From: Thinh Nguyen <Thinh.Nguyen@synopsys.com>
To: Thinh Nguyen <Thinh.Nguyen@synopsys.com>
Cc: JaeHun Jung <jh0801.jung@samsung.com>,
Felipe Balbi <balbi@kernel.org>,
Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
"open list:USB XHCI DRIVER" <linux-usb@vger.kernel.org>,
open list <linux-kernel@vger.kernel.org>,
Seungchull Suh <sc.suh@samsung.com>,
Daehwan Jung <dh10.jung@samsung.com>
Subject: Re: [PATCH] usb: dwc3: Clear DWC3_EVENT_PENDING when count is 0
Date: Mon, 9 Jan 2023 19:09:18 +0000 [thread overview]
Message-ID: <20230109190914.3blihjfjdcszazdd@synopsys.com> (raw)
In-Reply-To: <20230109183545.kqej7vmtf5o6ihq4@synopsys.com>
On Mon, Jan 09, 2023, Thinh Nguyen wrote:
> Hi,
>
> On Mon, Jan 02, 2023, JaeHun Jung wrote:
> > Sometimes very rarely, The count is 0 and the DWC3 flag is set has status.
> > It must not have these status. Because, It can make happen interrupt storming
> > status.
> > So, It have to clean up DWC3_EVENT_PENDING flags set when count is 0.
> > It means "There are no interrupts to handle.".
>
> Can you reword to include the explanation from my response [*] and add a
> fixes tag to 7441b273388b ("usb: dwc3: gadget: Fix event pending check")
>
Actually, the old problem that the commit 7441b273388b fixes may
resurface because of this change.
Can you try this instead?
diff --git a/drivers/usb/dwc3/gadget.c b/drivers/usb/dwc3/gadget.c
index 65500246323b..3c36dfdb88f0 100644
--- a/drivers/usb/dwc3/gadget.c
+++ b/drivers/usb/dwc3/gadget.c
@@ -5515,8 +5515,15 @@ static irqreturn_t dwc3_check_event_buf(struct dwc3_event_buffer *evt)
* irq event handler completes before caching new event to prevent
* losing events.
*/
- if (evt->flags & DWC3_EVENT_PENDING)
+ if (evt->flags & DWC3_EVENT_PENDING) {
+ if (!evt->count) {
+ u32 reg = dwc3_readl(dwc->regs, DWC3_GEVNTSIZ(0));
+
+ if (!(reg & DWC3_GEVNTSIZ_INTMASK))
+ evt->flags &= ~DWC3_EVENT_PENDING;
+ }
return IRQ_HANDLED;
+ }
This is a quick solution at the moment. Let me know if you have have a
better option. Note that reading a register may take some time, albeit
short. That's why we may want to keep the evt->count check to help with
performance.
Thanks,
Thinh
>
>
> >
> > (struct dwc3_event_buffer *) ev_buf = 0xFFFFFF883DBF1180 (
> > (void *) buf = 0xFFFFFFC00DBDD000 = end+0x337D000,
> > (void *) cache = 0xFFFFFF8839F54080,
> > (unsigned int) length = 0x1000,
> > (unsigned int) lpos = 0x0,
> > (unsigned int) count = 0x0,
> > (unsigned int) flags = 0x00000001,
> > (dma_addr_t) dma = 0x00000008BD7D7000,
> > (struct dwc3 *) dwc = 0xFFFFFF8839CBC880,
> > (u64) android_kabi_reserved1 = 0x0),
> >
> > (time = 47557628930999, irq = 165, fn = dwc3_interrupt, latency = 0, en = 1),
> > (time = 47557628931268, irq = 165, fn = dwc3_interrupt, latency = 0, en = 3),
> > (time = 47557628932383, irq = 165, fn = dwc3_interrupt, latency = 0, en = 1),
> > (time = 47557628932652, irq = 165, fn = dwc3_interrupt, latency = 0, en = 3),
> > (time = 47557628933768, irq = 165, fn = dwc3_interrupt, latency = 0, en = 1),
> > (time = 47557628934037, irq = 165, fn = dwc3_interrupt, latency = 0, en = 3),
> > (time = 47557628935152, irq = 165, fn = dwc3_interrupt, latency = 0, en = 1),
> > (time = 47557628935460, irq = 165, fn = dwc3_interrupt, latency = 0, en = 3),
> > (time = 47557628936575, irq = 165, fn = dwc3_interrupt, latency = 0, en = 1),
> > (time = 47557628936845, irq = 165, fn = dwc3_interrupt, latency = 0, en = 3),
> > (time = 47557628937960, irq = 165, fn = dwc3_interrupt, latency = 0, en = 1),
> > (time = 47557628938229, irq = 165, fn = dwc3_interrupt, latency = 0, en = 3),
> > (time = 47557628939345, irq = 165, fn = dwc3_interrupt, latency = 0, en = 1),
> > (time = 47557628939652, irq = 165, fn = dwc3_interrupt, latency = 0, en = 3),
> > (time = 47557628940768, irq = 165, fn = dwc3_interrupt, latency = 0, en = 1),
> > (time = 47557628941037, irq = 165, fn = dwc3_interrupt, latency = 0, en = 3),
> > (time = 47557628942152, irq = 165, fn = dwc3_interrupt, latency = 0, en = 1),
> > (time = 47557628942422, irq = 165, fn = dwc3_interrupt, latency = 0, en = 3),
> > (time = 47557628943537, irq = 165, fn = dwc3_interrupt, latency = 0, en = 1),
> > (time = 47557628943806, irq = 165, fn = dwc3_interrupt, latency = 0, en = 3),
> > (time = 47557628944922, irq = 165, fn = dwc3_interrupt, latency = 0, en = 1),
> > (time = 47557628945229, irq = 165, fn = dwc3_interrupt, latency = 0, en = 3),
> > (time = 47557628946345, irq = 165, fn = dwc3_interrupt, latency = 0, en = 1),
> > (time = 47557628946614, irq = 165, fn = dwc3_interrupt, latency = 0, en = 3),
> > (time = 47557628947729, irq = 165, fn = dwc3_interrupt, latency = 0, en = 1),
> > (time = 47557628947999, irq = 165, fn = dwc3_interrupt, latency = 0, en = 3),
> >
> > Signed-off-by: JaeHun Jung <jh0801.jung@samsung.com>
> > ---
> > drivers/usb/dwc3/gadget.c | 5 ++++-
> > 1 file changed, 4 insertions(+), 1 deletion(-)
> >
> > diff --git a/drivers/usb/dwc3/gadget.c b/drivers/usb/dwc3/gadget.c
> > index 789976567f9f..5d2d5a9b9915 100644
> > --- a/drivers/usb/dwc3/gadget.c
> > +++ b/drivers/usb/dwc3/gadget.c
> > @@ -4355,8 +4355,11 @@ static irqreturn_t dwc3_check_event_buf(struct dwc3_event_buffer *evt)
> > * irq event handler completes before caching new event to prevent
> > * losing events.
> > */
> > - if (evt->flags & DWC3_EVENT_PENDING)
> > + if (evt->flags & DWC3_EVENT_PENDING) {
> > + if (!evt->count)
> > + evt->flags &= ~DWC3_EVENT_PENDING;
> > return IRQ_HANDLED;
> > + }
> >
> > count = dwc3_readl(dwc->regs, DWC3_GEVNTCOUNT(0));
> > count &= DWC3_GEVNTCOUNT_MASK;
> > --
> > 2.31.1
> >
prev parent reply other threads:[~2023-01-09 19:10 UTC|newest]
Thread overview: 22+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <CGME20230102050839epcas2p4b9d09d926f9a14c3b8e8df2574d334c3@epcas2p4.samsung.com>
2023-01-02 5:08 ` [PATCH] usb: dwc3: Clear DWC3_EVENT_PENDING when count is 0 JaeHun Jung
2023-01-03 15:48 ` Felipe Balbi
2023-01-05 3:29 ` Linyu Yuan
2023-01-05 3:35 ` Linyu Yuan
2023-01-05 9:54 ` 정재훈
2023-01-06 3:13 ` Linyu Yuan
2023-01-09 18:28 ` Thinh Nguyen
2023-01-10 1:56 ` Linyu Yuan
2023-01-10 2:53 ` Thinh Nguyen
2023-01-10 3:05 ` Linyu Yuan
2023-01-10 3:13 ` Linyu Yuan
2023-01-10 7:38 ` Linyu Yuan
2023-01-11 0:00 ` Thinh Nguyen
2023-01-11 1:45 ` Linyu Yuan
2023-01-11 2:27 ` Thinh Nguyen
2023-01-31 6:38 ` Linyu Yuan
2023-02-01 18:57 ` Thinh Nguyen
2023-02-02 5:00 ` Linyu Yuan
2023-02-02 20:06 ` Thinh Nguyen
2023-02-03 2:18 ` Linyu Yuan
2023-01-09 18:35 ` Thinh Nguyen
2023-01-09 19:09 ` Thinh Nguyen [this message]
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=20230109190914.3blihjfjdcszazdd@synopsys.com \
--to=thinh.nguyen@synopsys.com \
--cc=balbi@kernel.org \
--cc=dh10.jung@samsung.com \
--cc=gregkh@linuxfoundation.org \
--cc=jh0801.jung@samsung.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-usb@vger.kernel.org \
--cc=sc.suh@samsung.com \
/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
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox