Commit 30129cf2 authored by David S. Miller's avatar David S. Miller

Merge branch 'for-davem' of git://git.kernel.org/pub/scm/linux/kernel/git/bwh/sfc-next

Ben Hutchings says:

====================
1. Merge sfc changes (only) accepted for 3.9.

2. PTP improvements from Laurence Evans.

3. Overhaul of RX buffer management:
- Always allocate pages, and enable scattering where possible
- Fit as many buffers as will fit into a page, rather than limiting to 2
- Introduce recycle rings to reduce the need for IOMMU mapping and
  unmapping

4. PCI error recovery (AER and EEH) implementation.

5. Fix a bug in RX filter replacement.

6. Fix configuration with 1 RX queue in the PF and multiple RX queues in
VFs.
====================
Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
parents e1733de2 1648a23f
This diff is collapsed.
......@@ -33,17 +33,22 @@ extern int efx_setup_tc(struct net_device *net_dev, u8 num_tc);
extern unsigned int efx_tx_max_skb_descs(struct efx_nic *efx);
/* RX */
extern void efx_rx_config_page_split(struct efx_nic *efx);
extern int efx_probe_rx_queue(struct efx_rx_queue *rx_queue);
extern void efx_remove_rx_queue(struct efx_rx_queue *rx_queue);
extern void efx_init_rx_queue(struct efx_rx_queue *rx_queue);
extern void efx_fini_rx_queue(struct efx_rx_queue *rx_queue);
extern void efx_rx_strategy(struct efx_channel *channel);
extern void efx_fast_push_rx_descriptors(struct efx_rx_queue *rx_queue);
extern void efx_rx_slow_fill(unsigned long context);
extern void __efx_rx_packet(struct efx_channel *channel,
struct efx_rx_buffer *rx_buf);
extern void efx_rx_packet(struct efx_rx_queue *rx_queue, unsigned int index,
extern void __efx_rx_packet(struct efx_channel *channel);
extern void efx_rx_packet(struct efx_rx_queue *rx_queue,
unsigned int index, unsigned int n_frags,
unsigned int len, u16 flags);
static inline void efx_rx_flush_packet(struct efx_channel *channel)
{
if (channel->rx_pkt_n_frags)
__efx_rx_packet(channel);
}
extern void efx_schedule_slow_fill(struct efx_rx_queue *rx_queue);
#define EFX_MAX_DMAQ_SIZE 4096UL
......@@ -67,6 +72,7 @@ extern void efx_schedule_slow_fill(struct efx_rx_queue *rx_queue);
extern int efx_probe_filters(struct efx_nic *efx);
extern void efx_restore_filters(struct efx_nic *efx);
extern void efx_remove_filters(struct efx_nic *efx);
extern void efx_filter_update_rx_scatter(struct efx_nic *efx);
extern s32 efx_filter_insert_filter(struct efx_nic *efx,
struct efx_filter_spec *spec,
bool replace);
......@@ -171,9 +177,9 @@ static inline void efx_device_detach_sync(struct efx_nic *efx)
* TX scheduler is stopped when we're done and before
* netif_device_present() becomes false.
*/
netif_tx_lock(dev);
netif_tx_lock_bh(dev);
netif_device_detach(dev);
netif_tx_unlock(dev);
netif_tx_unlock_bh(dev);
}
#endif /* EFX_EFX_H */
......@@ -137,8 +137,12 @@ enum efx_loopback_mode {
* Reset methods are numbered in order of increasing scope.
*
* @RESET_TYPE_INVISIBLE: Reset datapath and MAC (Falcon only)
* @RESET_TYPE_RECOVER_OR_ALL: Try to recover. Apply RESET_TYPE_ALL
* if unsuccessful.
* @RESET_TYPE_ALL: Reset datapath, MAC and PHY
* @RESET_TYPE_WORLD: Reset as much as possible
* @RESET_TYPE_RECOVER_OR_DISABLE: Try to recover. Apply RESET_TYPE_DISABLE if
* unsuccessful.
* @RESET_TYPE_DISABLE: Reset datapath, MAC and PHY; leave NIC disabled
* @RESET_TYPE_TX_WATCHDOG: reset due to TX watchdog
* @RESET_TYPE_INT_ERROR: reset due to internal error
......@@ -150,9 +154,11 @@ enum efx_loopback_mode {
*/
enum reset_type {
RESET_TYPE_INVISIBLE = 0,
RESET_TYPE_ALL = 1,
RESET_TYPE_WORLD = 2,
RESET_TYPE_DISABLE = 3,
RESET_TYPE_RECOVER_OR_ALL = 1,
RESET_TYPE_ALL = 2,
RESET_TYPE_WORLD = 3,
RESET_TYPE_RECOVER_OR_DISABLE = 4,
RESET_TYPE_DISABLE = 5,
RESET_TYPE_MAX_METHOD,
RESET_TYPE_TX_WATCHDOG,
RESET_TYPE_INT_ERROR,
......
......@@ -154,6 +154,7 @@ static const struct efx_ethtool_stat efx_ethtool_stats[] = {
EFX_ETHTOOL_UINT_CHANNEL_STAT(rx_tcp_udp_chksum_err),
EFX_ETHTOOL_UINT_CHANNEL_STAT(rx_mcast_mismatch),
EFX_ETHTOOL_UINT_CHANNEL_STAT(rx_frm_trunc),
EFX_ETHTOOL_UINT_CHANNEL_STAT(rx_nodesc_trunc),
};
/* Number of ethtool statistics */
......@@ -978,7 +979,8 @@ static int efx_ethtool_set_class_rule(struct efx_nic *efx,
rule->m_ext.data[1]))
return -EINVAL;
efx_filter_init_rx(&spec, EFX_FILTER_PRI_MANUAL, 0,
efx_filter_init_rx(&spec, EFX_FILTER_PRI_MANUAL,
efx->rx_scatter ? EFX_FILTER_FLAG_RX_SCATTER : 0,
(rule->ring_cookie == RX_CLS_FLOW_DISC) ?
0xfff : rule->ring_cookie);
......
......@@ -1546,10 +1546,6 @@ static int falcon_probe_nic(struct efx_nic *efx)
static void falcon_init_rx_cfg(struct efx_nic *efx)
{
/* Prior to Siena the RX DMA engine will split each frame at
* intervals of RX_USR_BUF_SIZE (32-byte units). We set it to
* be so large that that never happens. */
const unsigned huge_buf_size = (3 * 4096) >> 5;
/* RX control FIFO thresholds (32 entries) */
const unsigned ctrl_xon_thr = 20;
const unsigned ctrl_xoff_thr = 25;
......@@ -1557,10 +1553,15 @@ static void falcon_init_rx_cfg(struct efx_nic *efx)
efx_reado(efx, &reg, FR_AZ_RX_CFG);
if (efx_nic_rev(efx) <= EFX_REV_FALCON_A1) {
/* Data FIFO size is 5.5K */
/* Data FIFO size is 5.5K. The RX DMA engine only
* supports scattering for user-mode queues, but will
* split DMA writes at intervals of RX_USR_BUF_SIZE
* (32-byte units) even for kernel-mode queues. We
* set it to be so large that that never happens.
*/
EFX_SET_OWORD_FIELD(reg, FRF_AA_RX_DESC_PUSH_EN, 0);
EFX_SET_OWORD_FIELD(reg, FRF_AA_RX_USR_BUF_SIZE,
huge_buf_size);
(3 * 4096) >> 5);
EFX_SET_OWORD_FIELD(reg, FRF_AA_RX_XON_MAC_TH, 512 >> 8);
EFX_SET_OWORD_FIELD(reg, FRF_AA_RX_XOFF_MAC_TH, 2048 >> 8);
EFX_SET_OWORD_FIELD(reg, FRF_AA_RX_XON_TX_TH, ctrl_xon_thr);
......@@ -1569,7 +1570,7 @@ static void falcon_init_rx_cfg(struct efx_nic *efx)
/* Data FIFO size is 80K; register fields moved */
EFX_SET_OWORD_FIELD(reg, FRF_BZ_RX_DESC_PUSH_EN, 0);
EFX_SET_OWORD_FIELD(reg, FRF_BZ_RX_USR_BUF_SIZE,
huge_buf_size);
EFX_RX_USR_BUF_SIZE >> 5);
/* Send XON and XOFF at ~3 * max MTU away from empty/full */
EFX_SET_OWORD_FIELD(reg, FRF_BZ_RX_XON_MAC_TH, 27648 >> 8);
EFX_SET_OWORD_FIELD(reg, FRF_BZ_RX_XOFF_MAC_TH, 54272 >> 8);
......@@ -1815,6 +1816,7 @@ const struct efx_nic_type falcon_a1_nic_type = {
.evq_rptr_tbl_base = FR_AA_EVQ_RPTR_KER,
.max_dma_mask = DMA_BIT_MASK(FSF_AZ_TX_KER_BUF_ADDR_WIDTH),
.rx_buffer_padding = 0x24,
.can_rx_scatter = false,
.max_interrupt_mode = EFX_INT_MODE_MSI,
.phys_addr_channels = 4,
.timer_period_max = 1 << FRF_AB_TC_TIMER_VAL_WIDTH,
......@@ -1865,6 +1867,7 @@ const struct efx_nic_type falcon_b0_nic_type = {
.max_dma_mask = DMA_BIT_MASK(FSF_AZ_TX_KER_BUF_ADDR_WIDTH),
.rx_buffer_hash_size = 0x10,
.rx_buffer_padding = 0,
.can_rx_scatter = true,
.max_interrupt_mode = EFX_INT_MODE_MSIX,
.phys_addr_channels = 32, /* Hardware limit is 64, but the legacy
* interrupt handler only supports 32
......
This diff is collapsed.
......@@ -553,6 +553,7 @@
#define MC_CMD_PTP_MODE_V1_VLAN 0x1 /* enum */
#define MC_CMD_PTP_MODE_V2 0x2 /* enum */
#define MC_CMD_PTP_MODE_V2_VLAN 0x3 /* enum */
#define MC_CMD_PTP_MODE_V2_ENHANCED 0x4 /* enum */
/* MC_CMD_PTP_IN_DISABLE msgrequest */
#define MC_CMD_PTP_IN_DISABLE_LEN 8
......
......@@ -69,6 +69,12 @@
#define EFX_TXQ_TYPES 4
#define EFX_MAX_TX_QUEUES (EFX_TXQ_TYPES * EFX_MAX_CHANNELS)
/* Maximum possible MTU the driver supports */
#define EFX_MAX_MTU (9 * 1024)
/* Size of an RX scatter buffer. Small enough to pack 2 into a 4K page. */
#define EFX_RX_USR_BUF_SIZE 1824
/* Forward declare Precision Time Protocol (PTP) support structure. */
struct efx_ptp_data;
......@@ -206,25 +212,23 @@ struct efx_tx_queue {
/**
* struct efx_rx_buffer - An Efx RX data buffer
* @dma_addr: DMA base address of the buffer
* @skb: The associated socket buffer. Valid iff !(@flags & %EFX_RX_BUF_PAGE).
* @page: The associated page buffer.
* Will be %NULL if the buffer slot is currently free.
* @page: The associated page buffer. Valif iff @flags & %EFX_RX_BUF_PAGE.
* Will be %NULL if the buffer slot is currently free.
* @page_offset: Offset within page. Valid iff @flags & %EFX_RX_BUF_PAGE.
* @len: Buffer length, in bytes.
* @flags: Flags for buffer and packet state.
* @page_offset: If pending: offset in @page of DMA base address.
* If completed: offset in @page of Ethernet header.
* @len: If pending: length for DMA descriptor.
* If completed: received length, excluding hash prefix.
* @flags: Flags for buffer and packet state. These are only set on the
* first buffer of a scattered packet.
*/
struct efx_rx_buffer {
dma_addr_t dma_addr;
union {
struct sk_buff *skb;
struct page *page;
} u;
u16 page_offset;
u16 len;
u16 flags;
};
#define EFX_RX_BUF_PAGE 0x0001
#define EFX_RX_BUF_LAST_IN_PAGE 0x0001
#define EFX_RX_PKT_CSUMMED 0x0002
#define EFX_RX_PKT_DISCARD 0x0004
......@@ -260,14 +264,23 @@ struct efx_rx_page_state {
* @added_count: Number of buffers added to the receive queue.
* @notified_count: Number of buffers given to NIC (<= @added_count).
* @removed_count: Number of buffers removed from the receive queue.
* @scatter_n: Number of buffers used by current packet
* @page_ring: The ring to store DMA mapped pages for reuse.
* @page_add: Counter to calculate the write pointer for the recycle ring.
* @page_remove: Counter to calculate the read pointer for the recycle ring.
* @page_recycle_count: The number of pages that have been recycled.
* @page_recycle_failed: The number of pages that couldn't be recycled because
* the kernel still held a reference to them.
* @page_recycle_full: The number of pages that were released because the
* recycle ring was full.
* @page_ptr_mask: The number of pages in the RX recycle ring minus 1.
* @max_fill: RX descriptor maximum fill level (<= ring size)
* @fast_fill_trigger: RX descriptor fill level that will trigger a fast fill
* (<= @max_fill)
* @min_fill: RX descriptor minimum non-zero fill level.
* This records the minimum fill level observed when a ring
* refill was triggered.
* @alloc_page_count: RX allocation strategy counter.
* @alloc_skb_count: RX allocation strategy counter.
* @recycle_count: RX buffer recycle counter.
* @slow_fill: Timer used to defer efx_nic_generate_fill_event().
*/
struct efx_rx_queue {
......@@ -279,15 +292,22 @@ struct efx_rx_queue {
bool enabled;
bool flush_pending;
int added_count;
int notified_count;
int removed_count;
unsigned int added_count;
unsigned int notified_count;
unsigned int removed_count;
unsigned int scatter_n;
struct page **page_ring;
unsigned int page_add;
unsigned int page_remove;
unsigned int page_recycle_count;
unsigned int page_recycle_failed;
unsigned int page_recycle_full;
unsigned int page_ptr_mask;
unsigned int max_fill;
unsigned int fast_fill_trigger;
unsigned int min_fill;
unsigned int min_overfill;
unsigned int alloc_page_count;
unsigned int alloc_skb_count;
unsigned int recycle_count;
struct timer_list slow_fill;
unsigned int slow_fill_count;
};
......@@ -336,10 +356,6 @@ enum efx_rx_alloc_method {
* @event_test_cpu: Last CPU to handle interrupt or test event for this channel
* @irq_count: Number of IRQs since last adaptive moderation decision
* @irq_mod_score: IRQ moderation score
* @rx_alloc_level: Watermark based heuristic counter for pushing descriptors
* and diagnostic counters
* @rx_alloc_push_pages: RX allocation method currently in use for pushing
* descriptors
* @n_rx_tobe_disc: Count of RX_TOBE_DISC errors
* @n_rx_ip_hdr_chksum_err: Count of RX IP header checksum errors
* @n_rx_tcp_udp_chksum_err: Count of RX TCP and UDP checksum errors
......@@ -347,6 +363,12 @@ enum efx_rx_alloc_method {
* @n_rx_frm_trunc: Count of RX_FRM_TRUNC errors
* @n_rx_overlength: Count of RX_OVERLENGTH errors
* @n_skbuff_leaks: Count of skbuffs leaked due to RX overrun
* @n_rx_nodesc_trunc: Number of RX packets truncated and then dropped due to
* lack of descriptors
* @rx_pkt_n_frags: Number of fragments in next packet to be delivered by
* __efx_rx_packet(), or zero if there is none
* @rx_pkt_index: Ring index of first buffer for next packet to be delivered
* by __efx_rx_packet(), if @rx_pkt_n_frags != 0
* @rx_queue: RX queue for this channel
* @tx_queue: TX queues for this channel
*/
......@@ -371,9 +393,6 @@ struct efx_channel {
unsigned int rfs_filters_added;
#endif
int rx_alloc_level;
int rx_alloc_push_pages;
unsigned n_rx_tobe_disc;
unsigned n_rx_ip_hdr_chksum_err;
unsigned n_rx_tcp_udp_chksum_err;
......@@ -381,11 +400,10 @@ struct efx_channel {
unsigned n_rx_frm_trunc;
unsigned n_rx_overlength;
unsigned n_skbuff_leaks;
unsigned int n_rx_nodesc_trunc;
/* Used to pipeline received packets in order to optimise memory
* access with prefetches.
*/
struct efx_rx_buffer *rx_pkt;
unsigned int rx_pkt_n_frags;
unsigned int rx_pkt_index;
struct efx_rx_queue rx_queue;
struct efx_tx_queue tx_queue[EFX_TXQ_TYPES];
......@@ -410,7 +428,7 @@ struct efx_channel_type {
void (*post_remove)(struct efx_channel *);
void (*get_name)(struct efx_channel *, char *buf, size_t len);
struct efx_channel *(*copy)(const struct efx_channel *);
void (*receive_skb)(struct efx_channel *, struct sk_buff *);
bool (*receive_skb)(struct efx_channel *, struct sk_buff *);
bool keep_eventq;
};
......@@ -446,6 +464,7 @@ enum nic_state {
STATE_UNINIT = 0, /* device being probed/removed or is frozen */
STATE_READY = 1, /* hardware ready and netdev registered */
STATE_DISABLED = 2, /* device disabled due to hardware errors */
STATE_RECOVERY = 3, /* device recovering from PCI error */
};
/*
......@@ -684,10 +703,13 @@ struct vfdi_status;
* @n_channels: Number of channels in use
* @n_rx_channels: Number of channels used for RX (= number of RX queues)
* @n_tx_channels: Number of channels used for TX
* @rx_buffer_len: RX buffer length
* @rx_dma_len: Current maximum RX DMA length
* @rx_buffer_order: Order (log2) of number of pages for each RX buffer
* @rx_buffer_truesize: Amortised allocation size of an RX buffer,
* for use in sk_buff::truesize
* @rx_hash_key: Toeplitz hash key for RSS
* @rx_indir_table: Indirection table for RSS
* @rx_scatter: Scatter mode enabled for receives
* @int_error_count: Number of internal errors seen recently
* @int_error_expire: Time at which error count will be expired
* @irq_status: Interrupt status buffer
......@@ -800,10 +822,15 @@ struct efx_nic {
unsigned rss_spread;
unsigned tx_channel_offset;
unsigned n_tx_channels;
unsigned int rx_buffer_len;
unsigned int rx_dma_len;
unsigned int rx_buffer_order;
unsigned int rx_buffer_truesize;
unsigned int rx_page_buf_step;
unsigned int rx_bufs_per_page;
unsigned int rx_pages_per_batch;
u8 rx_hash_key[40];
u32 rx_indir_table[128];
bool rx_scatter;
unsigned int_error_count;
unsigned long int_error_expire;
......@@ -934,8 +961,9 @@ static inline unsigned int efx_port_num(struct efx_nic *efx)
* @evq_ptr_tbl_base: Event queue pointer table base address
* @evq_rptr_tbl_base: Event queue read-pointer table base address
* @max_dma_mask: Maximum possible DMA mask
* @rx_buffer_hash_size: Size of hash at start of RX buffer
* @rx_buffer_padding: Size of padding at end of RX buffer
* @rx_buffer_hash_size: Size of hash at start of RX packet
* @rx_buffer_padding: Size of padding at end of RX packet
* @can_rx_scatter: NIC is able to scatter packet to multiple buffers
* @max_interrupt_mode: Highest capability interrupt mode supported
* from &enum efx_init_mode.
