[Qemu-devel] [BUG] GTK terminal is broken

[Qemu-devel] [BUG] GTK terminal is broken

[Stefan Weil]

Hello,

maybe most developers will already have noticed that the terminal output
in QEMU's GTK user interface is broken. As far as I know, it never worked,
but as there are working alternatives, I did not care much and forgot to
report the issues.

See these snapshots which show the problems with QEMU's GTK terminals:

http://qemu.weilnetz.de/test/bugs/qemu-gtk-demo1.png

Here the first line of the QEMUmonitor is only partially visible.
Tested with Cygwin/X on W64 host, QEMU running on Debian Linux.
Other X servers (e.g. native X on Debian Linux) don't show this problem.


http://qemu.weilnetz.de/test/bugs/qemu-gtk-demo2.png

This shows crippled output of "help" in the QEMU monitor
(look for "default frequency" at the bottom of the page).
Tested with several X servers.


http://qemu.weilnetz.de/test/bugs/qemu-gtk-demo3.png

This shows crippled output in the QEMU serial console.
The Linux kernel boot messages are hardly readable.
Tested with several X servers, x86 and mips kernels,
GTK 2 and GTK 3.


Regards,

StefanW.

Re: [Qemu-devel] [BUG] GTK terminal is broken

[Anthony Liguori]

Stefan Weil <sw@weilnetz.de> writes:

> Hello,
>
> maybe most developers will already have noticed that the terminal output
> in QEMU's GTK user interface is broken. As far as I know, it never worked,
> but as there are working alternatives, I did not care much and forgot to
> report the issues.
>
> See these snapshots which show the problems with QEMU's GTK terminals:
>
> http://qemu.weilnetz.de/test/bugs/qemu-gtk-demo1.png
>
> Here the first line of the QEMUmonitor is only partially visible.
> Tested with Cygwin/X on W64 host, QEMU running on Debian Linux.
> Other X servers (e.g. native X on Debian Linux) don't show this
> problem.

It's a GTK/Cygwin/X bug.

We use a vbox without anything particularly fancy happening.  Looks like
something is calculating layout incorrectly.

Regards,

Anthony Liguori

> http://qemu.weilnetz.de/test/bugs/qemu-gtk-demo2.png
>
> This shows crippled output of "help" in the QEMU monitor
> (look for "default frequency" at the bottom of the page).
> Tested with several X servers.
>
>
> http://qemu.weilnetz.de/test/bugs/qemu-gtk-demo3.png
>
> This shows crippled output in the QEMU serial console.
> The Linux kernel boot messages are hardly readable.
> Tested with several X servers, x86 and mips kernels,
> GTK 2 and GTK 3.
>
>
> Regards,
>
> StefanW.

Re: [Qemu-devel] [BUG] GTK terminal is broken

[Stefan Weil]

Am 29.07.2013 20:05, schrieb Anthony Liguori:
> Stefan Weil <sw@weilnetz.de> writes:
>
>> Hello,
>>
>> maybe most developers will already have noticed that the terminal output
>> in QEMU's GTK user interface is broken. As far as I know, it never worked,
>> but as there are working alternatives, I did not care much and forgot to
>> report the issues.
>>
>> See these snapshots which show the problems with QEMU's GTK terminals:
>>
>> http://qemu.weilnetz.de/test/bugs/qemu-gtk-demo1.png
>>
>> Here the first line of the QEMUmonitor is only partially visible.
>> Tested with Cygwin/X on W64 host, QEMU running on Debian Linux.
>> Other X servers (e.g. native X on Debian Linux) don't show this
>> problem.
> It's a GTK/Cygwin/X bug.
>
> We use a vbox without anything particularly fancy happening.  Looks like
> something is calculating layout incorrectly.
>
> Regards,
>
> Anthony Liguori

I miss a comment on the more important next two points.
They are _not_ related to Cygwin / X.

Do you investigate them?


>> http://qemu.weilnetz.de/test/bugs/qemu-gtk-demo2.png
>>
>> This shows crippled output of "help" in the QEMU monitor
>> (look for "default frequency" at the bottom of the page).
>> Tested with several X servers.
>>
>>
>> http://qemu.weilnetz.de/test/bugs/qemu-gtk-demo3.png
>>
>> This shows crippled output in the QEMU serial console.
>> The Linux kernel boot messages are hardly readable.
>> Tested with several X servers, x86 and mips kernels,
>> GTK 2 and GTK 3.
>>
>>
>> Regards,
>>
>> StefanW.

