* Re: [PATCH 2.6.31-rc9] net: VMware virtual Ethernet NIC driver: vmxnet3
From: Greg KH @ 2009-09-29 0:20 UTC (permalink / raw)
To: Shreyas Bhatewara
Cc: linux-kernel@vger.kernel.org, netdev@vger.kernel.org,
Stephen Hemminger, David S. Miller, Jeff Garzik, Anthony Liguori,
Chris Wright, Andrew Morton, virtualization,
pv-drivers@vmware.com
In-Reply-To: <89E2752CFA8EC044846EB849981913410173CDFAF6@EXCH-MBX-4.vmware.com>
On Mon, Sep 28, 2009 at 04:56:45PM -0700, Shreyas Bhatewara wrote:
> Ethernet NIC driver for VMware's vmxnet3
>
> From: Shreyas Bhatewara <sbhatewara@vmware.com>
>
> This patch adds driver support for VMware's virtual Ethernet NIC : vmxnet3
> Guests running on VMware hypervisors supporting vmxnet3 device will thus
> have access to improved network functionalities and performance.
>
> Signed-off-by: Shreyas Bhatewara <sbhatewara@vmware.com>
I thought this was going to be submitted for the drivers/staging/ tree.
What happened?
thanks,
greg k-h
^ permalink raw reply
* Re: [Bonding-devel] [PATCH 4/4] bonding: add sysfs files to display tlb and alb hash table contents
From: Andy Gospodarek @ 2009-09-29 0:12 UTC (permalink / raw)
To: Stephen Hemminger; +Cc: Andy Gospodarek, netdev, fubar, bonding-devel
In-Reply-To: <20090928162237.6db0c9f5@nehalam>
On Mon, Sep 28, 2009 at 04:22:37PM -0700, Stephen Hemminger wrote:
> On Fri, 11 Sep 2009 17:13:17 -0400
> Andy Gospodarek <andy@greyhouse.net> wrote:
>
> >
> > bonding: add sysfs files to display tlb and alb hash table contents
> >
> > While debugging some problems with alb (mode 6) bonding I realized that
> > being able to output the contents of both hash tables would be helpful.
> > This is what the output looks like for the two files:
> >
> > device load
> > eth1 491
> > eth2 491
> > hash device last device tx bytes load next previous
> > 2 eth1 eth1 2254 491 0 0
> > 3 eth2 eth2 2744 491 0 0
> > 6 eth2 0 488 0 0
> > 8 eth2 0 461698 0 0
> > 1b eth2 0 249 0 0
> > eb eth2 0 21 0 0
> > ff eth2 0 22 0 0
> >
> > hash ip_src ip_dst mac_dst slave assign ntt
> > 2 10.0.3.2 10.0.3.11 00:e0:81:71:ee:a9 eth1 1 0
> > 3 10.0.3.2 10.0.3.10 00:e0:81:71:ee:a9 eth2 1 0
> > 8 10.0.3.2 10.0.3.1 00:e0:81:71:ee:a9 eth2 1 0
> >
> > These were a great help debugging the fixes I have just posted and they
> > might be helpful for others, so I decided to include them in my
> > patchset.
> >
> > Signed-off-by: Andy Gospodarek <andy@greyhouse.net>
>
> No.
>
> Please don't put formatted output in sysfs, it is not meant to be
> used like proc, there is supposed to be only one value per file.
Then based on the over 300 files in /sys/ that are more than 1 line on
my currently running kernel, it seems there is significant work to do.
Seemingly arbitrary requests like this are extremely annoying when the
current kernel violates them all over the place.
^ permalink raw reply
* Re: [PATCH 2/4 v3] bonding: make sure tx and rx hash tables stay in sync when using alb mode
From: Andy Gospodarek @ 2009-09-29 0:13 UTC (permalink / raw)
To: Jay Vosburgh; +Cc: Andy Gospodarek, netdev, bonding-devel
In-Reply-To: <1915.1254175794@death.nxdomain.ibm.com>
On Mon, Sep 28, 2009 at 03:09:54PM -0700, Jay Vosburgh wrote:
> Andy Gospodarek <andy@greyhouse.net> wrote:
>
> >On Fri, Sep 18, 2009 at 11:56:45AM -0400, Andy Gospodarek wrote:
> >> On Fri, Sep 18, 2009 at 11:36:22AM -0400, Andy Gospodarek wrote:
> >> > On Wed, Sep 16, 2009 at 04:36:09PM -0700, Jay Vosburgh wrote:
> >> > > Andy Gospodarek <andy@greyhouse.net> wrote:
> >> > >
> >> > > >
> >> > > >Subject: [PATCH] bonding: make sure tx and rx hash tables stay in sync when using alb mode
> >> > >
> >> > > When testing this, I'm getting a lockdep warning. It appears to
> >> > > be unhappy that tlb_choose_channel acquires the tx / rx hash table locks
> >> > > in the order tx then rx, but rlb_choose_channel -> alb_get_best_slave
> >> > > acquires the locks in the other order. I applied all four patches, but
> >> > > it looks like the change that trips lockdep is in this patch (#2).
> >> > >
> >> > > I haven't gotten an actual deadlock from this, although it seems
> >> > > plausible if there are two cpus in bond_alb_xmit at the same time, and
> >> > > one of them is sending an ARP.
> >> > >
> >> > > One fairly straightforward fix would be to combine the rx and tx
> >> > > hash table locks into a single lock. I suspect that wouldn't have any
> >> > > real performance penalty, since the rx hash table lock is generally not
> >> > > acquired very often (unlike the tx lock, which is taken for every packet
> >> > > that goes out).
> >> > >
> >> > > Also, FYI, two of the four patches had trailing whitespace. I
> >> > > believe it was #2 and #4.
> >> > >
> >> > > Thoughts?
> >> >
> >> > Jay,
> >> >
> >> > This patch should address both the the deadlock and whitespace conerns.
> >> > I ran a kernel with LOCKDEP enabled and saw no warnings while passing
> >> > traffic on the bond while pulling cables and while removing the module.
> >> > Here it is....
> >> >
> >>
> >> Adding the version and signed-off-by lines might be nice, eh?
> >>
> >> [PATCH v3] bonding: make sure tx and rx hash tables stay in sync when using alb mode
> >>
> >> I noticed that it was easy for alb (mode 6) bonding to get into a state
> >> where the tx hash-table and rx hash-table are out of sync (there is
> >> really nothing to keep them synchronized), and we will transmit traffic
> >> destined for a host on one slave and send ARP frames to the same slave
> >> from another interface using a different source MAC.
> >>
> >> There is no compelling reason to do this, so this patch makes sure the
> >> rx hash-table changes whenever the tx hash-table is updated based on
> >> device load. This patch also drops the code that does rlb re-balancing
> >> since the balancing will not be controlled by the tx hash-table based on
>
> In addition to my response in the other thread, I changed the
> "not" above to "now," which I suspect is what you meant.
>
You are correct. Thanks for catching that!
> >> transmit load. In order to address an issue found with the initial
> >> patch, I have also combined the rx and tx hash table lock into a single
> >> lock. This will facilitate moving these into a single table at some
> >> point.
> >>
> >> Signed-off-by: Andy Gospodarek <andy@greyhouse.net>
> >>
> >> ---
> >> drivers/net/bonding/bond_alb.c | 203 +++++++++++++++-------------------------
> >> drivers/net/bonding/bond_alb.h | 3 +-
> >> 2 files changed, 75 insertions(+), 131 deletions(-)
> >>
> >> diff --git a/drivers/net/bonding/bond_alb.c b/drivers/net/bonding/bond_alb.c
> >> index bcf25c6..04b7055 100644
> >> --- a/drivers/net/bonding/bond_alb.c
> >> +++ b/drivers/net/bonding/bond_alb.c
> >> @@ -111,6 +111,7 @@ static inline struct arp_pkt *arp_pkt(const struct sk_buff *skb)
> >>
> >> /* Forward declaration */
> >> static void alb_send_learning_packets(struct slave *slave, u8 mac_addr[]);
> >> +static struct slave *alb_get_best_slave(struct bonding *bond, u32 hash_index);
> >>
> >> static inline u8 _simple_hash(const u8 *hash_start, int hash_size)
> >> {
> >> @@ -124,18 +125,18 @@ static inline u8 _simple_hash(const u8 *hash_start, int hash_size)
> >> return hash;
> >> }
> >>
> >> -/*********************** tlb specific functions ***************************/
> >> -
> >> -static inline void _lock_tx_hashtbl(struct bonding *bond)
> >> +/********************* hash table lock functions *************************/
> >> +static inline void _lock_hashtbl(struct bonding *bond)
> >> {
> >> - spin_lock_bh(&(BOND_ALB_INFO(bond).tx_hashtbl_lock));
> >> + spin_lock_bh(&(BOND_ALB_INFO(bond).hashtbl_lock));
> >> }
> >>
> >> -static inline void _unlock_tx_hashtbl(struct bonding *bond)
> >> +static inline void _unlock_hashtbl(struct bonding *bond)
> >> {
> >> - spin_unlock_bh(&(BOND_ALB_INFO(bond).tx_hashtbl_lock));
> >> + spin_unlock_bh(&(BOND_ALB_INFO(bond).hashtbl_lock));
> >> }
> >>
> >> +/*********************** tlb specific functions ***************************/
> >> /* Caller must hold tx_hashtbl lock */
> >> static inline void tlb_init_table_entry(struct tlb_client_info *entry, int save_load)
> >> {
> >> @@ -163,7 +164,7 @@ static void tlb_clear_slave(struct bonding *bond, struct slave *slave, int save_
> >> struct tlb_client_info *tx_hash_table;
> >> u32 index;
> >>
> >> - _lock_tx_hashtbl(bond);
> >> + _lock_hashtbl(bond);
> >>
> >> /* clear slave from tx_hashtbl */
> >> tx_hash_table = BOND_ALB_INFO(bond).tx_hashtbl;
> >> @@ -180,7 +181,7 @@ static void tlb_clear_slave(struct bonding *bond, struct slave *slave, int save_
> >>
> >> tlb_init_slave(slave);
> >>
> >> - _unlock_tx_hashtbl(bond);
> >> + _unlock_hashtbl(bond);
> >> }
> >>
> >> /* Must be called before starting the monitor timer */
> >> @@ -191,7 +192,7 @@ static int tlb_initialize(struct bonding *bond)
> >> struct tlb_client_info *new_hashtbl;
> >> int i;
> >>
> >> - spin_lock_init(&(bond_info->tx_hashtbl_lock));
> >> + spin_lock_init(&(bond_info->hashtbl_lock));
> >>
> >> new_hashtbl = kzalloc(size, GFP_KERNEL);
> >> if (!new_hashtbl) {
> >> @@ -200,7 +201,7 @@ static int tlb_initialize(struct bonding *bond)
> >> bond->dev->name);
> >> return -1;
> >> }
> >> - _lock_tx_hashtbl(bond);
> >> + _lock_hashtbl(bond);
> >>
> >> bond_info->tx_hashtbl = new_hashtbl;
> >>
> >> @@ -208,7 +209,7 @@ static int tlb_initialize(struct bonding *bond)
> >> tlb_init_table_entry(&bond_info->tx_hashtbl[i], 1);
> >> }
> >>
> >> - _unlock_tx_hashtbl(bond);
> >> + _unlock_hashtbl(bond);
> >>
> >> return 0;
> >> }
> >> @@ -218,12 +219,12 @@ static void tlb_deinitialize(struct bonding *bond)
> >> {
> >> struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
> >>
> >> - _lock_tx_hashtbl(bond);
> >> + _lock_hashtbl(bond);
> >>
> >> kfree(bond_info->tx_hashtbl);
> >> bond_info->tx_hashtbl = NULL;
> >>
> >> - _unlock_tx_hashtbl(bond);
> >> + _unlock_hashtbl(bond);
> >> }
> >>
> >> /* Caller must hold bond lock for read */
> >> @@ -264,24 +265,6 @@ static struct slave *tlb_get_least_loaded_slave(struct bonding *bond)
> >> return least_loaded;
> >> }
> >>
> >> -/* Caller must hold bond lock for read and hashtbl lock */
> >> -static struct slave *tlb_get_best_slave(struct bonding *bond, u32 hash_index)
> >> -{
> >> - struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
> >> - struct tlb_client_info *tx_hash_table = bond_info->tx_hashtbl;
> >> - struct slave *last_slave = tx_hash_table[hash_index].last_slave;
> >> - struct slave *next_slave = NULL;
> >> -
> >> - if (last_slave && SLAVE_IS_OK(last_slave)) {
> >> - /* Use the last slave listed in the tx hashtbl if:
> >> - the last slave currently is essentially unloaded. */
> >> - if (SLAVE_TLB_INFO(last_slave).load < 10)
> >> - next_slave = last_slave;
> >> - }
> >> -
> >> - return next_slave ? next_slave : tlb_get_least_loaded_slave(bond);
> >> -}
> >> -
> >> /* Caller must hold bond lock for read */
> >> static struct slave *tlb_choose_channel(struct bonding *bond, u32 hash_index, u32 skb_len)
> >> {
> >> @@ -289,13 +272,12 @@ static struct slave *tlb_choose_channel(struct bonding *bond, u32 hash_index, u3
> >> struct tlb_client_info *hash_table;
> >> struct slave *assigned_slave;
> >>
> >> - _lock_tx_hashtbl(bond);
> >> + _lock_hashtbl(bond);
> >>
> >> hash_table = bond_info->tx_hashtbl;
> >> assigned_slave = hash_table[hash_index].tx_slave;
> >> if (!assigned_slave) {
> >> - assigned_slave = tlb_get_best_slave(bond, hash_index);
> >> -
> >> + assigned_slave = alb_get_best_slave(bond, hash_index);
> >> if (assigned_slave) {
> >> struct tlb_slave_info *slave_info =
> >> &(SLAVE_TLB_INFO(assigned_slave));
> >> @@ -319,20 +301,52 @@ static struct slave *tlb_choose_channel(struct bonding *bond, u32 hash_index, u3
> >> hash_table[hash_index].tx_bytes += skb_len;
> >> }
> >>
> >> - _unlock_tx_hashtbl(bond);
> >> + _unlock_hashtbl(bond);
> >>
> >> return assigned_slave;
> >> }
> >>
> >> /*********************** rlb specific functions ***************************/
> >> -static inline void _lock_rx_hashtbl(struct bonding *bond)
> >> +
> >> +/* Caller must hold bond lock for read and hashtbl lock */
> >> +static struct slave *rlb_update_rx_table(struct bonding *bond, struct slave *next_slave, u32 hash_index)
> >> {
> >> - spin_lock_bh(&(BOND_ALB_INFO(bond).rx_hashtbl_lock));
> >> + struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
> >> +
> >> + /* check rlb table and correct it if wrong */
> >> + if (bond_info->rlb_enabled) {
> >> + struct rlb_client_info *rx_client_info = &(bond_info->rx_hashtbl[hash_index]);
> >> +
> >> + /* if the new slave computed by tlb checks doesn't match rlb, stop rlb from using it */
> >> + if (next_slave && (next_slave != rx_client_info->slave))
> >> + rx_client_info->slave = next_slave;
> >> + }
> >> + return next_slave;
> >> }
> >>
> >> -static inline void _unlock_rx_hashtbl(struct bonding *bond)
> >> +/* Caller must hold bond lock for read and hashtbl lock */
> >> +static struct slave *alb_get_best_slave(struct bonding *bond, u32 hash_index)
> >> {
> >> - spin_unlock_bh(&(BOND_ALB_INFO(bond).rx_hashtbl_lock));
> >> + struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
> >> + struct tlb_client_info *tx_hash_table = bond_info->tx_hashtbl;
> >> + struct slave *last_slave = tx_hash_table[hash_index].last_slave;
> >> + struct slave *next_slave = NULL;
> >> +
> >> + /* presume the next slave will be the least loaded one */
> >> + next_slave = tlb_get_least_loaded_slave(bond);
> >> +
> >> + if (last_slave && SLAVE_IS_OK(last_slave)) {
> >> + /* Use the last slave listed in the tx hashtbl if:
> >> + the last slave currently is essentially unloaded. */
> >> + if (SLAVE_TLB_INFO(last_slave).load < 10)
> >> + next_slave = last_slave;
> >> + }
> >> +
> >> + /* update the rlb hashtbl if there was a previous entry */
> >> + if (bond_info->rlb_enabled)
> >> + rlb_update_rx_table(bond, next_slave, hash_index);
> >> +
> >> + return next_slave;
> >> }
> >>
> >> /* when an ARP REPLY is received from a client update its info
> >> @@ -344,7 +358,7 @@ static void rlb_update_entry_from_arp(struct bonding *bond, struct arp_pkt *arp)
> >> struct rlb_client_info *client_info;
> >> u32 hash_index;
> >>
> >> - _lock_rx_hashtbl(bond);
> >> + _lock_hashtbl(bond);
> >>
> >> hash_index = _simple_hash((u8*)&(arp->ip_src), sizeof(arp->ip_src));
> >> client_info = &(bond_info->rx_hashtbl[hash_index]);
> >> @@ -358,7 +372,7 @@ static void rlb_update_entry_from_arp(struct bonding *bond, struct arp_pkt *arp)
> >> bond_info->rx_ntt = 1;
> >> }
> >>
> >> - _unlock_rx_hashtbl(bond);
> >> + _unlock_hashtbl(bond);
> >> }
> >>
> >> static int rlb_arp_recv(struct sk_buff *skb, struct net_device *bond_dev, struct packet_type *ptype, struct net_device *orig_dev)
> >> @@ -402,38 +416,6 @@ out:
> >> return res;
> >> }
> >>
> >> -/* Caller must hold bond lock for read */
> >> -static struct slave *rlb_next_rx_slave(struct bonding *bond)
> >> -{
> >> - struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
> >> - struct slave *rx_slave, *slave, *start_at;
> >> - int i = 0;
> >> -
> >> - if (bond_info->next_rx_slave) {
> >> - start_at = bond_info->next_rx_slave;
> >> - } else {
> >> - start_at = bond->first_slave;
> >> - }
> >> -
> >> - rx_slave = NULL;
> >> -
> >> - bond_for_each_slave_from(bond, slave, i, start_at) {
> >> - if (SLAVE_IS_OK(slave)) {
> >> - if (!rx_slave) {
> >> - rx_slave = slave;
> >> - } else if (slave->speed > rx_slave->speed) {
> >> - rx_slave = slave;
> >> - }
> >> - }
> >> - }
> >> -
> >> - if (rx_slave) {
> >> - bond_info->next_rx_slave = rx_slave->next;
> >> - }
> >> -
> >> - return rx_slave;
> >> -}
> >> -
> >> /* teach the switch the mac of a disabled slave
> >> * on the primary for fault tolerance
> >> *
> >> @@ -468,14 +450,14 @@ static void rlb_clear_slave(struct bonding *bond, struct slave *slave)
> >> u32 index, next_index;
> >>
> >> /* clear slave from rx_hashtbl */
> >> - _lock_rx_hashtbl(bond);
> >> + _lock_hashtbl(bond);
> >>
> >> rx_hash_table = bond_info->rx_hashtbl;
> >> index = bond_info->rx_hashtbl_head;
> >> for (; index != RLB_NULL_INDEX; index = next_index) {
> >> next_index = rx_hash_table[index].next;
> >> if (rx_hash_table[index].slave == slave) {
> >> - struct slave *assigned_slave = rlb_next_rx_slave(bond);
> >> + struct slave *assigned_slave = alb_get_best_slave(bond, index);
> >>
> >> if (assigned_slave) {
> >> rx_hash_table[index].slave = assigned_slave;
> >> @@ -499,7 +481,7 @@ static void rlb_clear_slave(struct bonding *bond, struct slave *slave)
> >> }
> >> }
> >>
> >> - _unlock_rx_hashtbl(bond);
> >> + _unlock_hashtbl(bond);
> >>
> >> write_lock_bh(&bond->curr_slave_lock);
> >>
> >> @@ -558,7 +540,7 @@ static void rlb_update_rx_clients(struct bonding *bond)
> >> struct rlb_client_info *client_info;
> >> u32 hash_index;
> >>
> >> - _lock_rx_hashtbl(bond);
> >> + _lock_hashtbl(bond);
> >>
> >> hash_index = bond_info->rx_hashtbl_head;
> >> for (; hash_index != RLB_NULL_INDEX; hash_index = client_info->next) {
> >> @@ -576,7 +558,7 @@ static void rlb_update_rx_clients(struct bonding *bond)
> >> */
> >> bond_info->rlb_update_delay_counter = RLB_UPDATE_DELAY;
> >>
> >> - _unlock_rx_hashtbl(bond);
> >> + _unlock_hashtbl(bond);
> >> }
> >>
> >> /* The slave was assigned a new mac address - update the clients */
> >> @@ -587,7 +569,7 @@ static void rlb_req_update_slave_clients(struct bonding *bond, struct slave *sla
> >> int ntt = 0;
> >> u32 hash_index;
> >>
> >> - _lock_rx_hashtbl(bond);
> >> + _lock_hashtbl(bond);
> >>
> >> hash_index = bond_info->rx_hashtbl_head;
> >> for (; hash_index != RLB_NULL_INDEX; hash_index = client_info->next) {
> >> @@ -607,7 +589,7 @@ static void rlb_req_update_slave_clients(struct bonding *bond, struct slave *sla
> >> bond_info->rlb_update_retry_counter = RLB_UPDATE_RETRY;
> >> }
> >>
> >> - _unlock_rx_hashtbl(bond);
> >> + _unlock_hashtbl(bond);
> >> }
> >>
> >> /* mark all clients using src_ip to be updated */
> >> @@ -617,7 +599,7 @@ static void rlb_req_update_subnet_clients(struct bonding *bond, __be32 src_ip)
> >> struct rlb_client_info *client_info;
> >> u32 hash_index;
> >>
> >> - _lock_rx_hashtbl(bond);
> >> + _lock_hashtbl(bond);
> >>
> >> hash_index = bond_info->rx_hashtbl_head;
> >> for (; hash_index != RLB_NULL_INDEX; hash_index = client_info->next) {
> >> @@ -643,7 +625,7 @@ static void rlb_req_update_subnet_clients(struct bonding *bond, __be32 src_ip)
> >> }
> >> }
> >>
> >> - _unlock_rx_hashtbl(bond);
> >> + _unlock_hashtbl(bond);
> >> }
> >>
> >> /* Caller must hold both bond and ptr locks for read */
> >> @@ -655,7 +637,7 @@ static struct slave *rlb_choose_channel(struct sk_buff *skb, struct bonding *bon
> >> struct rlb_client_info *client_info;
> >> u32 hash_index = 0;
> >>
> >> - _lock_rx_hashtbl(bond);
> >> + _lock_hashtbl(bond);
> >>
> >> hash_index = _simple_hash((u8 *)&arp->ip_dst, sizeof(arp->ip_src));
> >> client_info = &(bond_info->rx_hashtbl[hash_index]);
> >> @@ -671,7 +653,7 @@ static struct slave *rlb_choose_channel(struct sk_buff *skb, struct bonding *bon
> >>
> >> assigned_slave = client_info->slave;
> >> if (assigned_slave) {
> >> - _unlock_rx_hashtbl(bond);
> >> + _unlock_hashtbl(bond);
> >> return assigned_slave;
> >> }
> >> } else {
> >> @@ -687,7 +669,7 @@ static struct slave *rlb_choose_channel(struct sk_buff *skb, struct bonding *bon
> >> }
> >> }
> >> /* assign a new slave */
> >> - assigned_slave = rlb_next_rx_slave(bond);
> >> + assigned_slave = alb_get_best_slave(bond, hash_index);
> >>
> >> if (assigned_slave) {
> >> client_info->ip_src = arp->ip_src;
> >> @@ -723,7 +705,7 @@ static struct slave *rlb_choose_channel(struct sk_buff *skb, struct bonding *bon
> >> }
> >> }
> >>
> >> - _unlock_rx_hashtbl(bond);
> >> + _unlock_hashtbl(bond);
> >>
> >> return assigned_slave;
> >> }
> >> @@ -771,36 +753,6 @@ static struct slave *rlb_arp_xmit(struct sk_buff *skb, struct bonding *bond)
> >> return tx_slave;
> >> }
> >>
> >> -/* Caller must hold bond lock for read */
> >> -static void rlb_rebalance(struct bonding *bond)
> >> -{
> >> - struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
> >> - struct slave *assigned_slave;
> >> - struct rlb_client_info *client_info;
> >> - int ntt;
> >> - u32 hash_index;
> >> -
> >> - _lock_rx_hashtbl(bond);
> >> -
> >> - ntt = 0;
> >> - hash_index = bond_info->rx_hashtbl_head;
> >> - for (; hash_index != RLB_NULL_INDEX; hash_index = client_info->next) {
> >> - client_info = &(bond_info->rx_hashtbl[hash_index]);
> >> - assigned_slave = rlb_next_rx_slave(bond);
> >> - if (assigned_slave && (client_info->slave != assigned_slave)) {
> >> - client_info->slave = assigned_slave;
> >> - client_info->ntt = 1;
> >> - ntt = 1;
> >> - }
> >> - }
> >> -
> >> - /* update the team's flag only after the whole iteration */
> >> - if (ntt) {
> >> - bond_info->rx_ntt = 1;
> >> - }
> >> - _unlock_rx_hashtbl(bond);
> >> -}
> >> -
> >> /* Caller must hold rx_hashtbl lock */
> >> static void rlb_init_table_entry(struct rlb_client_info *entry)
> >> {
> >> @@ -817,8 +769,6 @@ static int rlb_initialize(struct bonding *bond)
> >> int size = RLB_HASH_TABLE_SIZE * sizeof(struct rlb_client_info);
> >> int i;
> >>
> >> - spin_lock_init(&(bond_info->rx_hashtbl_lock));
> >> -
> >> new_hashtbl = kmalloc(size, GFP_KERNEL);
> >> if (!new_hashtbl) {
> >> printk(KERN_ERR DRV_NAME
> >> @@ -826,7 +776,7 @@ static int rlb_initialize(struct bonding *bond)
> >> bond->dev->name);
> >> return -1;
> >> }
> >> - _lock_rx_hashtbl(bond);
> >> + _lock_hashtbl(bond);
> >>
> >> bond_info->rx_hashtbl = new_hashtbl;
> >>
> >> @@ -836,7 +786,7 @@ static int rlb_initialize(struct bonding *bond)
> >> rlb_init_table_entry(bond_info->rx_hashtbl + i);
> >> }
> >>
> >> - _unlock_rx_hashtbl(bond);
> >> + _unlock_hashtbl(bond);
> >>
> >> /*initialize packet type*/
> >> pk_type->type = cpu_to_be16(ETH_P_ARP);
> >> @@ -855,13 +805,13 @@ static void rlb_deinitialize(struct bonding *bond)
> >>
> >> dev_remove_pack(&(bond_info->rlb_pkt_type));
> >>
> >> - _lock_rx_hashtbl(bond);
> >> + _lock_hashtbl(bond);
> >>
> >> kfree(bond_info->rx_hashtbl);
> >> bond_info->rx_hashtbl = NULL;
> >> bond_info->rx_hashtbl_head = RLB_NULL_INDEX;
> >>
> >> - _unlock_rx_hashtbl(bond);
> >> + _unlock_hashtbl(bond);
> >> }
> >>
> >> static void rlb_clear_vlan(struct bonding *bond, unsigned short vlan_id)
> >> @@ -869,7 +819,7 @@ static void rlb_clear_vlan(struct bonding *bond, unsigned short vlan_id)
> >> struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
> >> u32 curr_index;
> >>
> >> - _lock_rx_hashtbl(bond);
> >> + _lock_hashtbl(bond);
> >>
> >> curr_index = bond_info->rx_hashtbl_head;
> >> while (curr_index != RLB_NULL_INDEX) {
> >> @@ -894,7 +844,7 @@ static void rlb_clear_vlan(struct bonding *bond, unsigned short vlan_id)
> >> curr_index = next_index;
> >> }
> >>
> >> - _unlock_rx_hashtbl(bond);
> >> + _unlock_hashtbl(bond);
> >> }
> >>
> >> /*********************** tlb/rlb shared functions *********************/
> >> @@ -1521,11 +1471,6 @@ void bond_alb_monitor(struct work_struct *work)
> >> read_lock(&bond->lock);
> >> }
> >>
> >> - if (bond_info->rlb_rebalance) {
> >> - bond_info->rlb_rebalance = 0;
> >> - rlb_rebalance(bond);
> >> - }
> >> -
> >> /* check if clients need updating */
> >> if (bond_info->rx_ntt) {
> >> if (bond_info->rlb_update_delay_counter) {
> >> diff --git a/drivers/net/bonding/bond_alb.h b/drivers/net/bonding/bond_alb.h
> >> index b65fd29..09d755a 100644
> >> --- a/drivers/net/bonding/bond_alb.h
> >> +++ b/drivers/net/bonding/bond_alb.h
> >> @@ -90,7 +90,7 @@ struct tlb_slave_info {
> >> struct alb_bond_info {
> >> struct timer_list alb_timer;
> >> struct tlb_client_info *tx_hashtbl; /* Dynamically allocated */
> >> - spinlock_t tx_hashtbl_lock;
> >> + spinlock_t hashtbl_lock; /* lock for both tables */
> >> u32 unbalanced_load;
> >> int tx_rebalance_counter;
> >> int lp_counter;
> >> @@ -98,7 +98,6 @@ struct alb_bond_info {
> >> int rlb_enabled;
> >> struct packet_type rlb_pkt_type;
> >> struct rlb_client_info *rx_hashtbl; /* Receive hash table */
> >> - spinlock_t rx_hashtbl_lock;
> >> u32 rx_hashtbl_head;
> >> u8 rx_ntt; /* flag - need to transmit
> >> * to all rx clients
> >
> >Any thoughts on this, Jay?
>
> -J
>
> ---
> -Jay Vosburgh, IBM Linux Technology Center, fubar@us.ibm.com
> --
> To unsubscribe from this list: send the line "unsubscribe netdev" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply
* Re: [PATCH 2.6.31-rc9] net: VMware virtual Ethernet NIC driver: vmxnet3
From: David Miller @ 2009-09-29 0:08 UTC (permalink / raw)
To: sbhatewara
Cc: linux-kernel, netdev, shemminger, jgarzik, anthony, chrisw, greg,
akpm, virtualization, pv-drivers
In-Reply-To: <89E2752CFA8EC044846EB849981913410173CDFAF6@EXCH-MBX-4.vmware.com>
From: Shreyas Bhatewara <sbhatewara@vmware.com>
Date: Mon, 28 Sep 2009 16:56:45 -0700
> + uint32_t rxdIdx:12; /* Index of the RxDesc */
Don't use uint32_t et al. sized types, use "u32" and friends
throughout.
^ permalink raw reply
* [PATCH 2.6.31-rc9] net: VMware virtual Ethernet NIC driver: vmxnet3
From: Shreyas Bhatewara @ 2009-09-28 23:56 UTC (permalink / raw)
To: linux-kernel@vger.kernel.org, netdev@vger.kernel.org,
Stephen Hemminger, "David S. Miller
Cc: Anthony Liguori, Chris Wright, Greg Kroah-Hartman, Andrew Morton,
virtualization, pv-drivers@vmware.com
Ethernet NIC driver for VMware's vmxnet3
From: Shreyas Bhatewara <sbhatewara@vmware.com>
This patch adds driver support for VMware's virtual Ethernet NIC : vmxnet3
Guests running on VMware hypervisors supporting vmxnet3 device will thus
have access to improved network functionalities and performance.
Signed-off-by: Shreyas Bhatewara <sbhatewara@vmware.com>
---
Greetings.
The patch pasted below adds to Linux, driver support for VMware virtual Ethernet NIC : vmxnet3
About vmxnet3: VMware designed vmxnet3 a couple of years ago. It's being shipped with hypervisor in products since VMware Workstation 6.5 (9/2008) and ESX 4.0 (5/2009).
Some of the features of vmxnet3 are :
PCIe 2.0 compliant PCI device: Vendor ID 0x15ad, Device ID 0x07b0
INTx, MSI, MSI-X (25 vectors) interrupts
16 Rx queues, 8 Tx queues
Offloads: TCP/UDP checksum, TSO over IPv4/IPv6,
802.1q VLAN tag insertion, filtering, stripping
Multicast filtering, Jumbo Frames
Wake-on-LAN, PCI Power Management D0-D3 states
PXE-ROM for boot support
Please consider this for inclusion in the linux net tree. I will be glad to receive your review comments and answer queries in order to be accepted in mainline in 2.6.32 release cycle.
The patch applies to 2.6.31-rc9.
Thanking you.
Shreyas
---
diff --git a/MAINTAINERS b/MAINTAINERS
index 8dca9d8..c57a270 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -5490,6 +5490,12 @@ S: Maintained
F: drivers/vlynq/vlynq.c
F: include/linux/vlynq.h
+VMWARE VMXNET3 ETHERNET DRIVER
+M: Shreyas Bhatewara <pv-drivers@vmware.com>
+L: netdev@vger.kernel.org
+S: Maintained
+F: drivers/net/vmxnet3/
+
VOLTAGE AND CURRENT REGULATOR FRAMEWORK
M: Liam Girdwood <lrg@slimlogic.co.uk>
M: Mark Brown <broonie@opensource.wolfsonmicro.com>
diff --git a/drivers/net/Kconfig b/drivers/net/Kconfig
index 5ce7cba..703e0b6 100644
--- a/drivers/net/Kconfig
+++ b/drivers/net/Kconfig
@@ -3211,4 +3211,12 @@ config VIRTIO_NET
This is the virtual network driver for virtio. It can be used with
lguest or QEMU based VMMs (like KVM or Xen). Say Y or M.
+config VMXNET3
+ tristate "VMware VMXNET3 ethernet driver"
+ depends on PCI && X86
+ help
+ This driver supports VMware's vmxnet3 virtual ethernet NIC.
+ To compile this driver as a module, choose M here: the
+ module will be called vmxnet3.
+
endif # NETDEVICES
diff --git a/drivers/net/Makefile b/drivers/net/Makefile
index ead8cab..c146bc1 100644
--- a/drivers/net/Makefile
+++ b/drivers/net/Makefile
@@ -26,6 +26,7 @@ obj-$(CONFIG_TEHUTI) += tehuti.o
obj-$(CONFIG_ENIC) += enic/
obj-$(CONFIG_JME) += jme.o
obj-$(CONFIG_BE2NET) += benet/
+obj-$(CONFIG_VMXNET3) += vmxnet3/
gianfar_driver-objs := gianfar.o \
gianfar_ethtool.o \
diff --git a/drivers/net/vmxnet3/Makefile b/drivers/net/vmxnet3/Makefile
new file mode 100644
index 0000000..880f509
--- /dev/null
+++ b/drivers/net/vmxnet3/Makefile
@@ -0,0 +1,35 @@
+################################################################################
+#
+# Linux driver for VMware's vmxnet3 ethernet NIC.
+#
+# Copyright (C) 2007-2009, VMware, Inc. All Rights Reserved.
+#
+# This program is free software; you can redistribute it and/or modify it
+# under the terms of the GNU General Public License as published by the
+# Free Software Foundation; version 2 of the License and no later version.
+#
+# This program is distributed in the hope that it will be useful, but
+# WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or
+# NON INFRINGEMENT. See the GNU General Public License for more
+# details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, write to the Free Software
+# Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
+#
+# The full GNU General Public License is included in this distribution in
+# the file called "COPYING".
+#
+# Maintained by: Shreyas Bhatewara <pv-drivers@vmware.com>
+#
+#
+################################################################################
+
+#
+# Makefile for the VMware vmxnet3 ethernet NIC driver
+#
+
+obj-$(CONFIG_VMXNET3) += vmxnet3.o
+
+vmxnet3-objs := vmxnet3_drv.o vmxnet3_ethtool.o
diff --git a/drivers/net/vmxnet3/upt1_defs.h b/drivers/net/vmxnet3/upt1_defs.h
new file mode 100644
index 0000000..b50f91b
--- /dev/null
+++ b/drivers/net/vmxnet3/upt1_defs.h
@@ -0,0 +1,104 @@
+/*
+ * Linux driver for VMware's vmxnet3 ethernet NIC.
+ *
+ * Copyright (C) 2008-2009, VMware, Inc. All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License as published by the
+ * Free Software Foundation; version 2 of the License and no later version.
+ *
+ * This program 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, GOOD TITLE or
+ * NON INFRINGEMENT. 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, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
+ *
+ * The full GNU General Public License is included in this distribution in
+ * the file called "COPYING".
+ *
+ * Maintained by: Shreyas Bhatewara <pv-drivers@vmware.com>
+ *
+ */
+
+/* upt1_defs.h
+ *
+ * Definitions for Uniform Pass Through.
+ */
+
+#ifndef _UPT1_DEFS_H
+#define _UPT1_DEFS_H
+
+#define UPT1_MAX_TX_QUEUES 64
+#define UPT1_MAX_RX_QUEUES 64
+#define UPT1_MAX_INTRS (UPT1_MAX_TX_QUEUES + UPT1_MAX_RX_QUEUES)
+
+struct UPT1_TxStats {
+ uint64_t TSOPktsTxOK; /* TSO pkts post-segmentation */
+ uint64_t TSOBytesTxOK;
+ uint64_t ucastPktsTxOK;
+ uint64_t ucastBytesTxOK;
+ uint64_t mcastPktsTxOK;
+ uint64_t mcastBytesTxOK;
+ uint64_t bcastPktsTxOK;
+ uint64_t bcastBytesTxOK;
+ uint64_t pktsTxError;
+ uint64_t pktsTxDiscard;
+};
+
+struct UPT1_RxStats {
+ uint64_t LROPktsRxOK; /* LRO pkts */
+ uint64_t LROBytesRxOK; /* bytes from LRO pkts */
+ /* the following counters are for pkts from the wire, i.e., pre-LRO */
+ uint64_t ucastPktsRxOK;
+ uint64_t ucastBytesRxOK;
+ uint64_t mcastPktsRxOK;
+ uint64_t mcastBytesRxOK;
+ uint64_t bcastPktsRxOK;
+ uint64_t bcastBytesRxOK;
+ uint64_t pktsRxOutOfBuf;
+ uint64_t pktsRxError;
+};
+
+/* interrupt moderation level */
+#define UPT1_IML_NONE 0 /* no interrupt moderation */
+#define UPT1_IML_HIGHEST 7 /* least intr generated */
+#define UPT1_IML_ADAPTIVE 8 /* adpative intr moderation */
+
+/* values for UPT1_RSSConf.hashFunc */
+enum {
+ UPT1_RSS_HASH_TYPE_NONE = 0x0,
+ UPT1_RSS_HASH_TYPE_IPV4 = 0x01,
+ UPT1_RSS_HASH_TYPE_TCP_IPV4 = 0x02,
+ UPT1_RSS_HASH_TYPE_IPV6 = 0x04,
+ UPT1_RSS_HASH_TYPE_TCP_IPV6 = 0x08,
+};
+
+enum {
+ UPT1_RSS_HASH_FUNC_NONE = 0x0,
+ UPT1_RSS_HASH_FUNC_TOEPLITZ = 0x01,
+};
+
+#define UPT1_RSS_MAX_KEY_SIZE 40
+#define UPT1_RSS_MAX_IND_TABLE_SIZE 128
+
+struct UPT1_RSSConf {
+ uint16_t hashType;
+ uint16_t hashFunc;
+ uint16_t hashKeySize;
+ uint16_t indTableSize;
+ uint8_t hashKey[UPT1_RSS_MAX_KEY_SIZE];
+ uint8_t indTable[UPT1_RSS_MAX_IND_TABLE_SIZE];
+};
+
+/* features */
+enum {
+ UPT1_F_RXCSUM = 0x0001, /* rx csum verification */
+ UPT1_F_RSS = 0x0002,
+ UPT1_F_RXVLAN = 0x0004, /* VLAN tag stripping */
+ UPT1_F_LRO = 0x0008,
+};
+#endif
diff --git a/drivers/net/vmxnet3/vmxnet3_defs.h b/drivers/net/vmxnet3/vmxnet3_defs.h
new file mode 100644
index 0000000..a33a90b
--- /dev/null
+++ b/drivers/net/vmxnet3/vmxnet3_defs.h
@@ -0,0 +1,534 @@
+/*
+ * Linux driver for VMware's vmxnet3 ethernet NIC.
+ *
+ * Copyright (C) 2008-2009, VMware, Inc. All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License as published by the
+ * Free Software Foundation; version 2 of the License and no later version.
+ *
+ * This program 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, GOOD TITLE or
+ * NON INFRINGEMENT. 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, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
+ *
+ * The full GNU General Public License is included in this distribution in
+ * the file called "COPYING".
+ *
+ * Maintained by: Shreyas Bhatewara <pv-drivers@vmware.com>
+ *
+ */
+
+/*
+ * vmxnet3_defs.h --
+ */
+
+#ifndef _VMXNET3_DEFS_H_
+#define _VMXNET3_DEFS_H_
+
+#include "upt1_defs.h"
+
+/* all registers are 32 bit wide */
+/* BAR 1 */
+enum {
+ VMXNET3_REG_VRRS = 0x0, /* Vmxnet3 Revision Report Selection */
+ VMXNET3_REG_UVRS = 0x8, /* UPT Version Report Selection */
+ VMXNET3_REG_DSAL = 0x10, /* Driver Shared Address Low */
+ VMXNET3_REG_DSAH = 0x18, /* Driver Shared Address High */
+ VMXNET3_REG_CMD = 0x20, /* Command */
+ VMXNET3_REG_MACL = 0x28, /* MAC Address Low */
+ VMXNET3_REG_MACH = 0x30, /* MAC Address High */
+ VMXNET3_REG_ICR = 0x38, /* Interrupt Cause Register */
+ VMXNET3_REG_ECR = 0x40 /* Event Cause Register */
+};
+
+/* BAR 0 */
+enum {
+ VMXNET3_REG_IMR = 0x0, /* Interrupt Mask Register */
+ VMXNET3_REG_TXPROD = 0x600, /* Tx Producer Index */
+ VMXNET3_REG_RXPROD = 0x800, /* Rx Producer Index for ring 1 */
+ VMXNET3_REG_RXPROD2 = 0xA00 /* Rx Producer Index for ring 2 */
+};
+
+#define VMXNET3_PT_REG_SIZE 4096 /* BAR 0 */
+#define VMXNET3_VD_REG_SIZE 4096 /* BAR 1 */
+
+#define VMXNET3_REG_ALIGN 8 /* All registers are 8-byte aligned. */
+#define VMXNET3_REG_ALIGN_MASK 0x7
+
+/* I/O Mapped access to registers */
+#define VMXNET3_IO_TYPE_PT 0
+#define VMXNET3_IO_TYPE_VD 1
+#define VMXNET3_IO_ADDR(type, reg) (((type) << 24) | ((reg) & 0xFFFFFF))
+#define VMXNET3_IO_TYPE(addr) ((addr) >> 24)
+#define VMXNET3_IO_REG(addr) ((addr) & 0xFFFFFF)
+
+enum {
+ VMXNET3_CMD_FIRST_SET = 0xCAFE0000,
+ VMXNET3_CMD_ACTIVATE_DEV = VMXNET3_CMD_FIRST_SET,
+ VMXNET3_CMD_QUIESCE_DEV,
+ VMXNET3_CMD_RESET_DEV,
+ VMXNET3_CMD_UPDATE_RX_MODE,
+ VMXNET3_CMD_UPDATE_MAC_FILTERS,
+ VMXNET3_CMD_UPDATE_VLAN_FILTERS,
+ VMXNET3_CMD_UPDATE_RSSIDT,
+ VMXNET3_CMD_UPDATE_IML,
+ VMXNET3_CMD_UPDATE_PMCFG,
+ VMXNET3_CMD_UPDATE_FEATURE,
+ VMXNET3_CMD_LOAD_PLUGIN,
+
+ VMXNET3_CMD_FIRST_GET = 0xF00D0000,
+ VMXNET3_CMD_GET_QUEUE_STATUS = VMXNET3_CMD_FIRST_GET,
+ VMXNET3_CMD_GET_STATS,
+ VMXNET3_CMD_GET_LINK,
+ VMXNET3_CMD_GET_PERM_MAC_LO,
+ VMXNET3_CMD_GET_PERM_MAC_HI,
+ VMXNET3_CMD_GET_DID_LO,
+ VMXNET3_CMD_GET_DID_HI,
+ VMXNET3_CMD_GET_DEV_EXTRA_INFO,
+ VMXNET3_CMD_GET_CONF_INTR
+};
+
+struct Vmxnet3_TxDesc {
+ uint64_t addr;
+
+ uint32_t len:14;
+ uint32_t gen:1; /* generation bit */
+ uint32_t rsvd:1;
+ uint32_t dtype:1; /* descriptor type */
+ uint32_t ext1:1;
+ uint32_t msscof:14; /* MSS, checksum offset, flags */
+
+ uint32_t hlen:10; /* header len */
+ uint32_t om:2; /* offload mode */
+ uint32_t eop:1; /* End Of Packet */
+ uint32_t cq:1; /* completion request */
+ uint32_t ext2:1;
+ uint32_t ti:1; /* VLAN Tag Insertion */
+ uint32_t tci:16; /* Tag to Insert */
+};
+
+/* TxDesc.OM values */
+#define VMXNET3_OM_NONE 0
+#define VMXNET3_OM_CSUM 2
+#define VMXNET3_OM_TSO 3
+
+/* fields in TxDesc we access w/o using bit fields */
+#define VMXNET3_TXD_EOP_SHIFT 12
+#define VMXNET3_TXD_CQ_SHIFT 13
+#define VMXNET3_TXD_GEN_SHIFT 14
+
+#define VMXNET3_TXD_CQ (1 << VMXNET3_TXD_CQ_SHIFT)
+#define VMXNET3_TXD_EOP (1 << VMXNET3_TXD_EOP_SHIFT)
+#define VMXNET3_TXD_GEN (1 << VMXNET3_TXD_GEN_SHIFT)
+
+#define VMXNET3_HDR_COPY_SIZE 128
+
+
+struct Vmxnet3_TxDataDesc {
+ uint8_t data[VMXNET3_HDR_COPY_SIZE];
+};
+
+
+struct Vmxnet3_TxCompDesc {
+ uint32_t txdIdx:12; /* Index of the EOP TxDesc */
+ uint32_t ext1:20;
+
+ uint32_t ext2;
+ uint32_t ext3;
+
+ uint32_t rsvd:24;
+ uint32_t type:7; /* completion type */
+ uint32_t gen:1; /* generation bit */
+};
+
+
+struct Vmxnet3_RxDesc {
+ uint64_t addr;
+
+ uint32_t len:14;
+ uint32_t btype:1; /* Buffer Type */
+ uint32_t dtype:1; /* Descriptor type */
+ uint32_t rsvd:15;
+ uint32_t gen:1; /* Generation bit */
+
+ uint32_t ext1;
+};
+
+/* values of RXD.BTYPE */
+#define VMXNET3_RXD_BTYPE_HEAD 0 /* head only */
+#define VMXNET3_RXD_BTYPE_BODY 1 /* body only */
+
+/* fields in RxDesc we access w/o using bit fields */
+#define VMXNET3_RXD_BTYPE_SHIFT 14
+#define VMXNET3_RXD_GEN_SHIFT 31
+
+
+struct Vmxnet3_RxCompDesc {
+ uint32_t rxdIdx:12; /* Index of the RxDesc */
+ uint32_t ext1:2;
+ uint32_t eop:1; /* End of Packet */
+ uint32_t sop:1; /* Start of Packet */
+ uint32_t rqID:10; /* rx queue/ring ID */
+ uint32_t rssType:4; /* RSS hash type used */
+ uint32_t cnc:1; /* Checksum Not Calculated */
+ uint32_t ext2:1;
+
+ uint32_t rssHash; /* RSS hash value */
+
+ uint32_t len:14; /* data length */
+ uint32_t err:1; /* Error */
+ uint32_t ts:1; /* Tag is stripped */
+ uint32_t tci:16; /* Tag stripped */
+
+ uint32_t csum:16;
+ uint32_t tuc:1; /* TCP/UDP Checksum Correct */
+ uint32_t udp:1; /* UDP packet */
+ uint32_t tcp:1; /* TCP packet */
+ uint32_t ipc:1; /* IP Checksum Correct */
+ uint32_t v6:1; /* IPv6 */
+ uint32_t v4:1; /* IPv4 */
+ uint32_t frg:1; /* IP Fragment */
+ uint32_t fcs:1; /* Frame CRC correct */
+ uint32_t type:7; /* completion type */
+ uint32_t gen:1; /* generation bit */
+};
+
+/* fields in RxCompDesc we access via Vmxnet3_GenericDesc.dword[3] */
+#define VMXNET3_RCD_TUC_SHIFT 16
+#define VMXNET3_RCD_IPC_SHIFT 19
+
+/* fields in RxCompDesc we access via Vmxnet3_GenericDesc.qword[1] */
+#define VMXNET3_RCD_TYPE_SHIFT 56
+#define VMXNET3_RCD_GEN_SHIFT 63
+
+/* csum OK for TCP/UDP pkts over IP */
+#define VMXNET3_RCD_CSUM_OK (1 << VMXNET3_RCD_TUC_SHIFT | \
+ 1 << VMXNET3_RCD_IPC_SHIFT)
+
+/* value of RxCompDesc.rssType */
+enum {
+ VMXNET3_RCD_RSS_TYPE_NONE = 0,
+ VMXNET3_RCD_RSS_TYPE_IPV4 = 1,
+ VMXNET3_RCD_RSS_TYPE_TCPIPV4 = 2,
+ VMXNET3_RCD_RSS_TYPE_IPV6 = 3,
+ VMXNET3_RCD_RSS_TYPE_TCPIPV6 = 4,
+};
+
+/* a union for accessing all cmd/completion descriptors */
+union Vmxnet3_GenericDesc {
+ uint64_t qword[2];
+ uint32_t dword[4];
+ uint16_t word[8];
+ struct Vmxnet3_TxDesc txd;
+ struct Vmxnet3_RxDesc rxd;
+ struct Vmxnet3_TxCompDesc tcd;
+ struct Vmxnet3_RxCompDesc rcd;
+};
+
+#define VMXNET3_INIT_GEN 1
+
+/* Max size of a single tx buffer */
+#define VMXNET3_MAX_TX_BUF_SIZE (1 << 14)
+
+/* # of tx desc needed for a tx buffer size */
+#define VMXNET3_TXD_NEEDED(size) (((size) + VMXNET3_MAX_TX_BUF_SIZE - 1) / \
+ VMXNET3_MAX_TX_BUF_SIZE)
+
+/* max # of tx descs for a non-tso pkt */
+#define VMXNET3_MAX_TXD_PER_PKT 16
+
+/* Max size of a single rx buffer */
+#define VMXNET3_MAX_RX_BUF_SIZE ((1 << 14) - 1)
+/* Minimum size of a type 0 buffer */
+#define VMXNET3_MIN_T0_BUF_SIZE 128
+#define VMXNET3_MAX_CSUM_OFFSET 1024
+
+/* Ring base address alignment */
+#define VMXNET3_RING_BA_ALIGN 512
+#define VMXNET3_RING_BA_MASK (VMXNET3_RING_BA_ALIGN - 1)
+
+/* Ring size must be a multiple of 32 */
+#define VMXNET3_RING_SIZE_ALIGN 32
+#define VMXNET3_RING_SIZE_MASK (VMXNET3_RING_SIZE_ALIGN - 1)
+
+/* Max ring size */
+#define VMXNET3_TX_RING_MAX_SIZE 4096
+#define VMXNET3_TC_RING_MAX_SIZE 4096
+#define VMXNET3_RX_RING_MAX_SIZE 4096
+#define VMXNET3_RC_RING_MAX_SIZE 8192
+
+/* a list of reasons for queue stop */
+
+enum {
+ VMXNET3_ERR_NOEOP = 0x80000000, /* cannot find the EOP desc of a pkt */
+ VMXNET3_ERR_TXD_REUSE = 0x80000001, /* reuse TxDesc before tx completion */
+ VMXNET3_ERR_BIG_PKT = 0x80000002, /* too many TxDesc for a pkt */
+ VMXNET3_ERR_DESC_NOT_SPT = 0x80000003, /* descriptor type not supported */
+ VMXNET3_ERR_SMALL_BUF = 0x80000004, /* type 0 buffer too small */
+ VMXNET3_ERR_STRESS = 0x80000005, /* stress option firing in vmkernel */
+ VMXNET3_ERR_SWITCH = 0x80000006, /* mode switch failure */
+ VMXNET3_ERR_TXD_INVALID = 0x80000007, /* invalid TxDesc */
+};
+
+/* completion descriptor types */
+#define VMXNET3_CDTYPE_TXCOMP 0 /* Tx Completion Descriptor */
+#define VMXNET3_CDTYPE_RXCOMP 3 /* Rx Completion Descriptor */
+
+enum {
+ VMXNET3_GOS_BITS_UNK = 0, /* unknown */
+ VMXNET3_GOS_BITS_32 = 1,
+ VMXNET3_GOS_BITS_64 = 2,
+};
+
+#define VMXNET3_GOS_TYPE_LINUX 1
+
+/* All structures in DriverShared are padded to multiples of 8 bytes */
+
+
+struct Vmxnet3_GOSInfo {
+ uint32_t gosBits:2; /* 32-bit or 64-bit? */
+ uint32_t gosType:4; /* which guest */
+ uint32_t gosVer:16; /* gos version */
+ uint32_t gosMisc:10; /* other info about gos */
+};
+
+
+struct Vmxnet3_DriverInfo {
+ uint32_t version; /* driver version */
+ struct Vmxnet3_GOSInfo gos;
+ uint32_t vmxnet3RevSpt; /* vmxnet3 revision supported */
+ uint32_t uptVerSpt; /* upt version supported */
+};
+
+#define VMXNET3_REV1_MAGIC 0xbabefee1
+
+/*
+ * QueueDescPA must be 128 bytes aligned. It points to an array of
+ * Vmxnet3_TxQueueDesc followed by an array of Vmxnet3_RxQueueDesc.
+ * The number of Vmxnet3_TxQueueDesc/Vmxnet3_RxQueueDesc are specified by
+ * Vmxnet3_MiscConf.numTxQueues/numRxQueues, respectively.
+ */
+#define VMXNET3_QUEUE_DESC_ALIGN 128
+
+
+struct Vmxnet3_MiscConf {
+ struct Vmxnet3_DriverInfo driverInfo;
+ uint64_t uptFeatures;
+ uint64_t ddPA; /* driver data PA */
+ uint64_t queueDescPA; /* queue descriptor table PA */
+ uint32_t ddLen; /* driver data len */
+ uint32_t queueDescLen; /* queue desc. table len in bytes */
+ uint32_t mtu;
+ uint16_t maxNumRxSG;
+ uint8_t numTxQueues;
+ uint8_t numRxQueues;
+ uint32_t reserved[4];
+};
+
+
+struct Vmxnet3_TxQueueConf {
+ uint64_t txRingBasePA;
+ uint64_t dataRingBasePA;
+ uint64_t compRingBasePA;
+ uint64_t ddPA; /* driver data */
+ uint64_t reserved;
+ uint32_t txRingSize; /* # of tx desc */
+ uint32_t dataRingSize; /* # of data desc */
+ uint32_t compRingSize; /* # of comp desc */
+ uint32_t ddLen; /* size of driver data */
+ uint8_t intrIdx;
+ uint8_t _pad[7];
+};
+
+
+struct Vmxnet3_RxQueueConf {
+ uint64_t rxRingBasePA[2];
+ uint64_t compRingBasePA;
+ uint64_t ddPA; /* driver data */
+ uint64_t reserved;
+ uint32_t rxRingSize[2]; /* # of rx desc */
+ uint32_t compRingSize; /* # of rx comp desc */
+ uint32_t ddLen; /* size of driver data */
+ uint8_t intrIdx;
+ uint8_t _pad[7];
+};
+
+enum vmxnet3_intr_mask_mode {
+ VMXNET3_IMM_AUTO = 0,
+ VMXNET3_IMM_ACTIVE = 1,
+ VMXNET3_IMM_LAZY = 2
+};
+
+enum vmxnet3_intr_type {
+ VMXNET3_IT_AUTO = 0,
+ VMXNET3_IT_INTX = 1,
+ VMXNET3_IT_MSI = 2,
+ VMXNET3_IT_MSIX = 3
+};
+
+#define VMXNET3_MAX_TX_QUEUES 8
+#define VMXNET3_MAX_RX_QUEUES 16
+/* addition 1 for events */
+#define VMXNET3_MAX_INTRS 25
+
+
+struct Vmxnet3_IntrConf {
+ bool autoMask;
+ uint8_t numIntrs; /* # of interrupts */
+ uint8_t eventIntrIdx;
+ uint8_t modLevels[VMXNET3_MAX_INTRS]; /* moderation level for
+ * each intr */
+ uint32_t reserved[3];
+};
+
+/* one bit per VLAN ID, the size is in the units of uint32_t */
+#define VMXNET3_VFT_SIZE (4096 / (sizeof(u32) * 8))
+
+
+struct Vmxnet3_QueueStatus {
+ bool stopped;
+ uint8_t _pad[3];
+ uint32_t error;
+};
+
+
+struct Vmxnet3_TxQueueCtrl {
+ uint32_t txNumDeferred;
+ uint32_t txThreshold;
+ uint64_t reserved;
+};
+
+
+struct Vmxnet3_RxQueueCtrl {
+ bool updateRxProd;
+ uint8_t _pad[7];
+ uint64_t reserved;
+};
+
+enum {
+ VMXNET3_RXM_UCAST = 0x01, /* unicast only */
+ VMXNET3_RXM_MCAST = 0x02, /* multicast passing the filters */
+ VMXNET3_RXM_BCAST = 0x04, /* broadcast only */
+ VMXNET3_RXM_ALL_MULTI = 0x08, /* all multicast */
+ VMXNET3_RXM_PROMISC = 0x10 /* promiscuous */
+};
+
+struct Vmxnet3_RxFilterConf {
+ uint32_t rxMode; /* VMXNET3_RXM_xxx */
+ uint16_t mfTableLen; /* size of the multicast filter table */
+ uint16_t _pad1;
+ uint64_t mfTablePA; /* PA of the multicast filters table */
+ uint32_t vfTable[VMXNET3_VFT_SIZE]; /* vlan filter */
+};
+
+#define VMXNET3_PM_MAX_FILTERS 6
+#define VMXNET3_PM_MAX_PATTERN_SIZE 128
+#define VMXNET3_PM_MAX_MASK_SIZE (VMXNET3_PM_MAX_PATTERN_SIZE / 8)
+
+#define VMXNET3_PM_WAKEUP_MAGIC 0x01 /* wake up on magic pkts */
+#define VMXNET3_PM_WAKEUP_FILTER 0x02 /* wake up on pkts matching
+ * filters */
+
+
+struct Vmxnet3_PM_PktFilter {
+ uint8_t maskSize;
+ uint8_t patternSize;
+ uint8_t mask[VMXNET3_PM_MAX_MASK_SIZE];
+ uint8_t pattern[VMXNET3_PM_MAX_PATTERN_SIZE];
+ uint8_t pad[6];
+};
+
+
+struct Vmxnet3_PMConf {
+ uint16_t wakeUpEvents; /* VMXNET3_PM_WAKEUP_xxx */
+ uint8_t numFilters;
+ uint8_t pad[5];
+ struct Vmxnet3_PM_PktFilter filters[VMXNET3_PM_MAX_FILTERS];
+};
+
+
+struct Vmxnet3_VariableLenConfDesc {
+ uint32_t confVer;
+ uint32_t confLen;
+ uint64_t confPA;
+};
+
+
+struct Vmxnet3_DSDevRead {
+ /* read-only region for device, read by dev in response to a SET cmd */
+ struct Vmxnet3_MiscConf misc;
+ struct Vmxnet3_IntrConf intrConf;
+ struct Vmxnet3_RxFilterConf rxFilterConf;
+ struct Vmxnet3_VariableLenConfDesc rssConfDesc;
+ struct Vmxnet3_VariableLenConfDesc pmConfDesc;
+ struct Vmxnet3_VariableLenConfDesc pluginConfDesc;
+};
+
+
+struct Vmxnet3_TxQueueDesc {
+ struct Vmxnet3_TxQueueCtrl ctrl;
+ struct Vmxnet3_TxQueueConf conf;
+ /* Driver read after a GET command */
+ struct Vmxnet3_QueueStatus status;
+ struct UPT1_TxStats stats;
+ uint8_t _pad[88]; /* 128 aligned */
+};
+
+
+struct Vmxnet3_RxQueueDesc {
+ struct Vmxnet3_RxQueueCtrl ctrl;
+ struct Vmxnet3_RxQueueConf conf;
+ /* Driver read after a GET command */
+ struct Vmxnet3_QueueStatus status;
+ struct UPT1_RxStats stats;
+ uint8_t _pad[88]; /* 128 aligned */
+};
+
+
+struct Vmxnet3_DriverShared {
+ uint32_t magic;
+ uint32_t pad; /* make devRead start at 64bit boundaries */
+ struct Vmxnet3_DSDevRead devRead;
+ uint32_t ecr;
+ uint32_t reserved[5];
+};
+
+#define VMXNET3_ECR_RQERR (1 << 0)
+#define VMXNET3_ECR_TQERR (1 << 1)
+#define VMXNET3_ECR_LINK (1 << 2)
+#define VMXNET3_ECR_DIC (1 << 3)
+#define VMXNET3_ECR_DEBUG (1 << 4)
+
+/* flip the gen bit of a ring */
+#define VMXNET3_FLIP_RING_GEN(gen) ((gen) = (gen) ^ 0x1)
+
+/* only use this if moving the idx won't affect the gen bit */
+#define VMXNET3_INC_RING_IDX_ONLY(idx, ring_size) \
+ do {\
+ (idx)++;\
+ if (unlikely((idx) == (ring_size))) {\
+ (idx) = 0;\
+ } \
+ } while (0)
+
+#define VMXNET3_SET_VFTABLE_ENTRY(vfTable, vid) \
+ vfTable[vid >> 5] |= (1 << (vid & 31))
+#define VMXNET3_CLEAR_VFTABLE_ENTRY(vfTable, vid) \
+ vfTable[vid >> 5] &= ~(1 << (vid & 31))
+
+#define VMXNET3_VFTABLE_ENTRY_IS_SET(vfTable, vid) \
+ ((vfTable[vid >> 5] & (1 << (vid & 31))) != 0)
+
+#define VMXNET3_MAX_MTU 9000
+#define VMXNET3_MIN_MTU 60
+
+#define VMXNET3_LINK_UP (10000 << 16 | 1) /* 10 Gbps, up */
+#define VMXNET3_LINK_DOWN 0
+
+#endif /* _VMXNET3_DEFS_H_ */
diff --git a/drivers/net/vmxnet3/vmxnet3_drv.c b/drivers/net/vmxnet3/vmxnet3_drv.c
new file mode 100644
index 0000000..d9fa4e3
--- /dev/null
+++ b/drivers/net/vmxnet3/vmxnet3_drv.c
@@ -0,0 +1,2608 @@
+/*
+ * Linux driver for VMware's vmxnet3 ethernet NIC.
+ *
+ * Copyright (C) 2008-2009, VMware, Inc. All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License as published by the
+ * Free Software Foundation; version 2 of the License and no later version.
+ *
+ * This program 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, GOOD TITLE or
+ * NON INFRINGEMENT. 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, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
+ *
+ * The full GNU General Public License is included in this distribution in
+ * the file called "COPYING".
+ *
+ * Maintained by: Shreyas Bhatewara <pv-drivers@vmware.com>
+ *
+ */
+
+/*
+ * vmxnet3_drv.c --
+ *
+ * Linux driver for VMware's vmxnet3 NIC
+ */
+
+
+#include "vmxnet3_int.h"
+
+char vmxnet3_driver_name[] = "vmxnet3";
+#define VMXNET3_DRIVER_DESC "VMware vmxnet3 virtual NIC driver"
+
+
+/*
+ * PCI Device ID Table
+ * Last entry must be all 0s
+ */
+static const struct pci_device_id vmxnet3_pciid_table[] = {
+ {PCI_VDEVICE(VMWARE, PCI_DEVICE_ID_VMWARE_VMXNET3)},
+ {0}
+};
+
+MODULE_DEVICE_TABLE(pci, vmxnet3_pciid_table);
+
+static int disable_lro;
+static atomic_t devices_found;
+
+
+/*
+ * Enable/Disable the given intr
+ */
+static void
+vmxnet3_enable_intr(struct vmxnet3_adapter *adapter, unsigned intr_idx)
+{
+ VMXNET3_WRITE_BAR0_REG(adapter, VMXNET3_REG_IMR + intr_idx * 8, 0);
+}
+
+
+static void
+vmxnet3_disable_intr(struct vmxnet3_adapter *adapter, unsigned intr_idx)
+{
+ VMXNET3_WRITE_BAR0_REG(adapter, VMXNET3_REG_IMR + intr_idx * 8, 1);
+}
+
+
+/*
+ * Enable/Disable all intrs used by the device
+ */
+static void
+vmxnet3_enable_all_intrs(struct vmxnet3_adapter *adapter)
+{
+ int i;
+
+ for (i = 0; i < adapter->intr.num_intrs; i++)
+ vmxnet3_enable_intr(adapter, i);
+}
+
+
+static void
+vmxnet3_disable_all_intrs(struct vmxnet3_adapter *adapter)
+{
+ int i;
+
+ for (i = 0; i < adapter->intr.num_intrs; i++)
+ vmxnet3_disable_intr(adapter, i);
+}
+
+
+static void
+vmxnet3_ack_events(struct vmxnet3_adapter *adapter, u32 events)
+{
+ VMXNET3_WRITE_BAR1_REG(adapter, VMXNET3_REG_ECR, events);
+}
+
+
+static bool
+vmxnet3_tq_stopped(struct vmxnet3_tx_queue *tq, struct vmxnet3_adapter *adapter)
+{
+ return netif_queue_stopped(adapter->netdev);
+}
+
+
+static void
+vmxnet3_tq_start(struct vmxnet3_tx_queue *tq, struct vmxnet3_adapter *adapter)
+{
+ tq->stopped = false;
+ netif_start_queue(adapter->netdev);
+}
+
+
+static void
+vmxnet3_tq_wake(struct vmxnet3_tx_queue *tq, struct vmxnet3_adapter *adapter)
+{
+ tq->stopped = false;
+ netif_wake_queue(adapter->netdev);
+}
+
+
+static void
+vmxnet3_tq_stop(struct vmxnet3_tx_queue *tq, struct vmxnet3_adapter *adapter)
+{
+ tq->stopped = true;
+ tq->num_stop++;
+ netif_stop_queue(adapter->netdev);
+}
+
+
+/*
+ * Check the link state. This may start or stop the tx queue.
+ */
+static void
+vmxnet3_check_link(struct vmxnet3_adapter *adapter)
+{
+ u32 ret;
+
+ VMXNET3_WRITE_BAR1_REG(adapter, VMXNET3_REG_CMD, VMXNET3_CMD_GET_LINK);
+ ret = VMXNET3_READ_BAR1_REG(adapter, VMXNET3_REG_CMD);
+ adapter->link_speed = ret >> 16;
+ if (ret & 1) { /* Link is up. */
+ printk(KERN_INFO "%s: NIC Link is Up %d Mbps\n",
+ adapter->netdev->name, adapter->link_speed);
+ if (!netif_carrier_ok(adapter->netdev))
+ netif_carrier_on(adapter->netdev);
+
+ vmxnet3_tq_start(&adapter->tx_queue, adapter);
+ } else {
+ printk(KERN_INFO "%s: NIC Link is Down\n",
+ adapter->netdev->name);
+ if (netif_carrier_ok(adapter->netdev))
+ netif_carrier_off(adapter->netdev);
+
+ vmxnet3_tq_stop(&adapter->tx_queue, adapter);
+ }
+}
+
+
+static void
+vmxnet3_process_events(struct vmxnet3_adapter *adapter)
+{
+ u32 events = adapter->shared->ecr;
+ if (!events)
+ return;
+
+ vmxnet3_ack_events(adapter, events);
+
+ /* Check if link state has changed */
+ if (events & VMXNET3_ECR_LINK)
+ vmxnet3_check_link(adapter);
+
+ /* Check if there is an error on xmit/recv queues */
+ if (events & (VMXNET3_ECR_TQERR | VMXNET3_ECR_RQERR)) {
+ VMXNET3_WRITE_BAR1_REG(adapter, VMXNET3_REG_CMD,
+ VMXNET3_CMD_GET_QUEUE_STATUS);
+
+ if (adapter->tqd_start->status.stopped) {
+ printk(KERN_ERR "%s: tq error 0x%x\n",
+ adapter->netdev->name,
+ adapter->tqd_start->status.error);
+ }
+ if (adapter->rqd_start->status.stopped) {
+ printk(KERN_ERR "%s: rq error 0x%x\n",
+ adapter->netdev->name,
+ adapter->rqd_start->status.error);
+ }
+
+ schedule_work(&adapter->work);
+ }
+}
+
+
+static void
+vmxnet3_unmap_tx_buf(struct vmxnet3_tx_buf_info *tbi,
+ struct pci_dev *pdev)
+{
+ if (tbi->map_type == VMXNET3_MAP_SINGLE)
+ pci_unmap_single(pdev, tbi->dma_addr, tbi->len,
+ PCI_DMA_TODEVICE);
+ else if (tbi->map_type == VMXNET3_MAP_PAGE)
+ pci_unmap_page(pdev, tbi->dma_addr, tbi->len,
+ PCI_DMA_TODEVICE);
+ else
+ BUG_ON(tbi->map_type != VMXNET3_MAP_NONE);
+
+ tbi->map_type = VMXNET3_MAP_NONE; /* to help debugging */
+}
+
+
+static int
+vmxnet3_unmap_pkt(u32 eop_idx, struct vmxnet3_tx_queue *tq,
+ struct pci_dev *pdev, struct vmxnet3_adapter *adapter)
+{
+ struct sk_buff *skb;
+ int entries = 0;
+
+ /* no out of order completion */
+ BUG_ON(tq->buf_info[eop_idx].sop_idx != tq->tx_ring.next2comp);
+ BUG_ON(tq->tx_ring.base[eop_idx].txd.eop != 1);
+
+ dprintk(KERN_ERR "tx complete [%u %u]\n", tq->tx_ring.next2comp,
+ eop_idx);
+
+ skb = tq->buf_info[eop_idx].skb;
+ BUG_ON(skb == NULL);
+ tq->buf_info[eop_idx].skb = NULL;
+
+ VMXNET3_INC_RING_IDX_ONLY(eop_idx, tq->tx_ring.size);
+
+ while (tq->tx_ring.next2comp != eop_idx) {
+ vmxnet3_unmap_tx_buf(tq->buf_info + tq->tx_ring.next2comp,
+ pdev);
+
+ /* update next2comp w/o tx_lock. Since we are marking more,
+ * instead of less, tx ring entries avail, the worst case is
+ * that the tx routine incorrectly re-queues a pkt due to
+ * insufficient tx ring entries.
+ */
+ vmxnet3_cmd_ring_adv_next2comp(&tq->tx_ring);
+ entries++;
+ }
+
+ dev_kfree_skb_any(skb);
+ return entries;
+}
+
+
+static int
+vmxnet3_tq_tx_complete(struct vmxnet3_tx_queue *tq,
+ struct vmxnet3_adapter *adapter)
+{
+ int completed = 0;
+ union Vmxnet3_GenericDesc *gdesc;
+
+ gdesc = tq->comp_ring.base + tq->comp_ring.next2proc;
+ while (gdesc->tcd.gen == tq->comp_ring.gen) {
+ completed += vmxnet3_unmap_pkt(gdesc->tcd.txdIdx, tq,
+ adapter->pdev, adapter);
+
+ vmxnet3_comp_ring_adv_next2proc(&tq->comp_ring);
+ gdesc = tq->comp_ring.base + tq->comp_ring.next2proc;
+ }
+
+ if (completed) {
+ spin_lock(&tq->tx_lock);
+ if (unlikely(vmxnet3_tq_stopped(tq, adapter) &&
+ vmxnet3_cmd_ring_desc_avail(&tq->tx_ring) >
+ VMXNET3_WAKE_QUEUE_THRESHOLD(tq) &&
+ netif_carrier_ok(adapter->netdev))) {
+ vmxnet3_tq_wake(tq, adapter);
+ }
+ spin_unlock(&tq->tx_lock);
+ }
+ return completed;
+}
+
+
+static void
+vmxnet3_tq_cleanup(struct vmxnet3_tx_queue *tq,
+ struct vmxnet3_adapter *adapter)
+{
+ int i;
+
+ while (tq->tx_ring.next2comp != tq->tx_ring.next2fill) {
+ struct vmxnet3_tx_buf_info *tbi;
+ union Vmxnet3_GenericDesc *gdesc;
+
+ tbi = tq->buf_info + tq->tx_ring.next2comp;
+ gdesc = tq->tx_ring.base + tq->tx_ring.next2comp;
+
+ vmxnet3_unmap_tx_buf(tbi, adapter->pdev);
+ if (tbi->skb) {
+ dev_kfree_skb_any(tbi->skb);
+ tbi->skb = NULL;
+ }
+ vmxnet3_cmd_ring_adv_next2comp(&tq->tx_ring);
+ }
+
+ /* sanity check, verify all buffers are indeed unmapped and freed */
+ for (i = 0; i < tq->tx_ring.size; i++) {
+ BUG_ON(tq->buf_info[i].skb != NULL ||
+ tq->buf_info[i].map_type != VMXNET3_MAP_NONE);
+ }
+
+ tq->tx_ring.gen = VMXNET3_INIT_GEN;
+ tq->tx_ring.next2fill = tq->tx_ring.next2comp = 0;
+
+ tq->comp_ring.gen = VMXNET3_INIT_GEN;
+ tq->comp_ring.next2proc = 0;
+}
+
+
+void
+vmxnet3_tq_destroy(struct vmxnet3_tx_queue *tq,
+ struct vmxnet3_adapter *adapter)
+{
+ if (tq->tx_ring.base) {
+ pci_free_consistent(adapter->pdev, tq->tx_ring.size *
+ sizeof(struct Vmxnet3_TxDesc),
+ tq->tx_ring.base, tq->tx_ring.basePA);
+ tq->tx_ring.base = NULL;
+ }
+ if (tq->data_ring.base) {
+ pci_free_consistent(adapter->pdev, tq->data_ring.size *
+ sizeof(struct Vmxnet3_TxDataDesc),
+ tq->data_ring.base, tq->data_ring.basePA);
+ tq->data_ring.base = NULL;
+ }
+ if (tq->comp_ring.base) {
+ pci_free_consistent(adapter->pdev, tq->comp_ring.size *
+ sizeof(struct Vmxnet3_TxCompDesc),
+ tq->comp_ring.base, tq->comp_ring.basePA);
+ tq->comp_ring.base = NULL;
+ }
+ kfree(tq->buf_info);
+ tq->buf_info = NULL;
+}
+
+
+static void
+vmxnet3_tq_init(struct vmxnet3_tx_queue *tq,
+ struct vmxnet3_adapter *adapter)
+{
+ int i;
+
+ /* reset the tx ring contents to 0 and reset the tx ring states */
+ memset(tq->tx_ring.base, 0, tq->tx_ring.size *
+ sizeof(struct Vmxnet3_TxDesc));
+ tq->tx_ring.next2fill = tq->tx_ring.next2comp = 0;
+ tq->tx_ring.gen = VMXNET3_INIT_GEN;
+
+ memset(tq->data_ring.base, 0, tq->data_ring.size *
+ sizeof(struct Vmxnet3_TxDataDesc));
+
+ /* reset the tx comp ring contents to 0 and reset comp ring states */
+ memset(tq->comp_ring.base, 0, tq->comp_ring.size *
+ sizeof(struct Vmxnet3_TxCompDesc));
+ tq->comp_ring.next2proc = 0;
+ tq->comp_ring.gen = VMXNET3_INIT_GEN;
+
+ /* reset the bookkeeping data */
+ memset(tq->buf_info, 0, sizeof(tq->buf_info[0]) * tq->tx_ring.size);
+ for (i = 0; i < tq->tx_ring.size; i++)
+ tq->buf_info[i].map_type = VMXNET3_MAP_NONE;
+
+ /* stats are not reset */
+}
+
+
+static int
+vmxnet3_tq_create(struct vmxnet3_tx_queue *tq,
+ struct vmxnet3_adapter *adapter)
+{
+ BUG_ON(tq->tx_ring.size <= 0 || tq->data_ring.size != tq->tx_ring.size);
+ BUG_ON((tq->tx_ring.size & VMXNET3_RING_SIZE_MASK) != 0);
+ BUG_ON(tq->tx_ring.base || tq->data_ring.base ||
+ tq->comp_ring.base || tq->buf_info);
+
+ tq->tx_ring.base = pci_alloc_consistent(adapter->pdev, tq->tx_ring.size
+ * sizeof(struct Vmxnet3_TxDesc),
+ &tq->tx_ring.basePA);
+ if (!tq->tx_ring.base) {
+ printk(KERN_ERR "%s: failed to allocate tx ring\n",
+ adapter->netdev->name);
+ goto err;
+ }
+
+ tq->data_ring.base = pci_alloc_consistent(adapter->pdev,
+ tq->data_ring.size *
+ sizeof(struct Vmxnet3_TxDataDesc),
+ &tq->data_ring.basePA);
+ if (!tq->data_ring.base) {
+ printk(KERN_ERR "%s: failed to allocate data ring\n",
+ adapter->netdev->name);
+ goto err;
+ }
+
+ tq->comp_ring.base = pci_alloc_consistent(adapter->pdev,
+ tq->comp_ring.size *
+ sizeof(struct Vmxnet3_TxCompDesc),
+ &tq->comp_ring.basePA);
+ if (!tq->comp_ring.base) {
+ printk(KERN_ERR "%s: failed to allocate tx comp ring\n",
+ adapter->netdev->name);
+ goto err;
+ }
+
+ tq->buf_info = kcalloc(sizeof(tq->buf_info[0]), tq->tx_ring.size,
+ GFP_KERNEL);
+ if (!tq->buf_info) {
+ printk(KERN_ERR "%s: failed to allocate tx bufinfo\n",
+ adapter->netdev->name);
+ goto err;
+ }
+
+ return 0;
+
+err:
+ vmxnet3_tq_destroy(tq, adapter);
+ return -ENOMEM;
+}
+
+
+/*
+ * starting from ring->next2fill, allocate rx buffers for the given ring
+ * of the rx queue and update the rx desc. stop after @num_to_alloc buffers
+ * are allocated or allocation fails
+ */
+
+static int
+vmxnet3_rq_alloc_rx_buf(struct vmxnet3_rx_queue *rq, u32 ring_idx,
+ int num_to_alloc, struct vmxnet3_adapter *adapter)
+{
+ int num_allocated = 0;
+ struct vmxnet3_rx_buf_info *rbi_base = rq->buf_info[ring_idx];
+ struct vmxnet3_cmd_ring *ring = &rq->rx_ring[ring_idx];
+ u32 val;
+
+ while (num_allocated < num_to_alloc) {
+ struct vmxnet3_rx_buf_info *rbi;
+ union Vmxnet3_GenericDesc *gd;
+
+ rbi = rbi_base + ring->next2fill;
+ gd = ring->base + ring->next2fill;
+
+ if (rbi->buf_type == VMXNET3_RX_BUF_SKB) {
+ if (rbi->skb == NULL) {
+ rbi->skb = dev_alloc_skb(rbi->len +
+ NET_IP_ALIGN);
+ if (unlikely(rbi->skb == NULL)) {
+ rq->stats.rx_buf_alloc_failure++;
+ break;
+ }
+ rbi->skb->dev = adapter->netdev;
+
+ skb_reserve(rbi->skb, NET_IP_ALIGN);
+ rbi->dma_addr = pci_map_single(adapter->pdev,
+ rbi->skb->data, rbi->len,
+ PCI_DMA_FROMDEVICE);
+ } else {
+ /* rx buffer skipped by the device */
+ }
+ val = VMXNET3_RXD_BTYPE_HEAD << VMXNET3_RXD_BTYPE_SHIFT;
+ } else {
+ BUG_ON(rbi->buf_type != VMXNET3_RX_BUF_PAGE ||
+ rbi->len != PAGE_SIZE);
+
+ if (rbi->page == NULL) {
+ rbi->page = alloc_page(GFP_ATOMIC);
+ if (unlikely(rbi->page == NULL)) {
+ rq->stats.rx_buf_alloc_failure++;
+ break;
+ }
+ rbi->dma_addr = pci_map_page(adapter->pdev,
+ rbi->page, 0, PAGE_SIZE,
+ PCI_DMA_FROMDEVICE);
+ } else {
+ /* rx buffers skipped by the device */
+ }
+ val = VMXNET3_RXD_BTYPE_BODY << VMXNET3_RXD_BTYPE_SHIFT;
+ }
+
+ BUG_ON(rbi->dma_addr == 0);
+ gd->rxd.addr = rbi->dma_addr;
+ wmb();
+ gd->dword[2] = (ring->gen << VMXNET3_RXD_GEN_SHIFT) | val |
+ rbi->len;
+
+ num_allocated++;
+ vmxnet3_cmd_ring_adv_next2fill(ring);
+ }
+ rq->uncommitted[ring_idx] += num_allocated;
+
+ dprintk(KERN_ERR "alloc_rx_buf: %d allocated, next2fill %u, next2comp "
+ "%u, uncommited %u\n", num_allocated, ring->next2fill,
+ ring->next2comp, rq->uncommitted[ring_idx]);
+
+ /* so that the device can distinguish a full ring and an empty ring */
+ BUG_ON(num_allocated != 0 && ring->next2fill == ring->next2comp);
+
+ return num_allocated;
+}
+
+
+static void
+vmxnet3_append_frag(struct sk_buff *skb, struct Vmxnet3_RxCompDesc *rcd,
+ struct vmxnet3_rx_buf_info *rbi)
+{
+ struct skb_frag_struct *frag = skb_shinfo(skb)->frags +
+ skb_shinfo(skb)->nr_frags;
+
+ BUG_ON(skb_shinfo(skb)->nr_frags >= MAX_SKB_FRAGS);
+
+ frag->page = rbi->page;
+ frag->page_offset = 0;
+ frag->size = rcd->len;
+ skb->data_len += frag->size;
+ skb_shinfo(skb)->nr_frags++;
+}
+
+
+static void
+vmxnet3_map_pkt(struct sk_buff *skb, struct vmxnet3_tx_ctx *ctx,
+ struct vmxnet3_tx_queue *tq, struct pci_dev *pdev,
+ struct vmxnet3_adapter *adapter)
+{
+ u32 dw2, len;
+ unsigned long buf_offset;
+ int i;
+ union Vmxnet3_GenericDesc *gdesc;
+ struct vmxnet3_tx_buf_info *tbi = NULL;
+
+ BUG_ON(ctx->copy_size > skb_headlen(skb));
+
+ /* use the previous gen bit for the SOP desc */
+ dw2 = (tq->tx_ring.gen ^ 0x1) << VMXNET3_TXD_GEN_SHIFT;
+
+ ctx->sop_txd = tq->tx_ring.base + tq->tx_ring.next2fill;
+ gdesc = ctx->sop_txd; /* both loops below can be skipped */
+
+ /* no need to map the buffer if headers are copied */
+ if (ctx->copy_size) {
+ BUG_ON(ctx->sop_txd->txd.gen == tq->tx_ring.gen);
+
+ ctx->sop_txd->txd.addr = tq->data_ring.basePA +
+ tq->tx_ring.next2fill *
+ sizeof(struct Vmxnet3_TxDataDesc);
+ ctx->sop_txd->dword[2] = dw2 | ctx->copy_size;
+ ctx->sop_txd->dword[3] = 0;
+
+ tbi = tq->buf_info + tq->tx_ring.next2fill;
+ tbi->map_type = VMXNET3_MAP_NONE;
+
+ dprintk(KERN_ERR "txd[%u]: 0x%Lx 0x%x 0x%x\n",
+ tq->tx_ring.next2fill, ctx->sop_txd->txd.addr,
+ ctx->sop_txd->dword[2], ctx->sop_txd->dword[3]);
+ vmxnet3_cmd_ring_adv_next2fill(&tq->tx_ring);
+
+ /* use the right gen for non-SOP desc */
+ dw2 = tq->tx_ring.gen << VMXNET3_TXD_GEN_SHIFT;
+ }
+
+ /* linear part can use multiple tx desc if it's big */
+ len = skb_headlen(skb) - ctx->copy_size;
+ buf_offset = ctx->copy_size;
+ while (len) {
+ u32 buf_size;
+
+ buf_size = len > VMXNET3_MAX_TX_BUF_SIZE ?
+ VMXNET3_MAX_TX_BUF_SIZE : len;
+
+ tbi = tq->buf_info + tq->tx_ring.next2fill;
+ tbi->map_type = VMXNET3_MAP_SINGLE;
+ tbi->dma_addr = pci_map_single(adapter->pdev,
+ skb->data + buf_offset, buf_size,
+ PCI_DMA_TODEVICE);
+
+ tbi->len = buf_size; /* this automatically convert 2^14 to 0 */
+
+ gdesc = tq->tx_ring.base + tq->tx_ring.next2fill;
+ BUG_ON(gdesc->txd.gen == tq->tx_ring.gen);
+
+ gdesc->txd.addr = tbi->dma_addr;
+ gdesc->dword[2] = dw2 | buf_size;
+ gdesc->dword[3] = 0;
+
+ dprintk(KERN_ERR "txd[%u]: 0x%Lx 0x%x 0x%x\n",
+ tq->tx_ring.next2fill, gdesc->txd.addr,
+ gdesc->dword[2], gdesc->dword[3]);
+ vmxnet3_cmd_ring_adv_next2fill(&tq->tx_ring);
+ dw2 = tq->tx_ring.gen << VMXNET3_TXD_GEN_SHIFT;
+
+ len -= buf_size;
+ buf_offset += buf_size;
+ }
+
+ for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
+ struct skb_frag_struct *frag = &skb_shinfo(skb)->frags[i];
+
+ tbi = tq->buf_info + tq->tx_ring.next2fill;
+ tbi->map_type = VMXNET3_MAP_PAGE;
+ tbi->dma_addr = pci_map_page(adapter->pdev, frag->page,
+ frag->page_offset, frag->size,
+ PCI_DMA_TODEVICE);
+
+ tbi->len = frag->size;
+
+ gdesc = tq->tx_ring.base + tq->tx_ring.next2fill;
+ BUG_ON(gdesc->txd.gen == tq->tx_ring.gen);
+
+ gdesc->txd.addr = tbi->dma_addr;
+ gdesc->dword[2] = dw2 | frag->size;
+ gdesc->dword[3] = 0;
+
+ dprintk(KERN_ERR "txd[%u]: 0x%llu %u %u\n",
+ tq->tx_ring.next2fill, gdesc->txd.addr,
+ gdesc->dword[2], gdesc->dword[3]);
+ vmxnet3_cmd_ring_adv_next2fill(&tq->tx_ring);
+ dw2 = tq->tx_ring.gen << VMXNET3_TXD_GEN_SHIFT;
+ }
+
+ ctx->eop_txd = gdesc;
+
+ /* set the last buf_info for the pkt */
+ tbi->skb = skb;
+ tbi->sop_idx = ctx->sop_txd - tq->tx_ring.base;
+}
+
+
+/*
+ * parse and copy relevant protocol headers:
+ * For a tso pkt, relevant headers are L2/3/4 including options
+ * For a pkt requesting csum offloading, they are L2/3 and may include L4
+ * if it's a TCP/UDP pkt
+ *
+ * Returns:
+ * -1: error happens during parsing
+ * 0: protocol headers parsed, but too big to be copied
+ * 1: protocol headers parsed and copied
+ *
+ * Other effects:
+ * 1. related *ctx fields are updated.