* @phys_addr_channels: Number of channels with physically addressed
......@@ -983,6 +1011,7 @@ struct efx_nic_type {
u64 max_dma_mask;
unsigned int rx_buffer_hash_size;
unsigned int rx_buffer_padding;
bool can_rx_scatter;
unsigned int max_interrupt_mode;
unsigned int phys_addr_channels;
unsigned int timer_period_max;
......
......@@ -591,12 +591,22 @@ void efx_nic_init_rx(struct efx_rx_queue *rx_queue)
struct efx_nic *efx = rx_queue->efx;
bool is_b0 = efx_nic_rev(efx) >= EFX_REV_FALCON_B0;
bool iscsi_digest_en = is_b0;
bool jumbo_en;
/* For kernel-mode queues in Falcon A1, the JUMBO flag enables
* DMA to continue after a PCIe page boundary (and scattering
* is not possible). In Falcon B0 and Siena, it enables
* scatter.
*/
jumbo_en = !is_b0 || efx->rx_scatter;
netif_dbg(efx, hw, efx->net_dev,
"RX queue %d ring in special buffers %d-%d\n",
efx_rx_queue_index(rx_queue), rx_queue->rxd.index,
rx_queue->rxd.index + rx_queue->rxd.entries - 1);
rx_queue->scatter_n = 0;
/* Pin RX descriptor ring */
efx_init_special_buffer(efx, &rx_queue->rxd);
......@@ -613,8 +623,7 @@ void efx_nic_init_rx(struct efx_rx_queue *rx_queue)
FRF_AZ_RX_DESCQ_SIZE,
__ffs(rx_queue->rxd.entries),
FRF_AZ_RX_DESCQ_TYPE, 0 /* kernel queue */ ,
/* For >=B0 this is scatter so disable */
FRF_AZ_RX_DESCQ_JUMBO, !is_b0,
FRF_AZ_RX_DESCQ_JUMBO, jumbo_en,
FRF_AZ_RX_DESCQ_EN, 1);
efx_writeo_table(efx, &rx_desc_ptr, efx->type->rxd_ptr_tbl_base,
efx_rx_queue_index(rx_queue));
......@@ -968,13 +977,24 @@ static u16 efx_handle_rx_not_ok(struct efx_rx_queue *rx_queue,
EFX_RX_PKT_DISCARD : 0;
}
/* Handle receive events that are not in-order. */
static void
/* Handle receive events that are not in-order. Return true if this
* can be handled as a partial packet discard, false if it's more
* serious.
*/
static bool
efx_handle_rx_bad_index(struct efx_rx_queue *rx_queue, unsigned index)
{
struct efx_channel *channel = efx_rx_queue_channel(rx_queue);
struct efx_nic *efx = rx_queue->efx;
unsigned expected, dropped;
if (rx_queue->scatter_n &&
index == ((rx_queue->removed_count + rx_queue->scatter_n - 1) &
rx_queue->ptr_mask)) {
++channel->n_rx_nodesc_trunc;
return true;
}
expected = rx_queue->removed_count & rx_queue->ptr_mask;
dropped = (index - expected) & rx_queue->ptr_mask;
netif_info(efx, rx_err, efx->net_dev,
......@@ -983,6 +1003,7 @@ efx_handle_rx_bad_index(struct efx_rx_queue *rx_queue, unsigned index)
efx_schedule_reset(efx, EFX_WORKAROUND_5676(efx) ?
RESET_TYPE_RX_RECOVERY : RESET_TYPE_DISABLE);
return false;
}
/* Handle a packet received event
......@@ -998,7 +1019,7 @@ efx_handle_rx_event(struct efx_channel *channel, const efx_qword_t *event)
unsigned int rx_ev_desc_ptr, rx_ev_byte_cnt;
unsigned int rx_ev_hdr_type, rx_ev_mcast_pkt;
unsigned expected_ptr;
bool rx_ev_pkt_ok;
bool rx_ev_pkt_ok, rx_ev_sop, rx_ev_cont;
u16 flags;
struct efx_rx_queue *rx_queue;
struct efx_nic *efx = channel->efx;
......@@ -1006,21 +1027,56 @@ efx_handle_rx_event(struct efx_channel *channel, const efx_qword_t *event)
if (unlikely(ACCESS_ONCE(efx->reset_pending)))
return;
/* Basic packet information */
rx_ev_byte_cnt = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_BYTE_CNT);
rx_ev_pkt_ok = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_PKT_OK);
rx_ev_hdr_type = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_HDR_TYPE);
WARN_ON(EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_JUMBO_CONT));
WARN_ON(EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_SOP) != 1);
rx_ev_cont = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_JUMBO_CONT);
rx_ev_sop = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_SOP);
WARN_ON(EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_Q_LABEL) !=
channel->channel);
rx_queue = efx_channel_get_rx_queue(channel);
rx_ev_desc_ptr = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_DESC_PTR);
expected_ptr = rx_queue->removed_count & rx_queue->ptr_mask;
if (unlikely(rx_ev_desc_ptr != expected_ptr))
efx_handle_rx_bad_index(rx_queue, rx_ev_desc_ptr);
expected_ptr = ((rx_queue->removed_count + rx_queue->scatter_n) &
rx_queue->ptr_mask);
/* Check for partial drops and other errors */
if (unlikely(rx_ev_desc_ptr != expected_ptr) ||
unlikely(rx_ev_sop != (rx_queue->scatter_n == 0))) {
if (rx_ev_desc_ptr != expected_ptr &&
!efx_handle_rx_bad_index(rx_queue, rx_ev_desc_ptr))
return;
/* Discard all pending fragments */
if (rx_queue->scatter_n) {
efx_rx_packet(
rx_queue,
rx_queue->removed_count & rx_queue->ptr_mask,
rx_queue->scatter_n, 0, EFX_RX_PKT_DISCARD);
rx_queue->removed_count += rx_queue->scatter_n;
rx_queue->scatter_n = 0;
}
/* Return if there is no new fragment */
if (rx_ev_desc_ptr != expected_ptr)
return;
/* Discard new fragment if not SOP */
if (!rx_ev_sop) {
efx_rx_packet(
rx_queue,
rx_queue->removed_count & rx_queue->ptr_mask,
1, 0, EFX_RX_PKT_DISCARD);
++rx_queue->removed_count;
return;
}
}
++rx_queue->scatter_n;
if (rx_ev_cont)
return;
rx_ev_byte_cnt = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_BYTE_CNT);
rx_ev_pkt_ok = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_PKT_OK);
rx_ev_hdr_type = EFX_QWORD_FIELD(*event, FSF_AZ_RX_EV_HDR_TYPE);
if (likely(rx_ev_pkt_ok)) {
/* If packet is marked as OK and packet type is TCP/IP or
......@@ -1048,7 +1104,11 @@ efx_handle_rx_event(struct efx_channel *channel, const efx_qword_t *event)
channel->irq_mod_score += 2;
/* Handle received packet */
efx_rx_packet(rx_queue, rx_ev_desc_ptr, rx_ev_byte_cnt, flags);
efx_rx_packet(rx_queue,
rx_queue->removed_count & rx_queue->ptr_mask,
rx_queue->scatter_n, rx_ev_byte_cnt, flags);
rx_queue->removed_count += rx_queue->scatter_n;
rx_queue->scatter_n = 0;
}
/* If this flush done event corresponds to a &struct efx_tx_queue, then
......
......@@ -99,6 +99,9 @@
#define PTP_V2_VERSION_LENGTH 1
#define PTP_V2_VERSION_OFFSET 29
#define PTP_V2_UUID_LENGTH 8
#define PTP_V2_UUID_OFFSET 48
/* Although PTP V2 UUIDs are comprised a ClockIdentity (8) and PortNumber (2),
* the MC only captures the last six bytes of the clock identity. These values
* reflect those, not the ones used in the standard. The standard permits
......@@ -429,13 +432,10 @@ static int efx_ptp_process_times(struct efx_nic *efx, u8 *synch_buf,
unsigned number_readings = (response_length /
MC_CMD_PTP_OUT_SYNCHRONIZE_TIMESET_LEN);
unsigned i;
unsigned min;
unsigned min_set = 0;
unsigned total;
unsigned ngood = 0;
unsigned last_good = 0;
struct efx_ptp_data *ptp = efx->ptp_data;
bool min_valid = false;
u32 last_sec;
u32 start_sec;
struct timespec delta;
......@@ -443,35 +443,17 @@ static int efx_ptp_process_times(struct efx_nic *efx, u8 *synch_buf,
if (number_readings == 0)
return -EAGAIN;
/* Find minimum value in this set of results, discarding clearly
* erroneous results.
/* Read the set of results and increment stats for any results that
* appera to be erroneous.
*/
for (i = 0; i < number_readings; i++) {
efx_ptp_read_timeset(synch_buf, &ptp->timeset[i]);
synch_buf += MC_CMD_PTP_OUT_SYNCHRONIZE_TIMESET_LEN;
if (ptp->timeset[i].window > SYNCHRONISATION_GRANULARITY_NS) {
if (min_valid) {
if (ptp->timeset[i].window < min_set)
min_set = ptp->timeset[i].window;
} else {
min_valid = true;
min_set = ptp->timeset[i].window;
}
}
}
if (min_valid) {
if (ptp->base_sync_valid && (min_set > ptp->base_sync_ns))
min = ptp->base_sync_ns;
else
min = min_set;
} else {
min = SYNCHRONISATION_GRANULARITY_NS;
}
/* Discard excessively long synchronise durations. The MC times
* when it finishes reading the host time so the corrected window
* time should be fairly constant for a given platform.
/* 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.
*/
total = 0;
for (i = 0; i < number_readings; i++)
......@@ -489,8 +471,8 @@ static int efx_ptp_process_times(struct efx_nic *efx, u8 *synch_buf,
if (ngood == 0) {
netif_warn(efx, drv, efx->net_dev,
"PTP no suitable synchronisations %dns %dns\n",
ptp->base_sync_ns, min_set);
"PTP no suitable synchronisations %dns\n",
ptp->base_sync_ns);
return -EAGAIN;
}
......@@ -1006,43 +988,53 @@ bool efx_ptp_is_ptp_tx(struct efx_nic *efx, struct sk_buff *skb)
* the receive timestamp from the MC - this will probably occur after the
* packet arrival because of the processing in the MC.
*/
static void efx_ptp_rx(struct efx_channel *channel, struct sk_buff *skb)
static bool efx_ptp_rx(struct efx_channel *channel, struct sk_buff *skb)
{
struct efx_nic *efx = channel->efx;
struct efx_ptp_data *ptp = efx->ptp_data;
struct efx_ptp_match *match = (struct efx_ptp_match *)skb->cb;
u8 *data;
u8 *match_data_012, *match_data_345;
unsigned int version;
match->expiry = jiffies + msecs_to_jiffies(PKT_EVENT_LIFETIME_MS);
/* Correct version? */
if (ptp->mode == MC_CMD_PTP_MODE_V1) {
if (skb->len < PTP_V1_MIN_LENGTH) {
netif_receive_skb(skb);
return;
if (!pskb_may_pull(skb, PTP_V1_MIN_LENGTH)) {
return false;
}
version = ntohs(*(__be16 *)&skb->data[PTP_V1_VERSION_OFFSET]);
if (version != PTP_VERSION_V1) {
netif_receive_skb(skb);
return;
return false;
}
/* PTP V1 uses all six bytes of the UUID to match the packet
* to the timestamp
*/
match_data_012 = skb->data + PTP_V1_UUID_OFFSET;
match_data_345 = skb->data + PTP_V1_UUID_OFFSET + 3;
} else {
if (skb->len < PTP_V2_MIN_LENGTH) {
netif_receive_skb(skb);
return;
if (!pskb_may_pull(skb, PTP_V2_MIN_LENGTH)) {
return false;
}
version = skb->data[PTP_V2_VERSION_OFFSET];
BUG_ON(ptp->mode != MC_CMD_PTP_MODE_V2);
BUILD_BUG_ON(PTP_V1_UUID_OFFSET != PTP_V2_MC_UUID_OFFSET);
BUILD_BUG_ON(PTP_V1_UUID_LENGTH != PTP_V2_MC_UUID_LENGTH);
BUILD_BUG_ON(PTP_V1_SEQUENCE_OFFSET != PTP_V2_SEQUENCE_OFFSET);
BUILD_BUG_ON(PTP_V1_SEQUENCE_LENGTH != PTP_V2_SEQUENCE_LENGTH);
if ((version & PTP_VERSION_V2_MASK) != PTP_VERSION_V2) {
netif_receive_skb(skb);
return;
return false;
}
/* The original V2 implementation uses bytes 2-7 of
* the UUID to match the packet to the timestamp. This
* discards two of the bytes of the MAC address used
* to create the UUID (SF bug 33070). The PTP V2
* enhanced mode fixes this issue and uses bytes 0-2
* and byte 5-7 of the UUID.
*/
match_data_345 = skb->data + PTP_V2_UUID_OFFSET + 5;
if (ptp->mode == MC_CMD_PTP_MODE_V2) {
match_data_012 = skb->data + PTP_V2_UUID_OFFSET + 2;
} else {
match_data_012 = skb->data + PTP_V2_UUID_OFFSET + 0;
BUG_ON(ptp->mode != MC_CMD_PTP_MODE_V2_ENHANCED);
}
}
......@@ -1056,14 +1048,19 @@ static void efx_ptp_rx(struct efx_channel *channel, struct sk_buff *skb)
timestamps = skb_hwtstamps(skb);
memset(timestamps, 0, sizeof(*timestamps));
/* We expect the sequence number to be in the same position in
* the packet for PTP V1 and V2
*/
BUILD_BUG_ON(PTP_V1_SEQUENCE_OFFSET != PTP_V2_SEQUENCE_OFFSET);
BUILD_BUG_ON(PTP_V1_SEQUENCE_LENGTH != PTP_V2_SEQUENCE_LENGTH);
/* Extract UUID/Sequence information */
data = skb->data + PTP_V1_UUID_OFFSET;
match->words[0] = (data[0] |
(data[1] << 8) |
(data[2] << 16) |
(data[3] << 24));
match->words[1] = (data[4] |
(data[5] << 8) |
match->words[0] = (match_data_012[0] |
(match_data_012[1] << 8) |
(match_data_012[2] << 16) |
(match_data_345[0] << 24));
match->words[1] = (match_data_345[1] |
(match_data_345[2] << 8) |
(skb->data[PTP_V1_SEQUENCE_OFFSET +
PTP_V1_SEQUENCE_LENGTH - 1] <<
16));
......@@ -1073,6 +1070,8 @@ static void efx_ptp_rx(struct efx_channel *channel, struct sk_buff *skb)
skb_queue_tail(&ptp->rxq, skb);
queue_work(ptp->workwq, &ptp->work);
return true;
}
/* Transmit a PTP packet. This has to be transmitted by the MC
......@@ -1167,7 +1166,7 @@ static int efx_ptp_ts_init(struct efx_nic *efx, struct hwtstamp_config *init)
* timestamped
*/
init->rx_filter = HWTSTAMP_FILTER_PTP_V2_L4_EVENT;
new_mode = MC_CMD_PTP_MODE_V2;
new_mode = MC_CMD_PTP_MODE_V2_ENHANCED;
enable_wanted = true;
break;
case HWTSTAMP_FILTER_PTP_V2_EVENT:
......@@ -1186,7 +1185,14 @@ static int efx_ptp_ts_init(struct efx_nic *efx, struct hwtstamp_config *init)
if (init->tx_type != HWTSTAMP_TX_OFF)
enable_wanted = true;
/* Old versions of the firmware do not support the improved
* UUID filtering option (SF bug 33070). If the firmware does
* not accept the enhanced mode, fall back to the standard PTP
* v2 UUID filtering.
*/
rc = efx_ptp_change_mode(efx, enable_wanted, new_mode);
if ((rc != 0) && (new_mode == MC_CMD_PTP_MODE_V2_ENHANCED))
rc = efx_ptp_change_mode(efx, enable_wanted, MC_CMD_PTP_MODE_V2);
if (rc != 0)
return rc;
......
This diff is collapsed.
......@@ -202,7 +202,7 @@ static int siena_test_chip(struct efx_nic *efx, struct efx_self_tests *tests)
static enum reset_type siena_map_reset_reason(enum reset_type reason)
{
return RESET_TYPE_ALL;
return RESET_TYPE_RECOVER_OR_ALL;
}
static int siena_map_reset_flags(u32 *flags)
......@@ -245,6 +245,22 @@ static int siena_reset_hw(struct efx_nic *efx, enum reset_type method)
return efx_mcdi_reset_port(efx);
}
#ifdef CONFIG_EEH
/* When a PCI device is isolated from the bus, a subsequent MMIO read is
* required for the kernel EEH mechanisms to notice. As the Solarflare driver
* was written to minimise MMIO read (for latency) then a periodic call to check
* the EEH status of the device is required so that device recovery can happen
* in a timely fashion.
*/
static void siena_monitor(struct efx_nic *efx)
{
struct eeh_dev *eehdev =
of_node_to_eeh_dev(pci_device_to_OF_node(efx->pci_dev));
eeh_dev_check_failure(eehdev);
}
#endif
static int siena_probe_nvconfig(struct efx_nic *efx)
{
u32 caps = 0;
......@@ -398,6 +414,8 @@ static int siena_init_nic(struct efx_nic *efx)
EFX_SET_OWORD_FIELD(temp, FRF_BZ_RX_HASH_INSRT_HDR, 1);
EFX_SET_OWORD_FIELD(temp, FRF_BZ_RX_HASH_ALG, 1);
EFX_SET_OWORD_FIELD(temp, FRF_BZ_RX_IP_HASH, 1);
EFX_SET_OWORD_FIELD(temp, FRF_BZ_RX_USR_BUF_SIZE,
EFX_RX_USR_BUF_SIZE >> 5);
efx_writeo(efx, &temp, FR_AZ_RX_CFG);
/* Set hash key for IPv4 */
......@@ -665,7 +683,11 @@ const struct efx_nic_type siena_a0_nic_type = {
.init = siena_init_nic,
.dimension_resources = siena_dimension_resources,
.fini = efx_port_dummy_op_void,
#ifdef CONFIG_EEH
.monitor = siena_monitor,
#else
.monitor = NULL,
#endif
.map_reset_reason = siena_map_reset_reason,
.map_reset_flags = siena_map_reset_flags,
.reset = siena_reset_hw,
......@@ -698,6 +720,7 @@ const struct efx_nic_type siena_a0_nic_type = {
.max_dma_mask = DMA_BIT_MASK(FSF_AZ_TX_KER_BUF_ADDR_WIDTH),
.rx_buffer_hash_size = 0x10,
.rx_buffer_padding = 0,
.can_rx_scatter = true,
.max_interrupt_mode = EFX_INT_MODE_MSIX,
.phys_addr_channels = 32, /* Hardware limit is 64, but the legacy
* interrupt handler only supports 32
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment