Oliver Neukum wrote:
If you allow exactly 32768 the management information for kmalloc will
need additional space pushing the allocation to 64K.
How is this the case for allocations >= PAGE_SIZE? From my understanding
of the slab allocator, the size you give kmalloc() will be immediately
rounde
Am Donnerstag, 18. Mai 2006 20:33 schrieb Micah Dowty:
> 3. Isochronous transfers had a higher arbitrary buffer size
> limit than bulk/interrupt transfers. Isochronous transfers
> now use MAX_USBFS_BUFFER_SIZE, which has been raised to 32K.
> These limits are in place not to enforce the US
(Sorry for the duplicate- I should know better than to try
sending patches from Outlook... This should fix the word
wrapping.)
This patch cleans up a few arbitrary limits in usbfs:
1. The data stage of a control request was being limited
to 4096-8 bytes, when in actuality a 4096-byte data sta
On Thu, May 18, 2006 at 11:33:39AM -0700, Micah Dowty wrote:
> This patch cleans up a few arbitrary limits in usbfs:
>
> 1. The data stage of a control request was being limited
>to 4096-8 bytes, when in actuality a 4096-byte data stage
>is legal. This is necessary to write userspace drive
This patch cleans up a few arbitrary limits in usbfs:
1. The data stage of a control request was being limited
to 4096-8 bytes, when in actuality a 4096-byte data stage
is legal. This is necessary to write userspace drivers
for some devices, such as ATI's USB TV-Wonder tuner.
2. The arbi
