Skip to contentSkip to conversation

char: remove init_reset handling, more data per write()

Qemu-devel In review Amit Shah 3 participants 13 messages Nov 3, 2009 – Nov 4, 2009

lore
Share & tools
lore

Amit Shah

author
#0 lore raw

Hello,

These patches, all independent of each other, make the char backend a little better by removing the initial_reset handling that's unnecessary, bump up the amount of data that's sent per write() call to 4k from the current 1k and also renames the generic char open() command to reflect reality.

Please apply.

Amit

author
#1 lore raw

PATCH 1/3 char: don't limit data sent to backends to 1k per buffer

chardevs have a 'can_read' function via which backends specify the amount of data they can receive. When can_read returns > 0, apps can start sending data. However, each chardev driver here allows a max. of 1k bytes inspite of the backend being able to receive more.

The best we can do here is to allocate s->max_size bytes from the heap on each call (which is the number returned by the backend from the can_read call).

This is an intermediate step to bump up the bytes written in each call to 4k.

---
 qemu-char.c |   14 ++++++++------
 1 files changed, 8 insertions(+), 6 deletions(-)
qemu-char.c +8 -6
index 40bd7e8..1f63019 100644
@@ -97,6 +97,8 @@
 
 #include "qemu_socket.h"
 
+#define READ_BUF_LEN 4096
+
 /***********************************************************/
 /* character device */
 
@@ -172,7 +174,7 @@ void qemu_chr_accept_input(CharDriverState *s)
 
 void qemu_chr_printf(CharDriverState *s, const char *fmt, ...)
 {
-    char buf[4096];
+    char buf[READ_BUF_LEN];
     va_list ap;
     va_start(ap, fmt);
     vsnprintf(buf, sizeof(buf), fmt, ap);
@@ -555,7 +557,7 @@ static void fd_chr_read(void *opaque)
     CharDriverState *chr = opaque;
     FDCharDriver *s = chr->opaque;
     int size, len;
-    uint8_t buf[1024];
+    uint8_t buf[READ_BUF_LEN];
 
     len = sizeof(buf);
     if (len > s->max_size)
@@ -866,7 +868,7 @@ static void pty_chr_read(void *opaque)
     CharDriverState *chr = opaque;
     PtyCharDriver *s = chr->opaque;
     int size, len;
-    uint8_t buf[1024];
+    uint8_t buf[READ_BUF_LEN];
 
     len = sizeof(buf);
     if (len > s->read_bytes)
@@ -1554,7 +1556,7 @@ static void win_chr_readfile(CharDriverState *chr)
 {
     WinCharState *s = chr->opaque;
     int ret, err;
-    uint8_t buf[1024];
+    uint8_t buf[READ_BUF_LEN];
     DWORD size;
 
     ZeroMemory(&s->orecv, sizeof(s->orecv));
@@ -1760,7 +1762,7 @@ static CharDriverState *qemu_chr_open_win_file_out(QemuOpts *opts)
 
 typedef struct {
     int fd;
-    uint8_t buf[1024];
+    uint8_t buf[READ_BUF_LEN];
     int bufcnt;
     int bufptr;
     int max_size;
@@ -2020,7 +2022,7 @@ static void tcp_chr_read(void *opaque)
 {
     CharDriverState *chr = opaque;
     TCPCharDriver *s = chr->opaque;
-    uint8_t buf[1024];
+    uint8_t buf[READ_BUF_LEN];
     int len, size;
 
     if (!s->connected || s->max_size <= 0)
Signature · 1 line
1.6.2.5
author
#2 lore raw

PATCH 2/3 char: Remove special init_reset handling

The initial_reset sent to chardevs doesn't do much other than setting a bool to true. Char devices are interested in the open event and that gets sent whenever the device is opened.

Moreover, the reset logic breaks as and when qemu's bh scheduling changes.

