polling -- device polling support
options DEVICE_POLLING
options HZ=1000
``Device polling'' (polling for brevity) refers to a technique to handle
devices that does not rely on the latter to generate interrupts when they
need attention, but rather lets the CPU poll devices to service their
needs. This might seem inefficient and counterintuitive, but when done
properly, polling gives more control to the operating system on when and
how to handle devices, with a number of advantages in terms of system
responsivity and performance.
In particular, polling reduces the overhead for context switches which is
incurred when servicing interrupts, and gives more control on the scheduling
of the CPU between various tasks (user processes, software interrupts,
device handling) which ultimately reduces the chances of livelock
in the system.
PRINCIPLES OF OPERATION [Toc] [Back] In the normal, interrupt-based mode, devices generate an interrupt whenever
they need attention. This in turn causes a context switch and the
execution of an interrupt handler which performs whatever processing is
needed by the device. The duration of the interrupt handler is potentially
unbounded unless the device driver has been programmed with realtime
concerns in mind (which is generally not the case for FreeBSD drivers).
Furthermore, under heavy traffic, the system might be persistently
processing interrupts without being able to complete other work, either
in the kernel or in userland.
Polling disables interrupts by polling devices at appropriate times,
i.e., on clock interrupts, system calls and within the idle loop. This
way, the context switch overhead is removed. Furthermore, the operating
system can control accurately how much work to spend in handling device
events, and thus prevent livelock by reserving some amount of CPU to
other tasks.
Polling is enabled with a sysctl(8) variable kern.polling.enable whereas
the percentage of CPU cycles reserved to userland processes is controlled
by the sysctl(8) variable kern.polling.user_frac whose range is 0 to 100
(50 is the default value).
When polling is enabled, and provided that there is work to do, up to
kern.polling.user_frac percent of the CPU cycles is reserved to userland
tasks, the remaining fraction being available for device processing.
Enabling polling also changes the way network software interrupts are
scheduled, so there is never the risk of livelock because packets are not
processed to completion.
There are other variables which control or monitor the behaviour of
devices operating in polling mode, but they are unlikely to require modifications,
and are documented in the source file sys/kern/kern_poll.c.
Polling requires explicit modifications to the device drivers. As of
this writing, the dc(4), em(4), fxp(4), rl(4), and sis(4) devices are
supported, with other in the works. The modifications are rather
straightforward, consisting in the extraction of the inner part of the
interrupt service routine and writing a callback function, *_poll(),
which is invoked to probe the device for events and process them. See
the conditionally compiled sections of the devices mentioned above for
more details.
As in the worst case devices are only polled on clock interrupts, in
order to reduce the latency in processing packets, it is advisable to
increase the frequency of the clock to at least 1000 HZ.
Device polling was introduced in February 2002 by Luigi Rizzo
<[email protected]>.
FreeBSD 5.2.1 February 15, 2002 FreeBSD 5.2.1 [ Back ] |
