All of lore.kernel.org
 help / color / mirror / Atom feed
* [patch merge tree] netfront needs more headroom
@ 2005-12-07 15:48 Gerd Knorr
  2005-12-08 18:50 ` Christian Limpach
  0 siblings, 1 reply; 3+ messages in thread
From: Gerd Knorr @ 2005-12-07 15:48 UTC (permalink / raw)
  To: xen-devel; +Cc: Jan Beulich

[-- Attachment #1: Type: text/plain, Size: 436 bytes --]

   Hi,

Raised the headrum size from 200 to 512.  I've got "Received packet 
needs 120 bytes more headroom." in my syslog, so 200 is obviously too small.

Increased to 512 as the slab allocator gives out pieces of memory with 
sizes being powers of two, so the old allocation size already was 256 
and anything larger than that will end up being 512 anyway (unless we 
create our own memory pool with custom size ...)

cheers,

   Gerd

[-- Attachment #2: xen3-fixup-net-headroom --]
[-- Type: text/plain, Size: 644 bytes --]

Subject: xen netfront driver needs more headroom
From: kraxel@suse.de

Index: linux-2.6.14/drivers/xen/netfront/netfront.c
===================================================================
--- linux-2.6.14.orig/drivers/xen/netfront/netfront.c	2005-12-07 16:26:12.000000000 +0100
+++ linux-2.6.14/drivers/xen/netfront/netfront.c	2005-12-07 16:27:47.000000000 +0100
@@ -77,7 +77,7 @@
     } while (0)
 
 /* Allow headroom on each rx pkt for Ethernet header, alignment padding, ... */
-#define RX_HEADROOM 200
+#define RX_HEADROOM 512
 
 static unsigned long rx_pfn_array[NET_RX_RING_SIZE];
 static multicall_entry_t rx_mcl[NET_RX_RING_SIZE+1];

[-- Attachment #3: Type: text/plain, Size: 138 bytes --]

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [patch merge tree] netfront needs more headroom
  2005-12-07 15:48 [patch merge tree] netfront needs more headroom Gerd Knorr
@ 2005-12-08 18:50 ` Christian Limpach
  2005-12-09 11:23   ` Gerd Knorr
  0 siblings, 1 reply; 3+ messages in thread
From: Christian Limpach @ 2005-12-08 18:50 UTC (permalink / raw)
  To: Gerd Knorr; +Cc: xen-devel, Jan Beulich

On 12/7/05, Gerd Knorr <kraxel@suse.de> wrote:
> Raised the headrum size from 200 to 512.  I've got "Received packet
> needs 120 bytes more headroom." in my syslog, so 200 is obviously too small.
>
> Increased to 512 as the slab allocator gives out pieces of memory with
> sizes being powers of two, so the old allocation size already was 256
> and anything larger than that will end up being 512 anyway (unless we
> create our own memory pool with custom size ...)

We believe that this is fixed by recent changes to the network driver,
notably these changesets in xen-unstable:
linux-2.6-xen-sparse cset: 4e9c643968ffd0526258dc1ae7d1fee6123be718
linux-2.6-xen-sparse cset: c976c1ff806aa46c989fb5b1d84b311519fb0456

     christian

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [patch merge tree] netfront needs more headroom
  2005-12-08 18:50 ` Christian Limpach
@ 2005-12-09 11:23   ` Gerd Knorr
  0 siblings, 0 replies; 3+ messages in thread
From: Gerd Knorr @ 2005-12-09 11:23 UTC (permalink / raw)
  To: Christian.Limpach; +Cc: xen-devel, Jan Beulich

>> Raised the headrum size from 200 to 512.  I've got "Received packet
>> needs 120 bytes more headroom." in my syslog, so 200 is obviously too small.

> We believe that this is fixed by recent changes to the network driver,
> notably these changesets in xen-unstable:
> linux-2.6-xen-sparse cset: 4e9c643968ffd0526258dc1ae7d1fee6123be718
> linux-2.6-xen-sparse cset: c976c1ff806aa46c989fb5b1d84b311519fb0456

Yep, patch looks ok and the problem seems to be gone.

cheers,

   Gerd

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2005-12-09 11:23 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2005-12-07 15:48 [patch merge tree] netfront needs more headroom Gerd Knorr
2005-12-08 18:50 ` Christian Limpach
2005-12-09 11:23   ` Gerd Knorr

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.