svn commit: samba r21278 - in branches/SAMBA_3_0/source/smbd: .

[James Peach]

On 17/02/2007, at 3:32 AM, Volker Lendecke wrote:

> On Sun, Feb 11, 2007 at 05:59:53PM -0800, Jeremy Allison wrote:
>> The write code can potentially write 127k rather
>> than 65k so it might be interesting to compare
>> with that write size also. Can you run cachegrind
>> on both methods to see why the memcpy ends up
>> taking less time ?
>
> Hmmmm. Difficult. I just tried giving 127k as an argument to
> the Samba4 test. Samba3 does not like that request and
> neither does W2k3. Worse, issuing that request against
> leaves the file in an undeletable state. Rebooting now.

Eww. IIRC some versions of windows do accept large (>64K) writes, but  
not enough to make it worthwhile tring to figure out when it is safe  
for a client to use these.

> Do you have a sniff of a write of 127k?

I can get you one from the Mac OS X client.

--
James Peach | jpeach at samba.org