+ * 2. ctx->copy_size is # of bytes copied
+ * 3. the portion copied is guaranteed to be in the linear part
+ *
+ */
+static int
+vmxnet3_parse_and_copy_hdr(struct sk_buff *skb, struct vmxnet3_tx_queue *tq,
+ struct vmxnet3_tx_ctx *ctx,
+ struct vmxnet3_adapter *adapter)
+{
+ struct Vmxnet3_TxDataDesc *tdd;
+
+ if (ctx->mss) {
+ ctx->eth_ip_hdr_size = skb_transport_offset(skb);
+ ctx->l4_hdr_size = ((struct tcphdr *)
+ skb_transport_header(skb))->doff * 4;
+ ctx->copy_size = ctx->eth_ip_hdr_size + ctx->l4_hdr_size;
+ } else {
+ unsigned int pull_size;
+
+ if (skb->ip_summed == CHECKSUM_PARTIAL) {
+ ctx->eth_ip_hdr_size = skb_transport_offset(skb);
+
+ if (ctx->ipv4) {
+ struct iphdr *iph = (struct iphdr *)
+ skb_network_header(skb);
+ if (iph->protocol == IPPROTO_TCP) {
+ pull_size = ctx->eth_ip_hdr_size +
+ sizeof(struct tcphdr);
+
+ if (unlikely(!pskb_may_pull(skb,
+ pull_size))) {
+ goto err;
+ }
+ ctx->l4_hdr_size = ((struct tcphdr *)
+ skb_transport_header(skb))->doff * 4;
+ } else if (iph->protocol == IPPROTO_UDP) {
+ ctx->l4_hdr_size =
+ sizeof(struct udphdr);
+ } else {
+ ctx->l4_hdr_size = 0;
+ }
+ } else {
+ /* for simplicity, don't copy L4 headers */
+ ctx->l4_hdr_size = 0;
+ }
+ ctx->copy_size = ctx->eth_ip_hdr_size +
+ ctx->l4_hdr_size;
+ } else {
+ ctx->eth_ip_hdr_size = 0;
+ ctx->l4_hdr_size = 0;
+ /* copy as much as allowed */
+ ctx->copy_size = min((unsigned int)VMXNET3_HDR_COPY_SIZE
+ , skb_headlen(skb));
+ }
+
+ /* make sure headers are accessible directly */
+ if (unlikely(!pskb_may_pull(skb, ctx->copy_size)))
+ goto err;
+ }
+
+ if (unlikely(ctx->copy_size > VMXNET3_HDR_COPY_SIZE)) {
+ tq->stats.oversized_hdr++;
+ ctx->copy_size = 0;
+ return 0;
+ }
+
+ tdd = tq->data_ring.base + tq->tx_ring.next2fill;
+ BUG_ON(ctx->copy_size > skb_headlen(skb));
+
+ memcpy(tdd->data, skb->data, ctx->copy_size);
+ dprintk(KERN_ERR "copy %u bytes to dataRing[%u]\n",
+ ctx->copy_size, tq->tx_ring.next2fill);
+ return 1;
+
+err:
+ return -1;
+}
+
+
+static void
+vmxnet3_prepare_tso(struct sk_buff *skb,
+ struct vmxnet3_tx_ctx *ctx)
+{
+ struct tcphdr *tcph = (struct tcphdr *)skb_transport_header(skb);
+ if (ctx->ipv4) {
+ struct iphdr *iph = (struct iphdr *)skb_network_header(skb);
+ iph->check = 0;
+ tcph->check = ~csum_tcpudp_magic(iph->saddr, iph->daddr, 0,
+ IPPROTO_TCP, 0);
+ } else {
+ struct ipv6hdr *iph = (struct ipv6hdr *)skb_network_header(skb);
+ tcph->check = ~csum_ipv6_magic(&iph->saddr, &iph->daddr, 0,
+ IPPROTO_TCP, 0);
+ }
+}
+
+
+/*
+ * Transmits a pkt thru a given tq
+ * Returns:
+ * NETDEV_TX_OK: descriptors are setup successfully
+ * NETDEV_TX_OK: error occured, the pkt is dropped
+ * NETDEV_TX_BUSY: tx ring is full, queue is stopped
+ *
+ * Side-effects:
+ * 1. tx ring may be changed
+ * 2. tq stats may be updated accordingly
+ * 3. shared->txNumDeferred may be updated
+ */
+
+static int
+vmxnet3_tq_xmit(struct sk_buff *skb, struct vmxnet3_tx_queue *tq,
+ struct vmxnet3_adapter *adapter, struct net_device *netdev)
+{
+ int ret;
+ u32 count;
+ unsigned long flags;
+ struct vmxnet3_tx_ctx ctx;
+ union Vmxnet3_GenericDesc *gdesc;
+
+ /* conservatively estimate # of descriptors to use */
+ count = VMXNET3_TXD_NEEDED(skb_headlen(skb)) +
+ skb_shinfo(skb)->nr_frags + 1;
+
+ ctx.ipv4 = (skb->protocol == __constant_ntohs(ETH_P_IP));
+
+ ctx.mss = skb_shinfo(skb)->gso_size;
+ if (ctx.mss) {
+ if (skb_header_cloned(skb)) {
+ if (unlikely(pskb_expand_head(skb, 0, 0,
+ GFP_ATOMIC) != 0)) {
+ tq->stats.drop_tso++;
+ goto drop_pkt;
+ }
+ tq->stats.copy_skb_header++;
+ }
+ vmxnet3_prepare_tso(skb, &ctx);
+ } else {
+ if (unlikely(count > VMXNET3_MAX_TXD_PER_PKT)) {
+
+ /* non-tso pkts must not use more than
+ * VMXNET3_MAX_TXD_PER_PKT entries
+ */
+ if (skb_linearize(skb) != 0) {
+ tq->stats.drop_too_many_frags++;
+ goto drop_pkt;
+ }
+ tq->stats.linearized++;
+
+ /* recalculate the # of descriptors to use */
+ count = VMXNET3_TXD_NEEDED(skb_headlen(skb)) + 1;
+ }
+ }
+
+ ret = vmxnet3_parse_and_copy_hdr(skb, tq, &ctx, adapter);
+ if (ret >= 0) {
+ BUG_ON(ret <= 0 && ctx.copy_size != 0);
+ /* hdrs parsed, check against other limits */
+ if (ctx.mss) {
+ if (unlikely(ctx.eth_ip_hdr_size + ctx.l4_hdr_size >
+ VMXNET3_MAX_TX_BUF_SIZE)) {
+ goto hdr_too_big;
+ }
+ } else {
+ if (skb->ip_summed == CHECKSUM_PARTIAL) {
+ if (unlikely(ctx.eth_ip_hdr_size +
+ skb->csum_offset >
+ VMXNET3_MAX_CSUM_OFFSET)) {
+ goto hdr_too_big;
+ }
+ }
+ }
+ } else {
+ tq->stats.drop_hdr_inspect_err++;
+ goto drop_pkt;
+ }
+
+ spin_lock_irqsave(&tq->tx_lock, flags);
+
+ if (count > vmxnet3_cmd_ring_desc_avail(&tq->tx_ring)) {
+ tq->stats.tx_ring_full++;
+ dprintk(KERN_ERR "tx queue stopped on %s, next2comp %u"
+ " next2fill %u\n", adapter->netdev->name,
+ tq->tx_ring.next2comp, tq->tx_ring.next2fill);
+
+ vmxnet3_tq_stop(tq, adapter);
+ spin_unlock_irqrestore(&tq->tx_lock, flags);
+ return NETDEV_TX_BUSY;
+ }
+
+ /* fill tx descs related to addr & len */
+ vmxnet3_map_pkt(skb, &ctx, tq, adapter->pdev, adapter);
+
+ /* setup the EOP desc */
+ ctx.eop_txd->dword[3] = VMXNET3_TXD_CQ | VMXNET3_TXD_EOP;
+
+ /* setup the SOP desc */
+ gdesc = ctx.sop_txd;
+ if (ctx.mss) {
+ gdesc->txd.hlen = ctx.eth_ip_hdr_size + ctx.l4_hdr_size;
+ gdesc->txd.om = VMXNET3_OM_TSO;
+ gdesc->txd.msscof = ctx.mss;
+ tq->shared->txNumDeferred += (skb->len - gdesc->txd.hlen +
+ ctx.mss - 1) / ctx.mss;
+ } else {
+ if (skb->ip_summed == CHECKSUM_PARTIAL) {
+ gdesc->txd.hlen = ctx.eth_ip_hdr_size;
+ gdesc->txd.om = VMXNET3_OM_CSUM;
+ gdesc->txd.msscof = ctx.eth_ip_hdr_size +
+ skb->csum_offset;
+ } else {
+ gdesc->txd.om = 0;
+ gdesc->txd.msscof = 0;
+ }
+ tq->shared->txNumDeferred++;
+ }
+
+ if (vlan_tx_tag_present(skb)) {
+ gdesc->txd.ti = 1;
+ gdesc->txd.tci = vlan_tx_tag_get(skb);
+ }
+
+ wmb();
+
+ /* finally flips the GEN bit of the SOP desc */
+ gdesc->dword[2] ^= VMXNET3_TXD_GEN;
+ dprintk(KERN_ERR "txd[%u]: SOP 0x%Lx 0x%x 0x%x\n",
+ (u32)((union Vmxnet3_GenericDesc *)ctx.sop_txd -
+ tq->tx_ring.base), gdesc->txd.addr, gdesc->dword[2],
+ gdesc->dword[3]);
+
+ spin_unlock_irqrestore(&tq->tx_lock, flags);
+
+ if (tq->shared->txNumDeferred >= tq->shared->txThreshold) {
+ tq->shared->txNumDeferred = 0;
+ VMXNET3_WRITE_BAR0_REG(adapter, VMXNET3_REG_TXPROD,
+ tq->tx_ring.next2fill);
+ }
+ netdev->trans_start = jiffies;
+
+ return NETDEV_TX_OK;
+
+hdr_too_big:
+ tq->stats.drop_oversized_hdr++;
+drop_pkt:
+ tq->stats.drop_total++;
+ dev_kfree_skb(skb);
+ return NETDEV_TX_OK;
+}
+
+
+static int
+vmxnet3_xmit_frame(struct sk_buff *skb, struct net_device *netdev)
+{
+ struct vmxnet3_adapter *adapter = netdev_priv(netdev);
+ struct vmxnet3_tx_queue *tq = &adapter->tx_queue;
+
+ return vmxnet3_tq_xmit(skb, tq, adapter, netdev);
+}
+
+
+static void
+vmxnet3_rx_csum(struct vmxnet3_adapter *adapter,
+ struct sk_buff *skb,
+ union Vmxnet3_GenericDesc *gdesc)
+{
+ if (!gdesc->rcd.cnc && adapter->rxcsum) {
+ /* typical case: TCP/UDP over IP and both csums are correct */
+ if ((gdesc->dword[3] & VMXNET3_RCD_CSUM_OK) ==
+ VMXNET3_RCD_CSUM_OK) {
+ skb->ip_summed = CHECKSUM_UNNECESSARY;
+ BUG_ON(!(gdesc->rcd.tcp || gdesc->rcd.udp));
+ BUG_ON(!(gdesc->rcd.v4 || gdesc->rcd.v6));
+ BUG_ON(gdesc->rcd.frg);
+ } else {
+ if (gdesc->rcd.csum) {
+ skb->csum = htons(gdesc->rcd.csum);
+ skb->ip_summed = CHECKSUM_PARTIAL;
+ } else {
+ skb->ip_summed = CHECKSUM_NONE;
+ }
+ }
+ } else {
+ skb->ip_summed = CHECKSUM_NONE;
+ }
+}
+
+
+static void
+vmxnet3_rx_error(struct vmxnet3_rx_queue *rq, struct Vmxnet3_RxCompDesc *rcd,
+ struct vmxnet3_rx_ctx *ctx, struct vmxnet3_adapter *adapter)
+{
+ rq->stats.drop_err++;
+ if (!rcd->fcs)
+ rq->stats.drop_fcs++;
+
+ rq->stats.drop_total++;
+
+ /*
+ * We do not unmap and chain the rx buffer to the skb.
+ * We basically pretend this buffer is not used and will be recycled
+ * by vmxnet3_rq_alloc_rx_buf()
+ */
+
+ /*
+ * ctx->skb may be NULL if this is the first and the only one
+ * desc for the pkt
+ */
+ if (ctx->skb)
+ dev_kfree_skb_irq(ctx->skb);
+
+ ctx->skb = NULL;
+}
+
+
+static int
+vmxnet3_rq_rx_complete(struct vmxnet3_rx_queue *rq,
+ struct vmxnet3_adapter *adapter, int quota)
+{
+static u32 rxprod_reg[2] = {VMXNET3_REG_RXPROD, VMXNET3_REG_RXPROD2};
+ u32 num_rxd = 0;
+ struct Vmxnet3_RxCompDesc *rcd;
+ struct vmxnet3_rx_ctx *ctx = &rq->rx_ctx;
+
+ rcd = &rq->comp_ring.base[rq->comp_ring.next2proc].rcd;
+ while (rcd->gen == rq->comp_ring.gen) {
+ struct vmxnet3_rx_buf_info *rbi;
+ struct sk_buff *skb;
+ int num_to_alloc;
+ struct Vmxnet3_RxDesc *rxd;
+ u32 idx, ring_idx;
+
+ if (num_rxd >= quota) {
+ /* we may stop even before we see the EOP desc of
+ * the current pkt
+ */
+ break;
+ }
+ num_rxd++;
+
+ idx = rcd->rxdIdx;
+ ring_idx = rcd->rqID == rq->qid ? 0 : 1;
+
+ rxd = &rq->rx_ring[ring_idx].base[idx].rxd;
+ rbi = rq->buf_info[ring_idx] + idx;
+
+ BUG_ON(rcd->len > rxd->len);
+ BUG_ON(rxd->addr != rbi->dma_addr || rxd->len != rbi->len);
+
+ if (unlikely(rcd->eop && rcd->err)) {
+ vmxnet3_rx_error(rq, rcd, ctx, adapter);
+ goto rcd_done;
+ }
+
+ if (rcd->sop) { /* first buf of the pkt */
+ BUG_ON(rxd->btype != VMXNET3_RXD_BTYPE_HEAD ||
+ rcd->rqID != rq->qid);
+
+ BUG_ON(rbi->buf_type != VMXNET3_RX_BUF_SKB);
+ BUG_ON(ctx->skb != NULL || rbi->skb == NULL);
+
+ if (unlikely(rcd->len == 0)) {
+ /* Pretend the rx buffer is skipped. */
+ BUG_ON(!(rcd->sop && rcd->eop));
+ dprintk(KERN_ERR "rxRing[%u][%u] 0 length\n",
+ ring_idx, idx);
+ goto rcd_done;
+ }
+
+ ctx->skb = rbi->skb;
+ rbi->skb = NULL;
+
+ pci_unmap_single(adapter->pdev, rbi->dma_addr, rbi->len,
+ PCI_DMA_FROMDEVICE);
+
+ skb_put(ctx->skb, rcd->len);
+ } else {
+ BUG_ON(ctx->skb == NULL);
+ /* non SOP buffer must be type 1 in most cases */
+ if (rbi->buf_type == VMXNET3_RX_BUF_PAGE) {
+ BUG_ON(rxd->btype != VMXNET3_RXD_BTYPE_BODY);
+
+ if (rcd->len) {
+ pci_unmap_page(adapter->pdev,
+ rbi->dma_addr, rbi->len,
+ PCI_DMA_FROMDEVICE);
+
+ vmxnet3_append_frag(ctx->skb, rcd, rbi);
+ rbi->page = NULL;
+ }
+ } else {
+ /*
+ * The only time a non-SOP buffer is type 0 is
+ * when it's EOP and error flag is raised, which
+ * has already been handled.
+ */
+ BUG_ON(true);
+ }
+ }
+
+ skb = ctx->skb;
+ if (rcd->eop) {
+ skb->len += skb->data_len;
+ skb->truesize += skb->data_len;
+
+ vmxnet3_rx_csum(adapter, skb,
+ (union Vmxnet3_GenericDesc *)rcd);
+ skb->protocol = eth_type_trans(skb, adapter->netdev);
+
+ if (unlikely(adapter->vlan_grp && rcd->ts)) {
+ vlan_hwaccel_receive_skb(skb,
+ adapter->vlan_grp, rcd->tci);
+ } else {
+ netif_receive_skb(skb);
+ }
+
+ adapter->netdev->last_rx = jiffies;
+ ctx->skb = NULL;
+ }
+
+rcd_done:
+ /* device may skip some rx descs */
+ rq->rx_ring[ring_idx].next2comp = idx;
+ VMXNET3_INC_RING_IDX_ONLY(rq->rx_ring[ring_idx].next2comp,
+ rq->rx_ring[ring_idx].size);
+
+ /* refill rx buffers frequently to avoid starving the h/w */
+ num_to_alloc = vmxnet3_cmd_ring_desc_avail(rq->rx_ring +
+ ring_idx);
+ if (unlikely(num_to_alloc > VMXNET3_RX_ALLOC_THRESHOLD(rq,
+ ring_idx, adapter))) {
+ vmxnet3_rq_alloc_rx_buf(rq, ring_idx, num_to_alloc,
+ adapter);
+
+ /* if needed, update the register */
+ if (unlikely(rq->shared->updateRxProd)) {
+ VMXNET3_WRITE_BAR0_REG(adapter,
+ rxprod_reg[ring_idx] + rq->qid * 8,
+ rq->rx_ring[ring_idx].next2fill);
+ rq->uncommitted[ring_idx] = 0;
+ }
+ }
+
+ vmxnet3_comp_ring_adv_next2proc(&rq->comp_ring);
+ rcd = &rq->comp_ring.base[rq->comp_ring.next2proc].rcd;
+ }
+
+ return num_rxd;
+}
+
+
+static void
+vmxnet3_rq_cleanup(struct vmxnet3_rx_queue *rq,
+ struct vmxnet3_adapter *adapter)
+{
+ u32 i, ring_idx;
+ struct Vmxnet3_RxDesc *rxd;
+
+ for (ring_idx = 0; ring_idx < 2; ring_idx++) {
+ for (i = 0; i < rq->rx_ring[ring_idx].size; i++) {
+ rxd = &rq->rx_ring[ring_idx].base[i].rxd;
+
+ if (rxd->btype == VMXNET3_RXD_BTYPE_HEAD &&
+ rq->buf_info[ring_idx][i].skb) {
+ pci_unmap_single(adapter->pdev, rxd->addr,
+ rxd->len, PCI_DMA_FROMDEVICE);
+ dev_kfree_skb(rq->buf_info[ring_idx][i].skb);
+ rq->buf_info[ring_idx][i].skb = NULL;
+ } else if (rxd->btype == VMXNET3_RXD_BTYPE_BODY &&
+ rq->buf_info[ring_idx][i].page) {
+ pci_unmap_page(adapter->pdev, rxd->addr,
+ rxd->len, PCI_DMA_FROMDEVICE);
+ put_page(rq->buf_info[ring_idx][i].page);
+ rq->buf_info[ring_idx][i].page = NULL;
+ }
+ }
+
+ rq->rx_ring[ring_idx].gen = VMXNET3_INIT_GEN;
+ rq->rx_ring[ring_idx].next2fill =
+ rq->rx_ring[ring_idx].next2comp = 0;
+ rq->uncommitted[ring_idx] = 0;
+ }
+
+ rq->comp_ring.gen = VMXNET3_INIT_GEN;
+ rq->comp_ring.next2proc = 0;
+}
+
+
+void vmxnet3_rq_destroy(struct vmxnet3_rx_queue *rq,
+ struct vmxnet3_adapter *adapter)
+{
+ int i;
+ int j;
+
+ /* all rx buffers must have already been freed */
+ for (i = 0; i < 2; i++) {
+ if (rq->buf_info[i]) {
+ for (j = 0; j < rq->rx_ring[i].size; j++)
+ BUG_ON(rq->buf_info[i][j].page != NULL);
+ }
+ }
+
+
+ kfree(rq->buf_info[0]);
+
+ for (i = 0; i < 2; i++) {
+ if (rq->rx_ring[i].base) {
+ pci_free_consistent(adapter->pdev, rq->rx_ring[i].size
+ * sizeof(struct Vmxnet3_RxDesc),
+ rq->rx_ring[i].base,
+ rq->rx_ring[i].basePA);
+ rq->rx_ring[i].base = NULL;
+ }
+ rq->buf_info[i] = NULL;
+ }
+
+ if (rq->comp_ring.base) {
+ pci_free_consistent(adapter->pdev, rq->comp_ring.size *
+ sizeof(struct Vmxnet3_RxCompDesc),
+ rq->comp_ring.base, rq->comp_ring.basePA);
+ rq->comp_ring.base = NULL;
+ }
+}
+
+
+static int
+vmxnet3_rq_init(struct vmxnet3_rx_queue *rq,
+ struct vmxnet3_adapter *adapter)
+{
+ int i;
+
+ BUG_ON(adapter->rx_buf_per_pkt <= 0 ||
+ rq->rx_ring[0].size % adapter->rx_buf_per_pkt != 0);
+
+ /* initialize buf_info */
+ for (i = 0; i < rq->rx_ring[0].size; i++) {
+ BUG_ON(rq->buf_info[0][i].skb != NULL);
+
+ /* 1st buf for a pkt is skbuff */
+ if (i % adapter->rx_buf_per_pkt == 0) {
+ rq->buf_info[0][i].buf_type = VMXNET3_RX_BUF_SKB;
+ rq->buf_info[0][i].len = adapter->skb_buf_size;
+ } else { /* subsequent bufs for a pkt is frag */
+ rq->buf_info[0][i].buf_type = VMXNET3_RX_BUF_PAGE;
+ rq->buf_info[0][i].len = PAGE_SIZE;
+ }
+ }
+ for (i = 0; i < rq->rx_ring[1].size; i++) {
+ BUG_ON(rq->buf_info[1][i].page != NULL);
+ rq->buf_info[1][i].buf_type = VMXNET3_RX_BUF_PAGE;
+ rq->buf_info[1][i].len = PAGE_SIZE;
+ }
+
+ /* reset internal state and allocate buffers for both rings */
+ for (i = 0; i < 2; i++) {
+ rq->rx_ring[i].next2fill = rq->rx_ring[i].next2comp = 0;
+ rq->uncommitted[i] = 0;
+
+ memset(rq->rx_ring[i].base, 0, rq->rx_ring[i].size *
+ sizeof(struct Vmxnet3_RxDesc));
+ rq->rx_ring[i].gen = VMXNET3_INIT_GEN;
+ }
+ if (vmxnet3_rq_alloc_rx_buf(rq, 0, rq->rx_ring[0].size - 1,
+ adapter) == 0) {
+ /* at least has 1 rx buffer for the 1st ring */
+ return -ENOMEM;
+ }
+ vmxnet3_rq_alloc_rx_buf(rq, 1, rq->rx_ring[1].size - 1, adapter);
+
+ /* reset the comp ring */
+ rq->comp_ring.next2proc = 0;
+ memset(rq->comp_ring.base, 0, rq->comp_ring.size *
+ sizeof(struct Vmxnet3_RxCompDesc));
+ rq->comp_ring.gen = VMXNET3_INIT_GEN;
+
+ /* reset rxctx */
+ rq->rx_ctx.skb = NULL;
+
+ /* stats are not reset */
+ return 0;
+}
+
+
+static int
+vmxnet3_rq_create(struct vmxnet3_rx_queue *rq, struct vmxnet3_adapter *adapter)
+{
+ int i;
+ size_t sz;
+ struct vmxnet3_rx_buf_info *bi;
+
+ BUG_ON(rq->rx_ring[0].size % adapter->rx_buf_per_pkt != 0);
+
+ for (i = 0; i < 2; i++) {
+ BUG_ON((rq->rx_ring[i].size & VMXNET3_RING_SIZE_MASK) != 0);
+ BUG_ON(rq->rx_ring[i].base != NULL);
+
+ sz = rq->rx_ring[i].size * sizeof(struct Vmxnet3_RxDesc);
+ rq->rx_ring[i].base = pci_alloc_consistent(adapter->pdev, sz,
+ &rq->rx_ring[i].basePA);
+ if (!rq->rx_ring[i].base) {
+ printk(KERN_ERR "%s: failed to allocate rx ring %d\n",
+ adapter->netdev->name, i);
+ goto err;
+ }
+ }
+
+ sz = rq->comp_ring.size * sizeof(struct Vmxnet3_RxCompDesc);
+ BUG_ON(rq->comp_ring.base != NULL);
+ rq->comp_ring.base = pci_alloc_consistent(adapter->pdev, sz,
+ &rq->comp_ring.basePA);
+ if (!rq->comp_ring.base) {
+ printk(KERN_ERR "%s: failed to allocate rx comp ring\n",
+ adapter->netdev->name);
+ goto err;
+ }
+
+ BUG_ON(rq->buf_info[0] || rq->buf_info[1]);
+ sz = sizeof(struct vmxnet3_rx_buf_info) * (rq->rx_ring[0].size +
+ rq->rx_ring[1].size);
+ bi = kmalloc(sz, GFP_KERNEL);
+ if (!bi) {
+ printk(KERN_ERR "%s: failed to allocate rx bufinfo\n",
+ adapter->netdev->name);
+ goto err;
+ }
+ memset(bi, 0, sz);
+ rq->buf_info[0] = bi;
+ rq->buf_info[1] = bi + rq->rx_ring[0].size;
+
+ return 0;
+
+err:
+ vmxnet3_rq_destroy(rq, adapter);
+ return -ENOMEM;
+}
+
+
+static void
+vmxnet3_do_poll(struct vmxnet3_adapter *adapter, int budget, int *txd_done,
+ int *rxd_done)
+{
+ if (unlikely(adapter->shared->ecr))
+ vmxnet3_process_events(adapter);
+
+ *txd_done = vmxnet3_tq_tx_complete(&adapter->tx_queue, adapter);
+ *rxd_done = vmxnet3_rq_rx_complete(&adapter->rx_queue, adapter, budget);
+}
+
+
+static int
+vmxnet3_poll(struct napi_struct *napi, int budget)
+{
+ struct vmxnet3_adapter *adapter = container_of(napi,
+ struct vmxnet3_adapter, napi);
+ int rxd_done, txd_done;
+
+ vmxnet3_do_poll(adapter, budget, &txd_done, &rxd_done);
+
+ if (rxd_done < budget) {
+ napi_complete(napi);
+ vmxnet3_enable_intr(adapter, 0);
+ }
+ return rxd_done;
+}
+
+
+/* Interrupt handler for vmxnet3 */
+static irqreturn_t
+vmxnet3_intr(int irq, void *dev_id)
+{
+ struct net_device *dev = dev_id;
+ struct vmxnet3_adapter *adapter = netdev_priv(dev);
+
+ if (unlikely(adapter->intr.type == VMXNET3_IT_INTX)) {
+ u32 icr = VMXNET3_READ_BAR1_REG(adapter, VMXNET3_REG_ICR);
+ if (unlikely(icr == 0))
+ /* not ours */
+ return IRQ_NONE;
+ }
+
+
+ /* disable intr if needed */
+ if (adapter->intr.mask_mode == VMXNET3_IMM_ACTIVE)
+ vmxnet3_disable_intr(adapter, 0);
+
+ napi_schedule(&adapter->napi);
+
+ return IRQ_HANDLED;
+}
+
+#ifdef CONFIG_NET_POLL_CONTROLLER
+
+
+/* netpoll callback. */
+static void
+vmxnet3_netpoll(struct net_device *netdev)
+{
+ struct vmxnet3_adapter *adapter = netdev_priv(netdev);
+ int irq;
+
+ if (adapter->intr.type == VMXNET3_IT_MSIX)
+ irq = adapter->intr.msix_entries[0].vector;
+ else
+ irq = adapter->pdev->irq;
+
+ disable_irq(irq);
+ vmxnet3_intr(irq, netdev);
+ enable_irq(irq);
+}
+#endif
+
+static int
+vmxnet3_request_irqs(struct vmxnet3_adapter *adapter)
+{
+ int err;
+
+ if (adapter->intr.type == VMXNET3_IT_MSIX) {
+ /* we only use 1 MSI-X vector */
+ err = request_irq(adapter->intr.msix_entries[0].vector,
+ vmxnet3_intr, 0, adapter->netdev->name,
+ adapter->netdev);
+ } else if (adapter->intr.type == VMXNET3_IT_MSI) {
+ err = request_irq(adapter->pdev->irq, vmxnet3_intr, 0,
+ adapter->netdev->name, adapter->netdev);
+ } else {
+ BUG_ON(adapter->intr.type != VMXNET3_IT_INTX);
+
+ err = request_irq(adapter->pdev->irq, vmxnet3_intr,
+ IRQF_SHARED, adapter->netdev->name,
+ adapter->netdev);
+ }
+
+ if (err)
+ printk(KERN_ERR "Failed to request irq %s (intr type:%d), error"
+ ":%d\n", adapter->netdev->name, adapter->intr.type, err);
+
+
+ if (!err) {
+ int i;
+ /* init our intr settings */
+ for (i = 0; i < adapter->intr.num_intrs; i++)
+ adapter->intr.mod_levels[i] = UPT1_IML_ADAPTIVE;
+
+ /* next setup intr index for all intr sources */
+ adapter->tx_queue.comp_ring.intr_idx = 0;
+ adapter->rx_queue.comp_ring.intr_idx = 0;
+ adapter->intr.event_intr_idx = 0;
+
+ printk(KERN_INFO "%s: intr type %u, mode %u, %u vectors "
+ "allocated\n", adapter->netdev->name, adapter->intr.type,
+ adapter->intr.mask_mode, adapter->intr.num_intrs);
+ }
+
+ return err;
+}
+
+
+static void
+vmxnet3_free_irqs(struct vmxnet3_adapter *adapter)
+{
+ BUG_ON(adapter->intr.type == VMXNET3_IT_AUTO ||
+ adapter->intr.num_intrs <= 0);
+
+ switch (adapter->intr.type) {
+ case VMXNET3_IT_MSIX:
+ {
+ int i;
+
+ for (i = 0; i < adapter->intr.num_intrs; i++)
+ free_irq(adapter->intr.msix_entries[i].vector,
+ adapter->netdev);
+ break;
+ }
+ case VMXNET3_IT_MSI:
+ free_irq(adapter->pdev->irq, adapter->netdev);
+ break;
+ case VMXNET3_IT_INTX:
+ free_irq(adapter->pdev->irq, adapter->netdev);
+ break;
+ default:
+ BUG_ON(true);
+ }
+}
+
+
+static void
+vmxnet3_vlan_rx_register(struct net_device *netdev, struct vlan_group *grp)
+{
+ struct vmxnet3_adapter *adapter = netdev_priv(netdev);
+ struct Vmxnet3_DriverShared *shared = adapter->shared;
+ u32 *vfTable = adapter->shared->devRead.rxFilterConf.vfTable;
+
+ if (grp) {
+ /* add vlan rx stripping. */
+ if (adapter->netdev->features & NETIF_F_HW_VLAN_RX) {
+ int i;
+ struct Vmxnet3_DSDevRead *devRead = &shared->devRead;
+ adapter->vlan_grp = grp;
+
+ /* update FEATURES to device */
+ devRead->misc.uptFeatures |= UPT1_F_RXVLAN;
+ VMXNET3_WRITE_BAR1_REG(adapter, VMXNET3_REG_CMD,
+ VMXNET3_CMD_UPDATE_FEATURE);
+ /*
+ * Clear entire vfTable; then enable untagged pkts.
+ * Note: setting one entry in vfTable to non-zero turns
+ * on VLAN rx filtering.
+ */
+ for (i = 0; i < VMXNET3_VFT_SIZE; i++)
+ vfTable[i] = 0;
+
+ VMXNET3_SET_VFTABLE_ENTRY(vfTable, 0);
+ VMXNET3_WRITE_BAR1_REG(adapter, VMXNET3_REG_CMD,
+ VMXNET3_CMD_UPDATE_VLAN_FILTERS);
+ } else {
+ printk(KERN_ERR "%s: vlan_rx_register when device has "
+ "no NETIF_F_HW_VLAN_RX\n", netdev->name);
+ }
+ } else {
+ /* remove vlan rx stripping. */
+ struct Vmxnet3_DSDevRead *devRead = &shared->devRead;
+ adapter->vlan_grp = NULL;
+
+ if (devRead->misc.uptFeatures & UPT1_F_RXVLAN) {
+ int i;
+
+ for (i = 0; i < VMXNET3_VFT_SIZE; i++) {
+ /* clear entire vfTable; this also disables
+ * VLAN rx filtering
+ */
+ vfTable[i] = 0;
+ }
+ VMXNET3_WRITE_BAR1_REG(adapter, VMXNET3_REG_CMD,
+ VMXNET3_CMD_UPDATE_VLAN_FILTERS);
+
+ /* update FEATURES to device */
+ devRead->misc.uptFeatures &= ~UPT1_F_RXVLAN;
+ VMXNET3_WRITE_BAR1_REG(adapter, VMXNET3_REG_CMD,
+ VMXNET3_CMD_UPDATE_FEATURE);
+ }
+ }
+}
+
+
+static void
+vmxnet3_restore_vlan(struct vmxnet3_adapter *adapter)
+{
+ if (adapter->vlan_grp) {
+ u16 vid;
+ u32 *vfTable = adapter->shared->devRead.rxFilterConf.vfTable;
+ bool activeVlan = false;
+
+ for (vid = 0; vid < VLAN_GROUP_ARRAY_LEN; vid++) {
+ if (vlan_group_get_device(adapter->vlan_grp, vid)) {
+ VMXNET3_SET_VFTABLE_ENTRY(vfTable, vid);
+ activeVlan = true;
+ }
+ }
+ if (activeVlan) {
+ /* continue to allow untagged pkts */
+ VMXNET3_SET_VFTABLE_ENTRY(vfTable, 0);
+ }
+ }
+}
+
+
+static void
+vmxnet3_vlan_rx_add_vid(struct net_device *netdev, u16 vid)
+{
+ struct vmxnet3_adapter *adapter = netdev_priv(netdev);
+ u32 *vfTable = adapter->shared->devRead.rxFilterConf.vfTable;
+
+ VMXNET3_SET_VFTABLE_ENTRY(vfTable, vid);
+ VMXNET3_WRITE_BAR1_REG(adapter, VMXNET3_REG_CMD,
+ VMXNET3_CMD_UPDATE_VLAN_FILTERS);
+}
+
+
+static void
+vmxnet3_vlan_rx_kill_vid(struct net_device *netdev, u16 vid)
+{
+ struct vmxnet3_adapter *adapter = netdev_priv(netdev);
+ u32 *vfTable = adapter->shared->devRead.rxFilterConf.vfTable;
+
+ VMXNET3_CLEAR_VFTABLE_ENTRY(vfTable, vid);
+ VMXNET3_WRITE_BAR1_REG(adapter, VMXNET3_REG_CMD,
+ VMXNET3_CMD_UPDATE_VLAN_FILTERS);
+}
+
+
+static u8 *
+vmxnet3_copy_mc(struct net_device *netdev)
+{
+ u8 *buf = NULL;
+ u32 sz = netdev->mc_count * ETH_ALEN;
+
+ /* struct Vmxnet3_RxFilterConf.mfTableLen is u16. */
+ if (sz <= 0xffff) {
+ /* We may be called with BH disabled */
+ buf = kmalloc(sz, GFP_ATOMIC);
+ if (buf) {
+ int i;
+ struct dev_mc_list *mc = netdev->mc_list;
+
+ for (i = 0; i < netdev->mc_count; i++) {
+ BUG_ON(!mc);
+ memcpy(buf + i * ETH_ALEN, mc->dmi_addr,
+ ETH_ALEN);
+ mc = mc->next;
+ }
+ }
+ }
+ return buf;
+}
+
+
+static void
+vmxnet3_set_mc(struct net_device *netdev)
+{
+ struct vmxnet3_adapter *adapter = netdev_priv(netdev);
+ struct Vmxnet3_RxFilterConf *rxConf =
+ &adapter->shared->devRead.rxFilterConf;
+ u8 *new_table = NULL;
+ u32 new_mode = VMXNET3_RXM_UCAST;
+
+ if (netdev->flags & IFF_PROMISC)
+ new_mode |= VMXNET3_RXM_PROMISC;
+
+ if (netdev->flags & IFF_BROADCAST)
+ new_mode |= VMXNET3_RXM_BCAST;
+
+ if (netdev->flags & IFF_ALLMULTI)
+ new_mode |= VMXNET3_RXM_ALL_MULTI;
+ else
+ if (netdev->mc_count > 0) {
+ new_table = vmxnet3_copy_mc(netdev);
+ if (new_table) {
+ new_mode |= VMXNET3_RXM_MCAST;
+ rxConf->mfTableLen = netdev->mc_count *
+ ETH_ALEN;
+ rxConf->mfTablePA = virt_to_phys(new_table);
+ } else {
+ printk(KERN_INFO "%s: failed to copy mcast list"
+ ", setting ALL_MULTI\n", netdev->name);
+ new_mode |= VMXNET3_RXM_ALL_MULTI;
+ }
+ }
+
+
+ if (!(new_mode & VMXNET3_RXM_MCAST)) {
+ rxConf->mfTableLen = 0;
+ rxConf->mfTablePA = 0;
+ }
+
+ if (new_mode != rxConf->rxMode) {
+ rxConf->rxMode = new_mode;
+ VMXNET3_WRITE_BAR1_REG(adapter, VMXNET3_REG_CMD,
+ VMXNET3_CMD_UPDATE_RX_MODE);
+ }
+
+ VMXNET3_WRITE_BAR1_REG(adapter, VMXNET3_REG_CMD,
+ VMXNET3_CMD_UPDATE_MAC_FILTERS);
+
+ kfree(new_table);
+}
+
+
+/*
+ * Set up driver_shared based on settings in adapter.
+ */
+
+static void
+vmxnet3_setup_driver_shared(struct vmxnet3_adapter *adapter)
+{
+ struct Vmxnet3_DriverShared *shared = adapter->shared;
+ struct Vmxnet3_DSDevRead *devRead = &shared->devRead;
+ struct Vmxnet3_TxQueueConf *tqc;
+ struct Vmxnet3_RxQueueConf *rqc;
+ int i;
+
+ memset(shared, 0, sizeof(*shared));
+
+ /* driver settings */
+ shared->magic = VMXNET3_REV1_MAGIC;
+ devRead->misc.driverInfo.version = VMXNET3_DRIVER_VERSION_NUM;
+ devRead->misc.driverInfo.gos.gosBits = (sizeof(void *) == 4 ?
+ VMXNET3_GOS_BITS_32 : VMXNET3_GOS_BITS_64);
+ devRead->misc.driverInfo.gos.gosType = VMXNET3_GOS_TYPE_LINUX;
+ devRead->misc.driverInfo.vmxnet3RevSpt = 1;
+ devRead->misc.driverInfo.uptVerSpt = 1;
+
+ devRead->misc.ddPA = virt_to_phys(adapter);
+ devRead->misc.ddLen = sizeof(struct vmxnet3_adapter);
+
+ /* set up feature flags */
+ if (adapter->rxcsum)
+ devRead->misc.uptFeatures |= UPT1_F_RXCSUM;
+
+ if (adapter->lro) {
+ devRead->misc.uptFeatures |= UPT1_F_LRO;
+ devRead->misc.maxNumRxSG = 1 + MAX_SKB_FRAGS;
+ }
+ if ((adapter->netdev->features & NETIF_F_HW_VLAN_RX)
+ && adapter->vlan_grp) {
+ devRead->misc.uptFeatures |= UPT1_F_RXVLAN;
+ }
+
+ devRead->misc.mtu = adapter->netdev->mtu;
+ devRead->misc.queueDescPA = adapter->queue_desc_pa;
+ devRead->misc.queueDescLen = sizeof(struct Vmxnet3_TxQueueDesc) +
+ sizeof(struct Vmxnet3_RxQueueDesc);
+
+ /* tx queue settings */
+ BUG_ON(adapter->tx_queue.tx_ring.base == NULL);
+
+ devRead->misc.numTxQueues = 1;
+ tqc = &adapter->tqd_start->conf;
+ tqc->txRingBasePA = adapter->tx_queue.tx_ring.basePA;
+ tqc->dataRingBasePA = adapter->tx_queue.data_ring.basePA;
+ tqc->compRingBasePA = adapter->tx_queue.comp_ring.basePA;
+ tqc->ddPA = virt_to_phys(adapter->tx_queue.buf_info);
+ tqc->txRingSize = adapter->tx_queue.tx_ring.size;
+ tqc->dataRingSize = adapter->tx_queue.data_ring.size;
+ tqc->compRingSize = adapter->tx_queue.comp_ring.size;
+ tqc->ddLen = sizeof(struct vmxnet3_tx_buf_info) *
+ tqc->txRingSize;
+ tqc->intrIdx = adapter->tx_queue.comp_ring.intr_idx;
+
+ /* rx queue settings */
+ devRead->misc.numRxQueues = 1;
+ rqc = &adapter->rqd_start->conf;
+ rqc->rxRingBasePA[0] = adapter->rx_queue.rx_ring[0].basePA;
+ rqc->rxRingBasePA[1] = adapter->rx_queue.rx_ring[1].basePA;
+ rqc->compRingBasePA = adapter->rx_queue.comp_ring.basePA;
+ rqc->ddPA = virt_to_phys(adapter->rx_queue.buf_info);
+ rqc->rxRingSize[0] = adapter->rx_queue.rx_ring[0].size;
+ rqc->rxRingSize[1] = adapter->rx_queue.rx_ring[1].size;
+ rqc->compRingSize = adapter->rx_queue.comp_ring.size;
+ rqc->ddLen = sizeof(struct vmxnet3_rx_buf_info) *
+ (rqc->rxRingSize[0] + rqc->rxRingSize[1]);
+ rqc->intrIdx = adapter->rx_queue.comp_ring.intr_idx;
+
+ /* intr settings */
+ devRead->intrConf.autoMask = adapter->intr.mask_mode ==
+ VMXNET3_IMM_AUTO;
+ devRead->intrConf.numIntrs = adapter->intr.num_intrs;
+ for (i = 0; i < adapter->intr.num_intrs; i++)
+ devRead->intrConf.modLevels[i] = adapter->intr.mod_levels[i];
+
+ devRead->intrConf.eventIntrIdx = adapter->intr.event_intr_idx;
+
+ /* rx filter settings */
+ devRead->rxFilterConf.rxMode = 0;
+ vmxnet3_restore_vlan(adapter);
+ /* the rest are already zeroed */
+}
+
+
+int
+vmxnet3_activate_dev(struct vmxnet3_adapter *adapter)
+{
+ int err;
+ u32 ret;
+
+ dprintk(KERN_ERR "%s: skb_buf_size %d, rx_buf_per_pkt %d, ring sizes"
+ " %u %u %u\n", adapter->netdev->name, adapter->skb_buf_size,
+ adapter->rx_buf_per_pkt, adapter->tx_queue.tx_ring.size,
+ adapter->rx_queue.rx_ring[0].size,
+ adapter->rx_queue.rx_ring[1].size);
+
+ vmxnet3_tq_init(&adapter->tx_queue, adapter);
+ err = vmxnet3_rq_init(&adapter->rx_queue, adapter);
+ if (err) {
+ printk(KERN_ERR "Failed to init rx queue for %s: error %d\n",
+ adapter->netdev->name, err);
+ goto rq_err;
+ }
+
+ err = vmxnet3_request_irqs(adapter);
+ if (err) {
+ printk(KERN_ERR "Failed to setup irq for %s: error %d\n",
+ adapter->netdev->name, err);
+ goto irq_err;
+ }
+
+ vmxnet3_setup_driver_shared(adapter);
+
+ VMXNET3_WRITE_BAR1_REG(adapter, VMXNET3_REG_DSAL,
+ VMXNET3_GET_ADDR_LO(adapter->shared_pa));
+ VMXNET3_WRITE_BAR1_REG(adapter, VMXNET3_REG_DSAH,
+ VMXNET3_GET_ADDR_HI(adapter->shared_pa));
+
+ VMXNET3_WRITE_BAR1_REG(adapter, VMXNET3_REG_CMD,
+ VMXNET3_CMD_ACTIVATE_DEV);
+ ret = VMXNET3_READ_BAR1_REG(adapter, VMXNET3_REG_CMD);
+
+ if (ret != 0) {
+ printk(KERN_ERR "Failed to activate dev %s: error %u\n",
+ adapter->netdev->name, ret);
+ err = -EINVAL;
+ goto activate_err;
+ }
+ VMXNET3_WRITE_BAR0_REG(adapter, VMXNET3_REG_RXPROD,
+ adapter->rx_queue.rx_ring[0].next2fill);
+ VMXNET3_WRITE_BAR0_REG(adapter, VMXNET3_REG_RXPROD2,
+ adapter->rx_queue.rx_ring[1].next2fill);
+
+ /* Apply the rx filter settins last. */
+ vmxnet3_set_mc(adapter->netdev);
+
+ /*
+ * Check link state when first activating device. It will start the
+ * tx queue if the link is up.
+ */
+ vmxnet3_check_link(adapter);
+
+ napi_enable(&adapter->napi);
+ vmxnet3_enable_all_intrs(adapter);
+ clear_bit(VMXNET3_STATE_BIT_QUIESCED, &adapter->state);
+ return 0;
+
+activate_err:
+ VMXNET3_WRITE_BAR1_REG(adapter, VMXNET3_REG_DSAL, 0);
+ VMXNET3_WRITE_BAR1_REG(adapter, VMXNET3_REG_DSAH, 0);
+ vmxnet3_free_irqs(adapter);
+irq_err:
+rq_err:
+ /* free up buffers we allocated */
+ vmxnet3_rq_cleanup(&adapter->rx_queue, adapter);
+ return err;
+}
+
+
+void
+vmxnet3_reset_dev(struct vmxnet3_adapter *adapter)
+{
+ VMXNET3_WRITE_BAR1_REG(adapter, VMXNET3_REG_CMD, VMXNET3_CMD_RESET_DEV);
+}
+
+
+int
+vmxnet3_quiesce_dev(struct vmxnet3_adapter *adapter)
+{
+ if (test_and_set_bit(VMXNET3_STATE_BIT_QUIESCED, &adapter->state)) {
+ printk(KERN_INFO "%s: already quiesced\n",
+ adapter->netdev->name);
+ return 0;
+ }
+
+ VMXNET3_WRITE_BAR1_REG(adapter, VMXNET3_REG_CMD,
+ VMXNET3_CMD_QUIESCE_DEV);
+ vmxnet3_disable_all_intrs(adapter);
+
+ napi_disable(&adapter->napi);
+ netif_tx_disable(adapter->netdev);
+ adapter->link_speed = 0;
+ netif_carrier_off(adapter->netdev);
+
+ vmxnet3_tq_cleanup(&adapter->tx_queue, adapter);
+ vmxnet3_rq_cleanup(&adapter->rx_queue, adapter);
+ vmxnet3_free_irqs(adapter);
+ return 0;
+}
+
+
+static void
+vmxnet3_write_mac_addr(struct vmxnet3_adapter *adapter, u8 *mac)
+{
+ u32 tmp;
+
+ tmp = *(u32 *)mac;
+ VMXNET3_WRITE_BAR1_REG(adapter, VMXNET3_REG_MACL, tmp);
+
+ tmp = (mac[5] << 8) | mac[4];
+ VMXNET3_WRITE_BAR1_REG(adapter, VMXNET3_REG_MACH, tmp);
+}
+
+
+static int
+vmxnet3_set_mac_addr(struct net_device *netdev, void *p)
+{
+ struct sockaddr *addr = p;
+ struct vmxnet3_adapter *adapter = netdev_priv(netdev);
+
+ memcpy(netdev->dev_addr, addr->sa_data, netdev->addr_len);
+ vmxnet3_write_mac_addr(adapter, addr->sa_data);
+
+ return 0;
+}
+
+
+/* ==================== initialization and cleanup routines ============ */
+
+static int
+vmxnet3_alloc_pci_resources(struct vmxnet3_adapter *adapter, bool *dma64)
+{
+ int err;
+ unsigned long mmio_start, mmio_len;
+ struct pci_dev *pdev = adapter->pdev;
+
+ err = pci_enable_device(pdev);
+ if (err) {
+ printk(KERN_ERR "Failed to enable adapter %s: error %d\n",
+ pci_name(pdev), err);
+ return err;
+ }
+
+ if (pci_set_dma_mask(pdev, DMA_BIT_MASK(64)) == 0) {
+ if (pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(64)) != 0) {
+ printk(KERN_ERR "pci_set_consistent_dma_mask failed "
+ "for adapter %s\n", pci_name(pdev));
+ err = -EIO;
+ goto err_set_mask;
+ }
+ *dma64 = true;
+ } else {
+ if (pci_set_dma_mask(pdev, DMA_BIT_MASK(32)) != 0) {
+ printk(KERN_ERR "pci_set_dma_mask failed for adapter "
+ "%s\n", pci_name(pdev));
+ err = -EIO;
+ goto err_set_mask;
+ }
+ *dma64 = false;
+ }
+
+ err = pci_request_regions(pdev, vmxnet3_driver_name);
+ if (err) {
+ printk(KERN_ERR "Failed to request region for adapter %s: "
+ "error %d\n", pci_name(pdev), err);
+ goto err_set_mask;
+ }
+
+ pci_set_master(pdev);
+
+ mmio_start = pci_resource_start(pdev, 0);
+ mmio_len = pci_resource_len(pdev, 0);
+ adapter->hw_addr0 = ioremap(mmio_start, mmio_len);
+ if (!adapter->hw_addr0) {
+ printk(KERN_ERR "Failed to map bar0 for adapter %s\n",
+ pci_name(pdev));
+ err = -EIO;
+ goto err_ioremap;
+ }
+
+ mmio_start = pci_resource_start(pdev, 1);
+ mmio_len = pci_resource_len(pdev, 1);
+ adapter->hw_addr1 = ioremap(mmio_start, mmio_len);
+ if (!adapter->hw_addr1) {
+ printk(KERN_ERR "Failed to map bar1 for adapter %s\n",
+ pci_name(pdev));
+ err = -EIO;
+ goto err_bar1;
+ }
+ return 0;
+
+err_bar1:
+ iounmap(adapter->hw_addr0);
+err_ioremap:
+ pci_release_regions(pdev);
+err_set_mask:
+ pci_disable_device(pdev);
+ return err;
+}
+
+
+static void
+vmxnet3_free_pci_resources(struct vmxnet3_adapter *adapter)
+{
+ BUG_ON(!adapter->pdev);
+
+ iounmap(adapter->hw_addr0);
+ iounmap(adapter->hw_addr1);
+ pci_release_regions(adapter->pdev);
+ pci_disable_device(adapter->pdev);
+}
+
+
+static void
+vmxnet3_adjust_rx_ring_size(struct vmxnet3_adapter *adapter)
+{
+ size_t sz;
+
+ if (adapter->netdev->mtu <= VMXNET3_MAX_SKB_BUF_SIZE -
+ VMXNET3_MAX_ETH_HDR_SIZE) {
+ adapter->skb_buf_size = adapter->netdev->mtu +
+ VMXNET3_MAX_ETH_HDR_SIZE;
+ if (adapter->skb_buf_size < VMXNET3_MIN_T0_BUF_SIZE)
+ adapter->skb_buf_size = VMXNET3_MIN_T0_BUF_SIZE;
+
+ adapter->rx_buf_per_pkt = 1;
+ } else {
+ adapter->skb_buf_size = VMXNET3_MAX_SKB_BUF_SIZE;
+ sz = adapter->netdev->mtu - VMXNET3_MAX_SKB_BUF_SIZE +
+ VMXNET3_MAX_ETH_HDR_SIZE;
+ adapter->rx_buf_per_pkt = 1 + (sz + PAGE_SIZE - 1) / PAGE_SIZE;
+ }
+
+ /*
+ * for simplicity, force the ring0 size to be a multiple of
+ * rx_buf_per_pkt * VMXNET3_RING_SIZE_ALIGN
+ */
+ sz = adapter->rx_buf_per_pkt * VMXNET3_RING_SIZE_ALIGN;
+ adapter->rx_queue.rx_ring[0].size = (adapter->rx_queue.rx_ring[0].size +
+ sz - 1) / sz * sz;
+ adapter->rx_queue.rx_ring[0].size = min_t(u32,
+ adapter->rx_queue.rx_ring[0].size,
+ VMXNET3_RX_RING_MAX_SIZE / sz * sz);
+}
+
+
+int
+vmxnet3_create_queues(struct vmxnet3_adapter *adapter,
+ u32 tx_ring_size,
+ u32 rx_ring_size,
+ u32 rx_ring2_size)
+{
+ int err;
+
+ adapter->tx_queue.tx_ring.size = tx_ring_size;
+ adapter->tx_queue.data_ring.size = tx_ring_size;
+ adapter->tx_queue.comp_ring.size = tx_ring_size;
+ adapter->tx_queue.shared = &adapter->tqd_start->ctrl;
+ adapter->tx_queue.stopped = true;
+ err = vmxnet3_tq_create(&adapter->tx_queue, adapter);
+ if (err)
+ return err;
+
+ adapter->rx_queue.rx_ring[0].size = rx_ring_size;
+ adapter->rx_queue.rx_ring[1].size = rx_ring2_size;
+ vmxnet3_adjust_rx_ring_size(adapter);
+ adapter->rx_queue.comp_ring.size = adapter->rx_queue.rx_ring[0].size +
+ adapter->rx_queue.rx_ring[1].size;
+ adapter->rx_queue.qid = 0;
+ adapter->rx_queue.qid2 = 1;
+ adapter->rx_queue.shared = &adapter->rqd_start->ctrl;
+ err = vmxnet3_rq_create(&adapter->rx_queue, adapter);
+ if (err)
+ vmxnet3_tq_destroy(&adapter->tx_queue, adapter);
+
+ return err;
+}
+
+static int
+vmxnet3_open(struct net_device *netdev)
+{
+ struct vmxnet3_adapter *adapter;
+ int err;
+
+ adapter = netdev_priv(netdev);
+
+ spin_lock_init(&adapter->tx_queue.tx_lock);
+
+ err = vmxnet3_create_queues(adapter, VMXNET3_DEF_TX_RING_SIZE,
+ VMXNET3_DEF_RX_RING_SIZE,
+ VMXNET3_DEF_RX_RING_SIZE);
+ if (err)
+ goto queue_err;
+
+ err = vmxnet3_activate_dev(adapter);
+ if (err)
+ goto activate_err;
+
+ return 0;
+
+activate_err:
+ vmxnet3_rq_destroy(&adapter->rx_queue, adapter);
+ vmxnet3_tq_destroy(&adapter->tx_queue, adapter);
+queue_err:
+ return err;
+}
+
+
+static int
+vmxnet3_close(struct net_device *netdev)
+{
+ struct vmxnet3_adapter *adapter = netdev_priv(netdev);
+
+ /*
+ * Reset_work may be in the middle of resetting the device, wait for its
+ * completion.
+ */
+ while (test_and_set_bit(VMXNET3_STATE_BIT_RESETTING, &adapter->state))
+ msleep(1);
+
+ vmxnet3_quiesce_dev(adapter);
+
+ vmxnet3_rq_destroy(&adapter->rx_queue, adapter);
+ vmxnet3_tq_destroy(&adapter->tx_queue, adapter);
+
+ clear_bit(VMXNET3_STATE_BIT_RESETTING, &adapter->state);
+
+
+ return 0;
+}
+
+
+void
+vmxnet3_force_close(struct vmxnet3_adapter *adapter)
+{
+ /*
+ * we must clear VMXNET3_STATE_BIT_RESETTING, otherwise
+ * vmxnet3_close() will deadlock.
+ */
+ BUG_ON(test_bit(VMXNET3_STATE_BIT_RESETTING, &adapter->state));
+
+ /* we need to enable NAPI, otherwise dev_close will deadlock */
+ napi_enable(&adapter->napi);
+ dev_close(adapter->netdev);
+}
+
+
+static int
+vmxnet3_change_mtu(struct net_device *netdev, int new_mtu)
+{
+ struct vmxnet3_adapter *adapter = netdev_priv(netdev);
+ int err = 0;
+
+ if (new_mtu < VMXNET3_MIN_MTU || new_mtu > VMXNET3_MAX_MTU)
+ return -EINVAL;
+
+ if (new_mtu > 1500 && !adapter->jumbo_frame)
+ return -EINVAL;
+
+ netdev->mtu = new_mtu;
+
+ /*
+ * Reset_work may be in the middle of resetting the device, wait for its
+ * completion.
+ */
+ while (test_and_set_bit(VMXNET3_STATE_BIT_RESETTING, &adapter->state))
+ msleep(1);
+
+ if (netif_running(netdev)) {
+ vmxnet3_quiesce_dev(adapter);
+ vmxnet3_reset_dev(adapter);
+
+ /* we need to re-create the rx queue based on the new mtu */
+ vmxnet3_rq_destroy(&adapter->rx_queue, adapter);
+ vmxnet3_adjust_rx_ring_size(adapter);
+ adapter->rx_queue.comp_ring.size =
+ adapter->rx_queue.rx_ring[0].size +
+ adapter->rx_queue.rx_ring[1].size;
+ err = vmxnet3_rq_create(&adapter->rx_queue, adapter);
+ if (err) {
+ printk(KERN_ERR "%s: failed to re-create rx queue,"
+ " error %d. Closing it.\n", netdev->name, err);
+ goto out;
+ }
+
+ err = vmxnet3_activate_dev(adapter);
+ if (err) {
+ printk(KERN_ERR "%s: failed to re-activate, error %d. "
+ "Closing it\n", netdev->name, err);
+ goto out;
+ }
+ }
+
+out:
+ clear_bit(VMXNET3_STATE_BIT_RESETTING, &adapter->state);
+ if (err)
+ vmxnet3_force_close(adapter);
+
+ return err;
+}
+
+
+static void
+vmxnet3_declare_features(struct vmxnet3_adapter *adapter, bool dma64)
+{
+ struct net_device *netdev = adapter->netdev;
+
+ netdev->features = NETIF_F_SG |
+ NETIF_F_HW_CSUM |
+ NETIF_F_HW_VLAN_TX |
+ NETIF_F_HW_VLAN_RX |
+ NETIF_F_HW_VLAN_FILTER |
+ NETIF_F_TSO |
+ NETIF_F_TSO6;
+
+ printk(KERN_INFO "features: sg csum vlan jf tso tsoIPv6");
+
+ adapter->rxcsum = true;
+ adapter->jumbo_frame = true;
+
+ if (!disable_lro) {
+ adapter->lro = true;
+ printk(" lro");
+ }
+
+ if (dma64) {
+ netdev->features |= NETIF_F_HIGHDMA;
+ printk(" highDMA");
+ }
+
+ netdev->vlan_features = netdev->features;
+ printk("\n");
+}
+
+
+static void
+vmxnet3_read_mac_addr(struct vmxnet3_adapter *adapter, u8 *mac)
+{
+ u32 tmp;
+
+ tmp = VMXNET3_READ_BAR1_REG(adapter, VMXNET3_REG_MACL);
+ *(u32 *)mac = tmp;
+
+ tmp = VMXNET3_READ_BAR1_REG(adapter, VMXNET3_REG_MACH);
+ mac[4] = tmp & 0xff;
+ mac[5] = (tmp >> 8) & 0xff;
+}
+
+
+static void
+vmxnet3_alloc_intr_resources(struct vmxnet3_adapter *adapter)
+{
+ u32 cfg;
+
+ /* intr settings */
+ VMXNET3_WRITE_BAR1_REG(adapter, VMXNET3_REG_CMD,
+ VMXNET3_CMD_GET_CONF_INTR);
+ cfg = VMXNET3_READ_BAR1_REG(adapter, VMXNET3_REG_CMD);
+ adapter->intr.type = cfg & 0x3;
+ adapter->intr.mask_mode = (cfg >> 2) & 0x3;
+
+ if (adapter->intr.type == VMXNET3_IT_AUTO) {
+ int err;
+
+ adapter->intr.msix_entries[0].entry = 0;
+ err = pci_enable_msix(adapter->pdev, adapter->intr.msix_entries,
+ VMXNET3_LINUX_MAX_MSIX_VECT);
+ if (!err) {
+ adapter->intr.num_intrs = 1;
+ adapter->intr.type = VMXNET3_IT_MSIX;
+ return;
+ }
+
+ printk(KERN_INFO "Failed to enable MSI-X for %s, error %d, "
+ "try MSI\n", adapter->netdev->name, err);
+
+ err = pci_enable_msi(adapter->pdev);
+ if (!err) {
+ adapter->intr.num_intrs = 1;
+ adapter->intr.type = VMXNET3_IT_MSI;
+ return;
+ }
+
+ printk(KERN_INFO "Failed to enable MSI for %s, error %d, use "
+ "INTx\n", adapter->netdev->name, err);
+ }
+
+ adapter->intr.type = VMXNET3_IT_INTX;
+
+ /* INT-X related setting */
+ adapter->intr.num_intrs = 1;
+}
+
+
+static void
+vmxnet3_free_intr_resources(struct vmxnet3_adapter *adapter)
+{
+ if (adapter->intr.type == VMXNET3_IT_MSIX)
+ pci_disable_msix(adapter->pdev);
+ else if (adapter->intr.type == VMXNET3_IT_MSI)
+ pci_disable_msi(adapter->pdev);
+ else
+ BUG_ON(adapter->intr.type != VMXNET3_IT_INTX);
+}
+
+
+static void
+vmxnet3_tx_timeout(struct net_device *netdev)
+{
+ struct vmxnet3_adapter *adapter = netdev_priv(netdev);
+ adapter->tx_timeout_count++;
+
+ printk(KERN_ERR "%s: tx hang\n", adapter->netdev->name);
+ schedule_work(&adapter->work);
+}
+
+
+static void
+vmxnet3_reset_work(struct work_struct *data)
+{
+ struct vmxnet3_adapter *adapter;
+
+ adapter = container_of(data, struct vmxnet3_adapter, work);
+
+ /* if another thread is resetting the device, no need to proceed */
+ if (test_and_set_bit(VMXNET3_STATE_BIT_RESETTING, &adapter->state)) {
+ printk(KERN_INFO "%s: resetting already in progress\n",
+ adapter->netdev->name);
+ return;
+ }
+
+ /* if the device is closed, we must leave it alone */
+ if (netif_running(adapter->netdev)) {
+ printk(KERN_INFO "%s: resetting\n", adapter->netdev->name);
+ vmxnet3_quiesce_dev(adapter);
+ vmxnet3_reset_dev(adapter);
+ vmxnet3_activate_dev(adapter);
+ } else {
+ printk(KERN_INFO "%s: already closed\n", adapter->netdev->name);
+ }
+
+ clear_bit(VMXNET3_STATE_BIT_RESETTING, &adapter->state);
+}
+
+
+static int __devinit
+vmxnet3_probe_device(struct pci_dev *pdev,
+ const struct pci_device_id *id)
+{
+ static const struct net_device_ops vmxnet3_netdev_ops = {
+ .ndo_open = vmxnet3_open,
+ .ndo_stop = vmxnet3_close,
+ .ndo_start_xmit = vmxnet3_xmit_frame,
+ .ndo_set_mac_address = vmxnet3_set_mac_addr,
+ .ndo_change_mtu = vmxnet3_change_mtu,
+ .ndo_get_stats = vmxnet3_get_stats,
+ .ndo_tx_timeout = vmxnet3_tx_timeout,
+ .ndo_set_multicast_list = vmxnet3_set_mc,
+ .ndo_vlan_rx_register = vmxnet3_vlan_rx_register,
+ .ndo_vlan_rx_add_vid = vmxnet3_vlan_rx_add_vid,
+ .ndo_vlan_rx_kill_vid = vmxnet3_vlan_rx_kill_vid,
+# ifdef CONFIG_NET_POLL_CONTROLLER
+ .ndo_poll_controller = vmxnet3_netpoll,
+# endif
+ };
+ int err;
+ bool dma64 = false; /* stupid gcc */
+ u32 ver;
+ struct net_device *netdev;
+ struct vmxnet3_adapter *adapter;
+ u8 mac[ETH_ALEN];
+
+ netdev = alloc_etherdev(sizeof(struct vmxnet3_adapter));
+ if (!netdev) {
+ printk(KERN_ERR "Failed to alloc ethernet device for adapter "
+ "%s\n", pci_name(pdev));
+ return -ENOMEM;
+ }
+
+ pci_set_drvdata(pdev, netdev);
+ adapter = netdev_priv(netdev);
+ adapter->netdev = netdev;
+ adapter->pdev = pdev;
+
+ adapter->shared = pci_alloc_consistent(adapter->pdev,
+ sizeof(struct Vmxnet3_DriverShared),
+ &adapter->shared_pa);
+ if (!adapter->shared) {
+ printk(KERN_ERR "Failed to allocate memory for %s\n",
+ pci_name(pdev));
+ err = -ENOMEM;
+ goto err_alloc_shared;
+ }
+
+ adapter->tqd_start = pci_alloc_consistent(adapter->pdev,
+ sizeof(struct Vmxnet3_TxQueueDesc) +
+ sizeof(struct Vmxnet3_RxQueueDesc),
+ &adapter->queue_desc_pa);
+
+ if (!adapter->tqd_start) {
+ printk(KERN_ERR "Failed to allocate memory for %s\n",
+ pci_name(pdev));
+ err = -ENOMEM;
+ goto err_alloc_queue_desc;
+ }
+ adapter->rqd_start = (struct Vmxnet3_RxQueueDesc *)(adapter->tqd_start
+ + 1);
+
+ adapter->pm_conf = kmalloc(sizeof(struct Vmxnet3_PMConf), GFP_KERNEL);
+ if (adapter->pm_conf == NULL) {
+ printk(KERN_ERR "Failed to allocate memory for %s\n",
+ pci_name(pdev));
+ err = -ENOMEM;
+ goto err_alloc_pm;
+ }
+
+ err = vmxnet3_alloc_pci_resources(adapter, &dma64);
+ if (err < 0)
+ goto err_alloc_pci;
+
+ ver = VMXNET3_READ_BAR1_REG(adapter, VMXNET3_REG_VRRS);
+ if (ver & 1) {
+ VMXNET3_WRITE_BAR1_REG(adapter, VMXNET3_REG_VRRS, 1);
+ } else {
+ printk(KERN_ERR "Incompatible h/w version (0x%x) for adapter"
+ " %s\n", ver, pci_name(pdev));
+ err = -EBUSY;
+ goto err_ver;
+ }
+
+ ver = VMXNET3_READ_BAR1_REG(adapter, VMXNET3_REG_UVRS);
+ if (ver & 1) {
+ VMXNET3_WRITE_BAR1_REG(adapter, VMXNET3_REG_UVRS, 1);
+ } else {
+ printk(KERN_ERR "Incompatible upt version (0x%x) for "
+ "adapter %s\n", ver, pci_name(pdev));
+ err = -EBUSY;
+ goto err_ver;
+ }
+
+ vmxnet3_declare_features(adapter, dma64);
+
+ adapter->dev_number = atomic_read(&devices_found);
+ vmxnet3_alloc_intr_resources(adapter);
+
+ vmxnet3_read_mac_addr(adapter, mac);
+ memcpy(netdev->dev_addr, mac, netdev->addr_len);
+
+ netdev->netdev_ops = &vmxnet3_netdev_ops;
+ netdev->watchdog_timeo = 5 * HZ;
+ vmxnet3_set_ethtool_ops(netdev);
+
+ INIT_WORK(&adapter->work, vmxnet3_reset_work);
+
+ netif_napi_add(netdev, &adapter->napi, vmxnet3_poll, 64);
+ SET_NETDEV_DEV(netdev, &pdev->dev);
+ err = register_netdev(netdev);
+
+ if (err) {
+ printk(KERN_ERR "Failed to register adapter %s\n",
+ pci_name(pdev));
+ goto err_register;
+ }
+
+ set_bit(VMXNET3_STATE_BIT_QUIESCED, &adapter->state);
+ atomic_inc(&devices_found);
+ return 0;
+
+err_register:
+ vmxnet3_free_intr_resources(adapter);
+err_ver:
+ vmxnet3_free_pci_resources(adapter);
+err_alloc_pci:
+ kfree(adapter->pm_conf);
+err_alloc_pm:
+ pci_free_consistent(adapter->pdev, sizeof(struct Vmxnet3_TxQueueDesc) +
+ sizeof(struct Vmxnet3_RxQueueDesc),
+ adapter->tqd_start, adapter->queue_desc_pa);
+err_alloc_queue_desc:
+ pci_free_consistent(adapter->pdev, sizeof(struct Vmxnet3_DriverShared),
+ adapter->shared, adapter->shared_pa);
+err_alloc_shared:
+ pci_set_drvdata(pdev, NULL);
+ free_netdev(netdev);
+ return err;
+}
+
+
+static void __devexit
+vmxnet3_remove_device(struct pci_dev *pdev)
+{
+ struct net_device *netdev = pci_get_drvdata(pdev);
+ struct vmxnet3_adapter *adapter = netdev_priv(netdev);
+
+ flush_scheduled_work();
+
+ unregister_netdev(netdev);
+
+ vmxnet3_free_intr_resources(adapter);
+ vmxnet3_free_pci_resources(adapter);
+ kfree(adapter->pm_conf);
+ pci_free_consistent(adapter->pdev, sizeof(struct Vmxnet3_TxQueueDesc) +
+ sizeof(struct Vmxnet3_RxQueueDesc),
+ adapter->tqd_start, adapter->queue_desc_pa);
+ pci_free_consistent(adapter->pdev, sizeof(struct Vmxnet3_DriverShared),
+ adapter->shared, adapter->shared_pa);
+ free_netdev(netdev);
+}
+
+
+#ifdef CONFIG_PM
+
+static int
+vmxnet3_suspend(struct device *device)
+{
+ struct pci_dev *pdev = to_pci_dev(device);
+ struct net_device *netdev = pci_get_drvdata(pdev);
+ struct vmxnet3_adapter *adapter = netdev_priv(netdev);
+ struct Vmxnet3_PMConf *pmConf;
+ struct ethhdr *ehdr;
+ struct arphdr *ahdr;
+ u8 *arpreq;
+ struct in_device *in_dev;
+ struct in_ifaddr *ifa;
+ int i = 0;
+
+ if (!netif_running(netdev))
+ return 0;
+
+ vmxnet3_disable_all_intrs(adapter);
+ netif_device_detach(netdev);
+ netif_stop_queue(netdev);
+
+ /* Create wake-up filters. */
+ pmConf = adapter->pm_conf;
+ memset(pmConf, 0, sizeof(*pmConf));
+
+ if (adapter->wol & WAKE_UCAST) {
+ pmConf->filters[i].patternSize = ETH_ALEN;
+ pmConf->filters[i].maskSize = 1;
+ memcpy(pmConf->filters[i].pattern, netdev->dev_addr, ETH_ALEN);
+ pmConf->filters[i].mask[0] = 0x3F; /* LSB ETH_ALEN bits */
+
+ pmConf->wakeUpEvents |= VMXNET3_PM_WAKEUP_FILTER;
+ i++;
+ }
+
+ if (adapter->wol & WAKE_ARP) {
+ in_dev = in_dev_get(netdev);
+ if (!in_dev) {
+ dprintk(KERN_ERR "Cannot program WoL ARP filter for %s:"
+ " IPv4 not enabled.\n", netdev->name);
+ goto skip_arp;
+ }
+ ifa = (struct in_ifaddr *)in_dev->ifa_list;
+ if (!ifa) {
+ dprintk(KERN_ERR "Cannot program WoL ARP filter for %s:"
+ " no IPv4 address.\n", netdev->name);
+ in_dev_put(in_dev);
+ goto skip_arp;
+ }
+ pmConf->filters[i].patternSize = ETH_HLEN + /* Ethernet header*/
+ sizeof(struct arphdr) + /* ARP header */
+ 2 * ETH_ALEN + /* 2 Ethernet addresses*/
+ 2 * sizeof(u32); /*2 IPv4 addresses */
+ pmConf->filters[i].maskSize =
+ (pmConf->filters[i].patternSize - 1) / 8 + 1;
+
+ /* ETH_P_ARP in Ethernet header. */
+ ehdr = (struct ethhdr *)pmConf->filters[i].pattern;
+ ehdr->h_proto = htons(ETH_P_ARP);
+
+ /* ARPOP_REQUEST in ARP header. */
+ ahdr = (struct arphdr *)&pmConf->filters[i].pattern[ETH_HLEN];
+ ahdr->ar_op = htons(ARPOP_REQUEST);
+ arpreq = (u8 *)(ahdr + 1);
+
+ /* The Unicast IPv4 address in 'tip' field. */
+ arpreq += 2 * ETH_ALEN + sizeof(u32);
+ *(u32 *)arpreq = ifa->ifa_address;
+
+ /* The mask for the relevant bits. */
+ pmConf->filters[i].mask[0] = 0x00;
+ pmConf->filters[i].mask[1] = 0x30; /* ETH_P_ARP */
+ pmConf->filters[i].mask[2] = 0x30; /* ARPOP_REQUEST */
+ pmConf->filters[i].mask[3] = 0x00;
+ pmConf->filters[i].mask[4] = 0xC0; /* IPv4 TIP */
+ pmConf->filters[i].mask[5] = 0x03; /* IPv4 TIP */
+ in_dev_put(in_dev);
+
+ pmConf->wakeUpEvents |= VMXNET3_PM_WAKEUP_FILTER;
+ i++;
+ }
+
+skip_arp:
+ if (adapter->wol & WAKE_MAGIC)
+ pmConf->wakeUpEvents |= VMXNET3_PM_WAKEUP_MAGIC;
+
+ pmConf->numFilters = i;
+
+ adapter->shared->devRead.pmConfDesc.confVer = 1;
+ adapter->shared->devRead.pmConfDesc.confLen = sizeof(*pmConf);
+ adapter->shared->devRead.pmConfDesc.confPA = virt_to_phys(pmConf);
+
+ VMXNET3_WRITE_BAR1_REG(adapter, VMXNET3_REG_CMD,
+ VMXNET3_CMD_UPDATE_PMCFG);
+
+ pci_save_state(pdev);
+ pci_enable_wake(pdev, pci_choose_state(pdev, PMSG_SUSPEND),
+ adapter->wol);
+ pci_disable_device(pdev);
+ pci_set_power_state(pdev, pci_choose_state(pdev, PMSG_SUSPEND));
+
+ return 0;
+}
+
+
+static int
+vmxnet3_resume(struct device *device)
+{
+ int err;
+ struct pci_dev *pdev = to_pci_dev(device);
+ struct net_device *netdev = pci_get_drvdata(pdev);
+ struct vmxnet3_adapter *adapter = netdev_priv(netdev);
+ struct Vmxnet3_PMConf *pmConf;
+
+ if (!netif_running(netdev))
+ return 0;
+
+ /* Destroy wake-up filters. */
+ pmConf = adapter->pm_conf;
+ memset(pmConf, 0, sizeof(*pmConf));
+
+ adapter->shared->devRead.pmConfDesc.confVer = 1;
+ adapter->shared->devRead.pmConfDesc.confLen = sizeof(*pmConf);
+ adapter->shared->devRead.pmConfDesc.confPA = virt_to_phys(pmConf);
+
+ netif_device_attach(netdev);
+ pci_set_power_state(pdev, PCI_D0);
+ pci_restore_state(pdev);
+ err = pci_enable_device(pdev);
+ if (err != 0)
+ return err;
+
+ pci_enable_wake(pdev, PCI_D0, 0);
+
+ VMXNET3_WRITE_BAR1_REG(adapter, VMXNET3_REG_CMD,
+ VMXNET3_CMD_UPDATE_PMCFG);
+ vmxnet3_enable_all_intrs(adapter);
+
+ return 0;
+}
+
+static struct dev_pm_ops vmxnet3_pm_ops = {
+ .suspend = vmxnet3_suspend,
+ .resume = vmxnet3_resume,
+};
+#endif
+
+static struct pci_driver vmxnet3_driver = {
+ .name = vmxnet3_driver_name,
+ .id_table = vmxnet3_pciid_table,
+ .probe = vmxnet3_probe_device,
+ .remove = __devexit_p(vmxnet3_remove_device),
+#ifdef CONFIG_PM
+ .driver.pm = &vmxnet3_pm_ops,
+#endif
+};
+
+
+static int __init
+vmxnet3_init_module(void)
+{
+ printk(KERN_INFO "%s - version %s\n", VMXNET3_DRIVER_DESC,
+ VMXNET3_DRIVER_VERSION_REPORT);
+ return pci_register_driver(&vmxnet3_driver);
+}
+
+module_init(vmxnet3_init_module);
+
+
+static void
+vmxnet3_exit_module(void)
+{
+ pci_unregister_driver(&vmxnet3_driver);
+}
+
+module_exit(vmxnet3_exit_module);
+
+MODULE_AUTHOR("VMware, Inc.");
+MODULE_DESCRIPTION(VMXNET3_DRIVER_DESC);
+MODULE_LICENSE("GPL v2");
+MODULE_VERSION(VMXNET3_DRIVER_VERSION_STRING);
+
+/* This paramenter is used to control Large Receive Offload feature
+ * of the NIC. When set to non-zeora LRO is enabled.
+ */
+module_param(disable_lro, int, 0);
diff --git a/drivers/net/vmxnet3/vmxnet3_ethtool.c b/drivers/net/vmxnet3/vmxnet3_ethtool.c
new file mode 100644
index 0000000..490577f
--- /dev/null
+++ b/drivers/net/vmxnet3/vmxnet3_ethtool.c
@@ -0,0 +1,578 @@
+/*
+ * Linux driver for VMware's vmxnet3 ethernet NIC.
+ *
+ * Copyright (C) 2008-2009, VMware, Inc. All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License as published by the
+ * Free Software Foundation; version 2 of the License and no later version.
+ *
+ * This program 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, GOOD TITLE or
+ * NON INFRINGEMENT. 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, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
+ *
+ * The full GNU General Public License is included in this distribution in
+ * the file called "COPYING".
+ *
+ * Maintained by: Shreyas Bhatewara <pv-drivers@vmware.com>
+ *
+ */
+
+/*
+ * vmxnet3_ethtool.c --
+ *
+ * API to support ethtool for for VMXNET3 NIC
+ */
+
+
+#include "vmxnet3_int.h"
+
+struct vmxnet3_stat_desc {
+ char desc[ETH_GSTRING_LEN];
+ int offset;
+};
+
+
+static u32
+vmxnet3_get_rx_csum(struct net_device *netdev)
+{
+ struct vmxnet3_adapter *adapter = netdev_priv(netdev);
+ return adapter->rxcsum;
+}
+
+
+static int
+vmxnet3_set_rx_csum(struct net_device *netdev, u32 val)
+{
+ struct vmxnet3_adapter *adapter = netdev_priv(netdev);
+
+ if (adapter->rxcsum != val) {
+ adapter->rxcsum = val;
+ if (netif_running(netdev)) {
+ if (val)
+ adapter->shared->devRead.misc.uptFeatures |=
+ UPT1_F_RXCSUM;
+ else
+ adapter->shared->devRead.misc.uptFeatures &=
+ ~UPT1_F_RXCSUM;
+
+ VMXNET3_WRITE_BAR1_REG(adapter, VMXNET3_REG_CMD,
+ VMXNET3_CMD_UPDATE_FEATURE);
+ }
+ }
+ return 0;
+}
+
+
+static u32
+vmxnet3_get_tx_csum(struct net_device *netdev)
+{
+ return (netdev->features & NETIF_F_HW_CSUM) != 0;
+}
+
+
+static int
+vmxnet3_set_tx_csum(struct net_device *netdev, u32 val)
+{
+ if (val)
+ netdev->features |= NETIF_F_HW_CSUM;
+ else
+ netdev->features &= ~NETIF_F_HW_CSUM;
+
+ return 0;
+}
+
+
+static int
+vmxnet3_set_sg(struct net_device *netdev, u32 val)
+{
+ ethtool_op_set_sg(netdev, val);
+ return 0;
+}
+
+
+static int
+vmxnet3_set_tso(struct net_device *netdev, u32 val)
+{
+ ethtool_op_set_tso(netdev, val);
+ return 0;
+}
+
+
+/* per tq stats maintained by the device */
+static const struct vmxnet3_stat_desc
+vmxnet3_tq_dev_stats[] = {
+ /* description, offset */
+ { "TSO pkts tx", offsetof(struct UPT1_TxStats, TSOPktsTxOK) },
+ { "TSO bytes tx", offsetof(struct UPT1_TxStats, TSOBytesTxOK) },
+ { "ucast pkts tx", offsetof(struct UPT1_TxStats, ucastPktsTxOK) },
+ { "ucast bytes tx", offsetof(struct UPT1_TxStats, ucastBytesTxOK) },
+ { "mcast pkts tx", offsetof(struct UPT1_TxStats, mcastPktsTxOK) },
+ { "mcast bytes tx", offsetof(struct UPT1_TxStats, mcastBytesTxOK) },
+ { "bcast pkts tx", offsetof(struct UPT1_TxStats, bcastPktsTxOK) },
+ { "bcast bytes tx", offsetof(struct UPT1_TxStats, bcastBytesTxOK) },
+ { "pkts tx err", offsetof(struct UPT1_TxStats, pktsTxError) },
+ { "pkts tx discard", offsetof(struct UPT1_TxStats, pktsTxDiscard) },
+};
+
+/* per tq stats maintained by the driver */
+static const struct vmxnet3_stat_desc
+vmxnet3_tq_driver_stats[] = {
+ /* description, offset */
+ {"drv dropped tx total", offsetof(struct vmxnet3_tq_driver_stats,
+ drop_total) },
+ { " too many frags", offsetof(struct vmxnet3_tq_driver_stats,
+ drop_too_many_frags) },
+ { " giant hdr", offsetof(struct vmxnet3_tq_driver_stats,
+ drop_oversized_hdr) },
+ { " hdr err", offsetof(struct vmxnet3_tq_driver_stats,
+ drop_hdr_inspect_err) },
+ { " tso", offsetof(struct vmxnet3_tq_driver_stats,
+ drop_tso) },
+ { "ring full", offsetof(struct vmxnet3_tq_driver_stats,
+ tx_ring_full) },
+ { "pkts linearized", offsetof(struct vmxnet3_tq_driver_stats,
+ linearized) },
+ { "hdr cloned", offsetof(struct vmxnet3_tq_driver_stats,
+ copy_skb_header) },
+ { "giant hdr", offsetof(struct vmxnet3_tq_driver_stats,
+ oversized_hdr) },
+};
+
+/* per rq stats maintained by the device */
+static const struct vmxnet3_stat_desc
+vmxnet3_rq_dev_stats[] = {
+ { "LRO pkts rx", offsetof(struct UPT1_RxStats, LROPktsRxOK) },
+ { "LRO byte rx", offsetof(struct UPT1_RxStats, LROBytesRxOK) },
+ { "ucast pkts rx", offsetof(struct UPT1_RxStats, ucastPktsRxOK) },
+ { "ucast bytes rx", offsetof(struct UPT1_RxStats, ucastBytesRxOK) },
+ { "mcast pkts rx", offsetof(struct UPT1_RxStats, mcastPktsRxOK) },
+ { "mcast bytes rx", offsetof(struct UPT1_RxStats, mcastBytesRxOK) },
+ { "bcast pkts rx", offsetof(struct UPT1_RxStats, bcastPktsRxOK) },
+ { "bcast bytes rx", offsetof(struct UPT1_RxStats, bcastBytesRxOK) },
+ { "pkts rx out of buf", offsetof(struct UPT1_RxStats, pktsRxOutOfBuf) },
+ { "pkts rx err", offsetof(struct UPT1_RxStats, pktsRxError) },
+};
+
+/* per rq stats maintained by the driver */
+static const struct vmxnet3_stat_desc
+vmxnet3_rq_driver_stats[] = {
+ /* description, offset */
+ { "drv dropped rx total", offsetof(struct vmxnet3_rq_driver_stats,
+ drop_total) },
+ { " err", offsetof(struct vmxnet3_rq_driver_stats,
+ drop_err) },
+ { " fcs", offsetof(struct vmxnet3_rq_driver_stats,
+ drop_fcs) },
+ { "rx buf alloc fail", offsetof(struct vmxnet3_rq_driver_stats,
+ rx_buf_alloc_failure) },
+};
+
+/* gloabl stats maintained by the driver */
+static const struct vmxnet3_stat_desc
+vmxnet3_global_stats[] = {
+ /* description, offset */
+ { "tx timeout count", offsetof(struct vmxnet3_adapter,
+ tx_timeout_count) }
+};
+
+
+struct net_device_stats*
+vmxnet3_get_stats(struct net_device *netdev)
+{
+ struct vmxnet3_adapter *adapter;
+ struct vmxnet3_tq_driver_stats *drvTxStats;
+ struct vmxnet3_rq_driver_stats *drvRxStats;
+ struct UPT1_TxStats *devTxStats;
+ struct UPT1_RxStats *devRxStats;
+
+ adapter = netdev_priv(netdev);
+
+ /* Collect the dev stats into the shared area */
+ VMXNET3_WRITE_BAR1_REG(adapter, VMXNET3_REG_CMD, VMXNET3_CMD_GET_STATS);
+
+ /* Assuming that we have a single queue device */
+ devTxStats = &adapter->tqd_start->stats;
+ devRxStats = &adapter->rqd_start->stats;
+
+ /* Get access to the driver stats per queue */
+ drvTxStats = &adapter->tx_queue.stats;
+ drvRxStats = &adapter->rx_queue.stats;
+
+ memset(&adapter->net_stats, 0, sizeof(adapter->net_stats));
+
+ adapter->net_stats.rx_packets = devRxStats->ucastPktsRxOK +
+ devRxStats->mcastPktsRxOK +
+ devRxStats->bcastPktsRxOK;
+
+ adapter->net_stats.tx_packets = devTxStats->ucastPktsTxOK +
+ devTxStats->mcastPktsTxOK +
+ devTxStats->bcastPktsTxOK;
+
+ adapter->net_stats.rx_bytes = devRxStats->ucastBytesRxOK +
+ devRxStats->mcastBytesRxOK +
+ devRxStats->bcastBytesRxOK;
+
+ adapter->net_stats.tx_bytes = devTxStats->ucastBytesTxOK +
+ devTxStats->mcastBytesTxOK +
+ devTxStats->bcastBytesTxOK;
+
+ adapter->net_stats.rx_errors = devRxStats->pktsRxError;
+ adapter->net_stats.tx_errors = devTxStats->pktsTxError;
+ adapter->net_stats.rx_dropped = drvRxStats->drop_total;
+ adapter->net_stats.tx_dropped = drvTxStats->drop_total;
+ adapter->net_stats.multicast = devRxStats->mcastPktsRxOK;
+
+ return &adapter->net_stats;
+}
+
+static int
+vmxnet3_get_stats_count(struct net_device *netdev)
+{
+ return ARRAY_SIZE(vmxnet3_tq_dev_stats) +
+ ARRAY_SIZE(vmxnet3_tq_driver_stats) +
+ ARRAY_SIZE(vmxnet3_rq_dev_stats) +
+ ARRAY_SIZE(vmxnet3_rq_driver_stats) +
+ ARRAY_SIZE(vmxnet3_global_stats);
+}
+
+
+static int
+vmxnet3_get_regs_len(struct net_device *netdev)
+{
+ return 20 * sizeof(u32);
+}
+
+
+static void
+vmxnet3_get_drvinfo(struct net_device *netdev, struct ethtool_drvinfo *drvinfo)
+{
+ struct vmxnet3_adapter *adapter = netdev_priv(netdev);
+
+ strncpy(drvinfo->driver, vmxnet3_driver_name, sizeof(drvinfo->driver));
+ drvinfo->driver[sizeof(drvinfo->driver) - 1] = '\0';
+
+ strncpy(drvinfo->version, VMXNET3_DRIVER_VERSION_REPORT,
+ sizeof(drvinfo->version));
+ drvinfo->driver[sizeof(drvinfo->version) - 1] = '\0';
+
+ strncpy(drvinfo->fw_version, "N/A", sizeof(drvinfo->fw_version));
+ drvinfo->fw_version[sizeof(drvinfo->fw_version) - 1] = '\0';
+
+ strncpy(drvinfo->bus_info, pci_name(adapter->pdev),
+ ETHTOOL_BUSINFO_LEN);
+ drvinfo->n_stats = vmxnet3_get_stats_count(netdev);
+ drvinfo->testinfo_len = 0;
+ drvinfo->eedump_len = 0;
+ drvinfo->regdump_len = vmxnet3_get_regs_len(netdev);
+}
+
+
+static void
+vmxnet3_get_strings(struct net_device *netdev, u32 stringset, u8 *buf)
+{
+ if (stringset == ETH_SS_STATS) {
+ int i;
+
+ for (i = 0; i < ARRAY_SIZE(vmxnet3_tq_dev_stats); i++) {
+ memcpy(buf, vmxnet3_tq_dev_stats[i].desc,
+ ETH_GSTRING_LEN);
+ buf += ETH_GSTRING_LEN;
+ }
+ for (i = 0; i < ARRAY_SIZE(vmxnet3_tq_driver_stats); i++) {
+ memcpy(buf, vmxnet3_tq_driver_stats[i].desc,
+ ETH_GSTRING_LEN);
+ buf += ETH_GSTRING_LEN;
+ }
+ for (i = 0; i < ARRAY_SIZE(vmxnet3_rq_dev_stats); i++) {
+ memcpy(buf, vmxnet3_rq_dev_stats[i].desc,
+ ETH_GSTRING_LEN);
+ buf += ETH_GSTRING_LEN;
+ }
+ for (i = 0; i < ARRAY_SIZE(vmxnet3_rq_driver_stats); i++) {
+ memcpy(buf, vmxnet3_rq_driver_stats[i].desc,
+ ETH_GSTRING_LEN);
+ buf += ETH_GSTRING_LEN;
+ }
+ for (i = 0; i < ARRAY_SIZE(vmxnet3_global_stats); i++) {
+ memcpy(buf, vmxnet3_global_stats[i].desc,
+ ETH_GSTRING_LEN);
+ buf += ETH_GSTRING_LEN;
+ }
+ }
+}
+
+
+static void
+vmxnet3_get_ethtool_stats(struct net_device *netdev,
+ struct ethtool_stats *stats,
+ u64 *buf)
+{
+ struct vmxnet3_adapter *adapter = netdev_priv(netdev);
+ u8 *base;
+ int i;
+
+ VMXNET3_WRITE_BAR1_REG(adapter, VMXNET3_REG_CMD, VMXNET3_CMD_GET_STATS);
+
+ /* this does assume each counter is 64-bit wide */
+
+ base = (u8 *)&adapter->tqd_start->stats;
+ for (i = 0; i < ARRAY_SIZE(vmxnet3_tq_dev_stats); i++)
+ *buf++ = *(u64 *)(base + vmxnet3_tq_dev_stats[i].offset);
+
+ base = (u8 *)&adapter->tx_queue.stats;
+ for (i = 0; i < ARRAY_SIZE(vmxnet3_tq_driver_stats); i++)
+ *buf++ = *(u64 *)(base + vmxnet3_tq_driver_stats[i].offset);
+
+ base = (u8 *)&adapter->rqd_start->stats;
+ for (i = 0; i < ARRAY_SIZE(vmxnet3_rq_dev_stats); i++)
+ *buf++ = *(u64 *)(base + vmxnet3_rq_dev_stats[i].offset);
+
+ base = (u8 *)&adapter->rx_queue.stats;
+ for (i = 0; i < ARRAY_SIZE(vmxnet3_rq_driver_stats); i++)
+ *buf++ = *(u64 *)(base + vmxnet3_rq_driver_stats[i].offset);
+
+ base = (u8 *)adapter;
+ for (i = 0; i < ARRAY_SIZE(vmxnet3_global_stats); i++)
+ *buf++ = *(u64 *)(base + vmxnet3_global_stats[i].offset);
+}
+
+
+static void
+vmxnet3_get_regs(struct net_device *netdev, struct ethtool_regs *regs, void *p)
+{
+ struct vmxnet3_adapter *adapter = netdev_priv(netdev);
+ u32 *buf = p;
+
+ memset(p, 0, vmxnet3_get_regs_len(netdev));
+
+ regs->version = 1;
+
+ /* Update vmxnet3_get_regs_len if we want to dump more registers */
+
+ /* make each ring use multiple of 16 bytes */
+ buf[0] = adapter->tx_queue.tx_ring.next2fill;
+ buf[1] = adapter->tx_queue.tx_ring.next2comp;
+ buf[2] = adapter->tx_queue.tx_ring.gen;
+ buf[3] = 0;
+
+ buf[4] = adapter->tx_queue.comp_ring.next2proc;
+ buf[5] = adapter->tx_queue.comp_ring.gen;
+ buf[6] = adapter->tx_queue.stopped;
+ buf[7] = 0;
+
+ buf[8] = adapter->rx_queue.rx_ring[0].next2fill;
+ buf[9] = adapter->rx_queue.rx_ring[0].next2comp;
+ buf[10] = adapter->rx_queue.rx_ring[0].gen;
+ buf[11] = 0;
+
+ buf[12] = adapter->rx_queue.rx_ring[1].next2fill;
+ buf[13] = adapter->rx_queue.rx_ring[1].next2comp;
+ buf[14] = adapter->rx_queue.rx_ring[1].gen;
+ buf[15] = 0;
+
+ buf[16] = adapter->rx_queue.comp_ring.next2proc;
+ buf[17] = adapter->rx_queue.comp_ring.gen;
+ buf[18] = 0;
+ buf[19] = 0;
+}
+
+
+static void
+vmxnet3_get_wol(struct net_device *netdev, struct ethtool_wolinfo *wol)
+{
+ struct vmxnet3_adapter *adapter = netdev_priv(netdev);
+
+ wol->supported = WAKE_UCAST | WAKE_ARP | WAKE_MAGIC;
+ wol->wolopts = adapter->wol;
+}
+
+
+static int
+vmxnet3_set_wol(struct net_device *netdev, struct ethtool_wolinfo *wol)
+{
+ struct vmxnet3_adapter *adapter = netdev_priv(netdev);
+
+ if (wol->wolopts & (WAKE_PHY | WAKE_MCAST | WAKE_BCAST |
+ WAKE_MAGICSECURE)) {
+ return -EOPNOTSUPP;
+ }
+
+ adapter->wol = wol->wolopts;
+
+ device_set_wakeup_enable(&adapter->pdev->dev, adapter->wol);
+
+ return 0;
+}
+
+
+static int
+vmxnet3_get_settings(struct net_device *netdev, struct ethtool_cmd *ecmd)
+{
+ struct vmxnet3_adapter *adapter = netdev_priv(netdev);
+
+ ecmd->supported = SUPPORTED_10000baseT_Full | SUPPORTED_1000baseT_Full |
+ SUPPORTED_TP;
+ ecmd->advertising = ADVERTISED_TP;
+ ecmd->port = PORT_TP;
+ ecmd->transceiver = XCVR_INTERNAL;
+
+ if (adapter->link_speed) {
+ ecmd->speed = adapter->link_speed;
+ ecmd->duplex = DUPLEX_FULL;
+ } else {
+ ecmd->speed = -1;
+ ecmd->duplex = -1;
+ }
+ return 0;
+}
+
+
+static void
+vmxnet3_get_ringparam(struct net_device *netdev,
+ struct ethtool_ringparam *param)
+{
+ struct vmxnet3_adapter *adapter = netdev_priv(netdev);
+
+ param->rx_max_pending = VMXNET3_RX_RING_MAX_SIZE;
+ param->tx_max_pending = VMXNET3_TX_RING_MAX_SIZE;
+ param->rx_mini_max_pending = 0;
+ param->rx_jumbo_max_pending = 0;
+
+ param->rx_pending = adapter->rx_queue.rx_ring[0].size;
+ param->tx_pending = adapter->tx_queue.tx_ring.size;
+ param->rx_mini_pending = 0;
+ param->rx_jumbo_pending = 0;
+}
+
+
+static int
+vmxnet3_set_ringparam(struct net_device *netdev,
+ struct ethtool_ringparam *param)
+{
+ struct vmxnet3_adapter *adapter = netdev_priv(netdev);
+ u32 new_tx_ring_size, new_rx_ring_size;
+ u32 sz;
+ int err = 0;
+
+ if (param->tx_pending == 0 || param->tx_pending >
+ VMXNET3_TX_RING_MAX_SIZE) {
+ printk(KERN_ERR "%s: invalid tx ring size %u\n", netdev->name,
+ param->tx_pending);
+ return -EINVAL;
+ }
+ if (param->rx_pending == 0 || param->rx_pending >
+ VMXNET3_RX_RING_MAX_SIZE) {
+ printk(KERN_ERR "%s: invalid rx ring size %u\n", netdev->name,
+ param->rx_pending);
+ return -EINVAL;
+ }
+
+ /* round it up to a multiple of VMXNET3_RING_SIZE_ALIGN */
+ new_tx_ring_size = (param->tx_pending + VMXNET3_RING_SIZE_MASK) &
+ ~VMXNET3_RING_SIZE_MASK;
+ new_tx_ring_size = min_t(u32, new_tx_ring_size,
+ VMXNET3_TX_RING_MAX_SIZE);
+ BUG_ON(new_tx_ring_size > VMXNET3_TX_RING_MAX_SIZE);
+ BUG_ON(new_tx_ring_size % VMXNET3_RING_SIZE_ALIGN != 0);
+
+ /* ring0 has to be a multiple of
+ * rx_buf_per_pkt * VMXNET3_RING_SIZE_ALIGN
+ */
+ sz = adapter->rx_buf_per_pkt * VMXNET3_RING_SIZE_ALIGN;
+ new_rx_ring_size = (param->rx_pending + sz - 1) / sz * sz;
+ new_rx_ring_size = min_t(u32, new_rx_ring_size,
+ VMXNET3_RX_RING_MAX_SIZE / sz * sz);
+ BUG_ON(new_rx_ring_size > VMXNET3_RX_RING_MAX_SIZE);
+ BUG_ON(new_rx_ring_size % sz != 0);
+
+ if (new_tx_ring_size == adapter->tx_queue.tx_ring.size &&
+ new_rx_ring_size == adapter->rx_queue.rx_ring[0].size) {
+ return 0;
+ }
+
+ /*
+ * Reset_work may be in the middle of resetting the device, wait for its
+ * completion.
