Skip to contentSkip to conversation

more about serial ports: do they even work?

PATCH New Michael Tokarev 5 participants 16 messages Jan 26, 2009 – Feb 5, 2009 KVM

lore
Share & tools
lore
#0 lore raw

After some debugging and debugging, with a help Hollis Blanchard on #kvm@freenode, I discovered that kvm (or, rather, qemu) does not work correctly with serial ports, at least on linux. One problem report has already here, author Cc'd -- see e.g.
http://marc.info/?l=kvm&m=122995568009533&w=2 .

Here's what's going on.

When opening a host's port, kvm resets the status
lines, doing this:

ioctl(13, TIOCMGET, [TIOCM_DTR|TIOCM_RTS|TIOCM_CTS|TIOCM_DSR|0x4000]) ioctl(13, TIOCMSET, [TIOCM_DTR|TIOCM_RTS])

which results in the following set

ioctl(13, TIOCMGET, [TIOCM_DTR|TIOCM_RTS|TIOCM_CTS|TIOCM_DSR])

Note the difference between the default set and new one: the missing bit, 0x4000, which is unknown to strace(1) but is defined as TIOCM_OUT2 in linux headers.

After that change (resetting the TIOCM_OUT2 bit), no writes to the serial port works anymore, they're all gets accepted by host kernel and are buffered in the kernel.

After some time, when the kernel buffer fills up, and since the port (on host side) is opened in non-blocking mode, the writes starts returning EAGAIN, and kvm process starts endless loop, which were seen by David.

Here's the trivial program to demonstrate the idea:

---- cut ---
#include <sys/types.h>
#include <unistd.h>
#include <fcntl.h>
#include <sys/ioctl.h>
#include <termios.h>

int main(int argc, char **argv) {
fd = open("/dev/ttyS0", O_RDWR|O_NONBLOCK); fcntl(fd, F_SETFL, O_RDWR);
x = TIOCM_DTR|TIOCM_RTS|TIOCM_CTS|TIOCM_DSR // |0x4000 ;
ioctl(fd, TIOCMSET, &x);
ioctl(fd, TIOCMGET, &x);
write(fd, "at\r", 3);
read(fd, buf, 20);
close(fd);

return 0;
}
--- cut ---

Run it under strace while a dialup modem is connected to /dev/ttyS0 (i used this way for testing). It will stuck at read, and nothing will be written, even when write() will happily return 3. Un-comment the |0x4000 thing, and it will work.

I'm not sure what should be done with this, and how much this is linux-specific. But it is obvious that bit (TIOCM_OUT2) should be left in-place (after which the thing works), at least on linux.

Note that this bit is NOT shown in /proc/tty/driver/serial file (which shows other bits).

Note also that this file (/proc/tty/driver/serial) helps to see if any write were performed: compare the counters. In 'tx' there's number of bytes actually sent to device, as opposed to accepted by the kernel. When you write something to /dev/ttyS0, that number increases, IF that something actually reached the device.

Thanks.

/mjt

#1 lore raw

Michael Tokarev ?????:

Quoted 12 lines

After some debugging and debugging, with a help Hollis Blanchard on #kvm@freenode, I discovered that kvm (or, rather, qemu) does not work correctly with serial ports, at least on linux. One problem report has already here, author Cc'd -- see e.g.
http://marc.info/?l=kvm&m=122995568009533&w=2 .

Here's what's going on.

When opening a host's port, kvm resets the status lines, doing this:

Quoted 6 lines

ioctl(13, TIOCMGET, [TIOCM_DTR|TIOCM_RTS|TIOCM_CTS|TIOCM_DSR|0x4000]) ioctl(13, TIOCMSET, [TIOCM_DTR|TIOCM_RTS])

which results in the following set

ioctl(13, TIOCMGET, [TIOCM_DTR|TIOCM_RTS|TIOCM_CTS|TIOCM_DSR])

Here's the possible solution (NotAPAtch(tm)):

In kvm-xx/qemu/qemu-char.c:

    case CHR_IOCTL_SERIAL_SET_TIOCM:
        {
            int sarg = *(int *)arg;
            int targ = 0;        <==== change this 0 to 0x4000
            if (sarg | CHR_TIOCM_DTR)
                targ |= TIOCM_DTR;
            if (sarg | CHR_TIOCM_RTS)
                targ |= TIOCM_RTS;
            ioctl(s->fd_in, TIOCMSET, &targ);
        }
        break;

This is obviously a hack, esp. since this bit is not always present even on linux (after reading 8250.c driver).

Real fix will be, I guess, to read the full set first, and combine it with DTR|RTS received from guest, something like this:

    case CHR_IOCTL_SERIAL_SET_TIOCM:
        {
            int sarg = *(int *)arg;
            int targ = 0;
            ioctl(s->fd_in, TIOCMGET, &targ);
            if (!(sarg | CHR_TIOCM_DTR))
                targ &= ~TIOCM_DTR;
            if (!(sarg | CHR_TIOCM_RTS))
                targ ~= ~TIOCM_RTS;
            ioctl(s->fd_in, TIOCMSET, &targ);
        }
        break;

I.e., to always keep all the other bits, but allow changing DTR and RTS.

Again, I don't know how it's linux-specific, but it seems the solution above should work on other platforms just fine.

/mjt

#2 lore raw

I don't recall seeing a followup to this post.

To put Michael's second suggestion into patch form, the following fixes the problem for me:

qemu/qemu-char.c 2009-02-02 14:12:20.000000000 -0700 renamed from qemu/qemu-char.c 2008-12-14 06:16:27.000000000 -0700 +1 -0
@@ -1078,20 +1078,21 @@
             if (sarg | TIOCM_DTR)
                 *targ |= CHR_TIOCM_DTR;
             if (sarg | TIOCM_RTS)
                 *targ |= CHR_TIOCM_RTS;
         }
         break;
     case CHR_IOCTL_SERIAL_SET_TIOCM:
         {
             int sarg = *(int *)arg;
             int targ = 0;
+            ioctl(s->fd_in, TIOCMGET, &targ);
             if (sarg | CHR_TIOCM_DTR)
                 targ |= TIOCM_DTR;
             if (sarg | CHR_TIOCM_RTS)
                 targ |= TIOCM_RTS;
             ioctl(s->fd_in, TIOCMSET, &targ);
         }
         break;
     default:
         return -ENOTSUP;
     }

Is this approach palatable to folks?

david

Quoted 65 lines · Michael Tokarev

Michael Tokarev ?????:

Quoted 12 lines

After some debugging and debugging, with a help Hollis Blanchard on #kvm@freenode, I discovered that kvm (or, rather, qemu) does not work correctly with serial ports, at least on linux. One problem report has already here, author Cc'd -- see e.g.
http://marc.info/?l=kvm&m=122995568009533&w=2 .

Here's what's going on.

When opening a host's port, kvm resets the status lines, doing this:

Quoted 6 lines

ioctl(13, TIOCMGET, [TIOCM_DTR|TIOCM_RTS|TIOCM_CTS|TIOCM_DSR|0x4000]) ioctl(13, TIOCMSET, [TIOCM_DTR|TIOCM_RTS])

which results in the following set

ioctl(13, TIOCMGET, [TIOCM_DTR|TIOCM_RTS|TIOCM_CTS|TIOCM_DSR])

Here's the possible solution (NotAPAtch(tm)):

In kvm-xx/qemu/qemu-char.c:

    case CHR_IOCTL_SERIAL_SET_TIOCM:
        {
            int sarg = *(int *)arg;
            int targ = 0;        <==== change this 0 to 0x4000
            if (sarg | CHR_TIOCM_DTR)
                targ |= TIOCM_DTR;
            if (sarg | CHR_TIOCM_RTS)
                targ |= TIOCM_RTS;
            ioctl(s->fd_in, TIOCMSET, &targ);
        }
        break;

This is obviously a hack, esp. since this bit is not always present even on linux (after reading 8250.c driver).

Real fix will be, I guess, to read the full set first, and combine it with DTR|RTS received from guest, something like this:

    case
        {
            int sarg = *(int *)arg;
            int targ = 0;
            ioctl(s->fd_in, TIOCMGET, &targ);
            if (!(sarg | CHR_TIOCM_DTR))
                targ &= ~TIOCM_DTR;
            if (!(sarg | CHR_TIOCM_RTS))
                targ ~= ~TIOCM_RTS;
            ioctl(s->fd_in, TIOCMSET, &targ);
        }
        break;

I.e., to always keep all the other bits, but allow changing DTR and RTS.

Again, I don't know how it's linux-specific, but it seems the solution above should work on other platforms just fine.

/mjt

#3 lore raw
Quoted 9 lines · David S. Ahern

I don't recall seeing a followup to this post.

To put Michael's second suggestion into patch form, the following fixes the problem for me:

qemu/qemu-char.c 2009-02-02 14:12:20.000000000 -0700 renamed from qemu/qemu-char.c 2008-12-14 06:16:27.000000000 -0700 +0 -0
@@ -1078,20 +1078,21 @@
             if (sarg | TIOCM_DTR)

qemu/qemu-char.c 2009-02-02 14:12:20.000000000 -0700 line 1088

Quoted 10 lines
qemu/qemu-char.c 2009-02-02 14:12:20.000000000 -0700 renamed from qemu/qemu-char.c 2008-12-14 06:16:27.000000000 -0700 +1 -0
                 *targ |= CHR_TIOCM_DTR;
             if (sarg | TIOCM_RTS)
                 *targ |= CHR_TIOCM_RTS;
         }
         break;
     case CHR_IOCTL_SERIAL_SET_TIOCM:
         {
             int sarg = *(int *)arg;
             int targ = 0;
+            ioctl(s->fd_in, TIOCMGET, &targ);

here, one more operation is necessary:

               targ &= ~(TIOCM_DTR|TIOCM_RTS);
Quoted 11 lines
qemu/qemu-char.c 2009-02-02 14:12:20.000000000 -0700 renamed from qemu/qemu-char.c 2008-12-14 06:16:27.000000000 -0700 +0 -0
             if (sarg | CHR_TIOCM_DTR)
                 targ |= TIOCM_DTR;
             if (sarg | CHR_TIOCM_RTS)
                 targ |= TIOCM_RTS;
             ioctl(s->fd_in, TIOCMSET, &targ);
         }
         break;
     default:
         return -ENOTSUP;
     }
Quoted 3 lines

Is this approach palatable to folks?

david

/mjt

#4 lore raw
Quoted 22 lines · Michael Tokarev
Quoted 9 lines · David S. Ahern

I don't recall seeing a followup to this post.

To put Michael's second suggestion into patch form, the following fixes the problem for me:

qemu/qemu-char.c 2009-02-02 14:12:20.000000000 -0700 renamed from qemu/qemu-char.c 2008-12-14 06:16:27.000000000 -0700 +0 -0
@@ -1078,20 +1078,21 @@
             if (sarg | TIOCM_DTR)

qemu/qemu-char.c 2009-02-02 14:12:20.000000000 -0700 line 1088

Quoted 10 lines
qemu/qemu-char.c 2009-02-02 14:12:20.000000000 -0700 renamed from qemu/qemu-char.c 2008-12-14 06:16:27.000000000 -0700 +1 -0
                 *targ |= CHR_TIOCM_DTR;
             if (sarg | TIOCM_RTS)
                 *targ |= CHR_TIOCM_RTS;
         }
         break;
     case CHR_IOCTL_SERIAL_SET_TIOCM:
         {
             int sarg = *(int *)arg;
             int targ = 0;
+            ioctl(s->fd_in, TIOCMGET, &targ);

here, one more operation is necessary:

Quoted 2 lines
               targ &= ~(TIOCM_DTR|TIOCM_RTS);

Interesting. that change was not needed to fix my problem.

david

Quoted 16 lines
Quoted 11 lines
qemu/qemu-char.c 2009-02-02 14:12:20.000000000 -0700 renamed from qemu/qemu-char.c 2008-12-14 06:16:27.000000000 -0700 +0 -0
             if (sarg | CHR_TIOCM_DTR)
                 targ |= TIOCM_DTR;
             if (sarg | CHR_TIOCM_RTS)
                 targ |= TIOCM_RTS;
             ioctl(s->fd_in, TIOCMSET, &targ);
         }
         break;
     default:
         return -ENOTSUP;
     }
Quoted 3 lines

Is this approach palatable to folks?

david

/mjt

#5 lore raw
Quoted 24 lines · David S. Ahern
Quoted 21 lines · Michael Tokarev
Quoted 9 lines · David S. Ahern

I don't recall seeing a followup to this post.

To put Michael's second suggestion into patch form, the following fixes the problem for me:

qemu/qemu-char.c 2009-02-02 14:12:20.000000000 -0700 renamed from qemu/qemu-char.c 2008-12-14 06:16:27.000000000 -0700 +0 -0
@@ -1078,20 +1078,21 @@
             if (sarg | TIOCM_DTR)

qemu/qemu-char.c 2009-02-02 14:12:20.000000000 -0700 line 1088

Quoted 10 lines
qemu/qemu-char.c 2009-02-02 14:12:20.000000000 -0700 renamed from qemu/qemu-char.c 2008-12-14 06:16:27.000000000 -0700 +1 -0
                 *targ |= CHR_TIOCM_DTR;
             if (sarg | TIOCM_RTS)
                 *targ |= CHR_TIOCM_RTS;
         }
         break;
     case CHR_IOCTL_SERIAL_SET_TIOCM:
         {
             int sarg = *(int *)arg;
             int targ = 0;
+            ioctl(s->fd_in, TIOCMGET, &targ);

here, one more operation is necessary:

Quoted 1 line
               targ &= ~(TIOCM_DTR|TIOCM_RTS);
Quoted 1 line

Interesting. that change was not needed to fix my problem.

It just means you (or, rather, your guests) never really needed to DROP those signal lines, only to raise them.

Quoted 4 lines
Quoted 4 lines

qemu/qemu-char.c 2009-02-02 14:12:20.000000000 -0700 line 1092

Quoted 4 lines
qemu/qemu-char.c 2009-02-02 14:12:20.000000000 -0700 renamed from qemu/qemu-char.c 2008-12-14 06:16:27.000000000 -0700 +0 -0
             if (sarg | CHR_TIOCM_DTR)
                 targ |= TIOCM_DTR;
             if (sarg | CHR_TIOCM_RTS)
                 targ |= TIOCM_RTS;

Without that line above, the code never drops the two bits, once set they can't be "removed" anymore.

By the way, this is upstream qemu issue, not kvm one, and has to be pushed as such. Good you CC'd qemu list.

/mjt

#6 lore raw
Quoted 24 lines · Michael Tokarev
Quoted 23 lines · David S. Ahern
Quoted 22 lines · Michael Tokarev
Quoted 9 lines · David S. Ahern

I don't recall seeing a followup to this post.

To put Michael's second suggestion into patch form, the following fixes the problem for me:

qemu/qemu-char.c 2009-02-02 14:12:20.000000000 -0700 renamed from qemu/qemu-char.c 2008-12-14 06:16:27.000000000 -0700 +0 -0
@@ -1078,20 +1078,21 @@
             if (sarg | TIOCM_DTR)

qemu/qemu-char.c 2009-02-02 14:12:20.000000000 -0700 line 1088

Quoted 10 lines
qemu/qemu-char.c 2009-02-02 14:12:20.000000000 -0700 renamed from qemu/qemu-char.c 2008-12-14 06:16:27.000000000 -0700 +1 -0
                 *targ |= CHR_TIOCM_DTR;
             if (sarg | TIOCM_RTS)
                 *targ |= CHR_TIOCM_RTS;
         }
         break;
     case CHR_IOCTL_SERIAL_SET_TIOCM:
         {
             int sarg = *(int *)arg;
             int targ = 0;
+            ioctl(s->fd_in, TIOCMGET, &targ);

here, one more operation is necessary:

               targ &= ~(TIOCM_DTR|TIOCM_RTS);
Quoted 9 lines
Quoted 1 line

Interesting. that change was not needed to fix my problem.

It just means you (or, rather, your guests) never really needed to DROP those signal lines, only to raise them.

Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
qemu/qemu-char.c 2009-02-02 14:12:20.000000000 -0700 renamed from qemu/qemu-char.c 2008-12-14 06:16:27.000000000 -0700 +0 -0
             if (sarg | CHR_TIOCM_DTR)
                 targ |= TIOCM_DTR;
             if (sarg | CHR_TIOCM_RTS)
                 targ |= TIOCM_RTS;

Is this code really correct. If it is is there a comment somewhere describing why it's correct? I would have expected the lines above to be:

if (sarg & CHR_TIOCM_DTR)
targ |= TIOCM_DTR;
if (sarg & CHR_TIOCM_RTS)
targ |= TIOCM_RTS;

(Just an observation as this went past).

MM

Quoted 13 lines

Without that line above, the code never drops the two bits, once set they can't be "removed" anymore.

By the way, this is upstream qemu issue, not kvm one, and has to be pushed as such. Good you CC'd qemu list.

/mjt
--
To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

#7 lore raw
Quoted 38 lines · Mark Marshall
Quoted 24 lines · Michael Tokarev
Quoted 23 lines · David S. Ahern
Quoted 22 lines · Michael Tokarev
Quoted 9 lines · David S. Ahern

I don't recall seeing a followup to this post.

To put Michael's second suggestion into patch form, the following fixes the problem for me:

qemu/qemu-char.c 2009-02-02 14:12:20.000000000 -0700 renamed from qemu/qemu-char.c 2008-12-14 06:16:27.000000000 -0700 +0 -0
@@ -1078,20 +1078,21 @@
             if (sarg | TIOCM_DTR)

qemu/qemu-char.c 2009-02-02 14:12:20.000000000 -0700 line 1088

Quoted 10 lines
qemu/qemu-char.c 2009-02-02 14:12:20.000000000 -0700 renamed from qemu/qemu-char.c 2008-12-14 06:16:27.000000000 -0700 +1 -0
                 *targ |= CHR_TIOCM_DTR;
             if (sarg | TIOCM_RTS)
                 *targ |= CHR_TIOCM_RTS;
         }
         break;
     case CHR_IOCTL_SERIAL_SET_TIOCM:
         {
             int sarg = *(int *)arg;
             int targ = 0;
+            ioctl(s->fd_in, TIOCMGET, &targ);

here, one more operation is necessary:

               targ &= ~(TIOCM_DTR|TIOCM_RTS);
Quoted 9 lines
Quoted 1 line

Interesting. that change was not needed to fix my problem.

It just means you (or, rather, your guests) never really needed to DROP those signal lines, only to raise them.

Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
qemu/qemu-char.c 2009-02-02 14:12:20.000000000 -0700 renamed from qemu/qemu-char.c 2008-12-14 06:16:27.000000000 -0700 +0 -0
             if (sarg | CHR_TIOCM_DTR)
                 targ |= TIOCM_DTR;
             if (sarg | CHR_TIOCM_RTS)
                 targ |= TIOCM_RTS;

Is this code really correct. If it is is there a comment somewhere describing why it's correct? I would have expected the lines above to be:

Quoted 6 lines

if (sarg & CHR_TIOCM_DTR)
targ |= TIOCM_DTR;
if (sarg & CHR_TIOCM_RTS)
targ |= TIOCM_RTS;

(Just an observation as this went past).

Right. I don't understand the point of converting to an "internal" representation of TIOCM control bits.

CHR_IOCTL_SERIAL_SET_TIOCM is clearly broken as mentioned. It should at least preserve the bits it does not control.

qemu/qemu-char.c +2 -23
index ac431c7..66971e1 100644
@@ -1063,33 +1063,12 @@ static int tty_serial_ioctl(CharDriverState *chr, int cmd, void *arg)
         break;
     case CHR_IOCTL_SERIAL_GET_TIOCM:
         {
-            int sarg = 0;
-            int *targ = (int *)arg;
-            ioctl(s->fd_in, TIOCMGET, &sarg);
-            *targ = 0;
-            if (sarg | TIOCM_CTS)
-                *targ |= CHR_TIOCM_CTS;
-            if (sarg | TIOCM_CAR)
-                *targ |= CHR_TIOCM_CAR;
-            if (sarg | TIOCM_DSR)
-                *targ |= CHR_TIOCM_DSR;
-            if (sarg | TIOCM_RI)
-                *targ |= CHR_TIOCM_RI;
-            if (sarg | TIOCM_DTR)
-                *targ |= CHR_TIOCM_DTR;
-            if (sarg | TIOCM_RTS)
-                *targ |= CHR_TIOCM_RTS;
+            ioctl(s->fd_in, TIOCMGET, arg);
         }
         break;
     case CHR_IOCTL_SERIAL_SET_TIOCM:
         {
-            int sarg = *(int *)arg;
-            int targ = 0;
-            if (sarg | CHR_TIOCM_DTR)
-                targ |= TIOCM_DTR;
-            if (sarg | CHR_TIOCM_RTS)
-                targ |= TIOCM_RTS;
-            ioctl(s->fd_in, TIOCMSET, &targ);
+            ioctl(s->fd_in, TIOCMSET, arg);
         }
         break;
     default:
#8 lore raw
Quoted 1 line · Marcelo Tosatti

On Tue, Feb 03, 2009 at 08:41:14AM +0000, Mark Marshall wrote:

[]

Quoted 2 lines

Right. I don't understand the point of converting to an "internal" representation of TIOCM control bits.

<fun mode>
Well, the same goes for the IOCTL values themselves too -- like this CHR_IOCTL_SERIAL_SET_TIOCM itself. I mean, TIOCMSET is the right name for it ;) But see below.

Quoted 8 lines

CHR_IOCTL_SERIAL_SET_TIOCM is clearly broken as mentioned. It should at least preserve the bits it does not control.

qemu/qemu-char.c +0 -0
index ac431c7..66971e1 100644

qemu/qemu-char.c static int tty_serial_ioctl(CharDriverState *chr, int cmd, void *arg) · line 1068

Quoted 6 lines
qemu/qemu-char.c +1 -0
@@ -1063,33 +1063,12 @@ static int tty_serial_ioctl(CharDriverState *chr, int cmd, void *arg)
         break;
     case CHR_IOCTL_SERIAL_GET_TIOCM:
         {
+            ioctl(s->fd_in, TIOCMGET, arg);
         }
         break;

And those parens too, let it die, die! ;)
</fun mode>

Other than that, an.. excellent idea, I wanted to propose just that when I first saw all this stuff, but was somewhat afraid. And I *think* there's at least *some* sense. Qemu is a CPU emulator and may work on another arch where those bits are defined differently. Maybe that was the reason for all this converting - to be safe than sorry, so to say. No?

/mjt

#9 lore raw
Quoted 14 lines · Michael Tokarev
Quoted 1 line · Marcelo Tosatti

On Tue, Feb 03, 2009 at 08:41:14AM +0000, Mark Marshall wrote:

[]

Quoted 2 lines

Right. I don't understand the point of converting to an "internal" representation of TIOCM control bits.

<fun mode>
Well, the same goes for the IOCTL values themselves too -- like this CHR_IOCTL_SERIAL_SET_TIOCM itself. I mean, TIOCMSET is the right name for it ;) But see below.

Quoted 3 lines

CHR_IOCTL_SERIAL_SET_TIOCM is clearly broken as mentioned. It should at least preserve the bits it does not control.

Quoted 13 lines

qemu/qemu-char.c static int tty_serial_ioctl(CharDriverState *chr, int cmd, void *arg) · line 1068

Quoted 11 lines
qemu/qemu-char.c +1 -0
index ac431c7..66971e1 100644
@@ -1063,33 +1063,12 @@ static int tty_serial_ioctl(CharDriverState *chr, int cmd, void *arg)
         break;
     case CHR_IOCTL_SERIAL_GET_TIOCM:
         {
+            ioctl(s->fd_in, TIOCMGET, arg);
         }
         break;

And those parens too, let it die, die! ;)

Kill it!

Quoted 8 lines

</fun mode>

Other than that, an.. excellent idea, I wanted to propose just that when I first saw all this stuff, but was somewhat afraid. And I *think* there's at least *some* sense. Qemu is a CPU emulator and may work on another arch where those bits are defined differently. Maybe that was the reason for all this converting - to be safe than sorry, so to say. No?

Probably, yes.

Does it work for you?

#10 lore raw

Marcelo Tosatti wrote:
[]

Quoted 8 lines
Quoted 6 lines

Other than that, an.. excellent idea, I wanted to propose just that when I first saw all this stuff, but was somewhat afraid. And I *think* there's at least *some* sense. Qemu is a CPU emulator and may work on another arch where those bits are defined differently. Maybe that was the reason for all this converting - to be safe than sorry, so to say. No?

Probably, yes.

It's a question for qemu folks.

Quoted 1 line

Does it work for you?

Sure it does, since I only use it with kvm with the same arch in guest and host, even with the same operating system.

Wug. So much for so tiny issue.... Well, I'd not call non-working serial port a "tiny issue". Oh well.

/mjt

#11 lore raw
In reply to Michael Tokarev
Quoted 7 lines · Michael Tokarev

Other than that, an.. excellent idea, I wanted to propose just that when I first saw all this stuff, but was somewhat afraid. And I *think* there's at least *some* sense. Qemu is a CPU emulator and may work on another arch where those bits are defined differently. Maybe that was the reason for all this converting - to be safe than sorry, so to say. No?

Yes, this is exactly the reason why they were introduced in the first place. Let's suppose that the guest defines those constants differently: we need to parse them and covert them appropriately to the host format. CHR_IOCTL_SERIAL_SET_TIOCM and CHR_TIOCM_DTR correspond to the guest version of TIOCMSET and TIOCM_DTR and can be defined differently depending on the particular guest arch.

#12 lore raw
Quoted 10 lines · Stefano Stabellini
Quoted 7 lines · Michael Tokarev

Other than that, an.. excellent idea, I wanted to propose just that when I first saw all this stuff, but was somewhat afraid. And I *think* there's at least *some* sense. Qemu is a CPU emulator and may work on another arch where those bits are defined differently. Maybe that was the reason for all this converting - to be safe than sorry, so to say. No?

Quoted 6 lines

Yes, this is exactly the reason why they were introduced in the first place. Let's suppose that the guest defines those constants differently: we need to parse them and covert them appropriately to the host format. CHR_IOCTL_SERIAL_SET_TIOCM and CHR_TIOCM_DTR correspond to the guest version of TIOCMSET and TIOCM_DTR and can be defined differently depending on the particular guest arch.

The following works for me. It fixes the existing checks in place for the GET and replicates that for the SET. The ioctl initialization is needed in the SET is needed.

qemu-char.c (working copy) renamed from qemu-char.c (revision 6519) +19 -8
Index: trunk/qemu-char.c
===================================================================
@@ -1067,17 +1067,17 @@
             int *targ = (int *)arg;
             ioctl(s->fd_in, TIOCMGET, &sarg);
             *targ = 0;
-            if (sarg | TIOCM_CTS)
+            if (sarg & TIOCM_CTS)
                 *targ |= CHR_TIOCM_CTS;
-            if (sarg | TIOCM_CAR)
+            if (sarg & TIOCM_CAR)
                 *targ |= CHR_TIOCM_CAR;
-            if (sarg | TIOCM_DSR)
+            if (sarg & TIOCM_DSR)
                 *targ |= CHR_TIOCM_DSR;
-            if (sarg | TIOCM_RI)
+            if (sarg & TIOCM_RI)
                 *targ |= CHR_TIOCM_RI;
-            if (sarg | TIOCM_DTR)
+            if (sarg & TIOCM_DTR)
                 *targ |= CHR_TIOCM_DTR;
-            if (sarg | TIOCM_RTS)
+            if (sarg & TIOCM_RTS)
                 *targ |= CHR_TIOCM_RTS;
         }
         break;
@@ -1085,9 +1085,20 @@
         {
             int sarg = *(int *)arg;
             int targ = 0;
-            if (sarg | CHR_TIOCM_DTR)
+            ioctl(s->fd_in, TIOCMGET, &targ);
+            targ &= ~(CHR_TIOCM_CTS | CHR_TIOCM_CAR | CHR_TIOCM_DSR
+                     | CHR_TIOCM_RI | CHR_TIOCM_DTR | CHR_TIOCM_RTS);
+            if (sarg & CHR_TIOCM_CTS)
+                targ |= TIOCM_CTS;
+            if (sarg & CHR_TIOCM_CAR)
+                targ |= TIOCM_CAR;
+            if (sarg & CHR_TIOCM_DSR)
+                targ |= TIOCM_DSR;
+            if (sarg & CHR_TIOCM_RI)
+                targ |= TIOCM_RI;
+            if (sarg & CHR_TIOCM_DTR)
                 targ |= TIOCM_DTR;
-            if (sarg | CHR_TIOCM_RTS)
+            if (sarg & CHR_TIOCM_RTS)
                 targ |= TIOCM_RTS;
             ioctl(s->fd_in, TIOCMSET, &targ);
         }

david

#13 lore raw
Quoted 18 lines · David S. Ahern
Quoted 9 lines · Stefano Stabellini
Quoted 7 lines · Michael Tokarev

Other than that, an.. excellent idea, I wanted to propose just that when I first saw all this stuff, but was somewhat afraid. And I *think* there's at least *some* sense. Qemu is a CPU emulator and may work on another arch where those bits are defined differently. Maybe that was the reason for all this converting - to be safe than sorry, so to say. No?

Quoted 6 lines

Yes, this is exactly the reason why they were introduced in the first place. Let's suppose that the guest defines those constants differently: we need to parse them and covert them appropriately to the host format. CHR_IOCTL_SERIAL_SET_TIOCM and CHR_TIOCM_DTR correspond to the guest version of TIOCMSET and TIOCM_DTR and can be defined differently depending on the particular guest arch.

Quoted 3 lines

The following works for me. It fixes the existing checks in place for the GET and replicates that for the SET. The ioctl initialization is needed in the SET is needed.

It looks pretty good to me.

#14 lore raw
In reply to Michael Tokarev
Quoted 20 lines · Michael Tokarev
Quoted 10 lines · David S. Ahern
Quoted 8 lines · Michael Tokarev
Quoted 5 lines · David S. Ahern
     case CHR_IOCTL_SERIAL_SET_TIOCM:
         {
             int sarg = *(int *)arg;
             int targ = 0;

+ ioctl(s->fd_in, TIOCMGET, &targ);

here, one more operation is necessary:

               targ &= ~(TIOCM_DTR|TIOCM_RTS);

Interesting. that change was not needed to fix my problem.

It just means you (or, rather, your guests) never really needed to DROP those signal lines, only to raise them.

Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
             if (sarg | CHR_TIOCM_DTR)
                 targ |= TIOCM_DTR;
             if (sarg | CHR_TIOCM_RTS)
                 targ |= TIOCM_RTS;
Quoted 6 lines

Without that line above, the code never drops the two bits, once set they can't be "removed" anymore.

By the way, this is upstream qemu issue, not kvm one, and has to be pushed as such. Good you CC'd qemu list.

Either the two lines above or we could parse the whole set of possible flags, like we do in the CHR_IOCTL_SERIAL_GET_TIOCM case.

#15 lore raw

PATCH 1/1 : more about serial ports: do they even work?

In reply to Michael Tokarev
Quoted 6 lines · Michael Tokarev

After some debugging and debugging, with a help Hollis Blanchard on #kvm@freenode, I discovered that kvm (or, rather, qemu) does not work correctly with serial ports, at least on linux. One problem report has already here, author Cc'd -- see e.g.
http://marc.info/?l=kvm&m=122995568009533&w=2 .

... [quoted in full below]...

Ok, It's a real shame to see SO many wrong attempts to do it all, with so many idiotic mistakes all over... But c'est la vie, it seems... ;)

So here we go. Attached is a patch that fixes two problems with serial ports &qemu (yes it's a qemu issue, and, as far as I can see, both probs were here for a long time).

First is completely f*cked up flags reporting and setup for TIOCMGET and TIOCMSET ioctls, where ALL known flags were reported and set in case at least one flag is set, misusing "if(a|b) foo" instead of "if(a&b) foo" -- just a typo I assume, but heck of a typo... ;)

And second - for TIOCMSET it preserves other, unknown flags. Which fixes the problem that started it all, since there was a bit set internally in kernel which, when unset, makes serial port non-working, but TIOCMSET dropped all "other" bits on the floor.

And for this second one, I'm still unsure. The patch I'm sending only tries to remove TIOCM_DTR and _RTS bits (RTS is useless since it's controlled by the connected device, isn't it?), leaving all others, incl., say, CAR, RI, CTS etc, in place. The question is -- some of those bits are "input" lines, i.e., the ones controlled by the attached device, and I don't know if all platforms will ignore those instead of reporting error. I.e, maybe filter also those who are known "inputs"? And while we're at it, still, how about RTS?

Thanks!

/ashamed mjt

--- original content follows ---

Quoted 69 lines

Here's what's going on.

When opening a host's port, kvm resets the status
lines, doing this:

ioctl(13, TIOCMGET, [TIOCM_DTR|TIOCM_RTS|TIOCM_CTS|TIOCM_DSR|0x4000]) ioctl(13, TIOCMSET, [TIOCM_DTR|TIOCM_RTS])

which results in the following set

ioctl(13, TIOCMGET, [TIOCM_DTR|TIOCM_RTS|TIOCM_CTS|TIOCM_DSR])

Note the difference between the default set and new one: the missing bit, 0x4000, which is unknown to strace(1) but is defined as TIOCM_OUT2 in linux headers.

After that change (resetting the TIOCM_OUT2 bit), no writes to the serial port works anymore, they're all gets accepted by host kernel and are buffered in the kernel.

After some time, when the kernel buffer fills up, and since the port (on host side) is opened in non-blocking mode, the writes starts returning EAGAIN, and kvm process starts endless loop, which were seen by David.

Here's the trivial program to demonstrate the idea:

---- cut ---
#include <sys/types.h>
#include <unistd.h>
#include <fcntl.h>
#include <sys/ioctl.h>
#include <termios.h>

int main(int argc, char **argv) {
fd = open("/dev/ttyS0", O_RDWR|O_NONBLOCK); fcntl(fd, F_SETFL, O_RDWR);
x = TIOCM_DTR|TIOCM_RTS|TIOCM_CTS|TIOCM_DSR // |0x4000 ;
ioctl(fd, TIOCMSET, &x);
ioctl(fd, TIOCMGET, &x);
write(fd, "at\r", 3);
read(fd, buf, 20);
close(fd);

return 0;
}
--- cut ---

Run it under strace while a dialup modem is connected to /dev/ttyS0 (i used this way for testing). It will stuck at read, and nothing will be written, even when write() will happily return 3. Un-comment the |0x4000 thing, and it will work.

I'm not sure what should be done with this, and how much this is linux-specific. But it is obvious that bit (TIOCM_OUT2) should be left in-place (after which the thing works), at least on linux.

Note that this bit is NOT shown in /proc/tty/driver/serial file (which shows other bits).

Note also that this file (/proc/tty/driver/serial) helps to see if any write were performed: compare the counters. In 'tx' there's number of bytes actually sent to device, as opposed to accepted by the kernel. When you write something to /dev/ttyS0, that number increases, IF that something actually reached the device.

Quoted 3 lines

Thanks.

/mjt

qemu/qemu-char.c renamed from qemu/qemu-char.c.orig +10 -8
@@ -1067,17 +1067,17 @@ static int tty_serial_ioctl(CharDriverSt
             int *targ = (int *)arg;
             ioctl(s->fd_in, TIOCMGET, &sarg);
             *targ = 0;
-            if (sarg | TIOCM_CTS)
+            if (sarg & TIOCM_CTS)
                 *targ |= CHR_TIOCM_CTS;
-            if (sarg | TIOCM_CAR)
+            if (sarg & TIOCM_CAR)
                 *targ |= CHR_TIOCM_CAR;
-            if (sarg | TIOCM_DSR)
+            if (sarg & TIOCM_DSR)
                 *targ |= CHR_TIOCM_DSR;
-            if (sarg | TIOCM_RI)
+            if (sarg & TIOCM_RI)
                 *targ |= CHR_TIOCM_RI;
-            if (sarg | TIOCM_DTR)
+            if (sarg & TIOCM_DTR)
                 *targ |= CHR_TIOCM_DTR;
-            if (sarg | TIOCM_RTS)
+            if (sarg & TIOCM_RTS)
                 *targ |= CHR_TIOCM_RTS;
         }
         break;
@@ -1085,9 +1085,11 @@ static int tty_serial_ioctl(CharDriverSt
         {
             int sarg = *(int *)arg;
             int targ = 0;
-            if (sarg | CHR_TIOCM_DTR)
+            ioctl(s->fd_in, TIOCMGET, &targ);
+            targ &= ~(TIOCM_DTR|TIOCM_RTS);
+            if (sarg & CHR_TIOCM_DTR)
                 targ |= TIOCM_DTR;
-            if (sarg | CHR_TIOCM_RTS)
+            if (sarg & CHR_TIOCM_RTS)
                 targ |= TIOCM_RTS;
             ioctl(s->fd_in, TIOCMSET, &targ);
         }