On Mon, 17 Dec 2007 10:05:35 -0800
Ray Lee [EMAIL PROTECTED] wrote:
On Dec 17, 2007 9:55 AM, Eric Dumazet [EMAIL PROTECTED] wrote:
- mid = (last - first) / 2 + first;
+ while (low = high) {
+ mid = (low + high) / 2;
I think you just introduced a bug
James Nichols a écrit :
Here is a purely hypothethical (and in practice unlikely) idea:
Java opens up too many sockets (more than you really request) and the
kernel, for whatever reason, does not deliver packets to programs
which have maxed out their fds. Well it would already help if the
java
James Nichols a écrit :
Well... please dont start a flame war :(
Back to your SYN_SENT problem, I suppose the remote IP is known, so you
probably could post here the result of a tcdpump ?
tcpdump -p -n -s 1600 host IP_of_problematic_peer -c 500
Most probably remote peer received too many
cpuinfo_op; will be taken into account
in a separate patch, since its const status is arch dependant.
Signed-off-by: Eric Dumazet [EMAIL PROTECTED]
fs/proc/proc_misc.c |9 -
include/linux/seq_file.h | 11 +++
2 files changed, 11 insertions(+), 9 deletions(-)
diff --git a/fs
,%eax
1c: 8d 04 c3lea(%ebx,%eax,8),%eax
1f: 39 08 cmp%ecx,(%eax)
...
Signed-off-by: Eric Dumazet [EMAIL PROTECTED]
diff --git a/lib/extable.c b/lib/extable.c
index 463f456..179c087 100644
--- a/lib/extable.c
+++ b/lib/extable.c
@@ -57,10 +57,10
was found in groups_search()
(commit d74beb9f33a5f16d2965f11b275e401f225c949d ) and at that time I changed
some variables to unsigned int.
I believe that a more generic fix is to make sure NGROUPS_PER_BLOCK is unsigned.
Signed-off-by: Eric Dumazet [EMAIL PROTECTED]
diff --git a/include/linux
Andrew Morton a écrit :
On Wed, 19 Dec 2007 01:14:33 +0100
Eric Dumazet [EMAIL PROTECTED] wrote:
groups_sort() can be quite long if user loads a large gid table.
This is because GROUP_AT(group_info, some_integer) uses an integer divide.
So having to do XXX thousand divides during one syscall
Dotan Barak a écrit :
Hi all.
I noticed that the code segment of the user level in PPC64 machines
is in a VMA with a write permission enabled.
I'm using the following machine attributes:
*
Host Name : mtlsqt185
Host
James Nichols a écrit :
So you see outgoing SYN packets, but no SYN replies coming from the remote
peer ? (you mention ACKS, but the first packet received from the remote
peer should be a SYN+ACK),
Right, I meant to say SYN+ACK. I don't see them coming back.
So... Really unlikely a linux
James Nichols a écrit :
On 12/19/07, Eric Dumazet [EMAIL PROTECTED] wrote:
James Nichols a écrit :
So you see outgoing SYN packets, but no SYN replies coming from the remote
peer ? (you mention ACKS, but the first packet received from the remote
peer should be a SYN+ACK),
Right, I meant
On Fri, 21 Dec 2007 02:29:12 -0800
Andrew Morton [EMAIL PROTECTED] wrote:
On Fri, 14 Dec 2007 18:51:06 +0100 Eric Dumazet [EMAIL PROTECTED] wrote:
We can use ilog2() in fs/namespace.c to compute hash_bits and hash_mask at
compile time, not runtime.
Well noted.
[namespace.patch text
Holger Hoffstaette a écrit :
Hi -
This regular Linux user and lkml lurker just noticed data corruption in
ftp'ed files and narrowed it down to vsftpd using sendfile(). So far this
has never caused problems in the past; I have not noticed this with
2.6.22.x but may have missed it. I do remember
Marc Haber a écrit :
While debugging Exim4's GnuTLS interface, I recently found out that
reading from /dev/urandom depletes entropy as much as reading from
/dev/random would. This has somehow surprised me since I have always
believed that /dev/urandom has lower quality entropy than /dev/random,
that the nbits
estimation is less pessimistic, but also to
avoid injecting false entropy.
Signed-off-by: Eric Dumazet [EMAIL PROTECTED]
diff --git a/drivers/char/random.c b/drivers/char/random.c
index 5fee056..6eccfc9 100644
--- a/drivers/char/random.c
+++ b/drivers/char/random.c
@@ -550,8 +550,8
Ingo Molnar a écrit :
* Jie Chen [EMAIL PROTECTED] wrote:
I just ran the same test on two 2.6.24-rc4 kernels: one with
CONFIG_FAIR_GROUP_SCHED on and the other with CONFIG_FAIR_GROUP_SCHED
off. The odd behavior I described in my previous e-mails were still
there for both kernels. Let me know
Ingo Molnar a écrit :
* Eric Dumazet [EMAIL PROTECTED] wrote:
$ gcc -O2 -o burner burner.c
$ ./burner
Time to perform the unit of work on one thread is 0.040328 s
Time to perform the unit of work on 2 threads is 0.040221 s
ok, but this actually suggests that scheduling is fine
Matt Mackall a écrit :
On Tue, Dec 04, 2007 at 07:17:58PM +0100, Eric Dumazet wrote:
Alan Cox a ?crit :
No matter what you consider as being better, changing a 12 years old and
widely used userspace interface like /dev/urandom is simply not an
option.
Fixing it to be more efficient
Ulrich Drepper a écrit :
wing patches provide an alternative implementation of the
sys_indirect system call which has been discussed a few times.
This no system call allows us to extend existing system call
interfaces with adding more system calls.
I am wondering if some parts are missing from
Jie Chen a écrit :
Hi, there:
We have a simple pthread program that measures the synchronization
overheads for various synchronization mechanisms such as spin locks,
barriers (the barrier is implemented using queue-based barrier
algorithm) and so on. We have dual quad-core AMD opterons
Ulrich Drepper a écrit :
This patch adds support for setting the O_NONBLOCK flag of the file
descriptors returned by socket, socketpair, and accept.
Thanks Ulrich for this v5 series. I have two more questions.
1) Can the fd passing with recvmsg() on AF_UNIX also gets O_CLOEXEC support ?
Ulrich Drepper a écrit :
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Eric Dumazet wrote:
1) Can the fd passing with recvmsg() on AF_UNIX also gets O_CLOEXEC
support ?
Already there, see MSG_CMSG_CLOEXEC.
OK, but maybe for consistency, we might accept the two mechanisms.
The one added
sysctl (/proc/sys/fs/nr_open) wich defaults to
1024*1024, so that admins can decide to change this limit if their workload
needs it.
Signed-off-by: Eric Dumazet [EMAIL PROTECTED]
Cc: Alan Cox [EMAIL PROTECTED]
Signed-off-by: Andrew Morton [EMAIL PROTECTED]
Documentation/filesystems/proc.txt
[EMAIL PROTECTED] a écrit :
On Tue, 27 Nov 2007 08:09:19 +0100, Eric Dumazet said:
Changing NR_OPEN is not considered safe because of vmalloc space potential
exhaust.
Verbiage about this point...
+nr_open
+---
+
+Denotes the maximum number of file-handles a process can
+allocate
of vmalloc space potential
exhaust.
This patch introduces a new sysctl (/proc/sys/fs/nr_open) wich defaults to
1024*1024, so that admins can decide to change this limit if their workload
needs it.
Signed-off-by: Eric Dumazet [EMAIL PROTECTED]
Cc: Alan Cox [EMAIL PROTECTED]
Signed-off
vmlinux.before_patch
Signed-off-by: Eric Dumazet [EMAIL PROTECTED]
diff --git a/kernel/rcupdate.c b/kernel/rcupdate.c
index a66d4d1..11c815c 100644
--- a/kernel/rcupdate.c
+++ b/kernel/rcupdate.c
@@ -75,9 +75,9 @@ DEFINE_PER_CPU(struct rcu_data, rcu_bh_data) = { 0L };
/* Fake initialization required by compiler
While auditing proc_doulongvec_ms_jiffies_minmax() usage in kernel, I
found a bug
in drivers/parport/procfs.c, incorrectly using sizeof(int) instead of
sizeof(unsigned long)
Only 64bit arches are affected by this old bug.
Signed-off-by: Eric Dumazet [EMAIL PROTECTED]
diff --git a/drivers
min_sched_granularity_ns, max_sched_granularity_ns, min_wakeup_granularity_ns
and max_wakeup_granularity_ns are declared unsigned long.
This is incorrect since proc_dointvec_minmax() expects plain int guard values.
This bug only triggers on big endian 64 bit arches.
Signed-off-by: Eric Dumazet
Chris Friesen a écrit :
I retested it on an x86 machine and am seeing similar problems.
First, arp gives the arp table as expected:
[EMAIL PROTECTED]:/tftpboot/cnp/0-0-5-0/0-0-5-0 arp -n
Address HWtype HWaddress Flags MaskIface
172.24.0.9 ether
Chris Friesen a écrit :
Eric Dumazet wrote:
Chris Friesen a écrit :
Is this expected behaviour?
Probably not... Still a 2.6.14 kernel ?
Yep. Embedded hardware, so I'm unable to test with a more recent kernel.
And what is the version of ip command you have on this machine ?
ip -V
You
Adrian Bunk a écrit :
I tried the following patch with a full x86 .config [1]:
--- a/include/asm-x86/cache.h
+++ b/include/asm-x86/cache.h
-#define __read_mostly __attribute__((__section__(.data.read_mostly)))
+/* #define __read_mostly __attribute__((__section__(.data.read_mostly))) */
The
Christoph Lameter a écrit :
On Sat, 8 Dec 2007, Ingo Molnar wrote:
Good. Although we should perhaps look at that reported performance
problem with SLUB. It looks like SLUB will do a memclear() for the
area twice (first for the whole page, then for the thing it allocated)
for the slow case.
Matt Mackall a écrit :
On Thu, Dec 13, 2007 at 11:20:44PM +0100, Adrian Bunk wrote:
I tried the following patch with a full x86 .config [1]:
--- a/include/asm-x86/cache.h
+++ b/include/asm-x86/cache.h
-#define __read_mostly __attribute__((__section__(.data.read_mostly)))
+/* #define
Arnd Bergmann a écrit :
On Thursday 13 December 2007, Adrian Bunk wrote:
On Thu, Dec 13, 2007 at 11:29:08PM +0100, Andi Kleen wrote:
Adrian Bunk [EMAIL PROTECTED] writes:
-rwxrwxr-x 1 bunk bunk 46607243 2007-12-13 19:50 vmlinux.old
-rwxrwxr-x 1 bunk bunk 46598691 2007-12-13
We can use ilog2() in fs/namespace.c to compute hash_bits and hash_mask at
compile time, not runtime.
Signed-off-by: Eric Dumazet [EMAIL PROTECTED]
diff --git a/fs/namespace.c b/fs/namespace.c
index 0608388..835f14a 100644
--- a/fs/namespace.c
+++ b/fs/namespace.c
@@ -25,6 +25,7 @@
#include
Herbert Xu a écrit :
Ob Tue, Dec 04, 2007 at 12:17:23AM +1100, Herbert Xu wrote:
Never mind, we already have that in local_t and as Alexey correctly
points out, USER is still going to be the expensive variant with the
preempt_disable (well until BH gets threaded). So how about this patch?
I
-off-by: Eric Dumazet [EMAIL PROTECTED]
diff --git a/drivers/char/random.c b/drivers/char/random.c
index 5fee056..af48e86 100644
--- a/drivers/char/random.c
+++ b/drivers/char/random.c
@@ -256,14 +256,14 @@
* The minimum number of bits of entropy before we wake up a read on
* /dev/random. Should
Adrian Bunk a écrit :
On Sun, Dec 16, 2007 at 12:45:01PM +0100, Eric Dumazet wrote:
While examining vmlinux namelist on i686, I noticed :
c0581300 D random_table
c0581480 d input_pool
c0581580 d random_read_wakeup_thresh
c0581584 d random_write_wakeup_thresh
c0581600 d blocking_pool
Adrian Bunk a écrit :
On Sun, Dec 16, 2007 at 03:44:37PM +0100, Eric Dumazet wrote:
Adrian Bunk a écrit :
On Sun, Dec 16, 2007 at 12:45:01PM +0100, Eric Dumazet wrote:
While examining vmlinux namelist on i686, I noticed :
c0581300 D random_table
c0581480 d input_pool
c0581580 d
Matt Mackall a écrit :
On Sun, Dec 16, 2007 at 12:45:01PM +0100, Eric Dumazet wrote:
While examining vmlinux namelist on i686, I noticed :
c0581300 D random_table
c0581480 d input_pool
c0581580 d random_read_wakeup_thresh
c0581584 d random_write_wakeup_thresh
c0581600 d blocking_pool
Adrian Bunk a écrit :
On Sun, Dec 16, 2007 at 06:42:57PM +0100, Eric Dumazet wrote:
Adrian Bunk a écrit :
...
And even more funny, with gcc 4.2 and CONFIG_CC_OPTIMIZE_FOR_SIZE=y your
patch doesn't seem to make any space difference - are you using an older
compiler or even worse
.data.percpu22048 3227328512
Signed-off-by: Eric Dumazet [EMAIL PROTECTED]
arch/x86/kernel/cpu/common.c |2 +-
arch/x86/kernel/vmlinux_32.lds.S |1 +
include/asm-generic/percpu.h |4
include/asm-generic/vmlinux.lds.h |1 +
include/asm-x86
Andi Kleen a écrit :
They now look like
hal-resmgr[13791]: segfault at 3c rip 2b9c8caec182 rsp 7fff1e825d30 error 4 in
libacl.so.1.1.0[2b9c8caea000+6000]
This makes it easier to pinpoint bugs to specific libraries.
And printing the offset into a mapping also always allows to find the
On Thu, 3 Jan 2008 16:42:20 +0100 (CET)
Andi Kleen [EMAIL PROTECTED] wrote:
This allows to allocate memory really early before bootmem is setup.
And a symbol that can be tested by the preprocessor.
pgtable.h is probably not the best include for it, but also not the worst.
Cc: [EMAIL
422838 458752 5501532 53f25c vmlinux.before
4610534 422838 458752 5492124 53cd9c vmlinux.after
This saves 9408 bytes
Signed-off-by: Eric Dumazet [EMAIL PROTECTED]
diff --git a/arch/x86/kernel/acpi/wakeup_32.S b/arch/x86/kernel/acpi/wakeup_32.S
index 1e931aa..f53e327 100644
--- a/arch/x86/kernel
Marc Lehmann a écrit :
Hi!
I ran into what I see as unsolvable problems that make epoll useless as a
generic event mechanism.
I recently switched to libevent as event loop, and found that my programs
work fine when it is using select or poll, but work eratically or halt
when using epoll.
The
Marc Lehmann a écrit :
On Sat, Oct 27, 2007 at 10:23:17AM +0200, Eric Dumazet [EMAIL PROTECTED]
wrote:
In this case, the parent process works fine until the child closes fds,
after which the fds become unarmed in the parent too. This works as
I have no idea what exact problem you have
Marc Lehmann a écrit :
On Sat, Oct 27, 2007 at 11:22:25AM +0200, Eric Dumazet [EMAIL PROTECTED]
wrote:
If such a bug exists on your kernel, please fill a complete bug report,
giving details.
As this behaviour is clearly documented in the epoll manpage, why do you
think it is a bug? I think
David Schwartz a écrit :
6) Epoll removes the file from the set, when the *kernel* object gets
closed (internal use-count goes to zero)
With that in mind, how can the code snippet above trigger a removal from
the epoll set?
I don't see how that can be. Suppose I add fd 8 to an
1) hardcoded 10 value is used five times in places where NSEC_PER_SEC
might be more readable.
2) A conversion from nsec to msec uses the hardcoded 100 value, which is a
candidate for NSEC_PER_MSEC.
Signed-off-by: Eric Dumazet [EMAIL PROTECTED]
diff --git a/kernel/sched.c b
Christoph Lameter a écrit :
This patch increases the speed of the SLUB fastpath by
improving the per cpu allocator and makes it usable for SLUB.
Currently allocpercpu manages arrays of pointer to per cpu objects.
This means that is has to allocate the arrays and then populate them
as needed
Christoph Lameter a écrit :
+
+enum unit_type { FREE, END, USED };
+
+static u8 cpu_alloc_map[UNITS_PER_CPU] = { 1, };
You mean END here instead of 1 :)
+/*
+ * Allocate an object of a certain size
+ *
+ * Returns a per cpu pointer that must not be directly used.
+ */
+static void
Christoph Lameter a écrit :
Disguising costs a few cycles in the hot paths. So switch it off if
we are not debuggin.
Signed-off-by: Christoph Lameter [EMAIL PROTECTED]
---
include/linux/percpu.h |4
1 file changed, 4 insertions(+)
Index: linux-2.6/include/linux/percpu.h
Denis Cheng a écrit :
the temporary bf[127] char array is redundant, and the specified width 127 make
the output of /proc/net/route include many trailing spaces;
since most terminal's cols are less than 127, this made every fib entry occupy
two lines,
after applied this patch, the output of
Christoph Lameter a écrit :
On Fri, 8 Feb 2008, Eric Dumazet wrote:
And SLAB/SLUB allocators, even if only used from process context, want to
disable/re-enable interrupts...
Not any more. The new fastpath does allow avoiding interrupt
enable/disable and we will be hopefully able
Zhang, Yanmin a écrit :
Comparing with kernel 2.6.24, tbench result has regression with
2.6.25-rc1.
1) On 2 quad-core processor stoakley: 4%.
2) On 4 quad-core processor tigerton: more than 30%.
bisect located below patch.
b4ce92775c2e7ff9cf79cca4e0a19c8c5fd6287b is first bad commit
commit
Zhang, Yanmin a écrit :
On Fri, 2008-02-15 at 07:05 +0100, Eric Dumazet wrote:
Zhang, Yanmin a �crit :
Comparing with kernel 2.6.24, tbench result has regression with
2.6.25-rc1.
1) On 2 quad-core processor stoakley: 4%.
2) On 4 quad-core processor tigerton: more than 30%.
bisect
Marin Mitov a écrit :
Hi all,
As in: include/linux/netdevice.h (kernel-2.6.24.2) one finds:
static inline void __netif_tx_lock(struct net_device *dev, int cpu)
{
spin_lock(dev-_xmit_lock);
dev-xmit_lock_owner = cpu;
}
static inline void netif_tx_lock(struct net_device *dev)
{
On Thu, 2012-11-01 at 14:45 +0800, Chen Gang wrote:
Hello:
1) For Public Kernel:
A) in rt6_info_route function of net/ipv6/route.c
B) the length of rt-rt6i_dev-name is 16 (IFNAMSIZ)
C) using %16s is better than %8s (it will be more beautiful)
(also suggest to delete
On Tue, 2012-09-25 at 12:34 +1000, Stephen Rothwell wrote:
Hi all,
Today's linux-next merge of the net-next tree got a conflict in
net/ipv4/raw.c between commit ab43ed8b7490 (ipv4: raw: fix icmp_filter
()) from the net tree and commit 5640f7685831 (net: use a per task frag
allocator) from
On Tue, 2012-09-25 at 01:13 -0400, David Miller wrote:
From: Eric Dumazet eric.duma...@gmail.com
Date: Tue, 25 Sep 2012 07:10:42 +0200
Oops, my bad, net/ipv4/raw.c changes in 5640f7685831 (net: use a per
task frag allocator) should not be there :
I accidentally left a debugging
From: Eric Dumazet eduma...@google.com
Commit 5640f7685831 (net: use a per task frag allocator)
accidentally contained an unrelated change to net/ipv4/raw.c,
later committed (without the pr_err() debugging bits) in
net tree as commit ab43ed8b749 (ipv4: raw: fix icmp_filter())
This patch reverts
On Wed, 2012-10-03 at 10:47 +0100, Mel Gorman wrote:
On Tue, Oct 02, 2012 at 03:48:57PM -0700, Rick Jones wrote:
PS - I trust it is the receive-side throughput being reported/used
with UDP_STREAM :)
Good question. Now that I examine the scripts, it is in fact the sending
side that is
corrupted. If you're lucky, it might be
something that netdev will recognize as already fixed.
I have the same problem on the exact same hardware and found the cause:
Author: Eric Dumazet eric.duma...@gmail.com
Date: Tue Apr 10 20:08:39 2012 +
net: allow pskb_expand_head() to get
On Thu, 2012-10-04 at 18:02 +0200, Maxime Bizon wrote:
O
Since skb_recycle() resets skb-data using (skb-head + NET_SKB_PAD), a
recycled skb going multiple times through a path that needs to expand
skb head will get bigger and bigger each time, and you eventually end up
with an allocation
On Thu, 2012-10-04 at 19:09 +0200, Maxime Bizon wrote:
yes, on ipv6 forward path the default NET_SKB_PAD is too small, so each
packet forwarded has its headroom expanded, it is then recycled and gets
its original default headroom back, then it gets forwarded,
expanded, ...
Hmm, this sounds
On Thu, 2012-10-04 at 19:34 +0200, Maxime Bizon wrote:
On Thu, 2012-10-04 at 19:17 +0200, Eric Dumazet wrote:
yes, on ipv6 forward path the default NET_SKB_PAD is too small, so each
packet forwarded has its headroom expanded, it is then recycled and gets
its original default headroom
On Thu, 2012-10-04 at 20:05 -0700, ramesh.naga...@gmail.com wrote:
From: Ramesh Nagappa ramesh.naga...@ericsson.com
The retry loop in neigh_resolve_output() and neigh_connected_output()
call dev_hard_header() with out reseting the skb to network_header.
This causes the retry to fail with
On Thu, 2012-10-04 at 18:29 +0200, Eric Dumazet wrote:
On Thu, 2012-10-04 at 18:02 +0200, Maxime Bizon wrote:
On Fri, 2012-08-31 at 19:21 -0700, Hugh Dickins wrote:
Hi,
Francois is right that a GFP_ATOMIC allocation from pskb_expand_head()
is failing, which can easily happen
On Fri, 2012-10-05 at 12:49 +0200, Maxime Bizon wrote:
On Fri, 2012-10-05 at 09:41 +0200, Eric Dumazet wrote:
By the way, the commit you pointed has no effect on the reallocation
performed by pskb_expand_head() :
The commit has a side effect, because the problem appeared after
On Fri, 2012-10-05 at 14:22 +0200, Eric Dumazet wrote:
On Fri, 2012-10-05 at 12:49 +0200, Maxime Bizon wrote:
On Fri, 2012-10-05 at 09:41 +0200, Eric Dumazet wrote:
By the way, the commit you pointed has no effect on the reallocation
performed by pskb_expand_head() :
The commit
On Fri, 2012-10-05 at 14:37 +0200, Eric Dumazet wrote:
diff --git a/net/core/skbuff.c b/net/core/skbuff.c
index cdc2859..f6c1f52 100644
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -1053,11 +1053,22 @@ int pskb_expand_head(struct sk_buff *skb, int nhead,
int ntail,
{
int i
On Fri, 2012-10-05 at 14:51 +0200, Maxime Bizon wrote:
New convention would be : pass number of needed bytes after current
tail, not after current end.
Fully agree on this
Here is the proposal :
Change all occurrences of :
pskb_expand_head(skb, 0, 0, gfp)
to
pskb_realloc_head(skb,
On Fri, 2012-10-05 at 16:50 +0200, Maxime Bizon wrote:
On Fri, 2012-10-05 at 15:02 +0200, Eric Dumazet wrote:
On Fri, 2012-10-05 at 14:51 +0200, Maxime Bizon wrote:
New convention would be : pass number of needed bytes after current
tail, not after current end.
Fully agree
On Fri, 2012-10-05 at 17:15 +0200, Maxime Bizon wrote:
You think removing skb_recycle() is too big a change for stable ?
Driver change is simple, as recycling is not guaranteed today you have
this:
if (!try_recycle(skb))
skb = alloc_skb()
we just remove the try_recycle part, we
On Fri, 2012-10-05 at 12:33 -0400, Ramesh Nagappa wrote:
Hi Eric,
Yes, that is a good optimization. neigh_resolve_output() also has the
__skb_pull() outside the loop, is that required ? The changes would be
like ...
neigh_resolve_output()
...
-__skb_pull(skb,
On Sat, 2012-10-06 at 11:52 +, Chini, Georg (HP App Services) wrote:
Hello,
I have a network issue with kernel 3.6. My machine is set
up as router and with 3.6 it sometimes silently drops reply
packets instead of forwarding.
Example tcpdump on the router for ping www.heise.de from a
On Sat, 2012-10-06 at 17:52 +, Chini, Georg (HP App Services) wrote:
Hello,
Looks like the bug fixed by :
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux.git;a=commitdiff;h=f4ef85bbda96324785097356336bc79cdd37db0a
(David Miller will send it to stable team shortly)
On Tue, 2012-11-06 at 20:39 -0500, Dave Jones wrote:
On Tue, Nov 06, 2012 at 04:15:35PM -0800, Julius Werner wrote:
tcp_recvmsg contains a sanity check that WARNs when there is a gap
between the socket's copied_seq and the first buffer in the
sk_receive_queue. In theory, the TCP stack
On Wed, 2012-11-07 at 10:54 -0500, Dave Jones wrote:
It sounds more appropriate to me, instead of silently wedging the box.
At least with that approach we have a chance of finding out what happened.
Its quite the opposite.
If bug is still there 6 months after the commits that broke the
On Wed, 2012-11-07 at 11:43 -0500, Dave Jones wrote:
dude, look at the bug reports I just pointed you at.
People _are_ aware there are bugs there.
If I remember well, I helped to fix some of them.
If you turn that into a BUG() those reports would never have been filed.
How is that
On Wed, 2012-11-07 at 11:33 -0800, Julius Werner wrote:
tcp_recvmsg contains a sanity check that WARNs when there is a gap
between the socket's copied_seq and the first buffer in the
sk_receive_queue. In theory, the TCP stack makes sure that This Should
Never Happen (TM)... however, practice
On Wed, 2012-11-07 at 13:14 -0800, Julius Werner wrote:
What I find very sad in all this is that you didnt mention the driver
that was triggering this bug.
Sorry, I was just trying to keep this thread focussed on one patch.
The bug report that led me to this is publicly accessible at
On Wed, 2012-11-07 at 15:33 -0800, Eric Dumazet wrote:
So you probably are fighting a bug we already fixed in upstream kernel.
(commit c8628155ece363 tcp: reduce out_of_order memory use did not
played well with cloned skbs.)
This issue was already discussed on netdev in the past.
If you
On Fri, 2012-10-19 at 09:03 +0900, JoonSoo Kim wrote:
Hello, Eric.
Thank you very much for a kind comment about my question.
I have one more question related to network subsystem.
Please let me know what I misunderstand.
2012/10/14 Eric Dumazet eric.duma...@gmail.com:
In latest kernels
On Fri, 2012-10-19 at 11:50 +0400, Stanislav Kinsbursky wrote:
v4:
1) a couple of coding style fixes (lines over 80 characters)
v3:
1) hash calculation simlified to improve perfomance.
v2:
1) Hash table become RCU-friendly. Hash table search now done under RCU lock
protection.
This
On Fri, 2012-10-19 at 13:38 +0400, Stanislav Kinsbursky wrote:
19.10.2012 11:56, Eric Dumazet пишет:
I wonder if some applications relied on our idr, assuming they would get
low values for their timer id.
(We could imagine some applications use a table indexed by the timer id)
Hmm
On Mon, 2012-10-22 at 09:46 -0700, Peter LaDow wrote:
I posted this problem some time back on the linux-rt-users and
netfilter lists. Since then, we thought we had a workaround to avoid
this problem, so we dropped the issue. But now 5 months later, the
problem has reappeared. And this time
On Tue, 2012-10-23 at 05:38 +, Jongman Heo wrote:
Hmm,
I've just met the issue, with the commit 5640f768 reverted.
It seems that the issue does not always happen. So, my bisection may not be
correct.
At this moment, I don't have enough time to do bisection again..
Regards.
What
On Tue, 2012-10-23 at 11:29 +0900, JoonSoo Kim wrote:
2012/10/22 Christoph Lameter c...@linux.com:
On Sun, 21 Oct 2012, Joonsoo Kim wrote:
kmalloc() and kmalloc_node() of the SLUB isn't inlined when @flags =
__GFP_DMA.
This patch optimize this case,
so when @flags = __GFP_DMA, it
skinsbur...@parallels.com
---
SGTM
Signed-off-by: Eric Dumazet eduma...@google.com
Thanks !
--
To unsubscribe from this list: send the line unsubscribe linux-kernel in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read
On Tue, 2012-10-23 at 08:17 +, Jongman Heo wrote:
FYI, vmxnet3 driver is used for ethernet.
Yes, this driver needs some changes
#define VMXNET3_MAX_TX_BUF_SIZE (1 14)
Thats 16KB
As we can now provide up to 32KB fragments we broke something.
vmxnet3_tq_xmit() needs to split large
On Tue, 2012-10-23 at 03:02 -0700, Shreyas Bhatewara wrote:
Please dont top post on netdev or lkml
Well, actually the driver does split large frags into frags of
VMXNET3_MAX_TX_BUF_SIZE bytes each.
vmxnet3_drv.c
711 while (len) {
712 u32 buf_size;
713
714
On Tue, 2012-10-23 at 15:50 +0200, Eric Dumazet wrote:
Only the skb head is handled in the code you copy/pasted.
You need to generalize that to code in lines ~754
Then, the number of estimated descriptors is bad :
/* conservatively estimate # of descriptors to use */
count
On Tue, 2012-10-23 at 23:47 +0200, Thomas Gleixner wrote:
Not so good to me.
Signed-off-by: Eric Dumazet eduma...@google.com
And that should be either an Acked-by or a Reviewed-by. You can't sign
off on patches which have not been submitted or transported by you.
I actually gave some
On Wed, 2012-10-24 at 00:33 +0200, Thomas Gleixner wrote:
On Tue, 23 Oct 2012, Eric Dumazet wrote:
On Tue, 2012-10-23 at 23:47 +0200, Thomas Gleixner wrote:
Not so good to me.
Signed-off-by: Eric Dumazet eduma...@google.com
And that should be either an Acked
On Tue, 2012-10-23 at 17:15 -0700, Peter LaDow wrote:
(Sorry for the subject change, but I wanted to try and pull in those
who work on RT issues, and the subject didn't make that obvious.
Please search for the same subject without the RT Linux trailing
text.)
Well, more information. Even
On Wed, 2012-10-24 at 09:30 -0700, Peter LaDow wrote:
On Tue, Oct 23, 2012 at 9:32 PM, Eric Dumazet eric.duma...@gmail.com wrote:
Could you try following patch ?
Thanks for the suggestion. But I have a question about the patch below.
+ /* Note : cmpxchg() is a memory barrier, we
On Tue, 2012-10-09 at 10:52 +1100, Michael Neuling wrote:
The following patch:
acb600d net: remove skb recycling
added dev_free_skb() to drivers/net/ethernet/freescale/ucc_geth.c
This is a typo and should be dev_kfree_skb(). This fixes this.
Signed-off-by: Michael Neuling
On Tue, 2012-07-31 at 14:42 +0400, Konstantin Khlebnikov wrote:
A long time ago, in v2.4, VM_RESERVED kept swapout process off VMA,
currently it lost original meaning but still has some effects:
| effect | alternative flags
On Tue, 2012-10-09 at 15:36 +0300, Dan Carpenter wrote:
Add netdev to the CC list.
netdev already in the CC list by Borislav Petkov
Reporter was (kindly) requested to try 3.6-rc7 +, and we got no answer.
--
To unsubscribe from this list: send the line unsubscribe linux-kernel in
the body of
201 - 300 of 5670 matches
Mail list logo