Re: Mixing cached and not_cached access to the same file
Re: Mixing cached and not_cached access to the same file
- Subject: Re: Mixing cached and not_cached access to the same file
- From: Jim Luther <email@hidden>
- Date: Tue, 16 Oct 2012 09:54:31 -0700
I didn't say the way Carbon did it was the most efficient... that's just how it works. :-)
On Oct 16, 2012, at 9:31 AM, James Bucanek <email@hidden> wrote:
> Jim Luther <mailto:email@hidden> wrote (Monday, October 15, 2012 6:52 PM -0700):
>
>> The Carbon behavior is similar to your first model. Each read or write request
>> takes a lock on the struct that tracks an open file, toggles the current
>> caching behavior (which is tracked in the same struct) to match the request,
>> performs the read or write, and then unlocks the struct.
>
> Thanks, Jim. I wasn't sure if there was a lot of overhead associated with switching caching modes and really didn't want to spend days rewriting my application two different ways just to find out.
>
> The semaphore method works better for me anyway. It's just simpler managing a single file descriptor.
>
> James
> --
> James Bucanek
>
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Filesystem-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden