On 03.07.13, Ming Lei wrote:
diff --git a/drivers/usb/core/hcd.c b/drivers/usb/core/hcd.c
index 014dc99..26471cd 100644
--- a/drivers/usb/core/hcd.c
+++ b/drivers/usb/core/hcd.c
@@ -1648,6 +1635,72 @@ int usb_hcd_unlink_urb (struct urb *urb, int status)
On Fri, 8 Nov 2013, Sebastian Andrzej Siewior wrote:
On 03.07.13, Ming Lei wrote:
...
+ /*
+* We disable local IRQs here avoid possible deadlock because
+* drivers may call spin_lock() to hold lock which might be
+* acquired in one hard interrupt handler.
+*
+*
On Wed, Jul 03, 2013 at 10:53:07PM +0800, Ming Lei wrote:
1), for drivers, they don't care if the complete() is called in hard irq
context or softirq context
What about those 50+ patches you had that changed the locking for urb
callbacks? Aren't they needed before this patch can go in?
---
On Fri, Aug 2, 2013 at 11:28 AM, Greg Kroah-Hartman
gre...@linuxfoundation.org wrote:
On Wed, Jul 03, 2013 at 10:53:07PM +0800, Ming Lei wrote:
1), for drivers, they don't care if the complete() is called in hard irq
context or softirq context
What about those 50+ patches you had that changed
On Fri, Aug 02, 2013 at 11:30:54AM +0800, Ming Lei wrote:
On Fri, Aug 2, 2013 at 11:28 AM, Greg Kroah-Hartman
gre...@linuxfoundation.org wrote:
On Wed, Jul 03, 2013 at 10:53:07PM +0800, Ming Lei wrote:
1), for drivers, they don't care if the complete() is called in hard irq
context or
This patch implements the mechanism of giveback of URB in
tasklet context, so that hardware interrupt handling time for
usb host controller can be saved much, and HCD interrupt handling
can be simplified.
Motivations:
1), on some arch(such as ARM), DMA mapping/unmapping is a bit
time-consuming,
