linux-usb.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Ikjoon Jang <ikjn@chromium.org>
To: "Chunfeng Yun (云春峰)" <Chunfeng.Yun@mediatek.com>
Cc: "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	"linux-mediatek@lists.infradead.org"
	<linux-mediatek@lists.infradead.org>,
	"linux-usb@vger.kernel.org" <linux-usb@vger.kernel.org>,
	"Yaqii Wu (武亚奇)" <Yaqii.Wu@mediatek.com>,
	"mathias.nyman@intel.com" <mathias.nyman@intel.com>,
	"Eddie Hung (洪正鑫)" <Eddie.Hung@mediatek.com>,
	"linux-arm-kernel@lists.infradead.org"
	<linux-arm-kernel@lists.infradead.org>,
	"gregkh@linuxfoundation.org" <gregkh@linuxfoundation.org>,
	"matthias.bgg@gmail.com" <matthias.bgg@gmail.com>
Subject: Re: [PATCH next v2 3/6] usb: xhci-mtk: update fs bus bandwidth by bw_budget_table
Date: Mon, 30 Aug 2021 11:49:24 +0800	[thread overview]
Message-ID: <CAATdQgDKuO+bM7_tMy59ghhhjrgCp53kHgYbgEciriHNF3uKuA@mail.gmail.com> (raw)
In-Reply-To: <18cb18d7d0ea4a3bb46921aca88c53646f1b3764.camel@mediatek.com>

On Fri, Aug 27, 2021 at 5:49 PM Chunfeng Yun (云春峰)
<Chunfeng.Yun@mediatek.com> wrote:
>
> On Fri, 2021-08-27 at 17:14 +0800, Ikjoon Jang wrote:
> > On Fri, Aug 27, 2021 at 2:49 PM Chunfeng Yun (云春峰)
> > <Chunfeng.Yun@mediatek.com> wrote:
> > >
> > > On Thu, 2021-08-26 at 19:54 +0800, Ikjoon Jang wrote:
> > > > Hi Chunfeng,
> > > >
> > > > On Thu, Aug 26, 2021 at 10:52 AM Chunfeng Yun <
> > > > chunfeng.yun@mediatek.com> wrote:
> > > > >
> > > > > Use @bw_budget_table[] to update fs bus bandwidth due to
> > > > > not all microframes consume @bw_cost_per_microframe, see
> > > > > setup_sch_info().
> > > > >
> > > > > Signed-off-by: Chunfeng Yun <chunfeng.yun@mediatek.com>
> > > > > ---
> > > > > v2: new patch, move from another series
> > > > > ---
> > > > >  drivers/usb/host/xhci-mtk-sch.c | 17 +++++++----------
> > > > >  1 file changed, 7 insertions(+), 10 deletions(-)
> > > > >
> > > > > diff --git a/drivers/usb/host/xhci-mtk-sch.c
> > > > > b/drivers/usb/host/xhci-mtk-sch.c
> > > > > index cffcaf4dfa9f..83abd28269ca 100644
> > > > > --- a/drivers/usb/host/xhci-mtk-sch.c
> > > > > +++ b/drivers/usb/host/xhci-mtk-sch.c
> > > > > @@ -458,8 +458,8 @@ static int check_fs_bus_bw(struct
> > > > > mu3h_sch_ep_info *sch_ep, int offset)
> > > > >                  * Compared with hs bus, no matter what ep
> > > > > type,
> > > > >                  * the hub will always delay one uframe to send
> > > > > data
> > > > >                  */
> > > > > -               for (j = 0; j < sch_ep->cs_count; j++) {
> > > > > -                       tmp = tt->fs_bus_bw[base + j] + sch_ep-
> > > > > > bw_cost_per_microframe;
> > > > >
> > > > > +               for (j = 0; j < sch_ep->num_budget_microframes;
> > > > > j++) {
> > > > > +                       tmp = tt->fs_bus_bw[base + j] + sch_ep-
> > > > > > bw_budget_table[j];
> > > >
> > > > I'm worrying about this case with two endpoints,
> > > > * EP1OUT: isochronous, maxpacket=192: bw_budget_table[] = { 188,
> > > > 188,
> > > > 0, ... }
> > > > * EP2IN: interrupt, maxpacket=64: bw_budget_table[] = { 0, 0, 64,
> > > > 64,
> > > > ... }
> > > > (Is this correct bw_budget_table contents for those eps?)
> > >
> > > Yes, ep1out isoc use two uframe, ep2in intr use a extra cs;
> > > >
> > > > I'm not sure if it's okay for those two endpoints to be allocated
> > > > on the same u-frame slot.
> > > > Can you please check if this is okay for xhci-mtk?
> > >
> > > Already test it this afternoon, can transfer data rightly on our
> > > dvt
> > > env.
> > >
> > > > (I feel like I already asked the same questions many times.)
> > >
> > > Yes, as said before, prefer to use bw_budget_table[], if there is
> > > issue, we can fix it by building this table.
> >
> > So do you mean such an allocation shouldn't be a problem by IP
> > design?
> Yes, at least on our dvt platform

Did you check that your side also has a similar allocation
(SSPLIT-all sits between SSPLIT-start ~ -end for another ep)?
My audio headset doesn't work properly with this scheme.

>
> >
> > This patch starts to allow such an allocation (again).
> > But i remember my earlier tests showed that when those two eps
> > in the above example are allocated on the same u-frame slot,
> > xhci-mtk puts "SSPLIT for EP2" between
> > "SSPLIT-start and SSPLIT-end for EP1OUT transaction",
> > which is a spec violation.
>
> Which section in usb2.0 spec?

I think that's just a basic rule - if software wants to send 192 bytes
through a full-speed bus, HC should send OUT/DATA 192 bytes
continuously without inserting any other packets during that 192 bytes.
and usb2 11.14.2 mentions that TT has separated
Start-Split and Complete-Split buffers
but not tracked each transaction per endpoint basis.

>
> > Hub will generate bit stuffing errors on the
> > full-speed bus.
> which platform?

I remember it was mt8173.

And for bit stuffing errors I mentioned in the earlier mail.
when I read the spec again, 11.21 mentions that bit stuffing error
is generated when _a microframe_ should be passed without
corresponding SSPLIT-mid/end. So this is not the case and also
I'm not sure what will happen on the full-speed bus, sorry.
In my case what I can be sure of is that the audio output was
broken with those allotments.

What is the xhci-mtk's policy when there are more than two EPs
marked as the same u-frame offset like in the above example?

>
> >
> >
> > >
> > > Thanks
> > > >
> > > >
> > > > >                         if (tmp > FS_PAYLOAD_MAX)
> > > > >                                 return -ESCH_BW_OVERFLOW;
> > > > >                 }
> > > > > @@ -534,21 +534,18 @@ static void update_sch_tt(struct
> > > > > mu3h_sch_ep_info *sch_ep, bool used)
> > > > >  {
> > > > >         struct mu3h_sch_tt *tt = sch_ep->sch_tt;
> > > > >         u32 base, num_esit;
> > > > > -       int bw_updated;
> > > > >         int i, j;
> > > > >
> > > > >         num_esit = XHCI_MTK_MAX_ESIT / sch_ep->esit;
> > > > >
> > > > > -       if (used)
> > > > > -               bw_updated = sch_ep->bw_cost_per_microframe;
> > > > > -       else
> > > > > -               bw_updated = -sch_ep->bw_cost_per_microframe;
> > > > > -
> > > > >         for (i = 0; i < num_esit; i++) {
> > > > >                 base = sch_ep->offset + i * sch_ep->esit;
> > > > >
> > > > > -               for (j = 0; j < sch_ep->cs_count; j++)
> > > > > -                       tt->fs_bus_bw[base + j] += bw_updated;
> > > > > +               for (j = 0; j < sch_ep->num_budget_microframes;
> > > > > j++)
> > > > > +                       if (used)
> > > > > +                               tt->fs_bus_bw[base + j] +=
> > > > > sch_ep-
> > > > > > bw_budget_table[j];
> > > > >
> > > > > +                       else
> > > > > +                               tt->fs_bus_bw[base + j] -=
> > > > > sch_ep-
> > > > > > bw_budget_table[j];
> > > > >
> > > > >         }
> > > > >
> > > > >         if (used)
> > > > > --
> > > > > 2.18.0
> > > > >

  reply	other threads:[~2021-08-30  3:49 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-08-26  2:51 [PATCH next v2 1/6] Revert "usb: xhci-mtk: relax TT periodic bandwidth allocation" Chunfeng Yun
2021-08-26  2:51 ` [PATCH next v2 2/6] Revert "usb: xhci-mtk: Do not use xhci's virt_dev in drop_endpoint" Chunfeng Yun
2021-08-26  2:51 ` [PATCH next v2 3/6] usb: xhci-mtk: update fs bus bandwidth by bw_budget_table Chunfeng Yun
2021-08-26 11:54   ` Ikjoon Jang
2021-08-27  6:49     ` Chunfeng Yun (云春峰)
2021-08-27  9:14       ` Ikjoon Jang
2021-08-27  9:48         ` Chunfeng Yun (云春峰)
2021-08-30  3:49           ` Ikjoon Jang [this message]
2021-08-31  6:50             ` Chunfeng Yun (云春峰)
2021-08-26  2:51 ` [PATCH next v2 4/6] usb: xhci-mtk: add a member of num_esit Chunfeng Yun
2021-08-26  2:51 ` [PATCH next v2 5/6] usb: xhci-mtk: Do not use xhci's virt_dev in drop_endpoint Chunfeng Yun
2021-08-26  2:51 ` [PATCH next v2 6/6] usb: xhci-mtk: allow bandwidth table rollover Chunfeng Yun
2021-08-26 11:41 ` [PATCH next v2 1/6] Revert "usb: xhci-mtk: relax TT periodic bandwidth allocation" Greg Kroah-Hartman
2021-08-27  3:23   ` Chunfeng Yun (云春峰)
2021-08-27  9:46   ` Chunfeng Yun (云春峰)

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=CAATdQgDKuO+bM7_tMy59ghhhjrgCp53kHgYbgEciriHNF3uKuA@mail.gmail.com \
    --to=ikjn@chromium.org \
    --cc=Chunfeng.Yun@mediatek.com \
    --cc=Eddie.Hung@mediatek.com \
    --cc=Yaqii.Wu@mediatek.com \
    --cc=gregkh@linuxfoundation.org \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mediatek@lists.infradead.org \
    --cc=linux-usb@vger.kernel.org \
    --cc=mathias.nyman@intel.com \
    --cc=matthias.bgg@gmail.com \
    --subject='Re: [PATCH next v2 3/6] usb: xhci-mtk: update fs bus bandwidth by bw_budget_table' \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).