FLAG_RX_SCATTER : 0));
spec->dmaq_id = 0;
@@ -2456,20 +2456,13 @@ s32 efx_farch_filter_insert(struct efx_nic *efx,
rc = -EEXIST;
goto out;
}
- if (spec.priority < saved_spec->priority &&
- !(saved_spec->priority == EFX_FILTER_PRI_REQUIRED &&
- saved_spec->flags & EFX_FILTER_FLAG_RX_STACK)) {
+ if (spec.priority < saved_spec->priority) {
rc = -EPERM;
goto out;
}
- if (spec.flags & EFX_FILTER_FLAG_RX_STACK) {
- /* Just make sure it won't be removed */
- saved_spec->flags |= EFX_FILTER_FLAG_RX_STACK;
- rc = 0;
- goto out;
- }
- /* Retain the RX_STACK flag */
- spec.flags |= saved_spec->flags & EFX_FILTER_FLAG_RX_STACK;
+ if (saved_spec->priority == EFX_FILTER_PRI_AUTO ||
+ saved_spec->flags & EFX_FILTER_FLAG_RX_OVER_AUTO)
+ spec.flags |= EFX_FILTER_FLAG_RX_OVER_AUTO;
}
/* Insert the filter */
@@ -2553,7 +2546,7 @@ static int efx_farch_filter_remove(struct efx_nic *efx,
spec->priority > priority)
return -ENOENT;
- if (spec->flags & EFX_FILTER_FLAG_RX_STACK) {
+ if (spec->flags & EFX_FILTER_FLAG_RX_OVER_AUTO) {
efx_farch_filter_init_rx_for_stack(efx, spec);
efx_farch_filter_push_rx_config(efx);
} else {
@@ -2637,8 +2630,11 @@ efx_farch_filter_table_clear(struct efx_nic *efx,
unsigned int filter_idx;
spin_lock_bh(&efx->filter_lock);
- for (filter_idx = 0; filter_idx < table->size; ++filter_idx)
- efx_farch_filter_remove(efx, table, filter_idx, priority);
+ for (filter_idx = 0; filter_idx < table->size; ++filter_idx) {
+ if (table->spec[filter_idx].priority != EFX_FILTER_PRI_AUTO)
+ efx_farch_filter_remove(efx, table,
+ filter_idx, priority);
+ }
spin_unlock_bh(&efx->filter_lock);
}
diff --git a/drivers/net/ethernet/sfc/filter.h b/drivers/net/ethernet/sfc/filter.h
index 63c77a557178..3ef298d3c47e 100644
--- a/drivers/net/ethernet/sfc/filter.h
+++ b/drivers/net/ethernet/sfc/filter.h
@@ -59,12 +59,16 @@ enum efx_filter_match_flags {
/**
* enum efx_filter_priority - priority of a hardware filter specification
* @EFX_FILTER_PRI_HINT: Performance hint
+ * @EFX_FILTER_PRI_AUTO: Automatic filter based on device address list
+ * or hardware requirements. This may only be used by the filter
+ * implementation for each NIC type.
* @EFX_FILTER_PRI_MANUAL: Manually configured filter
* @EFX_FILTER_PRI_REQUIRED: Required for correct behaviour (user-level
* networking and SR-IOV)
*/
enum efx_filter_priority {
EFX_FILTER_PRI_HINT = 0,
+ EFX_FILTER_PRI_AUTO,
EFX_FILTER_PRI_MANUAL,
EFX_FILTER_PRI_REQUIRED,
};
@@ -78,19 +82,18 @@ enum efx_filter_priority {
* according to the indirection table.
* @EFX_FILTER_FLAG_RX_SCATTER: Enable DMA scatter on the receiving
* queue.
- * @EFX_FILTER_FLAG_RX_STACK: Indicates a filter inserted for the
- * network stack. The filter must have a priority of
- * %EFX_FILTER_PRI_REQUIRED. It can be steered by a replacement
- * request with priority %EFX_FILTER_PRI_MANUAL, and a removal
- * request with priority %EFX_FILTER_PRI_MANUAL will reset the
- * steering (but not remove the filter).
+ * @EFX_FILTER_FLAG_RX_OVER_AUTO: Indicates a filter that is
+ * overriding an automatic filter (priority
+ * %EFX_FILTER_PRI_AUTO). This may only be set by the filter
+ * implementation for each type. A removal request will restore
+ * the automatic filter in its place.
* @EFX_FILTER_FLAG_RX: Filter is for RX
* @EFX_FILTER_FLAG_TX: Filter is for TX
*/
enum efx_filter_flags {
EFX_FILTER_FLAG_RX_RSS = 0x01,
EFX_FILTER_FLAG_RX_SCATTER = 0x02,
- EFX_FILTER_FLAG_RX_STACK = 0x04,
+ EFX_FILTER_FLAG_RX_OVER_AUTO = 0x04,
EFX_FILTER_FLAG_RX = 0x08,
EFX_FILTER_FLAG_TX = 0x10,
};
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply related [flat|nested] 13+ messages in thread
* [PATCH net-next 04/11] sfc: Don't refer to 'stack' in filter implementation
2013-12-16 21:29 Pull request: sfc-next 2013-12-16 Ben Hutchings
` (2 preceding siblings ...)
2013-12-16 21:32 ` [PATCH net-next 03/11] sfc: Change priority and flags for automatic MAC filters Ben Hutchings
@ 2013-12-16 21:32 ` Ben Hutchings
2013-12-16 21:33 ` [PATCH net-next 05/11] sfc: Allow filter removal only with exactly matching priority Ben Hutchings
` (7 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: Ben Hutchings @ 2013-12-16 21:32 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linux-net-drivers
Change all the 'stack' naming to 'auto' (or other meaningful term);
the device address list is based on more than just what the network
stack wants, and the no-match filters aren't really what the stack
wants at all.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
---
drivers/net/ethernet/sfc/ef10.c | 104 +++++++++++++++++++--------------------
drivers/net/ethernet/sfc/farch.c | 8 +--
2 files changed, 56 insertions(+), 56 deletions(-)
diff --git a/drivers/net/ethernet/sfc/ef10.c b/drivers/net/ethernet/sfc/ef10.c
index e5eeac9063d2..2adc8e45461b 100644
--- a/drivers/net/ethernet/sfc/ef10.c
+++ b/drivers/net/ethernet/sfc/ef10.c
@@ -53,25 +53,25 @@ struct efx_ef10_filter_table {
struct {
unsigned long spec; /* pointer to spec plus flag bits */
-/* BUSY flag indicates that an update is in progress. STACK_OLD is
- * used to mark and sweep stack-owned MAC filters.
+/* BUSY flag indicates that an update is in progress. AUTO_OLD is
+ * used to mark and sweep MAC filters for the device address lists.
*/
#define EFX_EF10_FILTER_FLAG_BUSY 1UL
-#define EFX_EF10_FILTER_FLAG_STACK_OLD 2UL
+#define EFX_EF10_FILTER_FLAG_AUTO_OLD 2UL
#define EFX_EF10_FILTER_FLAGS 3UL
u64 handle; /* firmware handle */
} *entry;
wait_queue_head_t waitq;
/* Shadow of net_device address lists, guarded by mac_lock */
-#define EFX_EF10_FILTER_STACK_UC_MAX 32
-#define EFX_EF10_FILTER_STACK_MC_MAX 256
+#define EFX_EF10_FILTER_DEV_UC_MAX 32
+#define EFX_EF10_FILTER_DEV_MC_MAX 256
struct {
u8 addr[ETH_ALEN];
u16 id;
- } stack_uc_list[EFX_EF10_FILTER_STACK_UC_MAX],
- stack_mc_list[EFX_EF10_FILTER_STACK_MC_MAX];
- int stack_uc_count; /* negative for PROMISC */
- int stack_mc_count; /* negative for PROMISC/ALLMULTI */
+ } dev_uc_list[EFX_EF10_FILTER_DEV_UC_MAX],
+ dev_mc_list[EFX_EF10_FILTER_DEV_MC_MAX];
+ int dev_uc_count; /* negative for PROMISC */
+ int dev_mc_count; /* negative for PROMISC/ALLMULTI */
};
/* An arbitrary search limit for the software hash table */
@@ -2401,7 +2401,7 @@ found:
if (saved_spec->priority > EFX_FILTER_PRI_AUTO)
saved_spec->flags |= EFX_FILTER_FLAG_RX_OVER_AUTO;
table->entry[ins_index].spec &=
- ~EFX_EF10_FILTER_FLAG_STACK_OLD;
+ ~EFX_EF10_FILTER_FLAG_AUTO_OLD;
rc = ins_index;
goto out_unlock;
}
@@ -2514,13 +2514,13 @@ static void efx_ef10_filter_update_rx_scatter(struct efx_nic *efx)
}
/* Remove a filter.
- * If !stack_requested, remove by ID
- * If stack_requested, remove by index
+ * If !by_index, remove by ID
+ * If by_index, remove by index
* Filter ID may come from userland and must be range-checked.
*/
static int efx_ef10_filter_remove_internal(struct efx_nic *efx,
enum efx_filter_priority priority,
- u32 filter_id, bool stack_requested)
+ u32 filter_id, bool by_index)
{
unsigned int filter_idx = filter_id % HUNT_FILTER_TBL_ROWS;
struct efx_ef10_filter_table *table = efx->filter_state;
@@ -2547,7 +2547,7 @@ static int efx_ef10_filter_remove_internal(struct efx_nic *efx,
spec = efx_ef10_filter_entry_spec(table, filter_idx);
if (!spec ||
- (!stack_requested &&
+ (!by_index &&
efx_ef10_filter_rx_match_pri(table, spec->match_flags) !=
filter_id / HUNT_FILTER_TBL_ROWS)) {
rc = -ENOENT;
@@ -2558,7 +2558,7 @@ static int efx_ef10_filter_remove_internal(struct efx_nic *efx,
priority == EFX_FILTER_PRI_AUTO) {
/* Just remove flags */
spec->flags &= ~EFX_FILTER_FLAG_RX_OVER_AUTO;
- table->entry[filter_idx].spec &= ~EFX_EF10_FILTER_FLAG_STACK_OLD;
+ table->entry[filter_idx].spec &= ~EFX_EF10_FILTER_FLAG_AUTO_OLD;
rc = 0;
goto out_unlock;
}
@@ -2572,7 +2572,7 @@ static int efx_ef10_filter_remove_internal(struct efx_nic *efx,
spin_unlock_bh(&efx->filter_lock);
if (spec->flags & EFX_FILTER_FLAG_RX_OVER_AUTO) {
- /* Reset steering of a stack-owned filter */
+ /* Reset to an automatic filter */
struct efx_filter_spec new_spec = *spec;
@@ -3086,15 +3086,15 @@ static void efx_ef10_filter_sync_rx_mode(struct efx_nic *efx)
/* Mark old filters that may need to be removed */
spin_lock_bh(&efx->filter_lock);
- n = table->stack_uc_count < 0 ? 1 : table->stack_uc_count;
+ n = table->dev_uc_count < 0 ? 1 : table->dev_uc_count;
for (i = 0; i < n; i++) {
- filter_idx = table->stack_uc_list[i].id % HUNT_FILTER_TBL_ROWS;
- table->entry[filter_idx].spec |= EFX_EF10_FILTER_FLAG_STACK_OLD;
+ filter_idx = table->dev_uc_list[i].id % HUNT_FILTER_TBL_ROWS;
+ table->entry[filter_idx].spec |= EFX_EF10_FILTER_FLAG_AUTO_OLD;
}
- n = table->stack_mc_count < 0 ? 1 : table->stack_mc_count;
+ n = table->dev_mc_count < 0 ? 1 : table->dev_mc_count;
for (i = 0; i < n; i++) {
- filter_idx = table->stack_mc_list[i].id % HUNT_FILTER_TBL_ROWS;
- table->entry[filter_idx].spec |= EFX_EF10_FILTER_FLAG_STACK_OLD;
+ filter_idx = table->dev_mc_list[i].id % HUNT_FILTER_TBL_ROWS;
+ table->entry[filter_idx].spec |= EFX_EF10_FILTER_FLAG_AUTO_OLD;
}
spin_unlock_bh(&efx->filter_lock);
@@ -3103,28 +3103,28 @@ static void efx_ef10_filter_sync_rx_mode(struct efx_nic *efx)
*/
netif_addr_lock_bh(net_dev);
if (net_dev->flags & IFF_PROMISC ||
- netdev_uc_count(net_dev) >= EFX_EF10_FILTER_STACK_UC_MAX) {
- table->stack_uc_count = -1;
+ netdev_uc_count(net_dev) >= EFX_EF10_FILTER_DEV_UC_MAX) {
+ table->dev_uc_count = -1;
} else {
- table->stack_uc_count = 1 + netdev_uc_count(net_dev);
- memcpy(table->stack_uc_list[0].addr, net_dev->dev_addr,
+ table->dev_uc_count = 1 + netdev_uc_count(net_dev);
+ memcpy(table->dev_uc_list[0].addr, net_dev->dev_addr,
ETH_ALEN);
i = 1;
netdev_for_each_uc_addr(uc, net_dev) {
- memcpy(table->stack_uc_list[i].addr,
+ memcpy(table->dev_uc_list[i].addr,
uc->addr, ETH_ALEN);
i++;
}
}
if (net_dev->flags & (IFF_PROMISC | IFF_ALLMULTI) ||
- netdev_mc_count(net_dev) >= EFX_EF10_FILTER_STACK_MC_MAX) {
- table->stack_mc_count = -1;
+ netdev_mc_count(net_dev) >= EFX_EF10_FILTER_DEV_MC_MAX) {
+ table->dev_mc_count = -1;
} else {
- table->stack_mc_count = 1 + netdev_mc_count(net_dev);
- eth_broadcast_addr(table->stack_mc_list[0].addr);
+ table->dev_mc_count = 1 + netdev_mc_count(net_dev);
+ eth_broadcast_addr(table->dev_mc_list[0].addr);
i = 1;
netdev_for_each_mc_addr(mc, net_dev) {
- memcpy(table->stack_mc_list[i].addr,
+ memcpy(table->dev_mc_list[i].addr,
mc->addr, ETH_ALEN);
i++;
}
@@ -3132,27 +3132,27 @@ static void efx_ef10_filter_sync_rx_mode(struct efx_nic *efx)
netif_addr_unlock_bh(net_dev);
/* Insert/renew unicast filters */
- if (table->stack_uc_count >= 0) {
- for (i = 0; i < table->stack_uc_count; i++) {
+ if (table->dev_uc_count >= 0) {
+ for (i = 0; i < table->dev_uc_count; i++) {
efx_filter_init_rx(&spec, EFX_FILTER_PRI_AUTO,
EFX_FILTER_FLAG_RX_RSS,
0);
efx_filter_set_eth_local(&spec, EFX_FILTER_VID_UNSPEC,
- table->stack_uc_list[i].addr);
+ table->dev_uc_list[i].addr);
rc = efx_ef10_filter_insert(efx, &spec, true);
if (rc < 0) {
/* Fall back to unicast-promisc */
while (i--)
efx_ef10_filter_remove_safe(
efx, EFX_FILTER_PRI_AUTO,
- table->stack_uc_list[i].id);
- table->stack_uc_count = -1;
+ table->dev_uc_list[i].id);
+ table->dev_uc_count = -1;
break;
}
- table->stack_uc_list[i].id = rc;
+ table->dev_uc_list[i].id = rc;
}
}
