Re: Problem with slip connections after kernel upgrade on squeeze
Well, the old slip-open behavior was not correspond to the TTY logic.
So, is it possible to fix VPND instead of reverting the patch? Greg,
what do you think about it?
2011/7/22 Ben Hutchings <ben@decadent.org.uk>:
> On Fri, 2011-07-22 at 08:30 -0300, SUPORTE - Kernel Informática wrote:
>> I had a debian squeeze with 3 slips connection, sl0, sl1 and sl2 on
>> working VPND solution. Its using the kernel version
>> linux-image-2.6.32-5-amd64_2.6.32-31_amd64 so I upgrade it to
>> linux-image-2.6.32-5-amd64_2.6.32-35_amd64 and the slip interface dont
>> fork to sl1 any more! example: When I start the first connection its
>> works well with sl0 interface so when i start second and third
>> connections its bind interface sl0 too chasing the first one! they dont
>> fork to sl1 and sl2 any more!
>> Maybe this is a BUG on this patch?
>>
>> When I found this problem I stop to upgrade my others squeeze solution
>> using VPND with slip interface!
>
> vpnd (source: <http://vpnd.linuxsys.com/archive/vpnd-1.1.4.tar.gz>)
> expects to get the SLIP unit number when setting SLIP line discipline:
>
> /* set line discipline to SLIP, react to errors */
>
> #ifdef LINUX
> i=N_SLIP;
> if((anchor->proxy=ioctl(anchor->tty,TIOCSETD,&i))<0)
> JUMP("ioctl(TIOCSETD)",err3);
> #elif defined(FreeBSD) || defined(OSNetBSD) || defined(OSOpenBSD)
> ...
> /* create interface name */
>
> memset(&ifr,0,sizeof(ifr));
> sprintf(ifr.ifr_name,"sl%u",anchor->proxy);
>
> But in commit 057bef938896e6266ae24ec4266d24792d27c29a the behaviour of
> slip_open() and hence ioctl TIOCSETD was changed so that the unit number
> is no longer returned.
>
> It looks like applications are supposed to use the ioctl SIOCGIFNAME to
> find the net device name. But clearly the driver was previously
> returning the unit number and applications did depend on this. So I
> think this change has to be reverted and the TTY layer will have to
> accept positive return values from ldisc open().
>
> Ben.
>
> --
> Ben Hutchings
> Absolutum obsoletum. (If it works, it's out of date.) - Stafford Beer
>
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.10 (GNU/Linux)
>
> iQIVAwUBTinASee/yOyVhhEJAQqX8hAAoouFWS/Rm+mOXw8FPVGKtGQauUtILS37
> ZC1KCZizfITrTjlfXo+3f8hnAUzv/FlRLW6YBNFIaMhGgEnCV74lqYIyzyJ+K8iR
> ZXixzIKFhwn51dSgGVjUSJOUH5f66dHTn7IhlOXZ6qJWgNR7oNi4NWyX9F06M0gp
> fGPMzQnkdSWIk1+IwEo6RkOOq+vJR9Xgvxbyu7jiXD+KwydCBQChDcwDr8H5BFI9
> QBWmljc9+AoalrRBNd1ekyeIIbZFUOPDyFyBdGHRDf+myCVNhlg3Z+2Ea6fKDO4u
> RheGsAzicsBvWVkoZJUwXvHh+ysf7K3q8kPJn0YNAvEjlUPreOxXJxl97a/EPgtH
> D8uTkaI7YaBtIy14EbGDiF4JFwaUiIJQRdQhQUvutuCypdh8HzAgmMEDeb/ikmiz
> 66Hg6IMaRZMDzWndEIRRex3Bs1MgnRwktkYYZV9Lnb6xkf4OSK03lzF8ZP6tY2Ay
> Tf29vK2W0ji9hdbefUr0y4G2N2eo0ydYR3TUMvWE0xpjA3b37v6ypcVT3mCGAkwx
> lZb3CgX5d29EMRFpJYXlCtS2JzfHZOMoXXZU1+EMWbeRxHR//sSGGech0FHwNslK
> eJtMoMgn/pAStpaq8ZMnmU/+A40UIU2QoVgMpdpBmx7g1hvgqlA7rLDvOTtAdFpp
> cWqn78QK0Jc=
> =zCNx
> -----END PGP SIGNATURE-----
>
>
Reply to: