| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
| |
Makes new usb implementation work where submissions must be in order.
|
| |
|
|
|
|
| |
This was messing up the pad to 512 bytes logic for empty flushes.
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
| |
To properly do this, the Boost_LIBRARIES needs to be sanitized.
That will be saved for another commit...
|
|
|
|
| |
timeouts
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
| |
Once the send buffer size is increased,
there appears to be no performance advantage.
I would rather not have extra implementations.
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|\ |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
|/ |
|
|
|
|
|
|
|
| |
this fixes a bug when there are multiple mboard per device,
the occupancy int could be uninitialized to a bad value,
while subdev specs are being setup for the first time,
and cause allocation issues
|
| |
|
|
|
|
| |
10.4 compiler bug)
|