From: Chris Leech <[EMAIL PROTECTED]>
Date: Tue, 20 Dec 2005 23:10:07 -0800
> That could be a good way to deal with it. Actually, I should double
> check the length of tcp_skb_cb. I took a quick look and thought that
> there might be some room left there anyway, even though the comment in
> tcp
From: Eric Dumazet <[EMAIL PROTECTED]>
Date: Wed, 21 Dec 2005 07:20:43 +0100
> Please consider not enlarging cb[] if not CONFIG_NET_DMA ?
I also disagree with putting this thing in the cb[] at all.
Just put the dma_cookie_t explicitly into struct sk_buff,
protected by CONFIG_NET_DMA.
-
To unsubs
On Tue, Dec 20, 2005 at 09:17:34PM -0800, Chris Leech ([EMAIL PROTECTED]) wrote:
> DMA memcpy subsystem
> Provides an API for offloading memory copies to DMA devices.
> Along with client registration and DMA channel allocation, the main APIs are:
> dma_async_memcpy_buf_to_buf()
> dma_as
Sorry, bounced from vger the first time due to formatting
-- Forwarded message --
From: Chris Leech <[EMAIL PROTECTED]>
Date: Dec 20, 2005 11:05 PM
Subject: Re: [RFC][PATCH 4/5] I/OAT DMA support and TCP acceleration
To: Eric Dumazet <[EMAIL PROTECTED]>
Cc: lkml , netdev
, "Grover,
Chris Leech a écrit :
Structure changes for TCP recv offload to I/OAT
Adds an async_wait_queue and some additional fields to tcp_sock, a
copied_early flag to sb_buff and a dma_cookie_t to tcp_skb_cb
Renames cleanup_rbuf to tcp_cleanup_rbuf and makes it non-static so we
can call it from tcp_inpu
DMA memcpy subsystem
Provides an API for offloading memory copies to DMA devices.
Along with client registration and DMA channel allocation, the main APIs are:
dma_async_memcpy_buf_to_buf()
dma_async_memcpy_buf_to_pg()
dma_async_memcpy_pg_to_pg()
dma_async_memcpy_com
Structure changes for TCP recv offload to I/OAT
Adds an async_wait_queue and some additional fields to tcp_sock, a
copied_early flag to sb_buff and a dma_cookie_t to tcp_skb_cb
Renames cleanup_rbuf to tcp_cleanup_rbuf and makes it non-static so we
can call it from tcp_input.c
---
include/linu
Utility functions for offloading sk_buff to iovec copies
Provides for pinning user space pages in memory, copying to iovecs, and
copying from sk_buffs including fragmented and chained sk_buffs.
---
net/core/user_dma.c | 410 ++
net/core/Makefile
Setup the networking subsystem as a DMA client
Attempts to allocate per-CPU DMA channels
---
dev.c | 97 ++
1 files changed, 97 insertions(+)
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -113,6 +113,7 @@
#include /* N
TCP recv offload to I/OAT
Locks down user pages and sets up for the DMA in tcp_recvmsg, then calls
dma_async_try_early_copy in tcp_v4_do_rcv
---
tcp.c | 166 +---
tcp_input.c | 63 +-
tcp_ipv4.c | 16 +
3
Following up on the I/OAT patches that Andy posted on Nov 23, these
address some of the style concerns, add descriptive comments (kdoc
style) to many functions, remove some dead code, and most importantly
include our TCP recv offload changes.
This patch set does not include the driver for the I/OAT
Jens Osterkamp wrote:
Christoph Hellwig <[EMAIL PROTECTED]> wrote on 12/15/2005 02:40:06 PM:
On Thu, Dec 15, 2005 at 01:47:25PM +0100, Jens Osterkamp wrote:
The driver incorrectly used dma_addr_t to describe
HW structures and consequently broke when that type
was changed in 2.6.15-rc.
This
Christoph Hellwig <[EMAIL PROTECTED]> wrote on 12/15/2005 02:40:06 PM:
> On Thu, Dec 15, 2005 at 01:47:25PM +0100, Jens Osterkamp wrote:
> >
> > The driver incorrectly used dma_addr_t to describe
> > HW structures and consequently broke when that type
> > was changed in 2.6.15-rc.
> >
> > This cha
Number of new features for 2.6.16.
* limited ethtool
* set hardware address of bridge pseudo-interface
* speed detection when link carrier changes
--
Stephen Hemminger <[EMAIL PROTECTED]>
OSDL http://developer.osdl.org/~shemminger
-
To unsubscribe from this list: send the
Some people are using bridging to hide multiple machines from an ISP
that restricts by MAC address. So in that case allow the bridge mac
address to be set to any of the existing interfaces. I don't want to
allow any arbitrary value and confuse STP.
Signed-off-by: Stephen Hemminger <[EMAIL PROTECT
Add version info to bridge module.
Signed-off-by: Stephen Hemminger <[EMAIL PROTECTED]>
--- br-2.6.16.orig/net/bridge/br.c
+++ br-2.6.16/net/bridge/br.c
@@ -67,3 +67,4 @@ EXPORT_SYMBOL(br_should_route_hook);
module_init(br_init)
module_exit(br_deinit)
MODULE_LICENSE("GPL");
+MODULE_VERSION(BR_
Speed of a interface may not be available until carrier
is detected in the case of autonegotiation. To get the correct value
we need to recheck speed after carrier event. But the check needs to
be done in a context that is similar to normal ethtool interface (can sleep).
Also, delay check for 1ms
Add limited ethtool support to bridge to allow disabling
features.
Note: if underlying device does not support a feature (like checksum
offload), then the bridge device won't inherit it.
Signed-off-by: Stephen Hemminger <[EMAIL PROTECTED]>
--- br-2.6.16.orig/net/bridge/br_device.c
+++ br-2.6.16
While in the learning state, run filters but drop the result.
This prevents us from acquiring bad fdb entries in learning state.
Signed-off-by: Stephen Hemminger <[EMAIL PROTECTED]>
--- br-2.6.orig/net/bridge/br_input.c
+++ br-2.6/net/bridge/br_input.c
@@ -53,6 +53,11 @@ int br_handle_frame_finis
Version number change.
Signed-off-by: Stephen Hemminger <[EMAIL PROTECTED]>
--- sky2.orig/drivers/net/sky2.c
+++ sky2/drivers/net/sky2.c
@@ -57,7 +57,7 @@
#include "sky2.h"
#define DRV_NAME "sky2"
-#define DRV_VERSION"0.10"
+#define DRV_VERSION"0.11"
#de
It is possible for hardware to get confused when an oversized frame
is received. In that case, just drop the packet and increment a counter.
Signed-off-by: Stephen Hemminger <[EMAIL PROTECTED]>
--- sky2.orig/drivers/net/sky2.c
+++ sky2/drivers/net/sky2.c
@@ -1694,6 +1694,9 @@ static struct sk_buf
Don't die if we run out of memory on mtu or ring parameter change.
For other admin operations, don't rebuild Rx ring, just restart the PHY.
Signed-off-by: Stephen Hemminger <[EMAIL PROTECTED]>
--- sky2-netdev.orig/drivers/net/sky2.c
+++ sky2-netdev/drivers/net/sky2.c
@@ -481,6 +481,14 @@ static
Don't lose multicast addresses when link goes down or ring
parameters change.
Signed-off-by: Stephen Hemminger <[EMAIL PROTECTED]>
--- sky2-netdev.orig/drivers/net/sky2.c
+++ sky2-netdev/drivers/net/sky2.c
@@ -1478,8 +1478,6 @@ static void sky2_link_down(struct sky2_p
Incorporate error handling and fixes from skge 0.13 into sky2.
--
Stephen Hemminger <[EMAIL PROTECTED]>
OSDL http://developer.osdl.org/~shemminger
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.
Please apply to net-2.6 for 2.6.15 tree.
I hope to actually change this behaviour shortly but this will help
anybody grepping code at present.
