As discussed with Baolin Wang, Michal Nazarewicz and Felipe Balbi. I propose the forced buffer alignment of OUT endpoints USB requests. This is implemented by patches #1 and #2. That not just simplifies the driver code, but it also prevents nasty bugs when buflen is not aligned or even less than wMaxPacketSize. Patch #9 removes direct calls to usb_ep_alloc_request() and use alloc_ep_req() instead. If accepted, then we should apply to all other gadgets that uses usb_ep_alloc_request() when possible and encorage drivers to use it instead. Changes from v2: * Simplified logic in patch 1 * Added documentation to alloc_ep_req and free_ep_req * Improved commit message on patch 7 Changes from v1: * Added patches 1, 2, 7, 8 ,9. * Patch 3 removes max_t() for buffer alignment with wMaxPacketSize Felipe F. Tonello (9): usb: gadget: fix usb_ep_align_maybe endianness and new usb_ep_align usb: gadget: align buffer size when allocating for OUT endpoint usb: gadget: f_midi: remove alignment code for OUT endpoint usb: gadget: f_midi: defaults buflen sizes to 512 usb: gadget: f_midi: refactor state machine usb: gadget: f_midi: drop substreams when disabling endpoint usb: gadget: remove useless parameter in alloc_ep_req() usb: gadget: f_hid: use free_ep_req() usb: gadget: f_hid: use alloc_ep_req() drivers/usb/gadget/function/f_hid.c | 26 +--- drivers/usb/gadget/function/f_loopback.c | 9 +- drivers/usb/gadget/function/f_midi.c | 240 +++++++++++++++++------------ drivers/usb/gadget/function/f_sourcesink.c | 11 +- drivers/usb/gadget/legacy/gmidi.c | 2 +- drivers/usb/gadget/u_f.c | 6 +- drivers/usb/gadget/u_f.h | 17 +- include/linux/usb/gadget.h | 17 +- 8 files changed, 188 insertions(+), 140 deletions(-) -- 2.9.0 -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
