The patch below contains small changes to the USB Gadget Kconfig.
The main change is that multiple modular peripheral controllers are no
longer allowed (currently only one is there, but this may change).
cu
Adrian
--- linux-2.6.0-test9-mm4/drivers/usb/gadget/Kconfig.old 2003-11-23 17:17:51.000000000 +0100
+++ linux-2.6.0-test9-mm4/drivers/usb/gadget/Kconfig 2003-11-23 17:52:24.000000000 +0100
@@ -3,10 +3,9 @@
# (a) a peripheral controller, and
# (b) the gadget driver using it.
#
-# for 2.5 kbuild, drivers/usb/gadget/Kconfig
# source this at the end of drivers/usb/Kconfig
#
-menuconfig USB_GADGET
+config USB_GADGET
tristate "Support for USB Gadgets"
depends on EXPERIMENTAL
help
@@ -32,16 +31,24 @@
If in doubt, say "N" and don't enable these drivers; most people
don't have this kind of hardware (except maybe inside Linux PDAs).
+menu "USB Gadget options"
+ depends on USB_GADGET!=n
+
#
# USB Peripheral Controller Support
#
+
+#
+# only one gadget controller driver, linked statically or as a module
+# (depending on whether USB_GADGET is y or m)
+#
+
choice
prompt "USB Peripheral Controller Support"
- depends on USB_GADGET
-config USB_NET2280
- tristate "NetChip 2280 USB Peripheral Controller"
- depends on PCI && USB_GADGET
+config USB_TMP_NET2280
+ bool "NetChip 2280 USB Peripheral Controller"
+ depends on PCI
help
NetChip 2280 is a PCI based USB peripheral controller which
supports both full and high speed USB 2.0 data transfers.
@@ -56,19 +63,28 @@
endchoice
+config USB_NET2280
+ tristate
+ default USB_TMP_NET2280 && USB_GADGET
+
+
#
# USB Gadget Drivers
#
+
+#
+# at most one gadget driver statically linked,
+# OR any number of gadget drivers, linked as modules
+#
+
choice
prompt "USB Gadget Drivers"
depends on USB_GADGET
default USB_ETH
-# FIXME want a cleaner dependency/config approach for drivers.
-
config USB_ZERO
tristate "Gadget Zero (DEVELOPMENT)"
- depends on USB_GADGET && (USB_DUMMY_HCD || USB_NET2280 || USB_PXA2XX || USB_SA1100)
+ depends on USB_DUMMY_HCD || USB_NET2280 || USB_PXA2XX || USB_SA1100
help
Gadget Zero is a two-configuration device. It either sinks and
sources bulk data; or it loops back a configurable number of
@@ -110,7 +126,7 @@
config USB_ETH
tristate "Ethernet Gadget"
- depends on USB_GADGET && NET && (USB_DUMMY_HCD || USB_NET2280 || USB_PXA2XX || USB_SA1100)
+ depends on NET && (USB_DUMMY_HCD || USB_NET2280 || USB_PXA2XX || USB_SA1100)
help
This driver implements Ethernet style communication, in either
of two ways:
@@ -155,7 +171,7 @@
config USB_GADGETFS
tristate "Gadget Filesystem (EXPERIMENTAL)"
- depends on USB_GADGET && (USB_DUMMY_HCD || USB_NET2280 || USB_PXA2XX) && EXPERIMENTAL
+ depends on (USB_DUMMY_HCD || USB_NET2280 || USB_PXA2XX) && EXPERIMENTAL
help
This driver provides a filesystem based API that lets user mode
programs implement a single-configuration USB device, including
@@ -179,4 +195,4 @@
endchoice
-# endmenuconfig
+endmenu
Hi David,
first of all sorry for my late answer.
I'm currently wading through a bunch of mails in my INBOX that should
have been answered some time ago. :-(
On Mon, Dec 29, 2003 at 07:54:29PM -0800, David Brownell wrote:
> Adrian Bunk wrote:
> >The patch below contains small changes to the USB Gadget Kconfig.
