On Thu, 8 Nov 2007, Rainer Jochem wrote:
> @@ -620,6 +622,17 @@ ic_dhcp_init_options(u8 *options)
> *e++ = sizeof(ic_req_params);
> memcpy(e, ic_req_params, sizeof(ic_req_params));
> e += sizeof(ic_req_params);
> +
> + // Send it only if the ac
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
net/ipv4/tcp_input.c |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index 9fc9096..84bcdc9 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -1400,7 +
Hi,
Here are some changes that should go to net-2.6. Boot & simple
transfer tested. I'll do future cleanups that are now possible
in a different series to a next kernel version.
Reordering detection fails to take account that the reordered
skb may have pcount larger than 1. In such case the lowest of
them had the largest reordering, the old formula used the
highest of them which is pcount - 1 packets less reordered.
Signed-off-by: Ilpo Järvinen <[EMAIL PROTEC
t too sure so I left it there.
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
net/ipv4/tcp_input.c | 50 --
1 files changed, 32 insertions(+), 18 deletions(-)
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index 0f75757..9fc9096
core of the loop truly deals with a single
skb only, which also enables creation a more self-contained
of tcp_sacktag_one later on.
In addition, small reorganization of if branches was made.
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
net/ipv4/tcp_input.c
t could prevent
tcp_is_reno from going to tcp_xmit_retransmit_queue when queue is empty),
it should get rid of the crash and get the lost_out value for us as
well...
> .config:
...snip...
> # CONFIG_DEBUG_LIST is not set
Could you please add this one too, as there could be some list corru
n it
occurs though it's not the main problem itself.
Please keep the tcp_sack set to 0, and once you have at least one
stacktrace with it, you could try also with tcp_sack if the same thing
occurs there as well.
--
[PATCH] TCP DEBUG
- Check if empty queue is passed to xmit_retrans...
- Print
On Mon, 12 Nov 2007, Ilpo Järvinen wrote:
> Yeah, it's more likely a miscount somewhere rather than corruption but
> that wasn't obvious from the first mail...
>
> ...but alas, I haven't yet been able to come up with any theory on how
> a miscount could occur..
pletely non-existing.
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
Noticed this unrelated problem while thinking LOST bits and stuff...
net/ipv4/tcp_input.c |3 +++
1 files changed, 3 insertions(+), 0 deletions(-)
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index cc3
> > I just cannot now construct in my mind, end_seq <=
> > frto_highmark check does not match causing the lost_out
> > and LOST bits become out-of-sync due to clearing and
> > recounting in the loop.
> >
> > This may fix LOST-bit leak reported by Chazarai
k.
> >
> > Having LOST-bits integry ensured like done after commit
> > 23aeeec365dcf8bc87fae44c533e50d0bb4f23cc won't hurt. It may
> > still be useful in some other, possibly legimate, scenario.
> >
> > Reported by Chazarain Guillaume <[EMAIL PROTECTED]>.
> >
On Thu, 15 Nov 2007, Guillaume Chazarain wrote:
> David Miller <[EMAIL PROTECTED]> wrote:
>
> > Chazarain please let us know if it does indeed cure your
> > problem.
>
> Unfortunately, I couldn't manage to reproduce the problem with an
> unpatched kernel. But your investigation Ilpo was really i
Highest_sack_end_seq is no longer calculated in the loop,
thus it can be pushed to the worker function altogether
making that function independent of the sacktag.
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
net/ipv4/tcp_input.c | 20 +++-
1 files changed, 11 inse
markings after that occur
making the access pattern quite similar to the former fastpath
"special case".
DSACKs are special case that must always be walked.
