Fix SCO on some bluetooth adapters
Subsystems (2)
PATCH 1/1 Fix SCO on some bluetooth adapters
Some bluetooth adapters return an incorrect number of sco packets in READ_BUFFER_SIZE. Add a quirk hook to fix it from the driver side, and use the hook for a first adapter.
---
I tried to take Marcel Holtmann's remarks into account, but he seems unavailable to tell me whether he likes that version, so I'll send it to the whole shebang instead of just him this time.
The main point was "only the driver knows how to fix the problem", hence a function pointer quirk hook. This is definitively not limited to broadcom USB devices, googling for "SCO MTU 64:0" has interesting results.
Pavel, '4' doesn't seem to be a value that happens in any correctly answering adapter, while 8 happens often. Try "SCO MTU 64:4" and "SCO MTU 64:8" for comparison.
drivers/bluetooth/hci_usb.c | 12 ++++++++++++ drivers/bluetooth/hci_usb.h | 1 + include/net/bluetooth/hci_core.h | 1 + net/bluetooth/hci_event.c | 16 ++++++++++++++-- 4 files changed, 28 insertions(+), 2 deletions(-)
841d7690e75189803907fbc4616b984087e7f89c
drivers/bluetooth/hci_usb.c +12 -0
@@ -130,6 +130,9 @@ static struct usb_device_id blacklist_id/* CSR BlueCore Bluetooth Sniffer */ { USB_DEVICE(0x0a12, 0x0002), .driver_info = HCI_SNIFFER }, + /* Belkin F8T012 */ + { USB_DEVICE(0x050d, 0x0012), .driver_info = HCI_READ_BUFFER_SIZE }, + { }/* Terminating entry */ };
@@ -817,6 +820,12 @@ static void hci_usb_notify(struct hci_deBT_DBG ("%s evt %d" , hdev->name, evt); } +static void hci_usb_quirk_read_buffer_size (struct hci_dev *hdev) +{ +if (!hdev->sco_pkts) + hdev->sco_pkts =8 ; +} +static int hci_usb_probe (struct usb_interface *intf,const struct usb_device_id *id) {struct usb_device *udev =interface_to_usbdev (intf);
@@ -984,6 +993,9 @@ #endifif (reset || id->driver_info &HCI_RESET )set_bit (HCI_QUIRK_RESET_ON_INIT , &hdev->quirks); + +if (id->driver_info &HCI_READ_BUFFER_SIZE ) + hdev->quirk_read_buffer_size = hci_usb_quirk_read_buffer_size;if (id->driver_info &HCI_SNIFFER ) {if (le16_to_cpu (udev->descriptor.bcdDevice) >0x997 )
drivers/bluetooth/hci_usb.h +1 -0
@@ -35,6 +35,7 @@ #define HCI_CSR 0x08#define HCI_SNIFFER 0x10#define HCI_BCM92035 0x20#define HCI_BROKEN_ISOC 0x40 +#define HCI_READ_BUFFER_SIZE 0x80#define HCI_MAX_IFACE_NUM 3
include/net/bluetooth/hci_core.h +1 -0
@@ -134,6 +134,7 @@ struct hci_dev {void (*destruct)(struct hci_dev *hdev);void (*notify)(struct hci_dev *hdev,unsigned int evt);int (*ioctl)(struct hci_dev *hdev,unsigned int cmd,unsigned long arg); +void (*quirk_read_buffer_size)(struct hci_dev *hdev); };struct hci_conn {
net/bluetooth/hci_event.c +14 -2
@@ -321,8 +321,20 @@ static void hci_cc_info_param(struct hci hdev->acl_mtu =__le16_to_cpu (bs->acl_mtu); hdev->sco_mtu = bs->sco_mtu ? bs->sco_mtu :64 ; - hdev->acl_pkts = hdev->acl_cnt =__le16_to_cpu (bs->acl_max_pkt); - hdev->sco_pkts = hdev->sco_cnt =__le16_to_cpu (bs->sco_max_pkt); + hdev->acl_pkts =__le16_to_cpu (bs->acl_max_pkt); + hdev->sco_pkts =__le16_to_cpu (bs->sco_max_pkt); + +if (hdev->quirk_read_buffer_size) + hdev->quirk_read_buffer_size (hdev); + +if (!hdev->acl_mtu || !hdev->sco_mtu || !hdev->acl_pkts || !hdev->sco_pkts) { +printk (KERN_WARNING "Your Bluetooth adapter has an incorrect OCF_READ_BUFFER_SIZE reply (%d:%d, %d:%d)\n" , + hdev->acl_mtu, hdev->acl_pkts, hdev->sco_mtu, hdev->sco_pkts); +printk (KERN_WARNING "It won't work correctly. Please contact Marcel Holtmann <[email protected]> with information about your device to try workarounds.\n" ); + } + + hdev->acl_cnt = hdev->acl_pkts; + hdev->sco_cnt = hdev->sco_pkts;BT_DBG ("%s mtu: acl %d, sco %d max_pkt: acl %d, sco %d" , hdev->name, hdev->acl_mtu, hdev->sco_mtu, hdev->acl_pkts, hdev->sco_pkts);
Signature · 1 line
1.3.0.rc1.g40e9
Re: [PATCH] Fix SCO on some bluetooth adapters (Success report for Belkin F8T012)
Quoted 22 lines · Olivier Galibert
Some bluetooth adapters return an incorrect number of sco packets in
READ_BUFFER_SIZE. Add a quirk hook to fix it from the driver side, and use the hook for a first adapter.Signed-off-by Olivier Galibert <[email protected]>---
I tried to take Marcel Holtmann's remarks into account, but he seems unavailable to tell me whether he likes that version, so I'll send it to the whole shebang instead of just him this time.
The main point was "only the driver knows how to fix the problem", hence a function pointer quirk hook. This is definitively not limited to broadcom USB devices, googling for "SCO MTU 64:0" has interesting results.
Pavel, '4' doesn't seem to be a value that happens in any correctly answering adapter, while 8 happens often. Try "SCO MTU 64:4" and "SCO MTU 64:8" for comparison.
Quoted 7 lines
drivers/bluetooth/hci_usb.c | 12 ++++++++++++ drivers/bluetooth/hci_usb.h | 1 + include/net/bluetooth/hci_core.h | 1 + net/bluetooth/hci_event.c | 16 ++++++++++++++-- 4 files changed, 28 insertions(+), 2 deletions(-)
841d7690e75189803907fbc4616b984087e7f89c
Replying to an old email about a patch that hasn't made it yet (at least to 2.6.17.3).
Without this patch I can't use my Belkin F8T012 with the bluetooth alsa sound driver (snd-bt-sco)
with it, I can :)
Thanks Olivier.
Is this likely to be merged?
David
PS, I tried the patch here:
http://bluetooth-alsa.sourceforge.net/sco-mtu.patch
that is mentioned here:
http://bluetooth-alsa.sourceforge.net/
but it doesn't work for me.
Re: [PATCH] Fix SCO on some bluetooth adapters (Success report for Belkin F8T012)
Quoted 16 lines · On Sat, 01 Jul 2006 13:22:41 +0100 David Greaves <[email protected]> wrote:
Quoted 5 lines
Pavel, '4' doesn't seem to be a value that happens in any correctly answering adapter, while 8 happens often. Try "SCO MTU 64:4" and "SCO MTU 64:8" for comparison.
Quoted 7 lines
drivers/bluetooth/hci_usb.c | 12 ++++++++++++ drivers/bluetooth/hci_usb.h | 1 + include/net/bluetooth/hci_core.h | 1 + net/bluetooth/hci_event.c | 16 ++++++++++++++-- 4 files changed, 28 insertions(+), 2 deletions(-)
841d7690e75189803907fbc4616b984087e7f89cReplying to an old email about a patch that hasn't made it yet (at least to 2.6.17.3).
Quoted 8 lines
Without this patch I can't use my Belkin F8T012 with the bluetooth alsa sound driver (snd-bt-sco)
with it, I can :)
Thanks Olivier.
Is this likely to be merged?
It's floating around in -mm still. I resend it to various people when I feel like provoking a fight.
Marcel is working on a better way of fixing this problem.
But it's a teeny patch and `patch -R' is easy. If this patch doesn't break anything, perhaps we should temp-merge it?
From: Olivier Galibert <[email protected]>
Some bluetooth adapters return an incorrect number of sco packets in READ_BUFFER_SIZE. Fix it.
