Skip to contentSkip to conversation

pull request: bluetooth-next 2012-09-18

DISCUSSION Gustavo Padovan 3 participants 5 messages Sep 19, 2012 – Sep 24, 2012 Bluetooth drivers

lore
Share & tools
lore
#0 lore raw

Hi,

My last pull request to 3.7. There is nothin really exciting here. The biggest stuff is hci_connect() code re-write, the rest is just bugfixes, clean up and small improvements.

Please pull!

Gustavo

---
The following changes since commit 22c5649eef0fc37532e20c14d2656b28ca708a69:

p54: fix powerpc gcc warnings (2012-08-13 15:28:35 -0400)

are available in the git repository at:

git://git.kernel.org/pub/scm/linux/kernel/git/bluetooth/bluetooth-next master

for you to fetch changes up to 3c860f3d3cd75f0e1a33ba6eb9f96bdb2194ec41:

Bluetooth: Add support for BCM20702A0 [04ca, 2003] (2012-09-18 21:22:39 -0300)

----------------------------------------------------------------
Andrei Emeltchenko (7):
Bluetooth: trivial: Shorten variable scope
Bluetooth: trivial: Use preferred method for NULL check
Bluetooth: Remove unneeded zero init
Bluetooth: trivial: Make hci_chan_del return void
Bluetooth: trivial: Remove empty line
Bluetooth: debug: Print refcnt for hci_dev
Bluetooth: AMP: Add Read Data Block Size to amp_init

Gustavo Padovan (1):
Bluetooth: Add USB_VENDOR_AND_INTERFACE_INFO() for Broadcom/Foxconn

Jaroslav Resler (1):

      Bluetooth: Add support for BCM20702A0 [04ca, 2003]

Johan Hedberg (2):
Bluetooth: mgmt: Implement support for passkey notification
Bluetooth: Update management interface revision

Mikel Astiz (3):
Bluetooth: Add more HCI error codes
Bluetooth: Fix minor coding style in hci_event.c
Bluetooth: mgmt: Add device disconnect reason

Peter Senna Tschudin (5):
drivers/bluetooth/bluecard_cs.c: removes unnecessary semicolon drivers/bluetooth/hci_ldisc.c: removes unnecessary semicolon drivers/bluetooth/hci_ll.c: removes unnecessary semicolon drivers/bluetooth/hci_vhci.c: removes unnecessary semicolon drivers/bluetooth/btuart_cs.c: removes unnecessary semicolon

Sachin Kamat (1):
Bluetooth: Use module_platform_driver() in btwilink.c file

Syam Sidhardhan (2):
Bluetooth: debug: Correct the PSM printing
Bluetooth: Use kref for l2cap channel reference counting

Vinicius Costa Gomes (8):
Bluetooth: Remove some functions from being exported
Bluetooth: Rename LE and ACL connection functions
Bluetooth: Refactor LE connection into its own function
Bluetooth: Refactor ACL connection into its own function
Bluetooth: Refactor SCO connection into its own function
Bluetooth: Simplify a the connection type handling
Bluetooth: Add type information to the hci_connect() debug statement
Bluetooth: Fix establishing ESCO links

Wei Yongjun (1):
Bluetooth: btmrvl: remove pointless conditional before kfree_skb()

 drivers/bluetooth/bluecard_cs.c  |   2 +-
 drivers/bluetooth/btmrvl_sdio.c  |   3 +-
 drivers/bluetooth/btuart_cs.c    |   2 +-
 drivers/bluetooth/btusb.c        |   3 +-
 drivers/bluetooth/btwilink.c     |  16 +------
 drivers/bluetooth/hci_ldisc.c    |   2 +-
 drivers/bluetooth/hci_ll.c       |   2 +-
 drivers/bluetooth/hci_vhci.c     |   2 +-
 include/net/bluetooth/hci.h      |  21 +++++++++
 include/net/bluetooth/hci_core.h |  26 +++++------
 include/net/bluetooth/l2cap.h    |   3 +-
 include/net/bluetooth/mgmt.h     |  16 +++++++
 net/bluetooth/af_bluetooth.c     |  10 ++---
 net/bluetooth/hci_conn.c         | 100 +++++++++++++++++++++++++----------------
 net/bluetooth/hci_core.c         |   6 ++-
 net/bluetooth/hci_event.c        |  99 +++++++++++++++++++++++++++++++++++++---
 net/bluetooth/l2cap_core.c       |  17 +++----
 net/bluetooth/mgmt.c             |  28 +++++++++---
 18 files changed, 255 insertions(+), 103 deletions(-)
#1 lore raw
Quoted 2 lines · Gustavo Padovan

