Peter Zijlstra wrote:
You say "critical resource isolation", but it is not the case - consider
NFS over UDP - remote side will not stop sending just because receiving
socket code drops data due to OOM, or IPsec or compression, which can
requires reallocation. There is no "critical resource iso
Luca <[EMAIL PROTECTED]> wrote:
> I get a warning from lockdep during boot; 2.6.18-rc3 don't have this
> warning. I see a similar report in the archive (I haven't found time to
> test the patch...):
>
> http://marc.theaimsgroup.com/?l=linux-netdev&m=115506258902757&w=2
It's the same issue.
Thank
Alexey Kuznetsov <[EMAIL PROTECTED]> wrote:
>
> Probably, unicasts sent due to NLM_F_ECHO should somehow override
> rcvbuf limits.
Actually I think the only safe solution is to allocate a separate
socket for multicast messages. In other words, if you want reliable
unicast reception on a socket,
On Sat, 12 Aug 2006 06:30:14 +0400
Alexey Dobriyan <[EMAIL PROTECTED]> wrote:
> On Fri, Aug 11, 2006 at 06:49:43PM -0700, Andrew Morton wrote:
> > On Sat, 12 Aug 2006 04:46:23 +0400
> > Alexey Dobriyan <[EMAIL PROTECTED]> wrote:
> >
> > > +#ifdef CONFIG_PM
> > > static int amd8111e_suspend(struct
On Fri, Aug 11, 2006 at 06:49:43PM -0700, Andrew Morton wrote:
> On Sat, 12 Aug 2006 04:46:23 +0400
> Alexey Dobriyan <[EMAIL PROTECTED]> wrote:
>
> > +#ifdef CONFIG_PM
> > static int amd8111e_suspend(struct pci_dev *pci_dev, pm_message_t state)
> > {
> > struct net_device *dev = pci_get_drvd
On Sat, 12 Aug 2006 04:46:23 +0400
Alexey Dobriyan <[EMAIL PROTECTED]> wrote:
> +#ifdef CONFIG_PM
> static int amd8111e_suspend(struct pci_dev *pci_dev, pm_message_t state)
> {
> struct net_device *dev = pci_get_drvdata(pci_dev);
> @@ -1873,7 +1879,7 @@ static int amd8111e_resume(struc
Compile out dead code for CONFIG_PM=n users.
Signed-off-by: Alexey Dobriyan <[EMAIL PROTECTED]>
---
drivers/net/amd8111e.c | 10 +-
drivers/net/b44.c |4
drivers/net/bnx2.c |4
drivers/net/tg3.c |4
4 files changed, 21 insertions(+), 1 deletion(
On Fri, Aug 11, 2006 at 05:27:30PM -0700, David Miller wrote:
>
> E1000 wants 16K buffers for jumbo MTU settings.
>
> The reason is that the chip can only handle power-of-2 buffer
> sizes, and next hop from 9K is 16K.
>
> It is not possible to tell the chip to only accept 9K packets, you
> must
From: Alexey Kuznetsov <[EMAIL PROTECTED]>
Date: Fri, 11 Aug 2006 18:00:19 +0400
> > The e1000 issue is just one example of this, another
>
> What is this issue?
E1000 wants 16K buffers for jumbo MTU settings.
The reason is that the chip can only handle power-of-2 buffer
sizes, and next hop fro
From: Steven Whitehouse <[EMAIL PROTECTED]>
Date: Fri, 11 Aug 2006 16:54:06 +0100
> Here is a fix to the DECnet rules compare function where we used
> 32bit values rather than 16bit values. Spotted by Patrick McHardy.
>
> Cc: Patrick McHardy <[EMAIL PROTECTED]>
> Cc: Patrick Caulfield <[EMAIL PRO
From: Steven Whitehouse <[EMAIL PROTECTED]>
Date: Fri, 11 Aug 2006 15:27:42 +0100
> Here is a fix to Patrick McHardy's increase number of routing tables patch
> for DECnet. I did just test this and it appears to be working fine with
> this patch.
>
> Cc: Patrick McHardy <[EMAIL PROTECTED]>
> Cc:
From: Stephen Hemminger <[EMAIL PROTECTED]>
Date: Fri, 11 Aug 2006 15:13:29 -0700
> Some improvements to robust name interface. These API's are safe
> now by convention, but it is worth providing some safety checks
> against future bugs.
>
> Signed-off-by: Stephen Hemminger <[EMAIL PROTECTED]>
Some improvements to robust name interface. These API's are safe
now by convention, but it is worth providing some safety checks
against future bugs.
Signed-off-by: Stephen Hemminger <[EMAIL PROTECTED]>
---
net/core/dev.c |7 +--
1 file changed, 5 insertions(+), 2 deletions(-)
--- net-
> --- linux-2.6.18-rc4-orig/drivers/net/ehea/ehea.h 1969-12-31
> +extern void exit(int);
Should be able to remove that prototype :)
Anton
-
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.k
Hi,
> --- linux-2.6.18-rc4-orig/drivers/net/ehea/ehea_ethtool.c 1969-12-31
> +static void netdev_get_pauseparam(struct net_device *dev,
> + struct ethtool_pauseparam *pauseparam)
> +{
> + printk("get pauseparam\n");
> +}
There are a number of stubbed out e
* Alexey Kuznetsov <[EMAIL PROTECTED]> 2006-08-11 19:35
> Well, tc was supposed to use it, but this did not happen and
> it remained deficient.
Makes sense, especially for auto generated handles. I've been listening
to the notifications on a separate socket for this purpose. It would
make sense ho
Hi,
> drivers/net/ehea/ehea.h| 452
> +#define EHEA_DRIVER_NAME "IBM eHEA"
You are using this for ethtool get_drvinfo. Im not sure if it should
match the module name, and I worry about having a space in the name. Any
ideas on what we should be doing here?
> +#define NET_IP_ALIGN 0
S
Hi-
Jan-Bernd Themann wrote:
> +static inline long ehea_hcall_9arg_9ret(unsigned long opcode,
> + unsigned long arg1,
> + unsigned long arg2,
> + unsigned long arg3,
> +
Nathan Lynch wrote:
> Hope all the callers of this function are in non-atomic context (but I
> wasn't able to find any callers?).
Never mind, I somehow missed the users of ehea_hcall_9arg_9ret in this
patch, sorry.
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of
Hi,
> --- linux-2.6.18-rc4-orig/drivers/net/ehea/ehea_phyp.c1969-12-31
> 16:00:00.0 -0800
> +u64 ehea_h_alloc_resource_eq(const u64 hcp_adapter_handle,
...
> +u64 hipz_h_reregister_pmr(const u64 adapter_handle,
...
> +static inline int hcp_galpas_ctor(struct h_galpas *galpas,
B
Hi,
> --- linux-2.6.18-rc4-orig/drivers/net/ehea/ehea_main.c1969-12-31
> +#define DEB_PREFIX "main"
Doesnt appear to be used.
> +static struct net_device_stats *ehea_get_stats(struct net_device *dev)
...
> + cb2 = kzalloc(H_CB_ALIGNMENT, GFP_KERNEL);
I cant see where this gets fr
On Friday 11 August 2006 21:31, Linas Vepstas wrote:
> I put my name at the top when I was the primary author.
> I put Jim's name at the top when he was the primary author.
>
> Both names are there because I sat in Jim's office and used
> his keyboard. I got him to compile and run the tests on
>
Constify largish areas of firmware data in Tigon3 ethernet driver.
non-const:
lsmod:
tg3 101404 0
objdump -x:
.rodata 03e8
.data 4a0c
ls -l:
-rw-r--r-- 1 root root 114404 2006-08-19 21:36 drivers/net/tg3.ko
const:
lsmod:
tg3 101404 0
objdump -x
Sébastien Dugué wrote:
>aio completion notification
I looked over this now but I don't think I understand everything. Or I
don't see how it all is integrated. And no, I'm not looking at the
proposed glibc code since would mean being tainted.
> Details:
> ---
>
> A st
Hi Olof,
Olof Johansson <[EMAIL PROTECTED]> writes:
> On Fri, Aug 11, 2006 at 12:11:17PM -0500, Linas Vepstas wrote:
>
> > This patch adds version information as reported by
> > ethtool -i to the Spidernet driver.
>
> Why does a driver that's in the mainline kernel need to have a version
> num
On Fri, Aug 11, 2006 at 07:44:39PM +0200, Sam Ravnborg wrote:
> >
> > These have been well-tested over the last few weeks. Please apply.
> Hi Linas.
> Just noticed a nit-pick detail.
> The general rule is to add your Signed-off-by: at the bottom of the
> patch, so the top-most Signed-of-by: is al
On Fri, Aug 11, 2006 at 12:11:17PM -0500, Linas Vepstas wrote:
> This patch adds version information as reported by
> ethtool -i to the Spidernet driver.
Why does a driver that's in the mainline kernel need to have a version
number besides the kernel version?
I can understand it for drivers lik
On Fri, Aug 11, 2006 at 12:03:37PM -0500, Linas Vepstas wrote:
>
> The following series of patches implement some fixes and performance
> improvements for the Spedernet ethernet device driver. The high point
> of the patch series is some code to implement a low-watermark interrupt
> on the transmi
On Friday 11 August 2006 10:34, Johannes Berg wrote:
> This is running wireless-dev from yesterday. All I did was plug in a
> rt2500usb device into a usb port on a freshly booted system. I have a
> feeling that this is could be one of the problems reported earlier with
> the d80211 stack, but I
Linas Vepstas wrote:
The following series of patches implement some fixes and performance
improvements for the Spedernet ethernet device driver. The high point
of the patch series is some code to implement a low-watermark interrupt
on the transmit queue. The bundle of patches raises transmit perf
Turn off mis-interpretation of the queue-empty interrupt
status bit as an error. This bit is set as a part of
the previous low-watermark patch.
Signed-off-by: Linas Vepstas <[EMAIL PROTECTED]>
Signed-off-by: James K Lewis <[EMAIL PROTECTED]>
Cc: Utz Bacher <[EMAIL PROTECTED]>
Cc: Jens Osterkamp
This patch adds version information as reported by
ethtool -i to the Spidernet driver.
Signed-off-by: James K Lewis <[EMAIL PROTECTED]>
Signed-off-by: Linas Vepstas <[EMAIL PROTECTED]>
Cc: Utz Bacher <[EMAIL PROTECTED]>
Cc: Jens Osterkamp <[EMAIL PROTECTED]>
drivers/net/spider_net.c
This patch increases the Burst Address alignment from 64 to 1024 in the
Spidernet driver. This improves transmit performance for large packets
from about 100Mbps to 300-400Mbps.
Signed-off-by: James K Lewis <[EMAIL PROTECTED]>
Signed-off-by: Linas Vepstas <[EMAIL PROTECTED]>
Cc: Utz Bacher <[EMA
Implement basic low-watermark support for the transmit queue.
The basic idea of a low-watermark interrupt is as follows.
The device driver queues up a bunch of packets for the hardware
to transmit, and then kicks he hardware to get it started.
As the hardware drains the queue of pending, untrans
The following series of patches implement some fixes and performance
improvements for the Spedernet ethernet device driver. The high point
of the patch series is some code to implement a low-watermark interrupt
on the transmit queue. The bundle of patches raises transmit performance
from some emb
On Thu, 10 Aug 2006 20:25:40 -0400
Andy Gay <[EMAIL PROTECTED]> wrote:
> [IPROUTE]: Fix struct alignment with cris architecture
>
> gcc for the cris arch does not pad structures to the next multiple of 4
> bytes, as the i386 gcc does.
>
> This causes errors like this when displaying xfrm policie
Hi,
I get a warning from lockdep during boot; 2.6.18-rc3 don't have this
warning. I see a similar report in the archive (I haven't found time to
test the patch...):
http://marc.theaimsgroup.com/?l=linux-netdev&m=115506258902757&w=2
but my stacktrace is a bit different, so I'm reporting this one t
Mikey,
first of all thanks a lot for the effort you invested to review our code.
We're quite happy about the improvements we made due to your comments.
See our answers below.
Kind regards
Thomas
Michael Neuling wrote:
Please add comments to make the code more readable, especially at the
star
Here is a fix to the DECnet rules compare function where we used
32bit values rather than 16bit values. Spotted by Patrick McHardy.