+ */
+ while (test_and_set_bit(VMXNET3_STATE_BIT_RESETTING, &adapter->state))
+ msleep(1);
+
+ if (netif_running(netdev)) {
+ vmxnet3_quiesce_dev(adapter);
+ vmxnet3_reset_dev(adapter);
+
+ /* recreate the rx queue and the tx queue based on the
+ * new sizes */
+ vmxnet3_tq_destroy(&adapter->tx_queue, adapter);
+ vmxnet3_rq_destroy(&adapter->rx_queue, adapter);
+
+ err = vmxnet3_create_queues(adapter, new_tx_ring_size,
+ new_rx_ring_size, VMXNET3_DEF_RX_RING_SIZE);
+ if (err) {
+ /* failed, most likely because of OOM, try default
+ * size */
+ printk(KERN_ERR "%s: failed to apply new sizes, try the"
+ " default ones\n", netdev->name);
+ err = vmxnet3_create_queues(adapter,
+ VMXNET3_DEF_TX_RING_SIZE,
+ VMXNET3_DEF_RX_RING_SIZE,
+ VMXNET3_DEF_RX_RING_SIZE);
+ if (err) {
+ printk(KERN_ERR "%s: failed to create queues "
+ "with default sizes. Closing it\n",
+ netdev->name);
+ goto out;
+ }
+ }
+
+ err = vmxnet3_activate_dev(adapter);
+ if (err) {
+ printk(KERN_ERR "%s: failed to re-activate, error %d."
+ " Closing it\n", netdev->name, err);
+ goto out;
+ }
+ }
+
+out:
+ clear_bit(VMXNET3_STATE_BIT_RESETTING, &adapter->state);
+ if (err)
+ vmxnet3_force_close(adapter);
+
+ return err;
+}
+
+
+static struct ethtool_ops vmxnet3_ethtool_ops = {
+ .get_settings = vmxnet3_get_settings,
+ .get_drvinfo = vmxnet3_get_drvinfo,
+ .get_regs_len = vmxnet3_get_regs_len,
+ .get_regs = vmxnet3_get_regs,
+ .get_wol = vmxnet3_get_wol,
+ .set_wol = vmxnet3_set_wol,
+ .get_link = ethtool_op_get_link,
+ .get_rx_csum = vmxnet3_get_rx_csum,
+ .set_rx_csum = vmxnet3_set_rx_csum,
+ .get_tx_csum = vmxnet3_get_tx_csum,
+ .set_tx_csum = vmxnet3_set_tx_csum,
+ .get_sg = ethtool_op_get_sg,
+ .set_sg = vmxnet3_set_sg,
+ .get_tso = ethtool_op_get_tso,
+ .set_tso = vmxnet3_set_tso,
+ .get_strings = vmxnet3_get_strings,
+ .get_stats_count = vmxnet3_get_stats_count,
+ .get_ethtool_stats = vmxnet3_get_ethtool_stats,
+ .get_ringparam = vmxnet3_get_ringparam,
+ .set_ringparam = vmxnet3_set_ringparam,
+};
+
+void vmxnet3_set_ethtool_ops(struct net_device *netdev)
+{
+ SET_ETHTOOL_OPS(netdev, &vmxnet3_ethtool_ops);
+}
diff --git a/drivers/net/vmxnet3/vmxnet3_int.h b/drivers/net/vmxnet3/vmxnet3_int.h
new file mode 100644
index 0000000..c33d3d1
--- /dev/null
+++ b/drivers/net/vmxnet3/vmxnet3_int.h
@@ -0,0 +1,390 @@
+/*
+ * Linux driver for VMware's vmxnet3 ethernet NIC.
+ *
+ * Copyright (C) 2008-2009, VMware, Inc. All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License as published by the
+ * Free Software Foundation; version 2 of the License and no later version.
+ *
+ * This program 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, GOOD TITLE or
+ * NON INFRINGEMENT. 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, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
+ *
+ * The full GNU General Public License is included in this distribution in
+ * the file called "COPYING".
+ *
+ * Maintained by: Shreyas Bhatewara <pv-drivers@vmware.com>
+ *
+ */
+
+#ifndef _VMXNET3_INT_H
+#define _VMXNET3_INT_H
+
+#include <linux/types.h>
+#include <linux/ethtool.h>
+#include <linux/delay.h>
+#include <linux/netdevice.h>
+#include <linux/pci.h>
+#include <linux/ethtool.h>
+#include <linux/compiler.h>
+#include <linux/module.h>
+#include <linux/moduleparam.h>
+#include <linux/slab.h>
+#include <linux/spinlock.h>
+#include <linux/ioport.h>
+#include <linux/highmem.h>
+#include <linux/init.h>
+#include <linux/timer.h>
+#include <linux/skbuff.h>
+#include <linux/interrupt.h>
+#include <linux/workqueue.h>
+#include <linux/uaccess.h>
+#include <asm/dma.h>
+#include <asm/page.h>
+
+#include <linux/tcp.h>
+#include <linux/udp.h>
+#include <linux/ip.h>
+#include <linux/ipv6.h>
+#include <linux/in.h>
+#include <linux/etherdevice.h>
+#include <asm/checksum.h>
+#include <linux/if_vlan.h>
+#include <linux/if_arp.h>
+#include <linux/inetdevice.h>
+#include <linux/dst.h>
+
+#include "vmxnet3_defs.h"
+
+#ifdef DEBUG
+# define VMXNET3_DRIVER_VERSION_REPORT VMXNET3_DRIVER_VERSION_STRING"-NAPI(debug)"
+#else
+# define VMXNET3_DRIVER_VERSION_REPORT VMXNET3_DRIVER_VERSION_STRING"-NAPI"
+#endif
+
+
+/*
+ * Version numbers
+ */
+#define VMXNET3_DRIVER_VERSION_STRING "1.0.4.0-k"
+
+/* a 32-bit int, each byte encode a verion number in VMXNET3_DRIVER_VERSION */
+#define VMXNET3_DRIVER_VERSION_NUM 0x01000400
+
+
+/*
+ * Capabilities
+ */
+
+enum {
+ VMNET_CAP_SG = 0x0001, /* Can do scatter-gather transmits. */
+ VMNET_CAP_IP4_CSUM = 0x0002, /* Can checksum only TCP/UDP over
+ * IPv4 */
+ VMNET_CAP_HW_CSUM = 0x0004, /* Can checksum all packets. */
+ VMNET_CAP_HIGH_DMA = 0x0008, /* Can DMA to high memory. */
+ VMNET_CAP_TOE = 0x0010, /* Supports TCP/IP offload. */
+ VMNET_CAP_TSO = 0x0020, /* Supports TCP Segmentation
+ * offload */
+ VMNET_CAP_SW_TSO = 0x0040, /* Supports SW TCP Segmentation */
+ VMNET_CAP_VMXNET_APROM = 0x0080, /* Vmxnet APROM support */
+ VMNET_CAP_HW_TX_VLAN = 0x0100, /* Can we do VLAN tagging in HW */
+ VMNET_CAP_HW_RX_VLAN = 0x0200, /* Can we do VLAN untagging in HW */
+ VMNET_CAP_SW_VLAN = 0x0400, /* VLAN tagging/untagging in SW */
+ VMNET_CAP_WAKE_PCKT_RCV = 0x0800, /* Can wake on network packet recv? */
+ VMNET_CAP_ENABLE_INT_INLINE = 0x1000, /* Enable Interrupt Inline */
+ VMNET_CAP_ENABLE_HEADER_COPY = 0x2000, /* copy header for vmkernel */
+ VMNET_CAP_TX_CHAIN = 0x4000, /* Guest can use multiple tx entries
+ * for a pkt */
+ VMNET_CAP_RX_CHAIN = 0x8000, /* pkt can span multiple rx entries */
+ VMNET_CAP_LPD = 0x10000, /* large pkt delivery */
+ VMNET_CAP_BPF = 0x20000, /* BPF Support in VMXNET Virtual HW*/
+ VMNET_CAP_SG_SPAN_PAGES = 0x40000, /* Scatter-gather can span multiple*/
+ /* pages transmits */
+ VMNET_CAP_IP6_CSUM = 0x80000, /* Can do IPv6 csum offload. */
+ VMNET_CAP_TSO6 = 0x100000, /* TSO seg. offload for IPv6 pkts. */
+ VMNET_CAP_TSO256k = 0x200000, /* Can do TSO seg offload for */
+ /* pkts up to 256kB. */
+ VMNET_CAP_UPT = 0x400000 /* Support UPT */
+};
+
+/*
+ * PCI vendor and device IDs.
+ */
+#define PCI_VENDOR_ID_VMWARE 0x15AD
+#define PCI_DEVICE_ID_VMWARE_VMXNET3 0x07B0
+#define MAX_ETHERNET_CARDS 10
+#define MAX_PCI_PASSTHRU_DEVICE 6
+
+struct vmxnet3_cmd_ring {
+ union Vmxnet3_GenericDesc *base;
+ u32 size;
+ u32 next2fill;
+ u32 next2comp;
+ u8 gen;
+ dma_addr_t basePA;
+};
+
+static inline void
+vmxnet3_cmd_ring_adv_next2fill(struct vmxnet3_cmd_ring *ring)
+{
+ ring->next2fill++;
+ if (unlikely(ring->next2fill == ring->size)) {
+ ring->next2fill = 0;
+ VMXNET3_FLIP_RING_GEN(ring->gen);
+ }
+}
+
+static inline void
+vmxnet3_cmd_ring_adv_next2comp(struct vmxnet3_cmd_ring *ring)
+{
+ VMXNET3_INC_RING_IDX_ONLY(ring->next2comp, ring->size);
+}
+
+static inline int
+vmxnet3_cmd_ring_desc_avail(struct vmxnet3_cmd_ring *ring)
+{
+ return (ring->next2comp > ring->next2fill ? 0 : ring->size) +
+ ring->next2comp - ring->next2fill - 1;
+}
+
+struct vmxnet3_comp_ring {
+ union Vmxnet3_GenericDesc *base;
+ u32 size;
+ u32 next2proc;
+ u8 gen;
+ u8 intr_idx;
+ dma_addr_t basePA;
+};
+
+static inline void
+vmxnet3_comp_ring_adv_next2proc(struct vmxnet3_comp_ring *ring)
+{
+ ring->next2proc++;
+ if (unlikely(ring->next2proc == ring->size)) {
+ ring->next2proc = 0;
+ VMXNET3_FLIP_RING_GEN(ring->gen);
+ }
+}
+
+struct vmxnet3_tx_data_ring {
+ struct Vmxnet3_TxDataDesc *base;
+ u32 size;
+ dma_addr_t basePA;
+};
+
+enum vmxnet3_buf_map_type {
+ VMXNET3_MAP_INVALID = 0,
+ VMXNET3_MAP_NONE,
+ VMXNET3_MAP_SINGLE,
+ VMXNET3_MAP_PAGE,
+};
+
+struct vmxnet3_tx_buf_info {
+ u32 map_type;
+ u16 len;
+ u16 sop_idx;
+ dma_addr_t dma_addr;
+ struct sk_buff *skb;
+};
+
+struct vmxnet3_tq_driver_stats {
+ u64 drop_total; /* # of pkts dropped by the driver, the
+ * counters below track droppings due to
+ * different reasons
+ */
+ u64 drop_too_many_frags;
+ u64 drop_oversized_hdr;
+ u64 drop_hdr_inspect_err;
+ u64 drop_tso;
+
+ u64 tx_ring_full;
+ u64 linearized; /* # of pkts linearized */
+ u64 copy_skb_header; /* # of times we have to copy skb header */
+ u64 oversized_hdr;
+};
+
+struct vmxnet3_tx_ctx {
+ bool ipv4;
+ u16 mss;
+ u32 eth_ip_hdr_size; /* only valid for pkts requesting tso or csum
+ * offloading
+ */
+ u32 l4_hdr_size; /* only valid if mss != 0 */
+ u32 copy_size; /* # of bytes copied into the data ring */
+ union Vmxnet3_GenericDesc *sop_txd;
+ union Vmxnet3_GenericDesc *eop_txd;
+};
+
+struct vmxnet3_tx_queue {
+ spinlock_t tx_lock;
+ struct vmxnet3_cmd_ring tx_ring;
+ struct vmxnet3_tx_buf_info *buf_info;
+ struct vmxnet3_tx_data_ring data_ring;
+ struct vmxnet3_comp_ring comp_ring;
+ struct Vmxnet3_TxQueueCtrl *shared;
+ struct vmxnet3_tq_driver_stats stats;
+ bool stopped;
+ int num_stop; /* # of times the queue is
+ * stopped */
+} __attribute__((__aligned__(SMP_CACHE_BYTES)));
+
+enum vmxnet3_rx_buf_type {
+ VMXNET3_RX_BUF_NONE = 0,
+ VMXNET3_RX_BUF_SKB = 1,
+ VMXNET3_RX_BUF_PAGE = 2
+};
+
+struct vmxnet3_rx_buf_info {
+ enum vmxnet3_rx_buf_type buf_type;
+ u16 len;
+ union {
+ struct sk_buff *skb;
+ struct page *page;
+ };
+ dma_addr_t dma_addr;
+};
+
+struct vmxnet3_rx_ctx {
+ struct sk_buff *skb;
+ u32 sop_idx;
+};
+
+struct vmxnet3_rq_driver_stats {
+ u64 drop_total;
+ u64 drop_err;
+ u64 drop_fcs;
+ u64 rx_buf_alloc_failure;
+};
+
+struct vmxnet3_rx_queue {
+ struct vmxnet3_cmd_ring rx_ring[2];
+ struct vmxnet3_comp_ring comp_ring;
+ struct vmxnet3_rx_ctx rx_ctx;
+ u32 qid; /* rqID in RCD for buffer from 1st ring */
+ u32 qid2; /* rqID in RCD for buffer from 2nd ring */
+ u32 uncommitted[2]; /* # of buffers allocated since last RXPROD
+ * update */
+ struct vmxnet3_rx_buf_info *buf_info[2];
+ struct Vmxnet3_RxQueueCtrl *shared;
+ struct vmxnet3_rq_driver_stats stats;
+} __attribute__((__aligned__(SMP_CACHE_BYTES)));
+
+#define VMXNET3_LINUX_MAX_MSIX_VECT 1
+
+struct vmxnet3_intr {
+ enum vmxnet3_intr_mask_mode mask_mode;
+ enum vmxnet3_intr_type type; /* MSI-X, MSI, or INTx? */
+ u8 num_intrs; /* # of intr vectors */
+ u8 event_intr_idx; /* idx of the intr vector for event */
+ u8 mod_levels[VMXNET3_LINUX_MAX_MSIX_VECT]; /* moderation level */
+#ifdef CONFIG_PCI_MSI
+ struct msix_entry msix_entries[VMXNET3_LINUX_MAX_MSIX_VECT];
+#endif
+};
+
+#define VMXNET3_STATE_BIT_RESETTING 0
+#define VMXNET3_STATE_BIT_QUIESCED 1
+struct vmxnet3_adapter {
+ struct vmxnet3_tx_queue tx_queue;
+ struct vmxnet3_rx_queue rx_queue;
+ struct napi_struct napi;
+ struct vlan_group *vlan_grp;
+
+ struct vmxnet3_intr intr;
+
+ struct Vmxnet3_DriverShared *shared;
+ struct Vmxnet3_PMConf *pm_conf;
+ struct Vmxnet3_TxQueueDesc *tqd_start; /* first tx queue desc */
+ struct Vmxnet3_RxQueueDesc *rqd_start; /* first rx queue desc */
+ struct net_device *netdev;
+ struct net_device_stats net_stats;
+ struct pci_dev *pdev;
+
+ u8 *hw_addr0; /* for BAR 0 */
+ u8 *hw_addr1; /* for BAR 1 */
+
+ /* feature control */
+ bool rxcsum;
+ bool lro;
+ bool jumbo_frame;
+
+ /* rx buffer related */
+ unsigned skb_buf_size;
+ int rx_buf_per_pkt; /* only apply to the 1st ring */
+ dma_addr_t shared_pa;
+ dma_addr_t queue_desc_pa;
+
+ /* Wake-on-LAN */
+ u32 wol;
+
+ /* Link speed */
+ u32 link_speed; /* in mbps */
+
+ u64 tx_timeout_count;
+ struct work_struct work;
+
+ unsigned long state; /* VMXNET3_STATE_BIT_xxx */
+
+ int dev_number;
+};
+
+#define VMXNET3_WRITE_BAR0_REG(adapter, reg, val) \
+ writel((val), (adapter)->hw_addr0 + (reg))
+#define VMXNET3_READ_BAR0_REG(adapter, reg) \
+ readl((adapter)->hw_addr0 + (reg))
+
+#define VMXNET3_WRITE_BAR1_REG(adapter, reg, val) \
+ writel((val), (adapter)->hw_addr1 + (reg))
+#define VMXNET3_READ_BAR1_REG(adapter, reg) \
+ readl((adapter)->hw_addr1 + (reg))
+
+#define VMXNET3_WAKE_QUEUE_THRESHOLD(tq) (5)
+#define VMXNET3_RX_ALLOC_THRESHOLD(rq, ring_idx, adapter) \
+ ((rq)->rx_ring[ring_idx].size >> 3)
+
+#define VMXNET3_GET_ADDR_LO(dma) ((u32)(dma))
+#define VMXNET3_GET_ADDR_HI(dma) ((u32)(((u64)(dma)) >> 32))
+
+/* must be a multiple of VMXNET3_RING_SIZE_ALIGN */
+#define VMXNET3_DEF_TX_RING_SIZE 512
+#define VMXNET3_DEF_RX_RING_SIZE 256
+
+#define VMXNET3_MAX_ETH_HDR_SIZE 22
+#define VMXNET3_MAX_SKB_BUF_SIZE (3*1024)
+
+int
+vmxnet3_quiesce_dev(struct vmxnet3_adapter *adapter);
+
+int
+vmxnet3_activate_dev(struct vmxnet3_adapter *adapter);
+
+void
+vmxnet3_force_close(struct vmxnet3_adapter *adapter);
+
+void
+vmxnet3_reset_dev(struct vmxnet3_adapter *adapter);
+
+void
+vmxnet3_tq_destroy(struct vmxnet3_tx_queue *tq,
+ struct vmxnet3_adapter *adapter);
+
+void
+vmxnet3_rq_destroy(struct vmxnet3_rx_queue *rq,
+ struct vmxnet3_adapter *adapter);
+
+int
+vmxnet3_create_queues(struct vmxnet3_adapter *adapter,
+ u32 tx_ring_size, u32 rx_ring_size, u32 rx_ring2_size);
+
+extern void vmxnet3_set_ethtool_ops(struct net_device *netdev);
+extern struct net_device_stats *vmxnet3_get_stats(struct net_device *netdev);
+
+extern char vmxnet3_driver_name[];
+#endif
^ permalink raw reply related
* Re: [PATCH] /proc/net/tcp, overhead removed
From: Stephen Hemminger @ 2009-09-28 23:24 UTC (permalink / raw)
To: Eric Dumazet; +Cc: Yakov Lerner, netdev, Eric Dumazet, David Miller
In-Reply-To: <4AC13697.4090707@gmail.com>
On Tue, 29 Sep 2009 00:20:07 +0200
Eric Dumazet <eric.dumazet@gmail.com> wrote:
> Yakov Lerner a écrit :
> > On Sun, Sep 27, 2009 at 12:53, Eric Dumazet <eric.dumazet@gmail.com> wrote:
> >> Yakov Lerner a écrit :
> >>> /proc/net/tcp does 20,000 sockets in 60-80 milliseconds, with this patch.
> >>>
> >>> The overhead was in tcp_seq_start(). See analysis (3) below.
> >>> The patch is against Linus git tree (1). The patch is small.
> >>>
> >>> ------------ ----------- ------------------------------------
> >>> Before patch After patch 20,000 sockets (10,000 tw + 10,000 estab)(2)
> >>> ------------ ----------- ------------------------------------
> >>> 6 sec 0.06 sec dd bs=1k if=/proc/net/tcp >/dev/null
> >>> 1.5 sec 0.06 sec dd bs=4k if=/proc/net/tcp >/dev/null
> >>>
> >>> 1.9 sec 0.16 sec netstat -4ant >/dev/null
> >>> ------------ ----------- ------------------------------------
> >>>
> >>> This is ~ x25 improvement.
> >>> The new time is not dependent on read blockize.
> >>> Speed of netstat, naturally, improves, too; both -4 and -6.
> >>> /proc/net/tcp6 does 20,000 sockets in 100 millisec.
> >>>
> >>> (1) against git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git
> >>>
> >>> (2) Used 'manysock' utility to stress system with large number of sockets:
> >>> "manysock 10000 10000" - 10,000 tw + 10,000 estab ip4 sockets.
> >>> "manysock -6 10000 10000" - 10,000 tw + 10,000 estab ip6 sockets.
> >>> Found at http://ilerner.3b1.org/manysock/manysock.c
> >>>
> >>> (3) Algorithmic analysis.
> >>> Old algorithm.
> >>>
> >>> During 'cat </proc/net/tcp', tcp_seq_start() is called O(numsockets) times (4).
> >>> On average, every call to tcp_seq_start() scans half the whole hashtable. Ouch.
> >>> This is O(numsockets * hashsize). 95-99% of 'cat </proc/net/tcp' is spent in
> >>> tcp_seq_start()->tcp_get_idx. This overhead is eliminated by new algorithm,
> >>> which is O(numsockets + hashsize).
> >>>
> >>> New algorithm.
> >>>
> >>> New algorithms is O(numsockets + hashsize). We jump to the right
> >>> hash bucket in tcp_seq_start(), without scanning half the hash.
> >>> To jump right to the hash bucket corresponding to *pos in tcp_seq_start(),
> >>> we reuse three pieces of state (st->num, st->bucket, st->sbucket)
> >>> as follows:
> >>> - we check that requested pos >= last seen pos (st->num), the typical case.
> >>> - if so, we jump to bucket st->bucket
> >>> - to arrive to the right item after beginning of st->bucket, we
> >>> keep in st->sbucket the position corresponding to the beginning of
> >>> bucket.
> >>>
> >>> (4) Explanation of O( numsockets * hashsize) of old algorithm.
> >>>
> >>> tcp_seq_start() is called once for every ~7 lines of netstat output
> >>> if readsize is 1kb, or once for every ~28 lines if readsize >= 4kb.
> >>> Since record length of /proc/net/tcp records is 150 bytes, formula for
> >>> number of calls to tcp_seq_start() is
> >>> (numsockets * 150 / min(4096,readsize)).
> >>> Netstat uses 4kb readsize (newer versions), or 1kb (older versions).
> >>> Note that speed of old algorithm does not improve above 4kb blocksize.
> >>>
> >>> Speed of the new algorithm does not depend on blocksize.
> >>>
> >>> Speed of the new algorithm does not perceptibly depend on hashsize (which
> >>> depends on ramsize). Speed of old algorithm drops with bigger hashsize.
> >>>
> >>> (5) Reporting order.
> >>>
> >>> Reporting order is exactly same as before if hash does not change underfoot.
> >>> When hash elements come and go during report, reporting order will be
> >>> same as that of tcpdiag.
> >>>
> >>> Signed-off-by: Yakov Lerner <iler.ml@gmail.com>
Does the netlink interface used by ss command have the problem?
--
^ permalink raw reply
* Re: [Bonding-devel] [PATCH 4/4] bonding: add sysfs files to display tlb and alb hash table contents
From: Stephen Hemminger @ 2009-09-28 23:22 UTC (permalink / raw)
To: Andy Gospodarek; +Cc: netdev, fubar, bonding-devel
In-Reply-To: <20090911211317.GT8515@gospo.rdu.redhat.com>
On Fri, 11 Sep 2009 17:13:17 -0400
Andy Gospodarek <andy@greyhouse.net> wrote:
>
> bonding: add sysfs files to display tlb and alb hash table contents
>
> While debugging some problems with alb (mode 6) bonding I realized that
> being able to output the contents of both hash tables would be helpful.
> This is what the output looks like for the two files:
>
> device load
> eth1 491
> eth2 491
> hash device last device tx bytes load next previous
> 2 eth1 eth1 2254 491 0 0
> 3 eth2 eth2 2744 491 0 0
> 6 eth2 0 488 0 0
> 8 eth2 0 461698 0 0
> 1b eth2 0 249 0 0
> eb eth2 0 21 0 0
> ff eth2 0 22 0 0
>
> hash ip_src ip_dst mac_dst slave assign ntt
> 2 10.0.3.2 10.0.3.11 00:e0:81:71:ee:a9 eth1 1 0
> 3 10.0.3.2 10.0.3.10 00:e0:81:71:ee:a9 eth2 1 0
> 8 10.0.3.2 10.0.3.1 00:e0:81:71:ee:a9 eth2 1 0
>
> These were a great help debugging the fixes I have just posted and they
> might be helpful for others, so I decided to include them in my
> patchset.
>
> Signed-off-by: Andy Gospodarek <andy@greyhouse.net>
No.
Please don't put formatted output in sysfs, it is not meant to be
used like proc, there is supposed to be only one value per file.
Maybe put it on the end of the /proc/net/bonding/bond0 output or
use debugfs
^ permalink raw reply
* [PATCH] /proc/net/tcp, overhead removed
From: Yakov Lerner @ 2009-09-28 23:01 UTC (permalink / raw)
To: netdev, eric.dumazet, davem; +Cc: Yakov Lerner
Take 2.
"Sharp improvement in performance of /proc/net/tcp when number of
sockets is large and hashsize is large.
O(numsock * hashsize) time becomes O(numsock + hashsize). On slow
processors, speed difference can be x100 and more."
I must say that I'm not fully satisfied with my choice of "st->sbucket"
for the new preserved index. The better name would be "st->snum".
Re-using "st->sbucket" saves 4 bytes, and keeps the patch to one sourcefile.
But "st->sbucket" has different meaning in OPENREQ and LISTEN states;
this can be confusing.
Maybe better add "snum" member to struct tcp_iter_state ?
Shall I change subject when sending "take N+1", or keep the old subject ?
Signed-off-by: Yakov Lerner <iler.ml@gmail.com>
---
net/ipv4/tcp_ipv4.c | 35 +++++++++++++++++++++++++++++++++--
1 files changed, 33 insertions(+), 2 deletions(-)
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 7cda24b..e4c4f19 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -1994,13 +1994,14 @@ static inline int empty_bucket(struct tcp_iter_state *st)
hlist_nulls_empty(&tcp_hashinfo.ehash[st->bucket].twchain);
}
-static void *established_get_first(struct seq_file *seq)
+static void *established_get_first_after(struct seq_file *seq, int bucket)
{
struct tcp_iter_state *st = seq->private;
struct net *net = seq_file_net(seq);
void *rc = NULL;
- for (st->bucket = 0; st->bucket < tcp_hashinfo.ehash_size; ++st->bucket) {
+ for (st->bucket = bucket; st->bucket < tcp_hashinfo.ehash_size;
+ ++st->bucket) {
struct sock *sk;
struct hlist_nulls_node *node;
struct inet_timewait_sock *tw;
@@ -2010,6 +2011,8 @@ static void *established_get_first(struct seq_file *seq)
if (empty_bucket(st))
continue;
+ st->sbucket = st->num;
+
spin_lock_bh(lock);
sk_nulls_for_each(sk, node, &tcp_hashinfo.ehash[st->bucket].chain) {
if (sk->sk_family != st->family ||
@@ -2036,6 +2039,11 @@ out:
return rc;
}
+static void *established_get_first(struct seq_file *seq)
+{
+ return established_get_first_after(seq, 0);
+}
+
static void *established_get_next(struct seq_file *seq, void *cur)
{
struct sock *sk = cur;
@@ -2064,6 +2072,9 @@ get_tw:
while (++st->bucket < tcp_hashinfo.ehash_size &&
empty_bucket(st))
;
+
+ st->sbucket = st->num;
+
if (st->bucket >= tcp_hashinfo.ehash_size)
return NULL;
@@ -2107,6 +2118,7 @@ static void *tcp_get_idx(struct seq_file *seq, loff_t pos)
if (!rc) {
st->state = TCP_SEQ_STATE_ESTABLISHED;
+ st->sbucket = 0;
rc = established_get_idx(seq, pos);
}
@@ -2116,6 +2128,25 @@ static void *tcp_get_idx(struct seq_file *seq, loff_t pos)
static void *tcp_seq_start(struct seq_file *seq, loff_t *pos)
{
struct tcp_iter_state *st = seq->private;
+
+ if (*pos && *pos >= st->sbucket &&
+ (st->state == TCP_SEQ_STATE_ESTABLISHED ||
+ st->state == TCP_SEQ_STATE_TIME_WAIT)) {
+ void *cur;
+ int nskip;
+
+ /* for states estab and tw, st->sbucket is index (*pos) */
+ /* corresponding to the beginning of bucket st->bucket */
+
+ st->num = st->sbucket;
+ /* jump to st->bucket, then skip (*pos - st->sbucket) items */
+ st->state = TCP_SEQ_STATE_ESTABLISHED;
+ cur = established_get_first_after(seq, st->bucket);
+ for (nskip = *pos - st->num; cur && nskip > 0; --nskip)
+ cur = established_get_next(seq, cur);
+ return cur;
+ }
+
st->state = TCP_SEQ_STATE_LISTENING;
st->num = 0;
return *pos ? tcp_get_idx(seq, *pos - 1) : SEQ_START_TOKEN;
--
1.6.5.rc2
^ permalink raw reply related
* Re: [PATCH] /proc/net/tcp, overhead removed
From: Eric Dumazet @ 2009-09-28 22:20 UTC (permalink / raw)
To: Yakov Lerner; +Cc: netdev, Eric Dumazet, David Miller
In-Reply-To: <f36b08ee0909281510y282d621etb4264ecd92cbe8f0@mail.gmail.com>
Yakov Lerner a écrit :
> On Sun, Sep 27, 2009 at 12:53, Eric Dumazet <eric.dumazet@gmail.com> wrote:
>> Yakov Lerner a écrit :
>>> /proc/net/tcp does 20,000 sockets in 60-80 milliseconds, with this patch.
>>>
>>> The overhead was in tcp_seq_start(). See analysis (3) below.
>>> The patch is against Linus git tree (1). The patch is small.
>>>
>>> ------------ ----------- ------------------------------------
>>> Before patch After patch 20,000 sockets (10,000 tw + 10,000 estab)(2)
>>> ------------ ----------- ------------------------------------
>>> 6 sec 0.06 sec dd bs=1k if=/proc/net/tcp >/dev/null
>>> 1.5 sec 0.06 sec dd bs=4k if=/proc/net/tcp >/dev/null
>>>
>>> 1.9 sec 0.16 sec netstat -4ant >/dev/null
>>> ------------ ----------- ------------------------------------
>>>
>>> This is ~ x25 improvement.
>>> The new time is not dependent on read blockize.
>>> Speed of netstat, naturally, improves, too; both -4 and -6.
>>> /proc/net/tcp6 does 20,000 sockets in 100 millisec.
>>>
>>> (1) against git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git
>>>
>>> (2) Used 'manysock' utility to stress system with large number of sockets:
>>> "manysock 10000 10000" - 10,000 tw + 10,000 estab ip4 sockets.
>>> "manysock -6 10000 10000" - 10,000 tw + 10,000 estab ip6 sockets.
>>> Found at http://ilerner.3b1.org/manysock/manysock.c
>>>
>>> (3) Algorithmic analysis.
>>> Old algorithm.
>>>
>>> During 'cat </proc/net/tcp', tcp_seq_start() is called O(numsockets) times (4).
>>> On average, every call to tcp_seq_start() scans half the whole hashtable. Ouch.
>>> This is O(numsockets * hashsize). 95-99% of 'cat </proc/net/tcp' is spent in
>>> tcp_seq_start()->tcp_get_idx. This overhead is eliminated by new algorithm,
>>> which is O(numsockets + hashsize).
>>>
>>> New algorithm.
>>>
>>> New algorithms is O(numsockets + hashsize). We jump to the right
>>> hash bucket in tcp_seq_start(), without scanning half the hash.
>>> To jump right to the hash bucket corresponding to *pos in tcp_seq_start(),
>>> we reuse three pieces of state (st->num, st->bucket, st->sbucket)
>>> as follows:
>>> - we check that requested pos >= last seen pos (st->num), the typical case.
>>> - if so, we jump to bucket st->bucket
>>> - to arrive to the right item after beginning of st->bucket, we
>>> keep in st->sbucket the position corresponding to the beginning of
>>> bucket.
>>>
>>> (4) Explanation of O( numsockets * hashsize) of old algorithm.
>>>
>>> tcp_seq_start() is called once for every ~7 lines of netstat output
>>> if readsize is 1kb, or once for every ~28 lines if readsize >= 4kb.
>>> Since record length of /proc/net/tcp records is 150 bytes, formula for
>>> number of calls to tcp_seq_start() is
>>> (numsockets * 150 / min(4096,readsize)).
>>> Netstat uses 4kb readsize (newer versions), or 1kb (older versions).
>>> Note that speed of old algorithm does not improve above 4kb blocksize.
>>>
>>> Speed of the new algorithm does not depend on blocksize.
>>>
>>> Speed of the new algorithm does not perceptibly depend on hashsize (which
>>> depends on ramsize). Speed of old algorithm drops with bigger hashsize.
>>>
>>> (5) Reporting order.
>>>
>>> Reporting order is exactly same as before if hash does not change underfoot.
>>> When hash elements come and go during report, reporting order will be
>>> same as that of tcpdiag.
>>>
>>> Signed-off-by: Yakov Lerner <iler.ml@gmail.com>
>>> ---
>>> net/ipv4/tcp_ipv4.c | 26 ++++++++++++++++++++++++--
>>> 1 files changed, 24 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
>>> index 7cda24b..7d9421a 100644
>>> --- a/net/ipv4/tcp_ipv4.c
>>> +++ b/net/ipv4/tcp_ipv4.c
>>> @@ -1994,13 +1994,14 @@ static inline int empty_bucket(struct tcp_iter_state *st)
>>> hlist_nulls_empty(&tcp_hashinfo.ehash[st->bucket].twchain);
>>> }
>>>
>>> -static void *established_get_first(struct seq_file *seq)
>>> +static void *established_get_first_after(struct seq_file *seq, int bucket)
>>> {
>>> struct tcp_iter_state *st = seq->private;
>>> struct net *net = seq_file_net(seq);
>>> void *rc = NULL;
>>>
>>> - for (st->bucket = 0; st->bucket < tcp_hashinfo.ehash_size; ++st->bucket) {
>>> + for (st->bucket = bucket; st->bucket < tcp_hashinfo.ehash_size;
>>> + ++st->bucket) {
>>> struct sock *sk;
>>> struct hlist_nulls_node *node;
>>> struct inet_timewait_sock *tw;
>>> @@ -2036,6 +2037,11 @@ out:
>>> return rc;
>>> }
>>>
>>> +static void *established_get_first(struct seq_file *seq)
>>> +{
>>> + return established_get_first_after(seq, 0);
>>> +}
>>> +
>>> static void *established_get_next(struct seq_file *seq, void *cur)
>>> {
>>> struct sock *sk = cur;
>>> @@ -2045,6 +2051,7 @@ static void *established_get_next(struct seq_file *seq, void *cur)
>>> struct net *net = seq_file_net(seq);
>>>
>>> ++st->num;
>>> + st->sbucket = st->num;
>> Hello Yakov
>>
>> Intention of your patch is very good, but not currently working.
>>
>> It seems you believe there is at most one entry per hash slot or something like that
>>
>> Please reboot your test machine with "thash_entries=4096" so that tcp hash
>> size is 4096, and try to fill 20000 tcp sockets with a test program.
>>
>> then :
>>
>> # ss | wc -l
>> 20001
>> (ok)
>>
>> # cat /proc/net/tcp | wc -l
>> 22160
>> (not quite correct ...)
>>
>> # netstat -tn | wc -l
>> <never ends>
>>
>>
>> # dd if=/proc/net/tcp ibs=1024 | wc -l
>> <never ends>
>>
>>
>> Please send your next patch on netdev@vger.kernel.org , DaveM only , were netdev people
>> are reviewing netdev patches, there is no need include other people for first submissions.
>>
>> Thank you
>>
>>
>> #include <sys/types.h>
>> #include <sys/socket.h>
>> #include <netinet/in.h>
>> #include <string.h>
>> int fdlisten;
>> main()
>> {
>> int i;
>> struct sockaddr_in sockaddr;
>>
>> fdlisten = socket(AF_INET, SOCK_STREAM, 0);
>> memset(&sockaddr, 0, sizeof(sockaddr));
>> sockaddr.sin_family = AF_INET;
>> sockaddr.sin_port = htons(2222);
>> if (bind(fdlisten, (struct sockaddr *)&sockaddr, sizeof(sockaddr))== -1) {
>> perror("bind");
>> return 1;
>> }
>> if (listen(fdlisten, 10)== -1) {
>> perror("listen");
>> return 1;
>> }
>> if (fork() == 0) {
>> while (1) {
>> socklen_t len = sizeof(sockaddr);
>> int newfd = accept(fdlisten, (struct sockaddr *)&sockaddr, &len);
>> }
>> }
>> for (i = 0 ; i < 10000; i++) {
>> int fd = socket(AF_INET, SOCK_STREAM, 0);
>> if (fd == -1) {
>> perror("socket");
>> break;
>> }
>> connect(fd, (struct sockaddr *)&sockaddr, sizeof(sockaddr));
>> }
>> pause();
>> }
>>
>
> Hello Eric,
>
> I found the problem, thanks. I'll re-send after testing.
OK good !
>
> In the meantime, I'd like to ask you whether it makes sense to
> add the /proc/net entry, to switch between "old way" and "new way".
> The switch would allow quick compare/test between new way and
> old way not only by line count, but by full contents, without reboot.
>
Well, this switch wont be needed for patch validation, but it might help
you to test your patch of course.
Actually I found the error reading your patch, and I made a quick test to
confirm my understanding :)
See you tomorrow, its rather late here :)
^ permalink raw reply
* Re: [PATCH] /proc/net/tcp, overhead removed
From: Yakov Lerner @ 2009-09-28 22:10 UTC (permalink / raw)
To: netdev, Eric Dumazet, David Miller
In-Reply-To: <4ABF360E.7080301@gmail.com>
On Sun, Sep 27, 2009 at 12:53, Eric Dumazet <eric.dumazet@gmail.com> wrote:
> Yakov Lerner a écrit :
>> /proc/net/tcp does 20,000 sockets in 60-80 milliseconds, with this patch.
>>
>> The overhead was in tcp_seq_start(). See analysis (3) below.
>> The patch is against Linus git tree (1). The patch is small.
>>
>> ------------ ----------- ------------------------------------
>> Before patch After patch 20,000 sockets (10,000 tw + 10,000 estab)(2)
>> ------------ ----------- ------------------------------------
>> 6 sec 0.06 sec dd bs=1k if=/proc/net/tcp >/dev/null
>> 1.5 sec 0.06 sec dd bs=4k if=/proc/net/tcp >/dev/null
>>
>> 1.9 sec 0.16 sec netstat -4ant >/dev/null
>> ------------ ----------- ------------------------------------
>>
>> This is ~ x25 improvement.
>> The new time is not dependent on read blockize.
>> Speed of netstat, naturally, improves, too; both -4 and -6.
>> /proc/net/tcp6 does 20,000 sockets in 100 millisec.
>>
>> (1) against git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6.git
>>
>> (2) Used 'manysock' utility to stress system with large number of sockets:
>> "manysock 10000 10000" - 10,000 tw + 10,000 estab ip4 sockets.
>> "manysock -6 10000 10000" - 10,000 tw + 10,000 estab ip6 sockets.
>> Found at http://ilerner.3b1.org/manysock/manysock.c
>>
>> (3) Algorithmic analysis.
>> Old algorithm.
>>
>> During 'cat </proc/net/tcp', tcp_seq_start() is called O(numsockets) times (4).
>> On average, every call to tcp_seq_start() scans half the whole hashtable. Ouch.
>> This is O(numsockets * hashsize). 95-99% of 'cat </proc/net/tcp' is spent in
>> tcp_seq_start()->tcp_get_idx. This overhead is eliminated by new algorithm,
>> which is O(numsockets + hashsize).
>>
>> New algorithm.
>>
>> New algorithms is O(numsockets + hashsize). We jump to the right
>> hash bucket in tcp_seq_start(), without scanning half the hash.
>> To jump right to the hash bucket corresponding to *pos in tcp_seq_start(),
>> we reuse three pieces of state (st->num, st->bucket, st->sbucket)
>> as follows:
>> - we check that requested pos >= last seen pos (st->num), the typical case.
>> - if so, we jump to bucket st->bucket
>> - to arrive to the right item after beginning of st->bucket, we
>> keep in st->sbucket the position corresponding to the beginning of
>> bucket.
>>
>> (4) Explanation of O( numsockets * hashsize) of old algorithm.
>>
>> tcp_seq_start() is called once for every ~7 lines of netstat output
>> if readsize is 1kb, or once for every ~28 lines if readsize >= 4kb.
>> Since record length of /proc/net/tcp records is 150 bytes, formula for
>> number of calls to tcp_seq_start() is
>> (numsockets * 150 / min(4096,readsize)).
>> Netstat uses 4kb readsize (newer versions), or 1kb (older versions).
>> Note that speed of old algorithm does not improve above 4kb blocksize.
>>
>> Speed of the new algorithm does not depend on blocksize.
>>
>> Speed of the new algorithm does not perceptibly depend on hashsize (which
>> depends on ramsize). Speed of old algorithm drops with bigger hashsize.
>>
>> (5) Reporting order.
>>
>> Reporting order is exactly same as before if hash does not change underfoot.
>> When hash elements come and go during report, reporting order will be
>> same as that of tcpdiag.
>>
>> Signed-off-by: Yakov Lerner <iler.ml@gmail.com>
>> ---
>> net/ipv4/tcp_ipv4.c | 26 ++++++++++++++++++++++++--
>> 1 files changed, 24 insertions(+), 2 deletions(-)
>>
>> diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
>> index 7cda24b..7d9421a 100644
>> --- a/net/ipv4/tcp_ipv4.c
>> +++ b/net/ipv4/tcp_ipv4.c
>> @@ -1994,13 +1994,14 @@ static inline int empty_bucket(struct tcp_iter_state *st)
>> hlist_nulls_empty(&tcp_hashinfo.ehash[st->bucket].twchain);
>> }
>>
>> -static void *established_get_first(struct seq_file *seq)
>> +static void *established_get_first_after(struct seq_file *seq, int bucket)
>> {
>> struct tcp_iter_state *st = seq->private;
>> struct net *net = seq_file_net(seq);
>> void *rc = NULL;
>>
>> - for (st->bucket = 0; st->bucket < tcp_hashinfo.ehash_size; ++st->bucket) {
>> + for (st->bucket = bucket; st->bucket < tcp_hashinfo.ehash_size;
>> + ++st->bucket) {
>> struct sock *sk;
>> struct hlist_nulls_node *node;
>> struct inet_timewait_sock *tw;
>> @@ -2036,6 +2037,11 @@ out:
>> return rc;
>> }
>>
>> +static void *established_get_first(struct seq_file *seq)
>> +{
>> + return established_get_first_after(seq, 0);
>> +}
>> +
>> static void *established_get_next(struct seq_file *seq, void *cur)
>> {
>> struct sock *sk = cur;
>> @@ -2045,6 +2051,7 @@ static void *established_get_next(struct seq_file *seq, void *cur)
>> struct net *net = seq_file_net(seq);
>>
>> ++st->num;
>> + st->sbucket = st->num;
>
> Hello Yakov
>
> Intention of your patch is very good, but not currently working.
>
> It seems you believe there is at most one entry per hash slot or something like that
>
> Please reboot your test machine with "thash_entries=4096" so that tcp hash
> size is 4096, and try to fill 20000 tcp sockets with a test program.
>
> then :
>
> # ss | wc -l
> 20001
> (ok)
>
> # cat /proc/net/tcp | wc -l
> 22160
> (not quite correct ...)
>
> # netstat -tn | wc -l
> <never ends>
>
>
> # dd if=/proc/net/tcp ibs=1024 | wc -l
> <never ends>
>
>
> Please send your next patch on netdev@vger.kernel.org , DaveM only , were netdev people
> are reviewing netdev patches, there is no need include other people for first submissions.
>
> Thank you
>
>
> #include <sys/types.h>
> #include <sys/socket.h>
> #include <netinet/in.h>
> #include <string.h>
> int fdlisten;
> main()
> {
> int i;
> struct sockaddr_in sockaddr;
>
> fdlisten = socket(AF_INET, SOCK_STREAM, 0);
> memset(&sockaddr, 0, sizeof(sockaddr));
> sockaddr.sin_family = AF_INET;
> sockaddr.sin_port = htons(2222);
> if (bind(fdlisten, (struct sockaddr *)&sockaddr, sizeof(sockaddr))== -1) {
> perror("bind");
> return 1;
> }
> if (listen(fdlisten, 10)== -1) {
> perror("listen");
> return 1;
> }
> if (fork() == 0) {
> while (1) {
> socklen_t len = sizeof(sockaddr);
> int newfd = accept(fdlisten, (struct sockaddr *)&sockaddr, &len);
> }
> }
> for (i = 0 ; i < 10000; i++) {
> int fd = socket(AF_INET, SOCK_STREAM, 0);
> if (fd == -1) {
> perror("socket");
> break;
> }
> connect(fd, (struct sockaddr *)&sockaddr, sizeof(sockaddr));
> }
> pause();
> }
>
Hello Eric,
I found the problem, thanks. I'll re-send after testing.
In the meantime, I'd like to ask you whether it makes sense to
add the /proc/net entry, to switch between "old way" and "new way".
The switch would allow quick compare/test between new way and
old way not only by line count, but by full contents, without reboot.
Yakov
^ permalink raw reply
* Re: [PATCH 2/4 v3] bonding: make sure tx and rx hash tables stay in sync when using alb mode
From: Jay Vosburgh @ 2009-09-28 22:09 UTC (permalink / raw)
To: Andy Gospodarek; +Cc: netdev, bonding-devel
In-Reply-To: <20090928220020.GC4436@gospo.rdu.redhat.com>
Andy Gospodarek <andy@greyhouse.net> wrote:
>On Fri, Sep 18, 2009 at 11:56:45AM -0400, Andy Gospodarek wrote:
>> On Fri, Sep 18, 2009 at 11:36:22AM -0400, Andy Gospodarek wrote:
>> > On Wed, Sep 16, 2009 at 04:36:09PM -0700, Jay Vosburgh wrote:
>> > > Andy Gospodarek <andy@greyhouse.net> wrote:
>> > >
>> > > >
>> > > >Subject: [PATCH] bonding: make sure tx and rx hash tables stay in sync when using alb mode
>> > >
>> > > When testing this, I'm getting a lockdep warning. It appears to
>> > > be unhappy that tlb_choose_channel acquires the tx / rx hash table locks
>> > > in the order tx then rx, but rlb_choose_channel -> alb_get_best_slave
>> > > acquires the locks in the other order. I applied all four patches, but
>> > > it looks like the change that trips lockdep is in this patch (#2).
>> > >
>> > > I haven't gotten an actual deadlock from this, although it seems
>> > > plausible if there are two cpus in bond_alb_xmit at the same time, and
>> > > one of them is sending an ARP.
>> > >
>> > > One fairly straightforward fix would be to combine the rx and tx
>> > > hash table locks into a single lock. I suspect that wouldn't have any
>> > > real performance penalty, since the rx hash table lock is generally not
>> > > acquired very often (unlike the tx lock, which is taken for every packet
>> > > that goes out).
>> > >
>> > > Also, FYI, two of the four patches had trailing whitespace. I
>> > > believe it was #2 and #4.
>> > >
>> > > Thoughts?
>> >
>> > Jay,
>> >
>> > This patch should address both the the deadlock and whitespace conerns.
>> > I ran a kernel with LOCKDEP enabled and saw no warnings while passing
>> > traffic on the bond while pulling cables and while removing the module.
>> > Here it is....
>> >
>>
>> Adding the version and signed-off-by lines might be nice, eh?
>>
>> [PATCH v3] bonding: make sure tx and rx hash tables stay in sync when using alb mode
>>
>> I noticed that it was easy for alb (mode 6) bonding to get into a state
>> where the tx hash-table and rx hash-table are out of sync (there is
>> really nothing to keep them synchronized), and we will transmit traffic
>> destined for a host on one slave and send ARP frames to the same slave
>> from another interface using a different source MAC.
>>
>> There is no compelling reason to do this, so this patch makes sure the
>> rx hash-table changes whenever the tx hash-table is updated based on
>> device load. This patch also drops the code that does rlb re-balancing
>> since the balancing will not be controlled by the tx hash-table based on
In addition to my response in the other thread, I changed the
"not" above to "now," which I suspect is what you meant.
>> transmit load. In order to address an issue found with the initial
>> patch, I have also combined the rx and tx hash table lock into a single
>> lock. This will facilitate moving these into a single table at some
>> point.
>>
>> Signed-off-by: Andy Gospodarek <andy@greyhouse.net>
>>
>> ---
>> drivers/net/bonding/bond_alb.c | 203 +++++++++++++++-------------------------
>> drivers/net/bonding/bond_alb.h | 3 +-
>> 2 files changed, 75 insertions(+), 131 deletions(-)
>>
>> diff --git a/drivers/net/bonding/bond_alb.c b/drivers/net/bonding/bond_alb.c
>> index bcf25c6..04b7055 100644
>> --- a/drivers/net/bonding/bond_alb.c
>> +++ b/drivers/net/bonding/bond_alb.c
>> @@ -111,6 +111,7 @@ static inline struct arp_pkt *arp_pkt(const struct sk_buff *skb)
>>
>> /* Forward declaration */
>> static void alb_send_learning_packets(struct slave *slave, u8 mac_addr[]);
>> +static struct slave *alb_get_best_slave(struct bonding *bond, u32 hash_index);
>>
>> static inline u8 _simple_hash(const u8 *hash_start, int hash_size)
>> {
>> @@ -124,18 +125,18 @@ static inline u8 _simple_hash(const u8 *hash_start, int hash_size)
>> return hash;
>> }
>>
>> -/*********************** tlb specific functions ***************************/
>> -
>> -static inline void _lock_tx_hashtbl(struct bonding *bond)
>> +/********************* hash table lock functions *************************/
>> +static inline void _lock_hashtbl(struct bonding *bond)
>> {
>> - spin_lock_bh(&(BOND_ALB_INFO(bond).tx_hashtbl_lock));
>> + spin_lock_bh(&(BOND_ALB_INFO(bond).hashtbl_lock));
>> }
>>
>> -static inline void _unlock_tx_hashtbl(struct bonding *bond)
>> +static inline void _unlock_hashtbl(struct bonding *bond)
>> {
>> - spin_unlock_bh(&(BOND_ALB_INFO(bond).tx_hashtbl_lock));
>> + spin_unlock_bh(&(BOND_ALB_INFO(bond).hashtbl_lock));
>> }
>>
>> +/*********************** tlb specific functions ***************************/
>> /* Caller must hold tx_hashtbl lock */
>> static inline void tlb_init_table_entry(struct tlb_client_info *entry, int save_load)
>> {
>> @@ -163,7 +164,7 @@ static void tlb_clear_slave(struct bonding *bond, struct slave *slave, int save_
>> struct tlb_client_info *tx_hash_table;
>> u32 index;
>>
>> - _lock_tx_hashtbl(bond);
>> + _lock_hashtbl(bond);
>>
>> /* clear slave from tx_hashtbl */
>> tx_hash_table = BOND_ALB_INFO(bond).tx_hashtbl;
>> @@ -180,7 +181,7 @@ static void tlb_clear_slave(struct bonding *bond, struct slave *slave, int save_
>>
>> tlb_init_slave(slave);
>>
>> - _unlock_tx_hashtbl(bond);
>> + _unlock_hashtbl(bond);
>> }
>>
>> /* Must be called before starting the monitor timer */
>> @@ -191,7 +192,7 @@ static int tlb_initialize(struct bonding *bond)
>> struct tlb_client_info *new_hashtbl;
>> int i;
>>
>> - spin_lock_init(&(bond_info->tx_hashtbl_lock));
>> + spin_lock_init(&(bond_info->hashtbl_lock));
>>
>> new_hashtbl = kzalloc(size, GFP_KERNEL);
>> if (!new_hashtbl) {
>> @@ -200,7 +201,7 @@ static int tlb_initialize(struct bonding *bond)
>> bond->dev->name);
>> return -1;
>> }
>> - _lock_tx_hashtbl(bond);
>> + _lock_hashtbl(bond);
>>
>> bond_info->tx_hashtbl = new_hashtbl;
>>
>> @@ -208,7 +209,7 @@ static int tlb_initialize(struct bonding *bond)
>> tlb_init_table_entry(&bond_info->tx_hashtbl[i], 1);
>> }
>>
>> - _unlock_tx_hashtbl(bond);
>> + _unlock_hashtbl(bond);
>>
>> return 0;
>> }
>> @@ -218,12 +219,12 @@ static void tlb_deinitialize(struct bonding *bond)
>> {
>> struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
>>
>> - _lock_tx_hashtbl(bond);
>> + _lock_hashtbl(bond);
>>
>> kfree(bond_info->tx_hashtbl);
>> bond_info->tx_hashtbl = NULL;
>>
>> - _unlock_tx_hashtbl(bond);
>> + _unlock_hashtbl(bond);
>> }
>>
>> /* Caller must hold bond lock for read */
>> @@ -264,24 +265,6 @@ static struct slave *tlb_get_least_loaded_slave(struct bonding *bond)
>> return least_loaded;
>> }
>>
>> -/* Caller must hold bond lock for read and hashtbl lock */
>> -static struct slave *tlb_get_best_slave(struct bonding *bond, u32 hash_index)
>> -{
>> - struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
>> - struct tlb_client_info *tx_hash_table = bond_info->tx_hashtbl;
>> - struct slave *last_slave = tx_hash_table[hash_index].last_slave;
>> - struct slave *next_slave = NULL;
>> -
>> - if (last_slave && SLAVE_IS_OK(last_slave)) {
>> - /* Use the last slave listed in the tx hashtbl if:
>> - the last slave currently is essentially unloaded. */
>> - if (SLAVE_TLB_INFO(last_slave).load < 10)
>> - next_slave = last_slave;
>> - }
>> -
>> - return next_slave ? next_slave : tlb_get_least_loaded_slave(bond);
>> -}
>> -
>> /* Caller must hold bond lock for read */
>> static struct slave *tlb_choose_channel(struct bonding *bond, u32 hash_index, u32 skb_len)
>> {
>> @@ -289,13 +272,12 @@ static struct slave *tlb_choose_channel(struct bonding *bond, u32 hash_index, u3
>> struct tlb_client_info *hash_table;
>> struct slave *assigned_slave;
>>
>> - _lock_tx_hashtbl(bond);
>> + _lock_hashtbl(bond);
>>
>> hash_table = bond_info->tx_hashtbl;
>> assigned_slave = hash_table[hash_index].tx_slave;
>> if (!assigned_slave) {
>> - assigned_slave = tlb_get_best_slave(bond, hash_index);
>> -
>> + assigned_slave = alb_get_best_slave(bond, hash_index);
>> if (assigned_slave) {
>> struct tlb_slave_info *slave_info =
>> &(SLAVE_TLB_INFO(assigned_slave));
>> @@ -319,20 +301,52 @@ static struct slave *tlb_choose_channel(struct bonding *bond, u32 hash_index, u3
>> hash_table[hash_index].tx_bytes += skb_len;
>> }
>>
>> - _unlock_tx_hashtbl(bond);
>> + _unlock_hashtbl(bond);
>>
>> return assigned_slave;
>> }
>>
>> /*********************** rlb specific functions ***************************/
>> -static inline void _lock_rx_hashtbl(struct bonding *bond)
>> +
>> +/* Caller must hold bond lock for read and hashtbl lock */
>> +static struct slave *rlb_update_rx_table(struct bonding *bond, struct slave *next_slave, u32 hash_index)
>> {
>> - spin_lock_bh(&(BOND_ALB_INFO(bond).rx_hashtbl_lock));
>> + struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
>> +
>> + /* check rlb table and correct it if wrong */
>> + if (bond_info->rlb_enabled) {
>> + struct rlb_client_info *rx_client_info = &(bond_info->rx_hashtbl[hash_index]);
>> +
>> + /* if the new slave computed by tlb checks doesn't match rlb, stop rlb from using it */
>> + if (next_slave && (next_slave != rx_client_info->slave))
>> + rx_client_info->slave = next_slave;
>> + }
>> + return next_slave;
>> }
>>
>> -static inline void _unlock_rx_hashtbl(struct bonding *bond)
>> +/* Caller must hold bond lock for read and hashtbl lock */
>> +static struct slave *alb_get_best_slave(struct bonding *bond, u32 hash_index)
>> {
>> - spin_unlock_bh(&(BOND_ALB_INFO(bond).rx_hashtbl_lock));
>> + struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
>> + struct tlb_client_info *tx_hash_table = bond_info->tx_hashtbl;
>> + struct slave *last_slave = tx_hash_table[hash_index].last_slave;
>> + struct slave *next_slave = NULL;
>> +
>> + /* presume the next slave will be the least loaded one */
>> + next_slave = tlb_get_least_loaded_slave(bond);
>> +
>> + if (last_slave && SLAVE_IS_OK(last_slave)) {
>> + /* Use the last slave listed in the tx hashtbl if:
>> + the last slave currently is essentially unloaded. */
>> + if (SLAVE_TLB_INFO(last_slave).load < 10)
>> + next_slave = last_slave;
>> + }
>> +
>> + /* update the rlb hashtbl if there was a previous entry */
>> + if (bond_info->rlb_enabled)
>> + rlb_update_rx_table(bond, next_slave, hash_index);
>> +
>> + return next_slave;
>> }
>>
>> /* when an ARP REPLY is received from a client update its info
>> @@ -344,7 +358,7 @@ static void rlb_update_entry_from_arp(struct bonding *bond, struct arp_pkt *arp)
>> struct rlb_client_info *client_info;
>> u32 hash_index;
>>
>> - _lock_rx_hashtbl(bond);
>> + _lock_hashtbl(bond);
>>
>> hash_index = _simple_hash((u8*)&(arp->ip_src), sizeof(arp->ip_src));
>> client_info = &(bond_info->rx_hashtbl[hash_index]);
>> @@ -358,7 +372,7 @@ static void rlb_update_entry_from_arp(struct bonding *bond, struct arp_pkt *arp)
>> bond_info->rx_ntt = 1;
>> }
>>
>> - _unlock_rx_hashtbl(bond);
>> + _unlock_hashtbl(bond);
>> }
>>
>> static int rlb_arp_recv(struct sk_buff *skb, struct net_device *bond_dev, struct packet_type *ptype, struct net_device *orig_dev)
>> @@ -402,38 +416,6 @@ out:
>> return res;
>> }
>>
>> -/* Caller must hold bond lock for read */
>> -static struct slave *rlb_next_rx_slave(struct bonding *bond)
>> -{
>> - struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
>> - struct slave *rx_slave, *slave, *start_at;
>> - int i = 0;
>> -
>> - if (bond_info->next_rx_slave) {
>> - start_at = bond_info->next_rx_slave;
>> - } else {
>> - start_at = bond->first_slave;
>> - }
>> -
>> - rx_slave = NULL;
>> -
>> - bond_for_each_slave_from(bond, slave, i, start_at) {
>> - if (SLAVE_IS_OK(slave)) {
>> - if (!rx_slave) {
>> - rx_slave = slave;
>> - } else if (slave->speed > rx_slave->speed) {
>> - rx_slave = slave;
>> - }
>> - }
>> - }
>> -
>> - if (rx_slave) {
>> - bond_info->next_rx_slave = rx_slave->next;
>> - }
>> -
>> - return rx_slave;
>> -}
>> -
>> /* teach the switch the mac of a disabled slave
>> * on the primary for fault tolerance
>> *
>> @@ -468,14 +450,14 @@ static void rlb_clear_slave(struct bonding *bond, struct slave *slave)
>> u32 index, next_index;
>>
>> /* clear slave from rx_hashtbl */
>> - _lock_rx_hashtbl(bond);
>> + _lock_hashtbl(bond);
>>
>> rx_hash_table = bond_info->rx_hashtbl;
>> index = bond_info->rx_hashtbl_head;
>> for (; index != RLB_NULL_INDEX; index = next_index) {
>> next_index = rx_hash_table[index].next;
>> if (rx_hash_table[index].slave == slave) {
>> - struct slave *assigned_slave = rlb_next_rx_slave(bond);
>> + struct slave *assigned_slave = alb_get_best_slave(bond, index);
>>
>> if (assigned_slave) {
>> rx_hash_table[index].slave = assigned_slave;
>> @@ -499,7 +481,7 @@ static void rlb_clear_slave(struct bonding *bond, struct slave *slave)
>> }
>> }
>>
>> - _unlock_rx_hashtbl(bond);
>> + _unlock_hashtbl(bond);
>>
>> write_lock_bh(&bond->curr_slave_lock);
>>
>> @@ -558,7 +540,7 @@ static void rlb_update_rx_clients(struct bonding *bond)
>> struct rlb_client_info *client_info;
>> u32 hash_index;
>>
>> - _lock_rx_hashtbl(bond);
>> + _lock_hashtbl(bond);
>>
>> hash_index = bond_info->rx_hashtbl_head;
>> for (; hash_index != RLB_NULL_INDEX; hash_index = client_info->next) {
>> @@ -576,7 +558,7 @@ static void rlb_update_rx_clients(struct bonding *bond)
>> */
>> bond_info->rlb_update_delay_counter = RLB_UPDATE_DELAY;
>>
>> - _unlock_rx_hashtbl(bond);
>> + _unlock_hashtbl(bond);
>> }
>>
>> /* The slave was assigned a new mac address - update the clients */
>> @@ -587,7 +569,7 @@ static void rlb_req_update_slave_clients(struct bonding *bond, struct slave *sla
>> int ntt = 0;
>> u32 hash_index;
>>
>> - _lock_rx_hashtbl(bond);
>> + _lock_hashtbl(bond);
>>
>> hash_index = bond_info->rx_hashtbl_head;
>> for (; hash_index != RLB_NULL_INDEX; hash_index = client_info->next) {
>> @@ -607,7 +589,7 @@ static void rlb_req_update_slave_clients(struct bonding *bond, struct slave *sla
>> bond_info->rlb_update_retry_counter = RLB_UPDATE_RETRY;
>> }
>>
>> - _unlock_rx_hashtbl(bond);
>> + _unlock_hashtbl(bond);
>> }
>>
>> /* mark all clients using src_ip to be updated */
>> @@ -617,7 +599,7 @@ static void rlb_req_update_subnet_clients(struct bonding *bond, __be32 src_ip)
>> struct rlb_client_info *client_info;
>> u32 hash_index;
>>
>> - _lock_rx_hashtbl(bond);
>> + _lock_hashtbl(bond);
>>
>> hash_index = bond_info->rx_hashtbl_head;
>> for (; hash_index != RLB_NULL_INDEX; hash_index = client_info->next) {
>> @@ -643,7 +625,7 @@ static void rlb_req_update_subnet_clients(struct bonding *bond, __be32 src_ip)
>> }
>> }
>>
>> - _unlock_rx_hashtbl(bond);
>> + _unlock_hashtbl(bond);
>> }
>>
>> /* Caller must hold both bond and ptr locks for read */
>> @@ -655,7 +637,7 @@ static struct slave *rlb_choose_channel(struct sk_buff *skb, struct bonding *bon
>> struct rlb_client_info *client_info;
>> u32 hash_index = 0;
>>
>> - _lock_rx_hashtbl(bond);
>> + _lock_hashtbl(bond);
>>
>> hash_index = _simple_hash((u8 *)&arp->ip_dst, sizeof(arp->ip_src));
>> client_info = &(bond_info->rx_hashtbl[hash_index]);
>> @@ -671,7 +653,7 @@ static struct slave *rlb_choose_channel(struct sk_buff *skb, struct bonding *bon
>>
>> assigned_slave = client_info->slave;
>> if (assigned_slave) {
>> - _unlock_rx_hashtbl(bond);
>> + _unlock_hashtbl(bond);
>> return assigned_slave;
>> }
>> } else {
>> @@ -687,7 +669,7 @@ static struct slave *rlb_choose_channel(struct sk_buff *skb, struct bonding *bon
>> }
>> }
>> /* assign a new slave */
>> - assigned_slave = rlb_next_rx_slave(bond);
>> + assigned_slave = alb_get_best_slave(bond, hash_index);
>>
>> if (assigned_slave) {
>> client_info->ip_src = arp->ip_src;
>> @@ -723,7 +705,7 @@ static struct slave *rlb_choose_channel(struct sk_buff *skb, struct bonding *bon
>> }
>> }
>>
>> - _unlock_rx_hashtbl(bond);
>> + _unlock_hashtbl(bond);
>>
>> return assigned_slave;
>> }
>> @@ -771,36 +753,6 @@ static struct slave *rlb_arp_xmit(struct sk_buff *skb, struct bonding *bond)
>> return tx_slave;
>> }
>>
>> -/* Caller must hold bond lock for read */
>> -static void rlb_rebalance(struct bonding *bond)
>> -{
>> - struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
>> - struct slave *assigned_slave;
>> - struct rlb_client_info *client_info;
>> - int ntt;
>> - u32 hash_index;
>> -
>> - _lock_rx_hashtbl(bond);
>> -
>> - ntt = 0;
>> - hash_index = bond_info->rx_hashtbl_head;
>> - for (; hash_index != RLB_NULL_INDEX; hash_index = client_info->next) {
>> - client_info = &(bond_info->rx_hashtbl[hash_index]);
>> - assigned_slave = rlb_next_rx_slave(bond);
>> - if (assigned_slave && (client_info->slave != assigned_slave)) {
>> - client_info->slave = assigned_slave;
>> - client_info->ntt = 1;
>> - ntt = 1;
>> - }
>> - }
>> -
>> - /* update the team's flag only after the whole iteration */
>> - if (ntt) {
>> - bond_info->rx_ntt = 1;
>> - }
>> - _unlock_rx_hashtbl(bond);
>> -}
>> -
>> /* Caller must hold rx_hashtbl lock */
>> static void rlb_init_table_entry(struct rlb_client_info *entry)
>> {
>> @@ -817,8 +769,6 @@ static int rlb_initialize(struct bonding *bond)
>> int size = RLB_HASH_TABLE_SIZE * sizeof(struct rlb_client_info);
>> int i;
>>
>> - spin_lock_init(&(bond_info->rx_hashtbl_lock));
>> -
>> new_hashtbl = kmalloc(size, GFP_KERNEL);
>> if (!new_hashtbl) {
>> printk(KERN_ERR DRV_NAME
>> @@ -826,7 +776,7 @@ static int rlb_initialize(struct bonding *bond)
>> bond->dev->name);
>> return -1;
>> }
>> - _lock_rx_hashtbl(bond);
>> + _lock_hashtbl(bond);
>>
>> bond_info->rx_hashtbl = new_hashtbl;
>>
>> @@ -836,7 +786,7 @@ static int rlb_initialize(struct bonding *bond)
>> rlb_init_table_entry(bond_info->rx_hashtbl + i);
>> }
>>
>> - _unlock_rx_hashtbl(bond);
>> + _unlock_hashtbl(bond);
>>
>> /*initialize packet type*/
>> pk_type->type = cpu_to_be16(ETH_P_ARP);
>> @@ -855,13 +805,13 @@ static void rlb_deinitialize(struct bonding *bond)
>>
>> dev_remove_pack(&(bond_info->rlb_pkt_type));
>>
>> - _lock_rx_hashtbl(bond);
>> + _lock_hashtbl(bond);
>>
>> kfree(bond_info->rx_hashtbl);
>> bond_info->rx_hashtbl = NULL;
>> bond_info->rx_hashtbl_head = RLB_NULL_INDEX;
>>
>> - _unlock_rx_hashtbl(bond);
>> + _unlock_hashtbl(bond);
>> }
>>
>> static void rlb_clear_vlan(struct bonding *bond, unsigned short vlan_id)
>> @@ -869,7 +819,7 @@ static void rlb_clear_vlan(struct bonding *bond, unsigned short vlan_id)
>> struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
>> u32 curr_index;
>>
>> - _lock_rx_hashtbl(bond);
>> + _lock_hashtbl(bond);
>>
>> curr_index = bond_info->rx_hashtbl_head;
>> while (curr_index != RLB_NULL_INDEX) {
>> @@ -894,7 +844,7 @@ static void rlb_clear_vlan(struct bonding *bond, unsigned short vlan_id)
>> curr_index = next_index;
>> }
>>
>> - _unlock_rx_hashtbl(bond);
>> + _unlock_hashtbl(bond);
>> }
>>
>> /*********************** tlb/rlb shared functions *********************/
>> @@ -1521,11 +1471,6 @@ void bond_alb_monitor(struct work_struct *work)
>> read_lock(&bond->lock);
>> }
>>
>> - if (bond_info->rlb_rebalance) {
>> - bond_info->rlb_rebalance = 0;
>> - rlb_rebalance(bond);
>> - }
>> -
>> /* check if clients need updating */
>> if (bond_info->rx_ntt) {
>> if (bond_info->rlb_update_delay_counter) {
>> diff --git a/drivers/net/bonding/bond_alb.h b/drivers/net/bonding/bond_alb.h
>> index b65fd29..09d755a 100644
>> --- a/drivers/net/bonding/bond_alb.h
>> +++ b/drivers/net/bonding/bond_alb.h
>> @@ -90,7 +90,7 @@ struct tlb_slave_info {
>> struct alb_bond_info {
>> struct timer_list alb_timer;
>> struct tlb_client_info *tx_hashtbl; /* Dynamically allocated */
>> - spinlock_t tx_hashtbl_lock;
>> + spinlock_t hashtbl_lock; /* lock for both tables */
>> u32 unbalanced_load;
>> int tx_rebalance_counter;
>> int lp_counter;
>> @@ -98,7 +98,6 @@ struct alb_bond_info {
>> int rlb_enabled;
>> struct packet_type rlb_pkt_type;
>> struct rlb_client_info *rx_hashtbl; /* Receive hash table */
>> - spinlock_t rx_hashtbl_lock;
>> u32 rx_hashtbl_head;
>> u8 rx_ntt; /* flag - need to transmit
>> * to all rx clients
>
>Any thoughts on this, Jay?
-J
---
-Jay Vosburgh, IBM Linux Technology Center, fubar@us.ibm.com
^ permalink raw reply
* Re: [PATCH 4/4 v2] bonding: add sysfs files to display tlb and alb hash table contents
From: Jay Vosburgh @ 2009-09-28 22:06 UTC (permalink / raw)
To: Andy Gospodarek; +Cc: netdev, bonding-devel
In-Reply-To: <20090928220117.GD4436@gospo.rdu.redhat.com>
Andy Gospodarek <andy@greyhouse.net> wrote:
>On Fri, Sep 18, 2009 at 11:53:11AM -0400, Andy Gospodarek wrote:
>> On Fri, Sep 11, 2009 at 05:13:17PM -0400, Andy Gospodarek wrote:
>> >
>> > bonding: add sysfs files to display tlb and alb hash table contents
>> >
>> > While debugging some problems with alb (mode 6) bonding I realized that
>> > being able to output the contents of both hash tables would be helpful.
>> > This is what the output looks like for the two files:
>> >
>> > device load
>> > eth1 491
>> > eth2 491
>> > hash device last device tx bytes load next previous
>> > 2 eth1 eth1 2254 491 0 0
>> > 3 eth2 eth2 2744 491 0 0
>> > 6 eth2 0 488 0 0
>> > 8 eth2 0 461698 0 0
>> > 1b eth2 0 249 0 0
>> > eb eth2 0 21 0 0
>> > ff eth2 0 22 0 0
>> >
>> > hash ip_src ip_dst mac_dst slave assign ntt
>> > 2 10.0.3.2 10.0.3.11 00:e0:81:71:ee:a9 eth1 1 0
>> > 3 10.0.3.2 10.0.3.10 00:e0:81:71:ee:a9 eth2 1 0
>> > 8 10.0.3.2 10.0.3.1 00:e0:81:71:ee:a9 eth2 1 0
>> >
>> > These were a great help debugging the fixes I have just posted and they
>> > might be helpful for others, so I decided to include them in my
>> > patchset.
>> >
>> > Signed-off-by: Andy Gospodarek <andy@greyhouse.net>
>> >
>>
>> Needed to repost since patch 2/4 changed and first patch had whitespace
>> issues:
>>
>> [PATCH v2] bonding: add sysfs files to display tlb and alb hash table contents
>>
>> While debugging some problems with alb (mode 6) bonding I realized that
>> being able to output the contents of both hash tables would be helpful.
>> This is what the output looks like for the two files:
>>
>> device load
>> eth1 491
>> eth2 491
>> hash device last device tx bytes load next previous
>> 2 eth1 eth1 2254 491 0 0
>> 3 eth2 eth2 2744 491 0 0
>> 6 eth2 0 488 0 0
>> 8 eth2 0 461698 0 0
>> 1b eth2 0 249 0 0
>> eb eth2 0 21 0 0
>> ff eth2 0 22 0 0
>>
>> hash ip_src ip_dst mac_dst slave assign ntt
>> 2 10.0.3.2 10.0.3.11 00:e0:81:71:ee:a9 eth1 1 0
>> 3 10.0.3.2 10.0.3.10 00:e0:81:71:ee:a9 eth2 1 0
>> 8 10.0.3.2 10.0.3.1 00:e0:81:71:ee:a9 eth2 1 0
>>
>> These were a great help debugging the fixes I have just posted and they
>> might be helpful for others, so I decided to include them in my post.
>>
>> Signed-off-by: Andy Gospodarek <andy@greyhouse.net>
>>
>> ---
>> drivers/net/bonding/bond_alb.c | 61 ++++++++++++++++++++++++++++++++++++++
>> drivers/net/bonding/bond_alb.h | 2 +
>> drivers/net/bonding/bond_sysfs.c | 40 +++++++++++++++++++++++++
>> 3 files changed, 103 insertions(+), 0 deletions(-)
>>
>> diff --git a/drivers/net/bonding/bond_alb.c b/drivers/net/bonding/bond_alb.c
>> index 5d51489..adc5acd 100644
>> --- a/drivers/net/bonding/bond_alb.c
>> +++ b/drivers/net/bonding/bond_alb.c
>> @@ -750,6 +750,67 @@ static struct slave *rlb_arp_xmit(struct sk_buff *skb, struct bonding *bond)
>> return tx_slave;
>> }
>>
>> +int rlb_print_rx_hashtbl(struct bonding *bond, char *buf)
>> +{
>> + struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
>> + struct rlb_client_info *client_info;
>> + u32 hash_index;
>> + u32 count = 0;
>> +
>> + _lock_hashtbl(bond);
>> +
>> + count = sprintf(buf, "hash ip_src ip_dst mac_dst slave assign ntt\n");
>> + hash_index = bond_info->rx_hashtbl_head;
>> + for (; hash_index != RLB_NULL_INDEX; hash_index = client_info->next) {
>> + client_info = &(bond_info->rx_hashtbl[hash_index]);
>> + count += sprintf(buf + count,"%-4x %-15pi4 %-15pi4 %pM %-5s %-6d %d\n",
>> + hash_index,
>> + &client_info->ip_src,
>> + &client_info->ip_dst,
>> + client_info->mac_dst,
>> + client_info->slave->dev->name,
>> + client_info->assigned,
>> + client_info->ntt);
>> + }
>> +
>> + _unlock_hashtbl(bond);
>> + return count;
>> +}
>> +
>> +int tlb_print_tx_hashtbl(struct bonding *bond, char *buf)
>> +{
>> + struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
>> + u32 hash_index;
>> + u32 count = 0;
>> + struct slave *slave;
>> + int i;
>> +
>> + _lock_hashtbl(bond);
>> +
>> + count += sprintf(buf, "device load\n");
>> + bond_for_each_slave(bond, slave, i) {
>> + struct tlb_slave_info *slave_info = &(SLAVE_TLB_INFO(slave));
>> + count += sprintf(buf + count,"%-7s %d\n",slave->dev->name,slave_info->load);
>> + }
>> + count += sprintf(buf + count, "hash device last device tx bytes load next previous\n");
>> + for (hash_index = 0; hash_index < TLB_HASH_TABLE_SIZE; hash_index++) {
>> + struct tlb_client_info *client_info = &(bond_info->tx_hashtbl[hash_index]);
>> + if (client_info->tx_slave || client_info->last_slave) {
>> + count += sprintf(buf + count,"%-4x %-8s %-13s %-14d %-11d %-4x %d\n",
>> + hash_index,
>> + (client_info->tx_slave) ? client_info->tx_slave->dev->name : "",
>> + (client_info->last_slave) ? client_info->last_slave->dev->name : "",
>> + client_info->tx_bytes,
>> + client_info->load_history,
>> + (client_info->next != TLB_NULL_INDEX) ? client_info->next : 0,
>> + (client_info->prev != TLB_NULL_INDEX) ? client_info->prev : 0);
>> + }
>> + }
>> +
>> + _unlock_hashtbl(bond);
>> + return count;
>> +}
>> +
>> /* Caller must hold rx_hashtbl lock */
>> static void rlb_init_table_entry(struct rlb_client_info *entry)
>> {
>> diff --git a/drivers/net/bonding/bond_alb.h b/drivers/net/bonding/bond_alb.h
>> index 09d755a..57e761b 100644
>> --- a/drivers/net/bonding/bond_alb.h
>> +++ b/drivers/net/bonding/bond_alb.h
>> @@ -131,5 +131,7 @@ int bond_alb_xmit(struct sk_buff *skb, struct net_device *bond_dev);
>> void bond_alb_monitor(struct work_struct *);
>> int bond_alb_set_mac_address(struct net_device *bond_dev, void *addr);
>> void bond_alb_clear_vlan(struct bonding *bond, unsigned short vlan_id);
>> +int rlb_print_rx_hashtbl(struct bonding *bond, char *buf);
>> +int tlb_print_tx_hashtbl(struct bonding *bond, char *buf);
>> #endif /* __BOND_ALB_H__ */
>>
>> diff --git a/drivers/net/bonding/bond_sysfs.c b/drivers/net/bonding/bond_sysfs.c
>> index 55bf34f..1123e1f 100644
>> --- a/drivers/net/bonding/bond_sysfs.c
>> +++ b/drivers/net/bonding/bond_sysfs.c
>> @@ -1480,6 +1480,44 @@ static ssize_t bonding_show_ad_partner_mac(struct device *d,
>> static DEVICE_ATTR(ad_partner_mac, S_IRUGO, bonding_show_ad_partner_mac, NULL);
>>
>>
>> +/*
>> + * Show current tlb/alb tx hash table.
>> + */
>> +static ssize_t bonding_show_tlb_tx_hash(struct device *d,
>> + struct device_attribute *attr,
>> + char *buf)
>> +{
>> + int count = 0;
>> + struct bonding *bond = to_bond(d);
>> +
>> + if (bond->params.mode == BOND_MODE_ALB ||
>> + bond->params.mode == BOND_MODE_TLB) {
>> + count = tlb_print_tx_hashtbl(bond, buf);
>> + }
>> +
>> + return count;
>> +}
>> +static DEVICE_ATTR(tlb_tx_hash, S_IRUGO, bonding_show_tlb_tx_hash, NULL);
>> +
>> +
>> +/*
>> + * Show current alb rx hash table.
>> + */
>> +static ssize_t bonding_show_alb_rx_hash(struct device *d,
>> + struct device_attribute *attr,
>> + char *buf)
>> +{
>> + int count = 0;
>> + struct bonding *bond = to_bond(d);
>> +
>> + if (bond->params.mode == BOND_MODE_ALB) {
>> + count = rlb_print_rx_hashtbl(bond, buf);
>> + }
>> +
>> + return count;
>> +}
>> +static DEVICE_ATTR(alb_rx_hash, S_IRUGO, bonding_show_alb_rx_hash, NULL);
>> +
>>
>> static struct attribute *per_bond_attrs[] = {
>> &dev_attr_slaves.attr,
>> @@ -1505,6 +1543,8 @@ static struct attribute *per_bond_attrs[] = {
>> &dev_attr_ad_actor_key.attr,
>> &dev_attr_ad_partner_key.attr,
>> &dev_attr_ad_partner_mac.attr,
>> + &dev_attr_alb_rx_hash.attr,
>> + &dev_attr_tlb_tx_hash.attr,
>> NULL,
>> };
>>
>
>Any thoughts on this one as well, Jay?
>
I've been testing with them last Friday and today. Seem to work
ok so far. I've made a couple of minor changes (removed some remaining
vestiges of the rlb rebalance code, changed the mode of the sysfs hash
table files to 0400 after some testing with ping and concurrent "while 1
cat > /dev/null" loops).
Unless something awful comes up, I'll post them with my changes
later today or tomorrow.
-J
---
-Jay Vosburgh, IBM Linux Technology Center, fubar@us.ibm.com
^ permalink raw reply
* Re: [PATCH 4/4 v2] bonding: add sysfs files to display tlb and alb hash table contents
From: Andy Gospodarek @ 2009-09-28 22:01 UTC (permalink / raw)
To: netdev, fubar, bonding-devel
In-Reply-To: <20090918155311.GA4436@gospo.rdu.redhat.com>
On Fri, Sep 18, 2009 at 11:53:11AM -0400, Andy Gospodarek wrote:
> On Fri, Sep 11, 2009 at 05:13:17PM -0400, Andy Gospodarek wrote:
> >
> > bonding: add sysfs files to display tlb and alb hash table contents
> >
> > While debugging some problems with alb (mode 6) bonding I realized that
> > being able to output the contents of both hash tables would be helpful.
> > This is what the output looks like for the two files:
> >
> > device load
> > eth1 491
> > eth2 491
> > hash device last device tx bytes load next previous
> > 2 eth1 eth1 2254 491 0 0
> > 3 eth2 eth2 2744 491 0 0
> > 6 eth2 0 488 0 0
> > 8 eth2 0 461698 0 0
> > 1b eth2 0 249 0 0
> > eb eth2 0 21 0 0
> > ff eth2 0 22 0 0
> >
> > hash ip_src ip_dst mac_dst slave assign ntt
> > 2 10.0.3.2 10.0.3.11 00:e0:81:71:ee:a9 eth1 1 0
> > 3 10.0.3.2 10.0.3.10 00:e0:81:71:ee:a9 eth2 1 0
> > 8 10.0.3.2 10.0.3.1 00:e0:81:71:ee:a9 eth2 1 0
> >
> > These were a great help debugging the fixes I have just posted and they
> > might be helpful for others, so I decided to include them in my
> > patchset.
> >
> > Signed-off-by: Andy Gospodarek <andy@greyhouse.net>
> >
>
> Needed to repost since patch 2/4 changed and first patch had whitespace
> issues:
>
> [PATCH v2] bonding: add sysfs files to display tlb and alb hash table contents
>
> While debugging some problems with alb (mode 6) bonding I realized that
> being able to output the contents of both hash tables would be helpful.
> This is what the output looks like for the two files:
>
> device load
> eth1 491
> eth2 491
> hash device last device tx bytes load next previous
> 2 eth1 eth1 2254 491 0 0
> 3 eth2 eth2 2744 491 0 0
> 6 eth2 0 488 0 0
> 8 eth2 0 461698 0 0
> 1b eth2 0 249 0 0
> eb eth2 0 21 0 0
> ff eth2 0 22 0 0
>
> hash ip_src ip_dst mac_dst slave assign ntt
> 2 10.0.3.2 10.0.3.11 00:e0:81:71:ee:a9 eth1 1 0
> 3 10.0.3.2 10.0.3.10 00:e0:81:71:ee:a9 eth2 1 0
> 8 10.0.3.2 10.0.3.1 00:e0:81:71:ee:a9 eth2 1 0
>
> These were a great help debugging the fixes I have just posted and they
> might be helpful for others, so I decided to include them in my post.
>
> Signed-off-by: Andy Gospodarek <andy@greyhouse.net>
>
> ---
> drivers/net/bonding/bond_alb.c | 61 ++++++++++++++++++++++++++++++++++++++
> drivers/net/bonding/bond_alb.h | 2 +
> drivers/net/bonding/bond_sysfs.c | 40 +++++++++++++++++++++++++
> 3 files changed, 103 insertions(+), 0 deletions(-)
>
> diff --git a/drivers/net/bonding/bond_alb.c b/drivers/net/bonding/bond_alb.c
> index 5d51489..adc5acd 100644
> --- a/drivers/net/bonding/bond_alb.c
> +++ b/drivers/net/bonding/bond_alb.c
> @@ -750,6 +750,67 @@ static struct slave *rlb_arp_xmit(struct sk_buff *skb, struct bonding *bond)
> return tx_slave;
> }
>
> +int rlb_print_rx_hashtbl(struct bonding *bond, char *buf)
> +{
> + struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
> + struct rlb_client_info *client_info;
> + u32 hash_index;
> + u32 count = 0;
> +
> + _lock_hashtbl(bond);
> +
> + count = sprintf(buf, "hash ip_src ip_dst mac_dst slave assign ntt\n");
> + hash_index = bond_info->rx_hashtbl_head;
> + for (; hash_index != RLB_NULL_INDEX; hash_index = client_info->next) {
> + client_info = &(bond_info->rx_hashtbl[hash_index]);
> + count += sprintf(buf + count,"%-4x %-15pi4 %-15pi4 %pM %-5s %-6d %d\n",
> + hash_index,
> + &client_info->ip_src,
> + &client_info->ip_dst,
> + client_info->mac_dst,
> + client_info->slave->dev->name,
> + client_info->assigned,
> + client_info->ntt);
> + }
> +
> + _unlock_hashtbl(bond);
> + return count;
> +}
> +
> +int tlb_print_tx_hashtbl(struct bonding *bond, char *buf)
> +{
> + struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
> + u32 hash_index;
> + u32 count = 0;
> + struct slave *slave;
> + int i;
> +
> + _lock_hashtbl(bond);
> +
> + count += sprintf(buf, "device load\n");
> + bond_for_each_slave(bond, slave, i) {
> + struct tlb_slave_info *slave_info = &(SLAVE_TLB_INFO(slave));
> + count += sprintf(buf + count,"%-7s %d\n",slave->dev->name,slave_info->load);
> + }
> + count += sprintf(buf + count, "hash device last device tx bytes load next previous\n");
> + for (hash_index = 0; hash_index < TLB_HASH_TABLE_SIZE; hash_index++) {
> + struct tlb_client_info *client_info = &(bond_info->tx_hashtbl[hash_index]);
> + if (client_info->tx_slave || client_info->last_slave) {
> + count += sprintf(buf + count,"%-4x %-8s %-13s %-14d %-11d %-4x %d\n",
> + hash_index,
> + (client_info->tx_slave) ? client_info->tx_slave->dev->name : "",
> + (client_info->last_slave) ? client_info->last_slave->dev->name : "",
> + client_info->tx_bytes,
> + client_info->load_history,
> + (client_info->next != TLB_NULL_INDEX) ? client_info->next : 0,
> + (client_info->prev != TLB_NULL_INDEX) ? client_info->prev : 0);
> + }
> + }
> +
> + _unlock_hashtbl(bond);
> + return count;
> +}
> +
> /* Caller must hold rx_hashtbl lock */
> static void rlb_init_table_entry(struct rlb_client_info *entry)
> {
> diff --git a/drivers/net/bonding/bond_alb.h b/drivers/net/bonding/bond_alb.h
> index 09d755a..57e761b 100644
> --- a/drivers/net/bonding/bond_alb.h
> +++ b/drivers/net/bonding/bond_alb.h
> @@ -131,5 +131,7 @@ int bond_alb_xmit(struct sk_buff *skb, struct net_device *bond_dev);
> void bond_alb_monitor(struct work_struct *);
> int bond_alb_set_mac_address(struct net_device *bond_dev, void *addr);
> void bond_alb_clear_vlan(struct bonding *bond, unsigned short vlan_id);
> +int rlb_print_rx_hashtbl(struct bonding *bond, char *buf);
> +int tlb_print_tx_hashtbl(struct bonding *bond, char *buf);
> #endif /* __BOND_ALB_H__ */
>
> diff --git a/drivers/net/bonding/bond_sysfs.c b/drivers/net/bonding/bond_sysfs.c
> index 55bf34f..1123e1f 100644
> --- a/drivers/net/bonding/bond_sysfs.c
> +++ b/drivers/net/bonding/bond_sysfs.c
> @@ -1480,6 +1480,44 @@ static ssize_t bonding_show_ad_partner_mac(struct device *d,
> static DEVICE_ATTR(ad_partner_mac, S_IRUGO, bonding_show_ad_partner_mac, NULL);
>
>
> +/*
> + * Show current tlb/alb tx hash table.
> + */
> +static ssize_t bonding_show_tlb_tx_hash(struct device *d,
> + struct device_attribute *attr,
> + char *buf)
> +{
> + int count = 0;
> + struct bonding *bond = to_bond(d);
> +
> + if (bond->params.mode == BOND_MODE_ALB ||
> + bond->params.mode == BOND_MODE_TLB) {
> + count = tlb_print_tx_hashtbl(bond, buf);
> + }
> +
> + return count;
> +}
> +static DEVICE_ATTR(tlb_tx_hash, S_IRUGO, bonding_show_tlb_tx_hash, NULL);
> +
> +
> +/*
> + * Show current alb rx hash table.
> + */
> +static ssize_t bonding_show_alb_rx_hash(struct device *d,
> + struct device_attribute *attr,
> + char *buf)
> +{
> + int count = 0;
> + struct bonding *bond = to_bond(d);
> +
> + if (bond->params.mode == BOND_MODE_ALB) {
> + count = rlb_print_rx_hashtbl(bond, buf);
> + }
> +
> + return count;
> +}
> +static DEVICE_ATTR(alb_rx_hash, S_IRUGO, bonding_show_alb_rx_hash, NULL);
> +
>
> static struct attribute *per_bond_attrs[] = {
> &dev_attr_slaves.attr,
> @@ -1505,6 +1543,8 @@ static struct attribute *per_bond_attrs[] = {
> &dev_attr_ad_actor_key.attr,
> &dev_attr_ad_partner_key.attr,
> &dev_attr_ad_partner_mac.attr,
> + &dev_attr_alb_rx_hash.attr,
> + &dev_attr_tlb_tx_hash.attr,
> NULL,
> };
>
Any thoughts on this one as well, Jay?
^ permalink raw reply
* Re: [PATCH 2/4 v3] bonding: make sure tx and rx hash tables stay in sync when using alb mode
From: Andy Gospodarek @ 2009-09-28 22:00 UTC (permalink / raw)
To: Jay Vosburgh, netdev, bonding-devel
In-Reply-To: <20090918155645.GB4436@gospo.rdu.redhat.com>
On Fri, Sep 18, 2009 at 11:56:45AM -0400, Andy Gospodarek wrote:
> On Fri, Sep 18, 2009 at 11:36:22AM -0400, Andy Gospodarek wrote:
> > On Wed, Sep 16, 2009 at 04:36:09PM -0700, Jay Vosburgh wrote:
> > > Andy Gospodarek <andy@greyhouse.net> wrote:
> > >
> > > >
> > > >Subject: [PATCH] bonding: make sure tx and rx hash tables stay in sync when using alb mode
> > >
> > > When testing this, I'm getting a lockdep warning. It appears to
> > > be unhappy that tlb_choose_channel acquires the tx / rx hash table locks
> > > in the order tx then rx, but rlb_choose_channel -> alb_get_best_slave
> > > acquires the locks in the other order. I applied all four patches, but
> > > it looks like the change that trips lockdep is in this patch (#2).
> > >
> > > I haven't gotten an actual deadlock from this, although it seems
> > > plausible if there are two cpus in bond_alb_xmit at the same time, and
> > > one of them is sending an ARP.
> > >
> > > One fairly straightforward fix would be to combine the rx and tx
> > > hash table locks into a single lock. I suspect that wouldn't have any
> > > real performance penalty, since the rx hash table lock is generally not
> > > acquired very often (unlike the tx lock, which is taken for every packet
> > > that goes out).
> > >
> > > Also, FYI, two of the four patches had trailing whitespace. I
> > > believe it was #2 and #4.
> > >
> > > Thoughts?
> >
> > Jay,
> >
> > This patch should address both the the deadlock and whitespace conerns.
> > I ran a kernel with LOCKDEP enabled and saw no warnings while passing
> > traffic on the bond while pulling cables and while removing the module.
> > Here it is....
> >
>
> Adding the version and signed-off-by lines might be nice, eh?
>
> [PATCH v3] bonding: make sure tx and rx hash tables stay in sync when using alb mode
>
> I noticed that it was easy for alb (mode 6) bonding to get into a state
> where the tx hash-table and rx hash-table are out of sync (there is
> really nothing to keep them synchronized), and we will transmit traffic
> destined for a host on one slave and send ARP frames to the same slave
> from another interface using a different source MAC.
>
> There is no compelling reason to do this, so this patch makes sure the
> rx hash-table changes whenever the tx hash-table is updated based on
> device load. This patch also drops the code that does rlb re-balancing
> since the balancing will not be controlled by the tx hash-table based on
> transmit load. In order to address an issue found with the initial
> patch, I have also combined the rx and tx hash table lock into a single
> lock. This will facilitate moving these into a single table at some
> point.
>
> Signed-off-by: Andy Gospodarek <andy@greyhouse.net>
>
> ---
> drivers/net/bonding/bond_alb.c | 203 +++++++++++++++-------------------------
> drivers/net/bonding/bond_alb.h | 3 +-
> 2 files changed, 75 insertions(+), 131 deletions(-)
>
> diff --git a/drivers/net/bonding/bond_alb.c b/drivers/net/bonding/bond_alb.c
> index bcf25c6..04b7055 100644
> --- a/drivers/net/bonding/bond_alb.c
> +++ b/drivers/net/bonding/bond_alb.c
> @@ -111,6 +111,7 @@ static inline struct arp_pkt *arp_pkt(const struct sk_buff *skb)
>
> /* Forward declaration */
> static void alb_send_learning_packets(struct slave *slave, u8 mac_addr[]);
> +static struct slave *alb_get_best_slave(struct bonding *bond, u32 hash_index);
>
> static inline u8 _simple_hash(const u8 *hash_start, int hash_size)
> {
> @@ -124,18 +125,18 @@ static inline u8 _simple_hash(const u8 *hash_start, int hash_size)
> return hash;
> }
>
> -/*********************** tlb specific functions ***************************/
> -
> -static inline void _lock_tx_hashtbl(struct bonding *bond)
> +/********************* hash table lock functions *************************/
> +static inline void _lock_hashtbl(struct bonding *bond)
> {
> - spin_lock_bh(&(BOND_ALB_INFO(bond).tx_hashtbl_lock));
> + spin_lock_bh(&(BOND_ALB_INFO(bond).hashtbl_lock));
> }
>
> -static inline void _unlock_tx_hashtbl(struct bonding *bond)
> +static inline void _unlock_hashtbl(struct bonding *bond)
> {
> - spin_unlock_bh(&(BOND_ALB_INFO(bond).tx_hashtbl_lock));
> + spin_unlock_bh(&(BOND_ALB_INFO(bond).hashtbl_lock));
> }
>
> +/*********************** tlb specific functions ***************************/
> /* Caller must hold tx_hashtbl lock */
> static inline void tlb_init_table_entry(struct tlb_client_info *entry, int save_load)
> {
> @@ -163,7 +164,7 @@ static void tlb_clear_slave(struct bonding *bond, struct slave *slave, int save_
> struct tlb_client_info *tx_hash_table;
> u32 index;
>
> - _lock_tx_hashtbl(bond);
> + _lock_hashtbl(bond);
>
> /* clear slave from tx_hashtbl */
> tx_hash_table = BOND_ALB_INFO(bond).tx_hashtbl;
> @@ -180,7 +181,7 @@ static void tlb_clear_slave(struct bonding *bond, struct slave *slave, int save_
>
> tlb_init_slave(slave);
>
> - _unlock_tx_hashtbl(bond);
> + _unlock_hashtbl(bond);
> }
>
> /* Must be called before starting the monitor timer */
> @@ -191,7 +192,7 @@ static int tlb_initialize(struct bonding *bond)
> struct tlb_client_info *new_hashtbl;
> int i;
>
> - spin_lock_init(&(bond_info->tx_hashtbl_lock));
> + spin_lock_init(&(bond_info->hashtbl_lock));
>
> new_hashtbl = kzalloc(size, GFP_KERNEL);
> if (!new_hashtbl) {
> @@ -200,7 +201,7 @@ static int tlb_initialize(struct bonding *bond)
> bond->dev->name);
> return -1;
> }
> - _lock_tx_hashtbl(bond);
> + _lock_hashtbl(bond);
>
> bond_info->tx_hashtbl = new_hashtbl;
>
> @@ -208,7 +209,7 @@ static int tlb_initialize(struct bonding *bond)
> tlb_init_table_entry(&bond_info->tx_hashtbl[i], 1);
> }
>
> - _unlock_tx_hashtbl(bond);
> + _unlock_hashtbl(bond);
>
> return 0;
> }
> @@ -218,12 +219,12 @@ static void tlb_deinitialize(struct bonding *bond)
> {
> struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
>
> - _lock_tx_hashtbl(bond);
> + _lock_hashtbl(bond);
>
> kfree(bond_info->tx_hashtbl);
> bond_info->tx_hashtbl = NULL;
>
> - _unlock_tx_hashtbl(bond);
> + _unlock_hashtbl(bond);
> }
>
> /* Caller must hold bond lock for read */
> @@ -264,24 +265,6 @@ static struct slave *tlb_get_least_loaded_slave(struct bonding *bond)
> return least_loaded;
> }
>
> -/* Caller must hold bond lock for read and hashtbl lock */
> -static struct slave *tlb_get_best_slave(struct bonding *bond, u32 hash_index)
> -{
> - struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
> - struct tlb_client_info *tx_hash_table = bond_info->tx_hashtbl;
> - struct slave *last_slave = tx_hash_table[hash_index].last_slave;
> - struct slave *next_slave = NULL;
> -
> - if (last_slave && SLAVE_IS_OK(last_slave)) {
> - /* Use the last slave listed in the tx hashtbl if:
> - the last slave currently is essentially unloaded. */
> - if (SLAVE_TLB_INFO(last_slave).load < 10)
> - next_slave = last_slave;
> - }
> -
> - return next_slave ? next_slave : tlb_get_least_loaded_slave(bond);
> -}
> -
> /* Caller must hold bond lock for read */
> static struct slave *tlb_choose_channel(struct bonding *bond, u32 hash_index, u32 skb_len)
> {
> @@ -289,13 +272,12 @@ static struct slave *tlb_choose_channel(struct bonding *bond, u32 hash_index, u3
> struct tlb_client_info *hash_table;
> struct slave *assigned_slave;
>
> - _lock_tx_hashtbl(bond);
> + _lock_hashtbl(bond);
>
> hash_table = bond_info->tx_hashtbl;
> assigned_slave = hash_table[hash_index].tx_slave;
> if (!assigned_slave) {
> - assigned_slave = tlb_get_best_slave(bond, hash_index);
> -
> + assigned_slave = alb_get_best_slave(bond, hash_index);
> if (assigned_slave) {
> struct tlb_slave_info *slave_info =
> &(SLAVE_TLB_INFO(assigned_slave));
> @@ -319,20 +301,52 @@ static struct slave *tlb_choose_channel(struct bonding *bond, u32 hash_index, u3
> hash_table[hash_index].tx_bytes += skb_len;
> }
>
> - _unlock_tx_hashtbl(bond);
> + _unlock_hashtbl(bond);
>
> return assigned_slave;
> }
>
> /*********************** rlb specific functions ***************************/
> -static inline void _lock_rx_hashtbl(struct bonding *bond)
> +
> +/* Caller must hold bond lock for read and hashtbl lock */
> +static struct slave *rlb_update_rx_table(struct bonding *bond, struct slave *next_slave, u32 hash_index)
> {
> - spin_lock_bh(&(BOND_ALB_INFO(bond).rx_hashtbl_lock));
> + struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
> +
> + /* check rlb table and correct it if wrong */
> + if (bond_info->rlb_enabled) {
> + struct rlb_client_info *rx_client_info = &(bond_info->rx_hashtbl[hash_index]);
> +
> + /* if the new slave computed by tlb checks doesn't match rlb, stop rlb from using it */
> + if (next_slave && (next_slave != rx_client_info->slave))
> + rx_client_info->slave = next_slave;
> + }
> + return next_slave;
> }
>
> -static inline void _unlock_rx_hashtbl(struct bonding *bond)
> +/* Caller must hold bond lock for read and hashtbl lock */
> +static struct slave *alb_get_best_slave(struct bonding *bond, u32 hash_index)
> {
> - spin_unlock_bh(&(BOND_ALB_INFO(bond).rx_hashtbl_lock));
> + struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
> + struct tlb_client_info *tx_hash_table = bond_info->tx_hashtbl;
> + struct slave *last_slave = tx_hash_table[hash_index].last_slave;
> + struct slave *next_slave = NULL;
> +
> + /* presume the next slave will be the least loaded one */
> + next_slave = tlb_get_least_loaded_slave(bond);
> +
> + if (last_slave && SLAVE_IS_OK(last_slave)) {
> + /* Use the last slave listed in the tx hashtbl if:
> + the last slave currently is essentially unloaded. */
> + if (SLAVE_TLB_INFO(last_slave).load < 10)
> + next_slave = last_slave;
> + }
> +
> + /* update the rlb hashtbl if there was a previous entry */
> + if (bond_info->rlb_enabled)
> + rlb_update_rx_table(bond, next_slave, hash_index);
> +
> + return next_slave;
> }
>
> /* when an ARP REPLY is received from a client update its info
> @@ -344,7 +358,7 @@ static void rlb_update_entry_from_arp(struct bonding *bond, struct arp_pkt *arp)
> struct rlb_client_info *client_info;
> u32 hash_index;
>
> - _lock_rx_hashtbl(bond);
> + _lock_hashtbl(bond);
>
> hash_index = _simple_hash((u8*)&(arp->ip_src), sizeof(arp->ip_src));
> client_info = &(bond_info->rx_hashtbl[hash_index]);
> @@ -358,7 +372,7 @@ static void rlb_update_entry_from_arp(struct bonding *bond, struct arp_pkt *arp)
> bond_info->rx_ntt = 1;
> }
>
> - _unlock_rx_hashtbl(bond);
> + _unlock_hashtbl(bond);
> }
>
> static int rlb_arp_recv(struct sk_buff *skb, struct net_device *bond_dev, struct packet_type *ptype, struct net_device *orig_dev)
> @@ -402,38 +416,6 @@ out:
> return res;
> }
>
> -/* Caller must hold bond lock for read */
> -static struct slave *rlb_next_rx_slave(struct bonding *bond)
> -{
> - struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
> - struct slave *rx_slave, *slave, *start_at;
> - int i = 0;
> -
> - if (bond_info->next_rx_slave) {
> - start_at = bond_info->next_rx_slave;
> - } else {
> - start_at = bond->first_slave;
> - }
> -
> - rx_slave = NULL;
> -
> - bond_for_each_slave_from(bond, slave, i, start_at) {
> - if (SLAVE_IS_OK(slave)) {
> - if (!rx_slave) {
> - rx_slave = slave;
> - } else if (slave->speed > rx_slave->speed) {
> - rx_slave = slave;
> - }
> - }
> - }
> -
> - if (rx_slave) {
> - bond_info->next_rx_slave = rx_slave->next;
> - }
> -
> - return rx_slave;
> -}
> -
> /* teach the switch the mac of a disabled slave
> * on the primary for fault tolerance
> *
> @@ -468,14 +450,14 @@ static void rlb_clear_slave(struct bonding *bond, struct slave *slave)
> u32 index, next_index;
>
> /* clear slave from rx_hashtbl */
> - _lock_rx_hashtbl(bond);
> + _lock_hashtbl(bond);
>
> rx_hash_table = bond_info->rx_hashtbl;
> index = bond_info->rx_hashtbl_head;
> for (; index != RLB_NULL_INDEX; index = next_index) {
> next_index = rx_hash_table[index].next;
> if (rx_hash_table[index].slave == slave) {
> - struct slave *assigned_slave = rlb_next_rx_slave(bond);
> + struct slave *assigned_slave = alb_get_best_slave(bond, index);
>
> if (assigned_slave) {
> rx_hash_table[index].slave = assigned_slave;
> @@ -499,7 +481,7 @@ static void rlb_clear_slave(struct bonding *bond, struct slave *slave)
> }
> }
>
> - _unlock_rx_hashtbl(bond);
> + _unlock_hashtbl(bond);
>
> write_lock_bh(&bond->curr_slave_lock);
>
> @@ -558,7 +540,7 @@ static void rlb_update_rx_clients(struct bonding *bond)
> struct rlb_client_info *client_info;
> u32 hash_index;
>
> - _lock_rx_hashtbl(bond);
> + _lock_hashtbl(bond);
>
> hash_index = bond_info->rx_hashtbl_head;
> for (; hash_index != RLB_NULL_INDEX; hash_index = client_info->next) {
> @@ -576,7 +558,7 @@ static void rlb_update_rx_clients(struct bonding *bond)
> */
> bond_info->rlb_update_delay_counter = RLB_UPDATE_DELAY;
>
> - _unlock_rx_hashtbl(bond);
> + _unlock_hashtbl(bond);
> }
>
> /* The slave was assigned a new mac address - update the clients */
> @@ -587,7 +569,7 @@ static void rlb_req_update_slave_clients(struct bonding *bond, struct slave *sla
> int ntt = 0;
> u32 hash_index;
>
> - _lock_rx_hashtbl(bond);
> + _lock_hashtbl(bond);
>
> hash_index = bond_info->rx_hashtbl_head;
> for (; hash_index != RLB_NULL_INDEX; hash_index = client_info->next) {
> @@ -607,7 +589,7 @@ static void rlb_req_update_slave_clients(struct bonding *bond, struct slave *sla
> bond_info->rlb_update_retry_counter = RLB_UPDATE_RETRY;
> }
>
> - _unlock_rx_hashtbl(bond);
> + _unlock_hashtbl(bond);
> }
>
> /* mark all clients using src_ip to be updated */
> @@ -617,7 +599,7 @@ static void rlb_req_update_subnet_clients(struct bonding *bond, __be32 src_ip)
> struct rlb_client_info *client_info;
> u32 hash_index;
>
> - _lock_rx_hashtbl(bond);
> + _lock_hashtbl(bond);
>
> hash_index = bond_info->rx_hashtbl_head;
> for (; hash_index != RLB_NULL_INDEX; hash_index = client_info->next) {
> @@ -643,7 +625,7 @@ static void rlb_req_update_subnet_clients(struct bonding *bond, __be32 src_ip)
> }
> }
>
> - _unlock_rx_hashtbl(bond);
> + _unlock_hashtbl(bond);
> }
>
> /* Caller must hold both bond and ptr locks for read */
> @@ -655,7 +637,7 @@ static struct slave *rlb_choose_channel(struct sk_buff *skb, struct bonding *bon
> struct rlb_client_info *client_info;
> u32 hash_index = 0;
>
> - _lock_rx_hashtbl(bond);
> + _lock_hashtbl(bond);
>
> hash_index = _simple_hash((u8 *)&arp->ip_dst, sizeof(arp->ip_src));
> client_info = &(bond_info->rx_hashtbl[hash_index]);
> @@ -671,7 +653,7 @@ static struct slave *rlb_choose_channel(struct sk_buff *skb, struct bonding *bon
>
> assigned_slave = client_info->slave;
> if (assigned_slave) {
> - _unlock_rx_hashtbl(bond);
> + _unlock_hashtbl(bond);
> return assigned_slave;
> }
> } else {
> @@ -687,7 +669,7 @@ static struct slave *rlb_choose_channel(struct sk_buff *skb, struct bonding *bon
> }
> }
> /* assign a new slave */
> - assigned_slave = rlb_next_rx_slave(bond);
> + assigned_slave = alb_get_best_slave(bond, hash_index);
>
> if (assigned_slave) {
> client_info->ip_src = arp->ip_src;
> @@ -723,7 +705,7 @@ static struct slave *rlb_choose_channel(struct sk_buff *skb, struct bonding *bon
> }
> }
>
> - _unlock_rx_hashtbl(bond);
> + _unlock_hashtbl(bond);
>
> return assigned_slave;
> }
> @@ -771,36 +753,6 @@ static struct slave *rlb_arp_xmit(struct sk_buff *skb, struct bonding *bond)
> return tx_slave;
> }
>
> -/* Caller must hold bond lock for read */
> -static void rlb_rebalance(struct bonding *bond)
> -{
> - struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
> - struct slave *assigned_slave;
> - struct rlb_client_info *client_info;
> - int ntt;
> - u32 hash_index;
> -
> - _lock_rx_hashtbl(bond);
> -
> - ntt = 0;
> - hash_index = bond_info->rx_hashtbl_head;
> - for (; hash_index != RLB_NULL_INDEX; hash_index = client_info->next) {
> - client_info = &(bond_info->rx_hashtbl[hash_index]);
> - assigned_slave = rlb_next_rx_slave(bond);
> - if (assigned_slave && (client_info->slave != assigned_slave)) {
> - client_info->slave = assigned_slave;
> - client_info->ntt = 1;
> - ntt = 1;
> - }
> - }
> -
> - /* update the team's flag only after the whole iteration */
> - if (ntt) {
> - bond_info->rx_ntt = 1;
> - }
> - _unlock_rx_hashtbl(bond);
> -}
> -
> /* Caller must hold rx_hashtbl lock */
> static void rlb_init_table_entry(struct rlb_client_info *entry)
> {
> @@ -817,8 +769,6 @@ static int rlb_initialize(struct bonding *bond)
> int size = RLB_HASH_TABLE_SIZE * sizeof(struct rlb_client_info);
> int i;
>
> - spin_lock_init(&(bond_info->rx_hashtbl_lock));
> -
> new_hashtbl = kmalloc(size, GFP_KERNEL);
> if (!new_hashtbl) {
> printk(KERN_ERR DRV_NAME
> @@ -826,7 +776,7 @@ static int rlb_initialize(struct bonding *bond)
> bond->dev->name);
> return -1;
> }
> - _lock_rx_hashtbl(bond);
> + _lock_hashtbl(bond);
>
> bond_info->rx_hashtbl = new_hashtbl;
>
> @@ -836,7 +786,7 @@ static int rlb_initialize(struct bonding *bond)
> rlb_init_table_entry(bond_info->rx_hashtbl + i);
> }
>
> - _unlock_rx_hashtbl(bond);
> + _unlock_hashtbl(bond);
>
> /*initialize packet type*/
> pk_type->type = cpu_to_be16(ETH_P_ARP);
> @@ -855,13 +805,13 @@ static void rlb_deinitialize(struct bonding *bond)
>
> dev_remove_pack(&(bond_info->rlb_pkt_type));
>
> - _lock_rx_hashtbl(bond);
> + _lock_hashtbl(bond);
>
> kfree(bond_info->rx_hashtbl);
> bond_info->rx_hashtbl = NULL;
> bond_info->rx_hashtbl_head = RLB_NULL_INDEX;
>
> - _unlock_rx_hashtbl(bond);
> + _unlock_hashtbl(bond);
> }
>
> static void rlb_clear_vlan(struct bonding *bond, unsigned short vlan_id)
> @@ -869,7 +819,7 @@ static void rlb_clear_vlan(struct bonding *bond, unsigned short vlan_id)
> struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
> u32 curr_index;
>
> - _lock_rx_hashtbl(bond);
> + _lock_hashtbl(bond);
>
> curr_index = bond_info->rx_hashtbl_head;
> while (curr_index != RLB_NULL_INDEX) {
> @@ -894,7 +844,7 @@ static void rlb_clear_vlan(struct bonding *bond, unsigned short vlan_id)
> curr_index = next_index;
> }
>
> - _unlock_rx_hashtbl(bond);
> + _unlock_hashtbl(bond);
> }
>
> /*********************** tlb/rlb shared functions *********************/
> @@ -1521,11 +1471,6 @@ void bond_alb_monitor(struct work_struct *work)
> read_lock(&bond->lock);
> }
>
> - if (bond_info->rlb_rebalance) {
> - bond_info->rlb_rebalance = 0;
> - rlb_rebalance(bond);
> - }
> -
> /* check if clients need updating */
> if (bond_info->rx_ntt) {
> if (bond_info->rlb_update_delay_counter) {
> diff --git a/drivers/net/bonding/bond_alb.h b/drivers/net/bonding/bond_alb.h
> index b65fd29..09d755a 100644
> --- a/drivers/net/bonding/bond_alb.h
> +++ b/drivers/net/bonding/bond_alb.h
> @@ -90,7 +90,7 @@ struct tlb_slave_info {
> struct alb_bond_info {
> struct timer_list alb_timer;
> struct tlb_client_info *tx_hashtbl; /* Dynamically allocated */
> - spinlock_t tx_hashtbl_lock;
> + spinlock_t hashtbl_lock; /* lock for both tables */
> u32 unbalanced_load;
> int tx_rebalance_counter;
> int lp_counter;
> @@ -98,7 +98,6 @@ struct alb_bond_info {
> int rlb_enabled;
> struct packet_type rlb_pkt_type;
> struct rlb_client_info *rx_hashtbl; /* Receive hash table */
> - spinlock_t rx_hashtbl_lock;
> u32 rx_hashtbl_head;
> u8 rx_ntt; /* flag - need to transmit
> * to all rx clients
Any thoughts on this, Jay?
^ permalink raw reply
* Re: tg3 and Broadcom PHY driver
From: David Miller @ 2009-09-28 21:55 UTC (permalink / raw)
To: felix; +Cc: mcarlson, netdev
In-Reply-To: <4AC13036.8030506@embedded-sol.com>
From: Felix Radensky <felix@embedded-sol.com>
Date: Mon, 28 Sep 2009 23:52:54 +0200
> Yes, moving CONFIG_TIGON3 right after CONFIG_PHYLIB in
> drivers/net/Makefile fixes the problem for me.
Thanks for testing.
We really need to fix this generically.
Does anyone think that moving the MDIO/MII/PHY layer objects
to the top of drivers/net/Makefile will break anything?
If not, that's what we should do I think.
^ permalink raw reply
* Re: tg3 and Broadcom PHY driver
From: Felix Radensky @ 2009-09-28 21:52 UTC (permalink / raw)
To: David Miller; +Cc: mcarlson, netdev
In-Reply-To: <20090928.141527.193009676.davem@davemloft.net>
Hi, David
David Miller wrote:
> From: Felix Radensky <felix@embedded-sol.com>
> Date: Mon, 28 Sep 2009 22:53:15 +0200
>
>
>> Hi, Matt
>>
>> Matt Carlson wrote:
>>
>>> On Sat, Sep 26, 2009 at 02:32:18PM -0700, Felix Radensky wrote:
>>>
>>> Is the broadcom module also compiled into the kernel?
>>>
>>>
>>>
>> Yes.
>>
>
> I bet this is all because the tg3 driver is linked into the kernel
> before the PHY layer and drivers subdirectory or something like that.
>
> Link order determines the order in which built-in initializations
> occur (within the same init type).
>
Yes, moving CONFIG_TIGON3 right after CONFIG_PHYLIB in drivers/net/Makefile
fixes the problem for me.
Thanks a lot.
Felix.
^ permalink raw reply
* Re: pull request: wireless-2.6 2009-09-28
From: David Miller @ 2009-09-28 21:51 UTC (permalink / raw)
To: linville; +Cc: linux-wireless, netdev, linux-kernel
In-Reply-To: <20090928214047.GF4737@tuxdriver.com>
From: "John W. Linville" <linville@tuxdriver.com>
Date: Mon, 28 Sep 2009 17:40:48 -0400
> There are number of small fixes to core wireless infrastructure from
> Johannes Berg. These have been bounced around on the list quite a bit
> during the last several days, so I think they are solid. Also included
> is one that improves some debugging messages.
>
> There are a few fixes for the iwlwifi family of drivers, including a
> buffer overrun, a memory leak, and another debugging message fix.
>
> Arjan showed-up with a bounds checking fix in some ancient wext code. I
> included it due to the potential security implications.
>
> The sony-laptop fixes may seem a bit out of place, but they are related
> to rfkill. Also, they are "Acked-by: Mattia Dongili <malattia@linux.it>",
> who is the listed maintainer for that code. I put them at the end, so
> if you want you can pull from 8f1546cadf7ac5e9a40d54089a1c7302264ec49b
> instead of master.
>
> Please let me know if there are problems!
Pulled, thanks a lot John!
^ permalink raw reply
* pull request: wireless-2.6 2009-09-28
From: John W. Linville @ 2009-09-28 21:40 UTC (permalink / raw)
To: davem; +Cc: linux-wireless, netdev, linux-kernel
Dave,
This group of fixes is intended for the 2.6.32 train...
There are number of small fixes to core wireless infrastructure from
Johannes Berg. These have been bounced around on the list quite a bit
during the last several days, so I think they are solid. Also included
is one that improves some debugging messages.
There are a few fixes for the iwlwifi family of drivers, including a
buffer overrun, a memory leak, and another debugging message fix.
Arjan showed-up with a bounds checking fix in some ancient wext code. I
included it due to the potential security implications.
The sony-laptop fixes may seem a bit out of place, but they are related
to rfkill. Also, they are "Acked-by: Mattia Dongili <malattia@linux.it>",
who is the listed maintainer for that code. I put them at the end, so
if you want you can pull from 8f1546cadf7ac5e9a40d54089a1c7302264ec49b
instead of master.
Please let me know if there are problems!
Thanks,
John
---
Individual patches are available here:
http://www.kernel.org/pub/linux/kernel/people/linville/wireless-2.6/
---
The following changes since commit d1f8297a96b0d70f17704296a6666468f2087ce6:
Sascha Hlusiak (1):
Revert "sit: stateless autoconf for isatap"
are available in the git repository at:
ssh://master.kernel.org/pub/scm/linux/kernel/git/linville/wireless-2.6.git master
Alan Jenkins (2):
sony-laptop: check for rfkill hard block at load time
sony-laptop: re-read the rfkill state when resuming from suspend
Arjan van de Ven (1):
wext: Add bound checks for copy_from_user
Johannes Berg (5):
cfg80211: wext: don't display BSSID unless associated
cfg80211: don't set privacy w/o key
cfg80211: always get BSS
mac80211: improve/fix mlme messages
wext: add back wireless/ dir in sysfs for cfg80211 interfaces
Reinette Chatre (3):
iwlwifi: fix debugfs buffer handling
iwlwifi: fix memory leak in command queue handling
iwlwifi: fix 3945 ucode info retrieval after failure
drivers/net/wireless/iwlwifi/iwl-1000.c | 2 +
drivers/net/wireless/iwlwifi/iwl-3945.c | 2 +
drivers/net/wireless/iwlwifi/iwl-3945.h | 2 +
drivers/net/wireless/iwlwifi/iwl-4965.c | 2 +
drivers/net/wireless/iwlwifi/iwl-5000.c | 4 +
drivers/net/wireless/iwlwifi/iwl-6000.c | 2 +
drivers/net/wireless/iwlwifi/iwl-agn.c | 185 ++++++++++++++++++++++++++
drivers/net/wireless/iwlwifi/iwl-core.c | 187 +--------------------------
drivers/net/wireless/iwlwifi/iwl-core.h | 14 ++
drivers/net/wireless/iwlwifi/iwl-debugfs.c | 8 +-
drivers/net/wireless/iwlwifi/iwl-tx.c | 6 +
drivers/net/wireless/iwlwifi/iwl3945-base.c | 31 ++---
drivers/platform/x86/sony-laptop.c | 9 ++
include/net/wext.h | 1 +
net/core/net-sysfs.c | 12 +-
net/mac80211/mlme.c | 18 ++--
net/wireless/sme.c | 5 +-
net/wireless/wext-sme.c | 8 +-
net/wireless/wext.c | 11 +-
19 files changed, 274 insertions(+), 235 deletions(-)
diff --git a/drivers/net/wireless/iwlwifi/iwl-1000.c b/drivers/net/wireless/iwlwifi/iwl-1000.c
index a95caa0..2716b91 100644
--- a/drivers/net/wireless/iwlwifi/iwl-1000.c
+++ b/drivers/net/wireless/iwlwifi/iwl-1000.c
@@ -99,6 +99,8 @@ static struct iwl_lib_ops iwl1000_lib = {
.setup_deferred_work = iwl5000_setup_deferred_work,
.is_valid_rtc_data_addr = iwl5000_hw_valid_rtc_data_addr,
.load_ucode = iwl5000_load_ucode,
+ .dump_nic_event_log = iwl_dump_nic_event_log,
+ .dump_nic_error_log = iwl_dump_nic_error_log,
.init_alive_start = iwl5000_init_alive_start,
.alive_notify = iwl5000_alive_notify,
.send_tx_power = iwl5000_send_tx_power,
diff --git a/drivers/net/wireless/iwlwifi/iwl-3945.c b/drivers/net/wireless/iwlwifi/iwl-3945.c
index e9a685d..e70c5b0 100644
--- a/drivers/net/wireless/iwlwifi/iwl-3945.c
+++ b/drivers/net/wireless/iwlwifi/iwl-3945.c
@@ -2839,6 +2839,8 @@ static struct iwl_lib_ops iwl3945_lib = {
.txq_free_tfd = iwl3945_hw_txq_free_tfd,
.txq_init = iwl3945_hw_tx_queue_init,
.load_ucode = iwl3945_load_bsm,
+ .dump_nic_event_log = iwl3945_dump_nic_event_log,
+ .dump_nic_error_log = iwl3945_dump_nic_error_log,
.apm_ops = {
.init = iwl3945_apm_init,
.reset = iwl3945_apm_reset,
diff --git a/drivers/net/wireless/iwlwifi/iwl-3945.h b/drivers/net/wireless/iwlwifi/iwl-3945.h
index f240369..21679bf 100644
--- a/drivers/net/wireless/iwlwifi/iwl-3945.h
+++ b/drivers/net/wireless/iwlwifi/iwl-3945.h
@@ -209,6 +209,8 @@ extern int __must_check iwl3945_send_cmd(struct iwl_priv *priv,
struct iwl_host_cmd *cmd);
extern unsigned int iwl3945_fill_beacon_frame(struct iwl_priv *priv,
struct ieee80211_hdr *hdr,int left);
+extern void iwl3945_dump_nic_event_log(struct iwl_priv *priv);
+extern void iwl3945_dump_nic_error_log(struct iwl_priv *priv);
/*
* Currently used by iwl-3945-rs... look at restructuring so that it doesn't
diff --git a/drivers/net/wireless/iwlwifi/iwl-4965.c b/drivers/net/wireless/iwlwifi/iwl-4965.c
index 3259b88..a22a050 100644
--- a/drivers/net/wireless/iwlwifi/iwl-4965.c
+++ b/drivers/net/wireless/iwlwifi/iwl-4965.c
@@ -2298,6 +2298,8 @@ static struct iwl_lib_ops iwl4965_lib = {
.alive_notify = iwl4965_alive_notify,
.init_alive_start = iwl4965_init_alive_start,
.load_ucode = iwl4965_load_bsm,
+ .dump_nic_event_log = iwl_dump_nic_event_log,
+ .dump_nic_error_log = iwl_dump_nic_error_log,
.apm_ops = {
.init = iwl4965_apm_init,
.reset = iwl4965_apm_reset,
diff --git a/drivers/net/wireless/iwlwifi/iwl-5000.c b/drivers/net/wireless/iwlwifi/iwl-5000.c
index a6391c7..eb08f44 100644
--- a/drivers/net/wireless/iwlwifi/iwl-5000.c
+++ b/drivers/net/wireless/iwlwifi/iwl-5000.c
@@ -1535,6 +1535,8 @@ struct iwl_lib_ops iwl5000_lib = {
.rx_handler_setup = iwl5000_rx_handler_setup,
.setup_deferred_work = iwl5000_setup_deferred_work,
.is_valid_rtc_data_addr = iwl5000_hw_valid_rtc_data_addr,
+ .dump_nic_event_log = iwl_dump_nic_event_log,
+ .dump_nic_error_log = iwl_dump_nic_error_log,
.load_ucode = iwl5000_load_ucode,
.init_alive_start = iwl5000_init_alive_start,
.alive_notify = iwl5000_alive_notify,
@@ -1585,6 +1587,8 @@ static struct iwl_lib_ops iwl5150_lib = {
.rx_handler_setup = iwl5000_rx_handler_setup,
.setup_deferred_work = iwl5000_setup_deferred_work,
.is_valid_rtc_data_addr = iwl5000_hw_valid_rtc_data_addr,
+ .dump_nic_event_log = iwl_dump_nic_event_log,
+ .dump_nic_error_log = iwl_dump_nic_error_log,
.load_ucode = iwl5000_load_ucode,
.init_alive_start = iwl5000_init_alive_start,
.alive_notify = iwl5000_alive_notify,
diff --git a/drivers/net/wireless/iwlwifi/iwl-6000.c b/drivers/net/wireless/iwlwifi/iwl-6000.c
index 82b9c93..c295b8e 100644
--- a/drivers/net/wireless/iwlwifi/iwl-6000.c
+++ b/drivers/net/wireless/iwlwifi/iwl-6000.c
@@ -100,6 +100,8 @@ static struct iwl_lib_ops iwl6000_lib = {
.setup_deferred_work = iwl5000_setup_deferred_work,
.is_valid_rtc_data_addr = iwl5000_hw_valid_rtc_data_addr,
.load_ucode = iwl5000_load_ucode,
+ .dump_nic_event_log = iwl_dump_nic_event_log,
+ .dump_nic_error_log = iwl_dump_nic_error_log,
.init_alive_start = iwl5000_init_alive_start,
.alive_notify = iwl5000_alive_notify,
.send_tx_power = iwl5000_send_tx_power,
diff --git a/drivers/net/wireless/iwlwifi/iwl-agn.c b/drivers/net/wireless/iwlwifi/iwl-agn.c
index 00457bf..cdc07c4 100644
--- a/drivers/net/wireless/iwlwifi/iwl-agn.c
+++ b/drivers/net/wireless/iwlwifi/iwl-agn.c
@@ -1526,6 +1526,191 @@ static int iwl_read_ucode(struct iwl_priv *priv)
return ret;
}
+#ifdef CONFIG_IWLWIFI_DEBUG
+static const char *desc_lookup_text[] = {
+ "OK",
+ "FAIL",
+ "BAD_PARAM",
+ "BAD_CHECKSUM",
+ "NMI_INTERRUPT_WDG",
+ "SYSASSERT",
+ "FATAL_ERROR",
+ "BAD_COMMAND",
+ "HW_ERROR_TUNE_LOCK",
+ "HW_ERROR_TEMPERATURE",
+ "ILLEGAL_CHAN_FREQ",
+ "VCC_NOT_STABLE",
+ "FH_ERROR",
+ "NMI_INTERRUPT_HOST",
+ "NMI_INTERRUPT_ACTION_PT",
+ "NMI_INTERRUPT_UNKNOWN",
+ "UCODE_VERSION_MISMATCH",
+ "HW_ERROR_ABS_LOCK",
+ "HW_ERROR_CAL_LOCK_FAIL",
+ "NMI_INTERRUPT_INST_ACTION_PT",
+ "NMI_INTERRUPT_DATA_ACTION_PT",
+ "NMI_TRM_HW_ER",
+ "NMI_INTERRUPT_TRM",
+ "NMI_INTERRUPT_BREAK_POINT"
+ "DEBUG_0",
+ "DEBUG_1",
+ "DEBUG_2",
+ "DEBUG_3",
+ "UNKNOWN"
+};
+
+static const char *desc_lookup(int i)
+{
+ int max = ARRAY_SIZE(desc_lookup_text) - 1;
+
+ if (i < 0 || i > max)
+ i = max;
+
+ return desc_lookup_text[i];
+}
+
+#define ERROR_START_OFFSET (1 * sizeof(u32))
+#define ERROR_ELEM_SIZE (7 * sizeof(u32))
+
+void iwl_dump_nic_error_log(struct iwl_priv *priv)
+{
+ u32 data2, line;
+ u32 desc, time, count, base, data1;
+ u32 blink1, blink2, ilink1, ilink2;
+
+ if (priv->ucode_type == UCODE_INIT)
+ base = le32_to_cpu(priv->card_alive_init.error_event_table_ptr);
+ else
+ base = le32_to_cpu(priv->card_alive.error_event_table_ptr);
+
+ if (!priv->cfg->ops->lib->is_valid_rtc_data_addr(base)) {
+ IWL_ERR(priv, "Not valid error log pointer 0x%08X\n", base);
+ return;
+ }
+
+ count = iwl_read_targ_mem(priv, base);
+
+ if (ERROR_START_OFFSET <= count * ERROR_ELEM_SIZE) {
+ IWL_ERR(priv, "Start IWL Error Log Dump:\n");
+ IWL_ERR(priv, "Status: 0x%08lX, count: %d\n",
+ priv->status, count);
+ }
+
+ desc = iwl_read_targ_mem(priv, base + 1 * sizeof(u32));
+ blink1 = iwl_read_targ_mem(priv, base + 3 * sizeof(u32));
+ blink2 = iwl_read_targ_mem(priv, base + 4 * sizeof(u32));
+ ilink1 = iwl_read_targ_mem(priv, base + 5 * sizeof(u32));
+ ilink2 = iwl_read_targ_mem(priv, base + 6 * sizeof(u32));
+ data1 = iwl_read_targ_mem(priv, base + 7 * sizeof(u32));
+ data2 = iwl_read_targ_mem(priv, base + 8 * sizeof(u32));
+ line = iwl_read_targ_mem(priv, base + 9 * sizeof(u32));
+ time = iwl_read_targ_mem(priv, base + 11 * sizeof(u32));
+
+ IWL_ERR(priv, "Desc Time "
+ "data1 data2 line\n");
+ IWL_ERR(priv, "%-28s (#%02d) %010u 0x%08X 0x%08X %u\n",
+ desc_lookup(desc), desc, time, data1, data2, line);
+ IWL_ERR(priv, "blink1 blink2 ilink1 ilink2\n");
+ IWL_ERR(priv, "0x%05X 0x%05X 0x%05X 0x%05X\n", blink1, blink2,
+ ilink1, ilink2);
+
+}
+
+#define EVENT_START_OFFSET (4 * sizeof(u32))
+
+/**
+ * iwl_print_event_log - Dump error event log to syslog
+ *
+ */
+static void iwl_print_event_log(struct iwl_priv *priv, u32 start_idx,
+ u32 num_events, u32 mode)
+{
+ u32 i;
+ u32 base; /* SRAM byte address of event log header */
+ u32 event_size; /* 2 u32s, or 3 u32s if timestamp recorded */
+ u32 ptr; /* SRAM byte address of log data */
+ u32 ev, time, data; /* event log data */
+
+ if (num_events == 0)
+ return;
+ if (priv->ucode_type == UCODE_INIT)
+ base = le32_to_cpu(priv->card_alive_init.log_event_table_ptr);
+ else
+ base = le32_to_cpu(priv->card_alive.log_event_table_ptr);
+
+ if (mode == 0)
+ event_size = 2 * sizeof(u32);
+ else
+ event_size = 3 * sizeof(u32);
+
+ ptr = base + EVENT_START_OFFSET + (start_idx * event_size);
+
+ /* "time" is actually "data" for mode 0 (no timestamp).
+ * place event id # at far right for easier visual parsing. */
+ for (i = 0; i < num_events; i++) {
+ ev = iwl_read_targ_mem(priv, ptr);
+ ptr += sizeof(u32);
+ time = iwl_read_targ_mem(priv, ptr);
+ ptr += sizeof(u32);
+ if (mode == 0) {
+ /* data, ev */
+ IWL_ERR(priv, "EVT_LOG:0x%08x:%04u\n", time, ev);
+ } else {
+ data = iwl_read_targ_mem(priv, ptr);
+ ptr += sizeof(u32);
+ IWL_ERR(priv, "EVT_LOGT:%010u:0x%08x:%04u\n",
+ time, data, ev);
+ }
+ }
+}
+
+void iwl_dump_nic_event_log(struct iwl_priv *priv)
+{
+ u32 base; /* SRAM byte address of event log header */
+ u32 capacity; /* event log capacity in # entries */
+ u32 mode; /* 0 - no timestamp, 1 - timestamp recorded */
+ u32 num_wraps; /* # times uCode wrapped to top of log */
+ u32 next_entry; /* index of next entry to be written by uCode */
+ u32 size; /* # entries that we'll print */
+
+ if (priv->ucode_type == UCODE_INIT)
+ base = le32_to_cpu(priv->card_alive_init.log_event_table_ptr);
+ else
+ base = le32_to_cpu(priv->card_alive.log_event_table_ptr);
+
+ if (!priv->cfg->ops->lib->is_valid_rtc_data_addr(base)) {
+ IWL_ERR(priv, "Invalid event log pointer 0x%08X\n", base);
+ return;
+ }
+
+ /* event log header */
+ capacity = iwl_read_targ_mem(priv, base);
+ mode = iwl_read_targ_mem(priv, base + (1 * sizeof(u32)));
+ num_wraps = iwl_read_targ_mem(priv, base + (2 * sizeof(u32)));
+ next_entry = iwl_read_targ_mem(priv, base + (3 * sizeof(u32)));
+
+ size = num_wraps ? capacity : next_entry;
+
+ /* bail out if nothing in log */
+ if (size == 0) {
+ IWL_ERR(priv, "Start IWL Event Log Dump: nothing in log\n");
+ return;
+ }
+
+ IWL_ERR(priv, "Start IWL Event Log Dump: display count %d, wraps %d\n",
+ size, num_wraps);
+
+ /* if uCode has wrapped back to top of log, start at the oldest entry,
+ * i.e the next one that uCode would fill. */
+ if (num_wraps)
+ iwl_print_event_log(priv, next_entry,
+ capacity - next_entry, mode);
+ /* (then/else) start at top of log */
+ iwl_print_event_log(priv, 0, next_entry, mode);
+
+}
+#endif
+
/**
* iwl_alive_start - called after REPLY_ALIVE notification received
* from protocol/runtime uCode (initialization uCode's
diff --git a/drivers/net/wireless/iwlwifi/iwl-core.c b/drivers/net/wireless/iwlwifi/iwl-core.c
index fd26c0d..484d5c1 100644
--- a/drivers/net/wireless/iwlwifi/iwl-core.c
+++ b/drivers/net/wireless/iwlwifi/iwl-core.c
@@ -1309,189 +1309,6 @@ static void iwl_print_rx_config_cmd(struct iwl_priv *priv)
IWL_DEBUG_RADIO(priv, "u8[6] bssid_addr: %pM\n", rxon->bssid_addr);
IWL_DEBUG_RADIO(priv, "u16 assoc_id: 0x%x\n", le16_to_cpu(rxon->assoc_id));
}
-
-static const char *desc_lookup_text[] = {
- "OK",
- "FAIL",
- "BAD_PARAM",
- "BAD_CHECKSUM",
- "NMI_INTERRUPT_WDG",
- "SYSASSERT",
- "FATAL_ERROR",
- "BAD_COMMAND",
- "HW_ERROR_TUNE_LOCK",
- "HW_ERROR_TEMPERATURE",
- "ILLEGAL_CHAN_FREQ",
- "VCC_NOT_STABLE",
- "FH_ERROR",
- "NMI_INTERRUPT_HOST",
- "NMI_INTERRUPT_ACTION_PT",
- "NMI_INTERRUPT_UNKNOWN",
- "UCODE_VERSION_MISMATCH",
- "HW_ERROR_ABS_LOCK",
- "HW_ERROR_CAL_LOCK_FAIL",
- "NMI_INTERRUPT_INST_ACTION_PT",
- "NMI_INTERRUPT_DATA_ACTION_PT",
- "NMI_TRM_HW_ER",
- "NMI_INTERRUPT_TRM",
- "NMI_INTERRUPT_BREAK_POINT"
- "DEBUG_0",
- "DEBUG_1",
- "DEBUG_2",
- "DEBUG_3",
- "UNKNOWN"
-};
-
-static const char *desc_lookup(int i)
-{
- int max = ARRAY_SIZE(desc_lookup_text) - 1;
-
- if (i < 0 || i > max)
- i = max;
-
- return desc_lookup_text[i];
-}
-
-#define ERROR_START_OFFSET (1 * sizeof(u32))
-#define ERROR_ELEM_SIZE (7 * sizeof(u32))
-
-static void iwl_dump_nic_error_log(struct iwl_priv *priv)
-{
- u32 data2, line;
- u32 desc, time, count, base, data1;
- u32 blink1, blink2, ilink1, ilink2;
-
- if (priv->ucode_type == UCODE_INIT)
- base = le32_to_cpu(priv->card_alive_init.error_event_table_ptr);
- else
- base = le32_to_cpu(priv->card_alive.error_event_table_ptr);
-
- if (!priv->cfg->ops->lib->is_valid_rtc_data_addr(base)) {
- IWL_ERR(priv, "Not valid error log pointer 0x%08X\n", base);
- return;
- }
-
- count = iwl_read_targ_mem(priv, base);
-
- if (ERROR_START_OFFSET <= count * ERROR_ELEM_SIZE) {
- IWL_ERR(priv, "Start IWL Error Log Dump:\n");
- IWL_ERR(priv, "Status: 0x%08lX, count: %d\n",
- priv->status, count);
- }
-
- desc = iwl_read_targ_mem(priv, base + 1 * sizeof(u32));
- blink1 = iwl_read_targ_mem(priv, base + 3 * sizeof(u32));
- blink2 = iwl_read_targ_mem(priv, base + 4 * sizeof(u32));
- ilink1 = iwl_read_targ_mem(priv, base + 5 * sizeof(u32));
- ilink2 = iwl_read_targ_mem(priv, base + 6 * sizeof(u32));
- data1 = iwl_read_targ_mem(priv, base + 7 * sizeof(u32));
- data2 = iwl_read_targ_mem(priv, base + 8 * sizeof(u32));
- line = iwl_read_targ_mem(priv, base + 9 * sizeof(u32));
- time = iwl_read_targ_mem(priv, base + 11 * sizeof(u32));
-
- IWL_ERR(priv, "Desc Time "
- "data1 data2 line\n");
- IWL_ERR(priv, "%-28s (#%02d) %010u 0x%08X 0x%08X %u\n",
- desc_lookup(desc), desc, time, data1, data2, line);
- IWL_ERR(priv, "blink1 blink2 ilink1 ilink2\n");
- IWL_ERR(priv, "0x%05X 0x%05X 0x%05X 0x%05X\n", blink1, blink2,
- ilink1, ilink2);
-
-}
-
-#define EVENT_START_OFFSET (4 * sizeof(u32))
-
-/**
- * iwl_print_event_log - Dump error event log to syslog
- *
- */
-static void iwl_print_event_log(struct iwl_priv *priv, u32 start_idx,
- u32 num_events, u32 mode)
-{
- u32 i;
- u32 base; /* SRAM byte address of event log header */
- u32 event_size; /* 2 u32s, or 3 u32s if timestamp recorded */
- u32 ptr; /* SRAM byte address of log data */
- u32 ev, time, data; /* event log data */
-
- if (num_events == 0)
- return;
- if (priv->ucode_type == UCODE_INIT)
- base = le32_to_cpu(priv->card_alive_init.log_event_table_ptr);
- else
- base = le32_to_cpu(priv->card_alive.log_event_table_ptr);
-
- if (mode == 0)
- event_size = 2 * sizeof(u32);
- else
- event_size = 3 * sizeof(u32);
-
- ptr = base + EVENT_START_OFFSET + (start_idx * event_size);
-
- /* "time" is actually "data" for mode 0 (no timestamp).
- * place event id # at far right for easier visual parsing. */
- for (i = 0; i < num_events; i++) {
- ev = iwl_read_targ_mem(priv, ptr);
- ptr += sizeof(u32);
- time = iwl_read_targ_mem(priv, ptr);
- ptr += sizeof(u32);
- if (mode == 0) {
- /* data, ev */
- IWL_ERR(priv, "EVT_LOG:0x%08x:%04u\n", time, ev);
- } else {
- data = iwl_read_targ_mem(priv, ptr);
- ptr += sizeof(u32);
- IWL_ERR(priv, "EVT_LOGT:%010u:0x%08x:%04u\n",
- time, data, ev);
- }
- }
-}
-
-void iwl_dump_nic_event_log(struct iwl_priv *priv)
-{
- u32 base; /* SRAM byte address of event log header */
- u32 capacity; /* event log capacity in # entries */
- u32 mode; /* 0 - no timestamp, 1 - timestamp recorded */
- u32 num_wraps; /* # times uCode wrapped to top of log */
- u32 next_entry; /* index of next entry to be written by uCode */
- u32 size; /* # entries that we'll print */
-
- if (priv->ucode_type == UCODE_INIT)
- base = le32_to_cpu(priv->card_alive_init.log_event_table_ptr);
- else
- base = le32_to_cpu(priv->card_alive.log_event_table_ptr);
-
- if (!priv->cfg->ops->lib->is_valid_rtc_data_addr(base)) {
- IWL_ERR(priv, "Invalid event log pointer 0x%08X\n", base);
- return;
- }
-
- /* event log header */
- capacity = iwl_read_targ_mem(priv, base);
- mode = iwl_read_targ_mem(priv, base + (1 * sizeof(u32)));
- num_wraps = iwl_read_targ_mem(priv, base + (2 * sizeof(u32)));
- next_entry = iwl_read_targ_mem(priv, base + (3 * sizeof(u32)));
-
- size = num_wraps ? capacity : next_entry;
-
- /* bail out if nothing in log */
- if (size == 0) {
- IWL_ERR(priv, "Start IWL Event Log Dump: nothing in log\n");
- return;
- }
-
- IWL_ERR(priv, "Start IWL Event Log Dump: display count %d, wraps %d\n",
- size, num_wraps);
-
- /* if uCode has wrapped back to top of log, start at the oldest entry,
- * i.e the next one that uCode would fill. */
- if (num_wraps)
- iwl_print_event_log(priv, next_entry,
- capacity - next_entry, mode);
- /* (then/else) start at top of log */
- iwl_print_event_log(priv, 0, next_entry, mode);
-
-}
#endif
/**
* iwl_irq_handle_error - called for HW or SW error interrupt from card
@@ -1506,8 +1323,8 @@ void iwl_irq_handle_error(struct iwl_priv *priv)
#ifdef CONFIG_IWLWIFI_DEBUG
if (iwl_get_debug_level(priv) & IWL_DL_FW_ERRORS) {
- iwl_dump_nic_error_log(priv);
- iwl_dump_nic_event_log(priv);
+ priv->cfg->ops->lib->dump_nic_error_log(priv);
+ priv->cfg->ops->lib->dump_nic_event_log(priv);
iwl_print_rx_config_cmd(priv);
}
#endif
diff --git a/drivers/net/wireless/iwlwifi/iwl-core.h b/drivers/net/wireless/iwlwifi/iwl-core.h
index 7ff9ffb..e50103a 100644
--- a/drivers/net/wireless/iwlwifi/iwl-core.h
+++ b/drivers/net/wireless/iwlwifi/iwl-core.h
@@ -166,6 +166,8 @@ struct iwl_lib_ops {
int (*is_valid_rtc_data_addr)(u32 addr);
/* 1st ucode load */
int (*load_ucode)(struct iwl_priv *priv);
+ void (*dump_nic_event_log)(struct iwl_priv *priv);
+ void (*dump_nic_error_log)(struct iwl_priv *priv);
/* power management */
struct iwl_apm_ops apm_ops;
@@ -540,7 +542,19 @@ int iwl_pci_resume(struct pci_dev *pdev);
/*****************************************************
* Error Handling Debugging
******************************************************/
+#ifdef CONFIG_IWLWIFI_DEBUG
void iwl_dump_nic_event_log(struct iwl_priv *priv);
+void iwl_dump_nic_error_log(struct iwl_priv *priv);
+#else
+static inline void iwl_dump_nic_event_log(struct iwl_priv *priv)
+{
+}
+
+static inline void iwl_dump_nic_error_log(struct iwl_priv *priv)
+{
+}
+#endif
+
void iwl_clear_isr_stats(struct iwl_priv *priv);
/*****************************************************
diff --git a/drivers/net/wireless/iwlwifi/iwl-debugfs.c b/drivers/net/wireless/iwlwifi/iwl-debugfs.c
index fb84485..a198bcf 100644
--- a/drivers/net/wireless/iwlwifi/iwl-debugfs.c
+++ b/drivers/net/wireless/iwlwifi/iwl-debugfs.c
@@ -410,7 +410,7 @@ static ssize_t iwl_dbgfs_nvm_read(struct file *file,
pos += scnprintf(buf + pos, buf_size - pos, "0x%.4x ", ofs);
hex_dump_to_buffer(ptr + ofs, 16 , 16, 2, buf + pos,
buf_size - pos, 0);
- pos += strlen(buf);
+ pos += strlen(buf + pos);
if (buf_size - pos > 0)
buf[pos++] = '\n';
}
@@ -436,7 +436,7 @@ static ssize_t iwl_dbgfs_log_event_write(struct file *file,
if (sscanf(buf, "%d", &event_log_flag) != 1)
return -EFAULT;
if (event_log_flag == 1)
- iwl_dump_nic_event_log(priv);
+ priv->cfg->ops->lib->dump_nic_event_log(priv);
return count;
}
@@ -909,7 +909,7 @@ static ssize_t iwl_dbgfs_traffic_log_read(struct file *file,
"0x%.4x ", ofs);
hex_dump_to_buffer(ptr + ofs, 16, 16, 2,
buf + pos, bufsz - pos, 0);
- pos += strlen(buf);
+ pos += strlen(buf + pos);
if (bufsz - pos > 0)
buf[pos++] = '\n';
}
@@ -932,7 +932,7 @@ static ssize_t iwl_dbgfs_traffic_log_read(struct file *file,
"0x%.4x ", ofs);
hex_dump_to_buffer(ptr + ofs, 16, 16, 2,
buf + pos, bufsz - pos, 0);
- pos += strlen(buf);
+ pos += strlen(buf + pos);
if (bufsz - pos > 0)
buf[pos++] = '\n';
}
diff --git a/drivers/net/wireless/iwlwifi/iwl-tx.c b/drivers/net/wireless/iwlwifi/iwl-tx.c
index a7422e5..c189075 100644
--- a/drivers/net/wireless/iwlwifi/iwl-tx.c
+++ b/drivers/net/wireless/iwlwifi/iwl-tx.c
@@ -197,6 +197,12 @@ void iwl_cmd_queue_free(struct iwl_priv *priv)
pci_free_consistent(dev, priv->hw_params.tfd_size *
txq->q.n_bd, txq->tfds, txq->q.dma_addr);
+ /* deallocate arrays */
+ kfree(txq->cmd);
+ kfree(txq->meta);
+ txq->cmd = NULL;
+ txq->meta = NULL;
+
/* 0-fill queue descriptor structure */
memset(txq, 0, sizeof(*txq));
}
diff --git a/drivers/net/wireless/iwlwifi/iwl3945-base.c b/drivers/net/wireless/iwlwifi/iwl3945-base.c
index 4f2d439..c390dbd 100644
--- a/drivers/net/wireless/iwlwifi/iwl3945-base.c
+++ b/drivers/net/wireless/iwlwifi/iwl3945-base.c
@@ -1481,6 +1481,7 @@ static inline void iwl_synchronize_irq(struct iwl_priv *priv)
tasklet_kill(&priv->irq_tasklet);
}
+#ifdef CONFIG_IWLWIFI_DEBUG
static const char *desc_lookup(int i)
{
switch (i) {
@@ -1504,7 +1505,7 @@ static const char *desc_lookup(int i)
#define ERROR_START_OFFSET (1 * sizeof(u32))
#define ERROR_ELEM_SIZE (7 * sizeof(u32))
-static void iwl3945_dump_nic_error_log(struct iwl_priv *priv)
+void iwl3945_dump_nic_error_log(struct iwl_priv *priv)
{
u32 i;
u32 desc, time, count, base, data1;
@@ -1598,7 +1599,7 @@ static void iwl3945_print_event_log(struct iwl_priv *priv, u32 start_idx,
}
}
-static void iwl3945_dump_nic_event_log(struct iwl_priv *priv)
+void iwl3945_dump_nic_event_log(struct iwl_priv *priv)
{
u32 base; /* SRAM byte address of event log header */
u32 capacity; /* event log capacity in # entries */
@@ -1640,6 +1641,16 @@ static void iwl3945_dump_nic_event_log(struct iwl_priv *priv)
iwl3945_print_event_log(priv, 0, next_entry, mode);
}
+#else
+void iwl3945_dump_nic_event_log(struct iwl_priv *priv)
+{
+}
+
+void iwl3945_dump_nic_error_log(struct iwl_priv *priv)
+{
+}
+
+#endif
static void iwl3945_irq_tasklet(struct iwl_priv *priv)
{
@@ -3683,21 +3694,6 @@ static ssize_t dump_error_log(struct device *d,
static DEVICE_ATTR(dump_errors, S_IWUSR, NULL, dump_error_log);
-static ssize_t dump_event_log(struct device *d,
- struct device_attribute *attr,
- const char *buf, size_t count)
-{
- struct iwl_priv *priv = dev_get_drvdata(d);
- char *p = (char *)buf;
-
- if (p[0] == '1')
- iwl3945_dump_nic_event_log(priv);
-
- return strnlen(buf, count);
-}
-
-static DEVICE_ATTR(dump_events, S_IWUSR, NULL, dump_event_log);
-
/*****************************************************************************
*
* driver setup and tear down
@@ -3742,7 +3738,6 @@ static struct attribute *iwl3945_sysfs_entries[] = {
&dev_attr_antenna.attr,
&dev_attr_channels.attr,
&dev_attr_dump_errors.attr,
- &dev_attr_dump_events.attr,
&dev_attr_flags.attr,
&dev_attr_filter_flags.attr,
#ifdef CONFIG_IWL3945_SPECTRUM_MEASUREMENT
diff --git a/drivers/platform/x86/sony-laptop.c b/drivers/platform/x86/sony-laptop.c
index f9f68e0..afdbdaa 100644
--- a/drivers/platform/x86/sony-laptop.c
+++ b/drivers/platform/x86/sony-laptop.c
@@ -1041,6 +1041,9 @@ static int sony_nc_resume(struct acpi_device *device)
sony_backlight_update_status(sony_backlight_device) < 0)
printk(KERN_WARNING DRV_PFX "unable to restore brightness level\n");
+ /* re-read rfkill state */
+ sony_nc_rfkill_update();
+
return 0;
}
@@ -1078,6 +1081,8 @@ static int sony_nc_setup_rfkill(struct acpi_device *device,
struct rfkill *rfk;
enum rfkill_type type;
const char *name;
+ int result;
+ bool hwblock;
switch (nc_type) {
case SONY_WIFI:
@@ -1105,6 +1110,10 @@ static int sony_nc_setup_rfkill(struct acpi_device *device,
if (!rfk)
return -ENOMEM;
+ sony_call_snc_handle(0x124, 0x200, &result);
+ hwblock = !(result & 0x1);
+ rfkill_set_hw_state(rfk, hwblock);
+
err = rfkill_register(rfk);
if (err) {
rfkill_destroy(rfk);
diff --git a/include/net/wext.h b/include/net/wext.h
index 6d76a39..3f2b94d 100644
--- a/include/net/wext.h
+++ b/include/net/wext.h
@@ -14,6 +14,7 @@ extern int wext_handle_ioctl(struct net *net, struct ifreq *ifr, unsigned int cm
void __user *arg);
extern int compat_wext_handle_ioctl(struct net *net, unsigned int cmd,
unsigned long arg);
+extern struct iw_statistics *get_wireless_stats(struct net_device *dev);
#else
static inline int wext_proc_init(struct net *net)
{
diff --git a/net/core/net-sysfs.c b/net/core/net-sysfs.c
index 7d4c575..821d309 100644
--- a/net/core/net-sysfs.c
+++ b/net/core/net-sysfs.c
@@ -16,7 +16,7 @@
#include <net/sock.h>
#include <linux/rtnetlink.h>
#include <linux/wireless.h>
-#include <net/iw_handler.h>
+#include <net/wext.h>
#include "net-sysfs.h"
@@ -363,15 +363,13 @@ static ssize_t wireless_show(struct device *d, char *buf,
char *))
{
struct net_device *dev = to_net_dev(d);
- const struct iw_statistics *iw = NULL;
+ const struct iw_statistics *iw;
ssize_t ret = -EINVAL;
read_lock(&dev_base_lock);
if (dev_isalive(dev)) {
- if (dev->wireless_handlers &&
- dev->wireless_handlers->get_wireless_stats)
- iw = dev->wireless_handlers->get_wireless_stats(dev);
- if (iw != NULL)
+ iw = get_wireless_stats(dev);
+ if (iw)
ret = (*format)(iw, buf);
}
read_unlock(&dev_base_lock);
@@ -505,7 +503,7 @@ int netdev_register_kobject(struct net_device *net)
*groups++ = &netstat_group;
#ifdef CONFIG_WIRELESS_EXT_SYSFS
- if (net->wireless_handlers && net->wireless_handlers->get_wireless_stats)
+ if (net->wireless_handlers || net->ieee80211_ptr)
*groups++ = &wireless_group;
#endif
#endif /* CONFIG_SYSFS */
diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c
index 97a278a..8d26e9b 100644
--- a/net/mac80211/mlme.c
+++ b/net/mac80211/mlme.c
@@ -1388,8 +1388,8 @@ ieee80211_rx_mgmt_disassoc(struct ieee80211_sub_if_data *sdata,
reason_code = le16_to_cpu(mgmt->u.disassoc.reason_code);
- printk(KERN_DEBUG "%s: disassociated (Reason: %u)\n",
- sdata->dev->name, reason_code);
+ printk(KERN_DEBUG "%s: disassociated from %pM (Reason: %u)\n",
+ sdata->dev->name, mgmt->sa, reason_code);
ieee80211_set_disassoc(sdata, false);
return RX_MGMT_CFG80211_DISASSOC;
@@ -1675,7 +1675,7 @@ static void ieee80211_rx_mgmt_probe_resp(struct ieee80211_sub_if_data *sdata,
/* direct probe may be part of the association flow */
if (wk && wk->state == IEEE80211_MGD_STATE_PROBE) {
- printk(KERN_DEBUG "%s direct probe responded\n",
+ printk(KERN_DEBUG "%s: direct probe responded\n",
sdata->dev->name);
wk->tries = 0;
wk->state = IEEE80211_MGD_STATE_AUTH;
@@ -2502,9 +2502,6 @@ int ieee80211_mgd_deauth(struct ieee80211_sub_if_data *sdata,
struct ieee80211_mgd_work *wk;
const u8 *bssid = NULL;
- printk(KERN_DEBUG "%s: deauthenticating by local choice (reason=%d)\n",
- sdata->dev->name, req->reason_code);
-
mutex_lock(&ifmgd->mtx);
if (ifmgd->associated && &ifmgd->associated->cbss == req->bss) {
@@ -2532,6 +2529,9 @@ int ieee80211_mgd_deauth(struct ieee80211_sub_if_data *sdata,
mutex_unlock(&ifmgd->mtx);
+ printk(KERN_DEBUG "%s: deauthenticating from %pM by local choice (reason=%d)\n",
+ sdata->dev->name, bssid, req->reason_code);
+
ieee80211_send_deauth_disassoc(sdata, bssid,
IEEE80211_STYPE_DEAUTH, req->reason_code,
cookie);
@@ -2545,9 +2545,6 @@ int ieee80211_mgd_disassoc(struct ieee80211_sub_if_data *sdata,
{
struct ieee80211_if_managed *ifmgd = &sdata->u.mgd;
- printk(KERN_DEBUG "%s: disassociating by local choice (reason=%d)\n",
- sdata->dev->name, req->reason_code);
-
mutex_lock(&ifmgd->mtx);
/*
@@ -2561,6 +2558,9 @@ int ieee80211_mgd_disassoc(struct ieee80211_sub_if_data *sdata,
return -ENOLINK;
}
+ printk(KERN_DEBUG "%s: disassociating from %pM by local choice (reason=%d)\n",
+ sdata->dev->name, req->bss->bssid, req->reason_code);
+
ieee80211_set_disassoc(sdata, false);
mutex_unlock(&ifmgd->mtx);
diff --git a/net/wireless/sme.c b/net/wireless/sme.c
index 7fae7ee..93c3ed3 100644
--- a/net/wireless/sme.c
+++ b/net/wireless/sme.c
@@ -762,9 +762,8 @@ int __cfg80211_connect(struct cfg80211_registered_device *rdev,
wdev->conn->params.ssid = wdev->ssid;
wdev->conn->params.ssid_len = connect->ssid_len;
- /* don't care about result -- but fill bssid & channel */
- if (!wdev->conn->params.bssid || !wdev->conn->params.channel)
- bss = cfg80211_get_conn_bss(wdev);
+ /* see if we have the bss already */
+ bss = cfg80211_get_conn_bss(wdev);
wdev->sme_state = CFG80211_SME_CONNECTING;
wdev->connect_keys = connkeys;
diff --git a/net/wireless/wext-sme.c b/net/wireless/wext-sme.c
index bf72527..5615a88 100644
--- a/net/wireless/wext-sme.c
+++ b/net/wireless/wext-sme.c
@@ -30,7 +30,8 @@ int cfg80211_mgd_wext_connect(struct cfg80211_registered_device *rdev,
if (wdev->wext.keys) {
wdev->wext.keys->def = wdev->wext.default_key;
wdev->wext.keys->defmgmt = wdev->wext.default_mgmt_key;
- wdev->wext.connect.privacy = true;
+ if (wdev->wext.default_key != -1)
+ wdev->wext.connect.privacy = true;
}
if (!wdev->wext.connect.ssid_len)
@@ -229,8 +230,7 @@ int cfg80211_mgd_wext_giwessid(struct net_device *dev,
data->flags = 1;
data->length = wdev->wext.connect.ssid_len;
memcpy(ssid, wdev->wext.connect.ssid, data->length);
- } else
- data->flags = 0;
+ }
wdev_unlock(wdev);
return 0;
@@ -306,8 +306,6 @@ int cfg80211_mgd_wext_giwap(struct net_device *dev,
wdev_lock(wdev);
if (wdev->current_bss)
memcpy(ap_addr->sa_data, wdev->current_bss->pub.bssid, ETH_ALEN);
- else if (wdev->wext.connect.bssid)
- memcpy(ap_addr->sa_data, wdev->wext.connect.bssid, ETH_ALEN);
else
memset(ap_addr->sa_data, 0, ETH_ALEN);
wdev_unlock(wdev);
diff --git a/net/wireless/wext.c b/net/wireless/wext.c
index 5b4a0ce..60fe577 100644
--- a/net/wireless/wext.c
+++ b/net/wireless/wext.c
@@ -470,7 +470,7 @@ static iw_handler get_handler(struct net_device *dev, unsigned int cmd)
/*
* Get statistics out of the driver
*/
-static struct iw_statistics *get_wireless_stats(struct net_device *dev)
+struct iw_statistics *get_wireless_stats(struct net_device *dev)
{
/* New location */
if ((dev->wireless_handlers != NULL) &&
@@ -773,10 +773,13 @@ static int ioctl_standard_iw_point(struct iw_point *iwp, unsigned int cmd,
essid_compat = 1;
else if (IW_IS_SET(cmd) && (iwp->length != 0)) {
char essid[IW_ESSID_MAX_SIZE + 1];
+ unsigned int len;
+ len = iwp->length * descr->token_size;
- err = copy_from_user(essid, iwp->pointer,
- iwp->length *
- descr->token_size);
+ if (len > IW_ESSID_MAX_SIZE)
+ return -EFAULT;
+
+ err = copy_from_user(essid, iwp->pointer, len);
if (err)
return -EFAULT;
--
John W. Linville Someday the world will need a hero, and you
linville@tuxdriver.com might be all we have. Be ready.
^ permalink raw reply related
* Re: tg3 and Broadcom PHY driver
From: David Miller @ 2009-09-28 21:15 UTC (permalink / raw)
To: felix; +Cc: mcarlson, netdev
In-Reply-To: <4AC1223B.8070903@embedded-sol.com>
From: Felix Radensky <felix@embedded-sol.com>
Date: Mon, 28 Sep 2009 22:53:15 +0200
> Hi, Matt
>
> Matt Carlson wrote:
>> On Sat, Sep 26, 2009 at 02:32:18PM -0700, Felix Radensky wrote:
>>
>> Is the broadcom module also compiled into the kernel?
>>
>>
> Yes.
I bet this is all because the tg3 driver is linked into the kernel
before the PHY layer and drivers subdirectory or something like that.
Link order determines the order in which built-in initializations
occur (within the same init type).
^ permalink raw reply
* Re: [2.6.31-git17] WARNING: at kernel/hrtimer.c:648 hres_timers_resume+0x40/0x50()/WARNING: at drivers/base/sys.c:353 __sysdev_resume+0xc3/0xe0()
From: Rafael J. Wysocki @ 2009-09-28 21:13 UTC (permalink / raw)
To: Maciej Rutecki
Cc: Yong Zhang, Linux Kernel Mailing List, clemens,
venkatesh.pallipadi, gregkh, zambrano, davem, netdev
In-Reply-To: <8db1092f0909281308s36c35c80s65c18dcbcf9fff2b@mail.gmail.com>
On Monday 28 September 2009, Maciej Rutecki wrote:
> 2009/9/28 Maciej Rutecki <maciej.rutecki@gmail.com>:
> >
> > Add patch and remove previous:
> > http://unixy.pl/maciek/download/kernel/2.6.31-git17/gumis/dmesg-debug.txt
> >
> > s2disk&resume twice.
> >
> > no "timekeeping_resume() called with IRQs enabled!".
> >
> > I found some interesting thing, warnings appear only once, during
> > first s2disk, on second don't appear.
>
> Already I test 2.6-32-rc1 few times; warnings has gone. Any patches
> has been add since 2.6.31-git17?
Quite some of them, actually.
Thanks,
Rafael
^ permalink raw reply
* Re: tg3 and Broadcom PHY driver
From: Felix Radensky @ 2009-09-28 20:53 UTC (permalink / raw)
To: Matt Carlson; +Cc: netdev@vger.kernel.org
In-Reply-To: <20090928205226.GB12652@xw6200.broadcom.net>
Hi, Matt
Matt Carlson wrote:
> On Sat, Sep 26, 2009 at 02:32:18PM -0700, Felix Radensky wrote:
>
>> Hi,
>>
>> I've noticed that in linux-2.6.31 I have to make tg3 driver modular, due to
>> its dependency on Broadcom PHY driver. If both tg3 and PHY driver are
>> compiled into the kernel, tg3 fails to detect a PHY, apparently because PHY
>> driver is loaded later. I'm using BCM57760 on embedded powerpc platform
>> (MPC8536).
>>
>> How can I make tg3 work when it's compiled into the kernel ?
>>
>> Thanks.
>>
>
> Is the broadcom module also compiled into the kernel?
>
>
Yes.
^ permalink raw reply
* Re: tg3 and Broadcom PHY driver
From: Matt Carlson @ 2009-09-28 20:52 UTC (permalink / raw)
To: Felix Radensky; +Cc: netdev@vger.kernel.org
In-Reply-To: <4ABE8862.3060308@embedded-sol.com>
On Sat, Sep 26, 2009 at 02:32:18PM -0700, Felix Radensky wrote:
> Hi,
>
> I've noticed that in linux-2.6.31 I have to make tg3 driver modular, due to
> its dependency on Broadcom PHY driver. If both tg3 and PHY driver are
> compiled into the kernel, tg3 fails to detect a PHY, apparently because PHY
> driver is loaded later. I'm using BCM57760 on embedded powerpc platform
> (MPC8536).
>
> How can I make tg3 work when it's compiled into the kernel ?
>
> Thanks.
Is the broadcom module also compiled into the kernel?
^ permalink raw reply
* Re: tg3: Badness at kernel/mutex.c:207
From: Matt Carlson @ 2009-09-28 20:51 UTC (permalink / raw)
To: Felix Radensky; +Cc: netdev@vger.kernel.org
In-Reply-To: <4ABE85B9.3020304@embedded-sol.com>
On Sat, Sep 26, 2009 at 02:20:57PM -0700, Felix Radensky wrote:
> Hi,
>
> I'm running linux-2.6.31 on a custom MPC8536 based board with BCM57760 chip.
> Both tg3 driver, and Broadcom PHY driver are modules.
>
> Each time I run ifconfig eth2 up, I get the following error message:
>
> Badness at kernel/mutex.c:207
> NIP: c025132c LR: c0251314 CTR: c0251334
> REGS: efbedbd0 TRAP: 0700 Not tainted (2.6.31)
> MSR: 00029000 <EE,ME,CE> CR: 24020422 XER: 00000000
> TASK = efacce10[1080] 'ifconfig' THREAD: efbec000
> GPR00: 00000000 efbedc80 efacce10 00000001 00007020 00000002 00000000
> 00000200
> GPR08: 00029000 c0350000 c0330000 00000001 24020424 10057d94 000002a0
> 1000d82c
> GPR16: 1000d81c 1000d814 10010000 10050000 ef897a0c efbede18 ffff8914
> ef897a00
> GPR24: 00008000 c034b480 efbec000 efb0122c c0350000 efacce10 ef82d2c0
> efb01228
> NIP [c025132c] __mutex_lock_slowpath+0x1f0/0x1f8
> LR [c0251314] __mutex_lock_slowpath+0x1d8/0x1f8
> Call Trace:
> [efbedcd0] [c025134c] mutex_lock+0x18/0x34
> [efbedcf0] [f534a228] tg3_chip_reset+0x7cc/0x9f8 [tg3]
> [efbedd20] [f534a8f0] tg3_reset_hw+0x58/0x2360 [tg3]
> [efbedd70] [f5351dd4] tg3_open+0x610/0x910 [tg3]
> [efbeddb0] [c01e1c6c] dev_open+0x100/0x138
> [efbeddd0] [c01dff20] dev_change_flags+0x80/0x1ac
> [efbeddf0] [c02232cc] devinet_ioctl+0x648/0x824
> [efbede60] [c0223de4] inet_ioctl+0xcc/0xf8
> [efbede70] [c01cdf44] sock_ioctl+0x60/0x300
> [efbede90] [c008a35c] vfs_ioctl+0x34/0x8c
> [efbedea0] [c008a580] do_vfs_ioctl+0x88/0x724
> [efbedf10] [c008ac5c] sys_ioctl+0x40/0x74
> [efbedf40] [c000f814] ret_from_syscall+0x0/0x3c
> Instruction dump:
> 0fe00000 4bfffe80 801a000c 5409016f 4182fe60 4bf0f6d9 2f830000 41befe54
> 3d20c035 8009c2c0 2f800000 40befe44 <0fe00000> 4bfffe3c 9421ffe0 7c0802a6
>
> Does it indicate a real problem, or something that can be ignored ?
>
> Additional information from kernel log:
>
> tg3.c:v3.99 (April 20, 2009)
> tg3 0002:05:00.0: enabling bus mastering
> tg3 0002:05:00.0: PME# disabled
> tg3 mdio bus: probed
> eth2: Tigon3 [partno(BCM57760) rev 57780001] (PCI Express) MAC address
> 00:10:18:00:00:00
> eth2: attached PHY driver [Broadcom BCM57780] (mii_bus:phy_addr=500:01)
> eth2: RXcsums[1] LinkChgREG[0] MIirq[0] ASF[0] TSOcap[1]
> eth2: dma_rwctrl[76180000] dma_mask[64-bit]
> tg3 0002:05:00.0: PME# disabled
Yes, this is a real problem. The driver is taking the MDIO bus lock
while holding the device's own spinlock. I think I may have a
workaround. Let me test it and get back to you.
^ permalink raw reply
* Re: 2.6.31 regression: e1000e jumbo frames no longer work: 'Unsupported MTU setting'
From: Nix @ 2009-09-28 20:51 UTC (permalink / raw)
To: Alexander Duyck; +Cc: e1000-devel, netdev, bruce.w.allan, linux-kernel
In-Reply-To: <5f2db9d90909261914l3a927f4cya1dc5e4548688bce@mail.gmail.com>
On 27 Sep 2009, Alexander Duyck said:
> It looks like the problem is that the 82574 and 82583 seem to have
> their max_hw_frame_size values swapped. You might try applying the
> patch below. I am not sure if it will apply since I hand generated it
Applies fine: works fine. Thank you!
> using the git patch that seems to have introduced the problem, and I
> am sending the patch through an untested account that may mangle the
> patch.
Unmangled.
> I will see about submitting an official patch for this
> sometime next few days.
I wonder if this belongs in -stable? People with an 82574 who are using
jumbo frames may well find their networks not coming up, if like me
they set a bunch of properties in one /sbin/ip call.
------------------------------------------------------------------------------
Come build with us! The BlackBerry® Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay
ahead of the curve. Join us from November 9-12, 2009. Register now!
http://p.sf.net/sfu/devconf
^ permalink raw reply
* Re: [2.6.31-git17] WARNING: at kernel/hrtimer.c:648 hres_timers_resume+0x40/0x50()/WARNING: at drivers/base/sys.c:353 __sysdev_resume+0xc3/0xe0()
From: Maciej Rutecki @ 2009-09-28 20:08 UTC (permalink / raw)
To: Yong Zhang
Cc: Linux Kernel Mailing List, Rafael J. Wysocki, clemens,
venkatesh.pallipadi, gregkh, zambrano, davem, netdev
In-Reply-To: <8db1092f0909281138t18a379d1qdf999b0610ed6414@mail.gmail.com>
2009/9/28 Maciej Rutecki <maciej.rutecki@gmail.com>:
>
> Add patch and remove previous:
> http://unixy.pl/maciek/download/kernel/2.6.31-git17/gumis/dmesg-debug.txt
>
> s2disk&resume twice.
>
> no "timekeeping_resume() called with IRQs enabled!".
>
> I found some interesting thing, warnings appear only once, during
> first s2disk, on second don't appear.
Already I test 2.6-32-rc1 few times; warnings has gone. Any patches
has been add since 2.6.31-git17?
Regards
--
Maciej Rutecki
http://www.maciek.unixy.pl
^ permalink raw reply
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