- if (table->stack_uc_count < 0) {
+ if (table->dev_uc_count < 0) {
efx_filter_init_rx(&spec, EFX_FILTER_PRI_AUTO,
EFX_FILTER_FLAG_RX_RSS,
0);
@@ -3160,34 +3160,34 @@ static void efx_ef10_filter_sync_rx_mode(struct efx_nic *efx)
rc = efx_ef10_filter_insert(efx, &spec, true);
if (rc < 0) {
WARN_ON(1);
- table->stack_uc_count = 0;
+ table->dev_uc_count = 0;
} else {
- table->stack_uc_list[0].id = rc;
+ table->dev_uc_list[0].id = rc;
}
}
/* Insert/renew multicast filters */
- if (table->stack_mc_count >= 0) {
- for (i = 0; i < table->stack_mc_count; i++) {
+ if (table->dev_mc_count >= 0) {
+ for (i = 0; i < table->dev_mc_count; i++) {
efx_filter_init_rx(&spec, EFX_FILTER_PRI_AUTO,
EFX_FILTER_FLAG_RX_RSS,
0);
efx_filter_set_eth_local(&spec, EFX_FILTER_VID_UNSPEC,
- table->stack_mc_list[i].addr);
+ table->dev_mc_list[i].addr);
rc = efx_ef10_filter_insert(efx, &spec, true);
if (rc < 0) {
/* Fall back to multicast-promisc */
while (i--)
efx_ef10_filter_remove_safe(
efx, EFX_FILTER_PRI_AUTO,
- table->stack_mc_list[i].id);
- table->stack_mc_count = -1;
+ table->dev_mc_list[i].id);
+ table->dev_mc_count = -1;
break;
}
- table->stack_mc_list[i].id = rc;
+ table->dev_mc_list[i].id = rc;
}
}
- if (table->stack_mc_count < 0) {
+ if (table->dev_mc_count < 0) {
efx_filter_init_rx(&spec, EFX_FILTER_PRI_AUTO,
EFX_FILTER_FLAG_RX_RSS,
0);
@@ -3195,20 +3195,20 @@ static void efx_ef10_filter_sync_rx_mode(struct efx_nic *efx)
rc = efx_ef10_filter_insert(efx, &spec, true);
if (rc < 0) {
WARN_ON(1);
- table->stack_mc_count = 0;
+ table->dev_mc_count = 0;
} else {
- table->stack_mc_list[0].id = rc;
+ table->dev_mc_list[0].id = rc;
}
}
/* Remove filters that weren't renewed. Since nothing else
- * changes the STACK_OLD flag or removes these filters, we
+ * changes the AUTO_OLD flag or removes these filters, we
* don't need to hold the filter_lock while scanning for
* these filters.
*/
for (i = 0; i < HUNT_FILTER_TBL_ROWS; i++) {
if (ACCESS_ONCE(table->entry[i].spec) &
- EFX_EF10_FILTER_FLAG_STACK_OLD) {
+ EFX_EF10_FILTER_FLAG_AUTO_OLD) {
if (efx_ef10_filter_remove_internal(
efx, EFX_FILTER_PRI_AUTO, i, true) < 0)
remove_failed = true;
diff --git a/drivers/net/ethernet/sfc/farch.c b/drivers/net/ethernet/sfc/farch.c
index fbd923ddf546..378d6b968c4c 100644
--- a/drivers/net/ethernet/sfc/farch.c
+++ b/drivers/net/ethernet/sfc/farch.c
@@ -2184,8 +2184,8 @@ efx_farch_filter_to_gen_spec(struct efx_filter_spec *gen_spec,
}
static void
-efx_farch_filter_init_rx_for_stack(struct efx_nic *efx,
- struct efx_farch_filter_spec *spec)
+efx_farch_filter_init_rx_auto(struct efx_nic *efx,
+ struct efx_farch_filter_spec *spec)
{
/* If there's only one channel then disable RSS for non VF
* traffic, thereby allowing VFs to use RSS when the PF can't.
@@ -2547,7 +2547,7 @@ static int efx_farch_filter_remove(struct efx_nic *efx,
return -ENOENT;
if (spec->flags & EFX_FILTER_FLAG_RX_OVER_AUTO) {
- efx_farch_filter_init_rx_for_stack(efx, spec);
+ efx_farch_filter_init_rx_auto(efx, spec);
efx_farch_filter_push_rx_config(efx);
} else {
efx_farch_filter_table_clear_entry(efx, table, filter_idx);
@@ -2815,7 +2815,7 @@ int efx_farch_filter_table_probe(struct efx_nic *efx)
for (i = 0; i < EFX_FARCH_FILTER_SIZE_RX_DEF; i++) {
spec = &table->spec[i];
spec->type = EFX_FARCH_FILTER_UC_DEF + i;
- efx_farch_filter_init_rx_for_stack(efx, spec);
+ efx_farch_filter_init_rx_auto(efx, spec);
__set_bit(i, table->used_bitmap);
}
}
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply related [flat|nested] 13+ messages in thread
* [PATCH net-next 05/11] sfc: Allow filter removal only with exactly matching priority
2013-12-16 21:29 Pull request: sfc-next 2013-12-16 Ben Hutchings
` (3 preceding siblings ...)
2013-12-16 21:32 ` [PATCH net-next 04/11] sfc: Don't refer to 'stack' in filter implementation Ben Hutchings
@ 2013-12-16 21:33 ` Ben Hutchings
2013-12-16 21:33 ` [PATCH net-next 06/11] sfc: Implement efx_nic_type::filter_clear_rx operation for EF10 Ben Hutchings
` (6 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: Ben Hutchings @ 2013-12-16 21:33 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linux-net-drivers
Currently a higher priority client can remove a lower priority
client's filter with equal match-expression. This might happen if (a)
the higher priority client has a double-free bug, or (b) another
client with sufficient priority replaced and then removed an equal
filter, allowing the low priority client to insert an equal filter.
In neither case does it actually make sense to carry out the removal;
we should say the filter doesn't exist, as the filter currently
present is not the one that the high-priority client is referring to.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
---
drivers/net/ethernet/sfc/ef10.c | 2 +-
drivers/net/ethernet/sfc/farch.c | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/sfc/ef10.c b/drivers/net/ethernet/sfc/ef10.c
index 2adc8e45461b..155602c500d5 100644
--- a/drivers/net/ethernet/sfc/ef10.c
+++ b/drivers/net/ethernet/sfc/ef10.c
@@ -2563,7 +2563,7 @@ static int efx_ef10_filter_remove_internal(struct efx_nic *efx,
goto out_unlock;
}
- if (spec->priority > priority) {
+ if (spec->priority != priority) {
rc = -ENOENT;
goto out_unlock;
}
diff --git a/drivers/net/ethernet/sfc/farch.c b/drivers/net/ethernet/sfc/farch.c
index 378d6b968c4c..eed0741c2497 100644
--- a/drivers/net/ethernet/sfc/farch.c
+++ b/drivers/net/ethernet/sfc/farch.c
@@ -2543,7 +2543,7 @@ static int efx_farch_filter_remove(struct efx_nic *efx,
struct efx_farch_filter_spec *spec = &table->spec[filter_idx];
if (!test_bit(filter_idx, table->used_bitmap) ||
- spec->priority > priority)
+ spec->priority != priority)
return -ENOENT;
if (spec->flags & EFX_FILTER_FLAG_RX_OVER_AUTO) {
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply related [flat|nested] 13+ messages in thread
* [PATCH net-next 06/11] sfc: Implement efx_nic_type::filter_clear_rx operation for EF10
2013-12-16 21:29 Pull request: sfc-next 2013-12-16 Ben Hutchings
` (4 preceding siblings ...)
