Re: cdevsw close never called
Re: cdevsw close never called
- Subject: Re: cdevsw close never called
- From: mm w <email@hidden>
- Date: Sat, 7 Mar 2009 01:49:37 -0800
Hi Andrew, sorry for this I hadn't sleep all the previous night
I missed your point, for the record, I did and I am doing exactly what
you are doing anyway as I can see, maybe there are differences, I am
sure, but I never run unto your issue, I guess I was a bit lazy
because I used/reproduced for 99.9%
the ttys well-tested time API model and code, and my clients act like
your app or a term window, whatever it's opening and wiring to an
incremental device file using exactly the same semantic except I am
using KSL as prefix, stands for Kernel Scripting Language, yes the
device mechanism is the smallest part; the rest of the kext is hosting
the interpreter this is a lot inspired by APIC/ASL,
indeed why not offering the same concept/layer between the userspace
and the kernel as well as between the hardware and software, and yes
it's kicking the ass (sorry for the expression) to a lot of
"matusalem" Nix API (for instance, sysctl) and might have really
interesting further applications , and sure the concept can be bring
on any platform.
do hope you solved your issue by checking the vnode
Best Regards
On Fri, Mar 6, 2009 at 2:33 PM, Andrew Gallatin <email@hidden> wrote:
> Andrew Gallatin wrote:
>
>> The best I've been able to achieve so far is to
>> periodically reap state which look dead. But that
>> is just gross.
>
> BTW, the reaping I'm doing looks at the vnode associated
> with the device node. Periodically, the driver scans
> all open devices, and checks for any with a vnode whose
> vcount() is 0. If this is happens, then I know the race
> happened, and I can free up the resources associated with
> that device and allow it to be opened again. Does that
> seem reasonable?
>
> Drew
>
> _______________________________________________
> Do not post admin requests to the list. They will be ignored.
> Darwin-kernel mailing list (email@hidden)
> Help/Unsubscribe/Update your Subscription:
>
> This email sent to email@hidden
>
--
-mmw
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Darwin-kernel mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden