[Qemu-devel] [PATCH-for-4.1? 0/7] vl: Allow building with CONFIG_BLUETOOTH disabled

[Qemu-devel] [PATCH-for-4.1? 0/7] vl: Allow building with CONFIG_BLUETOOTH disabled

[Philippe Mathieu-Daudé]

A series of obvious patches to build without the deprecated
bluetooth devices. Still worth for 4.1 or too late?
It is clearly not a bugfix.

Regards,

Phil.

Philippe Mathieu-Daudé (7):
  hw/arm: Nokia N-series tablet requires Bluetooth
  hw/usb: Bluetooth HCI USB depends on USB & BLUETOOTH
  MAINTAINERS: Add a Bluetooth entry
  vl: Fix 'braces' coding style issues
  vl: Use qemu_strtoi() instead of strtol()
  vl: Extract bt_parse() into its own file
  hw/bt: Allow building with CONFIG_BLUETOOTH disabled

 MAINTAINERS         |   7 +++
 Makefile.objs       |   3 +-
 bt-opts.c           | 140 ++++++++++++++++++++++++++++++++++++++++++++
 bt-stubs.c          |  18 ++++++
 hw/arm/Kconfig      |   1 +
 hw/bt/Makefile.objs |   4 +-
 hw/usb/Kconfig      |   2 +-
 include/sysemu/bt.h |   3 +
 vl.c                | 121 --------------------------------------
 9 files changed, 174 insertions(+), 125 deletions(-)
 create mode 100644 bt-opts.c
 create mode 100644 bt-stubs.c

-- 
2.20.1

[Qemu-devel] [PATCH-for-4.1? 1/7] hw/arm: Nokia N-series tablet requires Bluetooth

[Philippe Mathieu-Daudé]

Missed in 4826ac6d192 because the Bluetooth devices are
currently always enabled.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
---
 hw/arm/Kconfig | 1 +
 1 file changed, 1 insertion(+)

diff --git a/hw/arm/Kconfig b/hw/arm/Kconfig
index ab65ecd216..11195251f3 100644
--- a/hw/arm/Kconfig
+++ b/hw/arm/Kconfig
@@ -106,6 +106,7 @@ config NSERIES
     select LM832X   # GPIO keyboard chip
     select TWL92230 # energy-management
     select TUSB6010
+    select BLUETOOTH
 
 config OMAP
     bool
-- 
2.20.1

Re: [Qemu-devel] [PATCH-for-4.1? 1/7] hw/arm: Nokia N-series tablet requires Bluetooth

[Thomas Huth]

On 12/07/2019 15.39, Philippe Mathieu-Daudé wrote:
> Missed in 4826ac6d192 because the Bluetooth devices are
> currently always enabled.
> 
> Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
> ---
>  hw/arm/Kconfig | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/hw/arm/Kconfig b/hw/arm/Kconfig
> index ab65ecd216..11195251f3 100644
> --- a/hw/arm/Kconfig
> +++ b/hw/arm/Kconfig
> @@ -106,6 +106,7 @@ config NSERIES
>      select LM832X   # GPIO keyboard chip
>      select TWL92230 # energy-management
>      select TUSB6010
> +    select BLUETOOTH

Looking at hw/bt/Makefile.objs the BLUETOOTH switch seems to be
completely unused?

 Thomas

[Qemu-devel] [PATCH-for-4.1 2/7] hw/usb: Bluetooth HCI USB depends on USB & BLUETOOTH

[Philippe Mathieu-Daudé]

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
---
 hw/usb/Kconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/hw/usb/Kconfig b/hw/usb/Kconfig
index 564305e283..1b435ec002 100644
--- a/hw/usb/Kconfig
+++ b/hw/usb/Kconfig
@@ -82,7 +82,7 @@ config USB_NETWORK
 config USB_BLUETOOTH
     bool
     default y
-    depends on USB
+    depends on USB && BLUETOOTH
 
 config USB_SMARTCARD
     bool
-- 
2.20.1

Re: [Qemu-devel] [PATCH-for-4.1 2/7] hw/usb: Bluetooth HCI USB depends on USB & BLUETOOTH

[Paolo Bonzini]

On 12/07/19 15:39, Philippe Mathieu-Daudé wrote:
> Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
> ---
>  hw/usb/Kconfig | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/hw/usb/Kconfig b/hw/usb/Kconfig
> index 564305e283..1b435ec002 100644
> --- a/hw/usb/Kconfig
> +++ b/hw/usb/Kconfig
> @@ -82,7 +82,7 @@ config USB_NETWORK
>  config USB_BLUETOOTH
>      bool
>      default y
> -    depends on USB
> +    depends on USB && BLUETOOTH
>  
>  config USB_SMARTCARD
>      bool
> 

Shouldn't it select BLUETOOTH instead?

Paolo

Re: [Qemu-devel] [PATCH-for-4.1 2/7] hw/usb: Bluetooth HCI USB depends on USB & BLUETOOTH

[Philippe Mathieu-Daudé]

On 7/12/19 3:58 PM, Paolo Bonzini wrote:
> On 12/07/19 15:39, Philippe Mathieu-Daudé wrote:
>> Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
>> ---
>>  hw/usb/Kconfig | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/hw/usb/Kconfig b/hw/usb/Kconfig
>> index 564305e283..1b435ec002 100644
>> --- a/hw/usb/Kconfig
>> +++ b/hw/usb/Kconfig
>> @@ -82,7 +82,7 @@ config USB_NETWORK
>>  config USB_BLUETOOTH
>>      bool
>>      default y
>> -    depends on USB
>> +    depends on USB && BLUETOOTH
>>  
>>  config USB_SMARTCARD
>>      bool
>>
> 
> Shouldn't it select BLUETOOTH instead?

I wondered but am not sure, it is default to Yes, but Bluetooth code is
deprecated and print a big warning when you use it, so I think this
device should not be selected as default, neither pull in the deprecated
bluetooth code.

So personally I'd respin without 'default y'. If you have a good reason
to use it, I can use 'select' instead.

Regards,

Phil.

Re: [Qemu-devel] [PATCH-for-4.1 2/7] hw/usb: Bluetooth HCI USB depends on USB & BLUETOOTH

[Paolo Bonzini]

On 12/07/19 16:16, Philippe Mathieu-Daudé wrote:
> I wondered but am not sure, it is default to Yes, but Bluetooth code is
> deprecated and print a big warning when you use it, so I think this
> device should not be selected as default, neither pull in the deprecated
> bluetooth code.
> 
> So personally I'd respin without 'default y'. If you have a good reason
> to use it, I can use 'select' instead.

There are two separate questions.  One is whether USB_BLUETOOTH should
select BLUETOOTH and the answer there is almost certainly yes (in the
words of kconfig.rst, BLUETOOTH is a "subsystem" while USB_BLUETOOTH is
a "device"; devices select the bus that the device provides).

The other is whether we want to enable USB_BLUETOOTH by default.  I
wouldn't have any problem there, but if we disable it basically no one
would ship/use it and we might as well delete the whole thing.

Paolo

Re: [Qemu-devel] [PATCH-for-4.1 2/7] hw/usb: Bluetooth HCI USB depends on USB & BLUETOOTH

[Philippe Mathieu-Daudé]

On 7/12/19 4:58 PM, Paolo Bonzini wrote:
> On 12/07/19 16:16, Philippe Mathieu-Daudé wrote:
>> I wondered but am not sure, it is default to Yes, but Bluetooth code is
>> deprecated and print a big warning when you use it, so I think this
>> device should not be selected as default, neither pull in the deprecated
>> bluetooth code.
>>
>> So personally I'd respin without 'default y'. If you have a good reason
>> to use it, I can use 'select' instead.
> 
> There are two separate questions.  One is whether USB_BLUETOOTH should
> select BLUETOOTH and the answer there is almost certainly yes (in the
> words of kconfig.rst, BLUETOOTH is a "subsystem" while USB_BLUETOOTH is
> a "device"; devices select the bus that the device provides).

Fair :)

> The other is whether we want to enable USB_BLUETOOTH by default.  I
> wouldn't have any problem there, but if we disable it basically no one
> would ship/use it and we might as well delete the whole thing.

The only user is the ARM Nokia N-series board, so if we set default=n,
the Bluetooth subsystem will be only be selected on arm-softmmu (and
aarch64-softmmu).

This seems a sane cleanup. If another board wants to use the bluetooth
code, it should probably move it out of the orphan status.

Re: [Qemu-devel] [PATCH-for-4.1 2/7] hw/usb: Bluetooth HCI USB depends on USB & BLUETOOTH

[Paolo Bonzini]

On 12/07/19 18:45, Philippe Mathieu-Daudé wrote:
> On 7/12/19 4:58 PM, Paolo Bonzini wrote:
>> The other is whether we want to enable USB_BLUETOOTH by default.  I
>> wouldn't have any problem there, but if we disable it basically no one
>> would ship/use it and we might as well delete the whole thing.
> 
> The only user is the ARM Nokia N-series board, so if we set default=n,
> the Bluetooth subsystem will be only be selected on arm-softmmu (and
> aarch64-softmmu).
> 
> This seems a sane cleanup. If another board wants to use the bluetooth
> code, it should probably move it out of the orphan status.

Fair! ;)

Paolo

Re: [Qemu-devel] [PATCH-for-4.1 2/7] hw/usb: Bluetooth HCI USB depends on USB & BLUETOOTH

[Philippe Mathieu-Daudé]

Hi Paolo,

On 7/12/19 7:31 PM, Paolo Bonzini wrote:
> On 12/07/19 18:45, Philippe Mathieu-Daudé wrote:
>> On 7/12/19 4:58 PM, Paolo Bonzini wrote:
>>> The other is whether we want to enable USB_BLUETOOTH by default.  I
>>> wouldn't have any problem there, but if we disable it basically no one
>>> would ship/use it and we might as well delete the whole thing.
>>
>> The only user is the ARM Nokia N-series board, so if we set default=n,
>> the Bluetooth subsystem will be only be selected on arm-softmmu (and
>> aarch64-softmmu).