> >
> >The main change is that multiple modular peripheral controllers are no
> >longer allowed (currently only one is there, but this may change).
>
> How about using this approach instead? It simplifies the kconfig
> for the gadget drivers by providing a boolean "which hardware"
> symbol, so gadget drivers don't need to make their own. The symbol
> that's synthetic is the one needed only by the Makefile.
>...
Your suggestion looks fine.
> - Dave
>...
cu
Adrian
--
"Is there not promise of rain?" Ling Tan asked suddenly out
of the darkness. There had been need of rain for many days.
"Only a promise," Lao Er said.
Pearl S. Buck - Dragon Seed
Hi,
On Mon, 29 Dec 2003, David Brownell wrote:
> How about using this approach instead? It simplifies the kconfig
> for the gadget drivers by providing a boolean "which hardware"
> symbol, so gadget drivers don't need to make their own. The symbol
> that's synthetic is the one needed only by the Makefile.
There are some strange things in there.
choice values can also be tristate symbols, so you wouldn't need the
separate defines, unless you really always want to compile only a single
controller (even as module).
The "default m if USB_GADGET = m" looks weird, if I understand them
correctly this should just be "depends on USB_GADGET", e.g.
config USB_NET2280
tristate
depends on USB_GADGET
default USB_GADGET_NET2280
this would also fix the menu structure and the drivers menu would appear
below the gadget option.
I'm also not sure about USB_PXA2XX_SMALL, as it also can be written as:
config USB_PXA2XX_SMALL
depends on USB_PXA2XX = y
default USB_ZERO = y || USB_ETH = y || USB_G_SERIAL
is this really intended?
The dependency "USB_DUMMY_HCD || USB_NET2280 || USB_PXA2XX || USB_SA1100
|| USB_GOKU" can be basically reduced to "USB_GADGET".
> Roman, this seems to trigger some kind of xconfig/menuconfig bug,
> since I can go down the list of hardware options (net2280, goku,
> dummy -- three, not the single one Adrian was working with) and
> each deselects the previous selection ... but then it's impossible
> to turn off the dummy, and select real hardware.
I can't reproduce this, it works fine here.
bye, Roman
#
# USB Gadget support on a system involves
# (a) a peripheral controller, and
# (b) the gadget driver using it.
#
menu "USB Gadget Support"
config USB_GADGET
tristate "Support for USB Gadgets"
help
USB is a master/slave protocol, organized with one master
host (such as a PC) controlling up to 127 peripheral devices.
The USB hardware is asymmetric, which makes it easier to set up:
you can't connect two "to-the-host" connectors to each other.
Linux can run in the host, or in the peripheral. In both cases
you need a low level bus controller driver, and some software
talking to it. Peripheral controllers are often discrete silicon,
or are integrated with the CPU in a microcontroller. The more
familiar host side controllers have names like like "EHCI", "OHCI",
or "UHCI", and are usually integrated into southbridges on PC
motherboards.
Enable this configuration option if you want to run Linux inside
a USB peripheral device. Configure one hardware driver for your
peripheral/device side bus controller, and a "gadget driver" for
your peripheral protocol. (If you use modular gadget drivers,
you may configure more than one.)
If in doubt, say "N" and don't enable these drivers; most people
don't have this kind of hardware (except maybe inside Linux PDAs).
#
# USB Peripheral Controller Support
#
choice
prompt "USB Peripheral Controller"
depends on USB_GADGET
help
A USB device uses a controller to talk to its host.
Systems should have only one such upstream link.
config USB_GADGET_NET2280
boolean "NetChip 2280"
depends on PCI
help
NetChip 2280 is a PCI based USB peripheral controller which
supports both full and high speed USB 2.0 data transfers.
It has six configurable endpoints, as well as endpoint zero
(for control transfers) and several endpoints with dedicated
functions.
Say "y" to link the driver statically, or "m" to build a
dynamically linked module called "net2280" and force all
gadget drivers to also be dynamically linked.
config USB_NET2280
tristate
depends on USB_GADGET_NET2280
default USB_GADGET
config USB_GADGET_PXA2XX
boolean "PXA 2xx or IXP 42x"
depends on ARCH_PXA || ARCH_IXP425
help
Intel's PXA 2xx series XScale ARM-5TE processors include
an integrated full speed USB 1.1 device controller. The
controller in the IXP 4xx series is register-compatible.
It has fifteen fixed-function endpoints, as well as endpoint
zero (for control transfers).
Say "y" to link the driver statically, or "m" to build a
dynamically linked module called "pxa2xx_udc" and force all
gadget drivers to also be dynamically linked.
config USB_PXA2XX
tristate
depends on USB_GADGET_PXA2XX
default USB_GADGET
# if there's only one gadget driver, using only two bulk endpoints,
# don't waste memory for the other endpoints
config USB_PXA2XX_SMALL
depends on USB_GADGET_PXA2XX
bool
default y if (USB_ZERO = y)
default y if (USB_ETH = y)
default y if (USB_G_SERIAL = y)
config USB_GADGET_GOKU
boolean "Toshiba TC86C001 'Goku-S'"
depends on PCI
help
The Toshiba TC86C001 is a PCI device which includes controllers
for full speed USB devices, IDE, I2C, SIO, plus a USB host (OHCI).
The device controller has three configurable (bulk or interrupt)
endpoints, plus endpoint zero (for control transfers).
Say "y" to link the driver statically, or "m" to build a
dynamically linked module called "goku_udc" and to force all
gadget drivers to also be dynamically linked.
config USB_GOKU
tristate
depends on USB_GADGET_GOKU
default USB_GADGET
# this could be built elsewhere (doesn't yet exist)
config USB_GADGET_SA1100
boolean "SA 1100"
depends on ARCH_SA1100
help
Intel's SA-1100 is an ARM-4 processor with an integrated
full speed USB 1.1 device controller.
It has two fixed-function endpoints, as well as endpoint
zero (for control transfers).
config USB_SA1100
tristate
depends on USB_GADGET_SA1100
default USB_GADGET
config USB_GADGET_DUMMY_HCD
boolean "Dummy HCD (DEVELOPMENT)"
depends on USB
help
This host controller driver emulates USB, looping all data transfer
requests back to a USB "gadget driver" in the same host. The host
side is the master; the gadget side is the slave. Gadget drivers
can be high, full, or low speed; and they have access to endpoints
like those from NET2280, PXA2xx, or SA1100 hardware.
This may help in some stages of creating a driver to embed in a
Linux device, since it lets you debug several parts of the gadget
driver without its hardware or drivers being involved.
Since such a gadget side driver needs to interoperate with a host
side Linux-USB device driver, this may help to debug both sides
of a USB protocol stack.
Say "y" to link the driver statically, or "m" to build a
dynamically linked module called "dummy_hcd" and force all
gadget drivers to also be dynamically linked.
config USB_DUMMY_HCD
tristate
depends on USB_GADGET_DUMMY_HCD
default USB_GADGET
endchoice
#
# USB Gadget Drivers
#
choice
tristate "USB Gadget Drivers"
depends on USB_GADGET
default USB_ETH
# FIXME Gadget drivers should now just #ifdef CONFIG_USB_GADGET_XXX;
# remove these other hardware flags
config USB_ZERO
tristate "Gadget Zero (DEVELOPMENT)"
depends on (USB_DUMMY_HCD || USB_NET2280 || USB_PXA2XX || USB_GOKU || USB_SA1100)
help
Gadget Zero is a two-configuration device. It either sinks and
sources bulk data; or it loops back a configurable number of
transfers. It also implements control requests, for "chapter 9"
conformance. The driver needs only two bulk-capable endpoints, so
it can work on top of most device-side usb controllers. It's
useful for testing, and is also a working example showing how
USB "gadget drivers" can be written.
Make this be the first driver you try using on top of any new
USB peripheral controller driver. Then you can use host-side
test software, like the "usbtest" driver, to put your hardware
and its driver through a basic set of functional tests.
Gadget Zero also works with the host-side "usb-skeleton" driver,
and with many kinds of host-side test software. You may need
to tweak product and vendor IDs before host software knows about
this device, and arrange to select an appropriate configuration.
Say "y" to link the driver statically, or "m" to build a
dynamically linked module called "g_zero".
config USB_ETH
tristate "Ethernet Gadget"
depends on NET && (USB_NET2280 || USB_PXA2XX || USB_GOKU || USB_SA1100)
help
This driver implements Ethernet style communication, in either
of two ways:
- The "Communication Device Class" (CDC) Ethernet Control Model.
That protocol is often avoided with pure Ethernet adapters, in
favor of simpler vendor-specific hardware, but is widely
supported by firmware for smart network devices.
- On hardware can't implement that protocol, a simpler approach
is used, placing fewer demands on USB.
Within the USB device, this gadget driver exposes a network device
"usbX", where X depends on what other networking devices you have.
Treat it like a two-node Ethernet link: host, and gadget.
The Linux-USB host-side "usbnet" driver interoperates with this
driver, so that deep I/O queues can be supported. On 2.4 kernels,
use "CDCEther" instead, if you're using the CDC option. That CDC
mode should also interoperate with standard CDC Ethernet class
drivers on other host operating systems.
Say "y" to link the driver statically, or "m" to build a
dynamically linked module called "g_ether".
config USB_GADGETFS
tristate "Gadget Filesystem (EXPERIMENTAL)"
depends on (USB_DUMMY_HCD || USB_NET2280 || USB_PXA2XX || USB_SA1100 || USB_GOKU) && EXPERIMENTAL
help
This driver provides a filesystem based API that lets user mode
programs implement a single-configuration USB device, including
endpoint I/O and control requests that don't relate to enumeration.
All endpoints, transfer speeds, and transfer types supported by
the hardware are available, through read() and write() calls.
Say "y" to link the driver statically, or "m" to build a
dynamically linked module called "gadgetfs".
config USB_FILE_STORAGE
tristate "File-backed Storage Gadget (DEVELOPMENT)"
depends on (USB_DUMMY_HCD || USB_NET2280 || USB_PXA2XX || USB_GOKU)
# we don't support the SA1100 because of its limitations
help
The File-backed Storage Gadget acts as a USB Mass Storage
disk drive. As its storage repository it can use a regular
file or a block device (in much the same way as the "loop"
device driver), specified as a module parameter.
Say "y" to link the driver statically, or "m" to build a
dynamically linked module called "g_file_storage".
config USB_FILE_STORAGE_TEST
bool "File-backed Storage Gadget test version"
depends on USB_FILE_STORAGE
default n
help
Say "y" to generate the larger testing version of the
File-backed Storage Gadget, useful for probing the
behavior of USB Mass Storage hosts. Not needed for
normal operation.
config USB_G_SERIAL
tristate "Serial Gadget"
depends on (USB_NET2280 || USB_PXA2XX || USB_GOKU || USB_SA1100)
help
The Serial Gadget talks to the Linux-USB generic serial driver.
Say "y" to link the driver statically, or "m" to build a
dynamically linked module called "g_serial".
endchoice
endmenu
Hi,
On Thu, 15 Jan 2004, David Brownell wrote:
> > choice values can also be tristate symbols, so you wouldn't need the
> > separate defines, unless you really always want to compile only a single
> > controller (even as module).
>
> That's it precisely. USB devices have only one (upstream) link;
> they're not like hosts. And its link to the controller isn't
> re-wired on the fly any more than, say, the MMU. Kconfig just
> needed some persuasion before it'd dance that way.
It's still weird. Where is the problem to compile all controllers as
module? At runtime you still have the possibility that only one of them
can be loaded and the next one would fail to load.
> And similar for net2280, pxa2xx, and so on. Either that, or moving it
> up higher in the text file, seems to have been the black magic that
> made the menu layout code behave.
To fix the menu layout it's probably the easiest to put most of it within
a "if USB_GADGET" ... "endif".
> > I'm also not sure about USB_PXA2XX_SMALL, as it also can be written as:
> >
> > config USB_PXA2XX_SMALL
> > depends on USB_PXA2XX = y
> > default USB_ZERO = y || USB_ETH = y || USB_G_SERIAL
> >
> > is this really intended?
>
> I'm not sure what you're asking. I wrote it with one line per
> driver that's less error-prone in case updates get merged. The
> latest version is more terse, but there are lots of ways to
> write that kind of logic.
The comment confuses me, I don't see how it tests that it's "only one" of
something.
> Reproduced it again here today, with a reasonably current 2.6.1
> tree on top of RH9 (plus some updated RPMs from RH). It's there
> in gconfig too. The workaround is "vi .config" and delete the
> sticky DUMMY_HCD entry, then re-configure.
It really works fine here, are you sure you don't have any additional
changes under scripts/kconfig? Did you try this on a different machine?
bye, Roman
Roman Zippel wrote:
> Hi,
>
> On Thu, 15 Jan 2004, David Brownell wrote:
>
>> > choice values can also be tristate symbols, so you wouldn't need the
>> > separate defines, unless you really always want to compile only a
>> > single controller (even as module).
>>
>> That's it precisely. USB devices have only one (upstream) link;
>> they're not like hosts. And its link to the controller isn't
>> re-wired on the fly any more than, say, the MMU. Kconfig just
>> needed some persuasion before it'd dance that way.
>
>
> It's still weird. Where is the problem to compile all controllers as
> module? At runtime you still have the possibility that only one of them
> can be loaded and the next one would fail to load.
The problem is something I accidently neglected to respond to in your
previous note. Those controllers aren't drop-in substitutes for one
another, and you seemed to be assuming they were.
>> > The dependency "USB_DUMMY_HCD || USB_NET2280 || USB_PXA2XX || USB_SA1100
>> > || USB_GOKU" can be basically reduced to "USB_GADGET".
That's where I accidentally trimmed out a response. It's wrong; those
controllers are not equivalent in general. Gadget drivers that work with
one controller will not automagically work with another. Two quick examples
should illustrate the point:
(i) A "video" gadget driver would requires isochronous transfer support.
But only two of those controllers support that (net2280, pxa2xx_udc).
So a video (or audio) class gadget driver would depend on configuring
an appropriate controller ... and even then, it'd probably compile in
more code if it's talking to a controller that supports high bandwidth
transfer modes (up to 24 MByte/sec) than one maxing ouat a 1 Mbyte/sec.
(ii) Essentially every controller has endpoints that differ from the others
in capabilities like addressing, direction, supported packet sizes
transfer modes, or throughput. All of those are things that MUST be
reported to the USB host in device descriptors.
The differences are currently all handled by conditional compilation.
Until a gadget driver has that support for a particular controller,
it shouldn't be possible to configure that combination.
There's some work afoot to do more autoconfiguration, but it won't be able
to handle all the issues for all drivers. I have hopes that at least some
simple configurations -- two endpoints for bulk data (IN/OUT) and one
for status (interrupt IN), for example -- will be able to autoconfigure
at some point.
>> Reproduced it again here today, with a reasonably current 2.6.1
>> tree on top of RH9 (plus some updated RPMs from RH). It's there
>> in gconfig too. The workaround is "vi .config" and delete the
>> sticky DUMMY_HCD entry, then re-configure.
>
>
> It really works fine here, are you sure you don't have any additional
> changes under scripts/kconfig? Did you try this on a different machine?
Absolutely certain. Yes, tried on multiple machines. Did you try
it on RH9? Since it's working for you, I wonder if that's the source
of the difference. Same bug in xconfig, gconfig, and menuconfig;
doesn't make a lot of sense to me but then I've not looked at that
code in any case.
- Dave