[PATCH][VT] fix bug that service os&vmx guest can't communicate with each other.

[PATCH][VT] fix bug that service os&vmx guest can't communicate with each other.

[Edwin Zhai]

fix bug that service os&vmx guest can't communicate with each other.
Signed-off-by: Edwin Zhai <edwin.zhai@intel.com>
Signed-off-by: Xiaofeng Ling <xiaofeng.ling@intel.com>

diff -r 04ca47c298b5 -r a80889c7da99 tools/ioemu/hw/pcnet.c
--- a/tools/ioemu/hw/pcnet.c	Thu Sep  1 21:30:51 2005
+++ b/tools/ioemu/hw/pcnet.c	Mon Sep  5 09:33:29 2005
@@ -380,10 +380,13 @@
     return sizeof(s->buffer)-16;
 }
 
+#define MIN_BUF_SIZE 60
+
 static void pcnet_receive(void *opaque, const uint8_t *buf, int size)
 {
     PCNetState *s = opaque;
     int is_padr = 0, is_bcast = 0, is_ladr = 0;
+    uint8_t buf1[60];
 
     if (CSR_DRX(s) || CSR_STOP(s) || CSR_SPND(s) || !size)
         return;
@@ -391,6 +394,14 @@
 #ifdef PCNET_DEBUG
     printf("pcnet_receive size=%d\n", size);
 #endif
+
+    /* if too small buffer, then expand it */
+    if (size < MIN_BUF_SIZE) {
+        memcpy(buf1, buf, size);
+        memset(buf1 + size, 0, MIN_BUF_SIZE - size);
+        buf = buf1;
+        size = MIN_BUF_SIZE;
+    }
 
     if (CSR_PROM(s) 
         || (is_padr=padr_match(s, buf, size)) 

Re: [PATCH][VT] fix bug that service os&vmx guest can't communicate with each other.

[Keir Fraser]

On 8 Sep 2005, at 11:06, Edwin Zhai wrote:

> fix bug that service os&vmx guest can't communicate with each other.
> Signed-off-by: Edwin Zhai <edwin.zhai@intel.com>
> Signed-off-by: Xiaofeng Ling <xiaofeng.ling@intel.com>

Why is special treatment needed for small packets? Is this simply 
masking a bug elsewhere -- the original qemu driver must have dealt 
with small packets (e.g., TCP acks) correctly.

  -- Keir

Re: [PATCH][VT] fix bug that service os&vmx guest can't communicate with each other.

[Zhai, Edwin]

Keir Fraser wrote:

>
> On 8 Sep 2005, at 11:06, Edwin Zhai wrote:
>
>> fix bug that service os&vmx guest can't communicate with each other.
>> Signed-off-by: Edwin Zhai <edwin.zhai@intel.com>
>> Signed-off-by: Xiaofeng Ling <xiaofeng.ling@intel.com>
>
>
> Why is special treatment needed for small packets? Is this simply 
> masking a bug elsewhere -- the original qemu driver must have dealt 
> with small packets (e.g., TCP acks) correctly.

extend the ether packet to 60 bytes else the guest ethernet driver will 
report error.
in fact, the patch is borrowed from qemu ne2000.

>
>  -- Keir
>

Re: [PATCH][VT] fix bug that service os&vmx guest can't communicate with each other.

[Keir Fraser]

On 8 Sep 2005, at 13:41, Zhai, Edwin wrote:

>> Why is special treatment needed for small packets? Is this simply 
>> masking a bug elsewhere -- the original qemu driver must have dealt 
>> with small packets (e.g., TCP acks) correctly.
>
> extend the ether packet to 60 bytes else the guest ethernet driver 
> will report error.
> in fact, the patch is borrowed from qemu ne2000.

Fair enough. I guess you will push this patch to the vx-testing tree?

  -- Keir

Re: [PATCH][VT] fix bug that service os&vmx guest can't communicate with each other.

[Zhai, Edwin]

Keir Fraser wrote:

>
> On 8 Sep 2005, at 13:41, Zhai, Edwin wrote:
>
>>> Why is special treatment needed for small packets? Is this simply 
>>> masking a bug elsewhere -- the original qemu driver must have dealt 
>>> with small packets (e.g., TCP acks) correctly.
>>
>>
>> extend the ether packet to 60 bytes else the guest ethernet driver 
>> will report error.
>> in fact, the patch is borrowed from qemu ne2000.
>
>
> Fair enough. I guess you will push this patch to the vx-testing tree?

it has already gone into xen-vtx-unstable tree :)

>
>  -- Keir
>