From: David Brownell <david-b@pacbell.net>
To: Linux Kernel list <linux-kernel@vger.kernel.org>
Cc: Felipe Balbi <felipebalbi@users.sourceforge.net>,
Bill Gatliff <bgat@billgatliff.com>,
Haavard Skinnemoen <hskinnemoen@atmel.com>,
Andrew Victor <andrew@sanpeople.com>,
Tony Lindgren <tony@atomide.com>,
Jean Delvare <khali@linux-fr.org>,
"eric miao" <eric.y.miao@gmail.com>,
Kevin Hilman <khilman@mvista.com>,
Paul Mundt <lethal@linux-sh.org>,
Ben Dooks <ben@trinity.fluff.org>
Subject: Re: [patch/rfc 1/4] GPIO implementation framework
Date: Mon, 5 Nov 2007 13:05:13 -0800 [thread overview]
Message-ID: <200711051305.13980.david-b@pacbell.net> (raw)
In-Reply-To: <200710291851.23821.david-b@pacbell.net>
On Monday 29 October 2007, David Brownell wrote:
>
> Provides new implementation infrastructure that platforms may choose to use
> when implementing the GPIO programming interface. Platforms can update their
> GPIO support to use this. The downside is slower access to non-inlined GPIOs;
> rarely a problem except when bitbanging some protocol.
I was asked just what that overhead *is* ... and it surprised me.
A summary of the results is appended to this note.
Fortuntely it turns out those problems all go away if the gpiolib
code uses a *raw* spinlock to guard its table lookups. With a raw
spinlock, any performance impact of gpiolib seems to be well under
a microsecond in this bitbang context (and not objectionable).
Preempt became free; enabling debug options had only a minor cost.
That's as it should be, since the only substantive changes were to
grab and release a lock, do one table lookup a bit differently, and
add one indirection function call ... changes which should not have
any visible performance impact on per-bit codepaths, and one might
expect to cost on the order of one dozen instructions.
So the next version of this code will include a few minor bugfixes,
and will also use a raw spinlock to protect that table. A raw lock
seems appropriate there in any case, since non-sleeping GPIOs should
be accessible from hardirq contexts even on RT kernels.
If anyone has any strong arguments against using a raw spinlock
to protect that table, it'd be nice to know them sooner rather
than later.
- Dave
SUMMARY:
Using the i2c-gpio driver on a preempt kernel with all the usual
kernel debug options enabled, the per-bit times (*) went up in a
bad way: from about 6.4 usec/bit (original GPIO code on this board)
up to about 11.2 usec/bit (just switching to gpiolib), which is
well into "objectionable overhead" territory for bit access.
Just enabling preempt shot the time up to 7.4 usec/bit ... which is
also objectionable (it's all-the-time overhead that is clearly
needless), but much less so.
Converting the table lock to be a raw spinlock essentially removed
all non-debug overheads. It took enabling all those debug options
plus internal gpiolib debugging overhead to get those times up to
the 7.4 usec/bit that previously applied even with just preempt.
(*) Those times being eyeballed medians; I didn't make time to find
a way to export a few thousand measurements from the tool and
do the math. The typical range was +/- one usec.
The numbers include udelay() calls, so the relevant point is
the time *delta* attributable only to increased gpiolib costs,
not the base time (with udelays). The delta probably reflects
on the order of four GPIO calls: set two different bits, clear
one of them, and read it to make sure it cleared.
> The upside is:
>
> * Providing two features which were "want to have (but OK to defer)" when
> GPIO interfaces were first discussed in November 2006:
>
> - A "struct gpio_chip" to plug in GPIOs that aren't directly supported
> by SOC platforms, but come from FPGAs or other multifunction devices
> (like UCB-1x00 GPIOs).
>
> - Full support for message-based GPIO expanders, needing a gpio_chip
> hookup; previous support for this part of the programming interface
> was just stubs. (One example: the widely used pcf8574 I2C chips,
> with 8 GPIOs each.)
>
> * Including a non-stub implementation of the gpio_{request,free}() calls,
> which makes those calls much more useful. The diagnostic labels are
> also recorded given DEBUG_FS, so /sys/kernel/debug/gpio can show a
> snapshot of all GPIOs known to this infrastructure.
>
> The driver programming interfaces introduced in 2.6.21 do not change at all;
> this new infrastructure is entirely below the covers.
next prev parent reply other threads:[~2007-11-05 21:05 UTC|newest]
Thread overview: 60+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <200710291809.29936.david-b@pacbell.net>
2007-10-30 1:51 ` [patch/rfc 1/4] GPIO implementation framework David Brownell
2007-11-05 21:05 ` David Brownell [this message]
2007-11-13 2:28 ` eric miao
2007-11-13 19:06 ` David Brownell
2007-11-14 0:57 ` eric miao
2007-11-14 1:00 ` eric miao
2007-11-14 1:02 ` eric miao
2007-11-14 1:03 ` eric miao
2007-11-14 1:04 ` eric miao
2007-11-14 1:04 ` eric miao
2007-11-14 4:36 ` David Brownell
2007-11-14 6:51 ` eric miao
2007-11-14 7:19 ` David Brownell
2007-11-14 7:36 ` eric miao
2007-11-17 10:38 ` Jean Delvare
2007-11-17 17:36 ` David Brownell
2007-11-20 15:20 ` Jean Delvare
2007-11-14 4:18 ` David Brownell
2007-11-14 6:46 ` eric miao
2007-11-14 3:28 ` David Brownell
2007-11-14 3:25 ` David Brownell
2007-11-14 3:53 ` David Brownell
2007-11-14 6:37 ` eric miao
2007-11-14 3:30 ` David Brownell
2007-11-14 6:40 ` eric miao
2007-11-14 7:08 ` David Brownell
2007-11-27 1:46 ` David Brownell
2007-11-27 10:58 ` eric miao
2007-11-27 17:26 ` David Brownell
2007-11-27 19:03 ` David Brownell
2007-11-27 19:29 ` David Brownell
2007-11-28 5:11 ` eric miao
2007-11-28 3:15 ` [patch/rfc 2.6.24-rc3-mm] gpiolib grows a gpio_desc David Brownell
2007-11-28 9:10 ` eric miao
2007-11-28 9:53 ` David Brownell
2007-10-30 1:51 ` [patch/rfc 2/4] pcf875x I2C GPIO expander driver David Brownell
2007-11-30 12:32 ` Jean Delvare
2007-11-30 13:04 ` Bill Gatliff
2007-11-30 13:36 ` Jean Delvare
2007-11-30 14:09 ` Bill Gatliff
2007-11-30 18:40 ` David Brownell
2007-11-30 20:13 ` Jean Delvare
2007-11-30 20:59 ` David Brownell
2008-04-04 2:06 ` Trent Piepho
2008-04-04 2:45 ` Ben Nizette
2008-04-04 3:33 ` Trent Piepho
2008-04-04 4:57 ` Ben Nizette
2008-04-05 4:05 ` userspace GPIO access (WAS: [patch/rfc 2/4] pcf875x ...) David Brownell
2008-04-07 17:56 ` Trent Piepho
2008-04-04 8:09 ` [patch/rfc 2/4] pcf875x I2C GPIO expander driver Jean Delvare
2008-04-04 19:07 ` Trent Piepho
2008-04-04 19:36 ` Jean Delvare
2008-04-04 20:18 ` Trent Piepho
2008-04-05 2:51 ` David Brownell
2008-04-05 2:53 ` David Brownell
2007-12-06 3:03 ` [patch/rfc 2/4] pcf857x " David Brownell
2007-12-06 23:17 ` Jean Delvare
2007-12-07 4:02 ` David Brownell
2007-10-30 1:53 ` [patch/rfc 3/4] DaVinci platform uses new GPIOLIB David Brownell
2007-10-30 1:54 ` [patch/rfc 4/4] DaVinci EVM uses pcf857x GPIO driver David Brownell
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=200711051305.13980.david-b@pacbell.net \
--to=david-b@pacbell.net \
--cc=andrew@sanpeople.com \
--cc=ben@trinity.fluff.org \
--cc=bgat@billgatliff.com \
--cc=eric.y.miao@gmail.com \
--cc=felipebalbi@users.sourceforge.net \
--cc=hskinnemoen@atmel.com \
--cc=khali@linux-fr.org \
--cc=khilman@mvista.com \
--cc=lethal@linux-sh.org \
--cc=linux-kernel@vger.kernel.org \
--cc=tony@atomide.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).