This is the worst possible way to fix it for several reasons:
- this is not a generic fix, it has to me activated explicitely by the driver
- this is not a driver-specific fix either, only one action is possible (change to 8), the driver can only enable it
- this is not a wildcard-able fix, because it overwrites the returned value no matter what, i.e. even when it is correct. So drivers without device id which need it, for instance uart, can't use it.
- there is no warning anywhere when the problematic condition is detected, so users of non-listed buggy hardware have no clue about what is going on, and the list of buggy hardware will be very hard to complete
- the fix is inconsistant with the sco_mtu handling
But that's exactly what Marcel Holtmann wants, so that's what he gets.
[akpm: this is permanacked, but will keep getting sent until the problem is fixed for real]
---
drivers/bluetooth/hci_usb.c | 6 ++++++ drivers/bluetooth/hci_usb.h | 1 + include/net/bluetooth/hci.h | 3 ++- net/bluetooth/hci_event.c | 10 ++++++++-- 4 files changed, 17 insertions(+), 3 deletions(-)
drivers/bluetooth/hci_usb.c renamed from drivers/bluetooth/hci_usb.c~fix-sco-on-some-bluetooth-adapters-2 +6 -0
@@ -129,6 +129,9 @@ static struct usb_device_id blacklist_id/* CSR BlueCore Bluetooth Sniffer */ { USB_DEVICE(0x0a12, 0x0002), .driver_info = HCI_SNIFFER }, + /* Belkin F8T012 */ + { USB_DEVICE(0x050d, 0x0012), .driver_info = HCI_READ_BUFFER_SIZE }, + { }/* Terminating entry */ };
@@ -984,6 +987,9 @@ static int hci_usb_probe(struct usb_inteif (reset || id->driver_info &HCI_RESET )set_bit (HCI_QUIRK_RESET_ON_INIT , &hdev->quirks); +if (id->driver_info &HCI_READ_BUFFER_SIZE ) +set_bit (HCI_QUIRK_FIX_SCO_PKTS , &hdev->quirks); +if (id->driver_info &HCI_SNIFFER ) {if (le16_to_cpu (udev->descriptor.bcdDevice) >0x997 )set_bit (HCI_QUIRK_RAW_DEVICE , &hdev->quirks);
drivers/bluetooth/hci_usb.h renamed from drivers/bluetooth/hci_usb.h~fix-sco-on-some-bluetooth-adapters-2 +1 -0
@@ -35,6 +35,7 @@#define HCI_SNIFFER 0x10#define HCI_BCM92035 0x20#define HCI_BROKEN_ISOC 0x40 +#define HCI_READ_BUFFER_SIZE 0x80#define HCI_MAX_IFACE_NUM 3
include/net/bluetooth/hci.h renamed from include/net/bluetooth/hci.h~fix-sco-on-some-bluetooth-adapters-2 +2 -1
@@ -54,7 +54,8 @@/* HCI device quirks */ enum {HCI_QUIRK_RESET_ON_INIT , -HCI_QUIRK_RAW_DEVICE +HCI_QUIRK_RAW_DEVICE , +HCI_QUIRK_FIX_SCO_PKTS };/* HCI device flags */
net/bluetooth/hci_event.c renamed from net/bluetooth/hci_event.c~fix-sco-on-some-bluetooth-adapters-2 +8 -2
@@ -320,8 +320,14 @@ static void hci_cc_info_param(struct hci hdev->acl_mtu =__le16_to_cpu (bs->acl_mtu); hdev->sco_mtu = bs->sco_mtu ? bs->sco_mtu :64 ; - hdev->acl_pkts = hdev->acl_cnt =__le16_to_cpu (bs->acl_max_pkt); - hdev->sco_pkts = hdev->sco_cnt =__le16_to_cpu (bs->sco_max_pkt); + hdev->acl_pkts =__le16_to_cpu (bs->acl_max_pkt); + hdev->sco_pkts =__le16_to_cpu (bs->sco_max_pkt); + +if (test_bit (HCI_QUIRK_FIX_SCO_PKTS , &hdev->quirks)) + hdev->sco_pkts =8 ; + + hdev->acl_cnt = hdev->acl_pkts; + hdev->sco_cnt = hdev->sco_pkts;BT_DBG ("%s mtu: acl %d, sco %d max_pkt: acl %d, sco %d" , hdev->name, hdev->acl_mtu, hdev->sco_mtu, hdev->acl_pkts, hdev->sco_pkts);
_