how slow is ioctl?

From: DazMe (z333r_at_yahoo.com)
Date: 09/21/05 

Date: 21 Sep 2005 07:41:24 -0700

Hi all,

I'm writing some kernel code and I have no option
AFAIK but to have apps communicate with my kernel
code via the ioctl interface. However I was curious
if anyone can tell me how much time it takes
to invoke an ioctl, because I'm pretty sure I'm
seeing a bit of a delay caused by it. The actual
kernel code is pretty fast. Hmm... has anyone
put a "no op" ioctl in place to permit determining
ioctl latency empirically?

Thanks.

Relevant Pages

  • Re: upcalls from kernel code to user space daemons
    ... >> and fails reconnecting the socket and the kernel code wants to see if ... > One way to do it (or is this what you meant by captive ioctl?) ... fill in data and return to userspace ... send the line "unsubscribe linux-kernel" in ...
    (Linux-Kernel)
  • Re: inode* from struct file*?
    ... I need to call an ioctl from inside the kernel, ... you want, and call whatever kernel code it calls directly, without the ... One of the fields of struct file leads to an inode. ...
    (comp.os.linux.development.system)
  • Re: Non blocking ioctl CDROMEJECT
    ... |> thread that is blocked in ioctl(). ... It will die when it comes out of the ... It's in kernel code, so it might ... | be taking signals even in kernel, ...
    (comp.os.linux.development.system)
  • Re: Non blocking ioctl CDROMEJECT
    ... > thread that is blocked in ioctl(). ... It's in kernel code, so it might ... be taking signals even in kernel, and die at once. ...
    (comp.os.linux.development.system)
  • Re: [Xen-devel] Re: [RFC] [PATCH] sysfs support for Xen attributes
    ... >> it has primarily an ioctl() interface. ... anything with an ioctl interface is probably a good cantidate ... > for a writable sysfs file. ... > in sysfs with a discrete, ...
    (Linux-Kernel)