eyJhb changed the topic of #nixos-on-your-router to: NixOS on your Router || https://logs.nix.samueldr.com/nixos-on-your-router
gchristensen has joined #nixos-on-your-router
{^_^} has joined #nixos-on-your-router
gchristensen has quit [Quit: WeeChat 2.9]
{^_^} has quit [Remote host closed the connection]
gchristensen has joined #nixos-on-your-router
{^_^} has joined #nixos-on-your-router
v0|d has quit [Ping timeout: 272 seconds]
teto has quit [Quit: WeeChat 3.0]
<NinjaTrappeur> ttps://store.alternativebit.fr/84e46a167e0dc55b90e70dacdf619007afcec5a92db6d61fdf99b3610a73d20e-UjwWiEY.png
<NinjaTrappeur> Famous last words :P
<NinjaTrappeur> hexa-: everytime I dig into that codebase, it ends up being a pretty unpleasant experience. It always makes me uneasy to know so much of the ecosystem relies on that piece of software.
<NinjaTrappeur> ^ does anybody here have an answer to that?
<hexa-> NinjaTrappeur: anything suspicious about that particular line?
<NinjaTrappeur> I was confused by the not-NULLed initializer.
<NinjaTrappeur> It was NULLed before this patch.
<NinjaTrappeur> After investigating this, it does not seem that we use this NULL anywhere. We loop through the capabilities using num_item.
<NinjaTrappeur> The part that still confuses me is the fact the capabilities array get init. to NULL here https://www.spinics.net/lists/hostap/msg06611.html while it does not there: https://w1.fi/cgit/hostap/patch/?id=7800725afb27397f7d6033d4969e2aeb61af4737
<hexa-> ah lol
<NinjaTrappeur> I was confused: both patches are claiming to come from the same email. andi- suggested the maintainer might have altered the patch.
<NinjaTrappeur> As long as we do not use NULL in any way to detect the boundary of the array, we'lle be ok.
<hexa-> yep, quite plausible
<hexa-> as long as we apply upstream patches :)
<andi-> that being all said I think the code is actually fine the way it is written. It is just a maint. burden I wouldn't want to have when mainting it..
<NinjaTrappeur> Paranoid me added a note to my local git checkout to double check we still do not rely on NULL for boundary detection before next release.
<NinjaTrappeur> yeah, sure. I was mostly worried about a vulnerability injected upstream here.
<andi-> as long as num_items fits into credentials we are fine as the function reading from the array actually doesn't check for NULL termination but takes the count
<andi-> It is a bit unfortunate that the array is the first thing on the stack as that means you could read/write into the other local variables but those aren't used after/while the array is written/send.
flokli has joined #nixos-on-your-router
<{^_^}> #111531 (by peterhoeg, 2 hours ago, open): firewall: support multicast
<hexa-> can we have more opinions here? I think a blanked multicast/broadcast accept feels a bit odd.
<hexa-> s/blanked/blanket/
thefloweringash has quit [Ping timeout: 244 seconds]
thefloweringash has joined #nixos-on-your-router
higebu_ has joined #nixos-on-your-router
higebu has quit [*.net *.split]
dtz has quit [*.net *.split]
higebu_ is now known as higebu
dtz has joined #nixos-on-your-router