linux/drivers/net/ethernet/atheros/alx/main.c

1606 lines
38 KiB
C
Raw Normal View History

/*
* Copyright (c) 2013 Johannes Berg <johannes@sipsolutions.net>
*
* This file is free software: you may copy, redistribute and/or modify it
* under the terms of the GNU General Public License as published by the
* Free Software Foundation, either version 2 of the License, or (at your
* option) any later version.
*
* This file is distributed in the hope that it will be useful, but
* WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*
* This file incorporates work covered by the following copyright and
* permission notice:
*
* Copyright (c) 2012 Qualcomm Atheros, Inc.
*
* Permission to use, copy, modify, and/or distribute this software for any
* purpose with or without fee is hereby granted, provided that the above
* copyright notice and this permission notice appear in all copies.
*
* THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
* WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
* MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
* ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
* WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
* ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
* OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
*/
#include <linux/module.h>
#include <linux/pci.h>
#include <linux/interrupt.h>
#include <linux/ip.h>
#include <linux/ipv6.h>
#include <linux/if_vlan.h>
#include <linux/mdio.h>
#include <linux/aer.h>
#include <linux/bitops.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <net/ip6_checksum.h>
#include <linux/crc32.h>
#include "alx.h"
#include "hw.h"
#include "reg.h"
const char alx_drv_name[] = "alx";
static void alx_free_txbuf(struct alx_priv *alx, int entry)
{
struct alx_buffer *txb = &alx->txq.bufs[entry];
if (dma_unmap_len(txb, size)) {
dma_unmap_single(&alx->hw.pdev->dev,
dma_unmap_addr(txb, dma),
dma_unmap_len(txb, size),
DMA_TO_DEVICE);
dma_unmap_len_set(txb, size, 0);
}
if (txb->skb) {
dev_kfree_skb_any(txb->skb);
txb->skb = NULL;
}
}
net: alx: use custom skb allocator This patch follows Eric Dumazet's commit 7b70176421 for Atheros atl1c driver to fix one exactly same bug in alx driver, that the network link will be lost in 1-5 minutes after the device is up. My laptop Lenovo Y580 with Atheros AR8161 ethernet device hit the same problem with kernel 4.4, and it will be cured by Jarod Wilson's commit c406700c for alx driver which get merged in 4.5. But there are still some alx devices can't function well even with Jarod's patch, while this patch could make them work fine. More details on https://bugzilla.kernel.org/show_bug.cgi?id=70761 The debug shows the issue is very likely to be related with the RX DMA address, specifically 0x...f80, if RX buffer get 0x...f80 several times, their will be RX overflow error and device will stop working. For kernel 4.5.0 with Jarod's patch which works fine with my AR8161/Lennov Y580, if I made some change to the __netdev_alloc_skb --> __alloc_page_frag() to make the allocated buffer can get an address with 0x...f80, then the same error happens. If I make it to 0x...f40 or 0x....fc0, everything will be still fine. So I tend to believe that the 0x..f80 address cause the silicon to behave abnormally. Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=70761 Cc: Eric Dumazet <edumazet@google.com> Cc: Johannes Berg <johannes@sipsolutions.net> Cc: Jarod Wilson <jarod@redhat.com> Signed-off-by: Feng Tang <feng.tang@intel.com> Tested-by: Ole Lukoie <olelukoie@mail.ru> Signed-off-by: David S. Miller <davem@davemloft.net>
2016-05-25 14:49:54 +08:00
static struct sk_buff *alx_alloc_skb(struct alx_priv *alx, gfp_t gfp)
{
struct sk_buff *skb;
struct page *page;
if (alx->rx_frag_size > PAGE_SIZE)
return __netdev_alloc_skb(alx->dev, alx->rxbuf_size, gfp);
page = alx->rx_page;
if (!page) {
alx->rx_page = page = alloc_page(gfp);
if (unlikely(!page))
return NULL;
alx->rx_page_offset = 0;
}
skb = build_skb(page_address(page) + alx->rx_page_offset,
alx->rx_frag_size);
if (likely(skb)) {
alx->rx_page_offset += alx->rx_frag_size;
if (alx->rx_page_offset >= PAGE_SIZE)
alx->rx_page = NULL;
else
get_page(page);
}
return skb;
}
static int alx_refill_rx_ring(struct alx_priv *alx, gfp_t gfp)
{
struct alx_rx_queue *rxq = &alx->rxq;
struct sk_buff *skb;
struct alx_buffer *cur_buf;
dma_addr_t dma;
u16 cur, next, count = 0;
next = cur = rxq->write_idx;
if (++next == alx->rx_ringsz)
next = 0;
cur_buf = &rxq->bufs[cur];
while (!cur_buf->skb && next != rxq->read_idx) {
struct alx_rfd *rfd = &rxq->rfd[cur];
net: alx: use custom skb allocator This patch follows Eric Dumazet's commit 7b70176421 for Atheros atl1c driver to fix one exactly same bug in alx driver, that the network link will be lost in 1-5 minutes after the device is up. My laptop Lenovo Y580 with Atheros AR8161 ethernet device hit the same problem with kernel 4.4, and it will be cured by Jarod Wilson's commit c406700c for alx driver which get merged in 4.5. But there are still some alx devices can't function well even with Jarod's patch, while this patch could make them work fine. More details on https://bugzilla.kernel.org/show_bug.cgi?id=70761 The debug shows the issue is very likely to be related with the RX DMA address, specifically 0x...f80, if RX buffer get 0x...f80 several times, their will be RX overflow error and device will stop working. For kernel 4.5.0 with Jarod's patch which works fine with my AR8161/Lennov Y580, if I made some change to the __netdev_alloc_skb --> __alloc_page_frag() to make the allocated buffer can get an address with 0x...f80, then the same error happens. If I make it to 0x...f40 or 0x....fc0, everything will be still fine. So I tend to believe that the 0x..f80 address cause the silicon to behave abnormally. Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=70761 Cc: Eric Dumazet <edumazet@google.com> Cc: Johannes Berg <johannes@sipsolutions.net> Cc: Jarod Wilson <jarod@redhat.com> Signed-off-by: Feng Tang <feng.tang@intel.com> Tested-by: Ole Lukoie <olelukoie@mail.ru> Signed-off-by: David S. Miller <davem@davemloft.net>
2016-05-25 14:49:54 +08:00
skb = alx_alloc_skb(alx, gfp);
if (!skb)
break;
dma = dma_map_single(&alx->hw.pdev->dev,
skb->data, alx->rxbuf_size,
DMA_FROM_DEVICE);
if (dma_mapping_error(&alx->hw.pdev->dev, dma)) {
dev_kfree_skb(skb);
break;
}
/* Unfortunately, RX descriptor buffers must be 4-byte
* aligned, so we can't use IP alignment.
*/
if (WARN_ON(dma & 3)) {
dev_kfree_skb(skb);
break;
}
cur_buf->skb = skb;
dma_unmap_len_set(cur_buf, size, alx->rxbuf_size);
dma_unmap_addr_set(cur_buf, dma, dma);
rfd->addr = cpu_to_le64(dma);
cur = next;
if (++next == alx->rx_ringsz)
next = 0;
cur_buf = &rxq->bufs[cur];
count++;
}
if (count) {
/* flush all updates before updating hardware */
wmb();
rxq->write_idx = cur;
alx_write_mem16(&alx->hw, ALX_RFD_PIDX, cur);
}
net: alx: use custom skb allocator This patch follows Eric Dumazet's commit 7b70176421 for Atheros atl1c driver to fix one exactly same bug in alx driver, that the network link will be lost in 1-5 minutes after the device is up. My laptop Lenovo Y580 with Atheros AR8161 ethernet device hit the same problem with kernel 4.4, and it will be cured by Jarod Wilson's commit c406700c for alx driver which get merged in 4.5. But there are still some alx devices can't function well even with Jarod's patch, while this patch could make them work fine. More details on https://bugzilla.kernel.org/show_bug.cgi?id=70761 The debug shows the issue is very likely to be related with the RX DMA address, specifically 0x...f80, if RX buffer get 0x...f80 several times, their will be RX overflow error and device will stop working. For kernel 4.5.0 with Jarod's patch which works fine with my AR8161/Lennov Y580, if I made some change to the __netdev_alloc_skb --> __alloc_page_frag() to make the allocated buffer can get an address with 0x...f80, then the same error happens. If I make it to 0x...f40 or 0x....fc0, everything will be still fine. So I tend to believe that the 0x..f80 address cause the silicon to behave abnormally. Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=70761 Cc: Eric Dumazet <edumazet@google.com> Cc: Johannes Berg <johannes@sipsolutions.net> Cc: Jarod Wilson <jarod@redhat.com> Signed-off-by: Feng Tang <feng.tang@intel.com> Tested-by: Ole Lukoie <olelukoie@mail.ru> Signed-off-by: David S. Miller <davem@davemloft.net>
2016-05-25 14:49:54 +08:00
return count;
}
static inline int alx_tpd_avail(struct alx_priv *alx)
{
struct alx_tx_queue *txq = &alx->txq;
if (txq->write_idx >= txq->read_idx)
return alx->tx_ringsz + txq->read_idx - txq->write_idx - 1;
return txq->read_idx - txq->write_idx - 1;
}
static bool alx_clean_tx_irq(struct alx_priv *alx)
{
struct alx_tx_queue *txq = &alx->txq;
u16 hw_read_idx, sw_read_idx;
unsigned int total_bytes = 0, total_packets = 0;
int budget = ALX_DEFAULT_TX_WORK;
sw_read_idx = txq->read_idx;
hw_read_idx = alx_read_mem16(&alx->hw, ALX_TPD_PRI0_CIDX);
if (sw_read_idx != hw_read_idx) {
while (sw_read_idx != hw_read_idx && budget > 0) {
struct sk_buff *skb;
skb = txq->bufs[sw_read_idx].skb;
if (skb) {
total_bytes += skb->len;
total_packets++;
budget--;
}
alx_free_txbuf(alx, sw_read_idx);
if (++sw_read_idx == alx->tx_ringsz)
sw_read_idx = 0;
}
txq->read_idx = sw_read_idx;
netdev_completed_queue(alx->dev, total_packets, total_bytes);
}
if (netif_queue_stopped(alx->dev) && netif_carrier_ok(alx->dev) &&
alx_tpd_avail(alx) > alx->tx_ringsz/4)
netif_wake_queue(alx->dev);
return sw_read_idx == hw_read_idx;
}
static void alx_schedule_link_check(struct alx_priv *alx)
{
schedule_work(&alx->link_check_wk);
}
static void alx_schedule_reset(struct alx_priv *alx)
{
schedule_work(&alx->reset_wk);
}
static int alx_clean_rx_irq(struct alx_priv *alx, int budget)
{
struct alx_rx_queue *rxq = &alx->rxq;
struct alx_rrd *rrd;
struct alx_buffer *rxb;
struct sk_buff *skb;
u16 length, rfd_cleaned = 0;
int work = 0;
while (work < budget) {
rrd = &rxq->rrd[rxq->rrd_read_idx];
if (!(rrd->word3 & cpu_to_le32(1 << RRD_UPDATED_SHIFT)))
break;
rrd->word3 &= ~cpu_to_le32(1 << RRD_UPDATED_SHIFT);
if (ALX_GET_FIELD(le32_to_cpu(rrd->word0),
RRD_SI) != rxq->read_idx ||
ALX_GET_FIELD(le32_to_cpu(rrd->word0),
RRD_NOR) != 1) {
alx_schedule_reset(alx);
return work;
}
rxb = &rxq->bufs[rxq->read_idx];
dma_unmap_single(&alx->hw.pdev->dev,
dma_unmap_addr(rxb, dma),
dma_unmap_len(rxb, size),
DMA_FROM_DEVICE);
dma_unmap_len_set(rxb, size, 0);
skb = rxb->skb;
rxb->skb = NULL;
if (rrd->word3 & cpu_to_le32(1 << RRD_ERR_RES_SHIFT) ||
rrd->word3 & cpu_to_le32(1 << RRD_ERR_LEN_SHIFT)) {
rrd->word3 = 0;
dev_kfree_skb_any(skb);
goto next_pkt;
}
length = ALX_GET_FIELD(le32_to_cpu(rrd->word3),
RRD_PKTLEN) - ETH_FCS_LEN;
skb_put(skb, length);
skb->protocol = eth_type_trans(skb, alx->dev);
skb_checksum_none_assert(skb);
if (alx->dev->features & NETIF_F_RXCSUM &&
!(rrd->word3 & (cpu_to_le32(1 << RRD_ERR_L4_SHIFT) |
cpu_to_le32(1 << RRD_ERR_IPV4_SHIFT)))) {
switch (ALX_GET_FIELD(le32_to_cpu(rrd->word2),
RRD_PID)) {
case RRD_PID_IPV6UDP:
case RRD_PID_IPV4UDP:
case RRD_PID_IPV4TCP:
case RRD_PID_IPV6TCP:
skb->ip_summed = CHECKSUM_UNNECESSARY;
break;
}
}
napi_gro_receive(&alx->napi, skb);
work++;
next_pkt:
if (++rxq->read_idx == alx->rx_ringsz)
rxq->read_idx = 0;
if (++rxq->rrd_read_idx == alx->rx_ringsz)
rxq->rrd_read_idx = 0;
if (++rfd_cleaned > ALX_RX_ALLOC_THRESH)
rfd_cleaned -= alx_refill_rx_ring(alx, GFP_ATOMIC);
}
if (rfd_cleaned)
alx_refill_rx_ring(alx, GFP_ATOMIC);
return work;
}
static int alx_poll(struct napi_struct *napi, int budget)
{
struct alx_priv *alx = container_of(napi, struct alx_priv, napi);
struct alx_hw *hw = &alx->hw;
unsigned long flags;
bool tx_complete;
int work;
tx_complete = alx_clean_tx_irq(alx);
work = alx_clean_rx_irq(alx, budget);
if (!tx_complete || work == budget)
return budget;
napi_complete(&alx->napi);
/* enable interrupt */
spin_lock_irqsave(&alx->irq_lock, flags);
alx->int_mask |= ALX_ISR_TX_Q0 | ALX_ISR_RX_Q0;
alx_write_mem32(hw, ALX_IMR, alx->int_mask);
spin_unlock_irqrestore(&alx->irq_lock, flags);
alx_post_write(hw);
return work;
}
static irqreturn_t alx_intr_handle(struct alx_priv *alx, u32 intr)
{
struct alx_hw *hw = &alx->hw;
bool write_int_mask = false;
spin_lock(&alx->irq_lock);
/* ACK interrupt */
alx_write_mem32(hw, ALX_ISR, intr | ALX_ISR_DIS);
intr &= alx->int_mask;
if (intr & ALX_ISR_FATAL) {
netif_warn(alx, hw, alx->dev,
"fatal interrupt 0x%x, resetting\n", intr);
alx_schedule_reset(alx);
goto out;
}
if (intr & ALX_ISR_ALERT)
netdev_warn(alx->dev, "alert interrupt: 0x%x\n", intr);
if (intr & ALX_ISR_PHY) {
/* suppress PHY interrupt, because the source
* is from PHY internal. only the internal status
* is cleared, the interrupt status could be cleared.
*/
alx->int_mask &= ~ALX_ISR_PHY;
write_int_mask = true;
alx_schedule_link_check(alx);
}
if (intr & (ALX_ISR_TX_Q0 | ALX_ISR_RX_Q0)) {
napi_schedule(&alx->napi);
/* mask rx/tx interrupt, enable them when napi complete */
alx->int_mask &= ~ALX_ISR_ALL_QUEUES;
write_int_mask = true;
}
if (write_int_mask)
alx_write_mem32(hw, ALX_IMR, alx->int_mask);
alx_write_mem32(hw, ALX_ISR, 0);
out:
spin_unlock(&alx->irq_lock);
return IRQ_HANDLED;
}
static irqreturn_t alx_intr_msi(int irq, void *data)
{
struct alx_priv *alx = data;
return alx_intr_handle(alx, alx_read_mem32(&alx->hw, ALX_ISR));
}
static irqreturn_t alx_intr_legacy(int irq, void *data)
{
struct alx_priv *alx = data;
struct alx_hw *hw = &alx->hw;
u32 intr;
intr = alx_read_mem32(hw, ALX_ISR);
if (intr & ALX_ISR_DIS || !(intr & alx->int_mask))
return IRQ_NONE;
return alx_intr_handle(alx, intr);
}
static void alx_init_ring_ptrs(struct alx_priv *alx)
{
struct alx_hw *hw = &alx->hw;
u32 addr_hi = ((u64)alx->descmem.dma) >> 32;
alx->rxq.read_idx = 0;
alx->rxq.write_idx = 0;
alx->rxq.rrd_read_idx = 0;
alx_write_mem32(hw, ALX_RX_BASE_ADDR_HI, addr_hi);
alx_write_mem32(hw, ALX_RRD_ADDR_LO, alx->rxq.rrd_dma);
alx_write_mem32(hw, ALX_RRD_RING_SZ, alx->rx_ringsz);
alx_write_mem32(hw, ALX_RFD_ADDR_LO, alx->rxq.rfd_dma);
alx_write_mem32(hw, ALX_RFD_RING_SZ, alx->rx_ringsz);
alx_write_mem32(hw, ALX_RFD_BUF_SZ, alx->rxbuf_size);
alx->txq.read_idx = 0;
alx->txq.write_idx = 0;
alx_write_mem32(hw, ALX_TX_BASE_ADDR_HI, addr_hi);
alx_write_mem32(hw, ALX_TPD_PRI0_ADDR_LO, alx->txq.tpd_dma);
alx_write_mem32(hw, ALX_TPD_RING_SZ, alx->tx_ringsz);
/* load these pointers into the chip */
alx_write_mem32(hw, ALX_SRAM9, ALX_SRAM_LOAD_PTR);
}
static void alx_free_txring_buf(struct alx_priv *alx)
{
struct alx_tx_queue *txq = &alx->txq;
int i;
if (!txq->bufs)
return;
for (i = 0; i < alx->tx_ringsz; i++)
alx_free_txbuf(alx, i);
memset(txq->bufs, 0, alx->tx_ringsz * sizeof(struct alx_buffer));
memset(txq->tpd, 0, alx->tx_ringsz * sizeof(struct alx_txd));
txq->write_idx = 0;
txq->read_idx = 0;
netdev_reset_queue(alx->dev);
}
static void alx_free_rxring_buf(struct alx_priv *alx)
{
struct alx_rx_queue *rxq = &alx->rxq;
struct alx_buffer *cur_buf;
u16 i;
if (rxq == NULL)
return;
for (i = 0; i < alx->rx_ringsz; i++) {
cur_buf = rxq->bufs + i;
if (cur_buf->skb) {
dma_unmap_single(&alx->hw.pdev->dev,
dma_unmap_addr(cur_buf, dma),
dma_unmap_len(cur_buf, size),
DMA_FROM_DEVICE);
dev_kfree_skb(cur_buf->skb);
cur_buf->skb = NULL;
dma_unmap_len_set(cur_buf, size, 0);
dma_unmap_addr_set(cur_buf, dma, 0);
}
}
rxq->write_idx = 0;
rxq->read_idx = 0;
rxq->rrd_read_idx = 0;
}
static void alx_free_buffers(struct alx_priv *alx)
{
alx_free_txring_buf(alx);
alx_free_rxring_buf(alx);
}
static int alx_reinit_rings(struct alx_priv *alx)
{
alx_free_buffers(alx);
alx_init_ring_ptrs(alx);
if (!alx_refill_rx_ring(alx, GFP_KERNEL))
return -ENOMEM;
return 0;
}
static void alx_add_mc_addr(struct alx_hw *hw, const u8 *addr, u32 *mc_hash)
{
u32 crc32, bit, reg;
crc32 = ether_crc(ETH_ALEN, addr);
reg = (crc32 >> 31) & 0x1;
bit = (crc32 >> 26) & 0x1F;
mc_hash[reg] |= BIT(bit);
}
static void __alx_set_rx_mode(struct net_device *netdev)
{
struct alx_priv *alx = netdev_priv(netdev);
struct alx_hw *hw = &alx->hw;
struct netdev_hw_addr *ha;
u32 mc_hash[2] = {};
if (!(netdev->flags & IFF_ALLMULTI)) {
netdev_for_each_mc_addr(ha, netdev)
alx_add_mc_addr(hw, ha->addr, mc_hash);
alx_write_mem32(hw, ALX_HASH_TBL0, mc_hash[0]);
alx_write_mem32(hw, ALX_HASH_TBL1, mc_hash[1]);
}
hw->rx_ctrl &= ~(ALX_MAC_CTRL_MULTIALL_EN | ALX_MAC_CTRL_PROMISC_EN);
if (netdev->flags & IFF_PROMISC)
hw->rx_ctrl |= ALX_MAC_CTRL_PROMISC_EN;
if (netdev->flags & IFF_ALLMULTI)
hw->rx_ctrl |= ALX_MAC_CTRL_MULTIALL_EN;
alx_write_mem32(hw, ALX_MAC_CTRL, hw->rx_ctrl);
}
static void alx_set_rx_mode(struct net_device *netdev)
{
__alx_set_rx_mode(netdev);
}
static int alx_set_mac_address(struct net_device *netdev, void *data)
{
struct alx_priv *alx = netdev_priv(netdev);
struct alx_hw *hw = &alx->hw;
struct sockaddr *addr = data;
if (!is_valid_ether_addr(addr->sa_data))
return -EADDRNOTAVAIL;
if (netdev->addr_assign_type & NET_ADDR_RANDOM)
netdev->addr_assign_type ^= NET_ADDR_RANDOM;
memcpy(netdev->dev_addr, addr->sa_data, netdev->addr_len);
memcpy(hw->mac_addr, addr->sa_data, netdev->addr_len);
alx_set_macaddr(hw, hw->mac_addr);
return 0;
}
static int alx_alloc_descriptors(struct alx_priv *alx)
{
alx->txq.bufs = kcalloc(alx->tx_ringsz,
sizeof(struct alx_buffer),
GFP_KERNEL);
if (!alx->txq.bufs)
return -ENOMEM;
alx->rxq.bufs = kcalloc(alx->rx_ringsz,
sizeof(struct alx_buffer),
GFP_KERNEL);
if (!alx->rxq.bufs)
goto out_free;
/* physical tx/rx ring descriptors
*
* Allocate them as a single chunk because they must not cross a
* 4G boundary (hardware has a single register for high 32 bits
* of addresses only)
*/
alx->descmem.size = sizeof(struct alx_txd) * alx->tx_ringsz +
sizeof(struct alx_rrd) * alx->rx_ringsz +
sizeof(struct alx_rfd) * alx->rx_ringsz;
alx->descmem.virt = dma_zalloc_coherent(&alx->hw.pdev->dev,
alx->descmem.size,
&alx->descmem.dma,
GFP_KERNEL);
if (!alx->descmem.virt)
goto out_free;
alx->txq.tpd = alx->descmem.virt;
alx->txq.tpd_dma = alx->descmem.dma;
/* alignment requirement for next block */
BUILD_BUG_ON(sizeof(struct alx_txd) % 8);
alx->rxq.rrd =
(void *)((u8 *)alx->descmem.virt +
sizeof(struct alx_txd) * alx->tx_ringsz);
alx->rxq.rrd_dma = alx->descmem.dma +
sizeof(struct alx_txd) * alx->tx_ringsz;
/* alignment requirement for next block */
BUILD_BUG_ON(sizeof(struct alx_rrd) % 8);
alx->rxq.rfd =
(void *)((u8 *)alx->descmem.virt +
sizeof(struct alx_txd) * alx->tx_ringsz +
sizeof(struct alx_rrd) * alx->rx_ringsz);
alx->rxq.rfd_dma = alx->descmem.dma +
sizeof(struct alx_txd) * alx->tx_ringsz +
sizeof(struct alx_rrd) * alx->rx_ringsz;
return 0;
out_free:
kfree(alx->txq.bufs);
kfree(alx->rxq.bufs);
return -ENOMEM;
}
static int alx_alloc_rings(struct alx_priv *alx)
{
int err;
err = alx_alloc_descriptors(alx);
if (err)
return err;
alx->int_mask &= ~ALX_ISR_ALL_QUEUES;
alx->int_mask |= ALX_ISR_TX_Q0 | ALX_ISR_RX_Q0;
netif_napi_add(alx->dev, &alx->napi, alx_poll, 64);
alx_reinit_rings(alx);
return 0;
}
static void alx_free_rings(struct alx_priv *alx)
{
netif_napi_del(&alx->napi);
alx_free_buffers(alx);
kfree(alx->txq.bufs);
kfree(alx->rxq.bufs);
net: alx: use custom skb allocator This patch follows Eric Dumazet's commit 7b70176421 for Atheros atl1c driver to fix one exactly same bug in alx driver, that the network link will be lost in 1-5 minutes after the device is up. My laptop Lenovo Y580 with Atheros AR8161 ethernet device hit the same problem with kernel 4.4, and it will be cured by Jarod Wilson's commit c406700c for alx driver which get merged in 4.5. But there are still some alx devices can't function well even with Jarod's patch, while this patch could make them work fine. More details on https://bugzilla.kernel.org/show_bug.cgi?id=70761 The debug shows the issue is very likely to be related with the RX DMA address, specifically 0x...f80, if RX buffer get 0x...f80 several times, their will be RX overflow error and device will stop working. For kernel 4.5.0 with Jarod's patch which works fine with my AR8161/Lennov Y580, if I made some change to the __netdev_alloc_skb --> __alloc_page_frag() to make the allocated buffer can get an address with 0x...f80, then the same error happens. If I make it to 0x...f40 or 0x....fc0, everything will be still fine. So I tend to believe that the 0x..f80 address cause the silicon to behave abnormally. Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=70761 Cc: Eric Dumazet <edumazet@google.com> Cc: Johannes Berg <johannes@sipsolutions.net> Cc: Jarod Wilson <jarod@redhat.com> Signed-off-by: Feng Tang <feng.tang@intel.com> Tested-by: Ole Lukoie <olelukoie@mail.ru> Signed-off-by: David S. Miller <davem@davemloft.net>
2016-05-25 14:49:54 +08:00
if (alx->rx_page) {
put_page(alx->rx_page);
alx->rx_page = NULL;
}
dma_free_coherent(&alx->hw.pdev->dev,
alx->descmem.size,
alx->descmem.virt,
alx->descmem.dma);
}
static void alx_config_vector_mapping(struct alx_priv *alx)
{
struct alx_hw *hw = &alx->hw;
alx_write_mem32(hw, ALX_MSI_MAP_TBL1, 0);
alx_write_mem32(hw, ALX_MSI_MAP_TBL2, 0);
alx_write_mem32(hw, ALX_MSI_ID_MAP, 0);
}
static void alx_irq_enable(struct alx_priv *alx)
{
struct alx_hw *hw = &alx->hw;
/* level-1 interrupt switch */
alx_write_mem32(hw, ALX_ISR, 0);
alx_write_mem32(hw, ALX_IMR, alx->int_mask);
alx_post_write(hw);
}
static void alx_irq_disable(struct alx_priv *alx)
{
struct alx_hw *hw = &alx->hw;
alx_write_mem32(hw, ALX_ISR, ALX_ISR_DIS);
alx_write_mem32(hw, ALX_IMR, 0);
alx_post_write(hw);
synchronize_irq(alx->hw.pdev->irq);
}
static int alx_request_irq(struct alx_priv *alx)
{
struct pci_dev *pdev = alx->hw.pdev;
struct alx_hw *hw = &alx->hw;
int err;
u32 msi_ctrl;
msi_ctrl = (hw->imt >> 1) << ALX_MSI_RETRANS_TM_SHIFT;
if (!pci_enable_msi(alx->hw.pdev)) {
alx->msi = true;
alx_write_mem32(hw, ALX_MSI_RETRANS_TIMER,
msi_ctrl | ALX_MSI_MASK_SEL_LINE);
err = request_irq(pdev->irq, alx_intr_msi, 0,
alx->dev->name, alx);
if (!err)
goto out;
net: alx: use custom skb allocator This patch follows Eric Dumazet's commit 7b70176421 for Atheros atl1c driver to fix one exactly same bug in alx driver, that the network link will be lost in 1-5 minutes after the device is up. My laptop Lenovo Y580 with Atheros AR8161 ethernet device hit the same problem with kernel 4.4, and it will be cured by Jarod Wilson's commit c406700c for alx driver which get merged in 4.5. But there are still some alx devices can't function well even with Jarod's patch, while this patch could make them work fine. More details on https://bugzilla.kernel.org/show_bug.cgi?id=70761 The debug shows the issue is very likely to be related with the RX DMA address, specifically 0x...f80, if RX buffer get 0x...f80 several times, their will be RX overflow error and device will stop working. For kernel 4.5.0 with Jarod's patch which works fine with my AR8161/Lennov Y580, if I made some change to the __netdev_alloc_skb --> __alloc_page_frag() to make the allocated buffer can get an address with 0x...f80, then the same error happens. If I make it to 0x...f40 or 0x....fc0, everything will be still fine. So I tend to believe that the 0x..f80 address cause the silicon to behave abnormally. Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=70761 Cc: Eric Dumazet <edumazet@google.com> Cc: Johannes Berg <johannes@sipsolutions.net> Cc: Jarod Wilson <jarod@redhat.com> Signed-off-by: Feng Tang <feng.tang@intel.com> Tested-by: Ole Lukoie <olelukoie@mail.ru> Signed-off-by: David S. Miller <davem@davemloft.net>
2016-05-25 14:49:54 +08:00
/* fall back to legacy interrupt */
pci_disable_msi(alx->hw.pdev);
}
alx_write_mem32(hw, ALX_MSI_RETRANS_TIMER, 0);
err = request_irq(pdev->irq, alx_intr_legacy, IRQF_SHARED,
alx->dev->name, alx);
out:
if (!err)
alx_config_vector_mapping(alx);
return err;
}
static void alx_free_irq(struct alx_priv *alx)
{
struct pci_dev *pdev = alx->hw.pdev;
free_irq(pdev->irq, alx);
if (alx->msi) {
pci_disable_msi(alx->hw.pdev);
alx->msi = false;
}
}
static int alx_identify_hw(struct alx_priv *alx)
{
struct alx_hw *hw = &alx->hw;
int rev = alx_hw_revision(hw);
if (rev > ALX_REV_C0)
return -EINVAL;
hw->max_dma_chnl = rev >= ALX_REV_B0 ? 4 : 2;
return 0;
}
static int alx_init_sw(struct alx_priv *alx)
{
struct pci_dev *pdev = alx->hw.pdev;
struct alx_hw *hw = &alx->hw;
int err;
net: alx: use custom skb allocator This patch follows Eric Dumazet's commit 7b70176421 for Atheros atl1c driver to fix one exactly same bug in alx driver, that the network link will be lost in 1-5 minutes after the device is up. My laptop Lenovo Y580 with Atheros AR8161 ethernet device hit the same problem with kernel 4.4, and it will be cured by Jarod Wilson's commit c406700c for alx driver which get merged in 4.5. But there are still some alx devices can't function well even with Jarod's patch, while this patch could make them work fine. More details on https://bugzilla.kernel.org/show_bug.cgi?id=70761 The debug shows the issue is very likely to be related with the RX DMA address, specifically 0x...f80, if RX buffer get 0x...f80 several times, their will be RX overflow error and device will stop working. For kernel 4.5.0 with Jarod's patch which works fine with my AR8161/Lennov Y580, if I made some change to the __netdev_alloc_skb --> __alloc_page_frag() to make the allocated buffer can get an address with 0x...f80, then the same error happens. If I make it to 0x...f40 or 0x....fc0, everything will be still fine. So I tend to believe that the 0x..f80 address cause the silicon to behave abnormally. Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=70761 Cc: Eric Dumazet <edumazet@google.com> Cc: Johannes Berg <johannes@sipsolutions.net> Cc: Jarod Wilson <jarod@redhat.com> Signed-off-by: Feng Tang <feng.tang@intel.com> Tested-by: Ole Lukoie <olelukoie@mail.ru> Signed-off-by: David S. Miller <davem@davemloft.net>
2016-05-25 14:49:54 +08:00
unsigned int head_size;
err = alx_identify_hw(alx);
if (err) {
dev_err(&pdev->dev, "unrecognized chip, aborting\n");
return err;
}
alx->hw.lnk_patch =
pdev->device == ALX_DEV_ID_AR8161 &&
pdev->subsystem_vendor == PCI_VENDOR_ID_ATTANSIC &&
pdev->subsystem_device == 0x0091 &&
pdev->revision == 0;
hw->smb_timer = 400;
hw->mtu = alx->dev->mtu;
net: alx: use custom skb allocator This patch follows Eric Dumazet's commit 7b70176421 for Atheros atl1c driver to fix one exactly same bug in alx driver, that the network link will be lost in 1-5 minutes after the device is up. My laptop Lenovo Y580 with Atheros AR8161 ethernet device hit the same problem with kernel 4.4, and it will be cured by Jarod Wilson's commit c406700c for alx driver which get merged in 4.5. But there are still some alx devices can't function well even with Jarod's patch, while this patch could make them work fine. More details on https://bugzilla.kernel.org/show_bug.cgi?id=70761 The debug shows the issue is very likely to be related with the RX DMA address, specifically 0x...f80, if RX buffer get 0x...f80 several times, their will be RX overflow error and device will stop working. For kernel 4.5.0 with Jarod's patch which works fine with my AR8161/Lennov Y580, if I made some change to the __netdev_alloc_skb --> __alloc_page_frag() to make the allocated buffer can get an address with 0x...f80, then the same error happens. If I make it to 0x...f40 or 0x....fc0, everything will be still fine. So I tend to believe that the 0x..f80 address cause the silicon to behave abnormally. Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=70761 Cc: Eric Dumazet <edumazet@google.com> Cc: Johannes Berg <johannes@sipsolutions.net> Cc: Jarod Wilson <jarod@redhat.com> Signed-off-by: Feng Tang <feng.tang@intel.com> Tested-by: Ole Lukoie <olelukoie@mail.ru> Signed-off-by: David S. Miller <davem@davemloft.net>
2016-05-25 14:49:54 +08:00
alx->rxbuf_size = ALX_MAX_FRAME_LEN(hw->mtu);
net: alx: use custom skb allocator This patch follows Eric Dumazet's commit 7b70176421 for Atheros atl1c driver to fix one exactly same bug in alx driver, that the network link will be lost in 1-5 minutes after the device is up. My laptop Lenovo Y580 with Atheros AR8161 ethernet device hit the same problem with kernel 4.4, and it will be cured by Jarod Wilson's commit c406700c for alx driver which get merged in 4.5. But there are still some alx devices can't function well even with Jarod's patch, while this patch could make them work fine. More details on https://bugzilla.kernel.org/show_bug.cgi?id=70761 The debug shows the issue is very likely to be related with the RX DMA address, specifically 0x...f80, if RX buffer get 0x...f80 several times, their will be RX overflow error and device will stop working. For kernel 4.5.0 with Jarod's patch which works fine with my AR8161/Lennov Y580, if I made some change to the __netdev_alloc_skb --> __alloc_page_frag() to make the allocated buffer can get an address with 0x...f80, then the same error happens. If I make it to 0x...f40 or 0x....fc0, everything will be still fine. So I tend to believe that the 0x..f80 address cause the silicon to behave abnormally. Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=70761 Cc: Eric Dumazet <edumazet@google.com> Cc: Johannes Berg <johannes@sipsolutions.net> Cc: Jarod Wilson <jarod@redhat.com> Signed-off-by: Feng Tang <feng.tang@intel.com> Tested-by: Ole Lukoie <olelukoie@mail.ru> Signed-off-by: David S. Miller <davem@davemloft.net>
2016-05-25 14:49:54 +08:00
head_size = SKB_DATA_ALIGN(alx->rxbuf_size + NET_SKB_PAD) +
SKB_DATA_ALIGN(sizeof(struct skb_shared_info));
alx->rx_frag_size = roundup_pow_of_two(head_size);
alx->tx_ringsz = 256;
alx->rx_ringsz = 512;
hw->imt = 200;
alx->int_mask = ALX_ISR_MISC;
hw->dma_chnl = hw->max_dma_chnl;
hw->ith_tpd = alx->tx_ringsz / 3;
hw->link_speed = SPEED_UNKNOWN;
hw->duplex = DUPLEX_UNKNOWN;
hw->adv_cfg = ADVERTISED_Autoneg |
ADVERTISED_10baseT_Half |
ADVERTISED_10baseT_Full |
ADVERTISED_100baseT_Full |
ADVERTISED_100baseT_Half |
ADVERTISED_1000baseT_Full;
hw->flowctrl = ALX_FC_ANEG | ALX_FC_RX | ALX_FC_TX;
hw->rx_ctrl = ALX_MAC_CTRL_WOLSPED_SWEN |
ALX_MAC_CTRL_MHASH_ALG_HI5B |
ALX_MAC_CTRL_BRD_EN |
ALX_MAC_CTRL_PCRCE |
ALX_MAC_CTRL_CRCE |
ALX_MAC_CTRL_RXFC_EN |
ALX_MAC_CTRL_TXFC_EN |
7 << ALX_MAC_CTRL_PRMBLEN_SHIFT;
return err;
}
static netdev_features_t alx_fix_features(struct net_device *netdev,
netdev_features_t features)
{
if (netdev->mtu > ALX_MAX_TSO_PKT_SIZE)
features &= ~(NETIF_F_TSO | NETIF_F_TSO6);
return features;
}
static void alx_netif_stop(struct alx_priv *alx)
{
netif_trans_update(alx->dev);
if (netif_carrier_ok(alx->dev)) {
netif_carrier_off(alx->dev);
netif_tx_disable(alx->dev);
napi_disable(&alx->napi);
}
}
static void alx_halt(struct alx_priv *alx)
{
struct alx_hw *hw = &alx->hw;
alx_netif_stop(alx);
hw->link_speed = SPEED_UNKNOWN;
hw->duplex = DUPLEX_UNKNOWN;
alx_reset_mac(hw);
/* disable l0s/l1 */
alx_enable_aspm(hw, false, false);
alx_irq_disable(alx);
alx_free_buffers(alx);
}
static void alx_configure(struct alx_priv *alx)
{
struct alx_hw *hw = &alx->hw;
alx_configure_basic(hw);
alx_disable_rss(hw);
__alx_set_rx_mode(alx->dev);
alx_write_mem32(hw, ALX_MAC_CTRL, hw->rx_ctrl);
}
static void alx_activate(struct alx_priv *alx)
{
/* hardware setting lost, restore it */
alx_reinit_rings(alx);
alx_configure(alx);
/* clear old interrupts */
alx_write_mem32(&alx->hw, ALX_ISR, ~(u32)ALX_ISR_DIS);
alx_irq_enable(alx);
alx_schedule_link_check(alx);
}
static void alx_reinit(struct alx_priv *alx)
{
ASSERT_RTNL();
alx_halt(alx);
alx_activate(alx);
}
static int alx_change_mtu(struct net_device *netdev, int mtu)
{
struct alx_priv *alx = netdev_priv(netdev);
int max_frame = ALX_MAX_FRAME_LEN(mtu);
net: alx: use custom skb allocator This patch follows Eric Dumazet's commit 7b70176421 for Atheros atl1c driver to fix one exactly same bug in alx driver, that the network link will be lost in 1-5 minutes after the device is up. My laptop Lenovo Y580 with Atheros AR8161 ethernet device hit the same problem with kernel 4.4, and it will be cured by Jarod Wilson's commit c406700c for alx driver which get merged in 4.5. But there are still some alx devices can't function well even with Jarod's patch, while this patch could make them work fine. More details on https://bugzilla.kernel.org/show_bug.cgi?id=70761 The debug shows the issue is very likely to be related with the RX DMA address, specifically 0x...f80, if RX buffer get 0x...f80 several times, their will be RX overflow error and device will stop working. For kernel 4.5.0 with Jarod's patch which works fine with my AR8161/Lennov Y580, if I made some change to the __netdev_alloc_skb --> __alloc_page_frag() to make the allocated buffer can get an address with 0x...f80, then the same error happens. If I make it to 0x...f40 or 0x....fc0, everything will be still fine. So I tend to believe that the 0x..f80 address cause the silicon to behave abnormally. Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=70761 Cc: Eric Dumazet <edumazet@google.com> Cc: Johannes Berg <johannes@sipsolutions.net> Cc: Jarod Wilson <jarod@redhat.com> Signed-off-by: Feng Tang <feng.tang@intel.com> Tested-by: Ole Lukoie <olelukoie@mail.ru> Signed-off-by: David S. Miller <davem@davemloft.net>
2016-05-25 14:49:54 +08:00
unsigned int head_size;
if ((max_frame < ALX_MIN_FRAME_SIZE) ||
(max_frame > ALX_MAX_FRAME_SIZE))
return -EINVAL;
if (netdev->mtu == mtu)
return 0;
netdev->mtu = mtu;
alx->hw.mtu = mtu;
alx->rxbuf_size = max(max_frame, ALX_DEF_RXBUF_SIZE);
net: alx: use custom skb allocator This patch follows Eric Dumazet's commit 7b70176421 for Atheros atl1c driver to fix one exactly same bug in alx driver, that the network link will be lost in 1-5 minutes after the device is up. My laptop Lenovo Y580 with Atheros AR8161 ethernet device hit the same problem with kernel 4.4, and it will be cured by Jarod Wilson's commit c406700c for alx driver which get merged in 4.5. But there are still some alx devices can't function well even with Jarod's patch, while this patch could make them work fine. More details on https://bugzilla.kernel.org/show_bug.cgi?id=70761 The debug shows the issue is very likely to be related with the RX DMA address, specifically 0x...f80, if RX buffer get 0x...f80 several times, their will be RX overflow error and device will stop working. For kernel 4.5.0 with Jarod's patch which works fine with my AR8161/Lennov Y580, if I made some change to the __netdev_alloc_skb --> __alloc_page_frag() to make the allocated buffer can get an address with 0x...f80, then the same error happens. If I make it to 0x...f40 or 0x....fc0, everything will be still fine. So I tend to believe that the 0x..f80 address cause the silicon to behave abnormally. Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=70761 Cc: Eric Dumazet <edumazet@google.com> Cc: Johannes Berg <johannes@sipsolutions.net> Cc: Jarod Wilson <jarod@redhat.com> Signed-off-by: Feng Tang <feng.tang@intel.com> Tested-by: Ole Lukoie <olelukoie@mail.ru> Signed-off-by: David S. Miller <davem@davemloft.net>
2016-05-25 14:49:54 +08:00
head_size = SKB_DATA_ALIGN(alx->rxbuf_size + NET_SKB_PAD) +
SKB_DATA_ALIGN(sizeof(struct skb_shared_info));
alx->rx_frag_size = roundup_pow_of_two(head_size);
netdev_update_features(netdev);
if (netif_running(netdev))
alx_reinit(alx);
return 0;
}
static void alx_netif_start(struct alx_priv *alx)
{
netif_tx_wake_all_queues(alx->dev);
napi_enable(&alx->napi);
netif_carrier_on(alx->dev);
}
static int __alx_open(struct alx_priv *alx, bool resume)
{
int err;
if (!resume)
netif_carrier_off(alx->dev);
err = alx_alloc_rings(alx);
if (err)
return err;
alx_configure(alx);
err = alx_request_irq(alx);
if (err)
goto out_free_rings;
/* clear old interrupts */
alx_write_mem32(&alx->hw, ALX_ISR, ~(u32)ALX_ISR_DIS);
alx_irq_enable(alx);
if (!resume)
netif_tx_start_all_queues(alx->dev);
alx_schedule_link_check(alx);
return 0;
out_free_rings:
alx_free_rings(alx);
return err;
}
static void __alx_stop(struct alx_priv *alx)
{
alx_halt(alx);
alx_free_irq(alx);
alx_free_rings(alx);
}
static const char *alx_speed_desc(struct alx_hw *hw)
{
switch (alx_speed_to_ethadv(hw->link_speed, hw->duplex)) {
case ADVERTISED_1000baseT_Full:
return "1 Gbps Full";
case ADVERTISED_100baseT_Full:
return "100 Mbps Full";
case ADVERTISED_100baseT_Half:
return "100 Mbps Half";
case ADVERTISED_10baseT_Full:
return "10 Mbps Full";
case ADVERTISED_10baseT_Half:
return "10 Mbps Half";
default:
return "Unknown speed";
}
}
static void alx_check_link(struct alx_priv *alx)
{
struct alx_hw *hw = &alx->hw;
unsigned long flags;
int old_speed;
u8 old_duplex;
int err;
/* clear PHY internal interrupt status, otherwise the main
* interrupt status will be asserted forever
*/
alx_clear_phy_intr(hw);
old_speed = hw->link_speed;
old_duplex = hw->duplex;
err = alx_read_phy_link(hw);
if (err < 0)
goto reset;
spin_lock_irqsave(&alx->irq_lock, flags);
alx->int_mask |= ALX_ISR_PHY;
alx_write_mem32(hw, ALX_IMR, alx->int_mask);
spin_unlock_irqrestore(&alx->irq_lock, flags);
if (old_speed == hw->link_speed)
return;
if (hw->link_speed != SPEED_UNKNOWN) {
netif_info(alx, link, alx->dev,
"NIC Up: %s\n", alx_speed_desc(hw));
alx_post_phy_link(hw);
alx_enable_aspm(hw, true, true);
alx_start_mac(hw);
if (old_speed == SPEED_UNKNOWN)
alx_netif_start(alx);
} else {
/* link is now down */
alx_netif_stop(alx);
netif_info(alx, link, alx->dev, "Link Down\n");
err = alx_reset_mac(hw);
if (err)
goto reset;
alx_irq_disable(alx);
/* MAC reset causes all HW settings to be lost, restore all */
err = alx_reinit_rings(alx);
if (err)
goto reset;
alx_configure(alx);
alx_enable_aspm(hw, false, true);
alx_post_phy_link(hw);
alx_irq_enable(alx);
}
return;
reset:
alx_schedule_reset(alx);
}
static int alx_open(struct net_device *netdev)
{
return __alx_open(netdev_priv(netdev), false);
}
static int alx_stop(struct net_device *netdev)
{
__alx_stop(netdev_priv(netdev));
return 0;
}
static void alx_link_check(struct work_struct *work)
{
struct alx_priv *alx;
alx = container_of(work, struct alx_priv, link_check_wk);
rtnl_lock();
alx_check_link(alx);
rtnl_unlock();
}
static void alx_reset(struct work_struct *work)
{
struct alx_priv *alx = container_of(work, struct alx_priv, reset_wk);
rtnl_lock();
alx_reinit(alx);
rtnl_unlock();
}
static int alx_tx_csum(struct sk_buff *skb, struct alx_txd *first)
{
u8 cso, css;
if (skb->ip_summed != CHECKSUM_PARTIAL)
return 0;
cso = skb_checksum_start_offset(skb);
if (cso & 1)
return -EINVAL;
css = cso + skb->csum_offset;
first->word1 |= cpu_to_le32((cso >> 1) << TPD_CXSUMSTART_SHIFT);
first->word1 |= cpu_to_le32((css >> 1) << TPD_CXSUMOFFSET_SHIFT);
first->word1 |= cpu_to_le32(1 << TPD_CXSUM_EN_SHIFT);
return 0;
}
static int alx_map_tx_skb(struct alx_priv *alx, struct sk_buff *skb)
{
struct alx_tx_queue *txq = &alx->txq;
struct alx_txd *tpd, *first_tpd;
dma_addr_t dma;
int maplen, f, first_idx = txq->write_idx;
first_tpd = &txq->tpd[txq->write_idx];
tpd = first_tpd;
maplen = skb_headlen(skb);
dma = dma_map_single(&alx->hw.pdev->dev, skb->data, maplen,
DMA_TO_DEVICE);
if (dma_mapping_error(&alx->hw.pdev->dev, dma))
goto err_dma;
dma_unmap_len_set(&txq->bufs[txq->write_idx], size, maplen);
dma_unmap_addr_set(&txq->bufs[txq->write_idx], dma, dma);
tpd->adrl.addr = cpu_to_le64(dma);
tpd->len = cpu_to_le16(maplen);
for (f = 0; f < skb_shinfo(skb)->nr_frags; f++) {
struct skb_frag_struct *frag;
frag = &skb_shinfo(skb)->frags[f];
if (++txq->write_idx == alx->tx_ringsz)
txq->write_idx = 0;
tpd = &txq->tpd[txq->write_idx];
tpd->word1 = first_tpd->word1;
maplen = skb_frag_size(frag);
dma = skb_frag_dma_map(&alx->hw.pdev->dev, frag, 0,
maplen, DMA_TO_DEVICE);
if (dma_mapping_error(&alx->hw.pdev->dev, dma))
goto err_dma;
dma_unmap_len_set(&txq->bufs[txq->write_idx], size, maplen);
dma_unmap_addr_set(&txq->bufs[txq->write_idx], dma, dma);
tpd->adrl.addr = cpu_to_le64(dma);
tpd->len = cpu_to_le16(maplen);
}
/* last TPD, set EOP flag and store skb */
tpd->word1 |= cpu_to_le32(1 << TPD_EOP_SHIFT);
txq->bufs[txq->write_idx].skb = skb;
if (++txq->write_idx == alx->tx_ringsz)
txq->write_idx = 0;
return 0;
err_dma:
f = first_idx;
while (f != txq->write_idx) {
alx_free_txbuf(alx, f);
if (++f == alx->tx_ringsz)
f = 0;
}
return -ENOMEM;
}
static netdev_tx_t alx_start_xmit(struct sk_buff *skb,
struct net_device *netdev)
{
struct alx_priv *alx = netdev_priv(netdev);
struct alx_tx_queue *txq = &alx->txq;
struct alx_txd *first;
int tpdreq = skb_shinfo(skb)->nr_frags + 1;
if (alx_tpd_avail(alx) < tpdreq) {
netif_stop_queue(alx->dev);
goto drop;
}
first = &txq->tpd[txq->write_idx];
memset(first, 0, sizeof(*first));
if (alx_tx_csum(skb, first))
goto drop;
if (alx_map_tx_skb(alx, skb) < 0)
goto drop;
netdev_sent_queue(alx->dev, skb->len);
/* flush updates before updating hardware */
wmb();
alx_write_mem16(&alx->hw, ALX_TPD_PRI0_PIDX, txq->write_idx);
if (alx_tpd_avail(alx) < alx->tx_ringsz/8)
netif_stop_queue(alx->dev);
return NETDEV_TX_OK;
drop:
dev_kfree_skb_any(skb);
return NETDEV_TX_OK;
}
static void alx_tx_timeout(struct net_device *dev)
{
struct alx_priv *alx = netdev_priv(dev);
alx_schedule_reset(alx);
}
static int alx_mdio_read(struct net_device *netdev,
int prtad, int devad, u16 addr)
{
struct alx_priv *alx = netdev_priv(netdev);
struct alx_hw *hw = &alx->hw;
u16 val;
int err;
if (prtad != hw->mdio.prtad)
return -EINVAL;
if (devad == MDIO_DEVAD_NONE)
err = alx_read_phy_reg(hw, addr, &val);
else
err = alx_read_phy_ext(hw, devad, addr, &val);
if (err)
return err;
return val;
}
static int alx_mdio_write(struct net_device *netdev,
int prtad, int devad, u16 addr, u16 val)
{
struct alx_priv *alx = netdev_priv(netdev);
struct alx_hw *hw = &alx->hw;
if (prtad != hw->mdio.prtad)
return -EINVAL;
if (devad == MDIO_DEVAD_NONE)
return alx_write_phy_reg(hw, addr, val);
return alx_write_phy_ext(hw, devad, addr, val);
}
static int alx_ioctl(struct net_device *netdev, struct ifreq *ifr, int cmd)
{
struct alx_priv *alx = netdev_priv(netdev);
if (!netif_running(netdev))
return -EAGAIN;
return mdio_mii_ioctl(&alx->hw.mdio, if_mii(ifr), cmd);
}
#ifdef CONFIG_NET_POLL_CONTROLLER
static void alx_poll_controller(struct net_device *netdev)
{
struct alx_priv *alx = netdev_priv(netdev);
if (alx->msi)
alx_intr_msi(0, alx);
else
alx_intr_legacy(0, alx);
}
#endif
static struct rtnl_link_stats64 *alx_get_stats64(struct net_device *dev,
struct rtnl_link_stats64 *net_stats)
{
struct alx_priv *alx = netdev_priv(dev);
struct alx_hw_stats *hw_stats = &alx->hw.stats;
spin_lock(&alx->stats_lock);
alx_update_hw_stats(&alx->hw);
net_stats->tx_bytes = hw_stats->tx_byte_cnt;
net_stats->rx_bytes = hw_stats->rx_byte_cnt;
net_stats->multicast = hw_stats->rx_mcast;
net_stats->collisions = hw_stats->tx_single_col +
hw_stats->tx_multi_col +
hw_stats->tx_late_col +
hw_stats->tx_abort_col;
net_stats->rx_errors = hw_stats->rx_frag +
hw_stats->rx_fcs_err +
hw_stats->rx_len_err +
hw_stats->rx_ov_sz +
hw_stats->rx_ov_rrd +
hw_stats->rx_align_err +
hw_stats->rx_ov_rxf;
net_stats->rx_fifo_errors = hw_stats->rx_ov_rxf;
net_stats->rx_length_errors = hw_stats->rx_len_err;
net_stats->rx_crc_errors = hw_stats->rx_fcs_err;
net_stats->rx_frame_errors = hw_stats->rx_align_err;
net_stats->rx_dropped = hw_stats->rx_ov_rrd;
net_stats->tx_errors = hw_stats->tx_late_col +
hw_stats->tx_abort_col +
hw_stats->tx_underrun +
hw_stats->tx_trunc;
net_stats->tx_aborted_errors = hw_stats->tx_abort_col;
net_stats->tx_fifo_errors = hw_stats->tx_underrun;
net_stats->tx_window_errors = hw_stats->tx_late_col;
net_stats->tx_packets = hw_stats->tx_ok + net_stats->tx_errors;
net_stats->rx_packets = hw_stats->rx_ok + net_stats->rx_errors;
spin_unlock(&alx->stats_lock);
return net_stats;
}
static const struct net_device_ops alx_netdev_ops = {
.ndo_open = alx_open,
.ndo_stop = alx_stop,
.ndo_start_xmit = alx_start_xmit,
.ndo_get_stats64 = alx_get_stats64,
.ndo_set_rx_mode = alx_set_rx_mode,
.ndo_validate_addr = eth_validate_addr,
.ndo_set_mac_address = alx_set_mac_address,
.ndo_change_mtu = alx_change_mtu,
.ndo_do_ioctl = alx_ioctl,
.ndo_tx_timeout = alx_tx_timeout,
.ndo_fix_features = alx_fix_features,
#ifdef CONFIG_NET_POLL_CONTROLLER
.ndo_poll_controller = alx_poll_controller,
#endif
};
static int alx_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
{
struct net_device *netdev;
struct alx_priv *alx;
struct alx_hw *hw;
bool phy_configured;
int bars, err;
err = pci_enable_device_mem(pdev);
if (err)
return err;
/* The alx chip can DMA to 64-bit addresses, but it uses a single
* shared register for the high 32 bits, so only a single, aligned,
* 4 GB physical address range can be used for descriptors.
*/
if (!dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(64))) {
dev_dbg(&pdev->dev, "DMA to 64-BIT addresses\n");
} else {
err = dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(32));
if (err) {
dev_err(&pdev->dev, "No usable DMA config, aborting\n");
goto out_pci_disable;
}
}
bars = pci_select_bars(pdev, IORESOURCE_MEM);
err = pci_request_selected_regions(pdev, bars, alx_drv_name);
if (err) {
dev_err(&pdev->dev,
"pci_request_selected_regions failed(bars:%d)\n", bars);
goto out_pci_disable;
}
pci_enable_pcie_error_reporting(pdev);
pci_set_master(pdev);
if (!pdev->pm_cap) {
dev_err(&pdev->dev,
"Can't find power management capability, aborting\n");
err = -EIO;
goto out_pci_release;
}
netdev = alloc_etherdev(sizeof(*alx));
if (!netdev) {
err = -ENOMEM;
goto out_pci_release;
}
SET_NETDEV_DEV(netdev, &pdev->dev);
alx = netdev_priv(netdev);
spin_lock_init(&alx->hw.mdio_lock);
spin_lock_init(&alx->irq_lock);
spin_lock_init(&alx->stats_lock);
alx->dev = netdev;
alx->hw.pdev = pdev;
alx->msg_enable = NETIF_MSG_LINK | NETIF_MSG_HW | NETIF_MSG_IFUP |
NETIF_MSG_TX_ERR | NETIF_MSG_RX_ERR | NETIF_MSG_WOL;
hw = &alx->hw;
pci_set_drvdata(pdev, alx);
hw->hw_addr = pci_ioremap_bar(pdev, 0);
if (!hw->hw_addr) {
dev_err(&pdev->dev, "cannot map device registers\n");
err = -EIO;
goto out_free_netdev;
}
netdev->netdev_ops = &alx_netdev_ops;
netdev->ethtool_ops = &alx_ethtool_ops;
netdev->irq = pdev->irq;
netdev->watchdog_timeo = ALX_WATCHDOG_TIME;
if (ent->driver_data & ALX_DEV_QUIRK_MSI_INTX_DISABLE_BUG)
pdev->dev_flags |= PCI_DEV_FLAGS_MSI_INTX_DISABLE_BUG;
err = alx_init_sw(alx);
if (err) {
dev_err(&pdev->dev, "net device private data init failed\n");
goto out_unmap;
}
alx_reset_pcie(hw);
phy_configured = alx_phy_configured(hw);
if (!phy_configured)
alx_reset_phy(hw);
err = alx_reset_mac(hw);
if (err) {
dev_err(&pdev->dev, "MAC Reset failed, error = %d\n", err);
goto out_unmap;
}
/* setup link to put it in a known good starting state */
if (!phy_configured) {
err = alx_setup_speed_duplex(hw, hw->adv_cfg, hw->flowctrl);
if (err) {
dev_err(&pdev->dev,
"failed to configure PHY speed/duplex (err=%d)\n",
err);
goto out_unmap;
}
}
netdev->hw_features = NETIF_F_SG | NETIF_F_HW_CSUM;
if (alx_get_perm_macaddr(hw, hw->perm_addr)) {
dev_warn(&pdev->dev,
"Invalid permanent address programmed, using random one\n");
eth_hw_addr_random(netdev);
memcpy(hw->perm_addr, netdev->dev_addr, netdev->addr_len);
}
memcpy(hw->mac_addr, hw->perm_addr, ETH_ALEN);
memcpy(netdev->dev_addr, hw->mac_addr, ETH_ALEN);
memcpy(netdev->perm_addr, hw->perm_addr, ETH_ALEN);
hw->mdio.prtad = 0;
hw->mdio.mmds = 0;
hw->mdio.dev = netdev;
hw->mdio.mode_support = MDIO_SUPPORTS_C45 |
MDIO_SUPPORTS_C22 |
MDIO_EMULATE_C22;
hw->mdio.mdio_read = alx_mdio_read;
hw->mdio.mdio_write = alx_mdio_write;
if (!alx_get_phy_info(hw)) {
dev_err(&pdev->dev, "failed to identify PHY\n");
err = -EIO;
goto out_unmap;
}
INIT_WORK(&alx->link_check_wk, alx_link_check);
INIT_WORK(&alx->reset_wk, alx_reset);
netif_carrier_off(netdev);
err = register_netdev(netdev);
if (err) {
dev_err(&pdev->dev, "register netdevice failed\n");
goto out_unmap;
}
netdev_info(netdev,
"Qualcomm Atheros AR816x/AR817x Ethernet [%pM]\n",
netdev->dev_addr);
return 0;
out_unmap:
iounmap(hw->hw_addr);
out_free_netdev:
free_netdev(netdev);
out_pci_release:
pci_release_selected_regions(pdev, bars);
out_pci_disable:
pci_disable_device(pdev);
return err;
}
static void alx_remove(struct pci_dev *pdev)
{
struct alx_priv *alx = pci_get_drvdata(pdev);
struct alx_hw *hw = &alx->hw;
cancel_work_sync(&alx->link_check_wk);
cancel_work_sync(&alx->reset_wk);
/* restore permanent mac address */
alx_set_macaddr(hw, hw->perm_addr);
unregister_netdev(alx->dev);
iounmap(hw->hw_addr);
pci_release_selected_regions(pdev,
pci_select_bars(pdev, IORESOURCE_MEM));
pci_disable_pcie_error_reporting(pdev);
pci_disable_device(pdev);
free_netdev(alx->dev);
}
#ifdef CONFIG_PM_SLEEP
static int alx_suspend(struct device *dev)
{
struct pci_dev *pdev = to_pci_dev(dev);
struct alx_priv *alx = pci_get_drvdata(pdev);
if (!netif_running(alx->dev))
return 0;
netif_device_detach(alx->dev);
__alx_stop(alx);
return 0;
}
static int alx_resume(struct device *dev)
{
struct pci_dev *pdev = to_pci_dev(dev);
struct alx_priv *alx = pci_get_drvdata(pdev);
struct alx_hw *hw = &alx->hw;
alx_reset_phy(hw);
if (!netif_running(alx->dev))
return 0;
netif_device_attach(alx->dev);
return __alx_open(alx, true);
}
static SIMPLE_DEV_PM_OPS(alx_pm_ops, alx_suspend, alx_resume);
#define ALX_PM_OPS (&alx_pm_ops)
#else
#define ALX_PM_OPS NULL
#endif
static pci_ers_result_t alx_pci_error_detected(struct pci_dev *pdev,
pci_channel_state_t state)
{
struct alx_priv *alx = pci_get_drvdata(pdev);
struct net_device *netdev = alx->dev;
pci_ers_result_t rc = PCI_ERS_RESULT_NEED_RESET;
dev_info(&pdev->dev, "pci error detected\n");
rtnl_lock();
if (netif_running(netdev)) {
netif_device_detach(netdev);
alx_halt(alx);
}
if (state == pci_channel_io_perm_failure)
rc = PCI_ERS_RESULT_DISCONNECT;
else
pci_disable_device(pdev);
rtnl_unlock();
return rc;
}
static pci_ers_result_t alx_pci_error_slot_reset(struct pci_dev *pdev)
{
struct alx_priv *alx = pci_get_drvdata(pdev);
struct alx_hw *hw = &alx->hw;
pci_ers_result_t rc = PCI_ERS_RESULT_DISCONNECT;
dev_info(&pdev->dev, "pci error slot reset\n");
rtnl_lock();
if (pci_enable_device(pdev)) {
dev_err(&pdev->dev, "Failed to re-enable PCI device after reset\n");
goto out;
}
pci_set_master(pdev);
alx_reset_pcie(hw);
if (!alx_reset_mac(hw))
rc = PCI_ERS_RESULT_RECOVERED;
out:
pci_cleanup_aer_uncorrect_error_status(pdev);
rtnl_unlock();
return rc;
}
static void alx_pci_error_resume(struct pci_dev *pdev)
{
struct alx_priv *alx = pci_get_drvdata(pdev);
struct net_device *netdev = alx->dev;
dev_info(&pdev->dev, "pci error resume\n");
rtnl_lock();
if (netif_running(netdev)) {
alx_activate(alx);
netif_device_attach(netdev);
}
rtnl_unlock();
}
static const struct pci_error_handlers alx_err_handlers = {
.error_detected = alx_pci_error_detected,
.slot_reset = alx_pci_error_slot_reset,
.resume = alx_pci_error_resume,
};
static const struct pci_device_id alx_pci_tbl[] = {
{ PCI_VDEVICE(ATTANSIC, ALX_DEV_ID_AR8161),
.driver_data = ALX_DEV_QUIRK_MSI_INTX_DISABLE_BUG },
{ PCI_VDEVICE(ATTANSIC, ALX_DEV_ID_E2200),
.driver_data = ALX_DEV_QUIRK_MSI_INTX_DISABLE_BUG },
{ PCI_VDEVICE(ATTANSIC, ALX_DEV_ID_E2400),
.driver_data = ALX_DEV_QUIRK_MSI_INTX_DISABLE_BUG },
{ PCI_VDEVICE(ATTANSIC, ALX_DEV_ID_AR8162),
.driver_data = ALX_DEV_QUIRK_MSI_INTX_DISABLE_BUG },
{ PCI_VDEVICE(ATTANSIC, ALX_DEV_ID_AR8171) },
{ PCI_VDEVICE(ATTANSIC, ALX_DEV_ID_AR8172) },
{}
};
static struct pci_driver alx_driver = {
.name = alx_drv_name,
.id_table = alx_pci_tbl,
.probe = alx_probe,
.remove = alx_remove,
.err_handler = &alx_err_handlers,
.driver.pm = ALX_PM_OPS,
};
module_pci_driver(alx_driver);
MODULE_DEVICE_TABLE(pci, alx_pci_tbl);
MODULE_AUTHOR("Johannes Berg <johannes@sipsolutions.net>");
MODULE_AUTHOR("Qualcomm Corporation, <nic-devel@qualcomm.com>");
MODULE_DESCRIPTION(
"Qualcomm Atheros(R) AR816x/AR817x PCI-E Ethernet Network Driver");
MODULE_LICENSE("GPL");