2013-12-16 21:33 ` [PATCH net-next 05/11] sfc: Allow filter removal only with exactly matching priority Ben Hutchings
@ 2013-12-16 21:33 ` Ben Hutchings
2013-12-16 21:33 ` [PATCH net-next 07/11] sfc: remove unused 'refcnt' from efx_rx_page_state Ben Hutchings
` (5 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: Ben Hutchings @ 2013-12-16 21:33 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linux-net-drivers
The operation can now fail, so change its return type to int.
Remove the inline wrapper while we're changing the signature.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
---
drivers/net/ethernet/sfc/ef10.c | 30 +++++++++++++++++++++++-------
drivers/net/ethernet/sfc/efx.c | 2 +-
drivers/net/ethernet/sfc/efx.h | 14 --------------
drivers/net/ethernet/sfc/farch.c | 3 ++-
drivers/net/ethernet/sfc/net_driver.h | 7 ++++---
drivers/net/ethernet/sfc/nic.h | 4 ++--
6 files changed, 32 insertions(+), 28 deletions(-)
diff --git a/drivers/net/ethernet/sfc/ef10.c b/drivers/net/ethernet/sfc/ef10.c
index 155602c500d5..b6a0c2133b8a 100644
--- a/drivers/net/ethernet/sfc/ef10.c
+++ b/drivers/net/ethernet/sfc/ef10.c
@@ -2519,7 +2519,7 @@ static void efx_ef10_filter_update_rx_scatter(struct efx_nic *efx)
* Filter ID may come from userland and must be range-checked.
*/
static int efx_ef10_filter_remove_internal(struct efx_nic *efx,
- enum efx_filter_priority priority,
+ unsigned int priority_mask,
u32 filter_id, bool by_index)
{
unsigned int filter_idx = filter_id % HUNT_FILTER_TBL_ROWS;
@@ -2555,7 +2555,7 @@ static int efx_ef10_filter_remove_internal(struct efx_nic *efx,
}
if (spec->flags & EFX_FILTER_FLAG_RX_OVER_AUTO &&
- priority == EFX_FILTER_PRI_AUTO) {
+ priority_mask == (1U << EFX_FILTER_PRI_AUTO)) {
/* Just remove flags */
spec->flags &= ~EFX_FILTER_FLAG_RX_OVER_AUTO;
table->entry[filter_idx].spec &= ~EFX_EF10_FILTER_FLAG_AUTO_OLD;
@@ -2563,7 +2563,7 @@ static int efx_ef10_filter_remove_internal(struct efx_nic *efx,
goto out_unlock;
}
- if (spec->priority != priority) {
+ if (!(priority_mask & (1U << spec->priority))) {
rc = -ENOENT;
goto out_unlock;
}
@@ -2619,7 +2619,8 @@ static int efx_ef10_filter_remove_safe(struct efx_nic *efx,
enum efx_filter_priority priority,
u32 filter_id)
{
- return efx_ef10_filter_remove_internal(efx, priority, filter_id, false);
+ return efx_ef10_filter_remove_internal(efx, 1U << priority,
+ filter_id, false);
}
static int efx_ef10_filter_get_safe(struct efx_nic *efx,
@@ -2645,10 +2646,24 @@ static int efx_ef10_filter_get_safe(struct efx_nic *efx,
return rc;
}
-static void efx_ef10_filter_clear_rx(struct efx_nic *efx,
+static int efx_ef10_filter_clear_rx(struct efx_nic *efx,
enum efx_filter_priority priority)
{
- /* TODO */
+ unsigned int priority_mask;
+ unsigned int i;
+ int rc;
+
+ priority_mask = (((1U << (priority + 1)) - 1) &
+ ~(1U << EFX_FILTER_PRI_AUTO));
+
+ for (i = 0; i < HUNT_FILTER_TBL_ROWS; i++) {
+ rc = efx_ef10_filter_remove_internal(efx, priority_mask,
+ i, true);
+ if (rc && rc != -ENOENT)
+ return rc;
+ }
+
+ return 0;
}
static u32 efx_ef10_filter_count_rx_used(struct efx_nic *efx,
@@ -3210,7 +3225,8 @@ static void efx_ef10_filter_sync_rx_mode(struct efx_nic *efx)
if (ACCESS_ONCE(table->entry[i].spec) &
EFX_EF10_FILTER_FLAG_AUTO_OLD) {
if (efx_ef10_filter_remove_internal(
- efx, EFX_FILTER_PRI_AUTO, i, true) < 0)
+ efx, 1U << EFX_FILTER_PRI_AUTO,
+ i, true) < 0)
remove_failed = true;
}
}
diff --git a/drivers/net/ethernet/sfc/efx.c b/drivers/net/ethernet/sfc/efx.c
index d6a927021523..ff0d9c624560 100644
--- a/drivers/net/ethernet/sfc/efx.c
+++ b/drivers/net/ethernet/sfc/efx.c
@@ -2151,7 +2151,7 @@ static int efx_set_features(struct net_device *net_dev, netdev_features_t data)
/* If disabling RX n-tuple filtering, clear existing filters */
if (net_dev->features & ~data & NETIF_F_NTUPLE)
- efx_filter_clear_rx(efx, EFX_FILTER_PRI_MANUAL);
+ return efx->type->filter_clear_rx(efx, EFX_FILTER_PRI_MANUAL);
return 0;
}
diff --git a/drivers/net/ethernet/sfc/efx.h b/drivers/net/ethernet/sfc/efx.h
index 3a67030c73ea..601224736b9b 100644
--- a/drivers/net/ethernet/sfc/efx.h
+++ b/drivers/net/ethernet/sfc/efx.h
@@ -134,20 +134,6 @@ efx_filter_get_filter_safe(struct efx_nic *efx,
return efx->type->filter_get_safe(efx, priority, filter_id, spec);
}
-/**
- * efx_farch_filter_clear_rx - remove RX filters by priority
- * @efx: NIC from which to remove the filters
- * @priority: Maximum priority to remove
- *
- * Remove all RX filters whose priority is less than or equal to the
- * given @priority and is not %EFX_FILTER_PRI_AUTO.
- */
-static inline void efx_filter_clear_rx(struct efx_nic *efx,
- enum efx_filter_priority priority)
-{
- return efx->type->filter_clear_rx(efx, priority);
-}
-
static inline u32 efx_filter_count_rx_used(struct efx_nic *efx,
enum efx_filter_priority priority)
{
diff --git a/drivers/net/ethernet/sfc/farch.c b/drivers/net/ethernet/sfc/farch.c
index eed0741c2497..f72489a105ca 100644
--- a/drivers/net/ethernet/sfc/farch.c
+++ b/drivers/net/ethernet/sfc/farch.c
@@ -2638,7 +2638,7 @@ efx_farch_filter_table_clear(struct efx_nic *efx,
spin_unlock_bh(&efx->filter_lock);
}
-void efx_farch_filter_clear_rx(struct efx_nic *efx,
+int efx_farch_filter_clear_rx(struct efx_nic *efx,
enum efx_filter_priority priority)
{
efx_farch_filter_table_clear(efx, EFX_FARCH_FILTER_TABLE_RX_IP,
@@ -2647,6 +2647,7 @@ void efx_farch_filter_clear_rx(struct efx_nic *efx,
priority);
efx_farch_filter_table_clear(efx, EFX_FARCH_FILTER_TABLE_RX_DEF,
priority);
+ return 0;
}
u32 efx_farch_filter_count_rx_used(struct efx_nic *efx,
diff --git a/drivers/net/ethernet/sfc/net_driver.h b/drivers/net/ethernet/sfc/net_driver.h
index 1b3c4e5207c5..a0de5c50a0e7 100644
--- a/drivers/net/ethernet/sfc/net_driver.h
+++ b/drivers/net/ethernet/sfc/net_driver.h
@@ -1044,7 +1044,8 @@ struct efx_mtd_partition {
* @filter_insert: add or replace a filter
* @filter_remove_safe: remove a filter by ID, carefully
* @filter_get_safe: retrieve a filter by ID, carefully
- * @filter_clear_rx: remove RX filters by priority
+ * @filter_clear_rx: Remove all RX filters whose priority is less than or
+ * equal to the given priority and is not %EFX_FILTER_PRI_AUTO
* @filter_count_rx_used: Get the number of filters in use at a given priority
* @filter_get_rx_id_limit: Get maximum value of a filter id, plus 1
* @filter_get_rx_ids: Get list of RX filters at a given priority
@@ -1166,8 +1167,8 @@ struct efx_nic_type {
int (*filter_get_safe)(struct efx_nic *efx,
enum efx_filter_priority priority,
u32 filter_id, struct efx_filter_spec *);
- void (*filter_clear_rx)(struct efx_nic *efx,
- enum efx_filter_priority priority);
+ int (*filter_clear_rx)(struct efx_nic *efx,
+ enum efx_filter_priority priority);
u32 (*filter_count_rx_used)(struct efx_nic *efx,
enum efx_filter_priority priority);
u32 (*filter_get_rx_id_limit)(struct efx_nic *efx);
diff --git a/drivers/net/ethernet/sfc/nic.h b/drivers/net/ethernet/sfc/nic.h
index c665ff6c8012..e25aaf6bda87 100644
--- a/drivers/net/ethernet/sfc/nic.h
+++ b/drivers/net/ethernet/sfc/nic.h
@@ -693,8 +693,8 @@ int efx_farch_filter_remove_safe(struct efx_nic *efx,
int efx_farch_filter_get_safe(struct efx_nic *efx,
enum efx_filter_priority priority, u32 filter_id,
struct efx_filter_spec *);
-void efx_farch_filter_clear_rx(struct efx_nic *efx,
- enum efx_filter_priority priority);
+int efx_farch_filter_clear_rx(struct efx_nic *efx,
+ enum efx_filter_priority priority);
u32 efx_farch_filter_count_rx_used(struct efx_nic *efx,
enum efx_filter_priority priority);
u32 efx_farch_filter_get_rx_id_limit(struct efx_nic *efx);
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply related [flat|nested] 13+ messages in thread
* [PATCH net-next 07/11] sfc: remove unused 'refcnt' from efx_rx_page_state
2013-12-16 21:29 Pull request: sfc-next 2013-12-16 Ben Hutchings
` (5 preceding siblings ...)
2013-12-16 21:33 ` [PATCH net-next 06/11] sfc: Implement efx_nic_type::filter_clear_rx operation for EF10 Ben Hutchings
@ 2013-12-16 21:33 ` Ben Hutchings
2013-12-16 21:34 ` [PATCH net-next 08/11] sfc: remove unused 'enum efx_rx_alloc_method' Ben Hutchings
` (4 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: Ben Hutchings @ 2013-12-16 21:33 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linux-net-drivers, Andrew Rybchenko
From: Andrew Rybchenko <Andrew.Rybchenko@oktetlabs.ru>
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
---
drivers/net/ethernet/sfc/net_driver.h | 3 ---
1 file changed, 3 deletions(-)
diff --git a/drivers/net/ethernet/sfc/net_driver.h b/drivers/net/ethernet/sfc/net_driver.h
index a0de5c50a0e7..bc1ff36c03e6 100644
--- a/drivers/net/ethernet/sfc/net_driver.h
+++ b/drivers/net/ethernet/sfc/net_driver.h
@@ -288,12 +288,9 @@ struct efx_rx_buffer {
* Used to facilitate sharing dma mappings between recycled rx buffers
* and those passed up to the kernel.
*
- * @refcnt: Number of struct efx_rx_buffer's referencing this page.
- * When refcnt falls to zero, the page is unmapped for dma
* @dma_addr: The dma address of this page.
*/
struct efx_rx_page_state {
- unsigned refcnt;
dma_addr_t dma_addr;
unsigned int __pad[0] ____cacheline_aligned;
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply related [flat|nested] 13+ messages in thread
* [PATCH net-next 08/11] sfc: remove unused 'enum efx_rx_alloc_method'
2013-12-16 21:29 Pull request: sfc-next 2013-12-16 Ben Hutchings
` (6 preceding siblings ...)
2013-12-16 21:33 ` [PATCH net-next 07/11] sfc: remove unused 'refcnt' from efx_rx_page_state Ben Hutchings
@ 2013-12-16 21:34 ` Ben Hutchings
2013-12-16 21:34 ` [PATCH net-next 09/11] sfc: Changed the statistic name emerg_{fetch,wait} to hlb_{fetch,wait} Ben Hutchings
` (3 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: Ben Hutchings @ 2013-12-16 21:34 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linux-net-drivers, Andrew Rybchenko
From: Andrew Rybchenko <Andrew.Rybchenko@oktetlabs.ru>
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
---
drivers/net/ethernet/sfc/net_driver.h | 6 ------
1 file changed, 6 deletions(-)
diff --git a/drivers/net/ethernet/sfc/net_driver.h b/drivers/net/ethernet/sfc/net_driver.h
index bc1ff36c03e6..af2b8c59a903 100644
--- a/drivers/net/ethernet/sfc/net_driver.h
+++ b/drivers/net/ethernet/sfc/net_driver.h
@@ -360,12 +360,6 @@ struct efx_rx_queue {
unsigned int slow_fill_count;
};
-enum efx_rx_alloc_method {
- RX_ALLOC_METHOD_AUTO = 0,
- RX_ALLOC_METHOD_SKB = 1,
- RX_ALLOC_METHOD_PAGE = 2,
-};
-
enum efx_sync_events_state {
SYNC_EVENTS_DISABLED = 0,
SYNC_EVENTS_QUIESCENT,
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply related [flat|nested] 13+ messages in thread
* [PATCH net-next 09/11] sfc: Changed the statistic name emerg_{fetch,wait} to hlb_{fetch,wait}
2013-12-16 21:29 Pull request: sfc-next 2013-12-16 Ben Hutchings
` (7 preceding siblings ...)
2013-12-16 21:34 ` [PATCH net-next 08/11] sfc: remove unused 'enum efx_rx_alloc_method' Ben Hutchings
@ 2013-12-16 21:34 ` Ben Hutchings
2013-12-16 21:34 ` [PATCH net-next 10/11] sfc: Add PTP counters to ethtool stats Ben Hutchings
` (2 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: Ben Hutchings @ 2013-12-16 21:34 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linux-net-drivers
From: Shradha Shah <sshah@solarflare.com>
The original names are unnecessarily alarming. Head-of-line blocking
is not an emergency.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
---
drivers/net/ethernet/sfc/ef10.c | 8 ++++----
drivers/net/ethernet/sfc/nic.h | 4 ++--
2 files changed, 6 insertions(+), 6 deletions(-)
diff --git a/drivers/net/ethernet/sfc/ef10.c b/drivers/net/ethernet/sfc/ef10.c
index b6a0c2133b8a..d352bdc4e53e 100644
--- a/drivers/net/ethernet/sfc/ef10.c
+++ b/drivers/net/ethernet/sfc/ef10.c
@@ -774,8 +774,8 @@ static const struct efx_hw_stat_desc efx_ef10_stat_desc[EF10_STAT_COUNT] = {
EF10_DMA_STAT(rx_dp_q_disabled_packets, RXDP_Q_DISABLED_PKTS),
EF10_DMA_STAT(rx_dp_di_dropped_packets, RXDP_DI_DROPPED_PKTS),
EF10_DMA_STAT(rx_dp_streaming_packets, RXDP_STREAMING_PKTS),
- EF10_DMA_STAT(rx_dp_emerg_fetch, RXDP_EMERGENCY_FETCH_CONDITIONS),
- EF10_DMA_STAT(rx_dp_emerg_wait, RXDP_EMERGENCY_WAIT_CONDITIONS),
+ EF10_DMA_STAT(rx_dp_hlb_fetch, RXDP_EMERGENCY_FETCH_CONDITIONS),
+ EF10_DMA_STAT(rx_dp_hlb_wait, RXDP_EMERGENCY_WAIT_CONDITIONS),
};
#define HUNT_COMMON_STAT_MASK ((1ULL << EF10_STAT_tx_bytes) | \
@@ -844,8 +844,8 @@ static const struct efx_hw_stat_desc efx_ef10_stat_desc[EF10_STAT_COUNT] = {
(1ULL << EF10_STAT_rx_dp_q_disabled_packets) | \
(1ULL << EF10_STAT_rx_dp_di_dropped_packets) | \
(1ULL << EF10_STAT_rx_dp_streaming_packets) | \
- (1ULL << EF10_STAT_rx_dp_emerg_fetch) | \
- (1ULL << EF10_STAT_rx_dp_emerg_wait))
+ (1ULL << EF10_STAT_rx_dp_hlb_fetch) | \
+ (1ULL << EF10_STAT_rx_dp_hlb_wait))
static u64 efx_ef10_raw_stat_mask(struct efx_nic *efx)
{
diff --git a/drivers/net/ethernet/sfc/nic.h b/drivers/net/ethernet/sfc/nic.h
index e25aaf6bda87..170ac6ee3ab2 100644
--- a/drivers/net/ethernet/sfc/nic.h
+++ b/drivers/net/ethernet/sfc/nic.h
@@ -412,8 +412,8 @@ enum {
EF10_STAT_rx_dp_q_disabled_packets,
EF10_STAT_rx_dp_di_dropped_packets,
EF10_STAT_rx_dp_streaming_packets,
- EF10_STAT_rx_dp_emerg_fetch,
- EF10_STAT_rx_dp_emerg_wait,
+ EF10_STAT_rx_dp_hlb_fetch,
+ EF10_STAT_rx_dp_hlb_wait,
EF10_STAT_COUNT
};
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply related [flat|nested] 13+ messages in thread
* [PATCH net-next 10/11] sfc: Add PTP counters to ethtool stats
2013-12-16 21:29 Pull request: sfc-next 2013-12-16 Ben Hutchings
` (8 preceding siblings ...)
2013-12-16 21:34 ` [PATCH net-next 09/11] sfc: Changed the statistic name emerg_{fetch,wait} to hlb_{fetch,wait} Ben Hutchings
@ 2013-12-16 21:34 ` Ben Hutchings
2013-12-16 21:35 ` [PATCH net-next 11/11] sfc: Fix RX drop filters for EF10 Ben Hutchings
2013-12-17 22:08 ` Pull request: sfc-next 2013-12-16 David Miller
11 siblings, 0 replies; 13+ messages in thread
From: Ben Hutchings @ 2013-12-16 21:34 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linux-net-drivers
These were implemented by Andrew Jackson and Laurence Evans but not
previously included in-tree.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
---
drivers/net/ethernet/sfc/ethtool.c | 8 ++-
drivers/net/ethernet/sfc/nic.h | 2 +
drivers/net/ethernet/sfc/ptp.c | 126 ++++++++++++++++++++++++++++++++++---
3 files changed, 127 insertions(+), 9 deletions(-)
diff --git a/drivers/net/ethernet/sfc/ethtool.c b/drivers/net/ethernet/sfc/ethtool.c
index bb2af8045281..f181522688b2 100644
--- a/drivers/net/ethernet/sfc/ethtool.c
+++ b/drivers/net/ethernet/sfc/ethtool.c
@@ -359,7 +359,8 @@ static int efx_ethtool_get_sset_count(struct net_device *net_dev,
switch (string_set) {
case ETH_SS_STATS:
return efx->type->describe_stats(efx, NULL) +
- EFX_ETHTOOL_SW_STAT_COUNT;
+ EFX_ETHTOOL_SW_STAT_COUNT +
+ efx_ptp_describe_stats(efx, NULL);
case ETH_SS_TEST:
return efx_ethtool_fill_self_tests(efx, NULL, NULL, NULL);
default:
@@ -380,6 +381,8 @@ static void efx_ethtool_get_strings(struct net_device *net_dev,
for (i = 0; i < EFX_ETHTOOL_SW_STAT_COUNT; i++)
strlcpy(strings + i * ETH_GSTRING_LEN,
efx_sw_stat_desc[i].name, ETH_GSTRING_LEN);
+ strings += EFX_ETHTOOL_SW_STAT_COUNT * ETH_GSTRING_LEN;
+ efx_ptp_describe_stats(efx, strings);
break;
case ETH_SS_TEST:
efx_ethtool_fill_self_tests(efx, NULL, strings, NULL);
@@ -429,8 +432,11 @@ static void efx_ethtool_get_stats(struct net_device *net_dev,
break;
}
}
+ data += EFX_ETHTOOL_SW_STAT_COUNT;
spin_unlock_bh(&efx->stats_lock);
+
+ efx_ptp_update_stats(efx, data);
}
static void efx_ethtool_self_test(struct net_device *net_dev,
diff --git a/drivers/net/ethernet/sfc/nic.h b/drivers/net/ethernet/sfc/nic.h
index 170ac6ee3ab2..a001fae1a8d7 100644
--- a/drivers/net/ethernet/sfc/nic.h
+++ b/drivers/net/ethernet/sfc/nic.h
@@ -566,6 +566,8 @@ int efx_ptp_change_mode(struct efx_nic *efx, bool enable_wanted,
unsigned int new_mode);
int efx_ptp_tx(struct efx_nic *efx, struct sk_buff *skb);
void efx_ptp_event(struct efx_nic *efx, efx_qword_t *ev);
+size_t efx_ptp_describe_stats(struct efx_nic *efx, u8 *strings);
+size_t efx_ptp_update_stats(struct efx_nic *efx, u64 *stats);
void efx_time_sync_event(struct efx_channel *channel, efx_qword_t *ev);
void __efx_rx_skb_attach_timestamp(struct efx_channel *channel,
struct sk_buff *skb);
diff --git a/drivers/net/ethernet/sfc/ptp.c b/drivers/net/ethernet/sfc/ptp.c
index 084e2d44790f..7aa070813960 100644
--- a/drivers/net/ethernet/sfc/ptp.c
+++ b/drivers/net/ethernet/sfc/ptp.c
@@ -255,6 +255,15 @@ struct efx_ptp_timeset {
* @nic_ts_enabled: Flag indicating if NIC generated TS events are handled
* @txbuf: Buffer for use when transmitting (PTP) packets to MC (avoids
* allocations in main data path).