Cc: Patrick McHardy <[EMAIL PROTECTED]>
Cc: Patrick Caulfield <[EMAIL PROTECTED]>
Signed-off-by: Steven Whitehouse <[EMAIL PROTECTED]>
diff --git a/net/decnet/dn_
On Fri, 11 Aug 2006 12:40:10 +0400
Evgeniy Polyakov <[EMAIL PROTECTED]> wrote:
>
> poll/select() notifications. Timer notifications.
>
> This patch includes generic poll/select and timer notifications.
>
> kevent_poll works simialr to epoll and has the same issues (callback
> is invoked not fro
Hello!
> I get your point and I see the value. Unfortunately, probably due to
> lack of documentation, this feature isn't used by any applications I
> know of.
Well, tc was supposed to use it, but this did not happen and
it remained deficient.
> We even put in the hacks to make identification o
Hi,
On Fri, Aug 11, 2006 at 05:22:17PM +0200, Patrick McHardy wrote:
> Steven Whitehouse wrote:
> > Here is a fix to Patrick McHardy's increase number of routing tables patch
> > for DECnet. I did just test this and it appears to be working fine with
> > this patch.
> >
> > Cc: Patrick McHardy <[
Steven Whitehouse wrote:
> Here is a fix to Patrick McHardy's increase number of routing tables patch
> for DECnet. I did just test this and it appears to be working fine with
> this patch.
>
> Cc: Patrick McHardy <[EMAIL PROTECTED]>
> Cc: Patrick Caulfield <[EMAIL PROTECTED]>
> Signed-off-by: Ste
Remove broken URLs (www.scyld.com) from network drivers' logging output.
URLs in comments and other strings are left intact.
Signed-off-by: Markus Dahms <[EMAIL PROTECTED]>
---
I was tired of always seeing an URL not working anymore on initialization
of 3c59x and natsemi. So this is an attempt t
Here is a fix to Patrick McHardy's increase number of routing tables patch
for DECnet. I did just test this and it appears to be working fine with
this patch.
Cc: Patrick McHardy <[EMAIL PROTECTED]>
Cc: Patrick Caulfield <[EMAIL PROTECTED]>
Signed-off-by: Steven Whitehouse <[EMAIL PROTECTED]>
di
Hello!
> I'd be interested in any opinions on the above mentioned effect.
Everything is right, it is exactly how it works.
Well, use another qdisc, which counts in bytes rather than in frames
(f.e. bfifo)
Set sndbuf small enough.
And if sndbuf*#senders is still too large, you have to use fair
Hello!
>> management schemes and to just wrap SKB's around
>> arbitrary pieces of data.
+
> and something clever like a special page_offset encoding
> means "use data, not page".
But for what purpose do you plan to use it?
> The e1000 issue is just one example of this, another
What is this iss
[PPP]: carrier/operstate support
Add support for setting carrier state and operstate of ppp devices.
The patch works as follows:
- Initially, ppp devices come up with carrier off and npmode set to NPMODE_DROP
- In non-demand mode (dialin or permanent connection), when the IP*CP
protocols trans
Hi,
+#define EHEA_EQE_SM_MECH_NUMBER EHEA_BMASK_IBM(48, 55)
+#define EHEA_EQE_SM_PORT_NUMBER EHEA_BMASK_IBM(56, 63)
+
+struct ehea_eqe {
+ u64 entry;
+};
ehea_ege.. what is that and why a struct if only 1 item? Comments
please.
There are send / receive queue elements (ehea_swqe
Hello!
> Did I miss some way that multiple file objects can point to the
> same socket inode?
Absolutely prohibited. Always was.
Apparently, sock_fasync() was cloned from tty_fasync(), that's the only
reason why it is so creepy.
Alexey
-
To unsubscribe from this list: send the line "unsubscribe
Hello,
a while ago I wrote a simple network load generator to inject datagrams or
frames at maximum rate into a network. Maybe I was mistaken but I expected
the socket's send operation to block, if the transmitting network device
becomes saturated (no matter if using UDP or PF_PACKET). However,
Hi Christian,
thanks for your comments, we'll send an updated patch set soon.
Jan-Bernd
Christian Borntraeger wrote:
Hi Jan-Bernd,
I had some minutes, here are some finding after a quick look.
On Wednesday 09 August 2006 10:38, you wrote:
+static struct net_device_stats *ehea_get_stats(stru
From: David Miller <[EMAIL PROTECTED]>
Date: Fri, 11 Aug 2006 03:15:16 -0700 (PDT)
> If my analysis is correct we can incredibly simplify sock_fasync().
I've also found more bugs in sock_fasync(), it's a real can of worms
:-)
It deviates from the return value policies used by othe
file_ops->fasy
What was that about locking not having problems? :P
I was writing a small program that (using ioctls)
* creates a new interface (using sysfs)
* sets the interface to monitor mode
* sets IFF_UP
* (1)
* sets IFF_DOWN
* (2)
* destroy interface (using sysfs)
That was fine, but then I wanted to see
From: Thomas Graf <[EMAIL PROTECTED]>
Date: Fri, 11 Aug 2006 12:38:22 +0200
> Looks good. You could have used nla_memcpy() instead of
> memcpy() but there is no functional difference in this case.
Right, because the code already verifies that the attribute
length is at least dev->addr_len, and th
* David Miller <[EMAIL PROTECTED]> 2006-08-10 23:02
> Thomas, there was a conflict which I did expect since I added
> all the 2.6.18 bug fixes into the tree. The conflict is for
> your netlink conversion of the iflink stuff.
>
> I had to fix the -stable tree IFLA_ADDRESS handling because
> dev->s
From: Evgeniy Polyakov <[EMAIL PROTECTED]>
Date: Fri, 11 Aug 2006 14:28:20 +0400
> On Fri, Aug 11, 2006 at 03:15:16AM -0700, David Miller ([EMAIL PROTECTED])
> wrote:
> > Did I miss some way that multiple file objects can point to the
> > same socket inode?
>
> What about dup and pipe?
Dup make
On Fri, Aug 11, 2006 at 03:15:16AM -0700, David Miller ([EMAIL PROTECTED])
wrote:
> Did I miss some way that multiple file objects can point to the
> same socket inode?
What about dup and pipe?
--
Evgeniy Polyakov
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
th
I was studying sock_fasync() and it definitely has a bunch
of questionable issues.
Well firstly, it duplicates fasync_helper() entirely.
The only difference is that sock_fasync() does socket
local locking which is better for performance. fasync_helper()
uses a global spinlock to protect the fasy
> Michal,
> This looks correct, but I think a better way to do it is:
>
> in_dev = inetdev_by_index(...)
> (void) ip_mc_leave_src()
> if (in_dev) {
> ip_mc_dec_group()
> in_dev_put()
> }
>
> That way, sflist internal
* Michael Tokarev <[EMAIL PROTECTED]> 2006-08-11 10:56
> And while we're at it... How about using table *names* instead of
> numbers in kernel too, a-la iptables? Once possible number of tables
> is large, and we're using hashes for tables now anyway, keeping a
> name inside the table structure w
On Fri, 11 August 2006 09:28:26 +0200, Thomas Klein wrote:
> Michael Neuling wrote:
> +static inline u32 map_swqe_size(u8 swqe_enc_size)
> +static inline u32|map_rwqe_size(u8 rwqe_enc_size)
> >
> Agreed. Functions were replaced by a single map_wqe_size() function.
Just a general thing, t
On Friday 11 August 2006 08:45, David Miller wrote:
> From: Krzysztof Oledzki <[EMAIL PROTECTED]>
> Date: Thu, 10 Aug 2006 20:18:23 +0200 (CEST)
>
> > OK, this patch really solves the bug from my report. Are there any
> > chances for similar fix in the net-2.6.19.git?
>
> I'm still thinking about t
From: Christophe Devriese <[EMAIL PROTECTED]>
Date: Fri, 11 Aug 2006 10:50:44 +0200
> What can I do to get you to apply this then ? This patch is about
> fixing a bug which is bothering me a lot.
You need to be patient while I review the problem.
Nothing you say will allow my brain to operate an
This is running wireless-dev from yesterday. All I did was plug in a
rt2500usb device into a usb port on a freshly booted system. I have a
feeling that this is could be one of the problems reported earlier with
the d80211 stack, but I haven't mastered the art of picking through
these traces yet
Generic event handling mechanism.
Changes from 'take7' patchset:
* new mmap interface (not tested, waiting for other changes to be acked)
- use nopage() method to dynamically substitue pages
- allocate new page for events only when new added kevent requres it
- do not use
Core files.
This patch includes core kevent files:
- userspace controlling
- kernelspace interfaces
- initialization
- notification state machines
Signed-off-by: Evgeniy Polyakov <[EMAIL PROTECTED]>
diff --git a/arch/i386/kernel/syscall_table.S b/arch/i386/kernel/syscall_table.S
index dd63d
poll/select() notifications. Timer notifications.
This patch includes generic poll/select and timer notifications.
kevent_poll works simialr to epoll and has the same issues (callback
is invoked not from internal state machine of the caller, but through
process awake).
Timer notifications can b
Attached (sorry, can't seem to figure out how to convince thunderbird to
not mangle things under windows) is a patch to make d80211 have some LED
triggers.
However, I'm not sure where I should call the _init and _exit functions
and if I really should be using the local struct. It seems I shoul
Hey,
I managed to crash it again :P
Here's approximately what I did:
johannes:/home/johannes# ifconfig wlan0 down
johannes:/home/johannes# cd /sys/class/ieee80211/phy0/
johannes:/sys/class/ieee80211/phy0# echo -n moni0 > add_iface
johannes:/sys/class/ieee80211/phy0# iwconfig wlan0 mode mast
From: Michael Tokarev <[EMAIL PROTECTED]>
Date: Fri, 11 Aug 2006 11:02:43 +0400
> Patrick McHardy wrote:
> > --- a/include/linux/rtnetlink.h
> > +++ b/include/linux/rtnetlink.h
>
> > +static inline u32 rtm_get_table(struct rtattr **rta, u8 table)
> > +{
> > + return RTA_GET_U32(rta[RTA_TABLE-1]
On Fri, Aug 11, 2006 at 12:04:54AM -0700, Andrew Morton ([EMAIL PROTECTED])
wrote:
> > This area can be decreased down to 70mb by reducing amount of
> > information placed into the buffer (only user's data and flags) without
> > additional hints.
> >
>
> 70MB is still very bad, naturally.
Actua
Michael Neuling wrote:
+static inline u32 map_swqe_size(u8 swqe_enc_size)
+{
+ return 128 << swqe_enc_size;
+}^
+ |
+static inline u32|map_rwqe_size(u8 rwqe_enc_size)
+{|
+ return 128 << rwqe_enc_size;
On Fri, 11 Aug 2006 10:30:21 +0400
Evgeniy Polyakov <[EMAIL PROTECTED]> wrote:
> On Thu, Aug 10, 2006 at 11:23:40PM -0700, Andrew Morton ([EMAIL PROTECTED])
> wrote:
> > On Fri, 11 Aug 2006 10:15:35 +0400
> > Evgeniy Polyakov <[EMAIL PROTECTED]> wrote:
> >
> > > On Thu, Aug 10, 2006 at 05:56:39P
Patrick McHardy wrote:
> --- a/include/linux/rtnetlink.h
> +++ b/include/linux/rtnetlink.h
> +static inline u32 rtm_get_table(struct rtattr **rta, u8 table)
> +{
> + return RTA_GET_U32(rta[RTA_TABLE-1]);
> +rtattr_failure:
> + return table;
> +}
> +
What's that ? ;)
/mjt
-
To unsubscrib
David Miller wrote:
> From: Michael Tokarev <[EMAIL PROTECTED]>
[]
>>> - Use u32 for routing table IDs everywhere inside the kernel
>> Just out of curiocity: why current limit of 2^31 isn't sufficient?
>> Or am I missing the point?
>
> The current limit is 256 because the table member of the struc
On Thu, Aug 10, 2006 at 11:38:26PM -0700, David Miller ([EMAIL PROTECTED])
wrote:
> From: Evgeniy Polyakov <[EMAIL PROTECTED]>
> Date: Fri, 11 Aug 2006 10:33:53 +0400
>
> > That requires mmap hacks to substitute pages in run-time without user
> > notifications. I do not expect it is a good soluti
77 matches
Mail list logo
