Mirrors_Framework/nuttx
1
0
Fork 0
mirror of https://github.com/apache/nuttx.git synced 2026-05-30 13:27:01 +08:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

USB hub: Add a configuration option to fallback to polling mode if the the HCD async method does not work. Don't call DRVR_CANCEL twice, and certainly not from the interrupt level. Add so checks so that we do not do some normal asynchronous actions if the hub has been disconnected

Browse Source
This commit is contained in:
Gregory Nutt
2015-04-30 11:30:01 -06:00
parent f36b833955
commit 1380945a13
2 changed files with 80 additions and 17 deletions
Download Patch File Download Diff File Expand all files Collapse all files
drivers/usbhost/Kconfig
+26
View File
@@ -51,6 +51,32 @@ config USBHOST_HUB
---help---
Select this option to build in support for USB hubs.
if USBHOST_HUB
config USBHOST_HUB_POLLMSEC
int "USB Hub Polling Interval (MSec)"
default 400
---help---
On higher end host controllers (OHCI and EHCI), the asynchronous,
interrupt IN transfers will pend until data is available from the
hub. On lower end host controllers (like STM32 and EFM32), the
transfer will fail immediately when the device NAKs the first
attempted interrupt IN transfer (with result == EGAIN) and the hub
class driver will fall back to polling the hub.
For the case of the higher end controllers, this polling interval
is not critical since it would really only be used in the event of
failures to communicate with the hub.
But for the lower end host controllers, the asynchronous transfers
are ineffective and this polling interval becomes a critical
parameter that must be tuned to tradeoff CPU usage with
responsiveness to hub-related events (It would, I suppose, be more
efficient to use synchronous transfers with these lower end host
controllers).
endif # USBHOST_HUB
config USBHOST_MSC
bool "Mass Storage Class Support"
default n