Using (1):

+    default y if BLUETOOTH
     depends on USB
+    select BLUETOOTH

I get:

    KconfigDataError: cycle found including BLUETOOTH

This works but doesn't follow kconfig.rst (2):

+    default y if BLUETOOTH
     depends on USB

This works (kconfig.rst compliant, 3):

+    default n
     depends on USB
+    select BLUETOOTH

Are you OK with (2) or you prefer (3)?

>>
>> This seems a sane cleanup. If another board wants to use the bluetooth
>> code, it should probably move it out of the orphan status.
> 
> Fair! ;)
> 
> Paolo
> 

[Qemu-devel] [PATCH-for-4.1? 3/7] MAINTAINERS: Add an entry for the Bluetooth devices

[Philippe Mathieu-Daudé]

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
---
 MAINTAINERS | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/MAINTAINERS b/MAINTAINERS
index cc9636b43a..5d8f27d9bd 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -1733,6 +1733,13 @@ F: hw/*/*xive*
 F: include/hw/*/*xive*
 F: docs/*/*xive*
 
+Bluetooth
+S: Orphan
+F: bt-*.c
+F: hw/bt/
+F: include/hw/bt.h
+F: include/sysemu/bt.h
+
 Subsystems
 ----------
 Audio
-- 
2.20.1

Re: [Qemu-devel] [PATCH-for-4.1? 3/7] MAINTAINERS: Add an entry for the Bluetooth devices

[Philippe Mathieu-Daudé]

Cc'ing qemu-trivial@ to salvage this patch.

On 7/12/19 3:39 PM, Philippe Mathieu-Daudé wrote:
> Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
> ---
>  MAINTAINERS | 7 +++++++
>  1 file changed, 7 insertions(+)
> 
> diff --git a/MAINTAINERS b/MAINTAINERS
> index cc9636b43a..5d8f27d9bd 100644
> --- a/MAINTAINERS
> +++ b/MAINTAINERS
> @@ -1733,6 +1733,13 @@ F: hw/*/*xive*
>  F: include/hw/*/*xive*
>  F: docs/*/*xive*
>  
> +Bluetooth
> +S: Orphan
> +F: bt-*.c
> +F: hw/bt/
> +F: include/hw/bt.h
> +F: include/sysemu/bt.h
> +
>  Subsystems
>  ----------
>  Audio
> 

[Qemu-devel] [PATCH-for-4.1? 4/7] vl: Fix 'braces' coding style issues

[Philippe Mathieu-Daudé]

Since we'll move this code around, fix its style first:

  ERROR: braces {} are necessary for all arms of this statement style

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
---
 vl.c | 27 ++++++++++++++++++---------
 1 file changed, 18 insertions(+), 9 deletions(-)

diff --git a/vl.c b/vl.c
index 5089fce6c5..96de4f74b9 100644
--- a/vl.c
+++ b/vl.c
@@ -920,8 +920,9 @@ static struct HCIInfo *hci_table[MAX_NICS];
 
 struct HCIInfo *qemu_next_hci(void)
 {
-    if (cur_hci == nb_hcis)
+    if (cur_hci == nb_hcis) {
         return &null_hci;
+    }
 
     return hci_table[cur_hci++];
 }
@@ -937,8 +938,9 @@ static int bt_hci_parse(const char *str)
     }
 
     hci = hci_init(str);
-    if (!hci)
+    if (!hci) {
         return -1;
+    }
 
     bdaddr.b[0] = 0x52;
     bdaddr.b[1] = 0x54;
@@ -957,9 +959,10 @@ static void bt_vhci_add(int vlan_id)
 {
     struct bt_scatternet_s *vlan = qemu_find_bt_vlan(vlan_id);
 
-    if (!vlan->slave)
+    if (!vlan->slave) {
         warn_report("adding a VHCI to an empty scatternet %i",
                     vlan_id);
+    }
 
     bt_vhci_init(bt_new_hci(vlan));
 }
@@ -984,12 +987,14 @@ static struct bt_device_s *bt_device_add(const char *opt)
 
     vlan = qemu_find_bt_vlan(vlan_id);
 
-    if (!vlan->slave)
+    if (!vlan->slave) {
         warn_report("adding a slave device to an empty scatternet %i",
                     vlan_id);
+    }
 
-    if (!strcmp(devname, "keyboard"))
+    if (!strcmp(devname, "keyboard")) {
         return bt_keyboard_init(vlan);
+    }
 
     error_report("unsupported bluetooth device '%s'", devname);
     return 0;
@@ -1002,9 +1007,11 @@ static int bt_parse(const char *opt)
 
     if (strstart(opt, "hci", &endp)) {
         if (!*endp || *endp == ',') {
-            if (*endp)
-                if (!strstart(endp, ",vlan=", 0))
+            if (*endp) {
+                if (!strstart(endp, ",vlan=", 0)) {
                     opt = endp + 1;
+                }
+            }
 
             return bt_hci_parse(opt);
        }
@@ -1021,14 +1028,16 @@ static int bt_parse(const char *opt)
                     error_report("bad parameter '%s'", endp + 1);
                     return 1;
                 }
-            } else
+            } else {
                 vlan = 0;
+            }
 
             bt_vhci_add(vlan);
             return 0;
         }
-    } else if (strstart(opt, "device:", &endp))
+    } else if (strstart(opt, "device:", &endp)) {
         return !bt_device_add(endp);
+    }
 
     error_report("bad bluetooth parameter '%s'", opt);
     return 1;
-- 
2.20.1

[Qemu-devel] [PATCH-for-4.1? 5/7] vl: Use qemu_strtoi() instead of strtol()

[Philippe Mathieu-Daudé]

Since we'll move this code around, fix this checkpatch warning
first:

  ERROR: consider using qemu_strtol in preference to strtol

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
---
 vl.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/vl.c b/vl.c
index 96de4f74b9..c7f4d01c99 100644
--- a/vl.c
+++ b/vl.c
@@ -978,8 +978,7 @@ static struct bt_device_s *bt_device_add(const char *opt)
     pstrcpy(devname, MIN(sizeof(devname), len), opt);
 
     if (endp) {
-        vlan_id = strtol(endp + 6, &endp, 0);
-        if (*endp) {
+        if (qemu_strtoi(endp + 6, (const char **)&endp, 0, &vlan_id) < 0) {
             error_report("unrecognised bluetooth vlan Id");
             return 0;
         }
@@ -1019,8 +1018,7 @@ static int bt_parse(const char *opt)
         if (!*endp || *endp == ',') {
             if (*endp) {
                 if (strstart(endp, ",vlan=", &p)) {
-                    vlan = strtol(p, (char **) &endp, 0);
-                    if (*endp) {
+                    if (qemu_strtoi(p, &endp, 0, &vlan) < 0) {
                         error_report("bad scatternet '%s'", p);
                         return 1;
                     }
-- 
2.20.1

[Qemu-devel] [PATCH-for-4.1? 6/7] vl: Extract bt_parse() into its own file

[Philippe Mathieu-Daudé]

This is required to be able to compile vl.c with CONFIG_BLUETOOTH
disabled.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
---
2300 lines less for the C preprocessor :)
---
 Makefile.objs       |   2 +-
 bt-opts.c           | 140 ++++++++++++++++++++++++++++++++++++++++++++
 include/sysemu/bt.h |   3 +
 vl.c                | 128 ----------------------------------------
 4 files changed, 144 insertions(+), 129 deletions(-)
 create mode 100644 bt-opts.c

diff --git a/Makefile.objs b/Makefile.objs
index 6a143dcd57..c2845a0efc 100644
--- a/Makefile.objs
+++ b/Makefile.objs
@@ -65,7 +65,7 @@ common-obj-y += replay/
 
 common-obj-y += ui/
 common-obj-m += ui/
-common-obj-y += bt-host.o bt-vhci.o
+common-obj-y += bt-host.o bt-vhci.o bt-opts.o
 bt-host.o-cflags := $(BLUEZ_CFLAGS)
 
 common-obj-y += dma-helpers.o
diff --git a/bt-opts.c b/bt-opts.c
new file mode 100644
index 0000000000..d14ae4836f
--- /dev/null
+++ b/bt-opts.c
@@ -0,0 +1,140 @@
+/*
+ * Bluetooth command line options.
+ *
+ * Copyright (C) 2008 Andrzej Zaborowski
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+
+#include "qemu/osdep.h"
+#include "qemu/error-report.h"
+#include "qemu/cutils.h"
+#include "sysemu/bt.h"
+#include "net/net.h"
+#include "hw/bt.h"
+
+static int nb_hcis;
+static int cur_hci;
+static struct HCIInfo *hci_table[MAX_NICS];
+
+struct HCIInfo *qemu_next_hci(void)
+{
+    if (cur_hci == nb_hcis) {
+        return &null_hci;
+    }
+
+    return hci_table[cur_hci++];
+}
+
+static int bt_hci_parse(const char *str)
+{
+    struct HCIInfo *hci;
+    bdaddr_t bdaddr;
+
+    if (nb_hcis >= MAX_NICS) {
+        error_report("too many bluetooth HCIs (max %i)", MAX_NICS);
+        return -1;
+    }
+
+    hci = hci_init(str);
+    if (!hci) {
+        return -1;
+    }
+
+    bdaddr.b[0] = 0x52;
+    bdaddr.b[1] = 0x54;
+    bdaddr.b[2] = 0x00;
+    bdaddr.b[3] = 0x12;
+    bdaddr.b[4] = 0x34;
+    bdaddr.b[5] = 0x56 + nb_hcis;
+    hci->bdaddr_set(hci, bdaddr.b);
+
+    hci_table[nb_hcis++] = hci;
+
+    return 0;
+}
+
+static void bt_vhci_add(int vlan_id)
+{
+    struct bt_scatternet_s *vlan = qemu_find_bt_vlan(vlan_id);
+
+    if (!vlan->slave) {
+        warn_report("adding a VHCI to an empty scatternet %i", vlan_id);
+    }
+
+    bt_vhci_init(bt_new_hci(vlan));
+}
+
+static struct bt_device_s *bt_device_add(const char *opt)
+{
+    struct bt_scatternet_s *vlan;
+    int vlan_id = 0;
+    char *endp = strstr(opt, ",vlan=");
+    int len = (endp ? endp - opt : strlen(opt)) + 1;
+    char devname[10];
+
+    pstrcpy(devname, MIN(sizeof(devname), len), opt);
+
+    if (endp) {
+        if (qemu_strtoi(endp + 6, (const char **)&endp, 0, &vlan_id) < 0) {
+            error_report("unrecognised bluetooth vlan Id");
+            return 0;
+        }
+    }
+
+    vlan = qemu_find_bt_vlan(vlan_id);
+
+    if (!vlan->slave) {
+        warn_report("adding a slave device to an empty scatternet %i",
+                    vlan_id);
+    }
+
+    if (!strcmp(devname, "keyboard")) {
+        return bt_keyboard_init(vlan);
+    }
+
+    error_report("unsupported bluetooth device '%s'", devname);
+    return 0;
+}
+
+int bt_parse(const char *opt)
+{
+    const char *endp, *p;
+    int vlan;
+
+    if (strstart(opt, "hci", &endp)) {
+        if (!*endp || *endp == ',') {
+            if (*endp) {
+                if (!strstart(endp, ",vlan=", 0)) {
+                    opt = endp + 1;
+                }
+            }
+
+            return bt_hci_parse(opt);
+       }
+    } else if (strstart(opt, "vhci", &endp)) {
+        if (!*endp || *endp == ',') {
+            if (*endp) {
+                if (strstart(endp, ",vlan=", &p)) {
+                    if (qemu_strtoi(p, &endp, 0, &vlan) < 0) {
+                        error_report("bad scatternet '%s'", p);
+                        return 1;
+                    }
+                } else {
+                    error_report("bad parameter '%s'", endp + 1);
+                    return 1;
+                }
+            } else {
+                vlan = 0;
+            }
+
+            bt_vhci_add(vlan);
+            return 0;
+        }
+    } else if (strstart(opt, "device:", &endp)) {
+        return !bt_device_add(endp);
+    }
+
+    error_report("bad bluetooth parameter '%s'", opt);
+    return 1;
+}
diff --git a/include/sysemu/bt.h b/include/sysemu/bt.h
index 2fd8c0f14b..5b73709c2d 100644
--- a/include/sysemu/bt.h
+++ b/include/sysemu/bt.h
@@ -17,4 +17,7 @@ typedef struct HCIInfo {
 struct HCIInfo *bt_host_hci(const char *id);
 struct HCIInfo *qemu_next_hci(void);
 
+/* TODO Convert bt-opts.c to opts_init() and remove from here. */
+int bt_parse(const char *opt);
+
 #endif
diff --git a/vl.c b/vl.c
index c7f4d01c99..3a6dd3ba2b 100644
--- a/vl.c
+++ b/vl.c
@@ -60,7 +60,6 @@ int main(int argc, char **argv)
 #include "hw/isa/isa.h"
 #include "hw/scsi/scsi.h"
 #include "hw/display/vga.h"
-#include "hw/bt.h"
 #include "sysemu/watchdog.h"
 #include "hw/firmware/smbios.h"
 #include "hw/acpi/acpi.h"
@@ -913,133 +912,6 @@ static void configure_rtc(QemuOpts *opts)
 }
 
 /***********************************************************/
-/* Bluetooth support */
-static int nb_hcis;
-static int cur_hci;
-static struct HCIInfo *hci_table[MAX_NICS];
-
-struct HCIInfo *qemu_next_hci(void)
-{
-    if (cur_hci == nb_hcis) {
-        return &null_hci;
-    }
-
-    return hci_table[cur_hci++];
-}
-
-static int bt_hci_parse(const char *str)
-{
-    struct HCIInfo *hci;
-    bdaddr_t bdaddr;
-
-    if (nb_hcis >= MAX_NICS) {
-        error_report("too many bluetooth HCIs (max %i)", MAX_NICS);
-        return -1;
-    }
-
-    hci = hci_init(str);
-    if (!hci) {
-        return -1;
-    }
-
-    bdaddr.b[0] = 0x52;
-    bdaddr.b[1] = 0x54;
-    bdaddr.b[2] = 0x00;
-    bdaddr.b[3] = 0x12;
-    bdaddr.b[4] = 0x34;
-    bdaddr.b[5] = 0x56 + nb_hcis;
-    hci->bdaddr_set(hci, bdaddr.b);
-
-    hci_table[nb_hcis++] = hci;
-
-    return 0;
-}
-
-static void bt_vhci_add(int vlan_id)
-{
-    struct bt_scatternet_s *vlan = qemu_find_bt_vlan(vlan_id);
-
-    if (!vlan->slave) {
-        warn_report("adding a VHCI to an empty scatternet %i",
-                    vlan_id);
-    }
-
-    bt_vhci_init(bt_new_hci(vlan));
-}
-
-static struct bt_device_s *bt_device_add(const char *opt)
-{
-    struct bt_scatternet_s *vlan;
-    int vlan_id = 0;
-    char *endp = strstr(opt, ",vlan=");
-    int len = (endp ? endp - opt : strlen(opt)) + 1;
-    char devname[10];
-
-    pstrcpy(devname, MIN(sizeof(devname), len), opt);
-
-    if (endp) {
-        if (qemu_strtoi(endp + 6, (const char **)&endp, 0, &vlan_id) < 0) {
-            error_report("unrecognised bluetooth vlan Id");
-            return 0;
-        }
-    }
-
-    vlan = qemu_find_bt_vlan(vlan_id);
-
-    if (!vlan->slave) {
-        warn_report("adding a slave device to an empty scatternet %i",
-                    vlan_id);
-    }
-
-    if (!strcmp(devname, "keyboard")) {
-        return bt_keyboard_init(vlan);
-    }
-
-    error_report("unsupported bluetooth device '%s'", devname);
-    return 0;
-}
-
-static int bt_parse(const char *opt)
-{
-    const char *endp, *p;
-    int vlan;
-
-    if (strstart(opt, "hci", &endp)) {
-        if (!*endp || *endp == ',') {
-            if (*endp) {
-                if (!strstart(endp, ",vlan=", 0)) {
-                    opt = endp + 1;
-                }
-            }
-
-            return bt_hci_parse(opt);
-       }
-    } else if (strstart(opt, "vhci", &endp)) {
-        if (!*endp || *endp == ',') {
-            if (*endp) {
-                if (strstart(endp, ",vlan=", &p)) {
-                    if (qemu_strtoi(p, &endp, 0, &vlan) < 0) {
-                        error_report("bad scatternet '%s'", p);
-                        return 1;
-                    }
-                } else {
-                    error_report("bad parameter '%s'", endp + 1);
-                    return 1;
-                }
-            } else {
-                vlan = 0;
-            }
-
-            bt_vhci_add(vlan);
-            return 0;
-        }
-    } else if (strstart(opt, "device:", &endp)) {
-        return !bt_device_add(endp);
-    }
-
-    error_report("bad bluetooth parameter '%s'", opt);
-    return 1;
-}
 
 static int parse_name(void *opaque, QemuOpts *opts, Error **errp)
 {
-- 
2.20.1

[Qemu-devel] [PATCH-for-4.1? 7/7] hw/bt: Allow building with CONFIG_BLUETOOTH disabled

[Philippe Mathieu-Daudé]

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
---
 Makefile.objs       |  3 ++-
 bt-stubs.c          | 18 ++++++++++++++++++
 hw/bt/Makefile.objs |  4 ++--
 3 files changed, 22 insertions(+), 3 deletions(-)
 create mode 100644 bt-stubs.c

diff --git a/Makefile.objs b/Makefile.objs
index c2845a0efc..98269e393a 100644
--- a/Makefile.objs
+++ b/Makefile.objs
@@ -65,8 +65,9 @@ common-obj-y += replay/
 
 common-obj-y += ui/
 common-obj-m += ui/
-common-obj-y += bt-host.o bt-vhci.o bt-opts.o
 bt-host.o-cflags := $(BLUEZ_CFLAGS)
+common-obj-$(CONFIG_BLUETOOTH) += bt-host.o bt-vhci.o bt-opts.o
+common-obj-$(call lnot,$(CONFIG_BLUETOOTH)) += bt-stubs.o
 
 common-obj-y += dma-helpers.o
 common-obj-y += vl.o
diff --git a/bt-stubs.c b/bt-stubs.c
new file mode 100644
index 0000000000..d3f5158d3b
--- /dev/null
+++ b/bt-stubs.c
@@ -0,0 +1,18 @@
+/*
+ * Bluetooth stubs.
+ *
+ * Copyright (c) 2019 Red Hat, Inc.
+ *
+ * SPDX-License-Identifier: GPL-2.0-or-later
+ */
+
+#include "qemu/osdep.h"
+#include "qemu/error-report.h"
+#include "sysemu/bt.h"
+
+int bt_parse(const char *opt)
+{
+    error_report("Bluetooth support is disabled");
+
+    return 1;
+}
diff --git a/hw/bt/Makefile.objs b/hw/bt/Makefile.objs
index 867a7d2e8a..46aec8e38e 100644
--- a/hw/bt/Makefile.objs
+++ b/hw/bt/Makefile.objs
@@ -1,3 +1,3 @@
-common-obj-y += core.o l2cap.o sdp.o hci.o hid.o
-common-obj-y += hci-csr.o
+common-obj-$(CONFIG_BLUETOOTH) += core.o l2cap.o sdp.o hci.o hid.o
+common-obj-$(CONFIG_BLUETOOTH) += hci-csr.o
 
-- 
2.20.1

Re: [Qemu-devel] [PATCH-for-4.1? 7/7] hw/bt: Allow building with CONFIG_BLUETOOTH disabled

[Thomas Huth]

