* Re: [patch] openvswitch: checking wrong variable in queue_userspace_packet()
From: David Miller @ 2012-05-13 19:47 UTC (permalink / raw)
To: jesse-l0M0P4e3n4LQT0dZR+AlfA
Cc: dev-yBygre7rU0TnMu66kgdUjQ, netdev-u79uwXL29TY76Z2rM5mHXA,
kernel-janitors-u79uwXL29TY76Z2rM5mHXA,
dan.carpenter-QHcLZuEGTsvQT0dZR+AlfA
In-Reply-To: <CAEP_g=_LojOio0E5K_vYBkO3BKdW9F=dDam4zwffyZPbSZJd2Q-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
From: Jesse Gross <jesse-l0M0P4e3n4LQT0dZR+AlfA@public.gmane.org>
Date: Sun, 13 May 2012 12:22:29 -0700
> On Sun, May 13, 2012 at 11:44 AM, Dan Carpenter
> <dan.carpenter-QHcLZuEGTsvQT0dZR+AlfA@public.gmane.org> wrote:
>> "skb" is non-NULL here, for example we dereference it in skb_clone().
>> The intent was to test "nskb" which was just set.
>>
>> Signed-off-by: Dan Carpenter <dan.carpenter-QHcLZuEGTsvQT0dZR+AlfA@public.gmane.org>
>
> Thanks Dan.
>
> Acked-by: Jesse Gross <jesse-l0M0P4e3n4LQT0dZR+AlfA@public.gmane.org>
Applied.
^ permalink raw reply
* Re: [patch] openvswitch: checking wrong variable in queue_userspace_packet()
From: Jesse Gross @ 2012-05-13 19:22 UTC (permalink / raw)
To: Dan Carpenter
Cc: dev-yBygre7rU0TnMu66kgdUjQ, netdev-u79uwXL29TY76Z2rM5mHXA,
kernel-janitors-u79uwXL29TY76Z2rM5mHXA, David S. Miller
In-Reply-To: <20120513184418.GB16541-mgFCXtclrQlZLf2FXnZxJA@public.gmane.org>
On Sun, May 13, 2012 at 11:44 AM, Dan Carpenter
<dan.carpenter-QHcLZuEGTsvQT0dZR+AlfA@public.gmane.org> wrote:
> "skb" is non-NULL here, for example we dereference it in skb_clone().
> The intent was to test "nskb" which was just set.
>
> Signed-off-by: Dan Carpenter <dan.carpenter-QHcLZuEGTsvQT0dZR+AlfA@public.gmane.org>
Thanks Dan.
Acked-by: Jesse Gross <jesse-l0M0P4e3n4LQT0dZR+AlfA@public.gmane.org>
^ permalink raw reply
* [patch] openvswitch: checking wrong variable in queue_userspace_packet()
From: Dan Carpenter @ 2012-05-13 18:44 UTC (permalink / raw)
To: Jesse Gross; +Cc: David S. Miller, dev, netdev, kernel-janitors
"skb" is non-NULL here, for example we dereference it in skb_clone().
The intent was to test "nskb" which was just set.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
diff --git a/net/openvswitch/datapath.c b/net/openvswitch/datapath.c
index fe28562..2c74daa 100644
--- a/net/openvswitch/datapath.c
+++ b/net/openvswitch/datapath.c
@@ -321,7 +321,7 @@ static int queue_userspace_packet(int dp_ifindex, struct sk_buff *skb,
return -ENOMEM;
nskb = __vlan_put_tag(nskb, vlan_tx_tag_get(nskb));
- if (!skb)
+ if (!nskb)
return -ENOMEM;
nskb->vlan_tci = 0;
^ permalink raw reply related
* [PATCH resend] sunrpc: fix kernel-doc warnings
From: Randy Dunlap @ 2012-05-13 17:35 UTC (permalink / raw)
To: linux-nfs-u79uwXL29TY76Z2rM5mHXA
Cc: Trond Myklebust, J. Bruce Fields, Andrew Morton,
netdev-u79uwXL29TY76Z2rM5mHXA
From: Randy Dunlap <rdunlap-/UHa2rfvQTnk1uMJSBkQmQ@public.gmane.org>
Fix kernel-doc warnings in sunrpc/rpc_pipe.c and
sunrpc/rpcb_clnt.c:
Warning(net/sunrpc/rpcb_clnt.c:428): No description found for parameter 'net'
Warning(net/sunrpc/rpcb_clnt.c:567): No description found for parameter 'net'
Warning(net/sunrpc/rpc_pipe.c:133): No description found for parameter 'pipe'
Warning(net/sunrpc/rpc_pipe.c:133): Excess function parameter 'inode' description in 'rpc_queue_upcall'
Warning(net/sunrpc/rpc_pipe.c:839): No description found for parameter 'pipe'
Warning(net/sunrpc/rpc_pipe.c:839): Excess function parameter 'ops' description in 'rpc_mkpipe_dentry'
Warning(net/sunrpc/rpc_pipe.c:839): Excess function parameter 'flags' description in 'rpc_mkpipe_dentry'
Warning(net/sunrpc/rpc_pipe.c:949): No description found for parameter 'dentry'
Warning(net/sunrpc/rpc_pipe.c:949): Excess function parameter 'clnt' description in 'rpc_remove_client_dir'
Signed-off-by: Randy Dunlap <rdunlap-/UHa2rfvQTnk1uMJSBkQmQ@public.gmane.org>
---
net/sunrpc/rpc_pipe.c | 8 +++-----
1 file changed, 3 insertions(+), 5 deletions(-)
--- lnx-34-rc3.orig/net/sunrpc/rpcb_clnt.c
+++ lnx-34-rc3/net/sunrpc/rpcb_clnt.c
@@ -394,6 +394,7 @@ static int rpcb_register_call(struct rpc
/**
* rpcb_register - set or unset a port registration with the local rpcbind svc
+ * @net: target network namespace
* @prog: RPC program number to bind
* @vers: RPC version number to bind
* @prot: transport protocol to register
@@ -521,6 +522,7 @@ static int rpcb_unregister_all_protofami
/**
* rpcb_v4_register - set or unset a port registration with the local rpcbind
+ * @net: target network namespace
* @program: RPC program number of service to (un)register
* @version: RPC version number of service to (un)register
* @address: address family, IP address, and port to (un)register
--- lnx-34-rc3.orig/net/sunrpc/rpc_pipe.c
+++ lnx-34-rc3/net/sunrpc/rpc_pipe.c
@@ -120,7 +120,7 @@ EXPORT_SYMBOL_GPL(rpc_pipe_generic_upcal
/**
* rpc_queue_upcall - queue an upcall message to userspace
- * @inode: inode of upcall pipe on which to queue given message
+ * @pipe: upcall pipe on which to queue given message
* @msg: message to queue
*
* Call with an @inode created by rpc_mkpipe() to queue an upcall.
@@ -819,9 +819,7 @@ static int rpc_rmdir_depopulate(struct d
* @parent: dentry of directory to create new "pipe" in
* @name: name of pipe
* @private: private data to associate with the pipe, for the caller's use
- * @ops: operations defining the behavior of the pipe: upcall, downcall,
- * release_pipe, open_pipe, and destroy_msg.
- * @flags: rpc_pipe flags
+ * @pipe: &rpc_pipe containing input parameters
*
* Data is made available for userspace to read by calls to
* rpc_queue_upcall(). The actual reads will result in calls to
@@ -943,7 +941,7 @@ struct dentry *rpc_create_client_dir(str
/**
* rpc_remove_client_dir - Remove a directory created with rpc_create_client_dir()
- * @clnt: rpc client
+ * @dentry: dentry for the pipe
*/
int rpc_remove_client_dir(struct dentry *dentry)
{
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply
* [PATCH RFC] tun: experimental zero copy tx support
From: Michael S. Tsirkin @ 2012-05-13 15:52 UTC (permalink / raw)
To: David S. Miller
Cc: Stephen Hemminger, Joe Perches, Jason Wang, netdev, linux-kernel,
Ian.Campbell, kvm
Let vhost-net utilize zero copy tx when the experimental
zero copy mode is enabled and when used with tun. This works on
top of the patchset 'copy aside frags with destructors' that I posted
previously. This is not using tcp so doesn't have the
issue with early skb cloning noticed by Ian.
For those that wish to test this with kvm, I intend to post a patchset +
git tree with just the necessary bits from the destructor patch
a bit later.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
drivers/net/tun.c | 125 ++++++++++++++++++++++++++++++++++++++++++++++++++--
1 files changed, 120 insertions(+), 5 deletions(-)
diff --git a/drivers/net/tun.c b/drivers/net/tun.c
index fe5cd2f3..eb10ee7 100644
--- a/drivers/net/tun.c
+++ b/drivers/net/tun.c
@@ -100,6 +100,8 @@ do { \
} while (0)
#endif
+#define GOODCOPY_LEN 128
+
#define FLT_EXACT_COUNT 8
struct tap_filter {
unsigned int count; /* Number of addrs. Zero means disabled */
@@ -602,8 +604,80 @@ static struct sk_buff *tun_alloc_skb(struct tun_struct *tun,
return skb;
}
+/* set skb frags from iovec, this can move to core network code for reuse */
+static int zerocopy_sg_from_iovec(struct sk_buff *skb, const struct iovec *from,
+ int offset, size_t count)
+{
+ int len = iov_length(from, count) - offset;
+ int copy = skb_headlen(skb);
+ int size, offset1 = 0;
+ int i = 0;
+
+ /* Skip over from offset */
+ while (count && (offset >= from->iov_len)) {
+ offset -= from->iov_len;
+ ++from;
+ --count;
+ }
+
+ /* copy up to skb headlen */
+ while (count && (copy > 0)) {
+ size = min_t(unsigned int, copy, from->iov_len - offset);
+ if (copy_from_user(skb->data + offset1, from->iov_base + offset,
+ size))
+ return -EFAULT;
+ if (copy > size) {
+ ++from;
+ --count;
+ }
+ copy -= size;
+ offset1 += size;
+ offset = 0;
+ }
+
+ if (len == offset1)
+ return 0;
+
+ while (count--) {
+ struct page *page[MAX_SKB_FRAGS];
+ int num_pages;
+ unsigned long base;
+
+ len = from->iov_len - offset1;
+ if (!len) {
+ offset1 = 0;
+ ++from;
+ continue;
+ }
+ base = (unsigned long)from->iov_base + offset1;
+ size = ((base & ~PAGE_MASK) + len + ~PAGE_MASK) >> PAGE_SHIFT;
+ num_pages = get_user_pages_fast(base, size, 0, &page[i]);
+ if ((num_pages != size) ||
+ (num_pages > MAX_SKB_FRAGS - skb_shinfo(skb)->nr_frags))
+ /* put_page is in skb free */
+ return -EFAULT;
+ skb->data_len += len;
+ skb->len += len;
+ skb->truesize += len;
+ atomic_add(len, &skb->sk->sk_wmem_alloc);
+ while (len) {
+ int off = base & ~PAGE_MASK;
+ int size = min_t(int, len, PAGE_SIZE - off);
+ __skb_fill_page_desc(skb, i, page[i], off, size);
+ skb_shinfo(skb)->nr_frags++;
+ /* increase sk_wmem_alloc */
+ base += size;
+ len -= size;
+ i++;
+ }
+ offset1 = 0;
+ ++from;
+ }
+ return 0;
+}
+
/* Get packet from user space buffer */
-static ssize_t tun_get_user(struct tun_struct *tun,
+static ssize_t tun_get_user(struct tun_struct *tun, void *msg_control,
const struct iovec *iv, size_t count,
int noblock)
{
@@ -612,6 +686,9 @@ static ssize_t tun_get_user(struct tun_struct *tun,
size_t len = count, align = NET_SKB_PAD;
struct virtio_net_hdr gso = { 0 };
int offset = 0;
+ int copylen;
+ bool zerocopy = false;
+ int err;
if (!(tun->flags & TUN_NO_PI)) {
if ((len -= sizeof(pi)) > count)
@@ -645,14 +722,47 @@ static ssize_t tun_get_user(struct tun_struct *tun,
return -EINVAL;
}
- skb = tun_alloc_skb(tun, align, len, gso.hdr_len, noblock);
+ if (msg_control)
+ zerocopy = true;
+
+ if (zerocopy) {
+ /* Userspace may produce vectors with count greater than
+ * MAX_SKB_FRAGS, so we need to linearize parts of the skb
+ * to let the rest of data to be fit in the frags.
+ */
+ if (count > MAX_SKB_FRAGS) {
+ copylen = iov_length(iv, count - MAX_SKB_FRAGS);
+ if (copylen < offset)
+ copylen = 0;
+ else
+ copylen -= offset;
+ } else
+ copylen = 0;
+ /* There are 256 bytes to be copied in skb, so there is enough
+ * room for skb expand head in case it is used.
+ * The rest of the buffer is mapped from userspace.
+ */
+ if (copylen < gso.hdr_len)
+ copylen = gso.hdr_len;
+ if (!copylen)
+ copylen = GOODCOPY_LEN;
+ } else
+ copylen = len;
+
+ skb = tun_alloc_skb(tun, align, copylen, gso.hdr_len, noblock);
if (IS_ERR(skb)) {
if (PTR_ERR(skb) != -EAGAIN)
tun->dev->stats.rx_dropped++;
return PTR_ERR(skb);
}
- if (skb_copy_datagram_from_iovec(skb, 0, iv, offset, len)) {
+ if (zerocopy) {
+ err = zerocopy_sg_from_iovec(skb, iv, offset, count);
+ skb_shinfo(skb)->tx_flags |= SKBTX_DEV_ZEROCOPY;
+ } else
+ err = skb_copy_datagram_from_iovec(skb, 0, iv, offset, len);
+
+ if (err) {
tun->dev->stats.rx_dropped++;
kfree_skb(skb);
return -EFAULT;
@@ -726,6 +836,10 @@ static ssize_t tun_get_user(struct tun_struct *tun,
skb_shinfo(skb)->gso_segs = 0;
}
+ /* copy skb_ubuf_info for callback when skb has no error */
+ if (zerocopy)
+ skb_shinfo(skb)->destructor_arg = msg_control;
+
netif_rx_ni(skb);
tun->dev->stats.rx_packets++;
@@ -746,7 +860,7 @@ static ssize_t tun_chr_aio_write(struct kiocb *iocb, const struct iovec *iv,
tun_debug(KERN_INFO, tun, "tun_chr_write %ld\n", count);
- result = tun_get_user(tun, iv, iov_length(iv, count),
+ result = tun_get_user(tun, NULL, iv, iov_length(iv, count),
file->f_flags & O_NONBLOCK);
tun_put(tun);
@@ -960,7 +1074,7 @@ static int tun_sendmsg(struct kiocb *iocb, struct socket *sock,
struct msghdr *m, size_t total_len)
{
struct tun_struct *tun = container_of(sock, struct tun_struct, socket);
- return tun_get_user(tun, m->msg_iov, total_len,
+ return tun_get_user(tun, m->msg_control, m->msg_iov, total_len,
m->msg_flags & MSG_DONTWAIT);
}
@@ -1130,6 +1244,7 @@ static int tun_set_iff(struct net *net, struct file *file, struct ifreq *ifr)
sock_init_data(&tun->socket, sk);
sk->sk_write_space = tun_sock_write_space;
sk->sk_sndbuf = INT_MAX;
+ sock_set_flag(sk, SOCK_ZEROCOPY);
tun_sk(sk)->tun = tun;
--
MST
^ permalink raw reply related
* Re: [patch] Re: qlge driver corrupting kernel memory
From: Mike Galbraith @ 2012-05-13 10:16 UTC (permalink / raw)
To: Thadeu Lima de Souza Cascardo; +Cc: netdev
In-Reply-To: <1336903839.7390.13.camel@marge.simpson.net>
Erm, with a quilt refresh you get the compiling version :)
glge: Fix double pci_free_consistent() upon tx_ring->q allocation failure
Let ql_free_tx_resources() do it's job. You are not helping.
Signed-off-by: Mike Galbraith <mgalbraith@suse.de>
---
drivers/net/qlge/qlge_main.c | 10 +++-------
1 file changed, 3 insertions(+), 7 deletions(-)
--- a/drivers/net/qlge/qlge_main.c
+++ b/drivers/net/qlge/qlge_main.c
@@ -2664,11 +2664,8 @@ static int ql_alloc_tx_resources(struct
pci_alloc_consistent(qdev->pdev, tx_ring->wq_size,
&tx_ring->wq_base_dma);
- if ((tx_ring->wq_base == NULL) ||
- tx_ring->wq_base_dma & WQ_ADDR_ALIGN) {
- QPRINTK(qdev, IFUP, ERR, "tx_ring alloc failed.\n");
- return -ENOMEM;
- }
+ if ((tx_ring->wq_base == NULL) || tx_ring->wq_base_dma & WQ_ADDR_ALIGN)
+ goto err;
tx_ring->q =
kmalloc(tx_ring->wq_len * sizeof(struct tx_ring_desc), GFP_KERNEL);
if (tx_ring->q == NULL)
@@ -2676,8 +2673,7 @@ static int ql_alloc_tx_resources(struct
return 0;
err:
- pci_free_consistent(qdev->pdev, tx_ring->wq_size,
- tx_ring->wq_base, tx_ring->wq_base_dma);
+ QPRINTK(qdev, IFUP, ERR, "tx_ring alloc failed.\n");
return -ENOMEM;
}
^ permalink raw reply
* [patch] Re: qlge driver corrupting kernel memory
From: Mike Galbraith @ 2012-05-13 10:10 UTC (permalink / raw)
To: Thadeu Lima de Souza Cascardo; +Cc: netdev
In-Reply-To: <1336736301.7361.144.camel@marge.simpson.net>
On Fri, 2012-05-11 at 13:38 +0200, Mike Galbraith wrote:
> On Tue, 2012-05-08 at 09:07 -0300, Thadeu Lima de Souza Cascardo wrote:
> > On Tue, May 08, 2012 at 01:00:18PM +0200, Mike Galbraith wrote:
> > > Greetings network wizards,
> > >
> > > $subject is happening in an 2.6.32 enterprise kernel with the driver
> > > updated to what looks to me to be 2.6.38 or so.
> > >
> > > Allegedly, IFF boxen are running dual CNAs with storage and LAN sharing
> > > a port, $subject happens fairly regularly. Rummaging in crashdumps
> > > seems to show corruption happens because we somehow end up stuffing
> > > loads of frags into skb_shared_info, scribbling all over the place.
> > >
> > > Before I proceed, what I know about skbs can be found here..
> > >
> > > http://vger.kernel.org/~davem/skb_data.html
> > >
> > > ..and that's the sum and total ;-)
> > >
> > > I guess the first thing I should ask is whether anyone has seen such
> > > scribbling with this driver. Known issue would be a case of happiness,
> > > but I doubt that will be the case from searching, so onward.
> > >
> >
> > Hi, Mike.
> >
> > From what you describe, I suspect this is related to this fix:
> >
> > http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=782428535e0819b5b7c9825cd3faa2ad37032a70
> >
> > Please, apply and report if that works for you.
>
> Nope, box exploded. I haven't seen a dump yet, but expect it'll be more
> of the same scribbling.
Something else popped up meanwhile. Shortly after tx_ring->q order 5
allocation failure and ql_release_adapter_resources(), BUG: Bad page
state has now arrived twice to muddy the water.
[ 3537.150327] Node 0 DMA: 2*4kB 2*8kB 1*16kB 2*32kB 2*64kB 1*128kB 0*256kB 0*512kB 0*1024kB 0*2048kB 0*4096kB = 360kB
[ 3537.150345] Node 0 DMA32: 318*4kB 144*8kB 89*16kB 17*32kB 3*64kB 1*128kB 0*256kB 0*512kB 0*1024kB 0*2048kB 0*4096kB = 4712kB
[ 3537.150364] 5248 total pagecache pages
[ 3537.150367] 211 pages in swap cache
[ 3537.150372] Swap cache stats: add 1437, delete 1226, find 1641/1752
[ 3537.150377] Free swap = 67109880kB
[ 3537.150381] Total swap = 67111528kB
[ 3537.152314] 73723 pages RAM
[ 3537.152319] 13128 pages reserved
[ 3537.152322] 4910 pages shared
[ 3537.152326] 22795 pages non-shared
[ 3537.152333] qlge 0000:04:00.0: ql_alloc_mem_resources: TX resource allocation failed.
[ 3537.152343] qlge 0000:04:00.0: ql_get_adapter_resources: Unable to allocate memory.
[ 3537.152499] qlge 0000:04:00.0: ql_set_mac_addr_reg: Adding UNICAST address 00:c0:dd:1a:46:ac at index 0 in the CAM.
[ 3537.440237] BUG: Bad page state in process ifdown-dhcp pfn:10940
[ 3537.440244] page:ffffea00003a0600 flags:0020000000000000 count:-1 mapcount:0 mapping:(null) index:0
[ 3537.440249] Pid: 4317, comm: ifdown-dhcp Tainted: G X 2.6.32.54-0.3.1.4242.0.TEST-default #1
[ 3537.440253] Call Trace:
[ 3537.440265] [<ffffffff810061dc>] dump_trace+0x6c/0x2d0
[ 3537.440271] [<ffffffff8139b366>] dump_stack+0x69/0x73
[ 3537.440279] [<ffffffff810badb3>] bad_page+0xe3/0x170
[ 3537.440284] [<ffffffff810bbedb>] prep_new_page+0xab/0x1b0
[ 3537.440289] [<ffffffff810bc2e4>] get_page_from_freelist+0x304/0x720
[ 3537.440295] [<ffffffff810bc9ba>] __alloc_pages_slowpath+0x11a/0x5f0
[ 3537.440300] [<ffffffff810bcfca>] __alloc_pages_nodemask+0x13a/0x140
[ 3537.440305] [<ffffffff810bbdd9>] __get_free_pages+0x9/0x50
[ 3537.440314] [<ffffffff8104ba62>] dup_task_struct+0x42/0x150
[ 3537.440320] [<ffffffff8104cc54>] copy_process+0xb4/0xe50
[ 3537.440324] [<ffffffff8104da7c>] do_fork+0x8c/0x3c0
[ 3537.440331] [<ffffffff81003263>] stub_clone+0x13/0x20
[ 3537.441094] DWARF2 unwinder stuck at stub_clone+0x13/0x20
[ 3537.441097]
[ 3537.441098] Leftover inexact backtrace:
[ 3537.441099]
[ 3537.441103] [<ffffffff81002f7b>] ? system_call_fastpath+0x16/0x1b
[ 3537.441107] Disabling lock debugging due to kernel taint
[ 3537.899545] bonding: bond0 is being deleted..
glge: Fix double pci_free_consistent() upon tx_ring->q allocation failure
Let ql_free_tx_resources() do it's job. You are not helping.
Signed-off-by: Mike Galbraith <mgalbraith@suse.de>
---
drivers/net/qlge/qlge_main.c | 10 +++-------
1 file changed, 3 insertions(+), 7 deletions(-)
--- a/drivers/net/qlge/qlge_main.c
+++ b/drivers/net/qlge/qlge_main.c
@@ -2664,11 +2664,8 @@ static int ql_alloc_tx_resources(struct
pci_alloc_consistent(qdev->pdev, tx_ring->wq_size,
&tx_ring->wq_base_dma);
- if ((tx_ring->wq_base == NULL) ||
- tx_ring->wq_base_dma & WQ_ADDR_ALIGN) {
- QPRINTK(qdev, IFUP, ERR, "tx_ring alloc failed.\n");
- return -ENOMEM;
- }
+ if ((tx_ring->wq_base == NULL) tx_ring->wq_base_dma & WQ_ADDR_ALIGN)
+ goto err;
tx_ring->q =
kmalloc(tx_ring->wq_len * sizeof(struct tx_ring_desc), GFP_KERNEL);
if (tx_ring->q == NULL)
@@ -2676,8 +2673,7 @@ static int ql_alloc_tx_resources(struct
return 0;
err:
- pci_free_consistent(qdev->pdev, tx_ring->wq_size,
- tx_ring->wq_base, tx_ring->wq_base_dma);
+ QPRINTK(qdev, IFUP, ERR, "tx_ring alloc failed.\n");
return -ENOMEM;
}
^ permalink raw reply
* Re: [PATCH RFC 1/6] skbuff: support per-page destructors in copy_ubufs
From: Michael S. Tsirkin @ 2012-05-13 10:10 UTC (permalink / raw)
To: Ian Campbell; +Cc: David Miller, netdev@vger.kernel.org, eric.dumazet@gmail.com
In-Reply-To: <1336802484.3891.24.camel@dagon.hellion.org.uk>
On Sat, May 12, 2012 at 07:01:24AM +0100, Ian Campbell wrote:
> On Fri, 2012-05-11 at 17:30 +0100, Michael S. Tsirkin wrote:
> > On Fri, May 11, 2012 at 03:08:36PM +0300, Michael S. Tsirkin wrote:
> > > On Fri, May 11, 2012 at 11:58:12AM +0100, Ian Campbell wrote:
> > > > On Fri, 2012-05-11 at 10:00 +0100, Ian Campbell wrote:
> > > > > I'm seeing copy_ubufs called in my remote NFS test, which I don't
> > > > > think I expected -- I'll investigate why this is happening today.
> > > >
> > > > It's tcp_transmit_skb which can (conditionally) call skb_clone
> > > > (backtrace below)
> > >
> > > Interesting. I didn't realise we clone skbs on data path:
> > > tcp_write_xmit calls tcp_transmit_skb with clone_it flag.
> > > Could someone comment on why we need to clone on good path
> > > like this?
> >
> > Hmm, it's in case we need to retransmit it later.
>
> I wonder if we could avoid the copy_ubuf in this particular clone path
> and have any subsequent calls to copy_ubufs use skb->fclone to determine
> if it can safely replace the frags?
>
> If it cannot then could it do a full copy of the skb (including new
> shinfo, new frag pages etc) as a fallback?
>
> Ian.
>
Yes I think we should call a variant of clone that avoids copy_ubuf on
the first transmit. But need to be careful we don't access the frag
list while it is being modified.
For example very roughly, maybe we could have copy_ubuf detect
packet clone is queued and take some lock?
On retransmit we could check and if we are not the only clone left
(which should be uncommon) trigger copy ubuf then.
Thoughts?
--
MST
^ permalink raw reply
* Re: [PATCH net-next] codel: use Newton method instead of sqrt() and divides
From: Eric Dumazet @ 2012-05-13 7:23 UTC (permalink / raw)
To: David Miller
Cc: dave.taht, netdev, nichols, van, codel, ycheng, mattmathis,
therbert, shemminger, nanditad
In-Reply-To: <20120512.175217.1632102067268101115.davem@davemloft.net>
From: Eric Dumazet <edumazet@google.com>
On Sat, 2012-05-12 at 17:52 -0400, David Miller wrote:
> Ok, fair enough.
Oh well, I sent my mail too late. The error made no sense after a good
night. Also, when Van says something, you can be fairly sure its right,
and if it's not, then you didn't understand what Van said ;)
16bit precision is OK, once the maths are correctly done in the userland
program I wrote yesterday...
count=16525, precision=16 bits, sqrt(scaled_count)=4113, reciprocal(sq)=1fde240, Newton=1fd0000
interval/sqrt(16525) =
777909 (float compute) // (u32)(interval/sqrt(count))
778020 (integer approx) // reciprocal_divide(interval, rec)
777926 (int_sqrt_div) // int_sqrt_div(interval, count)
776672 (Newton approx) // reciprocal_divide(interval, previnv << shift)
count=9889134, precision=16 bits, sqrt(scaled_count)=50315,
reciprocal(sq)=14d720, Newton=140000
interval/sqrt(9889134) =
31799 (float compute)
31799 (integer approx)
31799 (int_sqrt_div)
30517 (Newton approx)
And kernel code using u16 :
6a1: 0f b7 72 0a movzwl 0xa(%rdx),%esi
6a5: 8b 3a mov (%rdx),%edi
6a7: 83 c7 01 add $0x1,%edi
6aa: c1 e6 10 shl $0x10,%esi
6ad: 89 3a mov %edi,(%rdx) vars->count++
6af: 89 ff mov %edi,%edi
6b1: 89 f6 mov %esi,%esi
6b3: 48 89 f1 mov %rsi,%rcx
6b6: 48 0f af ce imul %rsi,%rcx
6ba: 48 c1 e9 20 shr $0x20,%rcx
6be: 48 0f af cf imul %rdi,%rcx
6c2: 48 bf 00 00 00 00 03 mov $0x300000000,%rdi
6c9: 00 00 00
6cc: 48 29 cf sub %rcx,%rdi
6cf: 48 89 f9 mov %rdi,%rcx
6d2: 48 c1 e9 02 shr $0x2,%rcx
6d6: 48 0f af ce imul %rsi,%rcx
6da: 48 c1 e9 2f shr $0x2f,%rcx
6de: 66 89 4a 0a mov %cx,0xa(%rdx)
Fell free to add following cleanup patch, if you like it ;)
Thanks
[PATCH net-next] codel: use u16 field instead of 31bits for rec_inv_sqrt
David pointed out gcc might generate poor code with 31bit fields.
Using u16 is more than enough and permits a better code output.
Also make the code intent more readable using constants, fixed point arithmetic
not being trivial for everybody.
Suggested-by: David Miller <davem@davemloft.net>
Signed-off-by: Eric Dumazet <edumazet@google.com>
---
include/net/codel.h | 25 +++++++++++++++----------
1 file changed, 15 insertions(+), 10 deletions(-)
diff --git a/include/net/codel.h b/include/net/codel.h
index bd8747c..7546517 100644
--- a/include/net/codel.h
+++ b/include/net/codel.h
@@ -133,13 +133,17 @@ struct codel_params {
struct codel_vars {
u32 count;
u32 lastcount;
- bool dropping:1;
- u32 rec_inv_sqrt:31;
+ bool dropping;
+ u16 rec_inv_sqrt;
codel_time_t first_above_time;
codel_time_t drop_next;
codel_time_t ldelay;
};
+#define REC_INV_SQRT_BITS (8 * sizeof(u16)) /* or sizeof_in_bits(rec_inv_sqrt) */
+/* needed shift to get a Q0.32 number from rec_inv_sqrt */
+#define REC_INV_SQRT_SHIFT (32 - REC_INV_SQRT_BITS)
+
/**
* struct codel_stats - contains codel shared variables and stats
* @maxpacket: largest packet we've seen so far
@@ -173,17 +177,18 @@ static void codel_stats_init(struct codel_stats *stats)
* http://en.wikipedia.org/wiki/Methods_of_computing_square_roots#Iterative_methods_for_reciprocal_square_roots
* new_invsqrt = (invsqrt / 2) * (3 - count * invsqrt^2)
*
- * Here, invsqrt is a fixed point number (< 1.0), 31bit mantissa)
+ * Here, invsqrt is a fixed point number (< 1.0), 32bit mantissa, aka Q0.32
*/
static void codel_Newton_step(struct codel_vars *vars)
{
- u32 invsqrt = vars->rec_inv_sqrt;
- u32 invsqrt2 = ((u64)invsqrt * invsqrt) >> 31;
- u64 val = (3LL << 31) - ((u64)vars->count * invsqrt2);
+ u32 invsqrt = ((u32)vars->rec_inv_sqrt) << REC_INV_SQRT_SHIFT;
+ u32 invsqrt2 = ((u64)invsqrt * invsqrt) >> 32;
+ u64 val = (3LL << 32) - ((u64)vars->count * invsqrt2);
- val = (val * invsqrt) >> 32;
+ val >>= 2; /* avoid overflow in following multiply */
+ val = (val * invsqrt) >> (32 - 2 + 1);
- vars->rec_inv_sqrt = val;
+ vars->rec_inv_sqrt = val >> REC_INV_SQRT_SHIFT;
}
/*
@@ -195,7 +200,7 @@ static codel_time_t codel_control_law(codel_time_t t,
codel_time_t interval,
u32 rec_inv_sqrt)
{
- return t + reciprocal_divide(interval, rec_inv_sqrt << 1);
+ return t + reciprocal_divide(interval, rec_inv_sqrt << REC_INV_SQRT_SHIFT);
}
@@ -326,7 +331,7 @@ static struct sk_buff *codel_dequeue(struct Qdisc *sch,
codel_Newton_step(vars);
} else {
vars->count = 1;
- vars->rec_inv_sqrt = 0x7fffffff;
+ vars->rec_inv_sqrt = ~0U >> REC_INV_SQRT_SHIFT;
}
vars->lastcount = vars->count;
vars->drop_next = codel_control_law(now, params->interval,
^ permalink raw reply related
* Re: [PATCH 1/2] Bluetooth: notify userspace of security level change
From: Gustavo Padovan @ 2012-05-13 6:22 UTC (permalink / raw)
To: linville; +Cc: davem, linux-wireless, linux-bluetooth, linux-kernel, netdev
In-Reply-To: <1336849910-29064-1-git-send-email-gustavo@padovan.org>
Hi,
* Gustavo Padovan <gustavo@padovan.org> [2012-05-12 16:11:49 -0300]:
> When the userspace request a security level change it needs to be notified
> of when the change is complete.
> This patch make the socket non writable while the security request is
> ongoing. If it succeeds POLL_OUT is emitted, otherwise the channel is
> disconnected.
I just sent a second version of this patches that includes comments Johan made
in the last e-mail for better understanding of what the problem really is.
Gustavo
^ permalink raw reply
* [PATCH 1/2] Bluetooth: notify userspace of security level change
From: Gustavo Padovan @ 2012-05-13 6:20 UTC (permalink / raw)
To: linville-2XuSBdqkA4R54TAoqtyWWQ
Cc: davem-fT/PcQaiUtIeIZ0/mPfg9Q,
linux-wireless-u79uwXL29TY76Z2rM5mHXA,
linux-bluetooth-u79uwXL29TY76Z2rM5mHXA,
linux-kernel-u79uwXL29TY76Z2rM5mHXA,
netdev-u79uwXL29TY76Z2rM5mHXA
In-Reply-To: <1336849910-29064-2-git-send-email-gustavo-THi1TnShQwVAfugRpC6u6w@public.gmane.org>
It fixes L2CAP socket based security level elevation during a
connection. The HID profile needs this (for keyboards) and it is the only
way to achieve the security level elevation when using the management
interface to talk to the kernel (hence the management enabling patch
being the one that exposes this issue).
It enables the userspace a security level change when the socket is
already connected and create a way to notify the socket the result of the
request. At the moment of the request the socket is made non writable, if
the request fails the connections closes, otherwise the socket is made
writable again, POLL_OUT is emmited.
Signed-off-by: Gustavo Padovan <gustavo-THi1TnShQwVAfugRpC6u6w@public.gmane.org>
Acked-by: Marcel Holtmann <marcel-kz+m5ild9QBg9hUCZPvPmw@public.gmane.org>
Signed-off-by: Johan Hedberg <johan.hedberg-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
---
include/net/bluetooth/bluetooth.h | 1 +
net/bluetooth/af_bluetooth.c | 2 +-
net/bluetooth/hci_event.c | 7 +++++++
net/bluetooth/l2cap_core.c | 5 +++++
net/bluetooth/l2cap_sock.c | 12 ++++++++----
5 files changed, 22 insertions(+), 5 deletions(-)
diff --git a/include/net/bluetooth/bluetooth.h b/include/net/bluetooth/bluetooth.h
index 262ebd1..a65910b 100644
--- a/include/net/bluetooth/bluetooth.h
+++ b/include/net/bluetooth/bluetooth.h
@@ -191,6 +191,7 @@ struct bt_sock {
struct list_head accept_q;
struct sock *parent;
u32 defer_setup;
+ bool suspended;
};
struct bt_sock_list {
diff --git a/net/bluetooth/af_bluetooth.c b/net/bluetooth/af_bluetooth.c
index 72eb187..6fb68a9 100644
--- a/net/bluetooth/af_bluetooth.c
+++ b/net/bluetooth/af_bluetooth.c
@@ -450,7 +450,7 @@ unsigned int bt_sock_poll(struct file *file, struct socket *sock, poll_table *wa
sk->sk_state == BT_CONFIG)
return mask;
- if (sock_writeable(sk))
+ if (!bt_sk(sk)->suspended && sock_writeable(sk))
mask |= POLLOUT | POLLWRNORM | POLLWRBAND;
else
set_bit(SOCK_ASYNC_NOSPACE, &sk->sk_socket->flags);
diff --git a/net/bluetooth/hci_event.c b/net/bluetooth/hci_event.c
index 7f87a70..ff38cc6 100644
--- a/net/bluetooth/hci_event.c
+++ b/net/bluetooth/hci_event.c
@@ -2040,6 +2040,12 @@ static inline void hci_encrypt_change_evt(struct hci_dev *hdev, struct sk_buff *
clear_bit(HCI_CONN_ENCRYPT_PEND, &conn->flags);
+ if (ev->status && conn->state == BT_CONNECTED) {
+ hci_acl_disconn(conn, 0x13);
+ hci_conn_put(conn);
+ goto unlock;
+ }
+
if (conn->state == BT_CONFIG) {
if (!ev->status)
conn->state = BT_CONNECTED;
@@ -2050,6 +2056,7 @@ static inline void hci_encrypt_change_evt(struct hci_dev *hdev, struct sk_buff *
hci_encrypt_cfm(conn, ev->status, ev->encrypt);
}
+unlock:
hci_dev_unlock(hdev);
}
diff --git a/net/bluetooth/l2cap_core.c b/net/bluetooth/l2cap_core.c
index 38d934a..c073533 100644
--- a/net/bluetooth/l2cap_core.c
+++ b/net/bluetooth/l2cap_core.c
@@ -4590,6 +4590,11 @@ int l2cap_security_cfm(struct hci_conn *hcon, u8 status, u8 encrypt)
if (!status && (chan->state == BT_CONNECTED ||
chan->state == BT_CONFIG)) {
+ struct sock *sk = chan->sk;
+
+ bt_sk(sk)->suspended = false;
+ sk->sk_state_change(sk);
+
l2cap_check_encryption(chan, encrypt);
l2cap_chan_unlock(chan);
continue;
diff --git a/net/bluetooth/l2cap_sock.c b/net/bluetooth/l2cap_sock.c
index 29122ed..04e7c17 100644
--- a/net/bluetooth/l2cap_sock.c
+++ b/net/bluetooth/l2cap_sock.c
@@ -592,10 +592,14 @@ static int l2cap_sock_setsockopt(struct socket *sock, int level, int optname, ch
sk->sk_state = BT_CONFIG;
chan->state = BT_CONFIG;
- /* or for ACL link, under defer_setup time */
- } else if (sk->sk_state == BT_CONNECT2 &&
- bt_sk(sk)->defer_setup) {
- err = l2cap_chan_check_security(chan);
+ /* or for ACL link */
+ } else if ((sk->sk_state == BT_CONNECT2 &&
+ bt_sk(sk)->defer_setup) ||
+ sk->sk_state == BT_CONNECTED) {
+ if (!l2cap_chan_check_security(chan))
+ bt_sk(sk)->suspended = true;
+ else
+ sk->sk_state_change(sk);
} else {
err = -EINVAL;
}
--
1.7.10.1
^ permalink raw reply related
* Re: pull request: bluetooth 2012-05-04
From: Johan Hedberg @ 2012-05-13 4:31 UTC (permalink / raw)
To: Gustavo Padovan, John W. Linville, David Miller,
linux-wireless-u79uwXL29TY76Z2rM5mHXA,
linux-bluetooth-u79uwXL29TY76Z2rM5mHXA,
linux-kernel-u79uwXL29TY76Z2rM5mHXA,
netdev-u79uwXL29TY76Z2rM5mHXA
In-Reply-To: <20120512190900.GA15956@joana>
Hi,
On Sat, May 12, 2012, Gustavo Padovan wrote:
> In my point of view there are two commits there that are really necessary:
>
> Gustavo Padovan (2):
> Bluetooth: report the right security level in getsockopt
>
> Johan Hedberg (2):
> Bluetooth: mgmt: Fix device_connected sending order
>
> They fix a userspace breakage caused by:
>
> Author: Marcel Holtmann <marcel-kz+m5ild9QBg9hUCZPvPmw@public.gmane.org>
> Date: Mon Feb 20 21:24:37 2012 +0100
>
> Bluetooth: Always enable management interface
>
> The management interface API has reached stable version 1.0 and thus
> it can now be always enabled. All future changes will be made backwards
> compatible.
>
> Signed-off-by: Marcel Holtmann <marcel-kz+m5ild9QBg9hUCZPvPmw@public.gmane.org>
> Signed-off-by: Johan Hedberg <johan.hedberg-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
>
>
> This cause no crash, but it make bluetooth keyboards stop to work with Linux.
> This a serious breakage IMO. I really would like to see at least these two
> patches in. The other fixes can wait.
I'd just like to add some further background since I don't think it's
that obvious how these are regression fixes:
The device_connected fix should be quite self-explanatory, but it's
actually a wider issue than just for keyboards. All profiles that do
incoming connection authorization (e.g. headsets) will break without it
with specific hardware. The reason it wasn't caught earlier is that it
only occurs with specific Bluetooth adapters.
As for the security level patch, this fixes L2CAP socket based security
level elevation during a connection. The HID profile needs this (for
keyboards) and it is the only way to achieve the security level
elevation when using the management interface to talk to the kernel
(hence the management enabling patch being the one that exposes this
issue).
Johan
--
To unsubscribe from this list: send the line "unsubscribe linux-wireless" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply
* Re: [PATCH net-next] etherdevice: Remove now unused compare_ether_addr_64bits
From: David Miller @ 2012-05-13 3:35 UTC (permalink / raw)
To: joe; +Cc: netdev, linux-kernel
In-Reply-To: <1336774866.13434.16.camel@joe2Laptop>
From: Joe Perches <joe@perches.com>
Date: Fri, 11 May 2012 15:21:06 -0700
> Move and invert the logic from the otherwise unused
> compare_ether_addr_64bits to ether_addr_equal_64bits.
>
> Neaten the logic in is_etherdev_addr.
>
> Signed-off-by: Joe Perches <joe@perches.com>
Applied, thanks Joe.
^ permalink raw reply
* Re: [PATCH 1/2 net v2] 6lowpan: rework data fetching from skb
From: David Miller @ 2012-05-13 3:28 UTC (permalink / raw)
To: alex.bluesman.smirnov; +Cc: netdev, eric.dumazet
In-Reply-To: <1336748290-25649-1-git-send-email-alex.bluesman.smirnov@gmail.com>
This and your other patch are not appropriate this late in
the -rc series, and Linus has ratched up his standards even
higher with todays' RC release.
You'll need to resubmit these for net-next.
^ permalink raw reply
* Re: [net-next v3 0/2][pull request] Intel Wired LAN Driver Updates
From: David Miller @ 2012-05-12 21:56 UTC (permalink / raw)
To: jeffrey.t.kirsher; +Cc: netdev, gospo, sassmann
In-Reply-To: <1336854637-6708-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Date: Sat, 12 May 2012 13:30:35 -0700
> This series of patches contains updates for igb. Most notably,
> the addition of i210 & i211 support.
>
> v2: fix up patch 2 based on feedback from David Miller
> v3: fix compiler warning
>
> The following are changes since commit 4b549a2ef4bef9965d97cbd992ba67930cd3e0fe:
> fq_codel: Fair Queue Codel AQM
> and are available in the git repository at:
> git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-next master
Ok, this works, pulled (for real) :-)
^ permalink raw reply
* Re: [PATCH net-next] codel: use Newton method instead of sqrt() and divides
From: David Miller @ 2012-05-12 21:52 UTC (permalink / raw)
To: eric.dumazet
Cc: dave.taht, netdev, nichols, van, codel, ycheng, mattmathis,
therbert, shemminger, nanditad
In-Reply-To: <1336859324.31653.1385.camel@edumazet-glaptop>
From: Eric Dumazet <eric.dumazet@gmail.com>
Date: Sat, 12 May 2012 23:48:44 +0200
> On Sat, 2012-05-12 at 16:45 -0400, David Miller wrote:
>> Using a u16 would also work for me.
>
> I tried it but it gives noticeable errors for count > 16000, and no
> speed gain.
...
> And if a flow is really agressive, count can grow above 10^6
>
>> > By the way, gcc on x86 generates nice "and 0xfffffffe,%eax" instruction
>> > for (vars->rec_inv_sqrt << 1).
>>
>> Yeah but what do stores of ->rec_inv_sqrt look like?
>
> The load is "shr %edi" as in :
> and the store an "or %ecx,%esi"
Ok, fair enough.
^ permalink raw reply
* Re: [PATCH net-next] codel: use Newton method instead of sqrt() and divides
From: Eric Dumazet @ 2012-05-12 21:48 UTC (permalink / raw)
To: David Miller
Cc: dave.taht, netdev, nichols, van, codel, ycheng, mattmathis,
therbert, shemminger, nanditad
In-Reply-To: <20120512.164513.1156706853054390966.davem@davemloft.net>
On Sat, 2012-05-12 at 16:45 -0400, David Miller wrote:
> From: Eric Dumazet <eric.dumazet@gmail.com>
> Date: Sat, 12 May 2012 22:40:56 +0200
>
> > 24 bit of precision for the reciprocal value is more than enough (Van
> > suggested 16 bits in fact), so we have actually room for 7 bits if
> > needed.
>
> Using a u16 would also work for me.
I tried it but it gives noticeable errors for count > 16000, and no
speed gain.
count=16525 scale=0 sqrt(scaled_count)=2056 reciprocal(sq)=1fe0200
Newton=235
interval/sqrt(16525) =
777909 (float compute)
778210 (integer approx)
777926 (int_sqrt_div)
862121 (Newton approx)
And if a flow is really agressive, count can grow above 10^6
>
> > By the way, gcc on x86 generates nice "and 0xfffffffe,%eax" instruction
> > for (vars->rec_inv_sqrt << 1).
>
> Yeah but what do stores of ->rec_inv_sqrt look like?
The load is "shr %edi" as in :
and the store an "or %ecx,%esi"
5f2: 8b 72 08 mov 0x8(%rdx),%esi
5f5: 44 8b 02 mov (%rdx),%r8d
5f8: 89 f7 mov %esi,%edi
5fa: 41 83 c0 01 add $0x1,%r8d vars->count + 1
5fe: 83 e6 01 and $0x1,%esi vars->dropping in esi
601: d1 ef shr %edi
603: 44 89 02 mov %r8d,(%rdx) vars->count++;
606: 45 89 c0 mov %r8d,%r8d
609: 89 ff mov %edi,%edi
60b: 48 89 f9 mov %rdi,%rcx
60e: 48 0f af cf imul %rdi,%rcx
612: 48 c1 e9 1f shr $0x1f,%rcx
616: 49 0f af c8 imul %r8,%rcx
61a: 49 b8 00 00 00 80 01 mov $0x180000000,%r8
621: 00 00 00
624: 49 29 c8 sub %rcx,%r8
627: 4c 89 c1 mov %r8,%rcx
62a: 48 0f af cf imul %rdi,%rcx
62e: 48 c1 e9 20 shr $0x20,%rcx
632: 01 c9 add %ecx,%ecx
634: 09 ce or %ecx,%esi combine the two fields
636: 89 72 08 mov %esi,0x8(%rdx) final store
Using 24bits generates roughly same code. (constants are different)
^ permalink raw reply
* Re: [PATCH net-next] codel: use Newton method instead of sqrt() and divides
From: David Miller @ 2012-05-12 20:45 UTC (permalink / raw)
To: eric.dumazet
Cc: dave.taht, netdev, nichols, van, codel, ycheng, mattmathis,
therbert, shemminger, nanditad
In-Reply-To: <1336855256.31653.1329.camel@edumazet-glaptop>
From: Eric Dumazet <eric.dumazet@gmail.com>
Date: Sat, 12 May 2012 22:40:56 +0200
> 24 bit of precision for the reciprocal value is more than enough (Van
> suggested 16 bits in fact), so we have actually room for 7 bits if
> needed.
Using a u16 would also work for me.
> By the way, gcc on x86 generates nice "and 0xfffffffe,%eax" instruction
> for (vars->rec_inv_sqrt << 1).
Yeah but what do stores of ->rec_inv_sqrt look like?
^ permalink raw reply
* Re: [PATCH v2 net-next] fq_codel: Fair Queue Codel AQM
From: Eric Dumazet @ 2012-05-12 20:42 UTC (permalink / raw)
To: David Miller
Cc: xiaosuo, netdev, dave.taht, nichols, van, therbert, mattmathis,
ycheng, shemminger, maze, nanditad
In-Reply-To: <20120512.155555.1011024268184559052.davem@davemloft.net>
On Sat, 2012-05-12 at 15:55 -0400, David Miller wrote:
> I applied this with the WARN_ON_ONCE(1) removed but there was another
> problem.
>
> When you include ping output in your commit message that "---" string
> told GIT that this was the end of the commit message when in fact
> there was more content including your signoff.
>
> I caught it and fixed it up, but please be more mindful of this in
> the future.
Oops... I feel sorry and I fear this will happen again, because I use a
lot "ping". I'll try to remember this.
Thanks a lot David.
^ permalink raw reply
* Re: [PATCH net-next] codel: use Newton method instead of sqrt() and divides
From: Eric Dumazet @ 2012-05-12 20:40 UTC (permalink / raw)
To: David Miller
Cc: dave.taht, netdev, nichols, van, codel, ycheng, mattmathis,
therbert, shemminger, nanditad
In-Reply-To: <20120512.155259.1178343836887150194.davem@davemloft.net>
On Sat, 2012-05-12 at 15:52 -0400, David Miller wrote:
> Applied but I never like that bitfield sharing for real integers.
>
> GCC makes a complete mess of it as it extracts and inserts the
> integer value into that bit field. You are guarenteed to get
> better code if you do this by hand in a full u32.
>
> Either that or just bite the bullet and use a completely seperate
> field, maybe we'll need more boolean states later.
I couldnt use a full u32 or else fq_codel cell was > 64 bytes (or I
would have to remove the 'dropped' field)
24 bit of precision for the reciprocal value is more than enough (Van
suggested 16 bits in fact), so we have actually room for 7 bits if
needed.
By the way, gcc on x86 generates nice "and 0xfffffffe,%eax" instruction
for (vars->rec_inv_sqrt << 1).
Thanks
^ permalink raw reply
* [net-next v3 2/2] igb: Add Support for new i210/i211 devices.
From: Jeff Kirsher @ 2012-05-12 20:30 UTC (permalink / raw)
To: davem; +Cc: Carolyn Wyborny, netdev, gospo, sassmann, Jeff Kirsher
In-Reply-To: <1336854637-6708-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Carolyn Wyborny <carolyn.wyborny@intel.com>
This patch adds new initialization functions and device support
for i210 and i211 devices.
v2: cleanup acquire/release nvm functions
v3: fix compiler warning
Signed-off-by: Carolyn Wyborny <carolyn.wyborny@intel.com>
Tested-by: Jeff Pieper <jeffrey.e.pieper@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/igb/Makefile | 3 +-
drivers/net/ethernet/intel/igb/e1000_82575.c | 182 ++++++--
drivers/net/ethernet/intel/igb/e1000_82575.h | 3 +-
drivers/net/ethernet/intel/igb/e1000_defines.h | 33 ++
drivers/net/ethernet/intel/igb/e1000_hw.h | 14 +
drivers/net/ethernet/intel/igb/e1000_i210.c | 603 ++++++++++++++++++++++++
drivers/net/ethernet/intel/igb/e1000_i210.h | 76 +++
drivers/net/ethernet/intel/igb/e1000_mac.c | 1 +
drivers/net/ethernet/intel/igb/e1000_nvm.c | 1 -
drivers/net/ethernet/intel/igb/e1000_phy.c | 147 ++++++-
drivers/net/ethernet/intel/igb/e1000_phy.h | 15 +
drivers/net/ethernet/intel/igb/e1000_regs.h | 14 +
drivers/net/ethernet/intel/igb/igb.h | 9 +-
drivers/net/ethernet/intel/igb/igb_ethtool.c | 103 +++-
drivers/net/ethernet/intel/igb/igb_main.c | 107 ++++-
drivers/net/ethernet/intel/igb/igb_ptp.c | 4 +
16 files changed, 1205 insertions(+), 110 deletions(-)
create mode 100644 drivers/net/ethernet/intel/igb/e1000_i210.c
create mode 100644 drivers/net/ethernet/intel/igb/e1000_i210.h
diff --git a/drivers/net/ethernet/intel/igb/Makefile b/drivers/net/ethernet/intel/igb/Makefile
index 4bd16e2..97c197f 100644
--- a/drivers/net/ethernet/intel/igb/Makefile
+++ b/drivers/net/ethernet/intel/igb/Makefile
@@ -33,6 +33,7 @@
obj-$(CONFIG_IGB) += igb.o
igb-objs := igb_main.o igb_ethtool.o e1000_82575.o \
- e1000_mac.o e1000_nvm.o e1000_phy.o e1000_mbx.o
+ e1000_mac.o e1000_nvm.o e1000_phy.o e1000_mbx.o \
+ e1000_i210.o
igb-$(CONFIG_IGB_PTP) += igb_ptp.o
diff --git a/drivers/net/ethernet/intel/igb/e1000_82575.c b/drivers/net/ethernet/intel/igb/e1000_82575.c
index b945992..e650839 100644
--- a/drivers/net/ethernet/intel/igb/e1000_82575.c
+++ b/drivers/net/ethernet/intel/igb/e1000_82575.c
@@ -36,6 +36,7 @@
#include "e1000_mac.h"
#include "e1000_82575.h"
+#include "e1000_i210.h"
static s32 igb_get_invariants_82575(struct e1000_hw *);
static s32 igb_acquire_phy_82575(struct e1000_hw *);
@@ -98,6 +99,8 @@ static bool igb_sgmii_uses_mdio_82575(struct e1000_hw *hw)
break;
case e1000_82580:
case e1000_i350:
+ case e1000_i210:
+ case e1000_i211:
reg = rd32(E1000_MDICNFG);
ext_mdio = !!(reg & E1000_MDICNFG_EXT_MDIO);
break;
@@ -152,6 +155,17 @@ static s32 igb_get_invariants_82575(struct e1000_hw *hw)
case E1000_DEV_ID_I350_SGMII:
mac->type = e1000_i350;
break;
+ case E1000_DEV_ID_I210_COPPER:
+ case E1000_DEV_ID_I210_COPPER_OEM1:
+ case E1000_DEV_ID_I210_COPPER_IT:
+ case E1000_DEV_ID_I210_FIBER:
+ case E1000_DEV_ID_I210_SERDES:
+ case E1000_DEV_ID_I210_SGMII:
+ mac->type = e1000_i210;
+ break;
+ case E1000_DEV_ID_I211_COPPER:
+ mac->type = e1000_i211;
+ break;
default:
return -E1000_ERR_MAC_INIT;
break;
@@ -184,26 +198,44 @@ static s32 igb_get_invariants_82575(struct e1000_hw *hw)
/* Set mta register count */
mac->mta_reg_count = 128;
/* Set rar entry count */
- mac->rar_entry_count = E1000_RAR_ENTRIES_82575;
- if (mac->type == e1000_82576)
+ switch (mac->type) {
+ case e1000_82576:
mac->rar_entry_count = E1000_RAR_ENTRIES_82576;
- if (mac->type == e1000_82580)
+ break;
+ case e1000_82580:
mac->rar_entry_count = E1000_RAR_ENTRIES_82580;
- if (mac->type == e1000_i350)
+ break;
+ case e1000_i350:
+ case e1000_i210:
+ case e1000_i211:
mac->rar_entry_count = E1000_RAR_ENTRIES_I350;
+ break;
+ default:
+ mac->rar_entry_count = E1000_RAR_ENTRIES_82575;
+ break;
+ }
/* reset */
if (mac->type >= e1000_82580)
mac->ops.reset_hw = igb_reset_hw_82580;
else
mac->ops.reset_hw = igb_reset_hw_82575;
+
+ if (mac->type >= e1000_i210) {
+ mac->ops.acquire_swfw_sync = igb_acquire_swfw_sync_i210;
+ mac->ops.release_swfw_sync = igb_release_swfw_sync_i210;
+ } else {
+ mac->ops.acquire_swfw_sync = igb_acquire_swfw_sync_82575;
+ mac->ops.release_swfw_sync = igb_release_swfw_sync_82575;
+ }
+
/* Set if part includes ASF firmware */
mac->asf_firmware_present = true;
/* Set if manageability features are enabled. */
mac->arc_subsystem_valid =
(rd32(E1000_FWSM) & E1000_FWSM_MODE_MASK)
? true : false;
- /* enable EEE on i350 parts */
- if (mac->type == e1000_i350)
+ /* enable EEE on i350 parts and later parts */
+ if (mac->type >= e1000_i350)
dev_spec->eee_disable = false;
else
dev_spec->eee_disable = true;
@@ -215,26 +247,6 @@ static s32 igb_get_invariants_82575(struct e1000_hw *hw)
/* NVM initialization */
eecd = rd32(E1000_EECD);
-
- nvm->opcode_bits = 8;
- nvm->delay_usec = 1;
- switch (nvm->override) {
- case e1000_nvm_override_spi_large:
- nvm->page_size = 32;
- nvm->address_bits = 16;
- break;
- case e1000_nvm_override_spi_small:
- nvm->page_size = 8;
- nvm->address_bits = 8;
- break;
- default:
- nvm->page_size = eecd & E1000_EECD_ADDR_BITS ? 32 : 8;
- nvm->address_bits = eecd & E1000_EECD_ADDR_BITS ? 16 : 8;
- break;
- }
-
- nvm->type = e1000_nvm_eeprom_spi;
-
size = (u16)((eecd & E1000_EECD_SIZE_EX_MASK) >>
E1000_EECD_SIZE_EX_SHIFT);
@@ -244,6 +256,33 @@ static s32 igb_get_invariants_82575(struct e1000_hw *hw)
*/
size += NVM_WORD_SIZE_BASE_SHIFT;
+ nvm->word_size = 1 << size;
+ if (hw->mac.type < e1000_i210) {
+ nvm->opcode_bits = 8;
+ nvm->delay_usec = 1;
+ switch (nvm->override) {
+ case e1000_nvm_override_spi_large:
+ nvm->page_size = 32;
+ nvm->address_bits = 16;
+ break;
+ case e1000_nvm_override_spi_small:
+ nvm->page_size = 8;
+ nvm->address_bits = 8;
+ break;
+ default:
+ nvm->page_size = eecd
+ & E1000_EECD_ADDR_BITS ? 32 : 8;
+ nvm->address_bits = eecd
+ & E1000_EECD_ADDR_BITS ? 16 : 8;
+ break;
+ }
+ if (nvm->word_size == (1 << 15))
+ nvm->page_size = 128;
+
+ nvm->type = e1000_nvm_eeprom_spi;
+ } else
+ nvm->type = e1000_nvm_flash_hw;
+
/*
* Check for invalid size
*/
@@ -251,32 +290,60 @@ static s32 igb_get_invariants_82575(struct e1000_hw *hw)
pr_notice("The NVM size is not valid, defaulting to 32K\n");
size = 15;
}
- nvm->word_size = 1 << size;
- if (nvm->word_size == (1 << 15))
- nvm->page_size = 128;
/* NVM Function Pointers */
- nvm->ops.acquire = igb_acquire_nvm_82575;
- if (nvm->word_size < (1 << 15))
- nvm->ops.read = igb_read_nvm_eerd;
- else
- nvm->ops.read = igb_read_nvm_spi;
-
- nvm->ops.release = igb_release_nvm_82575;
switch (hw->mac.type) {
case e1000_82580:
nvm->ops.validate = igb_validate_nvm_checksum_82580;
nvm->ops.update = igb_update_nvm_checksum_82580;
+ nvm->ops.acquire = igb_acquire_nvm_82575;
+ nvm->ops.release = igb_release_nvm_82575;
+ if (nvm->word_size < (1 << 15))
+ nvm->ops.read = igb_read_nvm_eerd;
+ else
+ nvm->ops.read = igb_read_nvm_spi;
+ nvm->ops.write = igb_write_nvm_spi;
break;
case e1000_i350:
nvm->ops.validate = igb_validate_nvm_checksum_i350;
nvm->ops.update = igb_update_nvm_checksum_i350;
+ nvm->ops.acquire = igb_acquire_nvm_82575;
+ nvm->ops.release = igb_release_nvm_82575;
+ if (nvm->word_size < (1 << 15))
+ nvm->ops.read = igb_read_nvm_eerd;
+ else
+ nvm->ops.read = igb_read_nvm_spi;
+ nvm->ops.write = igb_write_nvm_spi;
+ break;
+ case e1000_i210:
+ nvm->ops.validate = igb_validate_nvm_checksum_i210;
+ nvm->ops.update = igb_update_nvm_checksum_i210;
+ nvm->ops.acquire = igb_acquire_nvm_i210;
+ nvm->ops.release = igb_release_nvm_i210;
+ nvm->ops.read = igb_read_nvm_srrd_i210;
+ nvm->ops.valid_led_default = igb_valid_led_default_i210;
+ break;
+ case e1000_i211:
+ nvm->ops.acquire = igb_acquire_nvm_i210;
+ nvm->ops.release = igb_release_nvm_i210;
+ nvm->ops.read = igb_read_nvm_i211;
+ nvm->ops.valid_led_default = igb_valid_led_default_i210;
+ nvm->ops.validate = NULL;
+ nvm->ops.update = NULL;
+ nvm->ops.write = NULL;
break;
default:
nvm->ops.validate = igb_validate_nvm_checksum;
nvm->ops.update = igb_update_nvm_checksum;
+ nvm->ops.acquire = igb_acquire_nvm_82575;
+ nvm->ops.release = igb_release_nvm_82575;
+ if (nvm->word_size < (1 << 15))
+ nvm->ops.read = igb_read_nvm_eerd;
+ else
+ nvm->ops.read = igb_read_nvm_spi;
+ nvm->ops.write = igb_write_nvm_spi;
+ break;
}
- nvm->ops.write = igb_write_nvm_spi;
/* if part supports SR-IOV then initialize mailbox parameters */
switch (mac->type) {
@@ -314,9 +381,13 @@ static s32 igb_get_invariants_82575(struct e1000_hw *hw)
if (igb_sgmii_active_82575(hw) && !igb_sgmii_uses_mdio_82575(hw)) {
phy->ops.read_reg = igb_read_phy_reg_sgmii_82575;
phy->ops.write_reg = igb_write_phy_reg_sgmii_82575;
- } else if (hw->mac.type >= e1000_82580) {
+ } else if ((hw->mac.type == e1000_82580)
+ || (hw->mac.type == e1000_i350)) {
phy->ops.read_reg = igb_read_phy_reg_82580;
phy->ops.write_reg = igb_write_phy_reg_82580;
+ } else if (hw->phy.type >= e1000_phy_i210) {
+ phy->ops.read_reg = igb_read_phy_reg_gs40g;
+ phy->ops.write_reg = igb_write_phy_reg_gs40g;
} else {
phy->ops.read_reg = igb_read_phy_reg_igp;
phy->ops.write_reg = igb_write_phy_reg_igp;
@@ -345,6 +416,14 @@ static s32 igb_get_invariants_82575(struct e1000_hw *hw)
else
phy->ops.get_cable_length = igb_get_cable_length_m88;
+ if (phy->id == I210_I_PHY_ID) {
+ phy->ops.get_cable_length =
+ igb_get_cable_length_m88_gen2;
+ phy->ops.set_d0_lplu_state =
+ igb_set_d0_lplu_state_82580;
+ phy->ops.set_d3_lplu_state =
+ igb_set_d3_lplu_state_82580;
+ }
phy->ops.force_speed_duplex = igb_phy_force_speed_duplex_m88;
break;
case IGP03E1000_E_PHY_ID:
@@ -364,6 +443,15 @@ static s32 igb_get_invariants_82575(struct e1000_hw *hw)
phy->ops.set_d0_lplu_state = igb_set_d0_lplu_state_82580;
phy->ops.set_d3_lplu_state = igb_set_d3_lplu_state_82580;
break;
+ case I210_I_PHY_ID:
+ phy->type = e1000_phy_i210;
+ phy->ops.get_phy_info = igb_get_phy_info_m88;
+ phy->ops.check_polarity = igb_check_polarity_m88;
+ phy->ops.get_cable_length = igb_get_cable_length_m88_gen2;
+ phy->ops.set_d0_lplu_state = igb_set_d0_lplu_state_82580;
+ phy->ops.set_d3_lplu_state = igb_set_d3_lplu_state_82580;
+ phy->ops.force_speed_duplex = igb_phy_force_speed_duplex_m88;
+ break;
default:
return -E1000_ERR_PHY;
}
@@ -389,7 +477,7 @@ static s32 igb_acquire_phy_82575(struct e1000_hw *hw)
else if (hw->bus.func == E1000_FUNC_3)
mask = E1000_SWFW_PHY3_SM;
- return igb_acquire_swfw_sync_82575(hw, mask);
+ return hw->mac.ops.acquire_swfw_sync(hw, mask);
}
/**
@@ -410,7 +498,7 @@ static void igb_release_phy_82575(struct e1000_hw *hw)
else if (hw->bus.func == E1000_FUNC_3)
mask = E1000_SWFW_PHY3_SM;
- igb_release_swfw_sync_82575(hw, mask);
+ hw->mac.ops.release_swfw_sync(hw, mask);
}
/**
@@ -514,6 +602,8 @@ static s32 igb_get_phy_id_82575(struct e1000_hw *hw)
break;
case e1000_82580:
case e1000_i350:
+ case e1000_i210:
+ case e1000_i211:
mdic = rd32(E1000_MDICNFG);
mdic &= E1000_MDICNFG_PHY_MASK;
phy->addr = mdic >> E1000_MDICNFG_PHY_SHIFT;
@@ -780,14 +870,14 @@ static s32 igb_acquire_nvm_82575(struct e1000_hw *hw)
{
s32 ret_val;
- ret_val = igb_acquire_swfw_sync_82575(hw, E1000_SWFW_EEP_SM);
+ ret_val = hw->mac.ops.acquire_swfw_sync(hw, E1000_SWFW_EEP_SM);
if (ret_val)
goto out;
ret_val = igb_acquire_nvm(hw);
if (ret_val)
- igb_release_swfw_sync_82575(hw, E1000_SWFW_EEP_SM);
+ hw->mac.ops.release_swfw_sync(hw, E1000_SWFW_EEP_SM);
out:
return ret_val;
@@ -803,7 +893,7 @@ out:
static void igb_release_nvm_82575(struct e1000_hw *hw)
{
igb_release_nvm(hw);
- igb_release_swfw_sync_82575(hw, E1000_SWFW_EEP_SM);
+ hw->mac.ops.release_swfw_sync(hw, E1000_SWFW_EEP_SM);
}
/**
@@ -1174,7 +1264,6 @@ static s32 igb_init_hw_82575(struct e1000_hw *hw)
* is no link.
*/
igb_clear_hw_cntrs_82575(hw);
-
return ret_val;
}
@@ -1211,6 +1300,7 @@ static s32 igb_setup_copper_link_82575(struct e1000_hw *hw)
}
}
switch (hw->phy.type) {
+ case e1000_phy_i210:
case e1000_phy_m88:
if (hw->phy.id == I347AT4_E_PHY_ID ||
hw->phy.id == M88E1112_E_PHY_ID)
@@ -1851,7 +1941,7 @@ static s32 igb_reset_hw_82580(struct e1000_hw *hw)
/* Determine whether or not a global dev reset is requested */
if (global_device_reset &&
- igb_acquire_swfw_sync_82575(hw, swmbsw_mask))
+ hw->mac.ops.acquire_swfw_sync(hw, swmbsw_mask))
global_device_reset = false;
if (global_device_reset &&
@@ -1897,7 +1987,7 @@ static s32 igb_reset_hw_82580(struct e1000_hw *hw)
/* Release semaphore */
if (global_device_reset)
- igb_release_swfw_sync_82575(hw, swmbsw_mask);
+ hw->mac.ops.release_swfw_sync(hw, swmbsw_mask);
return ret_val;
}
diff --git a/drivers/net/ethernet/intel/igb/e1000_82575.h b/drivers/net/ethernet/intel/igb/e1000_82575.h
index b927d79..e85c453 100644
--- a/drivers/net/ethernet/intel/igb/e1000_82575.h
+++ b/drivers/net/ethernet/intel/igb/e1000_82575.h
@@ -55,10 +55,11 @@ extern void igb_rx_fifo_flush_82575(struct e1000_hw *hw);
#define E1000_SRRCTL_DROP_EN 0x80000000
#define E1000_SRRCTL_TIMESTAMP 0x40000000
+
#define E1000_MRQC_ENABLE_RSS_4Q 0x00000002
#define E1000_MRQC_ENABLE_VMDQ 0x00000003
-#define E1000_MRQC_ENABLE_VMDQ_RSS_2Q 0x00000005
#define E1000_MRQC_RSS_FIELD_IPV4_UDP 0x00400000
+#define E1000_MRQC_ENABLE_VMDQ_RSS_2Q 0x00000005
#define E1000_MRQC_RSS_FIELD_IPV6_UDP 0x00800000
#define E1000_MRQC_RSS_FIELD_IPV6_UDP_EX 0x01000000
diff --git a/drivers/net/ethernet/intel/igb/e1000_defines.h b/drivers/net/ethernet/intel/igb/e1000_defines.h
index 89eb1f8..6409f85 100644
--- a/drivers/net/ethernet/intel/igb/e1000_defines.h
+++ b/drivers/net/ethernet/intel/igb/e1000_defines.h
@@ -458,6 +458,7 @@
#define E1000_ERR_INVALID_ARGUMENT 16
#define E1000_ERR_NO_SPACE 17
#define E1000_ERR_NVM_PBA_SECTION 18
+#define E1000_ERR_INVM_VALUE_NOT_FOUND 19
/* Loop limit on how long we wait for auto-negotiation to complete */
#define COPPER_LINK_UP_LIMIT 10
@@ -595,6 +596,25 @@
#define E1000_EECD_AUTO_RD 0x00000200 /* NVM Auto Read done */
#define E1000_EECD_SIZE_EX_MASK 0x00007800 /* NVM Size */
#define E1000_EECD_SIZE_EX_SHIFT 11
+#define E1000_EECD_FLUPD_I210 0x00800000 /* Update FLASH */
+#define E1000_EECD_FLUDONE_I210 0x04000000 /* Update FLASH done*/
+#define E1000_FLUDONE_ATTEMPTS 20000
+#define E1000_EERD_EEWR_MAX_COUNT 512 /* buffered EEPROM words rw */
+#define E1000_I210_FIFO_SEL_RX 0x00
+#define E1000_I210_FIFO_SEL_TX_QAV(_i) (0x02 + (_i))
+#define E1000_I210_FIFO_SEL_TX_LEGACY E1000_I210_FIFO_SEL_TX_QAV(0)
+#define E1000_I210_FIFO_SEL_BMC2OS_TX 0x06
+#define E1000_I210_FIFO_SEL_BMC2OS_RX 0x01
+#define E1000_EECD_FLUPD_I210 0x00800000 /* Update FLASH */
+#define E1000_EECD_FLUDONE_I210 0x04000000 /* Update FLASH done*/
+#define E1000_FLUDONE_ATTEMPTS 20000
+#define E1000_EERD_EEWR_MAX_COUNT 512 /* buffered EEPROM words rw */
+#define E1000_I210_FIFO_SEL_RX 0x00
+#define E1000_I210_FIFO_SEL_TX_QAV(_i) (0x02 + (_i))
+#define E1000_I210_FIFO_SEL_TX_LEGACY E1000_I210_FIFO_SEL_TX_QAV(0)
+#define E1000_I210_FIFO_SEL_BMC2OS_TX 0x06
+#define E1000_I210_FIFO_SEL_BMC2OS_RX 0x01
+
/* Offset to data in NVM read/write registers */
#define E1000_NVM_RW_REG_DATA 16
@@ -613,6 +633,16 @@
#define NVM_CHECKSUM_REG 0x003F
#define NVM_COMPATIBILITY_REG_3 0x0003
#define NVM_COMPATIBILITY_BIT_MASK 0x8000
+#define NVM_MAC_ADDR 0x0000
+#define NVM_SUB_DEV_ID 0x000B
+#define NVM_SUB_VEN_ID 0x000C
+#define NVM_DEV_ID 0x000D
+#define NVM_VEN_ID 0x000E
+#define NVM_INIT_CTRL_2 0x000F
+#define NVM_INIT_CTRL_4 0x0013
+#define NVM_LED_1_CFG 0x001C
+#define NVM_LED_0_2_CFG 0x001F
+
#define E1000_NVM_CFG_DONE_PORT_0 0x040000 /* MNG config cycle done */
#define E1000_NVM_CFG_DONE_PORT_1 0x080000 /* ...for second port */
@@ -639,6 +669,7 @@
#define NVM_PBA_OFFSET_0 8
#define NVM_PBA_OFFSET_1 9
+#define NVM_RESERVED_WORD 0xFFFF
#define NVM_PBA_PTR_GUARD 0xFAFA
#define NVM_WORD_SIZE_BASE_SHIFT 6
@@ -696,6 +727,7 @@
#define I82580_I_PHY_ID 0x015403A0
#define I350_I_PHY_ID 0x015403B0
#define M88_VENDOR 0x0141
+#define I210_I_PHY_ID 0x01410C00
/* M88E1000 Specific Registers */
#define M88E1000_PHY_SPEC_CTRL 0x10 /* PHY Specific Control Register */
@@ -815,6 +847,7 @@
#define E1000_IPCNFG_EEE_100M_AN 0x00000004 /* EEE Enable 100M AN */
#define E1000_EEER_TX_LPI_EN 0x00010000 /* EEE Tx LPI Enable */
#define E1000_EEER_RX_LPI_EN 0x00020000 /* EEE Rx LPI Enable */
+#define E1000_EEER_FRC_AN 0x10000000 /* Enable EEE in loopback */
#define E1000_EEER_LPI_FC 0x00040000 /* EEE Enable on FC */
/* SerDes Control */
diff --git a/drivers/net/ethernet/intel/igb/e1000_hw.h b/drivers/net/ethernet/intel/igb/e1000_hw.h
index f67cbd3..c2a51dc 100644
--- a/drivers/net/ethernet/intel/igb/e1000_hw.h
+++ b/drivers/net/ethernet/intel/igb/e1000_hw.h
@@ -63,6 +63,13 @@ struct e1000_hw;
#define E1000_DEV_ID_I350_FIBER 0x1522
#define E1000_DEV_ID_I350_SERDES 0x1523
#define E1000_DEV_ID_I350_SGMII 0x1524
+#define E1000_DEV_ID_I210_COPPER 0x1533
+#define E1000_DEV_ID_I210_COPPER_OEM1 0x1534
+#define E1000_DEV_ID_I210_COPPER_IT 0x1535
+#define E1000_DEV_ID_I210_FIBER 0x1536
+#define E1000_DEV_ID_I210_SERDES 0x1537
+#define E1000_DEV_ID_I210_SGMII 0x1538
+#define E1000_DEV_ID_I211_COPPER 0x1539
#define E1000_REVISION_2 2
#define E1000_REVISION_4 4
@@ -83,6 +90,8 @@ enum e1000_mac_type {
e1000_82576,
e1000_82580,
e1000_i350,
+ e1000_i210,
+ e1000_i211,
e1000_num_macs /* List is 1-based, so subtract 1 for true count. */
};
@@ -117,6 +126,7 @@ enum e1000_phy_type {
e1000_phy_igp_3,
e1000_phy_ife,
e1000_phy_82580,
+ e1000_phy_i210,
};
enum e1000_bus_type {
@@ -313,6 +323,9 @@ struct e1000_mac_operations {
void (*rar_set)(struct e1000_hw *, u8 *, u32);
s32 (*read_mac_addr)(struct e1000_hw *);
s32 (*get_speed_and_duplex)(struct e1000_hw *, u16 *, u16 *);
+ s32 (*acquire_swfw_sync)(struct e1000_hw *, u16);
+ void (*release_swfw_sync)(struct e1000_hw *, u16);
+
};
struct e1000_phy_operations {
@@ -338,6 +351,7 @@ struct e1000_nvm_operations {
s32 (*write)(struct e1000_hw *, u16, u16, u16 *);
s32 (*update)(struct e1000_hw *);
s32 (*validate)(struct e1000_hw *);
+ s32 (*valid_led_default)(struct e1000_hw *, u16 *);
};
struct e1000_info {
diff --git a/drivers/net/ethernet/intel/igb/e1000_i210.c b/drivers/net/ethernet/intel/igb/e1000_i210.c
new file mode 100644
index 0000000..77a5f93
--- /dev/null
+++ b/drivers/net/ethernet/intel/igb/e1000_i210.c
@@ -0,0 +1,603 @@
+/*******************************************************************************
+
+ Intel(R) Gigabit Ethernet Linux driver
+ Copyright(c) 2007-2012 Intel Corporation.
+
+ This program is free software; you can redistribute it and/or modify it
+ under the terms and conditions of the GNU General Public License,
+ version 2, as published by the Free Software Foundation.
+
+ This program is distributed in the hope it will be useful, but WITHOUT
+ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ more details.
+
+ You should have received a copy of the GNU General Public License along with
+ this program; if not, 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".
+
+ Contact Information:
+ e1000-devel Mailing List <e1000-devel@lists.sourceforge.net>
+ Intel Corporation, 5200 N.E. Elam Young Parkway, Hillsboro, OR 97124-6497
+
+******************************************************************************/
+
+/* e1000_i210
+ * e1000_i211
+ */
+
+#include <linux/types.h>
+#include <linux/if_ether.h>
+
+#include "e1000_hw.h"
+#include "e1000_i210.h"
+
+static s32 igb_get_hw_semaphore_i210(struct e1000_hw *hw);
+static void igb_put_hw_semaphore_i210(struct e1000_hw *hw);
+static s32 igb_write_nvm_srwr(struct e1000_hw *hw, u16 offset, u16 words,
+ u16 *data);
+static s32 igb_pool_flash_update_done_i210(struct e1000_hw *hw);
+
+/**
+ * igb_acquire_nvm_i210 - Request for access to EEPROM
+ * @hw: pointer to the HW structure
+ *
+ * Acquire the necessary semaphores for exclusive access to the EEPROM.
+ * Set the EEPROM access request bit and wait for EEPROM access grant bit.
+ * Return successful if access grant bit set, else clear the request for
+ * EEPROM access and return -E1000_ERR_NVM (-1).
+ **/
+s32 igb_acquire_nvm_i210(struct e1000_hw *hw)
+{
+ return igb_acquire_swfw_sync_i210(hw, E1000_SWFW_EEP_SM);
+}
+
+/**
+ * igb_release_nvm_i210 - Release exclusive access to EEPROM
+ * @hw: pointer to the HW structure
+ *
+ * Stop any current commands to the EEPROM and clear the EEPROM request bit,
+ * then release the semaphores acquired.
+ **/
+void igb_release_nvm_i210(struct e1000_hw *hw)
+{
+ igb_release_swfw_sync_i210(hw, E1000_SWFW_EEP_SM);
+}
+
+/**
+ * igb_acquire_swfw_sync_i210 - Acquire SW/FW semaphore
+ * @hw: pointer to the HW structure
+ * @mask: specifies which semaphore to acquire
+ *
+ * Acquire the SW/FW semaphore to access the PHY or NVM. The mask
+ * will also specify which port we're acquiring the lock for.
+ **/
+s32 igb_acquire_swfw_sync_i210(struct e1000_hw *hw, u16 mask)
+{
+ u32 swfw_sync;
+ u32 swmask = mask;
+ u32 fwmask = mask << 16;
+ s32 ret_val = E1000_SUCCESS;
+ s32 i = 0, timeout = 200; /* FIXME: find real value to use here */
+
+ while (i < timeout) {
+ if (igb_get_hw_semaphore_i210(hw)) {
+ ret_val = -E1000_ERR_SWFW_SYNC;
+ goto out;
+ }
+
+ swfw_sync = rd32(E1000_SW_FW_SYNC);
+ if (!(swfw_sync & fwmask))
+ break;
+
+ /*
+ * Firmware currently using resource (fwmask)
+ */
+ igb_put_hw_semaphore_i210(hw);
+ mdelay(5);
+ i++;
+ }
+
+ if (i == timeout) {
+ hw_dbg("Driver can't access resource, SW_FW_SYNC timeout.\n");
+ ret_val = -E1000_ERR_SWFW_SYNC;
+ goto out;
+ }
+
+ swfw_sync |= swmask;
+ wr32(E1000_SW_FW_SYNC, swfw_sync);
+
+ igb_put_hw_semaphore_i210(hw);
+out:
+ return ret_val;
+}
+
+/**
+ * igb_release_swfw_sync_i210 - Release SW/FW semaphore
+ * @hw: pointer to the HW structure
+ * @mask: specifies which semaphore to acquire
+ *
+ * Release the SW/FW semaphore used to access the PHY or NVM. The mask
+ * will also specify which port we're releasing the lock for.
+ **/
+void igb_release_swfw_sync_i210(struct e1000_hw *hw, u16 mask)
+{
+ u32 swfw_sync;
+
+ while (igb_get_hw_semaphore_i210(hw) != E1000_SUCCESS)
+ ; /* Empty */
+
+ swfw_sync = rd32(E1000_SW_FW_SYNC);
+ swfw_sync &= ~mask;
+ wr32(E1000_SW_FW_SYNC, swfw_sync);
+
+ igb_put_hw_semaphore_i210(hw);
+}
+
+/**
+ * igb_get_hw_semaphore_i210 - Acquire hardware semaphore
+ * @hw: pointer to the HW structure
+ *
+ * Acquire the HW semaphore to access the PHY or NVM
+ **/
+static s32 igb_get_hw_semaphore_i210(struct e1000_hw *hw)
+{
+ u32 swsm;
+ s32 ret_val = E1000_SUCCESS;
+ s32 timeout = hw->nvm.word_size + 1;
+ s32 i = 0;
+
+ /* Get the FW semaphore. */
+ for (i = 0; i < timeout; i++) {
+ swsm = rd32(E1000_SWSM);
+ wr32(E1000_SWSM, swsm | E1000_SWSM_SWESMBI);
+
+ /* Semaphore acquired if bit latched */
+ if (rd32(E1000_SWSM) & E1000_SWSM_SWESMBI)
+ break;
+
+ udelay(50);
+ }
+
+ if (i == timeout) {
+ /* Release semaphores */
+ igb_put_hw_semaphore(hw);
+ hw_dbg("Driver can't access the NVM\n");
+ ret_val = -E1000_ERR_NVM;
+ goto out;
+ }
+
+out:
+ return ret_val;
+}
+
+/**
+ * igb_put_hw_semaphore_i210 - Release hardware semaphore
+ * @hw: pointer to the HW structure
+ *
+ * Release hardware semaphore used to access the PHY or NVM
+ **/
+static void igb_put_hw_semaphore_i210(struct e1000_hw *hw)
+{
+ u32 swsm;
+
+ swsm = rd32(E1000_SWSM);
+
+ swsm &= ~E1000_SWSM_SWESMBI;
+
+ wr32(E1000_SWSM, swsm);
+}
+
+/**
+ * igb_read_nvm_srrd_i210 - Reads Shadow Ram using EERD register
+ * @hw: pointer to the HW structure
+ * @offset: offset of word in the Shadow Ram to read
+ * @words: number of words to read
+ * @data: word read from the Shadow Ram
+ *
+ * Reads a 16 bit word from the Shadow Ram using the EERD register.
+ * Uses necessary synchronization semaphores.
+ **/
+s32 igb_read_nvm_srrd_i210(struct e1000_hw *hw, u16 offset, u16 words,
+ u16 *data)
+{
+ s32 status = E1000_SUCCESS;
+ u16 i, count;
+
+ /* We cannot hold synchronization semaphores for too long,
+ * because of forceful takeover procedure. However it is more efficient
+ * to read in bursts than synchronizing access for each word. */
+ for (i = 0; i < words; i += E1000_EERD_EEWR_MAX_COUNT) {
+ count = (words - i) / E1000_EERD_EEWR_MAX_COUNT > 0 ?
+ E1000_EERD_EEWR_MAX_COUNT : (words - i);
+ if (hw->nvm.ops.acquire(hw) == E1000_SUCCESS) {
+ status = igb_read_nvm_eerd(hw, offset, count,
+ data + i);
+ hw->nvm.ops.release(hw);
+ } else {
+ status = E1000_ERR_SWFW_SYNC;
+ }
+
+ if (status != E1000_SUCCESS)
+ break;
+ }
+
+ return status;
+}
+
+/**
+ * igb_write_nvm_srwr_i210 - Write to Shadow RAM using EEWR
+ * @hw: pointer to the HW structure
+ * @offset: offset within the Shadow RAM to be written to
+ * @words: number of words to write
+ * @data: 16 bit word(s) to be written to the Shadow RAM
+ *
+ * Writes data to Shadow RAM at offset using EEWR register.
+ *
+ * If e1000_update_nvm_checksum is not called after this function , the
+ * data will not be committed to FLASH and also Shadow RAM will most likely
+ * contain an invalid checksum.
+ *
+ * If error code is returned, data and Shadow RAM may be inconsistent - buffer
+ * partially written.
+ **/
+s32 igb_write_nvm_srwr_i210(struct e1000_hw *hw, u16 offset, u16 words,
+ u16 *data)
+{
+ s32 status = E1000_SUCCESS;
+ u16 i, count;
+
+ /* We cannot hold synchronization semaphores for too long,
+ * because of forceful takeover procedure. However it is more efficient
+ * to write in bursts than synchronizing access for each word. */
+ for (i = 0; i < words; i += E1000_EERD_EEWR_MAX_COUNT) {
+ count = (words - i) / E1000_EERD_EEWR_MAX_COUNT > 0 ?
+ E1000_EERD_EEWR_MAX_COUNT : (words - i);
+ if (hw->nvm.ops.acquire(hw) == E1000_SUCCESS) {
+ status = igb_write_nvm_srwr(hw, offset, count,
+ data + i);
+ hw->nvm.ops.release(hw);
+ } else {
+ status = E1000_ERR_SWFW_SYNC;
+ }
+
+ if (status != E1000_SUCCESS)
+ break;
+ }
+
+ return status;
+}
+
+/**
+ * igb_write_nvm_srwr - Write to Shadow Ram using EEWR
+ * @hw: pointer to the HW structure
+ * @offset: offset within the Shadow Ram to be written to
+ * @words: number of words to write
+ * @data: 16 bit word(s) to be written to the Shadow Ram
+ *
+ * Writes data to Shadow Ram at offset using EEWR register.
+ *
+ * If igb_update_nvm_checksum is not called after this function , the
+ * Shadow Ram will most likely contain an invalid checksum.
+ **/
+static s32 igb_write_nvm_srwr(struct e1000_hw *hw, u16 offset, u16 words,
+ u16 *data)
+{
+ struct e1000_nvm_info *nvm = &hw->nvm;
+ u32 i, k, eewr = 0;
+ u32 attempts = 100000;
+ s32 ret_val = E1000_SUCCESS;
+
+ /*
+ * A check for invalid values: offset too large, too many words,
+ * too many words for the offset, and not enough words.
+ */
+ if ((offset >= nvm->word_size) || (words > (nvm->word_size - offset)) ||
+ (words == 0)) {
+ hw_dbg("nvm parameter(s) out of bounds\n");
+ ret_val = -E1000_ERR_NVM;
+ goto out;
+ }
+
+ for (i = 0; i < words; i++) {
+ eewr = ((offset+i) << E1000_NVM_RW_ADDR_SHIFT) |
+ (data[i] << E1000_NVM_RW_REG_DATA) |
+ E1000_NVM_RW_REG_START;
+
+ wr32(E1000_SRWR, eewr);
+
+ for (k = 0; k < attempts; k++) {
+ if (E1000_NVM_RW_REG_DONE &
+ rd32(E1000_SRWR)) {
+ ret_val = E1000_SUCCESS;
+ break;
+ }
+ udelay(5);
+ }
+
+ if (ret_val != E1000_SUCCESS) {
+ hw_dbg("Shadow RAM write EEWR timed out\n");
+ break;
+ }
+ }
+
+out:
+ return ret_val;
+}
+
+/**
+ * igb_read_nvm_i211 - Read NVM wrapper function for I211
+ * @hw: pointer to the HW structure
+ * @address: the word address (aka eeprom offset) to read
+ * @data: pointer to the data read
+ *
+ * Wrapper function to return data formerly found in the NVM.
+ **/
+s32 igb_read_nvm_i211(struct e1000_hw *hw, u16 offset, u16 words,
+ u16 *data)
+{
+ s32 ret_val = E1000_SUCCESS;
+
+ /* Only the MAC addr is required to be present in the iNVM */
+ switch (offset) {
+ case NVM_MAC_ADDR:
+ ret_val = igb_read_invm_i211(hw, offset, &data[0]);
+ ret_val |= igb_read_invm_i211(hw, offset+1, &data[1]);
+ ret_val |= igb_read_invm_i211(hw, offset+2, &data[2]);
+ if (ret_val != E1000_SUCCESS)
+ hw_dbg("MAC Addr not found in iNVM\n");
+ break;
+ case NVM_ID_LED_SETTINGS:
+ case NVM_INIT_CTRL_2:
+ case NVM_INIT_CTRL_4:
+ case NVM_LED_1_CFG:
+ case NVM_LED_0_2_CFG:
+ igb_read_invm_i211(hw, offset, data);
+ break;
+ case NVM_COMPAT:
+ *data = ID_LED_DEFAULT_I210;
+ break;
+ case NVM_SUB_DEV_ID:
+ *data = hw->subsystem_device_id;
+ break;
+ case NVM_SUB_VEN_ID:
+ *data = hw->subsystem_vendor_id;
+ break;
+ case NVM_DEV_ID:
+ *data = hw->device_id;
+ break;
+ case NVM_VEN_ID:
+ *data = hw->vendor_id;
+ break;
+ default:
+ hw_dbg("NVM word 0x%02x is not mapped.\n", offset);
+ *data = NVM_RESERVED_WORD;
+ break;
+ }
+ return ret_val;
+}
+
+/**
+ * igb_read_invm_i211 - Reads OTP
+ * @hw: pointer to the HW structure
+ * @address: the word address (aka eeprom offset) to read
+ * @data: pointer to the data read
+ *
+ * Reads 16-bit words from the OTP. Return error when the word is not
+ * stored in OTP.
+ **/
+s32 igb_read_invm_i211(struct e1000_hw *hw, u16 address, u16 *data)
+{
+ s32 status = -E1000_ERR_INVM_VALUE_NOT_FOUND;
+ u32 invm_dword;
+ u16 i;
+ u8 record_type, word_address;
+
+ for (i = 0; i < E1000_INVM_SIZE; i++) {
+ invm_dword = rd32(E1000_INVM_DATA_REG(i));
+ /* Get record type */
+ record_type = INVM_DWORD_TO_RECORD_TYPE(invm_dword);
+ if (record_type == E1000_INVM_UNINITIALIZED_STRUCTURE)
+ break;
+ if (record_type == E1000_INVM_CSR_AUTOLOAD_STRUCTURE)
+ i += E1000_INVM_CSR_AUTOLOAD_DATA_SIZE_IN_DWORDS;
+ if (record_type == E1000_INVM_RSA_KEY_SHA256_STRUCTURE)
+ i += E1000_INVM_RSA_KEY_SHA256_DATA_SIZE_IN_DWORDS;
+ if (record_type == E1000_INVM_WORD_AUTOLOAD_STRUCTURE) {
+ word_address = INVM_DWORD_TO_WORD_ADDRESS(invm_dword);
+ if (word_address == (u8)address) {
+ *data = INVM_DWORD_TO_WORD_DATA(invm_dword);
+ hw_dbg("Read INVM Word 0x%02x = %x",
+ address, *data);
+ status = E1000_SUCCESS;
+ break;
+ }
+ }
+ }
+ if (status != E1000_SUCCESS)
+ hw_dbg("Requested word 0x%02x not found in OTP\n", address);
+ return status;
+}
+
+/**
+ * igb_validate_nvm_checksum_i210 - Validate EEPROM checksum
+ * @hw: pointer to the HW structure
+ *
+ * Calculates the EEPROM checksum by reading/adding each word of the EEPROM
+ * and then verifies that the sum of the EEPROM is equal to 0xBABA.
+ **/
+s32 igb_validate_nvm_checksum_i210(struct e1000_hw *hw)
+{
+ s32 status = E1000_SUCCESS;
+ s32 (*read_op_ptr)(struct e1000_hw *, u16, u16, u16 *);
+
+ if (hw->nvm.ops.acquire(hw) == E1000_SUCCESS) {
+
+ /*
+ * Replace the read function with semaphore grabbing with
+ * the one that skips this for a while.
+ * We have semaphore taken already here.
+ */
+ read_op_ptr = hw->nvm.ops.read;
+ hw->nvm.ops.read = igb_read_nvm_eerd;
+
+ status = igb_validate_nvm_checksum(hw);
+
+ /* Revert original read operation. */
+ hw->nvm.ops.read = read_op_ptr;
+
+ hw->nvm.ops.release(hw);
+ } else {
+ status = E1000_ERR_SWFW_SYNC;
+ }
+
+ return status;
+}
+
+
+/**
+ * igb_update_nvm_checksum_i210 - Update EEPROM checksum
+ * @hw: pointer to the HW structure
+ *
+ * Updates the EEPROM checksum by reading/adding each word of the EEPROM
+ * up to the checksum. Then calculates the EEPROM checksum and writes the
+ * value to the EEPROM. Next commit EEPROM data onto the Flash.
+ **/
+s32 igb_update_nvm_checksum_i210(struct e1000_hw *hw)
+{
+ s32 ret_val = E1000_SUCCESS;
+ u16 checksum = 0;
+ u16 i, nvm_data;
+
+ /*
+ * Read the first word from the EEPROM. If this times out or fails, do
+ * not continue or we could be in for a very long wait while every
+ * EEPROM read fails
+ */
+ ret_val = igb_read_nvm_eerd(hw, 0, 1, &nvm_data);
+ if (ret_val != E1000_SUCCESS) {
+ hw_dbg("EEPROM read failed\n");
+ goto out;
+ }
+
+ if (hw->nvm.ops.acquire(hw) == E1000_SUCCESS) {
+ /*
+ * Do not use hw->nvm.ops.write, hw->nvm.ops.read
+ * because we do not want to take the synchronization
+ * semaphores twice here.
