linux-kernel - Re: [PATCH v3 4/5] serial: 8250: Handle implementations not having TEMT interrupt using em485

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Mon, 18 May 2020 18:19:28 +0300
From:   Andy Shevchenko <andriy.shevchenko@...ux.intel.com>
To:     Heiko Stuebner <heiko@...ech.de>
Cc:     gregkh@...uxfoundation.org, jslaby@...e.com,
        matwey.kornilov@...il.com, giulio.benetti@...ronovasrl.com,
        lukas@...ner.de, linux-serial@...r.kernel.org,
        linux-kernel@...r.kernel.org,
        christoph.muellner@...obroma-systems.com,
        Heiko Stuebner <heiko.stuebner@...obroma-systems.com>
Subject: Re: [PATCH v3 4/5] serial: 8250: Handle implementations not having
 TEMT interrupt using em485

On Sun, May 17, 2020 at 11:56:09PM +0200, Heiko Stuebner wrote:
> From: Giulio Benetti <giulio.benetti@...ronovasrl.com>
> 
> Some 8250 ports have a TEMT interrupt but it's not a part of the 8250
> standard, instead only available on some implementations.
> 
> The current em485 implementation does not work on ports without it.
> The only chance to make it work is to loop-read on LSR register.
> 
> So add UART_CAP_TEMT to mark 8250 uarts having this interrupt,
> update all current em485 users with that capability and add
> a loop-reading during  __stop_tx_rs485() on uarts not having it.
> 
> As __stop_tx_rs485() can also be called from a hard-irq context the
> loop-reading is split. If the fifo clears in under 100us in
> __stop_tx_rs485() itself just the regular stop calls get executed.
> If it takes longer, re-use the existing stop-timer infrastructure
> but with only a 10us timer to again poll the LSR registers.
> 
> Signed-off-by: Giulio Benetti <giulio.benetti@...ronovasrl.com>

> [moved to use added UART_CAP_TEMT, use readx_poll_timeout]

I can't parse first part...

Also, shouldn't it be rather like
  [heiko: ...]
?

...

> +static inline int __get_lsr(struct uart_8250_port *p)
> +{
> +	return serial_in(p, UART_LSR);
> +}
> +
> +static inline int __wait_for_empty(struct uart_8250_port *p, u64 timeout_us)
> +{
> +	int lsr;
> +
> +	return readx_poll_timeout(__get_lsr, p, lsr,
> +				  (lsr & BOTH_EMPTY) == BOTH_EMPTY,
> +				  0, timeout_us);
> +}

...

> +			int ret = __wait_for_empty(p, 100);

Do you expect something different than 100? If no, perhaps for now just put it
inside the function as a constant?

> +			if (ret < 0) {
> +				restart = HRTIMER_RESTART;
> +				goto out;
> +			}

...

> +	} else if (!(p->capabilities & UART_CAP_TEMT) &&
> +		   __wait_for_empty(p, 100)) {

I would leave it on one line even if you leave 100 as a parameter, but it's up to you.

...

> +		if (p->capabilities & UART_CAP_TEMT) {
> +			if ((lsr & BOTH_EMPTY) != BOTH_EMPTY)
> +				return;
> +		}

if (a) {
	if (b) {
		...
	}
}

is equivalent to 

if (a && b) {
	...
}

But it's up to you which one to choose.

-- 
With Best Regards,
Andy Shevchenko