+ * @good_syncs: Number of successful synchronisations.
+ * @fast_syncs: Number of synchronisations requiring short delay
+ * @bad_syncs: Number of failed synchronisations.
+ * @sync_timeouts: Number of synchronisation timeouts
+ * @no_time_syncs: Number of synchronisations with no good times.
+ * @invalid_sync_windows: Number of sync windows with bad durations.
+ * @undersize_sync_windows: Number of corrected sync windows that are too small
+ * @oversize_sync_windows: Number of corrected sync windows that are too large
+ * @rx_no_timestamp: Number of packets received without a timestamp.
* @timeset: Last set of synchronisation statistics.
*/
struct efx_ptp_data {
@@ -300,6 +309,16 @@ struct efx_ptp_data {
struct workqueue_struct *pps_workwq;
bool nic_ts_enabled;
MCDI_DECLARE_BUF(txbuf, MC_CMD_PTP_IN_TRANSMIT_LENMAX);
+
+ unsigned int good_syncs;
+ unsigned int fast_syncs;
+ unsigned int bad_syncs;
+ unsigned int sync_timeouts;
+ unsigned int no_time_syncs;
+ unsigned int invalid_sync_windows;
+ unsigned int undersize_sync_windows;
+ unsigned int oversize_sync_windows;
+ unsigned int rx_no_timestamp;
struct efx_ptp_timeset
timeset[MC_CMD_PTP_OUT_SYNCHRONIZE_TIMESET_MAXNUM];
};
@@ -312,6 +331,78 @@ static int efx_phc_settime(struct ptp_clock_info *ptp,
static int efx_phc_enable(struct ptp_clock_info *ptp,
struct ptp_clock_request *request, int on);
+#define PTP_SW_STAT(ext_name, field_name) \
+ { #ext_name, 0, offsetof(struct efx_ptp_data, field_name) }
+#define PTP_MC_STAT(ext_name, mcdi_name) \
+ { #ext_name, 32, MC_CMD_PTP_OUT_STATUS_STATS_ ## mcdi_name ## _OFST }
+static const struct efx_hw_stat_desc efx_ptp_stat_desc[] = {
+ PTP_SW_STAT(ptp_good_syncs, good_syncs),
+ PTP_SW_STAT(ptp_fast_syncs, fast_syncs),
+ PTP_SW_STAT(ptp_bad_syncs, bad_syncs),
+ PTP_SW_STAT(ptp_sync_timeouts, sync_timeouts),
+ PTP_SW_STAT(ptp_no_time_syncs, no_time_syncs),
+ PTP_SW_STAT(ptp_invalid_sync_windows, invalid_sync_windows),
+ PTP_SW_STAT(ptp_undersize_sync_windows, undersize_sync_windows),
+ PTP_SW_STAT(ptp_oversize_sync_windows, oversize_sync_windows),
+ PTP_SW_STAT(ptp_rx_no_timestamp, rx_no_timestamp),
+ PTP_MC_STAT(ptp_tx_timestamp_packets, TX),
+ PTP_MC_STAT(ptp_rx_timestamp_packets, RX),
+ PTP_MC_STAT(ptp_timestamp_packets, TS),
+ PTP_MC_STAT(ptp_filter_matches, FM),
+ PTP_MC_STAT(ptp_non_filter_matches, NFM),
+};
+#define PTP_STAT_COUNT ARRAY_SIZE(efx_ptp_stat_desc)
+static const unsigned long efx_ptp_stat_mask[] = {
+ [0 ... BITS_TO_LONGS(PTP_STAT_COUNT) - 1] = ~0UL,
+};
+
+size_t efx_ptp_describe_stats(struct efx_nic *efx, u8 *strings)
+{
+ if (!efx->ptp_data)
+ return 0;
+
+ return efx_nic_describe_stats(efx_ptp_stat_desc, PTP_STAT_COUNT,
+ efx_ptp_stat_mask, strings);
+}
+
+size_t efx_ptp_update_stats(struct efx_nic *efx, u64 *stats)
+{
+ MCDI_DECLARE_BUF(inbuf, MC_CMD_PTP_IN_STATUS_LEN);
+ MCDI_DECLARE_BUF(outbuf, MC_CMD_PTP_OUT_STATUS_LEN);
+ size_t i;
+ int rc;
+
+ if (!efx->ptp_data)
+ return 0;
+
+ /* Copy software statistics */
+ for (i = 0; i < PTP_STAT_COUNT; i++) {
+ if (efx_ptp_stat_desc[i].dma_width)
+ continue;
+ stats[i] = *(unsigned int *)((char *)efx->ptp_data +
+ efx_ptp_stat_desc[i].offset);
+ }
+
+ /* Fetch MC statistics. We *must* fill in all statistics or
+ * risk leaking kernel memory to userland, so if the MCDI
+ * request fails we pretend we got zeroes.
+ */
+ MCDI_SET_DWORD(inbuf, PTP_IN_OP, MC_CMD_PTP_OP_STATUS);
+ MCDI_SET_DWORD(inbuf, PTP_IN_PERIPH_ID, 0);
+ rc = efx_mcdi_rpc(efx, MC_CMD_PTP, inbuf, sizeof(inbuf),
+ outbuf, sizeof(outbuf), NULL);
+ if (rc) {
+ netif_err(efx, hw, efx->net_dev,
+ "MC_CMD_PTP_OP_STATUS failed (%d)\n", rc);
+ memset(outbuf, 0, sizeof(outbuf));
+ }
+ efx_nic_update_stats(efx_ptp_stat_desc, PTP_STAT_COUNT,
+ efx_ptp_stat_mask,
+ stats, _MCDI_PTR(outbuf, 0), false);
+
+ return PTP_STAT_COUNT;
+}
+
/* For Siena platforms NIC time is s and ns */
static void efx_ptp_ns_to_s_ns(s64 ns, u32 *nic_major, u32 *nic_minor)
{
@@ -633,7 +724,8 @@ efx_ptp_process_times(struct efx_nic *efx, MCDI_DECLARE_STRUCT_PTR(synch_buf),
/* Read the set of results and find the last good host-MC
* synchronization result. The MC times when it finishes reading the
* host time so the corrected window time should be fairly constant
- * for a given platform.
+ * for a given platform. Increment stats for any results that appear
+ * to be erroneous.
*/
for (i = 0; i < number_readings; i++) {
s32 window, corrected;
@@ -658,9 +750,13 @@ efx_ptp_process_times(struct efx_nic *efx, MCDI_DECLARE_STRUCT_PTR(synch_buf),
* interrupt or other delay occurred between reading the system
* time and writing it to MC memory.
