In D28154#680209, @kib wrote:

Do we have an agreement there?

I think that switching the defaults, if ever, should be done by separate change indeed.

In D28154#677706, @trasz wrote:

I've tested this patch for both 64, and 32-bit Linux binaries, for both suid and sgid bit, and it works just fine.

One nit, though: I'd strongly prefer the sysctl to be enabled by default, ie to permit suid/sgid by default. Otherwise it will break some pretty common use cases.

Good to me

also remove links, add pfctlinput2 to obsolete files.

Left both .El as markj noted.

what if the user has no write access on cwd? or what if the socket name is '/xx/xxx/xxx'?
and who is responsible to unlink this socket after close?

In D29539#662228, @imp wrote:

I've never seen this lead to deadlock, but I'm also only ever using nvme with direct dispatch...

Fully support this change and motivation, propose to add a short description to the linux(4) manual about sysctl

Sure, done

In D29250#655108, @trasz wrote:

Also: you’re alive! \o/

In D29250#655107, @trasz wrote:

In D29250#655002, @dchagin wrote:

this can be done after PI futex

Funny you mention that, I’ve just hacked together a simple implementation of LOCK_PI and UNLOCK_PI. It doesn’t handle priorities in any way, but seems good enough to make some things work. I’ll try to clean it up and put up, unless you have something better?

this can be done after PI futex

Disagree, the KTR buffer can be examined in DDB what is really important.

fine, 10x Tijl

or implement native analog for kqueue/kevent? if it makes sense, I'm not sure)

limit bufsize by 1Mb

In D20575#476278, @markj wrote:

Seems like the limit could be removed by using an sbuf drain function?

Add process lock assert

In D23341#511764, @kib wrote:

Why do you need D23340 ? Why the current method is not sufficient ?

hi, could u please test D23339, D23340,D23341?

nice