Re: [Qemu-devel] [BUG] GTK terminal is broken

[Anthony Liguori]

Stefan Weil <sw@weilnetz.de> writes:

> Am 29.07.2013 20:05, schrieb Anthony Liguori:
>> Stefan Weil <sw@weilnetz.de> writes:
>>
>>> Hello,
>>>
>>> maybe most developers will already have noticed that the terminal output
>>> in QEMU's GTK user interface is broken. As far as I know, it never worked,
>>> but as there are working alternatives, I did not care much and forgot to
>>> report the issues.
>>>
>>> See these snapshots which show the problems with QEMU's GTK terminals:
>>>
>>> http://qemu.weilnetz.de/test/bugs/qemu-gtk-demo1.png
>>>
>>> Here the first line of the QEMUmonitor is only partially visible.
>>> Tested with Cygwin/X on W64 host, QEMU running on Debian Linux.
>>> Other X servers (e.g. native X on Debian Linux) don't show this
>>> problem.
>> It's a GTK/Cygwin/X bug.
>>
>> We use a vbox without anything particularly fancy happening.  Looks like
>> something is calculating layout incorrectly.
>>
>> Regards,
>>
>> Anthony Liguori
>
> I miss a comment on the more important next two points.
> They are _not_ related to Cygwin / X.
>
> Do you investigate them?

What you're seeing has always existed and is not GTK specific.  There is
not flow control on the outgoing side of serial emulation.  Data will
get dropped.

My guess is that your system is sufficiently slow and overloaded that
it's happening a lot for you.

Regards,

Anthony Liguori

>
>
>>> http://qemu.weilnetz.de/test/bugs/qemu-gtk-demo2.png
>>>
>>> This shows crippled output of "help" in the QEMU monitor
>>> (look for "default frequency" at the bottom of the page).
>>> Tested with several X servers.
>>>
>>>
>>> http://qemu.weilnetz.de/test/bugs/qemu-gtk-demo3.png
>>>
>>> This shows crippled output in the QEMU serial console.
>>> The Linux kernel boot messages are hardly readable.
>>> Tested with several X servers, x86 and mips kernels,
>>> GTK 2 and GTK 3.
>>>
>>>
>>> Regards,
>>>
>>> StefanW.

Re: [Qemu-devel] [BUG] GTK terminal is broken

[Stefan Weil]

Am 29.07.2013 20:55, schrieb Anthony Liguori:
> Stefan Weil <sw@weilnetz.de> writes:
>
>> Am 29.07.2013 20:05, schrieb Anthony Liguori:
>>> Stefan Weil <sw@weilnetz.de> writes:
>>>
>>>> Hello,
>>>>
>>>> maybe most developers will already have noticed that the terminal output
>>>> in QEMU's GTK user interface is broken. As far as I know, it never worked,
>>>> but as there are working alternatives, I did not care much and forgot to
>>>> report the issues.
>>>>
>>>> See these snapshots which show the problems with QEMU's GTK terminals:
>>>>
>>>> http://qemu.weilnetz.de/test/bugs/qemu-gtk-demo1.png
>>>>
>>>> Here the first line of the QEMUmonitor is only partially visible.
>>>> Tested with Cygwin/X on W64 host, QEMU running on Debian Linux.
>>>> Other X servers (e.g. native X on Debian Linux) don't show this
>>>> problem.
>>> It's a GTK/Cygwin/X bug.
>>>
>>> We use a vbox without anything particularly fancy happening.  Looks like
>>> something is calculating layout incorrectly.
>>>
>>> Regards,
>>>
>>> Anthony Liguori
>> I miss a comment on the more important next two points.
>> They are _not_ related to Cygwin / X.
>>
>> Do you investigate them?
> What you're seeing has always existed and is not GTK specific.  There is
> not flow control on the outgoing side of serial emulation.  Data will
> get dropped.
>
> My guess is that your system is sufficiently slow and overloaded that
> it's happening a lot for you.
>
> Regards,
>
> Anthony Liguori

If it were as simply as that, that would be fine, but alas,
it isn't. Even fast servers without any load show dropped
characters. There are dropped characters for QEMU's
monitor window (Alt-ctrl-2) and for Malta's line display
(Alt-Ctrl-5). Neither of these two should suffer from missing
flow control.

And even a current X.Org on Ubuntu shows the first bug
(crippled first line) here, so it is not Cygwin / X specific.

None of the Linux hosts which I use in my test is able
to display a correct Linux boot protocol on QEMU's
GTK serial console window. All of them work with SDL
or when the serial console is redirected to stdout.

Regards,

Stefan Weil

Re: [Qemu-devel] [BUG] GTK terminal is broken

[Andreas Färber]

Am 29.07.2013 21:27, schrieb Stefan Weil:
> Am 29.07.2013 20:55, schrieb Anthony Liguori:
>> Stefan Weil <sw@weilnetz.de> writes:
>>
>>> Am 29.07.2013 20:05, schrieb Anthony Liguori:
>>>> Stefan Weil <sw@weilnetz.de> writes:
>>>>
>>>>> Hello,
>>>>>
>>>>> maybe most developers will already have noticed that the terminal output
>>>>> in QEMU's GTK user interface is broken. As far as I know, it never worked,
>>>>> but as there are working alternatives, I did not care much and forgot to
>>>>> report the issues.
>>>>>
>>>>> See these snapshots which show the problems with QEMU's GTK terminals:
>>>>>
>>>>> http://qemu.weilnetz.de/test/bugs/qemu-gtk-demo1.png
>>>>>
>>>>> Here the first line of the QEMUmonitor is only partially visible.
>>>>> Tested with Cygwin/X on W64 host, QEMU running on Debian Linux.
>>>>> Other X servers (e.g. native X on Debian Linux) don't show this
>>>>> problem.
>>>> It's a GTK/Cygwin/X bug.
>>>>
>>>> We use a vbox without anything particularly fancy happening.  Looks like
>>>> something is calculating layout incorrectly.
>>>>
>>>> Regards,
>>>>
>>>> Anthony Liguori
>>> I miss a comment on the more important next two points.
>>> They are _not_ related to Cygwin / X.
>>>
>>> Do you investigate them?
>> What you're seeing has always existed and is not GTK specific.  There is
>> not flow control on the outgoing side of serial emulation.  Data will
>> get dropped.
>>
>> My guess is that your system is sufficiently slow and overloaded that
>> it's happening a lot for you.
>>
>> Regards,
>>
>> Anthony Liguori
> 
> If it were as simply as that, that would be fine, but alas,
> it isn't. Even fast servers without any load show dropped
> characters. There are dropped characters for QEMU's
> monitor window (Alt-ctrl-2) and for Malta's line display
> (Alt-Ctrl-5). Neither of these two should suffer from missing
> flow control.
> 
> And even a current X.Org on Ubuntu shows the first bug
> (crippled first line) here, so it is not Cygwin / X specific.

FWIW I've been seeing similar garbled-Gtk-output issues on openSUSE 12.3
but didn't investigate further. -monitor stdio has so far not failed me
on the same system by comparison.

One reproducible test case was the openrisc kernel that Jia provided
recently.

Regards,
Andreas

> 
> None of the Linux hosts which I use in my test is able
> to display a correct Linux boot protocol on QEMU's
> GTK serial console window. All of them work with SDL
> or when the serial console is redirected to stdout.
> 
> Regards,
> 
> Stefan Weil
> 
> 


-- 
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer; HRB 16746 AG Nürnberg

Re: [Qemu-devel] [BUG] GTK terminal is broken

[Stefan Weil]