---
 qemu-char.c |    9 ---------
 qemu-char.h |    1 -
 vl.c        |    1 -
 3 files changed, 0 insertions(+), 11 deletions(-)
qemu-char.c +0 -9
index 1f63019..d78bae3 100644
@@ -128,15 +128,6 @@ void qemu_chr_reset(CharDriverState *s)
     }
 }
 
-void qemu_chr_initial_reset(void)
-{
-    CharDriverState *chr;
-
-    QTAILQ_FOREACH(chr, &chardevs, next) {
-        qemu_chr_reset(chr);
-    }
-}
-
 int qemu_chr_write(CharDriverState *s, const uint8_t *buf, int len)
 {
     return s->chr_write(s, buf, len);
qemu-char.h +0 -1
index 05fe15d..b420111 100644
@@ -83,7 +83,6 @@ void qemu_chr_add_handlers(CharDriverState *s,
                            void *opaque);
 int qemu_chr_ioctl(CharDriverState *s, int cmd, void *arg);
 void qemu_chr_reset(CharDriverState *s);
-void qemu_chr_initial_reset(void);
 int qemu_chr_can_read(CharDriverState *s);
 void qemu_chr_read(CharDriverState *s, uint8_t *buf, int len);
 int qemu_chr_get_msgfd(CharDriverState *s);
vl.c +0 -1
index e57f58f..5ad4e42 100644
@@ -5750,7 +5750,6 @@ int main(int argc, char **argv, char **envp)
     }
 
     text_consoles_set_display(display_state);
-    qemu_chr_initial_reset();
 
     for (i = 0; i < MAX_MONITOR_DEVICES; i++) {
         if (monitor_devices[i] && monitor_hds[i]) {
Signature · 1 line
1.6.2.5
author
#3 lore raw

PATCH 3/3 char: rename qemu_chr_reset to qemu_chr_generic_open

This function sends out the OPENED event to backends that have drive the chardevs. The 'reset' is now a historical artifact and we can now just call the function for what it is.

---
 console.c   |    2 +-
 hw/baum.c   |    2 +-
 qemu-char.c |   22 +++++++++++-----------
 qemu-char.h |    2 +-
 4 files changed, 14 insertions(+), 14 deletions(-)
console.c +1 -1
index 9bbef59..82ddbe4 100644
@@ -1384,7 +1384,7 @@ static void text_console_do_init(CharDriverState *chr, DisplayState *ds, QemuOpt
     s->t_attrib = s->t_attrib_default;
     text_console_resize(s);
 
-    qemu_chr_reset(chr);
+    qemu_chr_generic_open(chr);
     if (chr->init)
         chr->init(chr);
 }
hw/baum.c +1 -1
index 8a12985..fa356ec 100644
@@ -614,7 +614,7 @@ CharDriverState *chr_baum_init(QemuOpts *opts)
 
     qemu_set_fd_handler(baum->brlapi_fd, baum_chr_read, NULL, baum);
 
-    qemu_chr_reset(chr);
+    qemu_chr_generic_open(chr);
 
     return chr;
 
qemu-char.c +11 -11
index d78bae3..5a81e8f 100644
@@ -112,7 +112,7 @@ static void qemu_chr_event(CharDriverState *s, int event)
     s->chr_event(s->handler_opaque, event);
 }
 
-static void qemu_chr_reset_bh(void *opaque)
+static void qemu_chr_generic_open_bh(void *opaque)
 {
     CharDriverState *s = opaque;
     qemu_chr_event(s, CHR_EVENT_OPENED);
@@ -120,10 +120,10 @@ static void qemu_chr_reset_bh(void *opaque)
     s->bh = NULL;
 }
 
-void qemu_chr_reset(CharDriverState *s)
+void qemu_chr_generic_open(CharDriverState *s)
 {
     if (s->bh == NULL) {
-	s->bh = qemu_bh_new(qemu_chr_reset_bh, s);
+	s->bh = qemu_bh_new(qemu_chr_generic_open_bh, s);
 	qemu_bh_schedule(s->bh);
     }
 }
@@ -610,7 +610,7 @@ static CharDriverState *qemu_chr_open_fd(int fd_in, int fd_out)
     chr->chr_update_read_handler = fd_chr_update_read_handler;
     chr->chr_close = fd_chr_close;
 
-    qemu_chr_reset(chr);
+    qemu_chr_generic_open(chr);
 
     return chr;
 }
@@ -910,7 +910,7 @@ static void pty_chr_state(CharDriverState *chr, int connected)
         qemu_mod_timer(s->timer, qemu_get_clock(rt_clock) + 1000);
     } else {
         if (!s->connected)
-            qemu_chr_reset(chr);
+            qemu_chr_generic_open(chr);
         s->connected = 1;
     }
 }
@@ -1184,7 +1184,7 @@ static CharDriverState *qemu_chr_open_tty(QemuOpts *opts)
         return NULL;
     }
     chr->chr_ioctl = tty_serial_ioctl;
-    qemu_chr_reset(chr);
+    qemu_chr_generic_open(chr);
     return chr;
 }
 #else  /* ! __linux__ && ! __sun__ */
@@ -1330,7 +1330,7 @@ static CharDriverState *qemu_chr_open_pp(QemuOpts *opts)
     chr->chr_close = pp_close;
     chr->opaque = drv;
 
-    qemu_chr_reset(chr);
+    qemu_chr_generic_open(chr);
 
     return chr;
 }
@@ -1611,7 +1611,7 @@ static CharDriverState *qemu_chr_open_win(QemuOpts *opts)
         free(chr);
         return NULL;
     }
-    qemu_chr_reset(chr);
+    qemu_chr_generic_open(chr);
     return chr;
 }
 
@@ -1711,7 +1711,7 @@ static CharDriverState *qemu_chr_open_win_pipe(QemuOpts *opts)
         free(chr);
         return NULL;
     }
-    qemu_chr_reset(chr);
+    qemu_chr_generic_open(chr);
     return chr;
 }
 
@@ -1725,7 +1725,7 @@ static CharDriverState *qemu_chr_open_win_file(HANDLE fd_out)
     s->hcom = fd_out;
     chr->opaque = s;
     chr->chr_write = win_chr_write;
-    qemu_chr_reset(chr);
+    qemu_chr_generic_open(chr);
     return chr;
 }
 
@@ -2052,7 +2052,7 @@ static void tcp_chr_connect(void *opaque)
     s->connected = 1;
     qemu_set_fd_handler2(s->fd, tcp_chr_read_poll,
                          tcp_chr_read, NULL, chr);
-    qemu_chr_reset(chr);
+    qemu_chr_generic_open(chr);
 }
 
 #define IACSET(x,a,b,c) x[0] = a; x[1] = b; x[2] = c;
qemu-char.h +1 -1
index b420111..9957db1 100644
@@ -82,7 +82,7 @@ void qemu_chr_add_handlers(CharDriverState *s,
                            IOEventHandler *fd_event,
                            void *opaque);
 int qemu_chr_ioctl(CharDriverState *s, int cmd, void *arg);
-void qemu_chr_reset(CharDriverState *s);
+void qemu_chr_generic_open(CharDriverState *s);
 int qemu_chr_can_read(CharDriverState *s);
 void qemu_chr_read(CharDriverState *s, uint8_t *buf, int len);
 int qemu_chr_get_msgfd(CharDriverState *s);
Signature · 1 line
1.6.2.5
#4 lore raw

[Qemu-devel] Re: [PATCH 2/3] char: Remove special init_reset handling

In reply to Amit Shah
Quoted 3 lines · Amit Shah

The initial_reset sent to chardevs doesn't do much other than setting a bool to true. Char devices are interested in the open event and that gets sent whenever the device is opened.

Have you checked with the monitor in all use cases (dedicated & muxed console, stdio & SDL console, etc.)? It was introduced once to fix a corner case, I think it's even documented...

Jan

Quoted 55 lines

Moreover, the reset logic breaks as and when qemu's bh scheduling changes.

---
 qemu-char.c |    9 ---------
 qemu-char.h |    1 -
 vl.c        |    1 -
 3 files changed, 0 insertions(+), 11 deletions(-)
qemu-char.c +0 -9
index 1f63019..d78bae3 100644
@@ -128,15 +128,6 @@ void qemu_chr_reset(CharDriverState *s)
     }
 }
 
-void qemu_chr_initial_reset(void)
-{
-    CharDriverState *chr;
-
-    QTAILQ_FOREACH(chr, &chardevs, next) {
-        qemu_chr_reset(chr);
-    }
-}
-
 int qemu_chr_write(CharDriverState *s, const uint8_t *buf, int len)
 {
     return s->chr_write(s, buf, len);
qemu-char.h +0 -1
index 05fe15d..b420111 100644
@@ -83,7 +83,6 @@ void qemu_chr_add_handlers(CharDriverState *s,
                            void *opaque);
 int qemu_chr_ioctl(CharDriverState *s, int cmd, void *arg);
 void qemu_chr_reset(CharDriverState *s);
-void qemu_chr_initial_reset(void);
 int qemu_chr_can_read(CharDriverState *s);
 void qemu_chr_read(CharDriverState *s, uint8_t *buf, int len);
 int qemu_chr_get_msgfd(CharDriverState *s);
vl.c +0 -1
index e57f58f..5ad4e42 100644
@@ -5750,7 +5750,6 @@ int main(int argc, char **argv, char **envp)
     }
 
     text_consoles_set_display(display_state);
-    qemu_chr_initial_reset();
 
     for (i = 0; i < MAX_MONITOR_DEVICES; i++) {
         if (monitor_devices[i] && monitor_hds[i]) {
author
#5 lore raw

[Qemu-devel] Re: [PATCH 2/3] char: Remove special init_reset handling

Quoted 7 lines · Jan Kiszka
Quoted 3 lines · Amit Shah

The initial_reset sent to chardevs doesn't do much other than setting a bool to true. Char devices are interested in the open event and that gets sent whenever the device is opened.

Have you checked with the monitor in all use cases (dedicated & muxed console, stdio & SDL console, etc.)? It was introduced once to fix a

I've checked with -monitor stdio, monitor in SDL and also chardevs using unix sockets.

I've not tried mux yet; I'll try that and report back. BTW if it ends up not working with this patch, it'd be broken in the current master as well.

Quoted 1 line

corner case, I think it's even documented...

Hm; I'll have to look really deep; haven't found any such thing yet.

Amit

author
#6 lore raw

[Qemu-devel] Re: [PATCH 2/3] char: Remove special init_reset handling

Quoted 9 lines · Amit Shah
Quoted 7 lines · Jan Kiszka
Quoted 3 lines · Amit Shah

The initial_reset sent to chardevs doesn't do much other than setting a bool to true. Char devices are interested in the open event and that gets sent whenever the device is opened.

Have you checked with the monitor in all use cases (dedicated & muxed console, stdio & SDL console, etc.)? It was introduced once to fix a

Quoted 6 lines

I've checked with -monitor stdio, monitor in SDL and also chardevs using unix sockets.

I've not tried mux yet; I'll try that and report back. BTW if it ends up not working with this patch, it'd be broken in the current master as well.

I tried with:

-chardev stdio,mux=on,id=mux -monitor chardev:mux -serial chardev:mux

The monitor prompt shows up as does the serial output.

(btw I've also tried closing and opening char devs and that works fine too)

Amit

#7 lore raw

[Qemu-devel] Re: [PATCH 2/3] char: Remove special init_reset handling

Quoted 15 lines · Amit Shah
Quoted 5 lines · Amit Shah
Quoted 4 lines · Jan Kiszka
Quoted 3 lines · Amit Shah

The initial_reset sent to chardevs doesn't do much other than setting a bool to true. Char devices are interested in the open event and that gets sent whenever the device is opened.

Quoted 8 lines
Quoted 2 lines

Have you checked with the monitor in all use cases (dedicated & muxed console, stdio & SDL console, etc.)? It was introduced once to fix a

I've checked with -monitor stdio, monitor in SDL and also chardevs using unix sockets.

I've not tried mux yet; I'll try that and report back. BTW if it ends up not working with this patch, it'd be broken in the current master as well.

Quoted 9 lines

I tried with:

-chardev stdio,mux=on,id=mux -monitor chardev:mux -serial chardev:mux

The monitor prompt shows up as does the serial output.

(btw I've also tried closing and opening char devs and that works fine too)

That sounds good. Then something must have changed since 2970a6c943, do you see what?

Jan

author
#8 lore raw

[Qemu-devel] Re: [PATCH 2/3] char: Remove special init_reset handling

Quoted 20 lines · Jan Kiszka
Quoted 19 lines · Amit Shah
Quoted 5 lines · Amit Shah
Quoted 4 lines · Jan Kiszka
Quoted 3 lines · Amit Shah

The initial_reset sent to chardevs doesn't do much other than setting a bool to true. Char devices are interested in the open event and that gets sent whenever the device is opened.

Quoted 8 lines
Quoted 2 lines

Have you checked with the monitor in all use cases (dedicated & muxed console, stdio & SDL console, etc.)? It was introduced once to fix a

I've checked with -monitor stdio, monitor in SDL and also chardevs using unix sockets.

I've not tried mux yet; I'll try that and report back. BTW if it ends up not working with this patch, it'd be broken in the current master as well.

I tried with:

-chardev stdio,mux=on,id=mux -monitor chardev:mux -serial chardev:mux

Quoted 7 lines
Quoted 4 lines

The monitor prompt shows up as does the serial output.

(btw I've also tried closing and opening char devs and that works fine too)

That sounds good. Then something must have changed since 2970a6c943, do you see what?

I think that depended on the resets being sent. I've now removed the need for resets altogether.

Amit

#9 lore raw

[Qemu-devel] Re: [PATCH 2/3] char: Remove special init_reset handling

Quoted 22 lines · Amit Shah
Quoted 21 lines · Jan Kiszka
Quoted 20 lines · Amit Shah
Quoted 5 lines · Amit Shah
Quoted 4 lines · Jan Kiszka
Quoted 3 lines · Amit Shah

The initial_reset sent to chardevs doesn't do much other than setting a bool to true. Char devices are interested in the open event and that gets sent whenever the device is opened.

Quoted 8 lines
Quoted 2 lines

Have you checked with the monitor in all use cases (dedicated & muxed console, stdio & SDL console, etc.)? It was introduced once to fix a

I've checked with -monitor stdio, monitor in SDL and also chardevs using unix sockets.

I've not tried mux yet; I'll try that and report back. BTW if it ends up not working with this patch, it'd be broken in the current master as well.

I tried with:

-chardev stdio,mux=on,id=mux -monitor chardev:mux -serial chardev:mux

The monitor prompt shows up as does the serial output.

Quoted 7 lines
Quoted 4 lines
Quoted 2 lines

(btw I've also tried closing and opening char devs and that works fine too)

That sounds good. Then something must have changed since 2970a6c943, do you see what?

I think that depended on the resets being sent. I've now removed the need for resets altogether.

No, this is in fact the reason why we no longer need it:

9a1e948129 (Introduce contexts for asynchronous callbacks)

As the initial reset of the char device that is marked pending on open is now no longer consumed by the IDE initialization, we can actually drop the later regeneration via qemu_chr_initial_reset. I just hope this stays like it is...

Jan

author
#10 lore raw

[Qemu-devel] Re: [PATCH 2/3] char: Remove special init_reset handling

Quoted 31 lines · Jan Kiszka
Quoted 22 lines · Amit Shah
Quoted 21 lines · Jan Kiszka
Quoted 20 lines · Amit Shah
Quoted 5 lines · Amit Shah
Quoted 4 lines · Jan Kiszka
Quoted 3 lines · Amit Shah

The initial_reset sent to chardevs doesn't do much other than setting a bool to true. Char devices are interested in the open event and that gets sent whenever the device is opened.

Quoted 8 lines
Quoted 2 lines

Have you checked with the monitor in all use cases (dedicated & muxed console, stdio & SDL console, etc.)? It was introduced once to fix a

I've checked with -monitor stdio, monitor in SDL and also chardevs using unix sockets.

I've not tried mux yet; I'll try that and report back. BTW if it ends up not working with this patch, it'd be broken in the current master as well.

I tried with:

-chardev stdio,mux=on,id=mux -monitor chardev:mux -serial chardev:mux

The monitor prompt shows up as does the serial output.

Quoted 7 lines
Quoted 4 lines
Quoted 2 lines

(btw I've also tried closing and opening char devs and that works fine too)

That sounds good. Then something must have changed since 2970a6c943, do you see what?

I think that depended on the resets being sent. I've now removed the need for resets altogether.

Quoted 8 lines

No, this is in fact the reason why we no longer need it:

9a1e948129 (Introduce contexts for asynchronous callbacks)

As the initial reset of the char device that is marked pending on open is now no longer consumed by the IDE initialization, we can actually drop the later regeneration via qemu_chr_initial_reset. I just hope this stays like it is...

I tested this even on a tree that doesn't have this patch.

I haven't really delved deep to see why this was added earlier -- the commit log only says very little. Plus my testing with the current tree works fine so I'm happy to mention these things in the commit log.

Amit

#11 lore raw

[Qemu-devel] Re: [PATCH 2/3] char: Remove special init_reset handling

Quoted 32 lines · Amit Shah
Quoted 31 lines · Jan Kiszka
Quoted 20 lines · Amit Shah
Quoted 19 lines · Jan Kiszka
Quoted 18 lines · Amit Shah
Quoted 5 lines · Amit Shah
Quoted 4 lines · Jan Kiszka
Quoted 3 lines · Amit Shah

The initial_reset sent to chardevs doesn't do much other than setting a bool to true. Char devices are interested in the open event and that gets sent whenever the device is opened.

Quoted 8 lines
Quoted 2 lines

Have you checked with the monitor in all use cases (dedicated & muxed console, stdio & SDL console, etc.)? It was introduced once to fix a

I've checked with -monitor stdio, monitor in SDL and also chardevs using unix sockets.

I've not tried mux yet; I'll try that and report back. BTW if it ends up not working with this patch, it'd be broken in the current master as well.

I tried with:

-chardev stdio,mux=on,id=mux -monitor chardev:mux -serial chardev:mux

Quoted 8 lines
Quoted 6 lines
Quoted 4 lines

The monitor prompt shows up as does the serial output.

(btw I've also tried closing and opening char devs and that works fine too)

That sounds good. Then something must have changed since 2970a6c943, do you see what?

I think that depended on the resets being sent. I've now removed the need for resets altogether.

No, this is in fact the reason why we no longer need it:

Quoted 8 lines
Quoted 6 lines

9a1e948129 (Introduce contexts for asynchronous callbacks)

As the initial reset of the char device that is marked pending on open is now no longer consumed by the IDE initialization, we can actually drop the later regeneration via qemu_chr_initial_reset. I just hope this stays like it is...

I tested this even on a tree that doesn't have this patch.

The whole things was (and maybe still is) fragile. So you have to take care what further patches with side effects you include. If you revert 9a1e948129 and remove qemu_chr_initial_reset, you get the old bug back again, ie. no prompt.

Quoted 4 lines

I haven't really delved deep to see why this was added earlier -- the commit log only says very little. Plus my testing with the current tree works fine so I'm happy to mention these things in the commit log.

I've nothing against your patch, specifically as it removes an obviously no longer needed workaround, not a feature. I just want to make sure that the current behavior is not only there by chance.

Jan

author
#12 lore raw

[Qemu-devel] Re: [PATCH 2/3] char: Remove special init_reset handling

Quoted 4 lines · Jan Kiszka

I've nothing against your patch, specifically as it removes an obviously no longer needed workaround, not a feature. I just want to make sure that the current behavior is not only there by chance.

Sure; I understand that. I'll be looking more at this code to free it from any surprises.

Amit