This patch introduces a configuration option USBMSC_WRMULTIPLE,
which is used to store multiple blocks into a larger chunk that
then gets written via the mmcsd_writemultiple() (in case mmcsd
is used).
The bottleneck with the current implementation is the poor
performance due to short block writes. USBMSC_DRVR_WRITE()
always writes only one sector (with eMMC that's usually 512 bytes).
eMMC devices usually erase much larger regions with near constant
time (see Jedec JESD84-B51, Extended CSD register byte [225],
SUPER_PAGE_SIZE): 'This register defines one or multiple of
programmable boundary unit that is programmed at the same time.'
If USBMSC_WRMULTIPLE is defined, then USBMSC_NWRREQS is used to
allocate the write buffer size. We don't want this to be the
default behavior yet as this may reveal unseen bugs in usb drivers
due to the faster overall performance.
Sample configurations with measured performance:
- Without USBMSC_WRMULTIPLE: 470 Kb/s
- With USBMSC_WRMULTIPLE, CONFIG_USBMSC_NWRREQS=4: 1.1 Mb/s
(dd with bs=2k)
- With USBMSC_WRMULTIPLE, CONFIG_USBMSC_NWRREQS=16: 5.2 Mb/s
(dd with bs=8k)
No doubt, this feature alone may make the mass storage work 10
times faster than before with eMMC cards.
Signed-off-by: Eero Nurkkala <eero.nurkkala@offcode.fi>
usbmsc.h uses the CONFIG_USBMSC_IFNOBASE configuration option,
but it cannot be set anywhere. Provide means to configure this
option which is useful when it's other than the default zero.
Signed-off-by: Eero Nurkkala <eero.nurkkala@offcode.fi>
iSerialNumber field in the device descriptor can be used to determining the
board when multiple boards connected to the same host. So add feature to change
serial string by board unique ID dynamically.
To use this feature, user must be implement the board_usbdev_serialstr() logic.
refs #13909
The new configuration CONFIG_USBMSC_NOT_STALL_BULKEP avoids to go into
stall state when :
- Receiving SCSI_CMD_INQUIRY with flags or pagecode != 0
- Receiving SCSI_CMD_MODESENSE6 and return the result data shorter than
the requested data size.
Instead of stall, it just ignores the error condition.
Originally the usb driver goes into the stall under the conditions
mentioned above to inform to the USB host that the requested SCSI feature
is not supported, or the result data is shorter than the requested data
size.
But, at this moment, RP2040 USB device driver cannot handle entering and
leaving stall state of the bulk endpoints well. Once stalls, the host -
device communication stops and cannot be resumed.
I have investigated to solve the issue and found that the existing
RP2040 USB device driver implementation, TinyUSB
(https://github.com/hathach/tinyusb) mass-storage class driver doesn't
enter the stall state like the treatment of this patch.
So, I introduce this new configuration as the RP2040 workaround.
In the future, when the RP2040 USB device driver is fixed and can handle
the bulk endpoint stall correctly, this patch should be reverted.
Summary:
- Change default value for RNDIS_NWRREQS based on NET_TCP_WRITE_BUFFERS
Impact:
- This commit affects RNDIS use cases
Testing:
- spresense:rndis with CONFIG_NET_TCP_WRITE_BUFFERS=y
Signed-off-by: Masayuki Ishikawa <Masayuki.Ishikawa@jp.sony.com>
This commit does two things:
1. First, it reorganizes the driver Kconfig files so that each is self contained. Before, a part of the driver configuration was in drivers/Kconfig and the rest was in in drivers/xyz/Konfig. Now, all of the driver configuration is consolitated in the latter.
2. Second, this commit correct numerous serious errors introduced in a previous reorganization of the driver Kconfig files. This was first noted by Nicholas Chin in PR270 for the case of the drivers/i2c/Kconfig but some examination indicates that the error was introduced into several other Kconfig files as well.
The nature of the introduced error was basically this:
- Nothing must intervene between the menuconfig selection and the following conditional configuration otpions.
- A previous PR erroneously introduced unconditional options between the menuconfig and the following confditional logic, thus corrupting the driver menus.
This error was easy to make because the driver Kconfig files were not well modularized. Making them fully self-contained should eliminate this kind of error in the future.
Initial OTG support for STM32H7 and some minor improvements
arch/arm/src/stm32h7/stm32h7x3xx_rcc.c: add support for HSI48 clock
configs/nucleo-f207zg: add support for USB and initialize CDCACM if configured
configs/nucleo-h743zi: add support for USB and initialize CDCACM if configured
drivers/usbdev/Kconfig: fix type for USBDEV_TRACE_INITIALIDSET
configs/teensy-3.x/usbnsh/defconfig: update config according to change in USBDEV_TRACE_INITIALIDSET
Approved-by: Gregory Nutt <gnutt@nuttx.org>
RNDIS composite support
* NuttX usb/composite.h: Forward-declare composite_devdesc_s.
This avoids "error: conflicting types for 'composite_initialize'"
on some versions of GCC. Because of the cross-inclusion between
usbdev.h and composite.h, the full declaration is not always
available.
* NuttX: USB Composite driver: Fix strid comparison
The last string ID used by composite driver is 4, and
the number of IDs used is 5 (0..4). The comparison
strid <= COMPOSITE_NSTRIDS caused composite driver to
reply with -EINVAL for id 5, even though it should be
available for subdevices to use.
* NuttX: RNDIS USB driver: Add support for composite configuration.
Approved-by: GregoryN <gnutt@nuttx.org>
These Microsoft-only descriptors help in loading the correct driver on Windows.
They are especially helpful to give libusb access to a custom device without
having to manually configure/install WinUSB driver.
With this change DFU interface works automatically on
Windows 10 with dfu-util 0.9 and libusb 1.0.22. On Windows 7
it still appears to need driver installation.
commit 69fcf3e849
Author: Alan Carvalho de Assis <acassis@gmail.com>
Date: Sun Jul 16 08:39:33 2017 -0600
Fix spark/stm32_composite.c: board_composite_connect cannot be static
commit 28eb253401
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Sun Jul 16 08:36:01 2017 -0600
Composite: Final review for coding style before merge
commit e6af1b9994
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Sun Jul 16 07:41:38 2017 -0600
Composite: Simplify some intiialization of data structures.
commit 771c367411
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Sun Jul 16 07:15:08 2017 -0600
Cosmetic changes to alignment.
commit 5d67ddda4e
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Sun Jul 16 07:00:48 2017 -0600
USBMSC: Add missing logic to saved device description.
commit 0729151d29
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Sat Jul 15 17:11:41 2017 -0600
Trivial, costmetic
commit 74b916ff84
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Sat Jul 15 14:50:29 2017 -0600
Composite: Private functions need to be marked static. Move static functions out of 'Public Functions' to 'Private Functions' where they belong. Disable composite configuration 1 in all STM32 F1 configurations.
commit cfaa4ece13
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Sat Jul 15 13:20:34 2017 -0600
Add some comments.
commit 8143563be6
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Sat Jul 15 12:33:10 2017 -0600
Spark: Need to condition out MSC logic in composite setup if there is no MSC in the composite.
commit 69d3a91ef1
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Sat Jul 15 11:03:33 2017 -0600
Composite: Remove all dependencies on CONFIG_SYSTEM_COMPOSITE_* configuration settings. Nothing in the OS can depend on external application settings.
commit 55a4388bbd
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Sat Jul 15 09:59:31 2017 -0600
All composite configurations now also support a dual CDC/ACM configuration.
commit 428f2147af
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Sat Jul 15 09:24:32 2017 -0600
Composite: Move board_msc* interfaces from apps/system/composite to the board specific OS logic where they belong.
commit f1cc168a5c
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Sat Jul 15 08:56:33 2017 -0600
Refresh all composite configurations.
commit 246afcaa10
Merge: 919877191d02c6672868
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Sat Jul 15 08:22:26 2017 -0600
Merge remote-tracking branch 'origin/master' into composite
commit 919877191d
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Sat Jul 15 08:20:02 2017 -0600
Composite setup: Remove useless board_cdc* wrapper.
commit 82129cf8c6
Merge: f2cb8b252a6537e4ea20
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Fri Jul 14 16:23:57 2017 -0600
Merge remote-tracking branch 'origin/master' into composite
commit f2cb8b252a
Author: Alan Carvalho de Assis <acassis@gmail.com>
Date: Fri Jul 14 10:19:35 2017 -0600
Composite: Fix for another cloned typo.
commit 676cfd526a
Author: Alan Carvalho de Assis <acassis@gmail.com>
Date: Fri Jul 14 09:11:37 2017 -0600
Composite: Fix some typos
commit 1ea0368c18
Author: Alan Carvalho de Assis <acassis@gmail.com>
Date: Fri Jul 14 09:10:18 2017 -0600
Composite: ./stm3210e-eval/src/stm32_composite.c
commit e485caced9
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Fri Jul 14 09:08:17 2017 -0600
Composite: I don't think the original code should have forced minor=0.
commit 6443c29621
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Fri Jul 14 07:15:38 2017 -0600
Composite: Flesh out support for all of other configurations that support composite.
commit 23cbc28b05
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Fri Jul 14 06:59:45 2017 -0600
Detangle use of board_xyzclassobject() and board_xyzuninitialize()
commit 1674cb8c8e
Merge: 6bc881a1925033a6def7
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Thu Jul 13 13:57:40 2017 -0600
Merge remote-tracking branch 'origin/master' into composite
commit 6bc881a192
Merge: fe3af4941d85b8d16d8c
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Tue Jul 11 12:24:07 2017 -0600
Merge remote-tracking branch 'origin/master' into composite
commit fe3af4941d
Merge: 0f9ad16e181bc0eea143
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Mon Jul 10 11:07:36 2017 -0600
Merge remote-tracking branch 'origin/master' into composite
commit 0f9ad16e18
Merge: a4cd90d4efaa2e9c15a5
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Fri Jul 7 20:26:53 2017 -0600
Merge remote-tracking branch 'origin/master' into composite
commit a4cd90d4ef
Merge: 8a4be7175e31f832d8c5
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Wed Jul 5 11:12:52 2017 -0600
Merge remote-tracking branch 'origin/master' into composite
commit 8a4be7175e
Merge: 18a32ed2caae1771454a
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Fri Jun 30 16:14:04 2017 -0600
Merge remote-tracking branch 'origin/master' into composite
commit 18a32ed2ca
Merge: aaa81ce4976d8df90b79
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Thu Jun 29 10:18:16 2017 -0600
Merge remote-tracking branch 'origin/master' into composite
commit aaa81ce497
Merge: 4eb548226b8cb4636bb1
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Mon Jun 26 11:56:11 2017 -0600
Merge remote-tracking branch 'origin/master' into composite
commit 4eb548226b
Merge: 2327f5a1b4dc8eec0b61
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Mon Jun 19 17:27:00 2017 -0600
Merge remote-tracking branch 'origin/master' into composite
commit 2327f5a1b4
Merge: 49cd279fc699bf0b522b
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Fri Jun 16 17:30:03 2017 -0600
Merge remote-tracking branch 'origin/master' into composite
commit 49cd279fc6
Merge: bb6a13f30a46f86982ee
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Wed Jun 14 09:17:49 2017 -0600
Merge remote-tracking branch 'origin/master' into composite
commit bb6a13f30a
Merge: 918480047aac93d4bda9
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Mon Jun 5 17:40:06 2017 -0600
Merge remote-tracking branch 'origin/master' into composite
commit 918480047a
Merge: e4d262436c4526cd665e
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Sat Jun 3 08:52:31 2017 -0600
Merge remote-tracking branch 'origin/master' into composite
commit e4d262436c
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Fri Jun 2 07:19:27 2017 -0600
SAMV71-Xult Composite: Now can switch between two different composite configurations dynamically.
commit 815257743d
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Fri Jun 2 07:11:57 2017 -0600
usbdev composite and SAMV7-Xult: Move board-specific USB composite configuration out of boardctl.c and into board-specific logic where it belongs. Add a configuration option to the boardctl() calls to support multiple composite device configurations dynamically.
commit ac13619dc5
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Thu Jun 1 17:03:58 2017 -0600
Cosmetic
commit 9dd41bdd2f
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Thu Jun 1 17:02:43 2017 -0600
Composite: More compile-related fixes
commit fc1438c95d
Merge: 049ccbfcbeff2b54a5e0
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Thu Jun 1 16:35:41 2017 -0600
Merge remote-tracking branch 'origin/master' into composite
commit 049ccbfcbe
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Thu Jun 1 16:35:16 2017 -0600
Composite: Add some structure definitions missed in first application of the patch.
commit ef33329e3a
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Thu Jun 1 16:14:46 2017 -0600
Add a warning
commit 89f77cd91a
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Thu Jun 1 16:11:27 2017 -0600
Fix some incomplete name changes
commit 0bb7af549a
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Thu Jun 1 15:09:50 2017 -0600
It is unnecessary to pack a structure that consists only of uint8_t data fields.
commit bd9b548914
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Thu Jun 1 15:05:41 2017 -0600
Remove COMPILE_TIME_ASSERTION
commit 7e6f481581
Author: Frank Benkert <Frank.Benkert@avat.de>
Date: Thu Jun 1 14:58:04 2017 -0600
Part II of the same big commit
commit dcc9b07715
Author: Frank Benkert <Frank.Benkert@avat.de>
Date: Thu Jun 1 14:08:22 2017 -0600
[[This is part 1 or several commits]]
We developed a huge Changeset over a year ago to make USB Composite configuration dynamical and be able to instanciate the CDCACM multiple times inside this device. We use this feature to switch between one in normal and up to three CDCACMs in maintenance boot.
The control path starts in “boardctl.c” where the configuration for the device is constructed. There are still a few issues which I’ll ask you to have a look at before this beast can be merged.
1. To be able to construct the data dynamically I have changed some USB-Structs to be packed. Maybe there are additional structs to change (just for completeness – not for current functionality).
2. I’ve added the Macro “COMPILE_TIME_ASSERTION” two times (in usbmsc_desc.c and in cdcacm_desc.c) to stay private. Maybe you’ll find a better place. It’s used to check the size of the structs against the assumptions.
3. I’ve changed the interface for some USB-Functions to receive also the dynamic configuration. Maybe this can be done more elegant.
4. The original NuttX (without the patch) seems to have problems with a Composite device holding a CDCACM and an MSC. The “USB SET CONFIGURATION” request does not to work at all. This makes the test fail under Windows and under Linux. Applying this patch doesn’t change anything – because it only changes the configuration – not the behavior. Maybe you’ll have a look at this problem before applying the patch.
