On 7/30/06, Shem Multinymous <[email protected]> wrote:
Alternatively, we can add an extra parameter to that new
syscall/ioctl: "block until the time is T+N and you have a refresh
that was received from the hardware at time T+M, whichever is later"
(where T is the current time and N>M).
That's semantically equivalent to an msleep(M) followed by the
original delayed_update(N-M), but will save one timer interrupt per
iteration in some cases (e.g., an event-based hardware data source).
No, wait, the explicit two-parameter version is far superior.
First, it makes it easier to poll multiple attributes. How'd you do
that with the one-parameter version, if you need to sleep separately
for each attribute?
Second, it gives the driver (or its sysfs support infrastructure) a
better picture of the future, which it can use to optimize the
hardware query scheduling. No, I don't think anyone will implement
*that* soon, but it's good to have an API that lets us add it later
without affecting userspace.
Is anyone interested in adding this "delayed updates" support to
sysfs, for the benefit of hwmon, tp_smapi and probably numerous other
drivers?
I have some ideas for how to do it so that that per-driver support is
really easy (or even free), but I don't know sysfs well enough to
implement it and don't have the resources.
Shem
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[Index of Archives]
[Kernel Newbies]
[Netfilter]
[Bugtraq]
[Photo]
[Stuff]
[Gimp]
[Yosemite News]
[MIPS Linux]
[ARM Linux]
[Linux Security]
[Linux RAID]
[Video 4 Linux]
[Linux for the blind]
[Linux Resources]