Signed-off-by: Ian McDonald <[EMAIL PROTECTED]>
---
diff --git a/net/dccp/ipv4.c b/net/dccp/ipv4.c
index ca03521..656e13e 100644
--- a/net/dccp/ipv4.c
+++
Here is a new feature for netem in 2.6.16. It adds the ability to
randomly corrupt packets with netem. A version was done by
Hagen Paul Pfeifer, but I redid it to handle the cases of backwards
compatibility with netlink interface and presence of hardware checksum
offload. It is useful for testing
From: jamal <[EMAIL PROTECTED]>
Date: Tue, 20 Dec 2005 08:47:12 -0500
> On Mon, 2005-19-12 at 23:38 -0800, David S. Miller wrote:
> > Every time we build a bundle, we add each xfrm_dst into a list
> > hung off of the xfrm_state.
> >
> "build a bundle" as in when the policy tmpl is resolved?
Cor
From: Chris Elmquist <[EMAIL PROTECTED]>
Date: Tue, 20 Dec 2005 15:15:03 -0600
> Resubmitting after recommendation to use GET_REG32_1() instead of
> GET_REG32_LOOP(..., 1). Retested. Problem remains fixed.
>
> Prevent tg3_get_regs() from reading reserved and undocumented registers
> at RX_CPU_B
Resubmitting after recommendation to use GET_REG32_1() instead of
GET_REG32_LOOP(..., 1). Retested. Problem remains fixed.
Prevent tg3_get_regs() from reading reserved and undocumented registers
at RX_CPU_BASE and TX_CPU_BASE offsets which caused hostile behavior
on PCIe platforms.
diff --git
On Tuesday (12/20/2005 at 10:48AM -0800), Michael Chan wrote:
>
> Yes, reading the reserved registers will cause timeouts on the PCIE bus
> and that will likely hang or cause NMIs.
yes, I can vouch for this ;-)
> I ACK this patch. But you should probably use GET_REG32_1() on the
> individual reg
On Tue, 2005-12-20 at 14:09 -0600, Chris Elmquist wrote:
> Howdy,
>
> I've been working with a four port PCIe card built from Broadcom 5714 with
> a 5704 behind the PCIx bridge that the 5714 externalizes. These cards
> are built by Silicom.
>
> When we would run 'ethtool -d' against either of th
Howdy,
I've been working with a four port PCIe card built from Broadcom 5714 with
a 5704 behind the PCIx bridge that the 5714 externalizes. These cards
are built by Silicom.
When we would run 'ethtool -d' against either of the secondary (ie,
5704's) interfaces, we would get all sorts of nasty PC
On Tue, 2005-20-12 at 06:04 +0100, Krzysztof Oledzki wrote:
>
> On Mon, 19 Dec 2005, jamal wrote:
>
> > On Mon, 2005-19-12 at 13:57 -0800, David S. Miller wrote:
> >> From: jamal <[EMAIL PROTECTED]>
> >> Date: Mon, 19 Dec 2005 08:17:19 -0500
> >>
> >>> Just an addendum: If this works it should be
On Mon, 2005-19-12 at 23:38 -0800, David S. Miller wrote:
> This patch below illustrates the general idea. It compiles,
> and that is all I guarentee :-)
>
> We add backpointers to the xfrm_dsts's that compose a bundle,
> back to the root of the bundle.
>
> Every time we build a bundle, we add ea
Ditto. I remember we had some discussion on this sometime back in the
netdev mailing list (Baruch was part of the discussion).
> -Original Message-
> From: David S. Miller [mailto:[EMAIL PROTECTED]
> Sent: Tuesday, December 20, 2005 4:09 AM
> To: [EMAIL PROTECTED]
> Cc: [EMAIL PROTECTED];
From: "Douglas Leith" <[EMAIL PROTECTED]>
Date: Tue, 20 Dec 2005 08:40:26 -
> Well some feedback to that effect might have been useful a while
> back Dave.
I gave him feedback on at least 5 seperate occaisions, both
publicly and in private correspondance.
Others have done so as well.
-
To un
36 matches
Mail list logo