+ */
+
+ for (i = 0; i < NVM_CHECKSUM_REG; i++) {
+ ret_val = igb_read_nvm_eerd(hw, i, 1, &nvm_data);
+ if (ret_val) {
+ hw->nvm.ops.release(hw);
+ hw_dbg("NVM Read Error while updating checksum.\n");
+ goto out;
+ }
+ checksum += nvm_data;
+ }
+ checksum = (u16) NVM_SUM - checksum;
+ ret_val = igb_write_nvm_srwr(hw, NVM_CHECKSUM_REG, 1,
+ &checksum);
+ if (ret_val != E1000_SUCCESS) {
+ hw->nvm.ops.release(hw);
+ hw_dbg("NVM Write Error while updating checksum.\n");
+ goto out;
+ }
+
+ hw->nvm.ops.release(hw);
+
+ ret_val = igb_update_flash_i210(hw);
+ } else {
+ ret_val = -E1000_ERR_SWFW_SYNC;
+ }
+out:
+ return ret_val;
+}
+
+/**
+ * igb_update_flash_i210 - Commit EEPROM to the flash
+ * @hw: pointer to the HW structure
+ *
+ **/
+s32 igb_update_flash_i210(struct e1000_hw *hw)
+{
+ s32 ret_val = E1000_SUCCESS;
+ u32 flup;
+
+ ret_val = igb_pool_flash_update_done_i210(hw);
+ if (ret_val == -E1000_ERR_NVM) {
+ hw_dbg("Flash update time out\n");
+ goto out;
+ }
+
+ flup = rd32(E1000_EECD) | E1000_EECD_FLUPD_I210;
+ wr32(E1000_EECD, flup);
+
+ ret_val = igb_pool_flash_update_done_i210(hw);
+ if (ret_val == E1000_SUCCESS)
+ hw_dbg("Flash update complete\n");
+ else
+ hw_dbg("Flash update time out\n");
+
+out:
+ return ret_val;
+}
+
+/**
+ * igb_pool_flash_update_done_i210 - Pool FLUDONE status.
+ * @hw: pointer to the HW structure
+ *
+ **/
+s32 igb_pool_flash_update_done_i210(struct e1000_hw *hw)
+{
+ s32 ret_val = -E1000_ERR_NVM;
+ u32 i, reg;
+
+ for (i = 0; i < E1000_FLUDONE_ATTEMPTS; i++) {
+ reg = rd32(E1000_EECD);
+ if (reg & E1000_EECD_FLUDONE_I210) {
+ ret_val = E1000_SUCCESS;
+ break;
+ }
+ udelay(5);
+ }
+
+ return ret_val;
+}
+
+/**
+ * igb_valid_led_default_i210 - Verify a valid default LED config
+ * @hw: pointer to the HW structure
+ * @data: pointer to the NVM (EEPROM)
+ *
+ * Read the EEPROM for the current default LED configuration. If the
+ * LED configuration is not valid, set to a valid LED configuration.
+ **/
+s32 igb_valid_led_default_i210(struct e1000_hw *hw, u16 *data)
+{
+ s32 ret_val;
+
+ ret_val = hw->nvm.ops.read(hw, NVM_ID_LED_SETTINGS, 1, data);
+ if (ret_val) {
+ hw_dbg("NVM Read Error\n");
+ goto out;
+ }
+
+ if (*data == ID_LED_RESERVED_0000 || *data == ID_LED_RESERVED_FFFF) {
+ switch (hw->phy.media_type) {
+ case e1000_media_type_internal_serdes:
+ *data = ID_LED_DEFAULT_I210_SERDES;
+ break;
+ case e1000_media_type_copper:
+ default:
+ *data = ID_LED_DEFAULT_I210;
+ break;
+ }
+ }
+out:
+ return ret_val;
+}
diff --git a/drivers/net/ethernet/intel/igb/e1000_i210.h b/drivers/net/ethernet/intel/igb/e1000_i210.h
new file mode 100644
index 0000000..5dc2bd3
--- /dev/null
+++ b/drivers/net/ethernet/intel/igb/e1000_i210.h
@@ -0,0 +1,76 @@
+/*******************************************************************************
+
+ Intel(R) Gigabit Ethernet Linux driver
+ Copyright(c) 2007-2012 Intel Corporation.
+
+ This program is free software; you can redistribute it and/or modify it
+ under the terms and conditions of the GNU General Public License,
+ version 2, as published by the Free Software Foundation.
+
+ This program is distributed in the hope it will be useful, but WITHOUT
+ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ more details.
+
+ You should have received a copy of the GNU General Public License along with
+ this program; if not, 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".
+
+ Contact Information:
+ e1000-devel Mailing List <e1000-devel@lists.sourceforge.net>
+ Intel Corporation, 5200 N.E. Elam Young Parkway, Hillsboro, OR 97124-6497
+
+*******************************************************************************/
+
+#ifndef _E1000_I210_H_
+#define _E1000_I210_H_
+
+extern s32 igb_update_flash_i210(struct e1000_hw *hw);
+extern s32 igb_update_nvm_checksum_i210(struct e1000_hw *hw);
+extern s32 igb_validate_nvm_checksum_i210(struct e1000_hw *hw);
+extern s32 igb_write_nvm_srwr_i210(struct e1000_hw *hw, u16 offset,
+ u16 words, u16 *data);
+extern s32 igb_read_nvm_srrd_i210(struct e1000_hw *hw, u16 offset,
+ u16 words, u16 *data);
+extern s32 igb_read_invm_i211(struct e1000_hw *hw, u16 address, u16 *data);
+extern s32 igb_acquire_swfw_sync_i210(struct e1000_hw *hw, u16 mask);
+extern void igb_release_swfw_sync_i210(struct e1000_hw *hw, u16 mask);
+extern s32 igb_acquire_nvm_i210(struct e1000_hw *hw);
+extern void igb_release_nvm_i210(struct e1000_hw *hw);
+extern s32 igb_valid_led_default_i210(struct e1000_hw *hw, u16 *data);
+extern s32 igb_read_nvm_i211(struct e1000_hw *hw, u16 offset, u16 words,
+ u16 *data);
+
+#define E1000_STM_OPCODE 0xDB00
+#define E1000_EEPROM_FLASH_SIZE_WORD 0x11
+
+#define INVM_DWORD_TO_RECORD_TYPE(invm_dword) \
+ (u8)((invm_dword) & 0x7)
+#define INVM_DWORD_TO_WORD_ADDRESS(invm_dword) \
+ (u8)(((invm_dword) & 0x0000FE00) >> 9)
+#define INVM_DWORD_TO_WORD_DATA(invm_dword) \
+ (u16)(((invm_dword) & 0xFFFF0000) >> 16)
+
+enum E1000_INVM_STRUCTURE_TYPE {
+ E1000_INVM_UNINITIALIZED_STRUCTURE = 0x00,
+ E1000_INVM_WORD_AUTOLOAD_STRUCTURE = 0x01,
+ E1000_INVM_CSR_AUTOLOAD_STRUCTURE = 0x02,
+ E1000_INVM_PHY_REGISTER_AUTOLOAD_STRUCTURE = 0x03,
+ E1000_INVM_RSA_KEY_SHA256_STRUCTURE = 0x04,
+ E1000_INVM_INVALIDATED_STRUCTURE = 0x0F,
+};
+
+#define E1000_INVM_RSA_KEY_SHA256_DATA_SIZE_IN_DWORDS 8
+#define E1000_INVM_CSR_AUTOLOAD_DATA_SIZE_IN_DWORDS 1
+
+#define ID_LED_DEFAULT_I210 ((ID_LED_OFF1_ON2 << 8) | \
+ (ID_LED_OFF1_OFF2 << 4) | \
+ (ID_LED_DEF1_DEF2))
+#define ID_LED_DEFAULT_I210_SERDES ((ID_LED_DEF1_DEF2 << 8) | \
+ (ID_LED_DEF1_DEF2 << 4) | \
+ (ID_LED_DEF1_DEF2))
+
+#endif
diff --git a/drivers/net/ethernet/intel/igb/e1000_mac.c b/drivers/net/ethernet/intel/igb/e1000_mac.c
index f57338a..819c145 100644
--- a/drivers/net/ethernet/intel/igb/e1000_mac.c
+++ b/drivers/net/ethernet/intel/igb/e1000_mac.c
@@ -658,6 +658,7 @@ s32 igb_setup_link(struct e1000_hw *hw)
ret_val = igb_set_fc_watermarks(hw);
out:
+
return ret_val;
}
diff --git a/drivers/net/ethernet/intel/igb/e1000_nvm.c b/drivers/net/ethernet/intel/igb/e1000_nvm.c
index fa2c6ba..aa5fcdf 100644
--- a/drivers/net/ethernet/intel/igb/e1000_nvm.c
+++ b/drivers/net/ethernet/intel/igb/e1000_nvm.c
@@ -710,4 +710,3 @@ s32 igb_update_nvm_checksum(struct e1000_hw *hw)
out:
return ret_val;
}
-
diff --git a/drivers/net/ethernet/intel/igb/e1000_phy.c b/drivers/net/ethernet/intel/igb/e1000_phy.c
index 789de5b..7be98b6 100644
--- a/drivers/net/ethernet/intel/igb/e1000_phy.c
+++ b/drivers/net/ethernet/intel/igb/e1000_phy.c
@@ -35,6 +35,7 @@ static s32 igb_phy_setup_autoneg(struct e1000_hw *hw);
static void igb_phy_force_speed_duplex_setup(struct e1000_hw *hw,
u16 *phy_ctrl);
static s32 igb_wait_autoneg(struct e1000_hw *hw);
+static s32 igb_set_master_slave_mode(struct e1000_hw *hw);
/* Cable length tables */
static const u16 e1000_m88_cable_length_table[] =
@@ -570,6 +571,11 @@ s32 igb_copper_link_setup_m88(struct e1000_hw *hw)
hw_dbg("Error committing the PHY changes\n");
goto out;
}
+ if (phy->type == e1000_phy_i210) {
+ ret_val = igb_set_master_slave_mode(hw);
+ if (ret_val)
+ return ret_val;
+ }
out:
return ret_val;
@@ -1213,12 +1219,22 @@ s32 igb_phy_force_speed_duplex_m88(struct e1000_hw *hw)
goto out;
if (!link) {
- if (hw->phy.type != e1000_phy_m88 ||
- hw->phy.id == I347AT4_E_PHY_ID ||
- hw->phy.id == M88E1112_E_PHY_ID) {
+ bool reset_dsp = true;
+
+ switch (hw->phy.id) {
+ case I347AT4_E_PHY_ID:
+ case M88E1112_E_PHY_ID:
+ case I210_I_PHY_ID:
+ reset_dsp = false;
+ break;
+ default:
+ if (hw->phy.type != e1000_phy_m88)
+ reset_dsp = false;
+ break;
+ }
+ if (!reset_dsp)
hw_dbg("Link taking longer than expected.\n");
- } else {
-
+ else {
/*
* We didn't get link.
* Reset the DSP and cross our fingers.
@@ -1243,7 +1259,8 @@ s32 igb_phy_force_speed_duplex_m88(struct e1000_hw *hw)
if (hw->phy.type != e1000_phy_m88 ||
hw->phy.id == I347AT4_E_PHY_ID ||
- hw->phy.id == M88E1112_E_PHY_ID)
+ hw->phy.id == M88E1112_E_PHY_ID ||
+ hw->phy.id == I210_I_PHY_ID)
goto out;
ret_val = phy->ops.read_reg(hw, M88E1000_EXT_PHY_SPEC_CTRL, &phy_data);
@@ -1441,6 +1458,7 @@ s32 igb_check_downshift(struct e1000_hw *hw)
u16 phy_data, offset, mask;
switch (phy->type) {
+ case e1000_phy_i210:
case e1000_phy_m88:
case e1000_phy_gg82563:
offset = M88E1000_PHY_SPEC_STATUS;
@@ -1476,7 +1494,7 @@ out:
*
* Polarity is determined based on the PHY specific status register.
**/
-static s32 igb_check_polarity_m88(struct e1000_hw *hw)
+s32 igb_check_polarity_m88(struct e1000_hw *hw)
{
struct e1000_phy_info *phy = &hw->phy;
s32 ret_val;
@@ -1665,6 +1683,7 @@ s32 igb_get_cable_length_m88_gen2(struct e1000_hw *hw)
u16 phy_data, phy_data2, index, default_page, is_cm;
switch (hw->phy.id) {
+ case I210_I_PHY_ID:
case I347AT4_E_PHY_ID:
/* Remember the original page select and set it to 7 */
ret_val = phy->ops.read_reg(hw, I347AT4_PAGE_SELECT,
@@ -2129,10 +2148,16 @@ s32 igb_phy_init_script_igp3(struct e1000_hw *hw)
void igb_power_up_phy_copper(struct e1000_hw *hw)
{
u16 mii_reg = 0;
+ u16 power_reg = 0;
/* The PHY will retain its settings across a power down/up cycle */
hw->phy.ops.read_reg(hw, PHY_CONTROL, &mii_reg);
mii_reg &= ~MII_CR_POWER_DOWN;
+ if (hw->phy.type == e1000_phy_i210) {
+ hw->phy.ops.read_reg(hw, GS40G_COPPER_SPEC, &power_reg);
+ power_reg &= ~GS40G_CS_POWER_DOWN;
+ hw->phy.ops.write_reg(hw, GS40G_COPPER_SPEC, power_reg);
+ }
hw->phy.ops.write_reg(hw, PHY_CONTROL, mii_reg);
}
@@ -2146,10 +2171,18 @@ void igb_power_up_phy_copper(struct e1000_hw *hw)
void igb_power_down_phy_copper(struct e1000_hw *hw)
{
u16 mii_reg = 0;
+ u16 power_reg = 0;
/* The PHY will retain its settings across a power down/up cycle */
hw->phy.ops.read_reg(hw, PHY_CONTROL, &mii_reg);
mii_reg |= MII_CR_POWER_DOWN;
+
+ /* i210 Phy requires an additional bit for power up/down */
+ if (hw->phy.type == e1000_phy_i210) {
+ hw->phy.ops.read_reg(hw, GS40G_COPPER_SPEC, &power_reg);
+ power_reg |= GS40G_CS_POWER_DOWN;
+ hw->phy.ops.write_reg(hw, GS40G_COPPER_SPEC, power_reg);
+ }
hw->phy.ops.write_reg(hw, PHY_CONTROL, mii_reg);
msleep(1);
}
@@ -2345,3 +2378,103 @@ s32 igb_get_cable_length_82580(struct e1000_hw *hw)
out:
return ret_val;
}
+
+/**
+ * igb_write_phy_reg_gs40g - Write GS40G PHY register
+ * @hw: pointer to the HW structure
+ * @offset: lower half is register offset to write to
+ * upper half is page to use.
+ * @data: data to write at register offset
+ *
+ * Acquires semaphore, if necessary, then writes the data to PHY register
+ * at the offset. Release any acquired semaphores before exiting.
+ **/
+s32 igb_write_phy_reg_gs40g(struct e1000_hw *hw, u32 offset, u16 data)
+{
+ s32 ret_val;
+ u16 page = offset >> GS40G_PAGE_SHIFT;
+
+ offset = offset & GS40G_OFFSET_MASK;
+ ret_val = hw->phy.ops.acquire(hw);
+ if (ret_val)
+ return ret_val;
+
+ ret_val = igb_write_phy_reg_mdic(hw, GS40G_PAGE_SELECT, page);
+ if (ret_val)
+ goto release;
+ ret_val = igb_write_phy_reg_mdic(hw, offset, data);
+
+release:
+ hw->phy.ops.release(hw);
+ return ret_val;
+}
+
+/**
+ * igb_read_phy_reg_gs40g - Read GS40G PHY register
+ * @hw: pointer to the HW structure
+ * @offset: lower half is register offset to read to
+ * upper half is page to use.
+ * @data: data to read at register offset
+ *
+ * Acquires semaphore, if necessary, then reads the data in the PHY register
+ * at the offset. Release any acquired semaphores before exiting.
+ **/
+s32 igb_read_phy_reg_gs40g(struct e1000_hw *hw, u32 offset, u16 *data)
+{
+ s32 ret_val;
+ u16 page = offset >> GS40G_PAGE_SHIFT;
+
+ offset = offset & GS40G_OFFSET_MASK;
+ ret_val = hw->phy.ops.acquire(hw);
+ if (ret_val)
+ return ret_val;
+
+ ret_val = igb_write_phy_reg_mdic(hw, GS40G_PAGE_SELECT, page);
+ if (ret_val)
+ goto release;
+ ret_val = igb_read_phy_reg_mdic(hw, offset, data);
+
+release:
+ hw->phy.ops.release(hw);
+ return ret_val;
+}
+
+/**
+ * igb_set_master_slave_mode - Setup PHY for Master/slave mode
+ * @hw: pointer to the HW structure
+ *
+ * Sets up Master/slave mode
+ **/
+static s32 igb_set_master_slave_mode(struct e1000_hw *hw)
+{
+ s32 ret_val;
+ u16 phy_data;
+
+ /* Resolve Master/Slave mode */
+ ret_val = hw->phy.ops.read_reg(hw, PHY_1000T_CTRL, &phy_data);
+ if (ret_val)
+ return ret_val;
+
+ /* load defaults for future use */
+ hw->phy.original_ms_type = (phy_data & CR_1000T_MS_ENABLE) ?
+ ((phy_data & CR_1000T_MS_VALUE) ?
