Re: discrepancy in ip(7) wrt. IP DF flag for UDP sockets

On Thu, Sep 22, 2011 at 06:15:53AM +0200, Michael Kerrisk wrote:

Ben, Neil,

On Tue, Sep 20, 2011 at 4:31 PM, Neil Horman [off-list ref] wrote:

quoted

On Tue, Sep 20, 2011 at 10:12:34AM -0400, Benjamin Poirier wrote:

quoted

On 11-09-20 09:38, Neil Horman wrote:

quoted

On Tue, Sep 20, 2011 at 09:29:54AM -0400, Benjamin Poirier wrote:

quoted

On 11-09-20 08:14, Michael Kerrisk wrote:

quoted

Hello Benjamin, Neil,

[snip]

quoted

quoted

quoted

Could you describe the required change in terms of how the man page
text should look--i.e., rewrite the passage as you think it should
look?

How about changing it to:
  IP_MTU_DISCOVER (since Linux 2.2)
  Set or receive the Path MTU Discovery setting for a socket. When
  enabled, the don't-fragment flag is set on all outgoing packets.
  Linux will perform Path MTU Discovery as defined in RFC 1191 on
  SOCK_STREAM sockets. For non-SOCK_STREAM sockets, it is the
  user's responsibility to packetize the data in MTU sized chunks
  and to do the retransmits if necessary. The kernel will reject
  (with EMSGSIZE) datagrams that are bigger than the known path
  MTU. The system-wide default is controlled by the
  /proc/sys/net/ipv4/ip_no_pmtu_disc file.

  Path MTU discovery flags   Meaning
  [...]

There are some differences between _DO and _WANT that are glossed over
in this description, but I suppose there's only so much detail you can
put in...

Thanks,
-Ben

Yeah, I think thats close, but its only the users responsibility to package
datagrams in mtu sized chunks if they force the dont fragment bit on.  If they
go wtih the default, the stack will fragment a datagram is it sees fit according
to the mtu of the path it traverses

Exactly. To get into this level of detail, I think we have to mention
the option value, not just enabled/disabled. Let's try like this:

      IP_MTU_DISCOVER (since Linux 2.2)
      Set or receive the Path MTU Discovery setting for a socket. When
      enabled, Linux will perform Path MTU Discovery as defined in RFC
      1191 on SOCK_STREAM sockets.  For non-SOCK_STREAM sockets,
      IP_PMTUDISC_DO forces the don't-fragment flag to be set on all
      outgoing packets.  It is the user's responsibility to packetize
      the data in MTU sized chunks and to do the retransmits if
      necessary.  The kernel will reject (with EMSGSIZE) datagrams
      that are bigger than the known path MTU.  IP_PMTUDISC_WANT will
      fragment a datagram if needed according to the path MTU or will
      set the don't-fragment flag otherwise.

      The system-wide default can be toggled between IP_PMTUDISC_WANT
      and IP_PMTUDISC_DONT by writting to the
      /proc/sys/net/ipv4/ip_no_pmtu_disc file.

Yes, that sounds good to me.  Thanks for doing this!
Acked-by: Neil Horman <redacted>

Ben, thanks for writing this, and Neil, thanks for reviewing it. I've
applied that change for man-pages-3.34.

Ben, I added one small piece ti the description of
/proc/sys/net/ipv4/ip_no_pmtu_disc. For completeness, I've reproduced
the entire text below. Perhaps you could take a quick scan, to make
sure that the changed text is consistent with the whole piece.

This looks good to me.  A minor nit in the chunk you added thoguh.  Calling the
MTU discovery value 'flags' suggests they can be or-together, or somehow
otherwise combined, which is non-sensical.  It may be better to use the term
values rather than flags.

Other than that
Acked-by: Neil Horman <redacted>

Thanks!
Neil

--
To unsubscribe from this list: send the line "unsubscribe linux-man" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html