Hi On 2015-03-17, Stefan Lippers-Hollmann wrote: > On 2015-03-17, Michael Biebl wrote: > > Am 17.03.2015 um 03:51 schrieb Stefan Lippers-Hollmann: [...] > There are two pending changes beyond this, but imho neither meet the unblock > criterias at this stage. [...] > - fixing a segfault when using hostapd for a non-wireless (ethernet) > interface (this is not exploitable, hostapd will just hard-fail to start > using a config file configuring driver=wired). Technically I'd consider > this to be RC, but given how rare using IEEE 8021X on wired ethernet > networks is in practice - especially on Debian compatible hardware > (rather than using commercial/ managed ethernet switches) - I don't think > this would warrant a freeze exemption (this bug has not been reported to > the Debian BTS). The bugfix[1] itself should be self-contained enough and > probably acceptable to the release team (not affecting the feature set > chosen for wpasupplicant-udeb). [...] > [1] hostapd: Verify VHT 160/80+80 MHz driver support > > http://w1.fi/cgit/hostap/commit/?id=7f0303d5b0bb425f3e7318a7016b55ba9e67f9de [...]
Small correction for a mistyped patch, the actual patch for the hostapd
with driver=wired regression would be:
[1] Fix hostapd operation without hw_mode driver data
http://w1.fi/cgit/hostap/commit/?id=e9b783d58c23a7bb50b2f25bce7157f1f3b5d58b
The afforementioned/ mispasted "hostapd: Verify VHT 160/80+80 MHz driver
support" introduced the bug by trying to enforce channel bandwidth
conditions for non-wireless interfaces as well (which can only fail on
a wired interface).
Regards
Stefan Lippers-Hollmann
pgpRG2wvIFY1u.pgp
Description: Digitale Signatur von OpenPGP