+ e1000_ms_force_master :
+ e1000_ms_force_slave) : e1000_ms_auto;
+
+ switch (hw->phy.ms_type) {
+ case e1000_ms_force_master:
+ phy_data |= (CR_1000T_MS_ENABLE | CR_1000T_MS_VALUE);
+ break;
+ case e1000_ms_force_slave:
+ phy_data |= CR_1000T_MS_ENABLE;
+ phy_data &= ~(CR_1000T_MS_VALUE);
+ break;
+ case e1000_ms_auto:
+ phy_data &= ~CR_1000T_MS_ENABLE;
+ /* fall-through */
+ default:
+ break;
+ }
+
+ return hw->phy.ops.write_reg(hw, PHY_1000T_CTRL, phy_data);
+}
diff --git a/drivers/net/ethernet/intel/igb/e1000_phy.h b/drivers/net/ethernet/intel/igb/e1000_phy.h
index 84172ea..34e4061 100644
--- a/drivers/net/ethernet/intel/igb/e1000_phy.h
+++ b/drivers/net/ethernet/intel/igb/e1000_phy.h
@@ -73,6 +73,9 @@ s32 igb_copper_link_setup_82580(struct e1000_hw *hw);
s32 igb_get_phy_info_82580(struct e1000_hw *hw);
s32 igb_phy_force_speed_duplex_82580(struct e1000_hw *hw);
s32 igb_get_cable_length_82580(struct e1000_hw *hw);
+s32 igb_read_phy_reg_gs40g(struct e1000_hw *hw, u32 offset, u16 *data);
+s32 igb_write_phy_reg_gs40g(struct e1000_hw *hw, u32 offset, u16 data);
+s32 igb_check_polarity_m88(struct e1000_hw *hw);
/* IGP01E1000 Specific Registers */
#define IGP01E1000_PHY_PORT_CONFIG 0x10 /* Port Config */
@@ -140,4 +143,16 @@ s32 igb_get_cable_length_82580(struct e1000_hw *hw);
#define E1000_CABLE_LENGTH_UNDEFINED 0xFF
+/* GS40G - I210 PHY defines */
+#define GS40G_PAGE_SELECT 0x16
+#define GS40G_PAGE_SHIFT 16
+#define GS40G_OFFSET_MASK 0xFFFF
+#define GS40G_PAGE_2 0x20000
+#define GS40G_MAC_REG2 0x15
+#define GS40G_MAC_LB 0x4140
+#define GS40G_MAC_SPEED_1G 0X0006
+#define GS40G_COPPER_SPEC 0x0010
+#define GS40G_CS_POWER_DOWN 0x0002
+#define GS40G_LINE_LB 0x4000
+
#endif
diff --git a/drivers/net/ethernet/intel/igb/e1000_regs.h b/drivers/net/ethernet/intel/igb/e1000_regs.h
index ccdf36d..35d1e4f 100644
--- a/drivers/net/ethernet/intel/igb/e1000_regs.h
+++ b/drivers/net/ethernet/intel/igb/e1000_regs.h
@@ -352,4 +352,18 @@
#define E1000_O2BGPTC 0x08FE4 /* OS2BMC packets received by BMC */
#define E1000_O2BSPC 0x0415C /* OS2BMC packets transmitted by host */
+#define E1000_SRWR 0x12018 /* Shadow Ram Write Register - RW */
+#define E1000_I210_FLMNGCTL 0x12038
+#define E1000_I210_FLMNGDATA 0x1203C
+#define E1000_I210_FLMNGCNT 0x12040
+
+#define E1000_I210_FLSWCTL 0x12048
+#define E1000_I210_FLSWDATA 0x1204C
+#define E1000_I210_FLSWCNT 0x12050
+
+#define E1000_I210_FLA 0x1201C
+
+#define E1000_INVM_DATA_REG(_n) (0x12120 + 4*(_n))
+#define E1000_INVM_SIZE 64 /* Number of INVM Data Registers */
+
#endif
diff --git a/drivers/net/ethernet/intel/igb/igb.h b/drivers/net/ethernet/intel/igb/igb.h
index 3758ad2..ae6d3f3 100644
--- a/drivers/net/ethernet/intel/igb/igb.h
+++ b/drivers/net/ethernet/intel/igb/igb.h
@@ -65,10 +65,13 @@ struct igb_adapter;
#define MAX_Q_VECTORS 8
/* Transmit and receive queues */
-#define IGB_MAX_RX_QUEUES (adapter->vfs_allocated_count ? 2 : \
- (hw->mac.type > e1000_82575 ? 8 : 4))
+#define IGB_MAX_RX_QUEUES ((adapter->vfs_allocated_count ? 2 : \
+ (hw->mac.type > e1000_82575 ? 8 : 4)))
+#define IGB_MAX_RX_QUEUES_I210 4
+#define IGB_MAX_RX_QUEUES_I211 2
#define IGB_MAX_TX_QUEUES 16
-
+#define IGB_MAX_TX_QUEUES_I210 4
+#define IGB_MAX_TX_QUEUES_I211 2
#define IGB_MAX_VF_MC_ENTRIES 30
#define IGB_MAX_VF_FUNCTIONS 8
#define IGB_MAX_VFTA_ENTRIES 128
diff --git a/drivers/net/ethernet/intel/igb/igb_ethtool.c b/drivers/net/ethernet/intel/igb/igb_ethtool.c
index faadf23..812d4f9 100644
--- a/drivers/net/ethernet/intel/igb/igb_ethtool.c
+++ b/drivers/net/ethernet/intel/igb/igb_ethtool.c
@@ -552,10 +552,13 @@ static void igb_get_regs(struct net_device *netdev,
regs_buff[548] = rd32(E1000_TDFT);
regs_buff[549] = rd32(E1000_TDFHS);
regs_buff[550] = rd32(E1000_TDFPC);
- regs_buff[551] = adapter->stats.o2bgptc;
- regs_buff[552] = adapter->stats.b2ospc;
- regs_buff[553] = adapter->stats.o2bspc;
- regs_buff[554] = adapter->stats.b2ogprc;
+
+ if (hw->mac.type > e1000_82580) {
+ regs_buff[551] = adapter->stats.o2bgptc;
+ regs_buff[552] = adapter->stats.b2ospc;
+ regs_buff[553] = adapter->stats.o2bspc;
+ regs_buff[554] = adapter->stats.b2ogprc;
+ }
if (hw->mac.type != e1000_82576)
return;
@@ -660,6 +663,9 @@ static int igb_set_eeprom(struct net_device *netdev,
if (eeprom->len == 0)
return -EOPNOTSUPP;
+ if (hw->mac.type == e1000_i211)
+ return -EOPNOTSUPP;
+
if (eeprom->magic != (hw->vendor_id | (hw->device_id << 16)))
return -EFAULT;
@@ -887,6 +893,36 @@ struct igb_reg_test {
#define TABLE64_TEST_LO 5
#define TABLE64_TEST_HI 6
+/* i210 reg test */
+static struct igb_reg_test reg_test_i210[] = {
+ { E1000_FCAL, 0x100, 1, PATTERN_TEST, 0xFFFFFFFF, 0xFFFFFFFF },
+ { E1000_FCAH, 0x100, 1, PATTERN_TEST, 0x0000FFFF, 0xFFFFFFFF },
+ { E1000_FCT, 0x100, 1, PATTERN_TEST, 0x0000FFFF, 0xFFFFFFFF },
+ { E1000_RDBAL(0), 0x100, 4, PATTERN_TEST, 0xFFFFFF80, 0xFFFFFFFF },
+ { E1000_RDBAH(0), 0x100, 4, PATTERN_TEST, 0xFFFFFFFF, 0xFFFFFFFF },
+ { E1000_RDLEN(0), 0x100, 4, PATTERN_TEST, 0x000FFF80, 0x000FFFFF },
+ /* RDH is read-only for i210, only test RDT. */
+ { E1000_RDT(0), 0x100, 4, PATTERN_TEST, 0x0000FFFF, 0x0000FFFF },
+ { E1000_FCRTH, 0x100, 1, PATTERN_TEST, 0x0000FFF0, 0x0000FFF0 },
+ { E1000_FCTTV, 0x100, 1, PATTERN_TEST, 0x0000FFFF, 0x0000FFFF },
+ { E1000_TIPG, 0x100, 1, PATTERN_TEST, 0x3FFFFFFF, 0x3FFFFFFF },
+ { E1000_TDBAL(0), 0x100, 4, PATTERN_TEST, 0xFFFFFF80, 0xFFFFFFFF },
+ { E1000_TDBAH(0), 0x100, 4, PATTERN_TEST, 0xFFFFFFFF, 0xFFFFFFFF },
+ { E1000_TDLEN(0), 0x100, 4, PATTERN_TEST, 0x000FFF80, 0x000FFFFF },
+ { E1000_TDT(0), 0x100, 4, PATTERN_TEST, 0x0000FFFF, 0x0000FFFF },
+ { E1000_RCTL, 0x100, 1, SET_READ_TEST, 0xFFFFFFFF, 0x00000000 },
+ { E1000_RCTL, 0x100, 1, SET_READ_TEST, 0x04CFB0FE, 0x003FFFFB },
+ { E1000_RCTL, 0x100, 1, SET_READ_TEST, 0x04CFB0FE, 0xFFFFFFFF },
+ { E1000_TCTL, 0x100, 1, SET_READ_TEST, 0xFFFFFFFF, 0x00000000 },
+ { E1000_RA, 0, 16, TABLE64_TEST_LO,
+ 0xFFFFFFFF, 0xFFFFFFFF },
+ { E1000_RA, 0, 16, TABLE64_TEST_HI,
+ 0x900FFFFF, 0xFFFFFFFF },
+ { E1000_MTA, 0, 128, TABLE32_TEST,
+ 0xFFFFFFFF, 0xFFFFFFFF },
+ { 0, 0, 0, 0, 0 }
+};
+
/* i350 reg test */
static struct igb_reg_test reg_test_i350[] = {
{ E1000_FCAL, 0x100, 1, PATTERN_TEST, 0xFFFFFFFF, 0xFFFFFFFF },
@@ -1109,6 +1145,11 @@ static int igb_reg_test(struct igb_adapter *adapter, u64 *data)
test = reg_test_i350;
toggle = 0x7FEFF3FF;
break;
+ case e1000_i210:
+ case e1000_i211:
+ test = reg_test_i210;
+ toggle = 0x7FEFF3FF;
+ break;
case e1000_82580:
test = reg_test_82580;
toggle = 0x7FEFF3FF;
@@ -1190,23 +1231,13 @@ static int igb_reg_test(struct igb_adapter *adapter, u64 *data)
static int igb_eeprom_test(struct igb_adapter *adapter, u64 *data)
{
- u16 temp;
- u16 checksum = 0;
- u16 i;
-
*data = 0;
- /* Read and add up the contents of the EEPROM */
- for (i = 0; i < (NVM_CHECKSUM_REG + 1); i++) {
- if ((adapter->hw.nvm.ops.read(&adapter->hw, i, 1, &temp)) < 0) {
- *data = 1;
- break;
- }
- checksum += temp;
- }
- /* If Checksum is not Correct return error else test passed */
- if ((checksum != (u16) NVM_SUM) && !(*data))
- *data = 2;
+ /* Validate eeprom on all parts but i211 */
+ if (adapter->hw.mac.type != e1000_i211) {
+ if (adapter->hw.nvm.ops.validate(&adapter->hw) < 0)
+ *data = 2;
+ }
return *data;
}
@@ -1272,6 +1303,8 @@ static int igb_intr_test(struct igb_adapter *adapter, u64 *data)
ics_mask = 0x77DCFED5;
break;
case e1000_i350:
+ case e1000_i210:
+ case e1000_i211:
ics_mask = 0x77DCFED5;
break;
default:
@@ -1438,23 +1471,35 @@ static int igb_integrated_phy_loopback(struct igb_adapter *adapter)
{
struct e1000_hw *hw = &adapter->hw;
u32 ctrl_reg = 0;
+ u16 phy_reg = 0;
hw->mac.autoneg = false;
- if (hw->phy.type == e1000_phy_m88) {
+ switch (hw->phy.type) {
+ case e1000_phy_m88:
/* Auto-MDI/MDIX Off */
igb_write_phy_reg(hw, M88E1000_PHY_SPEC_CTRL, 0x0808);
/* reset to update Auto-MDI/MDIX */
igb_write_phy_reg(hw, PHY_CONTROL, 0x9140);
/* autoneg off */
igb_write_phy_reg(hw, PHY_CONTROL, 0x8140);
- } else if (hw->phy.type == e1000_phy_82580) {
+ break;
+ case e1000_phy_82580:
/* enable MII loopback */
igb_write_phy_reg(hw, I82580_PHY_LBK_CTRL, 0x8041);
+ break;
+ case e1000_phy_i210:
+ /* set loopback speed in PHY */
+ igb_read_phy_reg(hw, (GS40G_PAGE_SELECT & GS40G_PAGE_2),
+ &phy_reg);
+ phy_reg |= GS40G_MAC_SPEED_1G;
+ igb_write_phy_reg(hw, (GS40G_PAGE_SELECT & GS40G_PAGE_2),
+ phy_reg);
+ ctrl_reg = rd32(E1000_CTRL_EXT);
+ default:
+ break;
}
- ctrl_reg = rd32(E1000_CTRL);
-
/* force 1000, set loopback */
igb_write_phy_reg(hw, PHY_CONTROL, 0x4140);
@@ -1467,7 +1512,7 @@ static int igb_integrated_phy_loopback(struct igb_adapter *adapter)
E1000_CTRL_FD | /* Force Duplex to FULL */
E1000_CTRL_SLU); /* Set link up enable bit */
- if (hw->phy.type == e1000_phy_m88)
+ if ((hw->phy.type == e1000_phy_m88) || (hw->phy.type == e1000_phy_i210))
ctrl_reg |= E1000_CTRL_ILOS; /* Invert Loss of Signal */
wr32(E1000_CTRL, ctrl_reg);
@@ -1475,7 +1520,7 @@ static int igb_integrated_phy_loopback(struct igb_adapter *adapter)
/* Disable the receiver on the PHY so when a cable is plugged in, the
* PHY does not begin to autoneg when a cable is reconnected to the NIC.
*/
- if (hw->phy.type == e1000_phy_m88)
+ if ((hw->phy.type == e1000_phy_m88) || (hw->phy.type == e1000_phy_i210))
igb_phy_disable_receiver(adapter);
udelay(500);
@@ -1740,6 +1785,14 @@ static int igb_loopback_test(struct igb_adapter *adapter, u64 *data)
*data = 0;
goto out;
}
+ if ((adapter->hw.mac.type == e1000_i210)
+ || (adapter->hw.mac.type == e1000_i210)) {
+ dev_err(&adapter->pdev->dev,
+ "Loopback test not supported "
+ "on this part at this time.\n");
+ *data = 0;
+ goto out;
+ }
*data = igb_setup_desc_rings(adapter);
if (*data)
goto out;
diff --git a/drivers/net/ethernet/intel/igb/igb_main.c b/drivers/net/ethernet/intel/igb/igb_main.c
index 80d52d2..ba21f9c 100644
--- a/drivers/net/ethernet/intel/igb/igb_main.c
+++ b/drivers/net/ethernet/intel/igb/igb_main.c
@@ -75,6 +75,11 @@ static const struct e1000_info *igb_info_tbl[] = {
};
static DEFINE_PCI_DEVICE_TABLE(igb_pci_tbl) = {
+ { PCI_VDEVICE(INTEL, E1000_DEV_ID_I211_COPPER), board_82575 },
+ { PCI_VDEVICE(INTEL, E1000_DEV_ID_I210_COPPER), board_82575 },
+ { PCI_VDEVICE(INTEL, E1000_DEV_ID_I210_FIBER), board_82575 },
+ { PCI_VDEVICE(INTEL, E1000_DEV_ID_I210_SERDES), board_82575 },
+ { PCI_VDEVICE(INTEL, E1000_DEV_ID_I210_SGMII), board_82575 },
{ PCI_VDEVICE(INTEL, E1000_DEV_ID_I350_COPPER), board_82575 },
{ PCI_VDEVICE(INTEL, E1000_DEV_ID_I350_FIBER), board_82575 },
{ PCI_VDEVICE(INTEL, E1000_DEV_ID_I350_SERDES), board_82575 },
@@ -641,6 +646,8 @@ static void igb_cache_ring_register(struct igb_adapter *adapter)
case e1000_82575:
case e1000_82580:
case e1000_i350:
+ case e1000_i210:
+ case e1000_i211:
default:
for (; i < adapter->num_rx_queues; i++)
adapter->rx_ring[i]->reg_idx = rbase_offset + i;
@@ -727,8 +734,11 @@ static int igb_alloc_queues(struct igb_adapter *adapter)
if (adapter->hw.mac.type >= e1000_82576)
set_bit(IGB_RING_FLAG_RX_SCTP_CSUM, &ring->flags);
- /* On i350, loopback VLAN packets have the tag byte-swapped. */
- if (adapter->hw.mac.type == e1000_i350)
+ /*
+ * On i350, i210, and i211, loopback VLAN packets
+ * have the tag byte-swapped.
+ * */
+ if (adapter->hw.mac.type >= e1000_i350)
set_bit(IGB_RING_FLAG_RX_LB_VLAN_BSWAP, &ring->flags);
adapter->rx_ring[i] = ring;
@@ -822,6 +832,8 @@ static void igb_assign_vector(struct igb_q_vector *q_vector, int msix_vector)
break;
case e1000_82580:
case e1000_i350:
+ case e1000_i210:
+ case e1000_i211:
/*
* On 82580 and newer adapters the scheme is similar to 82576
* however instead of ordering column-major we have things
@@ -888,6 +900,8 @@ static void igb_configure_msix(struct igb_adapter *adapter)
case e1000_82576:
case e1000_82580:
case e1000_i350:
+ case e1000_i210:
+ case e1000_i211:
/* Turn on MSI-X capability first, or our settings
* won't stick. And it will take days to debug. */
wr32(E1000_GPIE, E1000_GPIE_MSIX_MODE |
@@ -1034,6 +1048,11 @@ static int igb_set_interrupt_capability(struct igb_adapter *adapter)
if (!(adapter->flags & IGB_FLAG_QUEUE_PAIRS))
numvecs += adapter->num_tx_queues;
+ /* i210 and i211 can only have 4 MSIX vectors for rx/tx queues. */
+ if ((adapter->hw.mac.type == e1000_i210)
+ || (adapter->hw.mac.type == e1000_i211))
+ numvecs = 4;
+
/* store the number of vectors reserved for queues */
adapter->num_q_vectors = numvecs;
@@ -1041,6 +1060,7 @@ static int igb_set_interrupt_capability(struct igb_adapter *adapter)
numvecs++;
adapter->msix_entries = kcalloc(numvecs, sizeof(struct msix_entry),
GFP_KERNEL);
+
if (!adapter->msix_entries)
goto msi_only;
@@ -1631,6 +1651,8 @@ void igb_reset(struct igb_adapter *adapter)
pba &= E1000_RXPBS_SIZE_MASK_82576;
break;
case e1000_82575:
+ case e1000_i210:
+ case e1000_i211:
default:
pba = E1000_PBA_34K;
break;
@@ -1826,7 +1848,7 @@ static int __devinit igb_probe(struct pci_dev *pdev,
*/
if (pdev->is_virtfn) {
WARN(1, KERN_ERR "%s (%hx:%hx) should not be a VF!\n",
- pci_name(pdev), pdev->vendor, pdev->device);
+ pci_name(pdev), pdev->vendor, pdev->device);
return -EINVAL;
}
@@ -1980,11 +2002,16 @@ static int __devinit igb_probe(struct pci_dev *pdev,
* known good starting state */
hw->mac.ops.reset_hw(hw);
- /* make sure the NVM is good */
- if (hw->nvm.ops.validate(hw) < 0) {
- dev_err(&pdev->dev, "The NVM Checksum Is Not Valid\n");
- err = -EIO;
- goto err_eeprom;
+ /*
+ * make sure the NVM is good , i211 parts have special NVM that
+ * doesn't contain a checksum
+ */
+ if (hw->mac.type != e1000_i211) {
+ if (hw->nvm.ops.validate(hw) < 0) {
+ dev_err(&pdev->dev, "The NVM Checksum Is Not Valid\n");
+ err = -EIO;
+ goto err_eeprom;
+ }
}
/* copy the MAC address out of the NVM */
@@ -2118,6 +2145,8 @@ static int __devinit igb_probe(struct pci_dev *pdev,
adapter->num_rx_queues, adapter->num_tx_queues);
switch (hw->mac.type) {
case e1000_i350:
+ case e1000_i210:
+ case e1000_i211:
igb_set_eee_i350(hw);
break;
default:
@@ -2244,9 +2273,14 @@ static void __devinit igb_probe_vfs(struct igb_adapter * adapter)
{
#ifdef CONFIG_PCI_IOV
struct pci_dev *pdev = adapter->pdev;
+ struct e1000_hw *hw = &adapter->hw;
int old_vfs = igb_find_enabled_vfs(adapter);
int i;
+ /* Virtualization features not supported on i210 family. */
+ if ((hw->mac.type == e1000_i210) || (hw->mac.type == e1000_i211))
+ return;
+
if (old_vfs) {
dev_info(&pdev->dev, "%d pre-allocated VFs found - override "
"max_vfs setting of %d\n", old_vfs, max_vfs);
@@ -2258,6 +2292,7 @@ static void __devinit igb_probe_vfs(struct igb_adapter * adapter)
adapter->vf_data = kcalloc(adapter->vfs_allocated_count,
sizeof(struct vf_data_storage), GFP_KERNEL);
+
/* if allocation failed then we do not support SR-IOV */
if (!adapter->vf_data) {
adapter->vfs_allocated_count = 0;
@@ -2332,11 +2367,28 @@ static int __devinit igb_sw_init(struct igb_adapter *adapter)
} else
adapter->vfs_allocated_count = max_vfs;
break;
+ case e1000_i210:
+ case e1000_i211:
+ adapter->vfs_allocated_count = 0;
+ break;
default:
break;
}
#endif /* CONFIG_PCI_IOV */
- adapter->rss_queues = min_t(u32, IGB_MAX_RX_QUEUES, num_online_cpus());
+ switch (hw->mac.type) {
+ case e1000_i210:
+ adapter->rss_queues = min_t(u32, IGB_MAX_RX_QUEUES_I210,
+ num_online_cpus());
+ break;
+ case e1000_i211:
+ adapter->rss_queues = min_t(u32, IGB_MAX_RX_QUEUES_I211,
+ num_online_cpus());
+ break;
+ default:
+ adapter->rss_queues = min_t(u32, IGB_MAX_RX_QUEUES,
+ num_online_cpus());
+ break;
+ }
/* i350 cannot do RSS and SR-IOV at the same time */
if (hw->mac.type == e1000_i350 && adapter->vfs_allocated_count)
adapter->rss_queues = 1;
@@ -2366,7 +2418,7 @@ static int __devinit igb_sw_init(struct igb_adapter *adapter)
/* Explicitly disable IRQ since the NIC can be in any state. */
igb_irq_disable(adapter);
- if (hw->mac.type == e1000_i350)
+ if (hw->mac.type >= e1000_i350)
adapter->flags &= ~IGB_FLAG_DMAC;
set_bit(__IGB_DOWN, &adapter->state);
@@ -2819,6 +2871,17 @@ static void igb_setup_mrqc(struct igb_adapter *adapter)
/* Don't need to set TUOFL or IPOFL, they default to 1 */
wr32(E1000_RXCSUM, rxcsum);
+ /*
+ * Generate RSS hash based on TCP port numbers and/or
+ * IPv4/v6 src and dst addresses since UDP cannot be
+ * hashed reliably due to IP fragmentation
+ */
+
+ mrqc = E1000_MRQC_RSS_FIELD_IPV4 |
+ E1000_MRQC_RSS_FIELD_IPV4_TCP |
+ E1000_MRQC_RSS_FIELD_IPV6 |
+ E1000_MRQC_RSS_FIELD_IPV6_TCP |
+ E1000_MRQC_RSS_FIELD_IPV6_TCP_EX;
/* If VMDq is enabled then we set the appropriate mode for that, else
* we default to RSS so that an RSS hash is calculated per packet even
@@ -2834,25 +2897,15 @@ static void igb_setup_mrqc(struct igb_adapter *adapter)
wr32(E1000_VT_CTL, vtctl);
}
if (adapter->rss_queues > 1)
- mrqc = E1000_MRQC_ENABLE_VMDQ_RSS_2Q;
+ mrqc |= E1000_MRQC_ENABLE_VMDQ_RSS_2Q;
else
- mrqc = E1000_MRQC_ENABLE_VMDQ;
+ mrqc |= E1000_MRQC_ENABLE_VMDQ;
} else {
- mrqc = E1000_MRQC_ENABLE_RSS_4Q;
+ if (hw->mac.type != e1000_i211)
+ mrqc |= E1000_MRQC_ENABLE_RSS_4Q;
}
igb_vmm_control(adapter);
- /*
- * Generate RSS hash based on TCP port numbers and/or
- * IPv4/v6 src and dst addresses since UDP cannot be
- * hashed reliably due to IP fragmentation
- */
- mrqc |= E1000_MRQC_RSS_FIELD_IPV4 |
- E1000_MRQC_RSS_FIELD_IPV4_TCP |
- E1000_MRQC_RSS_FIELD_IPV6 |
- E1000_MRQC_RSS_FIELD_IPV6_TCP |
- E1000_MRQC_RSS_FIELD_IPV6_TCP_EX;
-
wr32(E1000_MRQC, mrqc);
}
@@ -3454,7 +3507,7 @@ static void igb_set_rx_mode(struct net_device *netdev)
* we will have issues with VLAN tag stripping not being done for frames
* that are only arriving because we are the default pool
*/
- if (hw->mac.type < e1000_82576)
+ if ((hw->mac.type < e1000_82576) || (hw->mac.type > e1000_i350))
return;
vmolr |= rd32(E1000_VMOLR(vfn)) &
@@ -3551,7 +3604,7 @@ static bool igb_thermal_sensor_event(struct e1000_hw *hw, u32 event)
bool ret = false;
u32 ctrl_ext, thstat;
- /* check for thermal sensor event on i350, copper only */
+ /* check for thermal sensor event on i350 copper only */
if (hw->mac.type == e1000_i350) {
thstat = rd32(E1000_THSTAT);
ctrl_ext = rd32(E1000_CTRL_EXT);
@@ -7027,6 +7080,8 @@ static void igb_vmm_control(struct igb_adapter *adapter)
switch (hw->mac.type) {
case e1000_82575:
+ case e1000_i210:
+ case e1000_i211:
default:
/* replication is not supported for 82575 */
return;
diff --git a/drivers/net/ethernet/intel/igb/igb_ptp.c b/drivers/net/ethernet/intel/igb/igb_ptp.c
index c9b71c5..d5ee7fa 100644
--- a/drivers/net/ethernet/intel/igb/igb_ptp.c
+++ b/drivers/net/ethernet/intel/igb/igb_ptp.c
@@ -262,6 +262,8 @@ void igb_ptp_init(struct igb_adapter *adapter)
struct e1000_hw *hw = &adapter->hw;
switch (hw->mac.type) {
+ case e1000_i210:
+ case e1000_i211:
case e1000_i350:
case e1000_82580:
adapter->caps.owner = THIS_MODULE;
@@ -362,6 +364,8 @@ void igb_systim_to_hwtstamp(struct igb_adapter *adapter,
unsigned long flags;
switch (adapter->hw.mac.type) {
+ case e1000_i210:
+ case e1000_i211:
case e1000_i350:
case e1000_82580:
case e1000_82576:
--
1.7.7.6
^ permalink raw reply related
* [net-next 1/2] igb: Add function and pointers for 82580 low power state settings.
From: Jeff Kirsher @ 2012-05-12 20:30 UTC (permalink / raw)
To: davem; +Cc: Carolyn Wyborny, netdev, gospo, sassmann, Jeff Kirsher
In-Reply-To: <1336854637-6708-1-git-send-email-jeffrey.t.kirsher@intel.com>
From: Carolyn Wyborny <carolyn.wyborny@intel.com>
82580 and later parts did not have low power setting functions. This patch
adds the specific functions, pointers and assignments for these low
power settings.
Signed-off-by: Carolyn Wyborny <carolyn.wyborny@intel.com>
Tested-by: Jeff Pieper <jeffrey.e.pieper@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
drivers/net/ethernet/intel/igb/e1000_82575.c | 94 ++++++++++++++++++++++++++
drivers/net/ethernet/intel/igb/e1000_phy.h | 7 ++
2 files changed, 101 insertions(+), 0 deletions(-)
diff --git a/drivers/net/ethernet/intel/igb/e1000_82575.c b/drivers/net/ethernet/intel/igb/e1000_82575.c
index 08bdc33..b945992 100644
--- a/drivers/net/ethernet/intel/igb/e1000_82575.c
+++ b/drivers/net/ethernet/intel/igb/e1000_82575.c
@@ -52,6 +52,8 @@ static s32 igb_write_phy_reg_82580(struct e1000_hw *, u32, u16);
static s32 igb_reset_hw_82575(struct e1000_hw *);
static s32 igb_reset_hw_82580(struct e1000_hw *);
static s32 igb_set_d0_lplu_state_82575(struct e1000_hw *, bool);
+static s32 igb_set_d0_lplu_state_82580(struct e1000_hw *, bool);
+static s32 igb_set_d3_lplu_state_82580(struct e1000_hw *, bool);
static s32 igb_setup_copper_link_82575(struct e1000_hw *);
static s32 igb_setup_serdes_link_82575(struct e1000_hw *);
static s32 igb_write_phy_reg_sgmii_82575(struct e1000_hw *, u32, u16);
@@ -359,6 +361,8 @@ static s32 igb_get_invariants_82575(struct e1000_hw *hw)
phy->ops.force_speed_duplex = igb_phy_force_speed_duplex_82580;
phy->ops.get_cable_length = igb_get_cable_length_82580;
phy->ops.get_phy_info = igb_get_phy_info_82580;
+ phy->ops.set_d0_lplu_state = igb_set_d0_lplu_state_82580;
+ phy->ops.set_d3_lplu_state = igb_set_d3_lplu_state_82580;
break;
default:
return -E1000_ERR_PHY;
@@ -674,6 +678,96 @@ out:
}
/**
+ * igb_set_d0_lplu_state_82580 - Set Low Power Linkup D0 state
+ * @hw: pointer to the HW structure
+ * @active: true to enable LPLU, false to disable
+ *
+ * Sets the LPLU D0 state according to the active flag. When
+ * activating LPLU this function also disables smart speed
+ * and vice versa. LPLU will not be activated unless the
+ * device autonegotiation advertisement meets standards of
+ * either 10 or 10/100 or 10/100/1000 at all duplexes.
+ * This is a function pointer entry point only called by
+ * PHY setup routines.
+ **/
+static s32 igb_set_d0_lplu_state_82580(struct e1000_hw *hw, bool active)
+{
+ struct e1000_phy_info *phy = &hw->phy;
+ s32 ret_val = 0;
+ u16 data;
+
+ data = rd32(E1000_82580_PHY_POWER_MGMT);
+
+ if (active) {
+ data |= E1000_82580_PM_D0_LPLU;
+
+ /* When LPLU is enabled, we should disable SmartSpeed */
+ data &= ~E1000_82580_PM_SPD;
+ } else {
+ data &= ~E1000_82580_PM_D0_LPLU;
+
+ /*
+ * LPLU and SmartSpeed are mutually exclusive. LPLU is used
+ * during Dx states where the power conservation is most
+ * important. During driver activity we should enable
+ * SmartSpeed, so performance is maintained.
+ */
+ if (phy->smart_speed == e1000_smart_speed_on)
+ data |= E1000_82580_PM_SPD;
+ else if (phy->smart_speed == e1000_smart_speed_off)
+ data &= ~E1000_82580_PM_SPD; }
+
+ wr32(E1000_82580_PHY_POWER_MGMT, data);
+ return ret_val;
+}
+
+/**
+ * igb_set_d3_lplu_state_82580 - Sets low power link up state for D3
+ * @hw: pointer to the HW structure
+ * @active: boolean used to enable/disable lplu
+ *
+ * Success returns 0, Failure returns 1
+ *
+ * The low power link up (lplu) state is set to the power management level D3
+ * and SmartSpeed is disabled when active is true, else clear lplu for D3
+ * and enable Smartspeed. LPLU and Smartspeed are mutually exclusive. LPLU
+ * is used during Dx states where the power conservation is most important.
+ * During driver activity, SmartSpeed should be enabled so performance is
+ * maintained.
+ **/
+s32 igb_set_d3_lplu_state_82580(struct e1000_hw *hw, bool active)
+{
+ struct e1000_phy_info *phy = &hw->phy;
+ s32 ret_val = 0;
+ u16 data;
+
+ data = rd32(E1000_82580_PHY_POWER_MGMT);
+
+ if (!active) {
+ data &= ~E1000_82580_PM_D3_LPLU;
+ /*
+ * LPLU and SmartSpeed are mutually exclusive. LPLU is used
+ * during Dx states where the power conservation is most
+ * important. During driver activity we should enable
+ * SmartSpeed, so performance is maintained.
+ */
+ if (phy->smart_speed == e1000_smart_speed_on)
+ data |= E1000_82580_PM_SPD;
+ else if (phy->smart_speed == e1000_smart_speed_off)
+ data &= ~E1000_82580_PM_SPD;
+ } else if ((phy->autoneg_advertised == E1000_ALL_SPEED_DUPLEX) ||
+ (phy->autoneg_advertised == E1000_ALL_NOT_GIG) ||
+ (phy->autoneg_advertised == E1000_ALL_10_SPEED)) {
+ data |= E1000_82580_PM_D3_LPLU;
+ /* When LPLU is enabled, we should disable SmartSpeed */
+ data &= ~E1000_82580_PM_SPD;
+ }
+
+ wr32(E1000_82580_PHY_POWER_MGMT, data);
+ return ret_val;
+}
+
+/**
* igb_acquire_nvm_82575 - Request for access to EEPROM
* @hw: pointer to the HW structure
*
diff --git a/drivers/net/ethernet/intel/igb/e1000_phy.h b/drivers/net/ethernet/intel/igb/e1000_phy.h
index 4c32ac6..84172ea 100644
--- a/drivers/net/ethernet/intel/igb/e1000_phy.h
+++ b/drivers/net/ethernet/intel/igb/e1000_phy.h
@@ -114,6 +114,13 @@ s32 igb_get_cable_length_82580(struct e1000_hw *hw);
/* I82580 PHY Diagnostics Status */
#define I82580_DSTATUS_CABLE_LENGTH 0x03FC
#define I82580_DSTATUS_CABLE_LENGTH_SHIFT 2
+
+/* 82580 PHY Power Management */
+#define E1000_82580_PHY_POWER_MGMT 0xE14
+#define E1000_82580_PM_SPD 0x0001 /* Smart Power Down */
+#define E1000_82580_PM_D0_LPLU 0x0002 /* For D0a states */
+#define E1000_82580_PM_D3_LPLU 0x0004 /* For all other states */
+
/* Enable flexible speed on link-up */
#define IGP02E1000_PM_D0_LPLU 0x0002 /* For D0a states */
#define IGP02E1000_PM_D3_LPLU 0x0004 /* For all other states */
--
1.7.7.6
^ permalink raw reply related
* [net-next v3 0/2][pull request] Intel Wired LAN Driver Updates
From: Jeff Kirsher @ 2012-05-12 20:30 UTC (permalink / raw)
To: davem; +Cc: Jeff Kirsher, netdev, gospo, sassmann
This series of patches contains updates for igb. Most notably,
the addition of i210 & i211 support.
v2: fix up patch 2 based on feedback from David Miller
v3: fix compiler warning
The following are changes since commit 4b549a2ef4bef9965d97cbd992ba67930cd3e0fe:
fq_codel: Fair Queue Codel AQM
and are available in the git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-next master
Carolyn Wyborny (2):
igb: Add function and pointers for 82580 low power state settings.
igb: Add Support for new i210/i211 devices.
drivers/net/ethernet/intel/igb/Makefile | 3 +-
drivers/net/ethernet/intel/igb/e1000_82575.c | 276 +++++++++--
drivers/net/ethernet/intel/igb/e1000_82575.h | 3 +-
drivers/net/ethernet/intel/igb/e1000_defines.h | 33 ++
drivers/net/ethernet/intel/igb/e1000_hw.h | 14 +
drivers/net/ethernet/intel/igb/e1000_i210.c | 603 ++++++++++++++++++++++++
drivers/net/ethernet/intel/igb/e1000_i210.h | 76 +++
drivers/net/ethernet/intel/igb/e1000_mac.c | 1 +
drivers/net/ethernet/intel/igb/e1000_nvm.c | 1 -
drivers/net/ethernet/intel/igb/e1000_phy.c | 147 ++++++-
drivers/net/ethernet/intel/igb/e1000_phy.h | 22 +
drivers/net/ethernet/intel/igb/e1000_regs.h | 14 +
drivers/net/ethernet/intel/igb/igb.h | 9 +-
drivers/net/ethernet/intel/igb/igb_ethtool.c | 103 +++-
drivers/net/ethernet/intel/igb/igb_main.c | 81 +++-
drivers/net/ethernet/intel/igb/igb_ptp.c | 4 +
16 files changed, 1293 insertions(+), 97 deletions(-)
create mode 100644 drivers/net/ethernet/intel/igb/e1000_i210.c
create mode 100644 drivers/net/ethernet/intel/igb/e1000_i210.h
--
1.7.7.6
^ permalink raw reply
* Re: [PATCH v2 net-next] fq_codel: Fair Queue Codel AQM
From: David Miller @ 2012-05-12 19:55 UTC (permalink / raw)
To: eric.dumazet
Cc: xiaosuo, netdev, dave.taht, nichols, van, therbert, mattmathis,
ycheng, shemminger, maze, nanditad
In-Reply-To: <1336774576.31653.289.camel@edumazet-glaptop>
From: Eric Dumazet <eric.dumazet@gmail.com>
Date: Sat, 12 May 2012 00:16:16 +0200
> On Fri, 2012-05-11 at 21:30 +0200, Eric Dumazet wrote:
>> From: Eric Dumazet <edumazet@google.com>
>
> ...
>
>> +static int fq_codel_dump_class_stats(struct Qdisc *sch, unsigned long cl,
>> + struct gnet_dump *d)
>> +{
>> + struct fq_codel_sched_data *q = qdisc_priv(sch);
>> + u32 idx = cl - 1;
>> + struct gnet_stats_queue qs = { 0 };
>> + struct tc_fq_codel_xstats xstats;
>> +
>> + WARN_ON_ONCE(1);
>> + if (idx < q->flows_cnt) {
>> + const struct fq_codel_flow *flow = &q->flows[idx];
>> + const struct sk_buff *skb = flow->head;
>
> Oh well, I forgot to remove this WARN_ON_ONCE(1)
I applied this with the WARN_ON_ONCE(1) removed but there was another
problem.
When you include ping output in your commit message that "---" string
told GIT that this was the end of the commit message when in fact
there was more content including your signoff.
I caught it and fixed it up, but please be more mindful of this in
the future.
Thanks.
^ permalink raw reply
* Re: [PATCH net-next] codel: use Newton method instead of sqrt() and divides
From: David Miller @ 2012-05-12 19:52 UTC (permalink / raw)
To: eric.dumazet
Cc: dave.taht, netdev, nichols, van, codel, ycheng, mattmathis,
therbert, shemminger, nanditad
In-Reply-To: <1336829533.31653.1108.camel@edumazet-glaptop>
From: Eric Dumazet <eric.dumazet@gmail.com>
Date: Sat, 12 May 2012 15:32:13 +0200
> From: Eric Dumazet <edumazet@google.com>
>
> As Van pointed out, interval/sqrt(count) can be implemented using
> multiplies only.
>
> http://en.wikipedia.org/wiki/Methods_of_computing_square_roots#Iterative_methods_for_reciprocal_square_roots
>
> This patch implements the Newton method and reciprocal divide.
>
> Total cost is 15 cycles instead of 120 on my Corei5 machine (64bit
> kernel).
>
> There is a small 'error' for count values < 5, but we don't really care.
>
> I reuse a hole in struct codel_vars :
> - pack the dropping boolean into one bit
> - use 31bit to store the reciprocal value of sqrt(count).
>
> Suggested-by: Van Jacobson <van@pollere.net>
> Signed-off-by: Eric Dumazet <edumazet@google.com>
Applied but I never like that bitfield sharing for real integers.
GCC makes a complete mess of it as it extracts and inserts the
integer value into that bit field. You are guarenteed to get
better code if you do this by hand in a full u32.
Either that or just bite the bullet and use a completely seperate
field, maybe we'll need more boolean states later.
^ 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