> Also, I've written a test for this new feature:
> http://sourceforge.net/p/strace/code/ci/fdfa72276a16d0bd42b795aef3ac6c8969210469/
>
Thank you Dmitry !
- Zubin
--
Meet PCI DSS 3.0 Compliance Requirements with EventL
On Thu, Sep 18, 2014 at 01:33:54AM +0400, Dmitry V. Levin wrote:
> Hi Zubin,
>
> On Tue, Aug 26, 2014 at 06:28:46PM +0200, Zubin Mithra wrote:
> > > I think it would be less confusing to print both src and dst address:port
> > > when dst address:port is non-zero, for example:
> > > accept(3<0.0.0.
Zubin, congratulation!
This is the feature really I need.
My understandig is that this is the initial step for developing a tool
unifying strace, lsof, and tcpdump:)
Masatake YAMATO
> Hi Zubin,
>
> On Tue, Aug 26, 2014 at 06:28:46PM +0200, Zubin Mithra wrote:
>> > I think it would be less conf
Hi Zubin,
On Tue, Aug 26, 2014 at 06:28:46PM +0200, Zubin Mithra wrote:
> > I think it would be less confusing to print both src and dst address:port
> > when dst address:port is non-zero, for example:
> > accept(3<0.0.0.0:7171>, {sa_family=AF_INET, sin_port=htons(48435),
> > sin_addr=inet_addr("
On Tue, Aug 26, 2014 at 06:30:10PM +0200, Zubin Mithra wrote:
> >> Hi Dmitry,
> >>
> >>> Actually, it is quite linux specific: SOCK_DIAG_BY_FAMILY needs tcp_diag
> >>> kernel module to handle IPPROTO_TCP, and udp_diag - for IPPROTO_UDP
> >>> sockets.
> >>
> >> I’m not sure I understand the proble
>> Hi Dmitry,
>>
>>> Actually, it is quite linux specific: SOCK_DIAG_BY_FAMILY needs tcp_diag
>>> kernel module to handle IPPROTO_TCP, and udp_diag - for IPPROTO_UDP
>>> sockets.
>>
>> I’m not sure I understand the problem here — could you tell me which kernel
>> version you had used so that I
> I think it would be less confusing to print both src and dst address:port
> when dst address:port is non-zero, for example:
> accept(3<0.0.0.0:7171>, {sa_family=AF_INET, sin_port=htons(48435),
> sin_addr=inet_addr("127.0.0.1")}, [16]) = 4<127.0.0.1:7171->127.0.0.1:48435>
>
Got it, thanks Dmitr
On Sun, Aug 24, 2014 at 04:38:51PM +0200, Zubin Mithra wrote:
> Hi Dmitry,
>
> > Actually, it is quite linux specific: SOCK_DIAG_BY_FAMILY needs tcp_diag
> > kernel module to handle IPPROTO_TCP, and udp_diag - for IPPROTO_UDP
> > sockets.
>
> I’m not sure I understand the problem here — could you
On Sun, Aug 24, 2014 at 04:44:23PM +0200, Zubin Mithra wrote:
> On 21-Aug-2014, at 4:27, Dmitry V. Levin wrote:
> > On Thu, Aug 21, 2014 at 02:53:46AM +0400, Dmitry V. Levin wrote:
> >> On Mon, Aug 11, 2014 at 02:21:01PM +0530, zubin.mit...@gmail.com wrote:
> > [...]
> >>if (!inet_ntop(diag_ms
Hi Dmitry,
On 21-Aug-2014, at 4:27, Dmitry V. Levin wrote:
> On Thu, Aug 21, 2014 at 02:53:46AM +0400, Dmitry V. Levin wrote:
>> On Mon, Aug 11, 2014 at 02:21:01PM +0530, zubin.mit...@gmail.com wrote:
> [...]
>> if (!inet_ntop(diag_msg->idiag_family, diag_msg->id.idiag_dst,
>>
Hi Dmitry,
> Actually, it is quite linux specific: SOCK_DIAG_BY_FAMILY needs tcp_diag
> kernel module to handle IPPROTO_TCP, and udp_diag - for IPPROTO_UDP
> sockets.
I’m not sure I understand the problem here — could you tell me which kernel
version you had used so that I could try compiling th
On Thu, Aug 21, 2014 at 02:53:46AM +0400, Dmitry V. Levin wrote:
> On Mon, Aug 11, 2014 at 02:21:01PM +0530, zubin.mit...@gmail.com wrote:
[...]
> if (!inet_ntop(diag_msg->idiag_family, diag_msg->id.idiag_dst,
> remote_addr_buf, size) || !remote_addr_buf[0])
>
On Thu, Aug 21, 2014 at 02:53:46AM +0400, Dmitry V. Levin wrote:
> On Mon, Aug 11, 2014 at 02:21:01PM +0530, zubin.mit...@gmail.com wrote:
[...]
> if (!inet_ntop(diag_msg->idiag_family, diag_msg->id.idiag_dst,
> remote_addr_buf, size) || !remote_addr_buf[0])
>
On Thu, Aug 21, 2014 at 02:53:46AM +0400, Dmitry V. Levin wrote:
[...]
> > + else if (nlh->nlmsg_type == NLMSG_ERROR)
> > + return -1;
>
>
> I tested this patch on a x86_64 system with a fresh kernel.
> It always returns NLMSG_ERROR.
> I suppose somethi
On Mon, Aug 11, 2014 at 02:21:01PM +0530, zubin.mit...@gmail.com wrote:
[...]
> diff --git a/configure.ac b/configure.ac
> index 9aeb3a6..21dcc0e 100644
> --- a/configure.ac
> +++ b/configure.ac
> @@ -228,8 +228,11 @@ AC_CHECK_HEADERS(m4_normalize([
> inttypes.h
> ioctls.h
> linux
>
>
>>
>>
> Unless I'm mistaken this header is present on all linux systems in
> /usr/include even with the dev packages installed.
>
>
Typo, sorry. "this header is *not* present on all"
Thanks,
-- zm
--
___
Hello all,
This is the latest patch for extended socket fd decoding. A few points :-
> +
> +#ifdef HAVE_LINUX_SOCK_DIAG_H
> +# include
> +#else
> +# define SOCK_DIAG_BY_FAMILY 20
> + struct sock_diag_req {
> + __u8sdiag_family;
> + __u8sdiag_protocol;
> + };
> +#endif
From: Zubin Mithra
* Makefile.am (strace_SOURCES): Add socketutils.c.
* configure.ac (AC_CHECK_HEADERS): Add linux/inet_diag.h,
linux/netlink.h, linux/sock_diag.h.
(AC_CHECK_TYPES): Add inet_diag_req_v2.
* defs.h (get_pagesize): Add declaration.
(printsockdetails): Add declaration.
* mem.c (get
From: Zubin Mithra
* Makefile.am (strace_SOURCES): Add socketutils.c.
* configure.ac (AC_CHECK_HEADERS): Add linux/inet_diag.h,
linux/netlink.h, linux/sock_diag.h.
(AC_CHECK_TYPES): Add inet_diag_req_v2.
* defs.h (get_pagesize): Add declaration.
(printsockdetails): Add declaration.
* mem.c (get
19 matches
Mail list logo