On 12/07/2019 15.39, Philippe Mathieu-Daudé wrote:
> Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
> ---
>  Makefile.objs       |  3 ++-
>  bt-stubs.c          | 18 ++++++++++++++++++
>  hw/bt/Makefile.objs |  4 ++--
>  3 files changed, 22 insertions(+), 3 deletions(-)
>  create mode 100644 bt-stubs.c
> 
> diff --git a/Makefile.objs b/Makefile.objs
> index c2845a0efc..98269e393a 100644
> --- a/Makefile.objs
> +++ b/Makefile.objs
> @@ -65,8 +65,9 @@ common-obj-y += replay/
>  
>  common-obj-y += ui/
>  common-obj-m += ui/
> -common-obj-y += bt-host.o bt-vhci.o bt-opts.o
>  bt-host.o-cflags := $(BLUEZ_CFLAGS)
> +common-obj-$(CONFIG_BLUETOOTH) += bt-host.o bt-vhci.o bt-opts.o
> +common-obj-$(call lnot,$(CONFIG_BLUETOOTH)) += bt-stubs.o
>  
>  common-obj-y += dma-helpers.o
>  common-obj-y += vl.o
> diff --git a/bt-stubs.c b/bt-stubs.c
> new file mode 100644
> index 0000000000..d3f5158d3b
> --- /dev/null
> +++ b/bt-stubs.c
> @@ -0,0 +1,18 @@
> +/*
> + * Bluetooth stubs.
> + *
> + * Copyright (c) 2019 Red Hat, Inc.
> + *
> + * SPDX-License-Identifier: GPL-2.0-or-later
> + */
> +
> +#include "qemu/osdep.h"
> +#include "qemu/error-report.h"
> +#include "sysemu/bt.h"
> +
> +int bt_parse(const char *opt)
> +{
> +    error_report("Bluetooth support is disabled");
> +
> +    return 1;
> +}
> diff --git a/hw/bt/Makefile.objs b/hw/bt/Makefile.objs
> index 867a7d2e8a..46aec8e38e 100644
> --- a/hw/bt/Makefile.objs
> +++ b/hw/bt/Makefile.objs
> @@ -1,3 +1,3 @@
> -common-obj-y += core.o l2cap.o sdp.o hci.o hid.o
> -common-obj-y += hci-csr.o
> +common-obj-$(CONFIG_BLUETOOTH) += core.o l2cap.o sdp.o hci.o hid.o
> +common-obj-$(CONFIG_BLUETOOTH) += hci-csr.o

... ok, so the BLUETOOTH switch gets used here. But I wonder whether we
should have separate switches for the backends (i.e. the files here) and
the guest devices (i.e. the ones in hw/bt/) ?

 Thomas

Re: [Qemu-devel] [PATCH-for-4.1? 0/7] vl: Allow building with CONFIG_BLUETOOTH disabled

[Richard Henderson]

On 7/12/19 3:39 PM, Philippe Mathieu-Daudé wrote:
> A series of obvious patches to build without the deprecated
> bluetooth devices. Still worth for 4.1 or too late?
> It is clearly not a bugfix.
> 
> Regards,
> 
> Phil.
> 
> Philippe Mathieu-Daudé (7):
>   hw/arm: Nokia N-series tablet requires Bluetooth
>   hw/usb: Bluetooth HCI USB depends on USB & BLUETOOTH
>   MAINTAINERS: Add a Bluetooth entry
>   vl: Fix 'braces' coding style issues
>   vl: Use qemu_strtoi() instead of strtol()
>   vl: Extract bt_parse() into its own file
>   hw/bt: Allow building with CONFIG_BLUETOOTH disabled

All of this looks plausible to me to go along with the 4.1 deprecation.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>


r~

Re: [Qemu-devel] [PATCH-for-4.1? 0/7] vl: Allow building with CONFIG_BLUETOOTH disabled

[Thomas Huth]

On 12/07/2019 15.39, Philippe Mathieu-Daudé wrote:
> A series of obvious patches to build without the deprecated
> bluetooth devices. Still worth for 4.1 or too late?
> It is clearly not a bugfix.

I wonder whether this series is worth the effort right now, or whether
we should simply nuke the bluetooth code after 4.1 has been released?

 Thomas

Re: [Qemu-devel] [PATCH-for-4.1? 0/7] vl: Allow building with CONFIG_BLUETOOTH disabled

[Philippe Mathieu-Daudé]

On 7/15/19 3:13 PM, Thomas Huth wrote:
> On 12/07/2019 15.39, Philippe Mathieu-Daudé wrote:
>> A series of obvious patches to build without the deprecated
>> bluetooth devices. Still worth for 4.1 or too late?
>> It is clearly not a bugfix.
> 
> I wonder whether this series is worth the effort right now, or whether
> we should simply nuke the bluetooth code after 4.1 has been released?

Well, perfect is the enemy of good :)

This series is already done and is an improvement to what we have.

Regarding nuking it, it depends on the Nokia N-series boards, they might
become useless without BT support.

Peter, as a maintainer of these boards, what do you prefer?

Thanks,

Phil.

Re: [Qemu-devel] [PATCH-for-4.1? 0/7] vl: Allow building with CONFIG_BLUETOOTH disabled

[Peter Maydell]

On Tue, 13 Aug 2019 at 15:01, Philippe Mathieu-Daudé <philmd@redhat.com> wrote:
>
> On 7/15/19 3:13 PM, Thomas Huth wrote:
> > On 12/07/2019 15.39, Philippe Mathieu-Daudé wrote:
> >> A series of obvious patches to build without the deprecated
> >> bluetooth devices. Still worth for 4.1 or too late?
> >> It is clearly not a bugfix.
> >
> > I wonder whether this series is worth the effort right now, or whether
> > we should simply nuke the bluetooth code after 4.1 has been released?
>
> Well, perfect is the enemy of good :)
>
> This series is already done and is an improvement to what we have.
>
> Regarding nuking it, it depends on the Nokia N-series boards, they might
> become useless without BT support.

Er, they're not useless at all without BT support. The BT
hardware is a really tiny part that I doubt many users of
the board models ever used. As long as we retain a "simulate
doing nothing much" model of the BT device to show the guest
I don't care whether the BT backend code disappears.

thanks
-- PMM

Re: [Qemu-devel] [PATCH-for-4.1? 0/7] vl: Allow building with CONFIG_BLUETOOTH disabled

[Philippe Mathieu-Daudé]

On 8/13/19 4:04 PM, Peter Maydell wrote:
> On Tue, 13 Aug 2019 at 15:01, Philippe Mathieu-Daudé <philmd@redhat.com> wrote:
>> On 7/15/19 3:13 PM, Thomas Huth wrote:
>>> On 12/07/2019 15.39, Philippe Mathieu-Daudé wrote:
>>>> A series of obvious patches to build without the deprecated
>>>> bluetooth devices. Still worth for 4.1 or too late?
>>>> It is clearly not a bugfix.
>>>
>>> I wonder whether this series is worth the effort right now, or whether
>>> we should simply nuke the bluetooth code after 4.1 has been released?
>>
>> Well, perfect is the enemy of good :)
>>
>> This series is already done and is an improvement to what we have.
>>
>> Regarding nuking it, it depends on the Nokia N-series boards, they might
>> become useless without BT support.
> 
> Er, they're not useless at all without BT support. The BT
> hardware is a really tiny part that I doubt many users of
> the board models ever used. As long as we retain a "simulate
> doing nothing much" model of the BT device to show the guest
> I don't care whether the BT backend code disappears.

OK, I won't insist then.

Re: [PATCH-for-4.1? 0/7] vl: Allow building with CONFIG_BLUETOOTH disabled

[Philippe Mathieu-Daudé]

On 8/14/19 2:45 PM, Philippe Mathieu-Daudé wrote:
> On 8/13/19 4:04 PM, Peter Maydell wrote:
>> On Tue, 13 Aug 2019 at 15:01, Philippe Mathieu-Daudé <philmd@redhat.com> wrote:
>>> On 7/15/19 3:13 PM, Thomas Huth wrote:
>>>> On 12/07/2019 15.39, Philippe Mathieu-Daudé wrote:
>>>>> A series of obvious patches to build without the deprecated
>>>>> bluetooth devices. Still worth for 4.1 or too late?
>>>>> It is clearly not a bugfix.
>>>>
>>>> I wonder whether this series is worth the effort right now, or whether
>>>> we should simply nuke the bluetooth code after 4.1 has been released?
>>>
>>> Well, perfect is the enemy of good :)
>>>
>>> This series is already done and is an improvement to what we have.
>>>
>>> Regarding nuking it, it depends on the Nokia N-series boards, they might
>>> become useless without BT support.
>>
>> Er, they're not useless at all without BT support. The BT
>> hardware is a really tiny part that I doubt many users of
>> the board models ever used. As long as we retain a "simulate
>> doing nothing much" model of the BT device to show the guest
>> I don't care whether the BT backend code disappears.
> 
> OK, I won't insist then.

Bluetooth is dead, long live BT!

v4.2.0-rc0 just got tagged. We should stop linking unmaintained dead 
code. If nobody step in to nuke BT, we should consider applying this 
series before we release QEMU 5.0 with dead Bluetooth. This approach is 
still better than burying our head in the sand.

Re: [PATCH-for-4.1? 0/7] vl: Allow building with CONFIG_BLUETOOTH disabled

[Thomas Huth]

On 07/11/2019 20.09, Philippe Mathieu-Daudé wrote:
[...]
> Bluetooth is dead, long live BT!
> 
> v4.2.0-rc0 just got tagged. We should stop linking unmaintained dead 
> code. If nobody step in to nuke BT, we should consider applying this 
> series before we release QEMU 5.0 with dead Bluetooth. This approach is 
> still better than burying our head in the sand.

FWIW, I'm planning to send some patches to remove the bluetooth code in 
the 5.0 development cycle. Unless you want to beat me to it. But I don't 
think it makes sense to still make this configurable. The bluetooth code 
has been marked as deprecated for a while now, and nobody spoke up that 
they are still using it (and as far as I can tell, it's currently not 
usable anymore anyway), so it should simply be removed now.

  Thomas