Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1758120Ab0FBMrV (ORCPT ); Wed, 2 Jun 2010 08:47:21 -0400 Received: from mailout3.w1.samsung.com ([210.118.77.13]:57176 "EHLO mailout3.w1.samsung.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1757226Ab0FBMo5 (ORCPT ); Wed, 2 Jun 2010 08:44:57 -0400 MIME-version: 1.0 Content-transfer-encoding: 8BIT Content-type: TEXT/PLAIN; CHARSET=EUC-KR Date: Wed, 02 Jun 2010 14:45:21 +0200 From: Michal Nazarewicz Subject: [PATCHv3 00/11] USB improvements and cleanpus, g_multi install mode To: linux-usb@vger.kernel.org, linux-usb@vger.kernel.org Cc: David Brownell , Kyungmin Park , Marek Szyprowski , linux-kernel@vger.kernel.org Message-id: X-Mailer: git-send-email 1.7.1 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 7672 Lines: 177 Hello again, It's me again with a bunch of patches but this time I bring some new functionality! Install mode! If you are interested only in this jump near the end of the mail. The first 7 patches are what I've posted previously. I'm resending it again since those are required for later patches plus there were some minor fixes. The patches are: USB: gadget: g_mass_storage: static data instead of dynamic allocation A micro optimisation preventing a kmalloc call() which would otherwise be always called at the very beginning for the same area size. This was submitted previously but Greg did not included it in his quilt tree (unless I overlooked something) so I'll posting it again for completeness. Sorry if that creates confusion. In my opinion this is worth including as it simplifies the code a bit and removes the call to kmalloc/kfree (hidden in fsg_common_init) which in my opinion is a good thing as well. USB: gadget: f_mass_storage: fsg_add() renamed to fsg_bind_config() USB: gadget: f_fs: functionfs_add() renamed to functionfs_bind_config() All other USB composite functions use foo_bind_config() to add function to USB configuration so changed the names for mass storage and FunctionFS as well. The inconstancy apparent in multifunction gadgets stroke me for quite some time now. I was always reluctant to submit patch thinking that it would be just stupid name change but finally couldn't stand it. ;) In case of mass storage the old (fsg_add()) function is still available but marked deprecated. In case of FFS the old (functionfs_add()) function has been removed. USB: gadget: composite: usb_string_ids_*() functions added USB: gadget: f_fs: use usb_string_ids_n() usb_string_ids_n() and usb_string_ids_tab() functions added for batch sting IDs requests. This simplifies the code where one would call usb_string_id() several times and each time has to check for error codes. With usb_string_ids_n() one can get a whole bunch of IDs with one call and with usb_string_ids_tab() fill an array full of usb_string structures. USB: gadget: g_multi: code clean up and refactoring USB: gadget: g_multi: added documentation and INF files g_multi clean up, fixes, etc. Also a bit of documentation for g_multi and an INF file tested with Windows XP SP3. Hopefully, someone will find it useful. Then, there is this patch David didn't seem impressed with: USB: gadget: g_multi: more configurable Adds a Kconfig option for each function provided by the g_multi. This may look like making g_multi complex but I see it more as creating an gadget that can be used to test various functionality as well as show that adding new functions is not so hard even in a complex gadget since the code adds a clear separation of each function so that one may easily see which parts are for each function. And, finally 3 new patches with the advertised install mode functionality: USB: gadget: composite: addad disconnect callback Adds an disconnect callback to the composite framework which is called whenever composite receives disconnect from the hardware driver. USB: gadget: f_mass_storage: added eject callback Adds a pre- and post-eject callbacks. The first may prevent eject from really happen even though no error is reported to the host so host things it succeeded. The second is called after the file is closed so one may use it for some funky stuff which is otherwise impossible while mass storage function holds the file or mounting another file. USB: gadget: g_multi: Install Mode added This adds the install mode to the g_multi. Not to repeat myself, quote from documentation: * Install mode The install mode makes the gadget appear as a plain mass storage device the first time it is connected (and after each disconnect). This lets one develop an "autorun" CD-ROM image with drivers and put it as the first logical unit. ** Workings of the install mode As you may know, mass storage gadget may provide several logical units and its easier to think of them as separate drives. When install mode is enabled, g_multi forces the first logical unit to be a read-only CD-ROM. When install mode is enabled but mass storage itself is not then exactly one logical unit is set. When an eject request is made on that logical unit, the file is not really closed but the gadget switches it's mode to the full flagged gadget with all the other functions. If mass storage is among them, the firs logical unit will be the CD-ROM image with drivers (which may be seen as a bad thing). When gadget is disconnected and connected afterwards it will work again in install mode. Some heuristics are used here -- if disconnection (or suspend) happens no longer then 10 seconds after last eject on the first logical unit then on next enumeration gadget will claim to be full flagged otherwise it'll stick to install mode. ** Interoperability with host As said, the idea behind install mode is that hosts that require drivers will be able to get them without the need for additional CD-ROM or another medium provided with the device. CD-ROM image should provide an "autorun" functionality witch will install drivers and eject the emulated CD-ROM to switch gadget into the other mode which will be handled by newly installed drivers. If drivers are installed already, they should "catch" the install mode device by product and vendor IDs and issue an eject. This mode is not very Linux-friendly though since Linux and Linux based systems have no notion of autorun (which from security point of view is a good thing) and there's no way of adding some file on the image which will make gadget eject the device. Fortunately, there's USB_ModeSwitch and/or udev which should handle it just fine. A single rule need to be added and everything should work fine. Documentation/usb/gadget_multi.txt | 221 +++++++++ Documentation/usb/gadget_multi_rndis.inf | 200 ++++++++ Documentation/usb/gadget_multi_serial.inf | 44 ++ drivers/usb/gadget/Kconfig | 88 +++- drivers/usb/gadget/composite.c | 73 +++- drivers/usb/gadget/f_fs.c | 36 +- drivers/usb/gadget/f_mass_storage.c | 122 +++-- drivers/usb/gadget/g_ffs.c | 2 +- drivers/usb/gadget/mass_storage.c | 17 +- drivers/usb/gadget/multi.c | 744 ++++++++++++++++++++++------- include/linux/usb/composite.h | 6 + include/linux/usb/functionfs.h | 6 +- 12 files changed, 1278 insertions(+), 281 deletions(-) create mode 100644 Documentation/usb/gadget_multi.txt create mode 100644 Documentation/usb/gadget_multi_rndis.inf create mode 100644 Documentation/usb/gadget_multi_serial.inf -- Best regards, _ _ | Humble Liege of Serenely Enlightened Majesty of o' \,=./ `o | Computer Science, Michał "mina86" Nazarewicz (o o) +----[mina86*mina86.com]---[mina86*jabber.org]----ooO--(_)--Ooo-- -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/