Skip to contentSkip to conversation

mos7840: fix chars_in_buffer() return value

PATCH In review Stas Sergeev 2 participants 4 messages Sep 24, 2016 – Sep 30, 2016

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 mos7840: fix chars_in_buffer() return value

The TIOCOUTQ ioctl calls chars_in_buffer(), and some apps depend on a correct behaviour of that.
mos7840 implements it wrongly: if you write just one char, TIOCOUTQ will return 32.
This patch should fix it by accounting the number of chars actually written.
This patch, unfortunately, misses the Tested-by tag.
The reporter didn't test it, and I don't have the hardware in question.

---
 drivers/usb/serial/mos7840.c | 9 ++++-----
 1 file changed, 4 insertions(+), 5 deletions(-)
drivers/usb/serial/mos7840.c +4 -5
index ed378fb..2a1eb07 100644
@@ -234,7 +234,7 @@ struct moschip_port {
 
 	spinlock_t pool_lock;
 	struct urb *write_urb_pool[NUM_URBS];
-	char busy[NUM_URBS];
+	int busy[NUM_URBS];
 	bool read_urb_busy;
 
 	/* For device(s) with LED indicator */
@@ -1139,8 +1139,7 @@ static int mos7840_chars_in_buffer(struct tty_struct *tty)
 	spin_lock_irqsave(&mos7840_port->pool_lock, flags);
 	for (i = 0; i < NUM_URBS; ++i) {
 		if (mos7840_port->busy[i]) {
-			struct urb *urb = mos7840_port->write_urb_pool[i];
-			chars += urb->transfer_buffer_length;
+			chars += mos7840_port->busy[i];
 		}
 	}
 	spin_unlock_irqrestore(&mos7840_port->pool_lock, flags);
@@ -1323,10 +1322,11 @@ static int mos7840_write(struct tty_struct *tty, struct usb_serial_port *port,
 	/* try to find a free urb in the list */
 	urb = NULL;
 
+	transfer_size = min(count, URB_TRANSFER_BUFFER_SIZE);
 	spin_lock_irqsave(&mos7840_port->pool_lock, flags);
 	for (i = 0; i < NUM_URBS; ++i) {
 		if (!mos7840_port->busy[i]) {
-			mos7840_port->busy[i] = 1;
+			mos7840_port->busy[i] = transfer_size;
 			urb = mos7840_port->write_urb_pool[i];
 			dev_dbg(&port->dev, "URB:%d\n", i);
 			break;
@@ -1345,7 +1345,6 @@ static int mos7840_write(struct tty_struct *tty, struct usb_serial_port *port,
 		if (!urb->transfer_buffer)
 			goto exit;
 	}
-	transfer_size = min(count, URB_TRANSFER_BUFFER_SIZE);
 
 	memcpy(urb->transfer_buffer, current_position, transfer_size);
 
Signature · 1 line
2.7.4
maintainer
#1 lore raw
Quoted 8 lines · Stas Sergeev

The TIOCOUTQ ioctl calls chars_in_buffer(), and some apps depend on a correct behaviour of that.
mos7840 implements it wrongly: if you write just one char, TIOCOUTQ will return 32.
This patch should fix it by accounting the number of chars actually written.
This patch, unfortunately, misses the Tested-by tag.
The reporter didn't test it, and I don't have the hardware in question.

Why do you think the driver returns 32b in chars_in_buffer after writing a single character?

As far as I can see, the driver correctly sums up the outstanding bytes in its queue (represented by the submitted URBs). There is a tiny window where the count may be a little off due to the write() implementation claiming the URB before initialising it, but I'm not sure anyone cares.

But specifically, if you write just one character, TIOCOUTQ does indeed return a correct count.

Also note that your patch would break the driver in case a line discipline forwards an empty write request.

Thanks,
Johan

#2 lore raw

29.09.2016 13:09, Johan Hovold пишет:

Quoted 11 lines · Stas Sergeev
Quoted 8 lines

The TIOCOUTQ ioctl calls chars_in_buffer(), and some apps depend on a correct behaviour of that.
mos7840 implements it wrongly: if you write just one char, TIOCOUTQ will return 32.
This patch should fix it by accounting the number of chars actually written.
This patch, unfortunately, misses the Tested-by tag.
The reporter didn't test it, and I don't have the hardware in question.

Why do you think the driver returns 32b in chars_in_buffer after writing a single character?

Hi Johan, this actually came from this ancient bug report:
https://bugzilla.kernel.org/show_bug.cgi?id=45791
I was trying to add you back then and now, but your e-mail doesn't seem to be registered in a bug tracker. The code in question was different when I submitted the original patch:
https://bugzilla.kernel.org/attachment.cgi?id=77241
Maybe the bug was since fixed.
In that case sorry for the noise.

maintainer
#3 lore raw
Quoted 11 lines · Stas Sergeev

29.09.2016 13:09, Johan Hovold пишет:

Quoted 9 lines · Stas Sergeev
Quoted 8 lines

The TIOCOUTQ ioctl calls chars_in_buffer(), and some apps depend on a correct behaviour of that.
mos7840 implements it wrongly: if you write just one char, TIOCOUTQ will return 32.
This patch should fix it by accounting the number of chars actually written.
This patch, unfortunately, misses the Tested-by tag.
The reporter didn't test it, and I don't have the hardware in question.

Quoted 7 lines
Quoted 2 lines

Why do you think the driver returns 32b in chars_in_buffer after writing a single character?

Hi Johan, this actually came from this ancient bug report:
https://bugzilla.kernel.org/show_bug.cgi?id=45791
I was trying to add you back then and now, but your e-mail doesn't seem to be registered in a bug tracker.

Yeah, Greg does good job at direction bug reports to the usb mailing list.

Quoted 4 lines

The code in question was different when I submitted the original patch:
https://bugzilla.kernel.org/attachment.cgi?id=77241
Maybe the bug was since fixed.

It does seem to have been addressed by commit 5c263b92f828 ("usb:
serial: mos7840: Fixup mos7840_chars_in_buffer()") around the time of your original report in 2012.

Quoted 1 line

In that case sorry for the noise.

No worries.

Thanks,
Johan