* [PATCH net-next v2 04/07] via-rhine: Support for byte queue limits
From: Tino Reichardt @ 2013-10-20 18:13 UTC (permalink / raw)
To: netdev, Roger Luethi
In-Reply-To: <1382292803-18875-1-git-send-email-milky-kernel@mcmilk.de>
Changes to via-rhine driver to use byte queue limits.
Signed-off-by: Tino Reichardt <milky-kernel@mcmilk.de>
---
drivers/net/ethernet/via/via-rhine.c | 17 +++++++++++++++++
1 file changed, 17 insertions(+)
diff --git a/drivers/net/ethernet/via/via-rhine.c b/drivers/net/ethernet/via/via-rhine.c
index bdf697b..067b4ab 100644
--- a/drivers/net/ethernet/via/via-rhine.c
+++ b/drivers/net/ethernet/via/via-rhine.c
@@ -53,6 +53,8 @@ static int rx_copybreak = 1518;
static int rx_copybreak;
#endif
+static bool bql_disable;
+
/* Work-around for broken BIOSes: they are unable to get the chip back out of
power state D3 so PXE booting fails. bootparam(7): via-rhine.avoid_D3=1 */
static bool avoid_D3;
@@ -128,9 +130,11 @@ MODULE_DESCRIPTION("VIA Rhine PCI Fast Ethernet driver");
MODULE_LICENSE("GPL");
module_param(debug, int, 0);
+module_param(bql_disable, bool, 0);
module_param(rx_copybreak, int, 0);
module_param(avoid_D3, bool, 0);
MODULE_PARM_DESC(debug, "VIA Rhine debug message flags");
+MODULE_PARM_DESC(bql_disable, "Disable Byte Queue Limits functionality (default: false)");
MODULE_PARM_DESC(rx_copybreak, "VIA Rhine copy breakpoint for copy-only-tiny-frames");
MODULE_PARM_DESC(avoid_D3, "Avoid power state D3 (work-around for broken BIOSes)");
@@ -1220,6 +1224,8 @@ static void alloc_tbufs(struct net_device* dev)
}
rp->tx_ring[i-1].next_desc = cpu_to_le32(rp->tx_ring_dma);
+ if (likely(bql_disable == false))
+ netdev_reset_queue(dev);
}
static void free_tbufs(struct net_device* dev)
@@ -1719,6 +1725,8 @@ static netdev_tx_t rhine_start_tx(struct sk_buff *skb,
/* lock eth irq */
wmb();
rp->tx_ring[entry].tx_status |= cpu_to_le32(DescOwn);
+ if (likely(bql_disable == false))
+ netdev_sent_queue(dev, skb->len);
wmb();
rp->cur_tx++;
@@ -1783,6 +1791,7 @@ static void rhine_tx(struct net_device *dev)
{
struct rhine_private *rp = netdev_priv(dev);
int txstatus = 0, entry = rp->dirty_tx % TX_RING_SIZE;
+ unsigned int pkts_compl = 0, bytes_compl = 0;
/* find and cleanup dirty tx descriptors */
while (rp->dirty_tx != rp->cur_tx) {
@@ -1830,10 +1839,18 @@ static void rhine_tx(struct net_device *dev)
rp->tx_skbuff[entry]->len,
PCI_DMA_TODEVICE);
}
+
+ bytes_compl += rp->tx_skbuff[entry]->len;
+ pkts_compl++;
dev_kfree_skb(rp->tx_skbuff[entry]);
+
rp->tx_skbuff[entry] = NULL;
entry = (++rp->dirty_tx) % TX_RING_SIZE;
}
+
+ if (likely(bql_disable == false))
+ netdev_completed_queue(dev, pkts_compl, bytes_compl);
+
if ((rp->cur_tx - rp->dirty_tx) < TX_QUEUE_LEN - 4)
netif_wake_queue(dev);
}
--
1.8.4.1
^ permalink raw reply related
* [PATCH net-next v2 05/07] via-velocity: Support for byte queue limits
From: Tino Reichardt @ 2013-10-20 18:13 UTC (permalink / raw)
To: netdev, Francois Romieu
In-Reply-To: <1382292803-18875-1-git-send-email-milky-kernel@mcmilk.de>
Changes to via-velocity driver to use byte queue limits.
Signed-off-by: Tino Reichardt <milky-kernel@mcmilk.de>
---
drivers/net/ethernet/via/via-velocity.c | 18 ++++++++++++++++++
1 file changed, 18 insertions(+)
diff --git a/drivers/net/ethernet/via/via-velocity.c b/drivers/net/ethernet/via/via-velocity.c
index d022bf9..b084404 100644
--- a/drivers/net/ethernet/via/via-velocity.c
+++ b/drivers/net/ethernet/via/via-velocity.c
@@ -1,3 +1,4 @@
+
/*
* This code is derived from the VIA reference driver (copyright message
* below) provided to Red Hat by VIA Networking Technologies, Inc. for
@@ -368,6 +369,11 @@ static int rx_copybreak = 200;
module_param(rx_copybreak, int, 0644);
MODULE_PARM_DESC(rx_copybreak, "Copy breakpoint for copy-only-tiny-frames");
+static bool bql_disable;
+module_param(bql_disable, bool, 0);
+MODULE_PARM_DESC(bql_disable,
+ "Disable Byte Queue Limits functionality (default: false)");
+
/*
* Internal board variants. At the moment we have only one
*/
@@ -1751,6 +1757,9 @@ static void velocity_free_tx_buf(struct velocity_info *vptr,
le16_to_cpu(pktlen), DMA_TO_DEVICE);
}
}
+
+ if (likely(bql_disable == false))
+ netdev_reset_queue(vptr->netdev);
dev_kfree_skb_irq(skb);
tdinfo->skb = NULL;
}
@@ -1915,6 +1924,7 @@ static int velocity_tx_srv(struct velocity_info *vptr)
int works = 0;
struct velocity_td_info *tdinfo;
struct net_device_stats *stats = &vptr->netdev->stats;
+ unsigned int pkts_compl = 0, bytes_compl = 0;
for (qnum = 0; qnum < vptr->tx.numq; qnum++) {
for (idx = vptr->tx.tail[qnum]; vptr->tx.used[qnum] > 0;
@@ -1946,6 +1956,8 @@ static int velocity_tx_srv(struct velocity_info *vptr)
} else {
stats->tx_packets++;
stats->tx_bytes += tdinfo->skb->len;
+ pkts_compl++;
+ bytes_compl += tdinfo->skb->len;
}
velocity_free_tx_buf(vptr, tdinfo, td);
vptr->tx.used[qnum]--;
@@ -1955,6 +1967,10 @@ static int velocity_tx_srv(struct velocity_info *vptr)
if (AVAIL_TD(vptr, qnum) < 1)
full = 1;
}
+
+ if (likely(bql_disable == false))
+ netdev_completed_queue(vptr->netdev, pkts_compl, bytes_compl);
+
/*
* Look to see if we should kick the transmit network
* layer for more work.
@@ -2641,6 +2657,8 @@ static netdev_tx_t velocity_xmit(struct sk_buff *skb,
td_ptr->td_buf[0].size |= TD_QUEUE;
mac_tx_queue_wake(vptr->mac_regs, qnum);
+ if (likely(bql_disable == false))
+ netdev_sent_queue(vptr->netdev, skb->len);
spin_unlock_irqrestore(&vptr->lock, flags);
out:
return NETDEV_TX_OK;
--
1.8.4.1
^ permalink raw reply related
* [PATCH net-next v2 06/07] 3c59x: Support for byte queue limits
From: Tino Reichardt @ 2013-10-20 18:13 UTC (permalink / raw)
To: netdev, Steffen Klassert
In-Reply-To: <1382292803-18875-1-git-send-email-milky-kernel@mcmilk.de>
Changes to 3c59x driver to use byte queue limits.
Signed-off-by: Tino Reichardt <milky-kernel@mcmilk.de>
---
drivers/net/ethernet/3com/3c59x.c | 47 ++++++++++++++++++++++++++++++++++-----
1 file changed, 41 insertions(+), 6 deletions(-)
diff --git a/drivers/net/ethernet/3com/3c59x.c b/drivers/net/ethernet/3com/3c59x.c
index ad5272b..6a7b77d 100644
--- a/drivers/net/ethernet/3com/3c59x.c
+++ b/drivers/net/ethernet/3com/3c59x.c
@@ -795,6 +795,7 @@ static int global_options = -1;
static int global_full_duplex = -1;
static int global_enable_wol = -1;
static int global_use_mmio = -1;
+static bool bql_disable;
/* Variables to work-around the Compaq PCI BIOS32 problem. */
static int compaq_ioaddr, compaq_irq, compaq_device_id = 0x5900;
@@ -819,6 +820,8 @@ module_param(compaq_device_id, int, 0);
module_param(watchdog, int, 0);
module_param(global_use_mmio, int, 0);
module_param_array(use_mmio, int, NULL, 0);
+module_param(bql_disable, bool, 0);
+
MODULE_PARM_DESC(debug, "3c59x debug level (0-6)");
MODULE_PARM_DESC(options, "3c59x: Bits 0-3: media type, bit 4: bus mastering, bit 9: full duplex");
MODULE_PARM_DESC(global_options, "3c59x: same as options, but applies to all NICs if options is unset");
@@ -836,6 +839,7 @@ MODULE_PARM_DESC(compaq_device_id, "3c59x PCI device ID (Compaq BIOS problem wor
MODULE_PARM_DESC(watchdog, "3c59x transmit timeout in milliseconds");
MODULE_PARM_DESC(global_use_mmio, "3c59x: same as use_mmio, but applies to all NICs if options is unset");
MODULE_PARM_DESC(use_mmio, "3c59x: use memory-mapped PCI I/O resource (0-1)");
+MODULE_PARM_DESC(bql_disable, "3c59x: Disable Byte Queue Limits functionality (default: false)");
#ifdef CONFIG_NET_POLL_CONTROLLER
static void poll_vortex(struct net_device *dev)
@@ -1726,6 +1730,9 @@ vortex_up(struct net_device *dev)
iowrite16(vp->intr_enable, ioaddr + EL3_CMD);
if (vp->cb_fn_base) /* The PCMCIA people are idiots. */
iowrite32(0x8000, vp->cb_fn_base + 4);
+
+ if (likely(bql_disable == false))
+ netdev_reset_queue(dev);
netif_start_queue (dev);
err_out:
return err;
@@ -2080,10 +2087,15 @@ vortex_start_xmit(struct sk_buff *skb, struct net_device *dev)
spin_unlock_irq(&vp->window_lock);
vp->tx_skb = skb;
iowrite16(StartDMADown, ioaddr + EL3_CMD);
+ if (likely(bql_disable == false))
+ netdev_sent_queue(dev, len);
/* netif_wake_queue() will be called at the DMADone interrupt. */
} else {
/* ... and the packet rounded to a doubleword. */
- iowrite32_rep(ioaddr + TX_FIFO, skb->data, (skb->len + 3) >> 2);
+ int len = (skb->len + 3) >> 2;
+ iowrite32_rep(ioaddr + TX_FIFO, skb->data, len);
+ if (likely(bql_disable == false))
+ netdev_sent_queue(dev, len);
dev_kfree_skb (skb);
if (ioread16(ioaddr + TxFree) > 1536) {
netif_start_queue (dev); /* AKPM: redundant? */
@@ -2094,7 +2106,6 @@ vortex_start_xmit(struct sk_buff *skb, struct net_device *dev)
}
}
-
/* Clear the Tx status stack. */
{
int tx_status;
@@ -2164,12 +2175,14 @@ boomerang_start_xmit(struct sk_buff *skb, struct net_device *dev)
vp->tx_ring[entry].frag[0].addr = cpu_to_le32(pci_map_single(VORTEX_PCI(vp), skb->data,
skb->len, PCI_DMA_TODEVICE));
vp->tx_ring[entry].frag[0].length = cpu_to_le32(skb->len | LAST_FRAG);
+ netdev_sent_queue(dev, skb->len);
} else {
- int i;
+ int i, len;
vp->tx_ring[entry].frag[0].addr = cpu_to_le32(pci_map_single(VORTEX_PCI(vp), skb->data,
skb_headlen(skb), PCI_DMA_TODEVICE));
vp->tx_ring[entry].frag[0].length = cpu_to_le32(skb_headlen(skb));
+ len = skb_headlen(skb);
for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
skb_frag_t *frag = &skb_shinfo(skb)->frags[i];
@@ -2180,16 +2193,21 @@ boomerang_start_xmit(struct sk_buff *skb, struct net_device *dev)
(void *)skb_frag_address(frag),
skb_frag_size(frag), PCI_DMA_TODEVICE));
- if (i == skb_shinfo(skb)->nr_frags-1)
+ if (i == skb_shinfo(skb)->nr_frags - 1) {
vp->tx_ring[entry].frag[i+1].length = cpu_to_le32(skb_frag_size(frag)|LAST_FRAG);
- else
+ len += skb_frag_size(frag) | LAST_FRAG;
+ } else {
vp->tx_ring[entry].frag[i+1].length = cpu_to_le32(skb_frag_size(frag));
+ len += skb_frag_size(frag);
+ }
}
+ netdev_sent_queue(dev, len);
}
#else
vp->tx_ring[entry].addr = cpu_to_le32(pci_map_single(VORTEX_PCI(vp), skb->data, skb->len, PCI_DMA_TODEVICE));
vp->tx_ring[entry].length = cpu_to_le32(skb->len | LAST_FRAG);
vp->tx_ring[entry].status = cpu_to_le32(skb->len | TxIntrUploaded);
+ netdev_sent_queue(dev, skb->len | LAST_FRAG);
#endif
spin_lock_irqsave(&vp->lock, flags);
@@ -2234,6 +2252,7 @@ vortex_interrupt(int irq, void *dev_id)
int status;
int work_done = max_interrupt_work;
int handled = 0;
+ unsigned bytes_compl = 0, pkts_compl = 0;
ioaddr = vp->ioaddr;
spin_lock(&vp->lock);
@@ -2279,8 +2298,12 @@ vortex_interrupt(int irq, void *dev_id)
if (status & DMADone) {
if (ioread16(ioaddr + Wn7_MasterStatus) & 0x1000) {
+ int len;
iowrite16(0x1000, ioaddr + Wn7_MasterStatus); /* Ack the event. */
- pci_unmap_single(VORTEX_PCI(vp), vp->tx_skb_dma, (vp->tx_skb->len + 3) & ~3, PCI_DMA_TODEVICE);
+ len = (vp->tx_skb->len + 3) & ~3;
+ pci_unmap_single(VORTEX_PCI(vp), vp->tx_skb_dma, len, PCI_DMA_TODEVICE);
+ bytes_compl += len;
+ pkts_compl++;
dev_kfree_skb_irq(vp->tx_skb); /* Release the transferred buffer */
if (ioread16(ioaddr + TxFree) > 1536) {
/*
@@ -2327,6 +2350,9 @@ vortex_interrupt(int irq, void *dev_id)
spin_unlock(&vp->window_lock);
+ if (likely(bql_disable == false))
+ netdev_completed_queue(dev, pkts_compl, bytes_compl);
+
if (vortex_debug > 4)
pr_debug("%s: exiting interrupt, status %4.4x.\n",
dev->name, status);
@@ -2348,6 +2374,7 @@ boomerang_interrupt(int irq, void *dev_id)
void __iomem *ioaddr;
int status;
int work_done = max_interrupt_work;
+ unsigned bytes_compl = 0, pkts_compl = 0;
ioaddr = vp->ioaddr;
@@ -2420,6 +2447,8 @@ boomerang_interrupt(int irq, void *dev_id)
pci_unmap_single(VORTEX_PCI(vp),
le32_to_cpu(vp->tx_ring[entry].addr), skb->len, PCI_DMA_TODEVICE);
#endif
+ bytes_compl += skb->len;
+ pkts_compl++;
dev_kfree_skb_irq(skb);
vp->tx_skbuff[entry] = NULL;
} else {
@@ -2467,6 +2496,10 @@ boomerang_interrupt(int irq, void *dev_id)
handler_exit:
vp->handling_irq = 0;
spin_unlock(&vp->lock);
+
+ if (likely(bql_disable == false))
+ netdev_completed_queue(dev, pkts_compl, bytes_compl);
+
return IRQ_HANDLED;
}
@@ -2660,6 +2693,8 @@ vortex_down(struct net_device *dev, int final_down)
struct vortex_private *vp = netdev_priv(dev);
void __iomem *ioaddr = vp->ioaddr;
+ if (likely(bql_disable == false))
+ netdev_reset_queue(dev);
netif_stop_queue (dev);
del_timer_sync(&vp->rx_oom_timer);
--
1.8.4.1
^ permalink raw reply related
* [PATCH net-next v2 07/07] natsemi: Support for byte queue limits
From: Tino Reichardt @ 2013-10-20 18:13 UTC (permalink / raw)
To: netdev, Greg Kroah-Hartman, David S. Miller, Jiri Pirko,
Bill Pemberton
In-Reply-To: <1382292803-18875-1-git-send-email-milky-kernel@mcmilk.de>
Changes to natsemi driver to use byte queue limits.
Signed-off-by: Tino Reichardt <milky-kernel@mcmilk.de>
---
drivers/net/ethernet/natsemi/natsemi.c | 21 +++++++++++++++++++++
1 file changed, 21 insertions(+)
diff --git a/drivers/net/ethernet/natsemi/natsemi.c b/drivers/net/ethernet/natsemi/natsemi.c
index 7a5e295..1f6efbc 100644
--- a/drivers/net/ethernet/natsemi/natsemi.c
+++ b/drivers/net/ethernet/natsemi/natsemi.c
@@ -71,6 +71,8 @@
NETIF_MSG_TX_ERR)
static int debug = -1;
+static bool bql_disable;
+
static int mtu;
/* Maximum number of multicast addresses to filter (vs. rx-all-multicast).
@@ -139,12 +141,15 @@ MODULE_LICENSE("GPL");
module_param(mtu, int, 0);
module_param(debug, int, 0);
+module_param(bql_disable, bool, 0);
module_param(rx_copybreak, int, 0);
module_param(dspcfg_workaround, int, 0);
module_param_array(options, int, NULL, 0);
module_param_array(full_duplex, int, NULL, 0);
MODULE_PARM_DESC(mtu, "DP8381x MTU (all boards)");
MODULE_PARM_DESC(debug, "DP8381x default debug level");
+MODULE_PARM_DESC(bql_disable,
+ "Disable Byte Queue Limits functionality (default: false)");
MODULE_PARM_DESC(rx_copybreak,
"DP8381x copy breakpoint for copy-only-tiny-frames");
MODULE_PARM_DESC(dspcfg_workaround, "DP8381x: control DspCfg workaround");
@@ -1974,6 +1979,9 @@ static void init_ring(struct net_device *dev)
np->tx_ring[i].cmd_status = 0;
}
+ if (likely(bql_disable == false))
+ netdev_reset_queue(dev);
+
/* 2) RX ring */
np->dirty_rx = 0;
np->cur_rx = RX_RING_SIZE;
@@ -2012,6 +2020,9 @@ static void drain_tx(struct net_device *dev)
}
np->tx_skbuff[i] = NULL;
}
+
+ if (likely(bql_disable == false))
+ netdev_reset_queue(dev);
}
static void drain_rx(struct net_device *dev)
@@ -2116,6 +2127,9 @@ static netdev_tx_t start_tx(struct sk_buff *skb, struct net_device *dev)
dev_kfree_skb_irq(skb);
dev->stats.tx_dropped++;
}
+
+ if (likely(bql_disable == false))
+ netdev_sent_queue(dev, skb->len);
spin_unlock_irqrestore(&np->lock, flags);
if (netif_msg_tx_queued(np)) {
@@ -2128,6 +2142,7 @@ static netdev_tx_t start_tx(struct sk_buff *skb, struct net_device *dev)
static void netdev_tx_done(struct net_device *dev)
{
struct netdev_private *np = netdev_priv(dev);
+ unsigned bytes_compl = 0, pkts_compl = 0;
for (; np->cur_tx - np->dirty_tx > 0; np->dirty_tx++) {
int entry = np->dirty_tx % TX_RING_SIZE;
@@ -2158,9 +2173,15 @@ static void netdev_tx_done(struct net_device *dev)
np->tx_skbuff[entry]->len,
PCI_DMA_TODEVICE);
/* Free the original skb. */
+ bytes_compl += np->tx_skbuff[entry]->len;
+ pkts_compl++;
dev_kfree_skb_irq(np->tx_skbuff[entry]);
np->tx_skbuff[entry] = NULL;
}
+
+ if (likely(bql_disable == false))
+ netdev_completed_queue(dev, pkts_compl, bytes_compl);
+
if (netif_queue_stopped(dev) &&
np->cur_tx - np->dirty_tx < TX_QUEUE_LEN - 4) {
/* The ring is no longer full, wake queue. */
--
1.8.4.1
^ permalink raw reply related
* Re: kernel BUG at net/core/skbuff.c:1048!
From: Hannes Frederic Sowa @ 2013-10-20 18:25 UTC (permalink / raw)
To: Patrick McHardy; +Cc: Wim Vandersmissen, netdev, kernelspam
In-Reply-To: <20131005100105.GA2457@macbook.localnet>
Hi Patrick!
On Sat, Oct 05, 2013 at 11:01:05AM +0100, Patrick McHardy wrote:
> On Fri, Oct 04, 2013 at 11:35:55AM +0200, Wim Vandersmissen wrote:
> > Hi,
> >
> > Got the following BUG when using ipv6 netfilter/conntrack/ipv6
> > forwarding and traffic flowing.
> >
> > No issues in 3.4.x, but triggered in 3.10.x (introduced in 3.7)
>
> Please also send me your 3.10.x config and the ip6tables rules you're
> using.
>
> Are you actually using IPv6 NAT?
Maybe the same bug was hit by someone else and was reported in kernel
bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=63331
I added him in Cc.
Greetings,
Hannes
^ permalink raw reply
* Re: [PATCH net 2/3] ipv6: fill rt6i_gateway with nexthop address
From: Hannes Frederic Sowa @ 2013-10-20 18:33 UTC (permalink / raw)
To: Julian Anastasov
Cc: David Miller, netdev, netfilter-devel, lvs-devel,
Hideaki YOSHIFUJI
In-Reply-To: <1382272985-1528-3-git-send-email-ja@ssi.bg>
On Sun, Oct 20, 2013 at 03:43:04PM +0300, Julian Anastasov wrote:
> Make sure rt6i_gateway contains nexthop information in
> all routes returned from lookup or when routes are directly
> attached to skb for generated ICMP packets.
>
> The effect of this patch should be a faster version of
> rt6_nexthop() and the consideration of local addresses as
> nexthop.
>
> Signed-off-by: Julian Anastasov <ja@ssi.bg>
> ---
> include/net/ip6_route.h | 6 ++----
> net/ipv6/ip6_output.c | 4 ++--
> net/ipv6/route.c | 8 ++++++--
> 3 files changed, 10 insertions(+), 8 deletions(-)
>
> diff --git a/include/net/ip6_route.h b/include/net/ip6_route.h
> index 481404a..2b786b7 100644
> --- a/include/net/ip6_route.h
> +++ b/include/net/ip6_route.h
> @@ -194,11 +194,9 @@ static inline int ip6_skb_dst_mtu(struct sk_buff *skb)
> skb_dst(skb)->dev->mtu : dst_mtu(skb_dst(skb));
> }
>
> -static inline struct in6_addr *rt6_nexthop(struct rt6_info *rt, struct in6_addr *dest)
> +static inline struct in6_addr *rt6_nexthop(struct rt6_info *rt)
> {
> - if (rt->rt6i_flags & RTF_GATEWAY || !ipv6_addr_any(&rt->rt6i_gateway))
> - return &rt->rt6i_gateway;
> - return dest;
> + return &rt->rt6i_gateway;
> }
>
Hm, looks good. Will test is soon.
One small remark:
Maybe we can kill rt6_nexthop completly and use rt6i_gateway directly so a
reader does not have to lookup this function.
Thanks,
Hannes
^ permalink raw reply
* Re: kernel BUG at net/core/skbuff.c:1048!
From: Patrick McHardy @ 2013-10-20 18:33 UTC (permalink / raw)
To: Wim Vandersmissen, netdev, kernelspam
In-Reply-To: <20131020182523.GF27787@order.stressinduktion.org>
On Sun, Oct 20, 2013 at 08:25:23PM +0200, Hannes Frederic Sowa wrote:
> Hi Patrick!
>
> On Sat, Oct 05, 2013 at 11:01:05AM +0100, Patrick McHardy wrote:
> > On Fri, Oct 04, 2013 at 11:35:55AM +0200, Wim Vandersmissen wrote:
> > > Hi,
> > >
> > > Got the following BUG when using ipv6 netfilter/conntrack/ipv6
> > > forwarding and traffic flowing.
> > >
> > > No issues in 3.4.x, but triggered in 3.10.x (introduced in 3.7)
> >
> > Please also send me your 3.10.x config and the ip6tables rules you're
> > using.
> >
> > Are you actually using IPv6 NAT?
>
> Maybe the same bug was hit by someone else and was reported in kernel
> bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=63331
>
> I added him in Cc.
Thanks, didn't really get to it yet. The problem is obviously that we're
rerouting the individual fragments, during which their headroom gets
expanded while a reference to them is held by the reassembled skb. Just
not sure yet how to fix it.
^ permalink raw reply
* Re: [PATCH net-next v2 01/07] 8139too: Support for byte queue limits
From: Joe Perches @ 2013-10-20 18:51 UTC (permalink / raw)
To: Tino Reichardt
Cc: netdev, David S. Miller, Jiri Pirko, Bill Pemberton,
Greg Kroah-Hartman
In-Reply-To: <1382292803-18875-2-git-send-email-milky-kernel@mcmilk.de>
On Sun, 2013-10-20 at 20:13 +0200, Tino Reichardt wrote:
> Changes to 8139too driver to use byte queue limits.
[]
> diff --git a/drivers/net/ethernet/realtek/8139too.c b/drivers/net/ethernet/realtek/8139too.c
[]
> @@ -1409,6 +1413,8 @@ static void rtl8139_hw_start (struct net_device *dev)
[]
> netdev_dbg(dev, "init buffer addresses\n");
> + if (likely(bql_disable == false))
> + netdev_reset_queue(dev);
Much prefer:
if (unlikely(bql_disable))
> @@ -1639,6 +1645,9 @@ static inline void rtl8139_tx_clear (struct rtl8139_private *tp)
[]
> + if (likely(bql_disable == false))
> + netdev_reset_queue(tp->dev);
etc...
^ permalink raw reply
* Re: [PATCH net-next v2 01/07] 8139too: Support for byte queue limits
From: Tino Reichardt @ 2013-10-20 19:14 UTC (permalink / raw)
To: netdev
In-Reply-To: <1382295089.2041.42.camel@joe-AO722>
* Joe Perches <joe@perches.com> wrote:
> On Sun, 2013-10-20 at 20:13 +0200, Tino Reichardt wrote:
> > Changes to 8139too driver to use byte queue limits.
> []
> > diff --git a/drivers/net/ethernet/realtek/8139too.c b/drivers/net/ethernet/realtek/8139too.c
> []
> > @@ -1409,6 +1413,8 @@ static void rtl8139_hw_start (struct net_device *dev)
> []
> > netdev_dbg(dev, "init buffer addresses\n");
> > + if (likely(bql_disable == false))
> > + netdev_reset_queue(dev);
>
> Much prefer:
>
> if (unlikely(bql_disable))
Okay, i will resend an updated v3 patchset.
--
Best regards, TR
^ permalink raw reply
* [PATCHSET net-next v3 00/07] Support for byte queue limits on various drivers
From: Tino Reichardt @ 2013-10-20 19:23 UTC (permalink / raw)
To: netdev, David S. Miller
Hello again,
this patchset is intended for net-next/linux-3.13.
It adds support for byte queue limits on various network drivers.
Since not all of them are fully tested by now, I added an bql_disable
module parameter, which can be used to disable the BQL code.
"if (likely(bql_disable == false))" was replaced by
"if (unlikely(bql_disable))" ...
Any comments are welcome, thanks.
Tino Reichardt
This BQL patchset contains the following patches by now:
0001-8139too-Support-for-byte-queue-limits.patch
0002-r8169-Support-for-byte-queue-limits.patch
0003-tulip-Support-for-byte-queue-limits.patch
0004-via-rhine-Support-for-byte-queue-limits.patch
0005-via-velocity-Support-for-byte-queue-limits.patch
0006-3c59x-Support-for-byte-queue-limits.patch
0007-natsemi-Support-for-byte-queue-limits.patch
---
drivers/net/ethernet/3com/3c59x.c | 47 +++++++++++++++++++++++++----
drivers/net/ethernet/dec/tulip/interrupt.c | 4 +++
drivers/net/ethernet/dec/tulip/tulip.h | 1 +
drivers/net/ethernet/dec/tulip/tulip_core.c | 9 +++++-
drivers/net/ethernet/natsemi/natsemi.c | 21 +++++++++++++
drivers/net/ethernet/realtek/8139too.c | 28 ++++++++++++++---
drivers/net/ethernet/realtek/r8169.c | 14 +++++++++
drivers/net/ethernet/via/via-rhine.c | 17 +++++++++++
drivers/net/ethernet/via/via-velocity.c | 18 +++++++++++
9 files changed, 147 insertions(+), 12 deletions(-)
^ permalink raw reply
* [PATCH net-next v3 01/07] 8139too: Support for byte queue limits
From: Tino Reichardt @ 2013-10-20 19:23 UTC (permalink / raw)
To: netdev, David S. Miller, Joe Perches, Jiri Pirko, Bill Pemberton,
Greg Kroah-Hartman
In-Reply-To: <1382296991-20289-1-git-send-email-milky-kernel@mcmilk.de>
Changes to 8139too driver to use byte queue limits.
Signed-off-by: Tino Reichardt <milky-kernel@mcmilk.de>
---
drivers/net/ethernet/realtek/8139too.c | 28 +++++++++++++++++++++++-----
1 file changed, 23 insertions(+), 5 deletions(-)
diff --git a/drivers/net/ethernet/realtek/8139too.c b/drivers/net/ethernet/realtek/8139too.c
index 3ccedeb..c17c12f 100644
--- a/drivers/net/ethernet/realtek/8139too.c
+++ b/drivers/net/ethernet/realtek/8139too.c
@@ -160,6 +160,8 @@ static int multicast_filter_limit = 32;
/* bitmapped message enable number */
static int debug = -1;
+static bool bql_disable;
+
/*
* Receive ring size
* Warning: 64K ring has hardware issues and may lock up.
@@ -626,10 +628,12 @@ module_param(multicast_filter_limit, int, 0);
module_param_array(media, int, NULL, 0);
module_param_array(full_duplex, int, NULL, 0);
module_param(debug, int, 0);
-MODULE_PARM_DESC (debug, "8139too bitmapped message enable number");
-MODULE_PARM_DESC (multicast_filter_limit, "8139too maximum number of filtered multicast addresses");
-MODULE_PARM_DESC (media, "8139too: Bits 4+9: force full duplex, bit 5: 100Mbps");
-MODULE_PARM_DESC (full_duplex, "8139too: Force full duplex for board(s) (1)");
+module_param(bql_disable, bool, 0);
+MODULE_PARM_DESC(debug, "8139too bitmapped message enable number");
+MODULE_PARM_DESC(multicast_filter_limit, "8139too maximum number of filtered multicast addresses");
+MODULE_PARM_DESC(media, "8139too: Bits 4+9: force full duplex, bit 5: 100Mbps");
+MODULE_PARM_DESC(full_duplex, "8139too: Force full duplex for board(s) (1)");
+MODULE_PARM_DESC(bql_disable, "8139too: Disable Byte Queue Limits functionality (default: false)");
static int read_eeprom (void __iomem *ioaddr, int location, int addr_len);
static int rtl8139_open (struct net_device *dev);
@@ -1409,6 +1413,8 @@ static void rtl8139_hw_start (struct net_device *dev)
}
netdev_dbg(dev, "init buffer addresses\n");
+ if (unlikely(bql_disable))
+ netdev_reset_queue(dev);
/* Lock Config[01234] and BMCR register writes */
RTL_W8 (Cfg9346, Cfg9346_Lock);
@@ -1639,6 +1645,9 @@ static inline void rtl8139_tx_clear (struct rtl8139_private *tp)
tp->cur_tx = 0;
tp->dirty_tx = 0;
+ if (unlikely(bql_disable))
+ netdev_reset_queue(tp->dev);
+
/* XXX account for unsent Tx packets in tp->stats.tx_dropped */
}
@@ -1729,10 +1738,13 @@ static netdev_tx_t rtl8139_start_xmit (struct sk_buff *skb,
* to make sure that the device sees the updated data.
*/
wmb();
+ len = max_t(unsigned int, len, (unsigned int)ETH_ZLEN);
RTL_W32_F (TxStatus0 + (entry * sizeof (u32)),
- tp->tx_flag | max(len, (unsigned int)ETH_ZLEN));
+ tp->tx_flag | len);
tp->cur_tx++;
+ if (unlikely(bql_disable))
+ netdev_sent_queue(dev, len);
if ((tp->cur_tx - NUM_TX_DESC) == tp->dirty_tx)
netif_stop_queue (dev);
@@ -1750,6 +1762,7 @@ static void rtl8139_tx_interrupt (struct net_device *dev,
void __iomem *ioaddr)
{
unsigned long dirty_tx, tx_left;
+ unsigned bytes_compl = 0, pkts_compl = 0;
assert (dev != NULL);
assert (ioaddr != NULL);
@@ -1792,6 +1805,8 @@ static void rtl8139_tx_interrupt (struct net_device *dev,
u64_stats_update_begin(&tp->tx_stats.syncp);
tp->tx_stats.packets++;
tp->tx_stats.bytes += txstatus & 0x7ff;
+ pkts_compl++;
+ bytes_compl += txstatus & 0x7ff;
u64_stats_update_end(&tp->tx_stats.syncp);
}
@@ -1807,6 +1822,9 @@ static void rtl8139_tx_interrupt (struct net_device *dev,
}
#endif /* RTL8139_NDEBUG */
+ if (unlikely(bql_disable))
+ netdev_completed_queue(dev, pkts_compl, bytes_compl);
+
/* only wake the queue if we did work, and the queue is stopped */
if (tp->dirty_tx != dirty_tx) {
tp->dirty_tx = dirty_tx;
--
1.8.4.1
^ permalink raw reply related
* [PATCH net-next v3 02/07] r8169: Support for byte queue limits
From: Tino Reichardt @ 2013-10-20 19:23 UTC (permalink / raw)
To: netdev, Realtek linux nic maintainers, Igor Maravic,
Francois Romieu
In-Reply-To: <1382296991-20289-1-git-send-email-milky-kernel@mcmilk.de>
Changes to r8169 driver to use byte queue limits.
Signed-off-by: Tino Reichardt <milky-kernel@mcmilk.de>
---
drivers/net/ethernet/realtek/r8169.c | 14 ++++++++++++++
1 file changed, 14 insertions(+)
diff --git a/drivers/net/ethernet/realtek/r8169.c b/drivers/net/ethernet/realtek/r8169.c
index 3397cee..8f12145 100644
--- a/drivers/net/ethernet/realtek/r8169.c
+++ b/drivers/net/ethernet/realtek/r8169.c
@@ -313,6 +313,7 @@ MODULE_DEVICE_TABLE(pci, rtl8169_pci_tbl);
static int rx_buf_sz = 16383;
static int use_dac;
+static bool bql_disable;
static struct {
u32 msg_enable;
} debug = { -1 };
@@ -818,6 +819,9 @@ MODULE_AUTHOR("Realtek and the Linux r8169 crew <netdev@vger.kernel.org>");
MODULE_DESCRIPTION("RealTek RTL-8169 Gigabit Ethernet driver");
module_param(use_dac, int, 0);
MODULE_PARM_DESC(use_dac, "Enable PCI DAC. Unsafe on 32 bit PCI slot.");
+module_param(bql_disable, bool, 0);
+MODULE_PARM_DESC(bql_disable,
+ "Disable Byte Queue Limits functionality (default: false)");
module_param_named(debug, debug.msg_enable, int, 0);
MODULE_PARM_DESC(debug, "Debug verbosity level (0=none, ..., 16=all)");
MODULE_LICENSE("GPL");
@@ -5841,6 +5845,8 @@ static void rtl8169_tx_clear(struct rtl8169_private *tp)
{
rtl8169_tx_clear_range(tp, tp->dirty_tx, NUM_TX_DESC);
tp->cur_tx = tp->dirty_tx = 0;
+ if (unlikely(bql_disable))
+ netdev_reset_queue(tp->dev);
}
static void rtl_reset_work(struct rtl8169_private *tp)
@@ -6017,6 +6023,8 @@ static netdev_tx_t rtl8169_start_xmit(struct sk_buff *skb,
txd->opts2 = cpu_to_le32(opts[1]);
skb_tx_timestamp(skb);
+ if (unlikely(bql_disable))
+ netdev_sent_queue(dev, skb->len);
wmb();
@@ -6116,6 +6124,7 @@ static void rtl8169_pcierr_interrupt(struct net_device *dev)
static void rtl_tx(struct net_device *dev, struct rtl8169_private *tp)
{
unsigned int dirty_tx, tx_left;
+ unsigned int pkts_compl = 0, bytes_compl = 0;
dirty_tx = tp->dirty_tx;
smp_rmb();
@@ -6138,6 +6147,9 @@ static void rtl_tx(struct net_device *dev, struct rtl8169_private *tp)
tp->tx_stats.packets++;
tp->tx_stats.bytes += tx_skb->skb->len;
u64_stats_update_end(&tp->tx_stats.syncp);
+
+ bytes_compl += tx_skb->skb->len;
+ pkts_compl++;
dev_kfree_skb(tx_skb->skb);
tx_skb->skb = NULL;
}
@@ -6155,6 +6167,8 @@ static void rtl_tx(struct net_device *dev, struct rtl8169_private *tp)
* ring status.
*/
smp_mb();
+ if (unlikely(bql_disable))
+ netdev_completed_queue(dev, pkts_compl, bytes_compl);
if (netif_queue_stopped(dev) &&
TX_FRAGS_READY_FOR(tp, MAX_SKB_FRAGS)) {
netif_wake_queue(dev);
--
1.8.4.1
^ permalink raw reply related
* [PATCH net-next v3 03/07] tulip: Support for byte queue limits
From: Tino Reichardt @ 2013-10-20 19:23 UTC (permalink / raw)
To: netdev, Grant Grundler
In-Reply-To: <1382296991-20289-1-git-send-email-milky-kernel@mcmilk.de>
Changes to tulip driver to use byte queue limits.
Signed-off-by: Tino Reichardt <milky-kernel@mcmilk.de>
---
drivers/net/ethernet/dec/tulip/interrupt.c | 4 ++++
drivers/net/ethernet/dec/tulip/tulip.h | 1 +
drivers/net/ethernet/dec/tulip/tulip_core.c | 9 ++++++++-
3 files changed, 13 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/dec/tulip/interrupt.c b/drivers/net/ethernet/dec/tulip/interrupt.c
index 92306b3..7084267 100644
--- a/drivers/net/ethernet/dec/tulip/interrupt.c
+++ b/drivers/net/ethernet/dec/tulip/interrupt.c
@@ -532,6 +532,7 @@ irqreturn_t tulip_interrupt(int irq, void *dev_instance)
#endif
unsigned int work_count = tulip_max_interrupt_work;
unsigned int handled = 0;
+ unsigned int bytes_compl = 0;
/* Let's see whether the interrupt really is for us */
csr5 = ioread32(ioaddr + CSR5);
@@ -634,6 +635,7 @@ irqreturn_t tulip_interrupt(int irq, void *dev_instance)
PCI_DMA_TODEVICE);
/* Free the original skb. */
+ bytes_compl += tp->tx_buffers[entry].skb->len;
dev_kfree_skb_irq(tp->tx_buffers[entry].skb);
tp->tx_buffers[entry].skb = NULL;
tp->tx_buffers[entry].mapping = 0;
@@ -802,6 +804,8 @@ irqreturn_t tulip_interrupt(int irq, void *dev_instance)
}
#endif /* CONFIG_TULIP_NAPI */
+ if (likely(tulip_bql_disable == false))
+ netdev_completed_queue(dev, tx, bytes_compl);
if ((missed = ioread32(ioaddr + CSR8) & 0x1ffff)) {
dev->stats.rx_dropped += missed & 0x10000 ? 0x10000 : missed;
}
diff --git a/drivers/net/ethernet/dec/tulip/tulip.h b/drivers/net/ethernet/dec/tulip/tulip.h
index 38431a1..3c62870 100644
--- a/drivers/net/ethernet/dec/tulip/tulip.h
+++ b/drivers/net/ethernet/dec/tulip/tulip.h
@@ -513,6 +513,7 @@ void comet_timer(unsigned long data);
/* tulip_core.c */
extern int tulip_debug;
+extern bool tulip_bql_disable;
extern const char * const medianame[];
extern const char tulip_media_cap[];
extern struct tulip_chip_table tulip_tbl[];
diff --git a/drivers/net/ethernet/dec/tulip/tulip_core.c b/drivers/net/ethernet/dec/tulip/tulip_core.c
index 4e8cfa2..d02eaa1 100644
--- a/drivers/net/ethernet/dec/tulip/tulip_core.c
+++ b/drivers/net/ethernet/dec/tulip/tulip_core.c
@@ -106,11 +106,13 @@ static int csr0 = 0x00A00000 | 0x4800;
/* Time in jiffies before concluding the transmitter is hung. */
#define TX_TIMEOUT (4*HZ)
-
MODULE_AUTHOR("The Linux Kernel Team");
MODULE_DESCRIPTION("Digital 21*4* Tulip ethernet driver");
MODULE_LICENSE("GPL");
MODULE_VERSION(DRV_VERSION);
+module_param(tulip_bql_disable, bool, 0);
+MODULE_PARM_DESC(tulip_bql_disable,
+ "Disable Byte Queue Limits functionality (default: false)");
module_param(tulip_debug, int, 0);
module_param(max_interrupt_work, int, 0);
module_param(rx_copybreak, int, 0);
@@ -123,6 +125,7 @@ int tulip_debug = TULIP_DEBUG;
#else
int tulip_debug = 1;
#endif
+bool tulip_bql_disable;
static void tulip_timer(unsigned long data)
{
@@ -703,6 +706,8 @@ tulip_start_xmit(struct sk_buff *skb, struct net_device *dev)
wmb();
tp->cur_tx++;
+ if (likely(tulip_bql_disable == false))
+ netdev_sent_queue(dev, skb->len);
/* Trigger an immediate transmit demand. */
iowrite32(0, tp->base_addr + CSR1);
@@ -746,6 +751,8 @@ static void tulip_clean_tx_ring(struct tulip_private *tp)
tp->tx_buffers[entry].skb = NULL;
tp->tx_buffers[entry].mapping = 0;
}
+ if (likely(tulip_bql_disable == false))
+ netdev_reset_queue(tp->dev);
}
static void tulip_down (struct net_device *dev)
--
1.8.4.1
^ permalink raw reply related
* [PATCH net-next v3 04/07] via-rhine: Support for byte queue limits
From: Tino Reichardt @ 2013-10-20 19:23 UTC (permalink / raw)
To: netdev, Roger Luethi
In-Reply-To: <1382296991-20289-1-git-send-email-milky-kernel@mcmilk.de>
Changes to via-rhine driver to use byte queue limits.
Signed-off-by: Tino Reichardt <milky-kernel@mcmilk.de>
---
drivers/net/ethernet/via/via-rhine.c | 17 +++++++++++++++++
1 file changed, 17 insertions(+)
diff --git a/drivers/net/ethernet/via/via-rhine.c b/drivers/net/ethernet/via/via-rhine.c
index bdf697b..067b4ab 100644
--- a/drivers/net/ethernet/via/via-rhine.c
+++ b/drivers/net/ethernet/via/via-rhine.c
@@ -53,6 +53,8 @@ static int rx_copybreak = 1518;
static int rx_copybreak;
#endif
+static bool bql_disable;
+
/* Work-around for broken BIOSes: they are unable to get the chip back out of
power state D3 so PXE booting fails. bootparam(7): via-rhine.avoid_D3=1 */
static bool avoid_D3;
@@ -128,9 +130,11 @@ MODULE_DESCRIPTION("VIA Rhine PCI Fast Ethernet driver");
MODULE_LICENSE("GPL");
module_param(debug, int, 0);
+module_param(bql_disable, bool, 0);
module_param(rx_copybreak, int, 0);
module_param(avoid_D3, bool, 0);
MODULE_PARM_DESC(debug, "VIA Rhine debug message flags");
+MODULE_PARM_DESC(bql_disable, "Disable Byte Queue Limits functionality (default: false)");
MODULE_PARM_DESC(rx_copybreak, "VIA Rhine copy breakpoint for copy-only-tiny-frames");
MODULE_PARM_DESC(avoid_D3, "Avoid power state D3 (work-around for broken BIOSes)");
@@ -1220,6 +1224,8 @@ static void alloc_tbufs(struct net_device* dev)
}
rp->tx_ring[i-1].next_desc = cpu_to_le32(rp->tx_ring_dma);
+ if (unlikely(bql_disable))
+ netdev_reset_queue(dev);
}
static void free_tbufs(struct net_device* dev)
@@ -1719,6 +1725,8 @@ static netdev_tx_t rhine_start_tx(struct sk_buff *skb,
/* lock eth irq */
wmb();
rp->tx_ring[entry].tx_status |= cpu_to_le32(DescOwn);
+ if (unlikely(bql_disable))
+ netdev_sent_queue(dev, skb->len);
wmb();
rp->cur_tx++;
@@ -1783,6 +1791,7 @@ static void rhine_tx(struct net_device *dev)
{
struct rhine_private *rp = netdev_priv(dev);
int txstatus = 0, entry = rp->dirty_tx % TX_RING_SIZE;
+ unsigned int pkts_compl = 0, bytes_compl = 0;
/* find and cleanup dirty tx descriptors */
while (rp->dirty_tx != rp->cur_tx) {
@@ -1830,10 +1839,18 @@ static void rhine_tx(struct net_device *dev)
rp->tx_skbuff[entry]->len,
PCI_DMA_TODEVICE);
}
+
+ bytes_compl += rp->tx_skbuff[entry]->len;
+ pkts_compl++;
dev_kfree_skb(rp->tx_skbuff[entry]);
+
rp->tx_skbuff[entry] = NULL;
entry = (++rp->dirty_tx) % TX_RING_SIZE;
}
+
+ if (unlikely(bql_disable))
+ netdev_completed_queue(dev, pkts_compl, bytes_compl);
+
if ((rp->cur_tx - rp->dirty_tx) < TX_QUEUE_LEN - 4)
netif_wake_queue(dev);
}
--
1.8.4.1
^ permalink raw reply related
* [PATCH net-next v3 05/07] via-velocity: Support for byte queue limits
From: Tino Reichardt @ 2013-10-20 19:23 UTC (permalink / raw)
To: netdev, Francois Romieu
In-Reply-To: <1382296991-20289-1-git-send-email-milky-kernel@mcmilk.de>
Changes to via-velocity driver to use byte queue limits.
Signed-off-by: Tino Reichardt <milky-kernel@mcmilk.de>
---
drivers/net/ethernet/via/via-velocity.c | 18 ++++++++++++++++++
1 file changed, 18 insertions(+)
diff --git a/drivers/net/ethernet/via/via-velocity.c b/drivers/net/ethernet/via/via-velocity.c
index d022bf9..b084404 100644
--- a/drivers/net/ethernet/via/via-velocity.c
+++ b/drivers/net/ethernet/via/via-velocity.c
@@ -1,3 +1,4 @@
+
/*
* This code is derived from the VIA reference driver (copyright message
* below) provided to Red Hat by VIA Networking Technologies, Inc. for
@@ -368,6 +369,11 @@ static int rx_copybreak = 200;
module_param(rx_copybreak, int, 0644);
MODULE_PARM_DESC(rx_copybreak, "Copy breakpoint for copy-only-tiny-frames");
+static bool bql_disable;
+module_param(bql_disable, bool, 0);
+MODULE_PARM_DESC(bql_disable,
+ "Disable Byte Queue Limits functionality (default: false)");
+
/*
* Internal board variants. At the moment we have only one
*/
@@ -1751,6 +1757,9 @@ static void velocity_free_tx_buf(struct velocity_info *vptr,
le16_to_cpu(pktlen), DMA_TO_DEVICE);
}
}
+
+ if (unlikely(bql_disable))
+ netdev_reset_queue(vptr->netdev);
dev_kfree_skb_irq(skb);
tdinfo->skb = NULL;
}
@@ -1915,6 +1924,7 @@ static int velocity_tx_srv(struct velocity_info *vptr)
int works = 0;
struct velocity_td_info *tdinfo;
struct net_device_stats *stats = &vptr->netdev->stats;
+ unsigned int pkts_compl = 0, bytes_compl = 0;
for (qnum = 0; qnum < vptr->tx.numq; qnum++) {
for (idx = vptr->tx.tail[qnum]; vptr->tx.used[qnum] > 0;
@@ -1946,6 +1956,8 @@ static int velocity_tx_srv(struct velocity_info *vptr)
} else {
stats->tx_packets++;
stats->tx_bytes += tdinfo->skb->len;
+ pkts_compl++;
+ bytes_compl += tdinfo->skb->len;
}
velocity_free_tx_buf(vptr, tdinfo, td);
vptr->tx.used[qnum]--;
@@ -1955,6 +1967,10 @@ static int velocity_tx_srv(struct velocity_info *vptr)
if (AVAIL_TD(vptr, qnum) < 1)
full = 1;
}
+
+ if (unlikely(bql_disable))
+ netdev_completed_queue(vptr->netdev, pkts_compl, bytes_compl);
+
/*
* Look to see if we should kick the transmit network
* layer for more work.
@@ -2641,6 +2657,8 @@ static netdev_tx_t velocity_xmit(struct sk_buff *skb,
td_ptr->td_buf[0].size |= TD_QUEUE;
mac_tx_queue_wake(vptr->mac_regs, qnum);
+ if (unlikely(bql_disable))
+ netdev_sent_queue(vptr->netdev, skb->len);
spin_unlock_irqrestore(&vptr->lock, flags);
out:
return NETDEV_TX_OK;
--
1.8.4.1
^ permalink raw reply related
* [PATCH net-next v3 06/07] 3c59x: Support for byte queue limits
From: Tino Reichardt @ 2013-10-20 19:23 UTC (permalink / raw)
To: netdev, Steffen Klassert
In-Reply-To: <1382296991-20289-1-git-send-email-milky-kernel@mcmilk.de>
Changes to 3c59x driver to use byte queue limits.
Signed-off-by: Tino Reichardt <milky-kernel@mcmilk.de>
---
drivers/net/ethernet/3com/3c59x.c | 47 ++++++++++++++++++++++++++++++++++-----
1 file changed, 41 insertions(+), 6 deletions(-)
diff --git a/drivers/net/ethernet/3com/3c59x.c b/drivers/net/ethernet/3com/3c59x.c
index ad5272b..6a7b77d 100644
--- a/drivers/net/ethernet/3com/3c59x.c
+++ b/drivers/net/ethernet/3com/3c59x.c
@@ -795,6 +795,7 @@ static int global_options = -1;
static int global_full_duplex = -1;
static int global_enable_wol = -1;
static int global_use_mmio = -1;
+static bool bql_disable;
/* Variables to work-around the Compaq PCI BIOS32 problem. */
static int compaq_ioaddr, compaq_irq, compaq_device_id = 0x5900;
@@ -819,6 +820,8 @@ module_param(compaq_device_id, int, 0);
module_param(watchdog, int, 0);
module_param(global_use_mmio, int, 0);
module_param_array(use_mmio, int, NULL, 0);
+module_param(bql_disable, bool, 0);
+
MODULE_PARM_DESC(debug, "3c59x debug level (0-6)");
MODULE_PARM_DESC(options, "3c59x: Bits 0-3: media type, bit 4: bus mastering, bit 9: full duplex");
MODULE_PARM_DESC(global_options, "3c59x: same as options, but applies to all NICs if options is unset");
@@ -836,6 +839,7 @@ MODULE_PARM_DESC(compaq_device_id, "3c59x PCI device ID (Compaq BIOS problem wor
MODULE_PARM_DESC(watchdog, "3c59x transmit timeout in milliseconds");
MODULE_PARM_DESC(global_use_mmio, "3c59x: same as use_mmio, but applies to all NICs if options is unset");
MODULE_PARM_DESC(use_mmio, "3c59x: use memory-mapped PCI I/O resource (0-1)");
+MODULE_PARM_DESC(bql_disable, "3c59x: Disable Byte Queue Limits functionality (default: false)");
#ifdef CONFIG_NET_POLL_CONTROLLER
static void poll_vortex(struct net_device *dev)
@@ -1726,6 +1730,9 @@ vortex_up(struct net_device *dev)
iowrite16(vp->intr_enable, ioaddr + EL3_CMD);
if (vp->cb_fn_base) /* The PCMCIA people are idiots. */
iowrite32(0x8000, vp->cb_fn_base + 4);
+
+ if (unlikely(bql_disable))
+ netdev_reset_queue(dev);
netif_start_queue (dev);
err_out:
return err;
@@ -2080,10 +2087,15 @@ vortex_start_xmit(struct sk_buff *skb, struct net_device *dev)
spin_unlock_irq(&vp->window_lock);
vp->tx_skb = skb;
iowrite16(StartDMADown, ioaddr + EL3_CMD);
+ if (unlikely(bql_disable))
+ netdev_sent_queue(dev, len);
/* netif_wake_queue() will be called at the DMADone interrupt. */
} else {
/* ... and the packet rounded to a doubleword. */
- iowrite32_rep(ioaddr + TX_FIFO, skb->data, (skb->len + 3) >> 2);
+ int len = (skb->len + 3) >> 2;
+ iowrite32_rep(ioaddr + TX_FIFO, skb->data, len);
+ if (unlikely(bql_disable))
+ netdev_sent_queue(dev, len);
dev_kfree_skb (skb);
if (ioread16(ioaddr + TxFree) > 1536) {
netif_start_queue (dev); /* AKPM: redundant? */
@@ -2094,7 +2106,6 @@ vortex_start_xmit(struct sk_buff *skb, struct net_device *dev)
}
}
-
/* Clear the Tx status stack. */
{
int tx_status;
@@ -2164,12 +2175,14 @@ boomerang_start_xmit(struct sk_buff *skb, struct net_device *dev)
vp->tx_ring[entry].frag[0].addr = cpu_to_le32(pci_map_single(VORTEX_PCI(vp), skb->data,
skb->len, PCI_DMA_TODEVICE));
vp->tx_ring[entry].frag[0].length = cpu_to_le32(skb->len | LAST_FRAG);
+ netdev_sent_queue(dev, skb->len);
} else {
- int i;
+ int i, len;
vp->tx_ring[entry].frag[0].addr = cpu_to_le32(pci_map_single(VORTEX_PCI(vp), skb->data,
skb_headlen(skb), PCI_DMA_TODEVICE));
vp->tx_ring[entry].frag[0].length = cpu_to_le32(skb_headlen(skb));
+ len = skb_headlen(skb);
for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
skb_frag_t *frag = &skb_shinfo(skb)->frags[i];
@@ -2180,16 +2193,21 @@ boomerang_start_xmit(struct sk_buff *skb, struct net_device *dev)
(void *)skb_frag_address(frag),
skb_frag_size(frag), PCI_DMA_TODEVICE));
- if (i == skb_shinfo(skb)->nr_frags-1)
+ if (i == skb_shinfo(skb)->nr_frags - 1) {
vp->tx_ring[entry].frag[i+1].length = cpu_to_le32(skb_frag_size(frag)|LAST_FRAG);
- else
+ len += skb_frag_size(frag) | LAST_FRAG;
+ } else {
vp->tx_ring[entry].frag[i+1].length = cpu_to_le32(skb_frag_size(frag));
+ len += skb_frag_size(frag);
+ }
}
+ netdev_sent_queue(dev, len);
}
#else
vp->tx_ring[entry].addr = cpu_to_le32(pci_map_single(VORTEX_PCI(vp), skb->data, skb->len, PCI_DMA_TODEVICE));
vp->tx_ring[entry].length = cpu_to_le32(skb->len | LAST_FRAG);
vp->tx_ring[entry].status = cpu_to_le32(skb->len | TxIntrUploaded);
+ netdev_sent_queue(dev, skb->len | LAST_FRAG);
#endif
spin_lock_irqsave(&vp->lock, flags);
@@ -2234,6 +2252,7 @@ vortex_interrupt(int irq, void *dev_id)
int status;
int work_done = max_interrupt_work;
int handled = 0;
+ unsigned bytes_compl = 0, pkts_compl = 0;
ioaddr = vp->ioaddr;
spin_lock(&vp->lock);
@@ -2279,8 +2298,12 @@ vortex_interrupt(int irq, void *dev_id)
if (status & DMADone) {
if (ioread16(ioaddr + Wn7_MasterStatus) & 0x1000) {
+ int len;
iowrite16(0x1000, ioaddr + Wn7_MasterStatus); /* Ack the event. */
- pci_unmap_single(VORTEX_PCI(vp), vp->tx_skb_dma, (vp->tx_skb->len + 3) & ~3, PCI_DMA_TODEVICE);
+ len = (vp->tx_skb->len + 3) & ~3;
+ pci_unmap_single(VORTEX_PCI(vp), vp->tx_skb_dma, len, PCI_DMA_TODEVICE);
+ bytes_compl += len;
+ pkts_compl++;
dev_kfree_skb_irq(vp->tx_skb); /* Release the transferred buffer */
if (ioread16(ioaddr + TxFree) > 1536) {
/*
@@ -2327,6 +2350,9 @@ vortex_interrupt(int irq, void *dev_id)
spin_unlock(&vp->window_lock);
+ if (unlikely(bql_disable))
+ netdev_completed_queue(dev, pkts_compl, bytes_compl);
+
if (vortex_debug > 4)
pr_debug("%s: exiting interrupt, status %4.4x.\n",
dev->name, status);
@@ -2348,6 +2374,7 @@ boomerang_interrupt(int irq, void *dev_id)
void __iomem *ioaddr;
int status;
int work_done = max_interrupt_work;
+ unsigned bytes_compl = 0, pkts_compl = 0;
ioaddr = vp->ioaddr;
@@ -2420,6 +2447,8 @@ boomerang_interrupt(int irq, void *dev_id)
pci_unmap_single(VORTEX_PCI(vp),
le32_to_cpu(vp->tx_ring[entry].addr), skb->len, PCI_DMA_TODEVICE);
#endif
+ bytes_compl += skb->len;
+ pkts_compl++;
dev_kfree_skb_irq(skb);
vp->tx_skbuff[entry] = NULL;
} else {
@@ -2467,6 +2496,10 @@ boomerang_interrupt(int irq, void *dev_id)
handler_exit:
vp->handling_irq = 0;
spin_unlock(&vp->lock);
+
+ if (unlikely(bql_disable))
+ netdev_completed_queue(dev, pkts_compl, bytes_compl);
+
return IRQ_HANDLED;
}
@@ -2660,6 +2693,8 @@ vortex_down(struct net_device *dev, int final_down)
struct vortex_private *vp = netdev_priv(dev);
void __iomem *ioaddr = vp->ioaddr;
+ if (unlikely(bql_disable))
+ netdev_reset_queue(dev);
netif_stop_queue (dev);
del_timer_sync(&vp->rx_oom_timer);
--
1.8.4.1
^ permalink raw reply related
* [PATCH net-next v3 07/07] natsemi: Support for byte queue limits
From: Tino Reichardt @ 2013-10-20 19:23 UTC (permalink / raw)
To: netdev, Greg Kroah-Hartman, David S. Miller, Jiri Pirko,
Bill Pemberton
In-Reply-To: <1382296991-20289-1-git-send-email-milky-kernel@mcmilk.de>
Changes to natsemi driver to use byte queue limits.
Signed-off-by: Tino Reichardt <milky-kernel@mcmilk.de>
---
drivers/net/ethernet/natsemi/natsemi.c | 21 +++++++++++++++++++++
1 file changed, 21 insertions(+)
diff --git a/drivers/net/ethernet/natsemi/natsemi.c b/drivers/net/ethernet/natsemi/natsemi.c
index 7a5e295..1f6efbc 100644
--- a/drivers/net/ethernet/natsemi/natsemi.c
+++ b/drivers/net/ethernet/natsemi/natsemi.c
@@ -71,6 +71,8 @@
NETIF_MSG_TX_ERR)
static int debug = -1;
+static bool bql_disable;
+
static int mtu;
/* Maximum number of multicast addresses to filter (vs. rx-all-multicast).
@@ -139,12 +141,15 @@ MODULE_LICENSE("GPL");
module_param(mtu, int, 0);
module_param(debug, int, 0);
+module_param(bql_disable, bool, 0);
module_param(rx_copybreak, int, 0);
module_param(dspcfg_workaround, int, 0);
module_param_array(options, int, NULL, 0);
module_param_array(full_duplex, int, NULL, 0);
MODULE_PARM_DESC(mtu, "DP8381x MTU (all boards)");
MODULE_PARM_DESC(debug, "DP8381x default debug level");
+MODULE_PARM_DESC(bql_disable,
+ "Disable Byte Queue Limits functionality (default: false)");
MODULE_PARM_DESC(rx_copybreak,
"DP8381x copy breakpoint for copy-only-tiny-frames");
MODULE_PARM_DESC(dspcfg_workaround, "DP8381x: control DspCfg workaround");
@@ -1974,6 +1979,9 @@ static void init_ring(struct net_device *dev)
np->tx_ring[i].cmd_status = 0;
}
+ if (unlikely(bql_disable))
+ netdev_reset_queue(dev);
+
/* 2) RX ring */
np->dirty_rx = 0;
np->cur_rx = RX_RING_SIZE;
@@ -2012,6 +2020,9 @@ static void drain_tx(struct net_device *dev)
}
np->tx_skbuff[i] = NULL;
}
+
+ if (unlikely(bql_disable))
+ netdev_reset_queue(dev);
}
static void drain_rx(struct net_device *dev)
@@ -2116,6 +2127,9 @@ static netdev_tx_t start_tx(struct sk_buff *skb, struct net_device *dev)
dev_kfree_skb_irq(skb);
dev->stats.tx_dropped++;
}
+
+ if (unlikely(bql_disable))
+ netdev_sent_queue(dev, skb->len);
spin_unlock_irqrestore(&np->lock, flags);
if (netif_msg_tx_queued(np)) {
@@ -2128,6 +2142,7 @@ static netdev_tx_t start_tx(struct sk_buff *skb, struct net_device *dev)
static void netdev_tx_done(struct net_device *dev)
{
struct netdev_private *np = netdev_priv(dev);
+ unsigned bytes_compl = 0, pkts_compl = 0;
for (; np->cur_tx - np->dirty_tx > 0; np->dirty_tx++) {
int entry = np->dirty_tx % TX_RING_SIZE;
@@ -2158,9 +2173,15 @@ static void netdev_tx_done(struct net_device *dev)
np->tx_skbuff[entry]->len,
PCI_DMA_TODEVICE);
/* Free the original skb. */
+ bytes_compl += np->tx_skbuff[entry]->len;
+ pkts_compl++;
dev_kfree_skb_irq(np->tx_skbuff[entry]);
np->tx_skbuff[entry] = NULL;
}
+
+ if (unlikely(bql_disable))
+ netdev_completed_queue(dev, pkts_compl, bytes_compl);
+
if (netif_queue_stopped(dev) &&
np->cur_tx - np->dirty_tx < TX_QUEUE_LEN - 4) {
/* The ring is no longer full, wake queue. */
--
1.8.4.1
^ permalink raw reply related
* Re: [PATCHSET net-next v3 00/07] Support for byte queue limits on various drivers
From: Joe Perches @ 2013-10-20 19:38 UTC (permalink / raw)
To: Tino Reichardt; +Cc: netdev, David S. Miller
In-Reply-To: <1382296991-20289-1-git-send-email-milky-kernel@mcmilk.de>
On Sun, 2013-10-20 at 21:23 +0200, Tino Reichardt wrote:
> Hello again,
Hi again Tino.
> "if (likely(bql_disable == false))" was replaced by
> "if (unlikely(bql_disable))" ...
Sorry for suggesting this, but it's not an
equivalent transform unless the then & else
paths are also inverted.
Here there are no else branches.
Doesn't this need to be?
from:
if (likely(bql_disable == false))
to:
if (likely(!bql_disable))
^ permalink raw reply
* Re: [PATCHSET net-next v3 00/07] Support for byte queue limits on various drivers
From: Tino Reichardt @ 2013-10-20 20:19 UTC (permalink / raw)
To: netdev
In-Reply-To: <1382296991-20289-1-git-send-email-milky-kernel@mcmilk.de>
* Tino Reichardt <milky-kernel@mcmilk.de> wrote:
> Hello again,
>
> this patchset is intended for net-next/linux-3.13.
>
> It adds support for byte queue limits on various network drivers.
>
> Since not all of them are fully tested by now, I added an bql_disable
> module parameter, which can be used to disable the BQL code.
>
> "if (likely(bql_disable == false))" was replaced by
> "if (unlikely(bql_disable))" ...
Please ignore Patchset v3... "unlikely(bql_disable)" is of cause not
correct :(
Please use "[PATCHSET net-next v3 00/07]" instead.