The local to recv_sack_cache copying could be more intelligent
w.r.t DSACKs which are likely to be there only once but that
i
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
net/ipv4/tcp_input.c |5 ++---
1 files changed, 2 insertions(+), 3 deletions(-)
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index b7af304..29fff81 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -1251,7 +
It is going to replace the sack fastpath hint quite soon... :-)
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
include/linux/tcp.h |6 --
include/net/tcp.h | 10 ++
net/ipv4/tcp_input.c | 12 ++--
net/ipv4/tcp_output.c | 19 ++-
4
Hi Dave,
Here's the sacktag recv_sack_cache usage rewrite which you were
interested to look at earlier, due to other fixes it has dragged
on this long... Besides that, couple of new bugs^W^Wcleanups &
tweaks are there as well :-). I'll probably have to summon create
tcp_sacktag_state patch back to
All intermediate conditions include it already, make them
simpler as well.
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
net/ipv4/tcp_input.c | 31 ++-
1 files changed, 14 insertions(+), 17 deletions(-)
diff --git a/net/ipv4/tcp_input.c b/ne
ation last for long enough to cause problems).
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
net/ipv4/tcp_input.c | 80 ++---
1 files changed, 62 insertions(+), 18 deletions(-)
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
inde
o TCPCB_RETRANS skbs were found,
thus having non-zero undo_marker in CA_Loss basically tells
that the R-bits still accurately reflect the current state
of TCP.
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
net/ipv4/tcp_input.c |3 ++-
1 files changed, 2 insertions(+), 1 deletions
Worker function that implements the main logic of
the inner-most loop of tcp_sacktag_write_queue().
Idea was originally presented by David S. Miller.
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
net/ipv4/tcp_input.c | 192 +-
1
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
include/linux/tcp.h |2 +-
net/ipv4/tcp_input.c | 85 ++
2 files changed, 52 insertions(+), 35 deletions(-)
diff --git a/include/linux/tcp.h b/include/linux/tcp.h
index 34acee6..7
This is not intented to go to mainline, provided just for those
who are interested enough about the algorithm internals during
a test.
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
include/linux/snmp.h | 19 +++
net/ipv4/proc.c | 19 +++
ne
Previously one of the in-block skip branches was missing it.
Also, drop it from tail-fully-processed case because the next
iteration will do exactly the same thing, i.e., process the
SACK block that contains the DSACK information.
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
ne
is fast-forward and mandatory markings after that occur
> > making the access pattern quite similar to the former fastpath
> > "special case".
> >
> > DSACKs are special case that must always be walked.
> >
> > The local to recv_sack_cache copying could be mor
When the abstraction functions got added, conversion was made
here incorrectly. As a result, the skb may end up pointing
to skb which got included to the probe skb and then was freed.
For it to trigger, however, transmit_skb must fail sending as
well.
Signed-off-by: Ilpo Järvinen <[EM
e skb may end up pointing
to skb which got included to the probe skb and then was freed.
For it to trigger, however, skb_transmit must fail sending as
well.
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
include/net/tcp.h |3 +++
net/ipv4/tcp_output.c |1 -
2 files changed, 3
On Mon, 19 Nov 2007, David Miller wrote:
> From: "Ilpo_Järvinen" <[EMAIL PROTECTED]>
> Date: Fri, 16 Nov 2007 15:44:40 +0200 (EET)
>
> > Besides the log(n) search for which patches already exists, only O(new
> > information) scan start..end range would be necessary. Or had you
> > something els
Hi all,
Here are two other things in MTU probe code that caught my
attention while attempting to figure out the sk_send_head
usage there (sent patch to that earlier). The latter here is
not strictly speaking a fix but the original code has striking
complexity to perform a query which
l queued.
Use of write_seq check guarantees that there's a valid skb in
send_head so I removed the extra check.
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
net/ipv4/tcp_output.c |7 +--
1 files changed, 1 insertions(+), 6 deletions(-)
diff --git a/net/ipv4/tcp_
account packets that are
necessary to trigger fast retransmit [RFC4821].
Location of snd_wnd < probe_size/size_needed check is bogus
because it will cause the other if() match as well (due to
snd_nxt >= snd_una invariant).
Removed dead obvious comment.
Signed-off-by: Ilpo Järvinen <[EMAIL
On Thu, 22 Nov 2007, Herbert Xu wrote:
> On Wed, Nov 21, 2007 at 06:01:27PM +0200, Ilpo Järvinen wrote:
> >
> > Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
>
> Thanks for the patch Ilpo!
>
> I've just got a couple small questions.
...Thanks for actua
Better place exists in update_send_head (other non-queue related
adjustments are done there as well) which is the only caller of
tcp_advance_send_head (now that the bogus call from mtu_probe is
gone).
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
include/net/tcp.h |5
On Fri, 23 Nov 2007, Herbert Xu wrote:
> On Fri, Nov 23, 2007 at 02:25:08PM +0200, Ilpo Järvinen wrote:
> >
> > Better place exists in update_send_head (other non-queue related
> > adjustments are done there as well) which is the only caller of
> > tcp_advance_send_he
':
include/linux/pcounter.h:87: error: 'struct pcounter' has no
member named 'value'
make[1]: *** [ipc/mqueue.o] Error 1
make: *** [ipc] Error 2
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
include/linux/pcounter.h |2 +-
1 files changed, 1 insertions(+), 1
similar advancement
is not problem because highest_sack was previosly put to point
a sacked skb.
These problems were located because of problem report from Matt
Mathis <[EMAIL PROTECTED]>.
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
...I verified result from time-seq graphs now, th
On Wed, 28 Nov 2007, Ilpo Järvinen wrote:
> @@ -1575,7 +1575,7 @@ tcp_sacktag_write_queue(struct sock *sk, struct sk_buff
> *ack_skb, u32 prior_snd_
> continue;
> }
>
> - if (!before(start_seq, tcp_highest_sack_seq(tp))) {
>
From: =?ISO-8859-1?q?Ilpo_J=E4rvinen?= <[EMAIL PROTECTED]>
Usually those skbs will have L set, not counting them as lost
retransmissions is misleading.
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
net/ipv4/tcp_input.c |2 +-
1 files changed, 1 insertions(+), 1 deleti
Hi all,
First, there are some rather trivial ones (up to [TCP]: Cleanup
local variables of clean _rtx_queue)...
They're followed by my current efforts to improve SACK processing
latencies with large windows, including RB-tree for fast
searching, problem space split to provide tight bounds that
re
marked with
FLAG_DATA_SACKED.
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
net/ipv4/tcp_input.c | 10 +++---
1 files changed, 3 insertions(+), 7 deletions(-)
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index ba05e16..6986a2d 100644
--- a/net/ipv4/tcp_input.c
++
From: =?ISO-8859-1?q?Ilpo_J=E4rvinen?= <[EMAIL PROTECTED]>
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
include/net/tcp.h|4 ++--
net/ipv4/tcp_bic.c |3 +--
net/ipv4/tcp_cong.c |2 +-
net/ipv4/tcp_cubic.c |3 +--
net/ipv4/tcp_highsp
From: =?ISO-8859-1?q?Ilpo_J=E4rvinen?= <[EMAIL PROTECTED]>
Bogus seqno compares just mislead, the code is identical for
both sides of the seqno compare (and was even executed just
once because of return in between).
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
net/ipv4
From: =?ISO-8859-1?q?Ilpo_J=E4rvinen?= <[EMAIL PROTECTED]>
They better be valid when call to write_queue functions is made
once things that follow are going in.
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
net/ipv4/tcp_output.c |3 ++-
1 files changed, 2 insertions(+),
From: =?ISO-8859-1?q?Ilpo_J=E4rvinen?= <[EMAIL PROTECTED]>
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
net/ipv4/tcp_input.c |4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index 3f0b6c7..365c6d4 1006
From: =?ISO-8859-1?q?Ilpo_J=E4rvinen?= <[EMAIL PROTECTED]>
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
net/ipv4/tcp_input.c | 26 --
1 files changed, 12 insertions(+), 14 deletions(-)
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
i
From: =?ISO-8859-1?q?Ilpo_J=E4rvinen?= <[EMAIL PROTECTED]>
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
net/ipv4/tcp_input.c | 48
1 files changed, 20 insertions(+), 28 deletions(-)
diff --git a/net/ipv4/tcp_input.
From: =?ISO-8859-1?q?Ilpo_J=E4rvinen?= <[EMAIL PROTECTED]>
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
include/net/tcp.h |5 +
net/ipv4/tcp_output.c |8
2 files changed, 9 insertions(+), 4 deletions(-)
diff --git a/include/net/tcp.h b/include/net
came from David S. Miller, included couple of bug
fixes from Tom Quetchenbach <[EMAIL PROTECTED]>.
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
include/net/tcp.h | 41 +
1 files changed, 41 insertions(+), 0 deletions(-)
diff --git a/in
From: =?ISO-8859-1?q?Ilpo_J=E4rvinen?= <[EMAIL PROTECTED]>
Pointing to the next skb is necessary to avoid referencing
already SACKed skbs which will soon be on a separate list.
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
include/net/tcp.
From: =?ISO-8859-1?q?Ilpo_J=E4rvinen?= <[EMAIL PROTECTED]>
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
include/net/tcp.h |8
net/ipv4/tcp_input.c | 18 +-
net/ipv4/tcp_output.c |8
3 files changed, 17 insertions(+), 17 deleti
From: =?ISO-8859-1?q?Ilpo_J=E4rvinen?= <[EMAIL PROTECTED]>
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
include/net/tcp.h | 10 +-
1 files changed, 5 insertions(+), 5 deletions(-)
diff --git a/include/net/tcp.h b/include/net/tcp.h
index 0883697..117fcc4 100644
-
From: =?ISO-8859-1?q?Ilpo_J=E4rvinen?= <[EMAIL PROTECTED]>
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
net/ipv4/tcp_input.c | 66 +-
1 files changed, 33 insertions(+), 33 deletions(-)
diff --git a/net/ipv4/tcp_input.
From: =?ISO-8859-1?q?Ilpo_J=E4rvinen?= <[EMAIL PROTECTED]>
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
net/ipv4/tcp_input.c | 30 +++---
1 files changed, 15 insertions(+), 15 deletions(-)
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
i
From: =?ISO-8859-1?q?Ilpo_J=E4rvinen?= <[EMAIL PROTECTED]>
The call must be done before any queue moves occur (done in
the later patch).
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
net/ipv4/tcp_input.c |6 +++---
1 files changed, 3 insertions(+), 3 deletions(-)
diff
From: =?ISO-8859-1?q?Ilpo_J=E4rvinen?= <[EMAIL PROTECTED]>
This work was mostly done by David S. Miller.
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
include/linux/skbuff.h |3 ++
include/linux/tcp.h |2 +
include/net/tcp.h
From: =?ISO-8859-1?q?Ilpo_J=E4rvinen?= <[EMAIL PROTECTED]>
This is necessary when SACKed and non-SACKed spaces become
separate in the later patch.
However, callers should be converted to behave without this
later on.
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
inclu
From: =?ISO-8859-1?q?Ilpo_J=E4rvinen?= <[EMAIL PROTECTED]>
Lines won't be that long and it's compiler's job to optimize
them.
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
net/ipv4/tcp_output.c |9 +++--
1 files changed, 3 insertions(+), 6 deletion
From: =?ISO-8859-1?q?Ilpo_J=E4rvinen?= <[EMAIL PROTECTED]>
Soon the skbs in the write_queue won't necessary be adjacent
ones because part of the queue is stored elsewhere. Verify that
there isn't hole between skbs when trying to combine two of
them.
Signed-off-by: Ilpo Järvinen &
nd tp->highest_sack may be obsolete after
this. At least the highest item of recv_sack_cache is very
likely redundant.
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
include/linux/tcp.h |2 +
include/net/tcp.h | 276 +--
On Sun, 2 Dec 2007, Herbert Xu wrote:
> On Sun, Dec 02, 2007 at 12:48:08AM +0200, Ilpo Järvinen wrote:
> >
> > This work was mostly done by David S. Miller.
>
> ...
>
> > @@ -253,6 +254,8 @@ struct sk_buff {
> > struct sk_buff *next;
&
On Sun, 2 Dec 2007, Herbert Xu wrote:
> On Sun, Dec 02, 2007 at 12:48:07AM +0200, Ilpo Järvinen wrote:
> >
> > @@ -1220,6 +1221,11 @@ static inline struct sk_buff
> > *tcp_write_queue_next(struct sock *sk, struct sk_bu
> > return skb->next;
> > }
&g
Should get the skb from the same queue. I had it first elsewhere
and missed this change while moving.
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
diff --git a/include/net/tcp.h b/include/net/tcp.h
index 7ae72c3..5929022 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -
On Mon, 3 Dec 2007, Wolfgang Walter wrote:
> with kernel 2.6.23.8 we saw a
>
> KERNEL: assertion ((int)tcp_packets_in_flight(tp) >= 0) failed at
> net/ipv4/tcp_input.c (1292)
Is this the only message? Are there any Leak printouts?
Any tweaking done to TCP related sysctls?
And for completeness,
On Mon, 3 Dec 2007, Lachlan Andrew wrote:
> On 03/12/2007, Stephen Hemminger <[EMAIL PROTECTED]> wrote:
> > On Mon, 3 Dec 2007 15:59:23 -0800
> > "Shao Liu" <[EMAIL PROTECTED]> wrote:
> > > And also a question, why the samples when SACK is active are outliers?
> >
> > Any sample with SACK is going
Hi Dave,
Here are couple of fixes to net-2.6. The first one is to FRTO
which may fix some corner-case bug if the hand-coded check and
the !tcp_may_send_now disagree. I didn't check the differences
that carefully so they might agree after considering what FRTO
overrides from there, however, it make
From: =?ISO-8859-1?q?Ilpo_J=E4rvinen?= <[EMAIL PROTECTED]>
The comment in tcp_nagle_test suggests that. This bug is very
very old, even 2.4.0 seems to have it.
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
net/ipv4/tcp_output.c |3 +--
1 files changed, 1 insertions(+),
e case of
tcp_reno_cong_avoid bails out early because it thinks that
TCP is not limited by cwnd.
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
net/ipv4/tcp_input.c |3 +--
1 files changed, 1 insertions(+), 2 deletions(-)
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
i
ous frto_counter past
tcp_may_send_now.
In addition, returns can then be combined.
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
net/ipv4/tcp_input.c | 14 +-
1 files changed, 5 insertions(+), 9 deletions(-)
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
in
On Tue, 4 Dec 2007, John Heffner wrote:
> Ilpo Järvinen wrote:
> > ...I'm still to figure out why tcp_cwnd_down uses snd_ssthresh/2
> > as lower bound even though the ssthresh was already halved, so snd_ssthresh
> > should suffice.
>
> I remember this coming
On Tue, 4 Dec 2007, John Heffner wrote:
> Ilpo Järvinen wrote:
>
> > ...Mind if I ask another similar one, any idea why prior_ssthresh is smaller
> > (3/4 of it) than cwnd used to be (see tcp_current_ssthresh)?
>
> Not sure on that one. I'm not aware of any publi
On Wed, 5 Dec 2007, David Miller wrote:
> From: "Ilpo_Järvinen" <[EMAIL PROTECTED]>
> Date: Tue, 4 Dec 2007 18:48:50 +0200
>
> > The comment in tcp_nagle_test suggests that. This bug is very
> > very old, even 2.4.0 seems to have it.
> >
> &g
Earlier resolution for NewReno's sacked_out should now keep
it small enough for this to become invariant-like check.
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
net/ipv4/tcp_input.c |5 +
1 files changed, 1 insertions(+), 4 deletions(-)
diff --git a/net/ipv4/tcp_in
On Wed, 5 Dec 2007, Ilpo Järvinen wrote:
>
> Earlier resolution for NewReno's sacked_out should now keep
> it small enough for this to become invariant-like check.
>
> Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
Forgot to modify it to the subject, this one to net-2.6.25 please.
--
i.
On Wed, 5 Dec 2007, Stephen Hemminger wrote:
> On Wed, 05 Dec 2007 08:53:07 -0800
> Joe Perches <[EMAIL PROTECTED]> wrote:
>
> > > it occurred to me that we might want to do something
> > > like a state change event generator.
> >
> > This could be a basis for an interesting TCP
> > performance
On Wed, 5 Dec 2007, David Miller wrote:
> Ilpo, I was pondering the kind of debugging one does to find
> congestion control issues and even SACK bugs and it's currently too
> painful because there is no standard way to track state changes.
That's definately true.
> I assume you're using somethin
On Wed, 5 Dec 2007, David Miller wrote:
> From: John Heffner <[EMAIL PROTECTED]>
> Date: Tue, 04 Dec 2007 13:42:41 -0500
>
> > Ilpo Järvinen wrote:
> > > ...I'm still to figure out why tcp_cwnd_down uses snd_ssthresh/2
> > > as lower bound even thoug
Fixes NewReno breakage that got introduced because reno ended up
into RFC3517 counting branch in tcp_mark_head_lost after the
split space tweaks. The fack branch would have worked, however,
coming lost marker updates would eventually add this anyway.
Signed-off-by: Ilpo Järvinen <[EM
On Thu, 6 Dec 2007, Lachlan Andrew wrote:
> On 04/12/2007, Ilpo Järvinen <[EMAIL PROTECTED]> wrote:
> > On Mon, 3 Dec 2007, Lachlan Andrew wrote:
> > >
> > > When SACK is active, the per-packet processing becomes more involved,
> > > tracking the list
On Fri, 7 Dec 2007, David Miller wrote:
> From: "Ilpo_Järvinen" <[EMAIL PROTECTED]>
> Date: Fri, 7 Dec 2007 13:05:46 +0200 (EET)
>
> > I guess if you get a large cumulative ACK, the amount of processing is
> > still overwhelming (added DaveM if he has some idea how to combat it).
> >
> > Even a
On Wed, 5 Dec 2007, David Miller wrote:
> From: Reuben Farrelly <[EMAIL PROTECTED]>
> Date: Thu, 06 Dec 2007 17:59:37 +1100
>
> > On 5/12/2007 4:17 PM, Andrew Morton wrote:
> > > - Lots of device IDs have been removed from the e1000 driver and moved
> > > over
> > > to e1000e. So if your e100
On Thu, 6 Dec 2007, David Miller wrote:
> From: "Ilpo_Järvinen" <[EMAIL PROTECTED]>
> Date: Thu, 6 Dec 2007 01:18:28 +0200 (EET)
>
> > On Wed, 5 Dec 2007, David Miller wrote:
> >
> > > I assume you're using something like carefully crafted printk's,
> > > kprobes, or even ad-hoc statistic counte
On Fri, 7 Dec 2007, Ilpo Järvinen wrote:
> On Fri, 7 Dec 2007, David Miller wrote:
>
> > From: "Ilpo_Järvinen" <[EMAIL PROTECTED]>
> > Date: Fri, 7 Dec 2007 13:05:46 +0200 (EET)
> >
> > > I guess if you get a large cumulative ACK, the amount
On Sun, 9 Dec 2007, Andrew Morton wrote:
>
> This is a mac g5. It is being used as a distcc server across the LAN
> when...
>
> windfarm: CPUs control loops started.
> EXT3-fs: INFO: recovery required on readonly filesystem.
> EXT3-fs: write access will be enabled during recovery.
> kjournald s
x27;t return it because it's on
the other list already). These manifest as fackets_out counting
error later on, the second-order effects are very hard to track,
so it may fix all out-standing TCP bug reports.
2) Prev == NULL check was wrong way around
3) Last skb's fack count was incorrectly
Added checks will catch most of the errors if the current
complex fack_count counting logic is flawed somewhere.
Fackets_out should always be advancable if highest_sack is too
because the fackets_out is nowadays accurate (and obviously it
must be smaller than packets_out).
Signed-off-by: Ilpo
> >
> > Fackets_out should always be advancable if highest_sack is too
> > because the fackets_out is nowadays accurate (and obviously it
> > must be smaller than packets_out).
> >
> > Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
>
> Applied to net
On Mon, 10 Dec 2007, Ilpo Järvinen wrote:
> Dave, please include this one to net-2.6.25.
...
> --
> [PATCH] [TCP]: Fix fack_count miscountings (multiple places)
I've better version of this coming up, so Dave please don't put this one
into net-2.6.25 (noticed that both t
losses in the same window when TCP
runs out new data. Now check really against the list heads
rather than a valid skb which should still be processed.
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
include/net/tcp.h |6 +++---
1 files changed, 3 insertions(+), 3 deletions(
Hi Dave,
Here are two TCP patches. The first one removes not so useful
parameter (my avoid GSO skb fragment RFC patch that I'll send
separately right after I get this out depends it). The second
applies cleanly only after the other fack_count fix patch.
--
i.
--
To unsubscribe from this list:
This makes flow more obvious in case of short-circuit and
removes need for prev double pointer & fc recount per queue
switch.
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
include/net/tcp.h | 54 ++--
1 files changed, 19 inserti
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
net/ipv4/tcp_input.c | 16 +++-
1 files changed, 7 insertions(+), 9 deletions(-)
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index 75a16b3..977c68c 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
avoided if the sender could see the upcoming future.
Not only SACK processing suffers but clean_rtx_queue as well
is considerable hit when the corresponding cumulative ACK
arrives.
Thus implement a local cache for a single skb to avoid enormous
splitting efforts while the latest SACK block is still gr
On Thu, 13 Dec 2007, Cedric Le Goater wrote:
> I got this one while compiling on NFS.
>
> C.
>
> kernel BUG at /home/legoater/linux/2.6.24-rc4-mm1/include/net/tcp.h:1480!
I'm not exactly sure what patches you have applied and which patches are
not, with rc4-mm1 there are two patches (first one
On Fri, 14 Dec 2007, Ilpo Järvinen wrote:
> So, I might soon prepare a revert patch for most of the questionable
> TCP parts and ask Dave to apply it (and drop them fully during next
> rebase) unless I suddently figure something out soon which explains
> all/most of the problems,
On Thu, 13 Dec 2007, Wolfgang Walter wrote:
> it happened again with your patch applied:
>
> WARNING: at net/ipv4/tcp_input.c:1018 tcp_sacktag_write_queue()
>
> Call Trace:
> [] tcp_sacktag_write_queue+0x7d0/0xa60
> [] add_partial+0x19/0x60
> [] tcp_ack+0x5a4/0x1d70
> [] tcp_rcv_established+0x
On Mon, 2 Jul 2007, David Miller wrote:
> From: "Ilpo_Järvinen" <[EMAIL PROTECTED]>
> Date: Sat, 16 Jun 2007 02:04:25 +0300 (EEST)
>
> > There are still some things I must think carefully in sacktag processing
> > since it does not validate start_seq and end_seq at all which can be
> > abused c
> does not implement this). However, that would require very
> > invasive change to fastretrans_alert (perhaps even slicing it to
> > two halves). Alternatively, all callers of tcp_packets_in_flight
> > (i.e., users that depend on sacked_out) should be postponed
> > until
Got bored to always recompile it for no reason.
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
drivers/net/bnx2.c |1 -
1 files changed, 0 insertions(+), 1 deletions(-)
diff --git a/drivers/net/bnx2.c b/drivers/net/bnx2.c
index 88b33c6..56244f6 100644
--- a/drivers/net/bnx2.c
No longer needed.
Signed-off-by: Ilpo Järvinen <[EMAIL PROTECTED]>
---
include/net/dn.h |1 -
1 files changed, 0 insertions(+), 1 deletions(-)
diff --git a/include/net/dn.h b/include/net/dn.h
index ac4ce90..6277783 100644
--- a/include/net/dn.h
+++ b/include/net/dn.h
@@ -3,
1 - 100 of 527 matches
Mail list logo