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
rounded up to the next cache size from kmalloc_sizes.h, which happens to
include exactly 32K.

Anyway, I think there are two questions to be answered here:

1. Is it worth using MAX_USBFS_BUFFER_SIZE for Isochronous transfers
   as well, for consistency with every other transfer type?

2. What value should MAX_USBFS_BUFFER_SIZE have?

I don't see a reason to allow only 16K for bulk transfers while we
allow 32K for isochronous URBs- but maybe there's something I'm missing.

At this point the argument is mostly academic to me- the overhead of using
16K URBs instead of 32K URBs doesn't bother me as much as the inconsistency
between bulk and iso bothers me.

Thanks,
--Micah



-------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
[email protected]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel

Reply via email to