On 7/12/19 4:24 PM, Michal Suchánek wrote: > On Fri, 12 Jul 2019 06:05:45 +0200 > Marek Vasut <ma...@denx.de> wrote: > >> On 7/10/19 5:47 PM, Michal Suchánek wrote: >>> On Fri, 5 Jul 2019 14:12:36 +0200 >>> Marek Vasut <ma...@denx.de> wrote: >>> >>>> On 7/5/19 12:44 PM, Michal Suchanek wrote: >>>>> Causes unbound key repeat on error otherwise. >>>>> >>>>> Signed-off-by: Michal Suchanek <msucha...@suse.de> >>>>> --- >>>>> v2: fix indentation >>>> >>>> What changed in V3 ? >>>> >>>>> --- >>>>> common/usb_kbd.c | 7 +++---- >>>>> 1 file changed, 3 insertions(+), 4 deletions(-) >>>>> >>>>> diff --git a/common/usb_kbd.c b/common/usb_kbd.c >>>>> index cc99c6be0720..fc9419e0238a 100644 >>>>> --- a/common/usb_kbd.c >>>>> +++ b/common/usb_kbd.c >>>>> @@ -339,10 +339,9 @@ static inline void usb_kbd_poll_for_event(struct >>>>> usb_device *dev) >>>>> struct usb_kbd_pdata *data = dev->privptr; >>>>> >>>>> /* Submit a interrupt transfer request */ >>>>> - usb_submit_int_msg(dev, data->intpipe, &data->new[0], data->intpktsize, >>>>> - data->intinterval); >>>>> - >>>>> - usb_kbd_irq_worker(dev); >>>>> + if (!usb_submit_int_msg(dev, data->intpipe, &data->new[0], >>>>> + data->intpktsize, data->intinterval)) >>>> >>>> This still doesn't propagate errors. >>> >>> Yes, it does not. I don't have a grand design that would make use of >>> these propagated errors so I don't know what errors to propagate. >> >> Each and every called of usb_kbd_poll_for_event() returns some return >> value, > A void >> so these functions should at least be aware of the new errors >> which might come from usb_kbd_poll_for_event(). > > No errors come from the function. The u-boot input flow is designed for > usb_kbd_poll_for_event handling errors internally. > > The function is only used by the usb_kbd specific testc and getc. The > input users are supposed to use testc to see if there is input > available and getc to get input (and it should not block if testc > indicated that there is input). The result of testc is treated as > boolean so there is no room for error handling. The result of getc is > treated as character input and I found no user that implements any kind > of error handling. Also at the getc time the testc error is no longer > available. > > If you want to redesign it to handle errors in the caller then please > do share the design. Without that it is not clear which errors make > sense to propagate.
Presumably, usb_kbd_tstc() should return 0 if usb_kbd_poll_for_event() returns an error code or somesuch ? _______________________________________________ U-Boot mailing list U-Boot@lists.denx.de https://lists.denx.de/listinfo/u-boot