Sorry for all the noise here ;)
--
Best regards, TR
^ permalink raw reply
* Fw: [Bug 63321] New: tcp_fastretrans_alert
From: Stephen Hemminger @ 2013-10-20 20:40 UTC (permalink / raw)
To: netdev
Begin forwarded message:
Date: Sun, 20 Oct 2013 06:20:25 -0700
From: "bugzilla-daemon@bugzilla.kernel.org" <bugzilla-daemon@bugzilla.kernel.org>
To: "stephen@networkplumber.org" <stephen@networkplumber.org>
Subject: [Bug 63321] New: tcp_fastretrans_alert
https://bugzilla.kernel.org/show_bug.cgi?id=63321
Bug ID: 63321
Summary: tcp_fastretrans_alert
Product: Networking
Version: 2.5
Kernel Version: 3.11.6
Hardware: x86-64
OS: Linux
Tree: Mainline
Status: NEW
Severity: normal
Priority: P1
Component: IPV4
Assignee: shemminger@linux-foundation.org
Reporter: warrior-elite@yandex.ru
Regression: No
[ 865.064799] ------------[ cut here ]------------
[ 865.064813] WARNING: CPU: 0 PID: 0 at
/home/apw/COD/linux/net/ipv4/tcp_input.c:2711
tcp_fastretrans_alert+0x5df/0x620()
[ 865.064815] Modules linked in: pci_stub vboxpci(OF) vboxnetadp(OF)
vboxnetflt(OF) joydev hid_generic usbhid hid vboxdrv(OF) binfmt_misc
snd_hda_codec_realtek ppdev uvcvideo videobuf2_vmalloc videobuf2_memops
videobuf2_core videodev snd_usb_audio snd_usbmidi_lib nvidia(POF) snd_hda_intel
snd_hda_codec snd_hwdep snd_pcm kvm_amd kvm snd_page_alloc snd_seq_midi
snd_seq_midi_event snd_rawmidi snd_seq microcode snd_seq_device snd_timer snd
serio_raw soundcore edac_core k10temp edac_mce_amd drm sp5100_tco ohci_pci
i2c_piix4 wmi mac_hid parport_pc lp parport pata_acpi ahci r8169 pata_atiixp
libahci mii
[ 865.064843] CPU: 0 PID: 0 Comm: swapper/0 Tainted: PF O
3.11.6-031106-generic #201310181453
[ 865.064845] Hardware name: Gigabyte Technology Co., Ltd.
GA-MA770-UD3/GA-MA770-UD3, BIOS FKb 07/08/2010
[ 865.064846] 0000000000000a97 ffff88012fc039a0 ffffffff81714a56
000000000000259d
[ 865.064849] 0000000000000000 ffff88012fc039e0 ffffffff8106539c
ffff88012fc03a00
[ 865.064851] ffff880114f45400 000000000000000c 0000000000004120
0000000000000001
[ 865.064853] Call Trace:
[ 865.064855] <IRQ> [<ffffffff81714a56>] dump_stack+0x46/0x58
[ 865.064861] [<ffffffff8106539c>] warn_slowpath_common+0x8c/0xc0
[ 865.064863] [<ffffffff810653ea>] warn_slowpath_null+0x1a/0x20
[ 865.064865] [<ffffffff8165f9df>] tcp_fastretrans_alert+0x5df/0x620
[ 865.064867] [<ffffffff816608c8>] tcp_ack+0x498/0x630
[ 865.064868] [<ffffffff8166109a>] tcp_rcv_established+0x31a/0x780
[ 865.064871] [<ffffffff8166b3b5>] tcp_v4_do_rcv+0x155/0x240
[ 865.064873] [<ffffffff8166d230>] tcp_v4_rcv+0x5f0/0x770
[ 865.064875] [<ffffffff81099f42>] ? try_to_wake_up+0x192/0x210
[ 865.064878] [<ffffffff81647a6e>] ip_local_deliver_finish+0xce/0x250
[ 865.064880] [<ffffffff81647d7b>] ip_local_deliver+0x4b/0x90
[ 865.064881] [<ffffffff81647761>] ip_rcv_finish+0x121/0x360
[ 865.064883] [<ffffffff81647ff6>] ip_rcv+0x236/0x350
[ 865.064886] [<ffffffff81611a2a>] __netif_receive_skb_core+0x5ca/0x720
[ 865.064888] [<ffffffff81611ba1>] __netif_receive_skb+0x21/0x70
[ 865.064890] [<ffffffff81611c13>] netif_receive_skb+0x23/0x90
[ 865.064891] [<ffffffff81612330>] napi_gro_receive+0xb0/0x130
[ 865.064898] [<ffffffffa00217f2>] rtl_rx+0xe2/0x320 [r8169]
[ 865.064903] [<ffffffffa00254c5>] rtl8169_poll+0x105/0x120 [r8169]
[ 865.064905] [<ffffffff8108d00e>] ? hrtimer_get_next_event+0xce/0xd0
[ 865.064907] [<ffffffff81612ca4>] net_rx_action+0x134/0x260
[ 865.064910] [<ffffffff814734c1>] ? add_interrupt_randomness+0x41/0x190
[ 865.064913] [<ffffffff8106a510>] __do_softirq+0xe0/0x280
[ 865.064915] [<ffffffff8172005e>] ? _raw_spin_lock+0xe/0x20
[ 865.064918] [<ffffffff8172ae9c>] call_softirq+0x1c/0x30
[ 865.064921] [<ffffffff81015e35>] do_softirq+0x65/0xa0
[ 865.064923] [<ffffffff8106a82e>] irq_exit+0x9e/0xc0
[ 865.064924] [<ffffffff8172b7a3>] do_IRQ+0x63/0xe0
[ 865.064926] [<ffffffff8172066d>] common_interrupt+0x6d/0x6d
[ 865.064927] <EOI> [<ffffffff810504f6>] ? native_safe_halt+0x6/0x10
[ 865.064932] [<ffffffff8101c803>] default_idle+0x23/0xe0
[ 865.064934] [<ffffffff8101c93a>] amd_e400_idle+0x7a/0x100
[ 865.064936] [<ffffffff8101d096>] arch_cpu_idle+0x26/0x30
[ 865.064938] [<ffffffff810ba77e>] cpu_idle_loop+0x7e/0x250
[ 865.064940] [<ffffffff810ba9bb>] cpu_startup_entry+0x6b/0x70
[ 865.064943] [<ffffffff816fcbd7>] rest_init+0x77/0x80
[ 865.064946] [<ffffffff81d24f23>] start_kernel+0x407/0x414
[ 865.064948] [<ffffffff81d249ae>] ? do_early_param+0x87/0x87
[ 865.064950] [<ffffffff81d24120>] ? early_idt_handlers+0x120/0x120
[ 865.064952] [<ffffffff81d245e8>] x86_64_start_reservations+0x2a/0x2c
[ 865.064954] [<ffffffff81d246f2>] x86_64_start_kernel+0x108/0x117
[ 865.064955] ---[ end trace 7fed0fa1e4d5f06f ]---
--
You are receiving this mail because:
You are the assignee for the bug.
^ permalink raw reply
* Re: [PATCHSET net-next v3 00/07] Support for byte queue limits on various drivers
From: Andi Kleen @ 2013-10-20 21:13 UTC (permalink / raw)
To: Tino Reichardt; +Cc: netdev, David S. Miller
In-Reply-To: <1382296991-20289-1-git-send-email-milky-kernel@mcmilk.de>
Tino Reichardt <milky-kernel@mcmilk.de> writes:
>
> Since not all of them are fully tested by now, I added an bql_disable
> module parameter, which can be used to disable the BQL code.
>
> "if (likely(bql_disable == false))" was replaced by
> "if (unlikely(bql_disable))" ...
If it's untested it should probably be disabled by default,
until someone tests it.
-Andi
--
ak@linux.intel.com -- Speaking for myself only
^ permalink raw reply
* Re: [PATCHSET net-next v3 00/07] Support for byte queue limits on various drivers
From: Francois Romieu @ 2013-10-20 22:07 UTC (permalink / raw)
To: Tino Reichardt; +Cc: netdev, David S. Miller
In-Reply-To: <1382296991-20289-1-git-send-email-milky-kernel@mcmilk.de>
Tino Reichardt <milky-kernel@mcmilk.de> :
[...]
> this patchset is intended for net-next/linux-3.13.
>
> It adds support for byte queue limits on various network drivers.
>
> Since not all of them are fully tested by now, I added an bql_disable
> module parameter, which can be used to disable the BQL code.
(regarding the r8169 changes)
The module parameter is pointless: you previously said you tested the
r8169 changes, there is a whole release cycle ahead and we already
have enough kernels that can be choosen amongst if some specific
feature turns unexpectedly wrong.
This way you won't have to ponder if you should keep the original
author's Signed-off-by in the r8169 driver or not.
--
Ueimor
^ permalink raw reply
* Re: [PATCH net-next] bonding: move bond-specific init after enslave happens
From: Ding Tianhong @ 2013-10-21 1:35 UTC (permalink / raw)
To: Veaceslav Falico; +Cc: netdev, jiri, Jay Vosburgh, Andy Gospodarek
In-Reply-To: <1382273273-27344-1-git-send-email-vfalico@redhat.com>
On 2013/10/20 20:47, Veaceslav Falico wrote:
> As Jiri noted, currently we first do all bonding-specific initialization
> (specifically - bond_select_active_slave(bond)) before we actually attach
> the slave (so that it becomes visible through bond_for_each_slave() and
> friends). This might result in bond_select_active_slave() not seeing the
> first/new slave and, thus, not actually selecting an active slave.
>
> Fix this by moving all the bond-related init part after we've actually
> completely initialized and linked (via bond_master_upper_dev_link()) the
> new slave.
>
> After this we have all the initialization of the new slave *before*
> linking, and all the stuff that needs to be done on bonding *after* it. It
> has also a bonus effect - we can remove the locking on the new slave init
> completely, and only use it for bond_select_active_slave().
>
> Reported-by: Jiri Pirko <jiri@resnulli.us>
> CC: Jay Vosburgh <fubar@us.ibm.com>
> CC: Andy Gospodarek <andy@greyhouse.net>
> Signed-off-by: Veaceslav Falico <vfalico@redhat.com>
> ---
> drivers/net/bonding/bond_main.c | 29 ++++++++++-------------------
> 1 file changed, 10 insertions(+), 19 deletions(-)
>
> diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
> index d90734f..047c0fb 100644
> --- a/drivers/net/bonding/bond_main.c
> +++ b/drivers/net/bonding/bond_main.c
> @@ -1471,22 +1471,14 @@ int bond_enslave(struct net_device *bond_dev, struct net_device *slave_dev)
> goto err_close;
> }
>
> - write_lock_bh(&bond->lock);
> -
> prev_slave = bond_last_slave(bond);
> bond_attach_slave(bond, new_slave);
>
> new_slave->delay = 0;
> new_slave->link_failure_count = 0;
>
> - write_unlock_bh(&bond->lock);
> -
> - bond_compute_features(bond);
> -
> bond_update_speed_duplex(new_slave);
>
> - read_lock(&bond->lock);
> -
> new_slave->last_arp_rx = jiffies -
> (msecs_to_jiffies(bond->params.arp_interval) + 1);
> for (i = 0; i < BOND_MAX_ARP_TARGETS; i++)
> @@ -1547,12 +1539,9 @@ int bond_enslave(struct net_device *bond_dev, struct net_device *slave_dev)
> }
> }
>
> - write_lock_bh(&bond->curr_slave_lock);
> -
> switch (bond->params.mode) {
> case BOND_MODE_ACTIVEBACKUP:
> bond_set_slave_inactive_flags(new_slave);
> - bond_select_active_slave(bond);
> break;
> case BOND_MODE_8023AD:
> /* in 802.3ad mode, the internal mechanism
> @@ -1578,7 +1567,6 @@ int bond_enslave(struct net_device *bond_dev, struct net_device *slave_dev)
> case BOND_MODE_ALB:
> bond_set_active_slave(new_slave);
> bond_set_slave_inactive_flags(new_slave);
> - bond_select_active_slave(bond);
> break;
> default:
> pr_debug("This slave is always active in trunk mode\n");
> @@ -1596,10 +1584,6 @@ int bond_enslave(struct net_device *bond_dev, struct net_device *slave_dev)
> break;
> } /* switch(bond_mode) */
>
> - write_unlock_bh(&bond->curr_slave_lock);
> -
> - bond_set_carrier(bond);
> -
> #ifdef CONFIG_NET_POLL_CONTROLLER
> slave_dev->npinfo = bond->dev->npinfo;
> if (slave_dev->npinfo) {
> @@ -1614,8 +1598,6 @@ int bond_enslave(struct net_device *bond_dev, struct net_device *slave_dev)
> }
> #endif
>
> - read_unlock(&bond->lock);
> -
> res = netdev_rx_handler_register(slave_dev, bond_handle_frame,
> new_slave);
> if (res) {
> @@ -1629,6 +1611,16 @@ int bond_enslave(struct net_device *bond_dev, struct net_device *slave_dev)
> goto err_unregister;
> }
>
> + bond_compute_features(bond);
> + bond_set_carrier(bond);
> +
> + if (USES_PRIMARY(bond->params.mode)) {
> + read_lock(&bond->lock);
> + write_lock_bh(&bond->curr_slave_lock);
> + bond_select_active_slave(bond);
> + write_unlock_bh(&bond->curr_slave_lock);
> + read_unlock(&bond->lock);
> + }
>
agree to move the lock, and I think bond_attach_slave() should add here,
as it look more logical, the slave_cnt should not add before the slave truly
add to the bond.
Regards.
Ding
> pr_info("%s: enslaving %s as a%s interface with a%s link.\n",
> bond_dev->name, slave_dev->name,
> @@ -1686,7 +1678,6 @@ err_free:
> kfree(new_slave);
>
> err_undo_flags:
> - bond_compute_features(bond);
> /* Enslave of first slave has failed and we need to fix master's mac */
> if (!bond_has_slaves(bond) &&
> ether_addr_equal(bond_dev->dev_addr, slave_dev->dev_addr))
>
^ permalink raw reply
* [PATCH 00/15] net: ethernet: remove unnecessary pci_set_drvdata() part 2
From: Jingoo Han @ 2013-10-21 2:08 UTC (permalink / raw)
To: 'David S. Miller'; +Cc: netdev, 'Jingoo Han'
Since commit 0998d0631001288a5974afc0b2a5f568bcdecb4d
(device-core: Ensure drvdata = NULL when no driver is bound),
the driver core clears the driver data to NULL after device_release
or on probe failure. Thus, it is not needed to manually clear the
device driver data to NULL.
---
drivers/net/ethernet/dec/tulip/de2104x.c | 1 -
drivers/net/ethernet/dec/tulip/dmfe.c | 3 ---
drivers/net/ethernet/dec/tulip/tulip_core.c | 1 -
drivers/net/ethernet/dec/tulip/uli526x.c | 2 --
drivers/net/ethernet/dec/tulip/winbond-840.c | 3 ---
drivers/net/ethernet/dec/tulip/xircom_cb.c | 2 --
drivers/net/ethernet/dlink/dl2k.c | 1 -
drivers/net/ethernet/dlink/sundance.c | 2 --
drivers/net/ethernet/emulex/benet/be_main.c | 2 --
drivers/net/ethernet/fealnx.c | 1 -
drivers/net/ethernet/icplus/ipg.c | 1 -
drivers/net/ethernet/intel/e100.c | 2 --
drivers/net/ethernet/jme.c | 2 --
drivers/net/ethernet/marvell/skge.c | 2 --
drivers/net/ethernet/marvell/sky2.c | 3 ---
drivers/net/ethernet/micrel/ksz884x.c | 2 --
drivers/net/ethernet/myricom/myri10ge/myri10ge.c | 1 -
drivers/net/ethernet/natsemi/natsemi.c | 2 --
drivers/net/ethernet/neterion/s2io.c | 2 --
drivers/net/ethernet/neterion/vxge/vxge-main.c | 2 --
drivers/net/ethernet/packetengines/hamachi.c | 1 -
drivers/net/ethernet/packetengines/yellowfin.c | 2 --
22 files changed, 40 deletions(-)
^ permalink raw reply
* [PATCH 01/15] net: tulip: remove unnecessary pci_set_drvdata()
From: Jingoo Han @ 2013-10-21 2:09 UTC (permalink / raw)
To: 'David S. Miller'; +Cc: netdev, 'Jingoo Han'
In-Reply-To: <003801cece02$6abb0160$40310420$%han@samsung.com>
The driver core clears the driver data to NULL after device_release
or on probe failure. Thus, it is not needed to manually clear the
device driver data to NULL.
Signed-off-by: Jingoo Han <jg1.han@samsung.com>
---
drivers/net/ethernet/dec/tulip/de2104x.c | 1 -
drivers/net/ethernet/dec/tulip/dmfe.c | 3 ---
drivers/net/ethernet/dec/tulip/tulip_core.c | 1 -
drivers/net/ethernet/dec/tulip/uli526x.c | 2 --
drivers/net/ethernet/dec/tulip/winbond-840.c | 3 ---
drivers/net/ethernet/dec/tulip/xircom_cb.c | 2 --
6 files changed, 12 deletions(-)
diff --git a/drivers/net/ethernet/dec/tulip/de2104x.c b/drivers/net/ethernet/dec/tulip/de2104x.c
index eaab73c..38148b0 100644
--- a/drivers/net/ethernet/dec/tulip/de2104x.c
+++ b/drivers/net/ethernet/dec/tulip/de2104x.c
@@ -2110,7 +2110,6 @@ static void de_remove_one(struct pci_dev *pdev)
iounmap(de->regs);
pci_release_regions(pdev);
pci_disable_device(pdev);
- pci_set_drvdata(pdev, NULL);
free_netdev(dev);
}
diff --git a/drivers/net/ethernet/dec/tulip/dmfe.c b/drivers/net/ethernet/dec/tulip/dmfe.c
index 8313930..5ad9e3e 100644
--- a/drivers/net/ethernet/dec/tulip/dmfe.c
+++ b/drivers/net/ethernet/dec/tulip/dmfe.c
@@ -523,7 +523,6 @@ err_out_res:
err_out_disable:
pci_disable_device(pdev);
err_out_free:
- pci_set_drvdata(pdev, NULL);
free_netdev(dev);
return err;
@@ -548,8 +547,6 @@ static void dmfe_remove_one(struct pci_dev *pdev)
db->buf_pool_ptr, db->buf_pool_dma_ptr);
pci_release_regions(pdev);
free_netdev(dev); /* free board information */
-
- pci_set_drvdata(pdev, NULL);
}
DMFE_DBUG(0, "dmfe_remove_one() exit", 0);
diff --git a/drivers/net/ethernet/dec/tulip/tulip_core.c b/drivers/net/ethernet/dec/tulip/tulip_core.c
index 4e8cfa2..add05f1 100644
--- a/drivers/net/ethernet/dec/tulip/tulip_core.c
+++ b/drivers/net/ethernet/dec/tulip/tulip_core.c
@@ -1939,7 +1939,6 @@ static void tulip_remove_one(struct pci_dev *pdev)
pci_iounmap(pdev, tp->base_addr);
free_netdev (dev);
pci_release_regions (pdev);
- pci_set_drvdata (pdev, NULL);
/* pci_power_off (pdev, -1); */
}
diff --git a/drivers/net/ethernet/dec/tulip/uli526x.c b/drivers/net/ethernet/dec/tulip/uli526x.c
index 93845af..a5397b1 100644
--- a/drivers/net/ethernet/dec/tulip/uli526x.c
+++ b/drivers/net/ethernet/dec/tulip/uli526x.c
@@ -429,7 +429,6 @@ err_out_release:
err_out_disable:
pci_disable_device(pdev);
err_out_free:
- pci_set_drvdata(pdev, NULL);
free_netdev(dev);
return err;
@@ -450,7 +449,6 @@ static void uli526x_remove_one(struct pci_dev *pdev)
db->buf_pool_ptr, db->buf_pool_dma_ptr);
pci_release_regions(pdev);
pci_disable_device(pdev);
- pci_set_drvdata(pdev, NULL);
free_netdev(dev);
}
diff --git a/drivers/net/ethernet/dec/tulip/winbond-840.c b/drivers/net/ethernet/dec/tulip/winbond-840.c
index c7b04ec..62fe512 100644
--- a/drivers/net/ethernet/dec/tulip/winbond-840.c
+++ b/drivers/net/ethernet/dec/tulip/winbond-840.c
@@ -468,7 +468,6 @@ static int w840_probe1(struct pci_dev *pdev, const struct pci_device_id *ent)
return 0;
err_out_cleardev:
- pci_set_drvdata(pdev, NULL);
pci_iounmap(pdev, ioaddr);
err_out_free_res:
pci_release_regions(pdev);
@@ -1542,8 +1541,6 @@ static void w840_remove1(struct pci_dev *pdev)
pci_iounmap(pdev, np->base_addr);
free_netdev(dev);
}
-
- pci_set_drvdata(pdev, NULL);
}
#ifdef CONFIG_PM
diff --git a/drivers/net/ethernet/dec/tulip/xircom_cb.c b/drivers/net/ethernet/dec/tulip/xircom_cb.c
index 9b84cb0..ab7ebac 100644
--- a/drivers/net/ethernet/dec/tulip/xircom_cb.c
+++ b/drivers/net/ethernet/dec/tulip/xircom_cb.c
@@ -289,7 +289,6 @@ out:
err_unmap:
pci_iounmap(pdev, private->ioaddr);
reg_fail:
- pci_set_drvdata(pdev, NULL);
dma_free_coherent(d, 8192, private->tx_buffer, private->tx_dma_handle);
tx_buf_fail:
dma_free_coherent(d, 8192, private->rx_buffer, private->rx_dma_handle);
@@ -317,7 +316,6 @@ static void xircom_remove(struct pci_dev *pdev)
unregister_netdev(dev);
pci_iounmap(pdev, card->ioaddr);
- pci_set_drvdata(pdev, NULL);
dma_free_coherent(d, 8192, card->tx_buffer, card->tx_dma_handle);
dma_free_coherent(d, 8192, card->rx_buffer, card->rx_dma_handle);
free_netdev(dev);
--
1.7.10.4
^ permalink raw reply related
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox