Hi,
Vincent Pelletier <plr.vincent@xxxxxxxxx> writes:
> On Mon, 16 Jan 2017 14:48:31 +0000, Vincent Pelletier
> <plr.vincent@xxxxxxxxx> wrote:
>> 3) I declared 4 endpoints (2 IN, 2 OUT). And I went one level deeper
>> down the rabbit hole: now enumeration fails with this message on
>> host:
>
> Aaand... I dug this part of the hole myself. In the library part of my
> python module, and I was looking at the test program. Shame on me.
>
> Next tests: After a bit of patching f_fs so it accepts creating 30
> endpoints, more endpoints seem to have issues above the 7th (ignoring
> direction).
>
> 1IN bandwidth: 0 B/s (0.20s)
> 1OUT bandwidth: 30230053 B/s (0.21s)
> 2IN bandwidth: 29541962 B/s (0.22s)
> 2OUT bandwidth: 29995671 B/s (0.22s)
> 3IN bandwidth: 28787881 B/s (0.22s)
> 3OUT bandwidth: 30428794 B/s (0.22s)
> 4IN bandwidth: 29768777 B/s (0.22s)
> 4OUT bandwidth: 30377747 B/s (0.22s)
> 5IN bandwidth: 29521290 B/s (0.22s)
> 5OUT bandwidth: 30133012 B/s (0.22s)
> 6IN bandwidth: 29662245 B/s (0.22s)
> 6OUT bandwidth: 30335271 B/s (0.22s)
> 7IN bandwidth: 29467489 B/s (0.22s)
> 7OUT bandwidth: 30387586 B/s (0.22s)
> 8IN bandwidth: 0 B/s (0.20s)
> 8OUT bandwidth: 0 B/s (0.20s)
> 9IN bandwidth: 0 B/s (0.20s)
> 9OUT bandwidth: 30174377 B/s (0.22s)
> 10IN bandwidth: 0 B/s (0.20s)
> 10OUT bandwidth: 30389703 B/s (0.22s)
> 11IN bandwidth: 0 B/s (0.20s)
> 11OUT bandwidth: 29947720 B/s (0.21s)
> 12IN bandwidth: 0 B/s (0.20s)
> 12OUT bandwidth: 29386298 B/s (0.21s)
> 13IN bandwidth: 0 B/s (0.20s)
> 13OUT bandwidth: 30097119 B/s (0.22s)
> 14IN bandwidth: 0 B/s (0.20s)
> 14OUT bandwidth: 30099782 B/s (0.22s)
> 15IN bandwidth: 0 B/s (0.20s)
> 15OUT bandwidth: 31122053 B/s (0.21s)
>
> Would this pattern ring a bell ?
it could be that we're ran out of IN endpoints. There's a maximum number
to how many IN endpoints we can have enabled at one time and, currently,
dwc3 is not enforcing that in any way (I'll get that sorted out for
v4.12, v4.11 is already too late)
> I did the stall test again (on all endpoints and on 8IN + 9IN), and the
> halted endpoints behave as expected.
> I also verified the right data reaches/comes from the right endpoint
> (for those which work), without error.
> I did not check with my protocol analyser yet.
>
> I attach the patch against f_fs.c as an RFC.
>
> Regards,
> --
> Vincent Pelletier
> From fc77bb86dfbf088ce4a63510867210c2f2d8b0af Mon Sep 17 00:00:00 2001
> Message-Id: <fc77bb86dfbf088ce4a63510867210c2f2d8b0af.1484661254.git.plr.vincent@xxxxxxxxx>
> From: Vincent Pelletier <plr.vincent@xxxxxxxxx>
> Date: Tue, 17 Jan 2017 12:57:39 +0000
> Subject: usb: gadget: f_fs: Accept up to 30 endpoints.
>
> It is allowed by the USB specification to enabled same-address, opposite-
> direction endpoints simultaneously, which means 30 non-zero endpoints
> are allowed. So double eps_addrmap length to 30.
> The original code only accepted 14 descriptors, because the first
> eps_addrmap entry is unused (it is a placeholder for endpoint zero).
> So increase eps_addrmap length by one to 31.
>
> Signed-off-by: Vincent Pelletier <plr.vincent@xxxxxxxxx>
> ---
> drivers/usb/gadget/function/f_fs.c | 2 +-
> drivers/usb/gadget/function/u_fs.h | 2 +-
> 2 files changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/usb/gadget/function/f_fs.c b/drivers/usb/gadget/function/f_fs.c
> index 5c91a6f4613b..206184b561b7 100644
> --- a/drivers/usb/gadget/function/f_fs.c
> +++ b/drivers/usb/gadget/function/f_fs.c
> @@ -2101,7 +2101,7 @@ static int __ffs_data_do_entity(enum ffs_entity_type type,
> case FFS_ENDPOINT:
> d = (void *)desc;
> helper->eps_count++;
> - if (helper->eps_count >= 15)
> + if (helper->eps_count >= 31)
> return -EINVAL;
> /* Check if descriptors for any speed were already parsed */
> if (!helper->ffs->eps_count && !helper->ffs->interfaces_count)
> diff --git a/drivers/usb/gadget/function/u_fs.h b/drivers/usb/gadget/function/u_fs.h
> index 60139854e0b1..8eaf473d1ac0 100644
> --- a/drivers/usb/gadget/function/u_fs.h
> +++ b/drivers/usb/gadget/function/u_fs.h
> @@ -247,7 +247,7 @@ struct ffs_data {
>
> unsigned user_flags;
>
> - u8 eps_addrmap[15];
> + u8 eps_addrmap[31];
please send this as a proper patch using git send-email. Also, please
define a macro for this constant. If we ever need to change it again,
it'll change in one place.
thanks
--
balbi
Attachment:
signature.asc
Description: PGP signature
