On Thu, Feb 07, 2008, Robert Collins wrote:
>
> On Thu, 2008-02-07 at 14:08 +0900, Adrian Chadd wrote:
> > On Thu, Feb 07, 2008, Robert Collins wrote:
> >
> > > > Still, this is one of those "death of a thousand cuts" method of
> > > > killing performance..
> > >
> > > Right, I haven't seen the
On Thu, 2008-02-07 at 14:08 +0900, Adrian Chadd wrote:
> On Thu, Feb 07, 2008, Robert Collins wrote:
>
> > > Still, this is one of those "death of a thousand cuts" method of
> > > killing performance..
> >
> > Right, I haven't seen the commit; care to mail the diff?
>
> Which? I just looked at
On Thu, Feb 07, 2008, Robert Collins wrote:
> > Still, this is one of those "death of a thousand cuts" method of
> > killing performance..
>
> Right, I haven't seen the commit; care to mail the diff?
Which? I just looked at the places where Amos is calling GetAddrInfo()
and FreeAddrInfo(); more
On Thu, 2008-02-07 at 12:21 +0900, Adrian Chadd wrote:
>
> Well, I haven't removed the temporary malloc/free pair, whatever its
> called;
> I've just removed Amos' workaround in src/comm.cc so it doesn't leak
> on my
> system whilst I profile.
>
> Still, this is one of those "death of a thousand
On Wed, Feb 06, 2008, Alex Rousskov wrote:
> Ouch. I did not realize the two ended up in different classes.
yeah, neither did I until I started toying with it.
> doZeroOnPush has nothing to do with allocation and is not even used by
> MemAllocator, but it has to be there and not in MemPool becau
On Thu, 2008-02-07 at 09:13 +0900, Adrian Chadd wrote:
> Updated patch:
>
> http://www.creative.net.au/diffs/20080207-squid3-nonzero-buffers-2.diff
>
> doZeroOnPush can't be private as it then seems to be inaccessible from
> MemPool::push().
Ouch. I did not realize the two ended up in different
On Thu, Feb 07, 2008, Robert Collins wrote:
> >
> > I've removed that hack, and things work fine for me. Ubuntu 7.01 here,
> > x86_64.
>
> There's a useful c++ thing called 'placement new' - its used when you
> have classes that you want to use on the stack, but still have tight
> abstraction bo
On Thu, 2008-02-07 at 11:56 +0900, Adrian Chadd wrote:
> Argh, this temporary malloc/free pair is peppered throughout the codebase!
> Grr.
>
> I've removed that hack, and things work fine for me. Ubuntu 7.01 here, x86_64.
There's a useful c++ thing called 'placement new' - its used when you
ha
On Thu, Feb 07, 2008, Amos Jeffries wrote:
> > Amos, could you please poke the leak in IPAddress:GetAddrInfo a little?
>
> I'll give it another try. But don't hold your breath too long on that one.
>
> Did you see my note on it?
> I tracked that leak backwards from two segfaults to the connect()
On Thu, Feb 07, 2008, Amos Jeffries wrote:
> > Amos, could you please poke the leak in IPAddress:GetAddrInfo a little?
>
> I'll give it another try. But don't hold your breath too long on that one.
Why are you trying to allocate the structure on invocation of GetAddrInfo() ?
Wouldn't it be bette
On Thu, Feb 07, 2008, Amos Jeffries wrote:
> > Amos, could you please poke the leak in IPAddress:GetAddrInfo a little?
>
> I'll give it another try. But don't hold your breath too long on that one.
>
> Did you see my note on it?
> I tracked that leak backwards from two segfaults to the connect()
> Updated patch:
>
> http://www.creative.net.au/diffs/20080207-squid3-nonzero-buffers-2.diff
>
> doZeroOnPush can't be private as it then seems to be inaccessible from
> MemPool::push().
>
Try setting up a friend' relationship between MemPool and MemAllocator
classes. Or creating an inline accesso
> Amos, could you please poke the leak in IPAddress:GetAddrInfo a little?
I'll give it another try. But don't hold your breath too long on that one.
Did you see my note on it?
I tracked that leak backwards from two segfaults to the connect() system
call. Then I tracked the allocation from my init
> On Sat, Jan 26, 2008 at 01:16:14PM +1300, Amos Jeffries wrote:
>> Frank Fegert wrote:
>> >On Fri, Jan 25, 2008 at 09:25:24AM +1300, Amos Jeffries wrote:
>> >>>i did a quick hack and patched Solaris privileges support into
>> pinger.c
>> >>>from squid-2.6.STABLE18. This should allow to run pinger
Thanks, I'll try that out.
I can see henrik's point; this *could* be solved by improving the
documentation (and in any case, it needs it). However, I tend to like
what I think this patch is going to do; it's the least surprising thing.
Cheers,
On 03/02/2008, at 6:55 AM, Adrian Chadd wrote
Updated patch:
http://www.creative.net.au/diffs/20080207-squid3-nonzero-buffers-2.diff
doZeroOnPush can't be private as it then seems to be inaccessible from
MemPool::push().
Adrian
On Wed, Feb 06, 2008, Alex Rousskov wrote:
> Adrian,
>
> A few comments regarding the nonzero patch you po
Amos, could you please poke the leak in IPAddress:GetAddrInfo a little?
Thanks,
Adrian
==19601==
==19601== ERROR SUMMARY: 9353 errors from 5 contexts (suppressed: 8 from 2)
==19601== malloc/free: in use at exit: 36,429,514 bytes in 14,790 blocks.
==19601== malloc/free: 114,225 allocs, 99,435
>> Hi all,
>>
>> I wrote a small perl script which fixes (some of) the astyle problems.
>>
>> Before use it, you must adjust the $ASTYLE_BIN variable at the beggining
>> of the formatter.pl file.
>>
>> I am running it using the following command:
>> # find . -name "*.cc" -exec formater.pl \{\} \
> Hi all,
>
> I wrote a small perl script which fixes (some of) the astyle problems.
>
> Before use it, you must adjust the $ASTYLE_BIN variable at the beggining
> of the formatter.pl file.
>
> I am running it using the following command:
> # find . -name "*.cc" -exec formater.pl \{\} \;
> but i
On Sat, Jan 26, 2008 at 01:16:14PM +1300, Amos Jeffries wrote:
> Frank Fegert wrote:
> >On Fri, Jan 25, 2008 at 09:25:24AM +1300, Amos Jeffries wrote:
> >>>i did a quick hack and patched Solaris privileges support into pinger.c
> >>>from squid-2.6.STABLE18. This should allow to run pinger w/o setui
Hi all,
I wrote a small perl script which fixes (some of) the astyle problems.
Before use it, you must adjust the $ASTYLE_BIN variable at the beggining
of the formatter.pl file.
I am running it using the following command:
# find . -name "*.cc" -exec formater.pl \{\} \;
but it can take multip
Adrian,
A few comments regarding the nonzero patch you posted on IRC:
http://www.creative.net.au/diffs/20080207-squid3-nonzero-buffers.diff
First of all, it is a move in the right direction, of course. If you
verified that it works, it can be committed as is. Here are a few
nitpicks since you
22 matches
Mail list logo