Am 29.07.2013 22:34, schrieb Andreas Färber:
> Am 29.07.2013 21:27, schrieb Stefan Weil:
>> Am 29.07.2013 20:55, schrieb Anthony Liguori:
>>> Stefan Weil <sw@weilnetz.de> writes:
>>>
>>>> Am 29.07.2013 20:05, schrieb Anthony Liguori:
>>>>> Stefan Weil <sw@weilnetz.de> writes:
>>>>>
>>>>>> Hello,
>>>>>>
>>>>>> maybe most developers will already have noticed that the terminal output
>>>>>> in QEMU's GTK user interface is broken. As far as I know, it never worked,
>>>>>> but as there are working alternatives, I did not care much and forgot to
>>>>>> report the issues.
>>>>>>
>>>>>> See these snapshots which show the problems with QEMU's GTK terminals:
>>>>>>
>>>>>> http://qemu.weilnetz.de/test/bugs/qemu-gtk-demo1.png
>>>>>>
>>>>>> Here the first line of the QEMUmonitor is only partially visible.
>>>>>> Tested with Cygwin/X on W64 host, QEMU running on Debian Linux.
>>>>>> Other X servers (e.g. native X on Debian Linux) don't show this
>>>>>> problem.
>>>>> It's a GTK/Cygwin/X bug.
>>>>>
>>>>> We use a vbox without anything particularly fancy happening.  Looks like
>>>>> something is calculating layout incorrectly.
>>>>>
>>>>> Regards,
>>>>>
>>>>> Anthony Liguori
>>>> I miss a comment on the more important next two points.
>>>> They are _not_ related to Cygwin / X.
>>>>
>>>> Do you investigate them?
>>> What you're seeing has always existed and is not GTK specific.  There is
>>> not flow control on the outgoing side of serial emulation.  Data will
>>> get dropped.
>>>
>>> My guess is that your system is sufficiently slow and overloaded that
>>> it's happening a lot for you.
>>>
>>> Regards,
>>>
>>> Anthony Liguori
>> If it were as simply as that, that would be fine, but alas,
>> it isn't. Even fast servers without any load show dropped
>> characters. There are dropped characters for QEMU's
>> monitor window (Alt-ctrl-2) and for Malta's line display
>> (Alt-Ctrl-5). Neither of these two should suffer from missing
>> flow control.
>>
>> And even a current X.Org on Ubuntu shows the first bug
>> (crippled first line) here, so it is not Cygwin / X specific.
> FWIW I've been seeing similar garbled-Gtk-output issues on openSUSE 12.3
> but didn't investigate further. -monitor stdio has so far not failed me
> on the same system by comparison.
>
> One reproducible test case was the openrisc kernel that Jia provided
> recently.
>
> Regards,
> Andreas

Maybe the problems only occur with TCG, but not with KVM.
That might explain why I get them more often than Anthony :-)

Up to now, I did not try GTK terminals with KVM because my
typical test scenario uses MIPS or ARM boards on Intel hosts.

Regards,
Stefan

Re: [Qemu-devel] [BUG] GTK terminal is broken

[Anthony Liguori]

Stefan Weil <sw@weilnetz.de> writes:

> Am 29.07.2013 22:34, schrieb Andreas Färber:
>> Am 29.07.2013 21:27, schrieb Stefan Weil:
>>> Am 29.07.2013 20:55, schrieb Anthony Liguori:
>>>> Stefan Weil <sw@weilnetz.de> writes:
>>>>
>>>>> Am 29.07.2013 20:05, schrieb Anthony Liguori:
>>>>>> Stefan Weil <sw@weilnetz.de> writes:
>>>>>>
>>>>>>> Hello,
>>>>>>>
>>>>>>> maybe most developers will already have noticed that the terminal output
>>>>>>> in QEMU's GTK user interface is broken. As far as I know, it never worked,
>>>>>>> but as there are working alternatives, I did not care much and forgot to
>>>>>>> report the issues.
>>>>>>>
>>>>>>> See these snapshots which show the problems with QEMU's GTK terminals:
>>>>>>>
>>>>>>> http://qemu.weilnetz.de/test/bugs/qemu-gtk-demo1.png
>>>>>>>
>>>>>>> Here the first line of the QEMUmonitor is only partially visible.
>>>>>>> Tested with Cygwin/X on W64 host, QEMU running on Debian Linux.
>>>>>>> Other X servers (e.g. native X on Debian Linux) don't show this
>>>>>>> problem.
>>>>>> It's a GTK/Cygwin/X bug.
>>>>>>
>>>>>> We use a vbox without anything particularly fancy happening.  Looks like
>>>>>> something is calculating layout incorrectly.
>>>>>>
>>>>>> Regards,
>>>>>>
>>>>>> Anthony Liguori
>>>>> I miss a comment on the more important next two points.
>>>>> They are _not_ related to Cygwin / X.
>>>>>
>>>>> Do you investigate them?
>>>> What you're seeing has always existed and is not GTK specific.  There is
>>>> not flow control on the outgoing side of serial emulation.  Data will
>>>> get dropped.
>>>>
>>>> My guess is that your system is sufficiently slow and overloaded that
>>>> it's happening a lot for you.
>>>>
>>>> Regards,
>>>>
>>>> Anthony Liguori
>>> If it were as simply as that, that would be fine, but alas,
>>> it isn't. Even fast servers without any load show dropped
>>> characters. There are dropped characters for QEMU's
>>> monitor window (Alt-ctrl-2) and for Malta's line display
>>> (Alt-Ctrl-5). Neither of these two should suffer from missing
>>> flow control.
>>>
>>> And even a current X.Org on Ubuntu shows the first bug
>>> (crippled first line) here, so it is not Cygwin / X specific.
>> FWIW I've been seeing similar garbled-Gtk-output issues on openSUSE 12.3
>> but didn't investigate further. -monitor stdio has so far not failed me
>> on the same system by comparison.
>>
>> One reproducible test case was the openrisc kernel that Jia provided
>> recently.
>>
>> Regards,
>> Andreas
>
> Maybe the problems only occur with TCG, but not with KVM.
> That might explain why I get them more often than Anthony :-)
>
> Up to now, I did not try GTK terminals with KVM because my
> typical test scenario uses MIPS or ARM boards on Intel hosts.

x86 vs. other will come into play too.

The serial driver for x86 has some basic backoff logic.  I suspect
others don't.

My guess why you don't see this on -serial stdio is that it probably is
using synchronous I/O.

Regards,

Anthony Liguori

>
> Regards,
> Stefan

Re: [Qemu-devel] [BUG] GTK terminal is broken

[Andreas Färber]

Am 29.07.2013 22:34, schrieb Andreas Färber:
> FWIW I've been seeing similar garbled-Gtk-output issues on openSUSE 12.3
> but didn't investigate further. -monitor stdio has so far not failed me
> on the same system by comparison.

-serial stdio obviously. :)

> One reproducible test case was the openrisc kernel that Jia provided
> recently.
> 
> Regards,
> Andreas

-- 
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer; HRB 16746 AG Nürnberg

Re: [Qemu-devel] [BUG] GTK terminal is broken

[Laszlo Ersek]

[-- Attachment #1: Type: text/plain, Size: 3454 bytes --]

On 07/29/13 20:18, Stefan Weil wrote:
> Am 29.07.2013 20:05, schrieb Anthony Liguori:
>> Stefan Weil <sw@weilnetz.de> writes:
>>
>>> Hello,
>>>
>>> maybe most developers will already have noticed that the terminal output
>>> in QEMU's GTK user interface is broken. As far as I know, it never worked,
>>> but as there are working alternatives, I did not care much and forgot to
>>> report the issues.
>>>
>>> See these snapshots which show the problems with QEMU's GTK terminals:
>>>
>>> http://qemu.weilnetz.de/test/bugs/qemu-gtk-demo1.png
>>>
>>> Here the first line of the QEMUmonitor is only partially visible.
>>> Tested with Cygwin/X on W64 host, QEMU running on Debian Linux.
>>> Other X servers (e.g. native X on Debian Linux) don't show this
>>> problem.
>> It's a GTK/Cygwin/X bug.
>>
>> We use a vbox without anything particularly fancy happening.  Looks like
>> something is calculating layout incorrectly.
>>
>> Regards,
>>
>> Anthony Liguori
> 
> I miss a comment on the more important next two points.
> They are _not_ related to Cygwin / X.
> 
> Do you investigate them?

I did some investigation, and the results are not pretty. I'm using
RHEL-6.4 host OS (kernel, gtk, vte etc), with fresh upstream qemu. The
command line (started as a mere user --> TCG) was

./qemu-system-x86_64 -cdrom
/home/virt-images/isos/Fedora-18-x86_64-Live-XFCE.iso

Please find the attached debug patch (it has a small functional change
as well, I'll discuss it below).

(a) The patch extends the qemu_hexdump() function with ASCII strings. I
wanted to ask you to run your tests with this patch applied, and try to
correlate the hexdumps with the breakage on the screen.

However, when testing the patch myself, I ran into problems.

(b) The patch adds a qemu_set_nonblock(slave_fd) call to gd_vc_init(). I
tested the patch both with and without this hunk.

The key is to produce big output for VTE. For that purpose I waited
until the Fedora 18 kernel enabled paging, and then issued "info tlb".

("info tlb" should be familiar from a recent monitor bugfix, as test
case / trigger, but in bleeding edge qemu the monitor is OK.)

So, if I do not add (b), the debug patch starts to spew data to stderr,
and after a while qemu hangs hard, in the write() call in gd_vc_chr_write().

If I add (b), then the same happens initially, followed by qemu spinning
(ie. entering and exiting rapidly, without making any progress)
gd_vc_chr_write(), with write() returning -1/EAGAIN.

VTE seems to expect the owner of the slave side to run in a different
process or thread. While browsing vte_terminal_io_read(), I found some
throttling code in the read path, where VTE tries to keep one child
process (maybe running in one tab?) to monopolize it.

I think this could explain the deadlock / busy wait behavior. In qemu
both sides of the terminal are manipulated by the same thread, and both
file descriptors are (should be) added to the same GLib main loop. At
some point VTE would be able read the master side but refuses to, the
pty buffer becomes full, the slave-side write() in gd_vc_chr_write()
blocks, and VTE never gets back control to release the throttle.

This doesn't explain why qemu never recovers when the slave side is set
to non-blocking mode (ie. with (b) added). I suspect that although we
manage to run circles around the main loop in this case, we either never
reach the master-fd read in VTE, or VTE always decides it's still too
early to read.

Laszlo

[-- Attachment #2: 0001-gd_vc_chr_write-print-hexdump-set-nonblock.patch --]
[-- Type: text/plain, Size: 3128 bytes --]

>From a598fb00caf89a02eb64145cdada559f9405235b Mon Sep 17 00:00:00 2001
From: Laszlo Ersek <lersek@redhat.com>
Date: Mon, 29 Jul 2013 21:27:04 +0200
Subject: [PATCH] gd_vc_chr_write(): print hexdump, set nonblock


Signed-off-by: Laszlo Ersek <lersek@redhat.com>
---
 ui/gtk.c       |   14 +++++++++++++-
 util/hexdump.c |   39 ++++++++++++++++++++++++++++++++-------
 2 files changed, 45 insertions(+), 8 deletions(-)

diff --git a/ui/gtk.c b/ui/gtk.c
index c38146f..c88bd37 100644
--- a/ui/gtk.c
+++ b/ui/gtk.c
@@ -35,6 +35,7 @@
 #define LOCALEDIR "po"
 
 #include "qemu-common.h"
+#include "qemu/sockets.h"
 
 #ifdef CONFIG_PRAGMA_DIAGNOSTIC_AVAILABLE
 /* Work around an -Wstrict-prototypes warning in GTK headers */
@@ -1119,8 +1120,18 @@ static gboolean gd_focus_out_event(GtkWidget *widget,
 static int gd_vc_chr_write(CharDriverState *chr, const uint8_t *buf, int len)
 {
     VirtualConsole *vc = chr->opaque;
+    int written, errno_save;
 
-    return write(vc->fd, buf, len);
+    errno = 0;
+    written = write(vc->fd, buf, len);
+    errno_save = errno;
+
+    fprintf(stderr, "%s: len=%d written=%d errno=%d\n", __FUNCTION__, len,
+            written, errno);
+    qemu_hexdump((const char *)buf, stderr, __FUNCTION__, len);
+
+    errno = errno_save;
+    return written;
 }
 
 static int nb_vcs;
@@ -1213,6 +1224,7 @@ static GSList *gd_vc_init(GtkDisplayState *s, VirtualConsole *vc, int index, GSL
     vte_terminal_set_size(VTE_TERMINAL(vc->terminal), 80, 25);
 
     vc->fd = slave_fd;
+    qemu_set_nonblock(slave_fd);
     vc->chr->opaque = vc;
     vc->scrolled_window = scrolled_window;
 
diff --git a/util/hexdump.c b/util/hexdump.c
index 969b340..994fd69 100644
--- a/util/hexdump.c
+++ b/util/hexdump.c
@@ -13,25 +13,50 @@
  * GNU GPL, version 2 or (at your option) any later version.
  */
 
+#include <ctype.h>
 #include "qemu-common.h"
 
+static void asciiize(const char *buf, FILE *fp, size_t size, unsigned *col)
+{
+    size_t i;
+
+    while (*col < 4 + 1 + 4 * (1 + 4 * (1 + 2))) {
+        fputc(' ', fp);
+        ++*col;
+    }
+    *col = 0;
+
+    fprintf(fp, "  '");
+    for (i = 0; i < size; ++i) {
+      unsigned char c = buf[i];
+
+      fputc(isprint(c) ? c : '.', fp);
+    }
+    fprintf(fp, "'\n");
+}
+
 void qemu_hexdump(const char *buf, FILE *fp, const char *prefix, size_t size)
 {
     unsigned int b;
+    unsigned pos, col;
 
     for (b = 0; b < size; b++) {
-        if ((b % 16) == 0) {
+        pos = b % 16;
+
+        if (pos == 0) {
             fprintf(fp, "%s: %04x:", prefix, b);
+            col = 5;
         }
         if ((b % 4) == 0) {
-            fprintf(fp, " ");
+            col += fprintf(fp, " ");
         }
-        fprintf(fp, " %02x", (unsigned char)buf[b]);
-        if ((b % 16) == 15) {
-            fprintf(fp, "\n");
+        col += fprintf(fp, " %02x", (unsigned char)buf[b]);
+        if (pos == 15) {
+          asciiize(buf + (b - 15), fp, 16, &col);
         }
     }
-    if ((b % 16) != 0) {
-        fprintf(fp, "\n");
+    pos = b % 16;
+    if (pos != 0) {
+        asciiize(buf + (b - pos), fp, pos, &col);
     }
 }
-- 
1.7.1

Re: [Qemu-devel] [BUG] GTK terminal is broken

[Stefan Weil]

Am 30.07.2013 00:35, schrieb Laszlo Ersek:
> On 07/29/13 20:18, Stefan Weil wrote:
>> Am 29.07.2013 20:05, schrieb Anthony Liguori:
>>> Stefan Weil <sw@weilnetz.de> writes:
>>>
>>>> Hello,
>>>>
>>>> maybe most developers will already have noticed that the terminal output
>>>> in QEMU's GTK user interface is broken. As far as I know, it never worked,
>>>> but as there are working alternatives, I did not care much and forgot to
>>>> report the issues.
>>>>
>>>> See these snapshots which show the problems with QEMU's GTK terminals:
>>>>
>>>> http://qemu.weilnetz.de/test/bugs/qemu-gtk-demo1.png
>>>>
>>>> Here the first line of the QEMUmonitor is only partially visible.
>>>> Tested with Cygwin/X on W64 host, QEMU running on Debian Linux.
>>>> Other X servers (e.g. native X on Debian Linux) don't show this
>>>> problem.
>>> It's a GTK/Cygwin/X bug.
>>>
>>> We use a vbox without anything particularly fancy happening.  Looks like
>>> something is calculating layout incorrectly.
>>>
>>> Regards,
>>>
>>> Anthony Liguori
>> I miss a comment on the more important next two points.
>> They are _not_ related to Cygwin / X.
>>
>> Do you investigate them?
> I did some investigation, and the results are not pretty. I'm using
> RHEL-6.4 host OS (kernel, gtk, vte etc), with fresh upstream qemu. The
> command line (started as a mere user --> TCG) was
>
> ./qemu-system-x86_64 -cdrom
> /home/virt-images/isos/Fedora-18-x86_64-Live-XFCE.iso
>
> Please find the attached debug patch (it has a small functional change
> as well, I'll discuss it below).
>
> (a) The patch extends the qemu_hexdump() function with ASCII strings. I
> wanted to ask you to run your tests with this patch applied, and try to
> correlate the hexdumps with the breakage on the screen.
>
> However, when testing the patch myself, I ran into problems.
>
> (b) The patch adds a qemu_set_nonblock(slave_fd) call to gd_vc_init(). I
> tested the patch both with and without this hunk.
>
> The key is to produce big output for VTE. For that purpose I waited
> until the Fedora 18 kernel enabled paging, and then issued "info tlb".
>
> ("info tlb" should be familiar from a recent monitor bugfix, as test
> case / trigger, but in bleeding edge qemu the monitor is OK.)
>
> So, if I do not add (b), the debug patch starts to spew data to stderr,
> and after a while qemu hangs hard, in the write() call in gd_vc_chr_write().
>
> If I add (b), then the same happens initially, followed by qemu spinning
> (ie. entering and exiting rapidly, without making any progress)
> gd_vc_chr_write(), with write() returning -1/EAGAIN.
>
> VTE seems to expect the owner of the slave side to run in a different
> process or thread. While browsing vte_terminal_io_read(), I found some
> throttling code in the read path, where VTE tries to keep one child
> process (maybe running in one tab?) to monopolize it.
>
> I think this could explain the deadlock / busy wait behavior. In qemu
> both sides of the terminal are manipulated by the same thread, and both
> file descriptors are (should be) added to the same GLib main loop. At
> some point VTE would be able read the master side but refuses to, the
> pty buffer becomes full, the slave-side write() in gd_vc_chr_write()
> blocks, and VTE never gets back control to release the throttle.
>
> This doesn't explain why qemu never recovers when the slave side is set
> to non-blocking mode (ie. with (b) added). I suspect that although we
> manage to run circles around the main loop in this case, we either never
> reach the master-fd read in VTE, or VTE always decides it's still too
> early to read.
>
> Laszlo

Thanks for your analysis. For me, it confirms that terminal
emulation with GTK is currently unusable.

Here is one more obvious bug (tested on Linux):

1. Run qemu-system-i386 -L pc-bios with GTK UI.
2. Select "Show tabs".
3. Unselect "Show tabs".

In 3, blank lines are added on top and on the bottom of the BIOS boot text,
so the screen in 3 is higher than in 1 although both should be identical.

Stefan

Re: [Qemu-devel] [BUG] GTK terminal is broken

[Stefan Weil]

Am 29.07.2013 18:04, schrieb Stefan Weil:
> Hello,
>
> maybe most developers will already have noticed that the terminal output
> in QEMU's GTK user interface is broken. As far as I know, it never worked,
> but as there are working alternatives, I did not care much and forgot to
> report the issues.
>
> See these snapshots which show the problems with QEMU's GTK terminals:
>
> http://qemu.weilnetz.de/test/bugs/qemu-gtk-demo1.png
>
> Here the first line of the QEMUmonitor is only partially visible.
> Tested with Cygwin/X on W64 host, QEMU running on Debian Linux.
> Other X servers (e.g. native X on Debian Linux) don't show this problem.
>
>
> http://qemu.weilnetz.de/test/bugs/qemu-gtk-demo2.png
>
> This shows crippled output of "help" in the QEMU monitor
> (look for "default frequency" at the bottom of the page).
> Tested with several X servers.
>
>
> http://qemu.weilnetz.de/test/bugs/qemu-gtk-demo3.png
>
> This shows crippled output in the QEMU serial console.
> The Linux kernel boot messages are hardly readable.
> Tested with several X servers, x86 and mips kernels,
> GTK 2 and GTK 3.
>
>
> Regards,
>
> StefanW.
>

This is just a reminder that the GTK terminal is still broken in QEMU 1.7.
It only seems to work with KVM and in the monitor window, but it is
unusable in serial consoles which are typically needed for embedded
system emulation (ARM, MIPS, ...).

For GTK, it's not a regression because it never worked. All known hosts are
affected. The serial console remains usable with SDL or stdio.

See this link for more details:
http://lists.nongnu.org/archive/html/qemu-devel/2013-07/msg05740.html

Regards,
Stefan

PS. For MinGW, I compile QEMU without console support because there is
no VTE for that platform.

Re: [Qemu-devel] [BUG] GTK terminal is broken

[Stefan Weil]

Am 14.11.2013 22:32, schrieb Stefan Weil:
> Am 29.07.2013 18:04, schrieb Stefan Weil:
>> Hello,
>>
>> maybe most developers will already have noticed that the terminal output
>> in QEMU's GTK user interface is broken. As far as I know, it never worked,
>> but as there are working alternatives, I did not care much and forgot to
>> report the issues.
>>
>> See these snapshots which show the problems with QEMU's GTK terminals:
>>
>> http://qemu.weilnetz.de/test/bugs/qemu-gtk-demo1.png
>>
>> Here the first line of the QEMUmonitor is only partially visible.
>> Tested with Cygwin/X on W64 host, QEMU running on Debian Linux.
>> Other X servers (e.g. native X on Debian Linux) don't show this problem.
>>
>>
>> http://qemu.weilnetz.de/test/bugs/qemu-gtk-demo2.png
>>
>> This shows crippled output of "help" in the QEMU monitor
>> (look for "default frequency" at the bottom of the page).
>> Tested with several X servers.
>>
>>
>> http://qemu.weilnetz.de/test/bugs/qemu-gtk-demo3.png
>>
>> This shows crippled output in the QEMU serial console.
>> The Linux kernel boot messages are hardly readable.
>> Tested with several X servers, x86 and mips kernels,
>> GTK 2 and GTK 3.
>>
>>
>> Regards,
>>
>> StefanW.
>>
> This is just a reminder that the GTK terminal is still broken in QEMU 1.7.
> It only seems to work with KVM and in the monitor window, but it is
> unusable in serial consoles which are typically needed for embedded
> system emulation (ARM, MIPS, ...).
>
> For GTK, it's not a regression because it never worked. All known hosts are
> affected. The serial console remains usable with SDL or stdio.
>
> See this link for more details:
> http://lists.nongnu.org/archive/html/qemu-devel/2013-07/msg05740.html
>
> Regards,
> Stefan
>
> PS. For MinGW, I compile QEMU without console support because there is
> no VTE for that platform.

Sorry, I used Anthony's old e-mail address, please fix when replying.

Here is another bug with GTK (any host), this time in the framebuffer
window:

run qemu-system-i386 (BIOS only), enable tabs in the menu, disable
tabs,compare the initial window with the BIOS messages with the final
one - they differ in height.

Stefan