Chris Leech <[EMAIL PROTECTED]> wrote:
>
> Locks down user pages and sets up for DMA in tcp_recvmsg, then calls
> dma_async_try_early_copy in tcp_v4_do_rcv
>
+#ifdef CONFIG_NET_DMA
+#ifdef CONFIG_NET_DMA
+#ifdef CONFIG_NET_DMA
+#ifdef CONFIG_NET_DMA
+#ifdef CONFIG_NET_DMA
+#ifdef CONFIG_NET_DMA
+
Chris Leech <[EMAIL PROTECTED]> wrote:
>
> Any socket recv of less than this ammount will not be offloaded
>
> ...
>
> +int sysctl_tcp_dma_copybreak = NET_DMA_DEFAULT_COPYBREAK;
Is it appropriate that this tunable be kernel-wide, rather than more
finely-grained?
-
To unsubscribe from this list:
Chris Leech <[EMAIL PROTECTED]> wrote:
>
> +#ifdef CONFIG_NET_DMA
> +#include
> +#endif
Please move the ifdefs into the header and include it unconditionally
(entire patchset).
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to [EMAIL PROTECTED]
Mor
Chris Leech <[EMAIL PROTECTED]> wrote:
>
> +
> +#define NUM_PAGES_SPANNED(start, length) \
> + ((PAGE_ALIGN((unsigned long)start + length) - \
> + ((unsigned long)start & PAGE_MASK)) >> PAGE_SHIFT)
static inline all-lower-case functions are much nicer.
> +/*
> + * Lock down all the iovec
Andreas Schwab writes:
> I suppose the NIC in the PowerMac G5 can do GigE, yet when plugged into a
> GB switch it is only willing to talk 100MB with it. Any idea why? Kernel
> is 2.6.16-rc5-git2.
It does 1000Mb/s here...
# ethtool eth0
Settings for eth0:
Supported ports: [ TP MII ]
From: Evgeniy Polyakov <[EMAIL PROTECTED]>
Date: Sun, 5 Mar 2006 04:43:25 +0300
> According to investigation made for kevent based FS AIO reading,
> get_user_pages() performange graph looks like sqrt() function
> with plato starting on about 64-80 pages on Xeon 2.4Ghz with 1Gb of ram,
> while memc
On Sat, Mar 04, 2006 at 01:41:44PM -0800, David S. Miller ([EMAIL PROTECTED])
wrote:
> From: Jan Engelhardt <[EMAIL PROTECTED]>
> Date: Sat, 4 Mar 2006 19:46:22 +0100 (MET)
>
> > Does this buy the normal standard desktop user anything?
>
> Absolutely, it optimizes end-node performance.
It reall
On Friday 03 March 2006 16:26, Adrian Bunk wrote:
> Two remarks regarding the new IPW_QOS option:
> - it should be named IPW2200_QOS (similar to the other IPW2200_*
> options)
done.
> - please add a help text
i could add some stuff about WMM to its help text, but I think someone more
involved
On Fri, Mar 03, 2006 at 01:42:36PM -0800, Chris Leech wrote:
> diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
> index 13abfa2..b792048 100644
> --- a/net/ipv4/tcp.c
> +++ b/net/ipv4/tcp.c
> @@ -262,6 +262,9 @@
> #include
> #include
> #include
> +#ifdef CONFIG_NET_DMA
> +#include
> +#endif
#if
Alexey Dobriyan wrote:
From: Eric Sesterhenn <[EMAIL PROTECTED]>
memset() is called before check.
Signed-off-by: Eric Sesterhenn <[EMAIL PROTECTED]>
Signed-off-by: Alexey Dobriyan <[EMAIL PROTECTED]>
---
drivers/net/chelsio/espi.c |4 +---
1 file changed, 1 insertion(+), 3 deletions(-)
-
From: Eric Sesterhenn <[EMAIL PROTECTED]>
memset() is called before check.
Signed-off-by: Eric Sesterhenn <[EMAIL PROTECTED]>
Signed-off-by: Alexey Dobriyan <[EMAIL PROTECTED]>
---
drivers/net/chelsio/espi.c |4 +---
1 file changed, 1 insertion(+), 3 deletions(-)
--- a/drivers/net/chelsio/
From: Jan Engelhardt <[EMAIL PROTECTED]>
Date: Sat, 4 Mar 2006 19:46:22 +0100 (MET)
> Does this buy the normal standard desktop user anything?
Absolutely, it optimizes end-node performance.
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to [EMAIL PROT
On 3/2/06, Stefan Rompf <[EMAIL PROTECTED]> wrote:
> this makes me curious: What's the reason that e1000 driver intercepts and
> somehow parses outgoing DHCP packets on the 82573?
For those machines that are AMT (Intel Active Manangement Technology)
enabled with an 82573, we have to tell the firmw
On Sat, 2006-03-04 at 15:53 +0100, Andreas Schwab wrote:
> [Sorry for duplicate posting, I've used the wrong list address.]
>
> I suppose the NIC in the PowerMac G5 can do GigE, yet when plugged into a
> GB switch it is only willing to talk 100MB with it. Any idea why? Kernel
> is 2.6.16-rc5-git
jamal wrote:
> On Sat, 2006-04-03 at 17:56 +0100, Patrick McHardy wrote:
>
>>Adrian Bunk wrote:
>>
>>>This option should IMHO no longer depend on EXPERIMENTAL.
>>>
>>
>>Yesterday I managed to crash my machine playing around with tc actions
>>within minutes. I haven't looked into it yet, but it see
On Fri, Mar 03, 2006 at 01:42:20PM -0800, Chris Leech wrote:
> +void dma_async_device_unregister(struct dma_device* device)
> +{
...
> + kref_put(&device->refcount, dma_async_device_cleanup);
> + wait_for_completion(&device->done);
> +}
This looks like a bug: device is dereferenced after i
On Sat, 2006-04-03 at 17:56 +0100, Patrick McHardy wrote:
> Adrian Bunk wrote:
> > This option should IMHO no longer depend on EXPERIMENTAL.
> >
> >
> > Signed-off-by: Adrian Bunk <[EMAIL PROTECTED]>
> >
> > ---
> >
> > This patch was already sent on:
> > - 12 Feb 2006
>
> Yesterday I managed
>This patch series is the first full release of the Intel(R) I/O
>Acceleration Technology (I/OAT) for Linux. It includes an in kernel API
>for offloading memory copies to hardware, a driver for the I/OAT DMA memcpy
>engine, and changes to the TCP stack to offload copies of received
>networking da
hi,
this patch converts drivers/net to kzalloc usage.
Compile tested with allyes config.
It also fixes a bug in drivers/net/chelsio/espi.c, because
it called the memset() before checking if kmalloc failed.
Signed-off-by: Eric Sesterhenn <[EMAIL PROTECTED]>
--- linux-2.6.16-rc5-mm1/drivers/net/e
The following stuff is queued for 2.6.17 [really 2.6.16-git1] at
rsync://rsync.kernel.org/pub/scm/linux/kernel/git/jgarzik/netdev-2.6.git
branch 'upstream':
Documentation/DocBook/sis900.tmpl | 585 --
Documentation/networking/sis900.txt| 257
Documentation/DocBook/
Mark Brown wrote:
This patch converts the natsemi driver to use NAPI. It was originally
based on one written by Harald Welte, though it has since been modified
quite a bit, most extensively in order to remove the ability to disable
NAPI since none of the other drivers seem to provide that functi
Kirsher, Jeffrey T wrote:
Sorry. The duplex control register is used for setting the driver and
is not necessary for debug purposes. The value of the duplex control
register is what the register's current value is and may not reflect the
correct status of the current connection. That is what t
Arthur Kepner wrote:
The mac_addr variable doesn't get reset between
(re)additions of multicast addresses. One byte
of all multicast addresses (except the first)
can be incorrect.
Signed-off-by: Arthur Kepner <[EMAIL PROTECTED]>
applied. For the future, please don't use attachments!
Your mai
Adrian Bunk wrote:
> This option should IMHO no longer depend on EXPERIMENTAL.
>
>
> Signed-off-by: Adrian Bunk <[EMAIL PROTECTED]>
>
> ---
>
> This patch was already sent on:
> - 12 Feb 2006
Yesterday I managed to crash my machine playing around with tc actions
within minutes. I haven't looke
On Fri, Mar 03, Jesse Brandeburg wrote:
> thanks for testing, I wonder if there is something wrong with my 630, oh well.
You have them in SMP mode (non-partitioned)? Maybe that makes a
difference.
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to [EM
This option should IMHO no longer depend on EXPERIMENTAL.
Signed-off-by: Adrian Bunk <[EMAIL PROTECTED]>
---
This patch was already sent on:
- 12 Feb 2006
--- linux-2.6.16-rc2-mm1-full/net/sched/Kconfig.old 2006-02-12
02:21:30.0 +0100
+++ linux-2.6.16-rc2-mm1-full/net/sched/Kconfi
This documentation is mostly obsolete, and should therefore either be
updated or removed (this patch does the latter).
Signed-off-by: Adrian Bunk <[EMAIL PROTECTED]>
---
This patch was already sent on:
- 18 Feb 2006
Documentation/DocBook/Makefile |2
Documentation/DocBook/sis900.tm
[Sorry for duplicate posting, I've used the wrong list address.]
I suppose the NIC in the PowerMac G5 can do GigE, yet when plugged into a
GB switch it is only willing to talk 100MB with it. Any idea why? Kernel
is 2.6.16-rc5-git2.
# lsprop /proc/device-tree/[EMAIL PROTECTED],f200/[EMAIL PR
It appears sockaddr_in.sin_zero is not zeroed during certain operations
returning IPv4 socket names, namely:
- getsockopt(...SO_ORIGINAL_DST...) (2.4 and 2.6)
see getorigdst() in net/ipv4/netfilter/ip_conntrack_core.c
(+ in net/ipv4/netfilter/nf_conntrack_l3proto_ipv4.c in 2.6?!)
- getsockna
This patch makes the needlessly global function ipw_qos_current_mode()
static.
Signed-off-by: Adrian Bunk <[EMAIL PROTECTED]>
--- linux-2.6.16-rc5-mm2-full/drivers/net/wireless/ipw2200.c.old
2006-03-03 17:49:37.0 +0100
+++ linux-2.6.16-rc5-mm2-full/drivers/net/wireless/ipw2200.c
On Fri, Mar 03, 2006 at 02:39:22PM -0800, Chris Leech ([EMAIL PROTECTED]) wrote:
> > Patch #2 didn't make it. Too big for the list?
>
> Could be, it's the largest of the series. I've attached the gziped
> patch. I can try and split this up for the future.
How can owner of cb_chan->common.devic
On Fri, Mar 03, 2006 at 01:42:34PM -0800, Chris Leech wrote:
> Any socket recv of less than this ammount will not be offloaded
There is no documentation update.
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at
On Fri, Mar 03, 2006 at 09:24:46PM -0500, Jeff Garzik wrote:
> Adrian Bunk wrote:
> >Signed-off-by: Adrian Bunk <[EMAIL PROTECTED]>
> >
> >---
> >
> >This patch was already sent on:
> >- 4 Feb 2006
> >
> > Documentation/feature-removal-schedule.txt |6 ++
> > drivers/net/eepro100.c
33 matches
Mail list logo
