"Access and friends provide a good first guess if a future open might succeed." Might be good to insert at the start. Or "access, eaccess and faccessat provide..." and change the list at the start of the next sentence to "These system calls".

Wrong tree

Looks great

Even though we relocate this binary, this value is invariant. And less than 32 bits, though that shouldn't matter here

In D45215#1031630, @imp wrote:

What does nvmecontrol identify ndaX say here? What's the optimal I/O boundary? And maybe we should just have a knob to disable trying to use it.

nvmecontrol identify nda4 | grep Opt
Optimal I/O Boundary:        256 blocks

What does nvmecontrol identify ndaX say here? What's the optimal I/O boundary? And maybe we should just have a knob to disable trying to use it.

Respectfully, this code is non-standards-conforming, since the standard doesn't allow multiple variable sized elements in a structure.

It's all one sentence one line with a couple minor tweaks that are easy to review.

Yes

Then a quick line in yhe commit messafe saying its a fix

I don't know.. has soft updates cooked enough ? :)

I'd be more inclined to land the other review, but until it does, this isn't bad.

S/lube/line/. Stupid autocowrecked

Other arch? At the very least a lube in the commit about why not relevant there

Yea. I didn't like it either.
I could move the timeouts into qpair. Then there'd be no more computing the values... I'd have to plumb the sysctl too... it would be cleaner. I could also put the soft timeout in there too. Then it wouldn't be so arbitrary.

A 32bit freebsd cat reading /compat/linux/mumble

Generally i like it. It does seem to mix memory leak / style fixed in with the new functionality.