Johan Hedberg (2):
Bluetooth: mgmt: Implement support for passkey notification

Too late now... but why did we allow this a stable fix? I get its a feature that is important and likely overlooked / someone had a brain fart, but from what I gather this is not on v3.5.4 and yet may make it to v3.5.5. At least for backporting this was a bitch:

/home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3290:52:
error: dereferencing pointer to incomplete type /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3294:6:
error: ‘struct hci_conn’ has no member named ‘passkey_notify’ /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3294:25:
error: dereferencing pointer to incomplete type /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3295:6:
error: ‘struct hci_conn’ has no member named ‘passkey_entered’ /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3298:3:
error: implicit declaration of function ‘mgmt_user_passkey_notify’ [-Werror=implicit-function-declaration] /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3299:27:
error: ‘struct hci_conn’ has no member named ‘passkey_notify’ /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3300:11:
error: ‘struct hci_conn’ has no member named ‘passkey_entered’ /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:
In function ‘hci_keypress_notify_evt’:
/home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3310:52:
error: dereferencing pointer to incomplete type /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3314:12:
error: dereferencing pointer to incomplete type /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3315:7:
error: ‘HCI_KEYPRESS_STARTED’ undeclared (first use in this function) /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3315:7:
note: each undeclared identifier is reported only once for each
function it appears in /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3316:7:
error: ‘struct hci_conn’ has no member named ‘passkey_entered’ /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3319:7:
error: ‘HCI_KEYPRESS_ENTERED’ undeclared (first use in this function) /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3320:7:
error: ‘struct hci_conn’ has no member named ‘passkey_entered’ /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3323:7:
error: ‘HCI_KEYPRESS_ERASED’ undeclared (first use in this function) /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3324:7:
error: ‘struct hci_conn’ has no member named ‘passkey_entered’ /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3327:7:
error: ‘HCI_KEYPRESS_CLEARED’ undeclared (first use in this function) /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3328:7:
error: ‘struct hci_conn’ has no member named ‘passkey_entered’ /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3331:7:
error: ‘HCI_KEYPRESS_COMPLETED’ undeclared (first use in this
function) /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3337:27:
error: ‘struct hci_conn’ has no member named ‘passkey_notify’ /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3338:11:
error: ‘struct hci_conn’ has no member named ‘passkey_entered’ /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:
In function ‘hci_event_packet’:
/home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3701:7:
error: ‘HCI_EV_USER_PASSKEY_NOTIFY’ undeclared (first use in this
function) /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3705:7:
error: ‘HCI_EV_KEYPRESS_NOTIFY’ undeclared (first use in this
function)

I'd much prefer to see things like these identified as 'important' but not 'fuck, this fixes that crash'. The fact that kernel maintainers may be anal about these sorts of patches is *good* for many reasons and figuring out a way to categorize, identify and allow users to reap benefits of backporting suck fixes / cherry picking a full solution is what something that should not interfere and hopefully be transparent to Linux upstream development.

Here's one way I dealt with these type of patches:

https://backports.wiki.kernel.org/index.php/Documentation/compat-drivers/additional-patches

If you look at the patch 'Bluetooth: mgmt: Implement support for passkey' its a clear "feature" implementation, not a fucking fix.

Luis

#2 lore raw
Quoted 74 lines · Luis R. Rodriguez
Quoted 2 lines · Gustavo Padovan

Johan Hedberg (2):
Bluetooth: mgmt: Implement support for passkey notification

Too late now... but why did we allow this a stable fix? I get its a feature that is important and likely overlooked / someone had a brain fart, but from what I gather this is not on v3.5.4 and yet may make it to v3.5.5. At least for backporting this was a bitch:

/home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3290:52:
error: dereferencing pointer to incomplete type /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3294:6:
error: ‘struct hci_conn’ has no member named ‘passkey_notify’ /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3294:25:
error: dereferencing pointer to incomplete type /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3295:6:
error: ‘struct hci_conn’ has no member named ‘passkey_entered’ /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3298:3:
error: implicit declaration of function ‘mgmt_user_passkey_notify’ [-Werror=implicit-function-declaration] /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3299:27:
error: ‘struct hci_conn’ has no member named ‘passkey_notify’ /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3300:11:
error: ‘struct hci_conn’ has no member named ‘passkey_entered’ /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:
In function ‘hci_keypress_notify_evt’:
/home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3310:52:
error: dereferencing pointer to incomplete type /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3314:12:
error: dereferencing pointer to incomplete type /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3315:7:
error: ‘HCI_KEYPRESS_STARTED’ undeclared (first use in this function) /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3315:7:
note: each undeclared identifier is reported only once for each
function it appears in /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3316:7:
error: ‘struct hci_conn’ has no member named ‘passkey_entered’ /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3319:7:
error: ‘HCI_KEYPRESS_ENTERED’ undeclared (first use in this function) /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3320:7:
error: ‘struct hci_conn’ has no member named ‘passkey_entered’ /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3323:7:
error: ‘HCI_KEYPRESS_ERASED’ undeclared (first use in this function) /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3324:7:
error: ‘struct hci_conn’ has no member named ‘passkey_entered’ /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3327:7:
error: ‘HCI_KEYPRESS_CLEARED’ undeclared (first use in this function) /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3328:7:
error: ‘struct hci_conn’ has no member named ‘passkey_entered’ /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3331:7:
error: ‘HCI_KEYPRESS_COMPLETED’ undeclared (first use in this
function) /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3337:27:
error: ‘struct hci_conn’ has no member named ‘passkey_notify’ /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3338:11:
error: ‘struct hci_conn’ has no member named ‘passkey_entered’ /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:
In function ‘hci_event_packet’:
/home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3701:7:
error: ‘HCI_EV_USER_PASSKEY_NOTIFY’ undeclared (first use in this
function) /home/mcgrof/staging/v3.5.4/compat-wireless-3.5.3-2-snpc/net/bluetooth/hci_event.c:3705:7:
error: ‘HCI_EV_KEYPRESS_NOTIFY’ undeclared (first use in this
function)

I'd much prefer to see things like these identified as 'important' but not 'fuck, this fixes that crash'. The fact that kernel maintainers may be anal about these sorts of patches is *good* for many reasons and figuring out a way to categorize, identify and allow users to reap benefits of backporting suck fixes / cherry picking a full solution is what something that should not interfere and hopefully be transparent to Linux upstream development.

Quoted 6 lines

Here's one way I dealt with these type of patches:

https://backports.wiki.kernel.org/index.php/Documentation/compat-drivers/additional-patches

If you look at the patch 'Bluetooth: mgmt: Implement support for passkey' its a clear "feature" implementation, not a fucking fix.

Yes, this patch is wrongly marked as stable, it is not even going to 3.5. I'm just removing the tag from it and resending the pull request to John.

Gustavo

#3 lore raw
Quoted 11 lines · Gustavo Padovan
Quoted 9 lines · Luis R. Rodriguez
Quoted 2 lines · Gustavo Padovan

Johan Hedberg (2):
Bluetooth: mgmt: Implement support for passkey notification

Too late now... but why did we allow this a stable fix? I get its a feature that is important and likely overlooked / someone had a brain fart, but from what I gather this is not on v3.5.4 and yet may make it to v3.5.5. At least for backporting this was a bitch:

<snip>

Quoted 2 lines

Yes, this patch is wrongly marked as stable, it is not even going to 3.5. I'm just removing the tag from it and resending the pull request to John.

Hmmm...well, I already pulled it, and I just pushed it out. Can we live with it?

John

Signature · 2 lines
John W. Linville		Someday the world will need a hero, and you
[email protected]			might be all we have.  Be ready.
#4 lore raw
Quoted 13 lines · John W. Linville
Quoted 11 lines · Gustavo Padovan
Quoted 9 lines · Luis R. Rodriguez
Quoted 2 lines · Gustavo Padovan

Johan Hedberg (2):
Bluetooth: mgmt: Implement support for passkey notification

Too late now... but why did we allow this a stable fix? I get its a feature that is important and likely overlooked / someone had a brain fart, but from what I gather this is not on v3.5.4 and yet may make it to v3.5.5. At least for backporting this was a bitch:

Quoted 6 lines

<snip>

Quoted 2 lines

Yes, this patch is wrongly marked as stable, it is not even going to 3.5. I'm just removing the tag from it and resending the pull request to John.

Hmmm...well, I already pulled it, and I just pushed it out. Can we live with it?

Yes, that would be ok for me.

Gustavo