*/
- if (window >= SYNCHRONISATION_GRANULARITY_NS &&
- corrected < MAX_SYNCHRONISATION_NS &&
- corrected >= ptp->min_synchronisation_ns) {
+ if (window < SYNCHRONISATION_GRANULARITY_NS) {
+ ++ptp->invalid_sync_windows;
+ } else if (corrected >= MAX_SYNCHRONISATION_NS) {
+ ++ptp->undersize_sync_windows;
+ } else if (corrected < ptp->min_synchronisation_ns) {
+ ++ptp->oversize_sync_windows;
+ } else {
ngood++;
last_good = i;
}
@@ -741,6 +837,11 @@ static int efx_ptp_synchronize(struct efx_nic *efx, unsigned int num_readings)
loops++;
}
+ if (loops <= 1)
+ ++ptp->fast_syncs;
+ if (!time_before(jiffies, timeout))
+ ++ptp->sync_timeouts;
+
if (ACCESS_ONCE(*start))
efx_ptp_send_times(efx, &last_time);
@@ -749,9 +850,20 @@ static int efx_ptp_synchronize(struct efx_nic *efx, unsigned int num_readings)
MC_CMD_PTP_IN_SYNCHRONIZE_LEN,
synch_buf, sizeof(synch_buf),
&response_length);
- if (rc == 0)
+ if (rc == 0) {
rc = efx_ptp_process_times(efx, synch_buf, response_length,
&last_time);
+ if (rc == 0)
+ ++ptp->good_syncs;
+ else
+ ++ptp->no_time_syncs;
+ }
+
+ /* Increment the bad syncs counter if the synchronize fails, whatever
+ * the reason.
+ */
+ if (rc != 0)
+ ++ptp->bad_syncs;
return rc;
}
@@ -907,9 +1019,7 @@ static void efx_ptp_process_events(struct efx_nic *efx, struct sk_buff_head *q)
__skb_queue_tail(q, skb);
} else if (time_after(jiffies, match->expiry)) {
match->state = PTP_PACKET_STATE_TIMED_OUT;
- if (net_ratelimit())
- netif_warn(efx, rx_err, efx->net_dev,
- "PTP packet - no timestamp seen\n");
+ ++ptp->rx_no_timestamp;
__skb_queue_tail(q, skb);
} else {
/* Replace unprocessed entry and stop */
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply related [flat|nested] 13+ messages in thread
* [PATCH net-next 11/11] sfc: Fix RX drop filters for EF10
2013-12-16 21:29 Pull request: sfc-next 2013-12-16 Ben Hutchings
` (9 preceding siblings ...)
2013-12-16 21:34 ` [PATCH net-next 10/11] sfc: Add PTP counters to ethtool stats Ben Hutchings
@ 2013-12-16 21:35 ` Ben Hutchings
2013-12-17 22:08 ` Pull request: sfc-next 2013-12-16 David Miller
11 siblings, 0 replies; 13+ messages in thread
From: Ben Hutchings @ 2013-12-16 21:35 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linux-net-drivers
When we insert an filter, the firmware checks that the given RX queue
index is in range even if it will not be used. In case we're
inserting a drop filter, pass the value 0.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
---
drivers/net/ethernet/sfc/ef10.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/sfc/ef10.c b/drivers/net/ethernet/sfc/ef10.c
index d352bdc4e53e..4dfc2296600d 100644
--- a/drivers/net/ethernet/sfc/ef10.c
+++ b/drivers/net/ethernet/sfc/ef10.c
@@ -2245,7 +2245,9 @@ static void efx_ef10_filter_push_prep(struct efx_nic *efx,
MC_CMD_FILTER_OP_IN_RX_DEST_HOST);
MCDI_SET_DWORD(inbuf, FILTER_OP_IN_TX_DEST,
MC_CMD_FILTER_OP_IN_TX_DEST_DEFAULT);
- MCDI_SET_DWORD(inbuf, FILTER_OP_IN_RX_QUEUE, spec->dmaq_id);
+ MCDI_SET_DWORD(inbuf, FILTER_OP_IN_RX_QUEUE,
+ spec->dmaq_id == EFX_FILTER_RX_DMAQ_ID_DROP ?
+ 0 : spec->dmaq_id);
MCDI_SET_DWORD(inbuf, FILTER_OP_IN_RX_MODE,
(spec->flags & EFX_FILTER_FLAG_RX_RSS) ?
MC_CMD_FILTER_OP_IN_RX_MODE_RSS :
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply related [flat|nested] 13+ messages in thread
* Re: Pull request: sfc-next 2013-12-16
2013-12-16 21:29 Pull request: sfc-next 2013-12-16 Ben Hutchings
` (10 preceding siblings ...)
2013-12-16 21:35 ` [PATCH net-next 11/11] sfc: Fix RX drop filters for EF10 Ben Hutchings
@ 2013-12-17 22:08 ` David Miller
11 siblings, 0 replies; 13+ messages in thread
From: David Miller @ 2013-12-17 22:08 UTC (permalink / raw)
To: bhutchings; +Cc: netdev, linux-net-drivers
From: Ben Hutchings <bhutchings@solarflare.com>
Date: Mon, 16 Dec 2013 21:29:10 +0000
> Miscellaneous changes for 3.14:
>
> 1. Add more information to some WARN messages.
> 2. Refactor pushing of RSS configuration, from Andrew Rybchenko.
> 3. Refactor handling of automatic (device address list) vs manual (RX
> NFC) MAC filters.
> 4. Implement clearing of manual RX filters on EF10 when ntuple offload
> is disabled.
> 5. Remove definitions that are unused since the RX buffer allocation
> changes, from Andrew Rybchenko.
> 6. Improve naming of some statistics, from Shradha Shah.
> 7. Add statistics for PTP support code.
> 8. Fix insertion of RX drop filters on EF10.
Pulled, thanks Ben.
^ permalink raw reply [flat|nested] 13+ messages in thread
end of thread, other threads:[~2013-12-17 22:08 UTC | newest]
Thread overview: 13+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2013-12-16 21:29 Pull request: sfc-next 2013-12-16 Ben Hutchings
2013-12-16 21:31 ` [PATCH net-next 01/11] sfc: Add more information to many warnings using WARN() and netdev_WARN() Ben Hutchings
2013-12-16 21:31 ` [PATCH net-next 02/11] sfc: Change efx_nic_type::rx_push_indir_table to push hash key as well Ben Hutchings
2013-12-16 21:32 ` [PATCH net-next 03/11] sfc: Change priority and flags for automatic MAC filters Ben Hutchings
2013-12-16 21:32 ` [PATCH net-next 04/11] sfc: Don't refer to 'stack' in filter implementation Ben Hutchings
2013-12-16 21:33 ` [PATCH net-next 05/11] sfc: Allow filter removal only with exactly matching priority Ben Hutchings
2013-12-16 21:33 ` [PATCH net-next 06/11] sfc: Implement efx_nic_type::filter_clear_rx operation for EF10 Ben Hutchings
2013-12-16 21:33 ` [PATCH net-next 07/11] sfc: remove unused 'refcnt' from efx_rx_page_state Ben Hutchings
2013-12-16 21:34 ` [PATCH net-next 08/11] sfc: remove unused 'enum efx_rx_alloc_method' Ben Hutchings
2013-12-16 21:34 ` [PATCH net-next 09/11] sfc: Changed the statistic name emerg_{fetch,wait} to hlb_{fetch,wait} Ben Hutchings
2013-12-16 21:34 ` [PATCH net-next 10/11] sfc: Add PTP counters to ethtool stats Ben Hutchings
2013-12-16 21:35 ` [PATCH net-next 11/11] sfc: Fix RX drop filters for EF10 Ben Hutchings
2013-12-17 22:08 ` Pull request: sfc-next 2013-12-16 David Miller
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).