public inbox for linux-serial@vger.kernel.org
 help / color / mirror / Atom feed
From: Hugo Villeneuve <hugo@hugovil.com>
To: Tapio Reijonen <tapio.reijonen@vaisala.com>
Cc: Jiri Slaby <jirislaby@kernel.org>,
	Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	Alexander Shiyan <shc_work@mail.ru>,
	Hugo Villeneuve <hvilleneuve@dimonoff.com>,
	linux-kernel@vger.kernel.org, linux-serial@vger.kernel.org
Subject: Re: [PATCH] serial: max310x: improve interrupt handling
Date: Thu, 4 Sep 2025 12:36:42 -0400	[thread overview]
Message-ID: <20250904123642.c8dd7bd6bf693ccbb5f6989b@hugovil.com> (raw)
In-Reply-To: <e88f3c74-2ea0-4266-b5f7-62b87a1987c5@vaisala.com>

Hi,

On Thu, 4 Sep 2025 14:50:16 +0300
Tapio Reijonen <tapio.reijonen@vaisala.com> wrote:

> Hi,
> 
> On 9/4/25 10:53, Jiri Slaby wrote:
> > On 03. 09. 25, 11:23, Tapio Reijonen wrote:
> >> When there is a heavy load of receiving characters to all
> >> four UART's, the warning 'Hardware RX FIFO overrun' is
> >> sometimes detected.
> >> The current implementation always service first UART3 until
> >> no more interrupt and then service another UARTs.
> >>
> >> This commit improve interrupt service routine to handle all
> >> interrupt sources, e.g. UARTs when a global IRQ is detected.
> >>
> >> Signed-off-by: Tapio Reijonen <tapio.reijonen@vaisala.com>
> >> ---
> >>   drivers/tty/serial/max310x.c | 21 ++++++++++++++++-----
> >>   1 file changed, 16 insertions(+), 5 deletions(-)
> >>
> >> diff --git a/drivers/tty/serial/max310x.c b/drivers/tty/serial/max310x.c
> >> index 
> >> ce260e9949c3c268e706b2615d6fc01adc21e49b..3234ed7c688ff423d25a007ed8b938b249ae0b82 100644
> >> --- a/drivers/tty/serial/max310x.c
> >> +++ b/drivers/tty/serial/max310x.c
> >> @@ -824,15 +824,26 @@ static irqreturn_t max310x_ist(int irq, void 
> >> *dev_id)
> >>       if (s->devtype->nr > 1) {
> >>           do {
> >> -            unsigned int val = ~0;
> >> +            unsigned int val;
> >> +            unsigned int global_irq = ~0;
> >> +            int port;
> >>               WARN_ON_ONCE(regmap_read(s->regmap,
> >> -                         MAX310X_GLOBALIRQ_REG, &val));
> >> -            val = ((1 << s->devtype->nr) - 1) & ~val;
> >> +                MAX310X_GLOBALIRQ_REG, &global_irq));
> >> +
> >> +            val = ((1 << s->devtype->nr) - 1) & ~global_irq;
> > 
> > This is horrid. Use GENMASK() (or BIT() below) instead. Likely, you want 
> > a local var storing the mask (the first part before the &).
> > 
> val = GENMASK(s->devtype->nr - 1, 0) & ~global_irq;>>               if 
> (!val)
> >>                   break;
> >> -            if (max310x_port_irq(s, fls(val) - 1) == IRQ_HANDLED)
> >> -                handled = true;
> >> +
> >> +            do {
> >> +                port = fls(val) - 1;
> >> +                if (max310x_port_irq(s, port) == IRQ_HANDLED)
> >> +                    handled = true;
> >> +
> >> +                global_irq |= 1 << port;
> >> +                val = ((1 << s->devtype->nr) - 1) & ~global_irq;
> >> +            } while (val);
> > 
> > Actually, does it have to be from the end? I am thinking of 
> > for_each_and_bit()...
> > 
> >>           } while (1);
> >>       } else {
> >>           if (max310x_port_irq(s, 0) == IRQ_HANDLED)

Combining what I suggested earlier, with Jiri's comments but using 
for_each_clear_bit() allow to get rid of the original horrid mask:

---
        if (s->devtype->nr > 1) {
                do {
                        unsigned int val = ~0;
+                       unsigned long channel;
+                       unsigned long irq;
+                       bool read_reg_done = true;
 
                        WARN_ON_ONCE(regmap_read(s->regmap,
                                                 MAX310X_GLOBALIRQ_REG, &val));
-                       val = ((1 << s->devtype->nr) - 1) & ~val;
-                       if (!val)
+                       irq = val;
+
+                       for_each_clear_bit(channel, &irq, s->devtype->nr) {
+                               read_reg_done = false;
+
+                               if (max310x_port_irq(s, channel) == IRQ_HANDLED)
+                                       handled = true;
+                       }
+
+                       if (read_reg_done)
                                break;
-                       if (max310x_port_irq(s, fls(val) - 1) == IRQ_HANDLED)
-                               handled = true;
                } while (1);
---


-- 
Hugo Villeneuve

  reply	other threads:[~2025-09-04 16:37 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-09-03  9:23 [PATCH] serial: max310x: improve interrupt handling Tapio Reijonen
2025-09-03 15:37 ` Hugo Villeneuve
2025-09-04  7:23   ` Tapio Reijonen
2025-09-04  7:53 ` Jiri Slaby
2025-09-04 11:50   ` Tapio Reijonen
2025-09-04 16:36     ` Hugo Villeneuve [this message]
2025-09-05  5:11       ` Tapio Reijonen

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=20250904123642.c8dd7bd6bf693ccbb5f6989b@hugovil.com \
    --to=hugo@hugovil.com \
    --cc=gregkh@linuxfoundation.org \
    --cc=hvilleneuve@dimonoff.com \
    --cc=jirislaby@kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-serial@vger.kernel.org \
    --cc=shc_work@mail.ru \
    --cc=tapio.reijonen@vaisala.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