pipe: Don't oops when pipe filesystem isn't mounted
In-Reply-To: <[email protected]>
Quoted 6 lines · Andrew Morton
Quoted 2 lines
Prevent oops when an app tries to create a pipe while pipefs is not mounted.
That's pretty lame. It means that pipes just won't work, so people who are using pipes in their initramfs setups will just get mysterious failures running userspace on a crippled kernel.
I know, I just wanted to keep the issue alive. :)
Quoted 6 lines
I think the bug really is the running of
populate_rootfs()before running the initcalls, in init/main.c:init(). It's just more sensible to start running userspace after the initcalls have been run. Statically-linked drivers which want to load firmware files will lose. To fix that we'd need a new callback. It could be with a new linker section or perhaps simply a notifier chain.
Why not create a new initcall category for things that must run before early userspace?
Signature · 1 line
MBTI: IXTP
Quoted 2 lines · Chuck Ebbert
Why not create a new initcall category for things that must run before early userspace?
Why do you want to continue with papering over the root cause? Pick some janitor, let him write something that implements something like make style dependencies for initcalls.
Then you can get rid of all that foo_initcall stuff.
It surely needs work to get all that done.
Quoted 10 lines · Olaf Hering
Quoted 2 lines · Chuck Ebbert
Why not create a new initcall category for things that must run before early userspace?
Why do you want to continue with papering over the root cause? Pick some janitor, let him write something that implements something like make style dependencies for initcalls.
Then you can get rid of all thatfoo_initcallstuff.
It surely needs work to get all that done.
I would argue that _this_ is papering over the root cause. Namely, too complex ordering requirements. Growing a technics for allowing them to fester is an odd solution...