[Bug 60824] [PATCH][regression] Cambridge Silicon Radio, Ltd Bluetooth Dongle unusable

[bugzilla-daemon@bugzilla.kernel.org]

https://bugzilla.kernel.org/show_bug.cgi?id=60824

--- Comment #36 from Sergey Kondakov (virtuousfox@gmail.com) ---
(In reply to Arthur Fragoso from comment #35)
> The code for these devices are bellow.
> 
> You are right, the patch is way too old for this.
> 
> I will probably buy a different device while we wait for someone with more
> knowledge to fix this.
>...
>       /* Detect controllers which aren't real CSR ones. */
>       if (le16_to_cpu(rp->manufacturer) != 10 ||
>           le16_to_cpu(rp->lmp_subver) == 0x0c5c) {
>...

Luckily, I still have my old 2.1 dongle.

It seems that this check is too specific, mine has 0x811 subversion but the
real problem is idiotic notion of holding all BT devices to some imaginary
standard of compliant vendor-approved behaviour and creating blacklists for
actual devices only if someone from BT maintainers have heard something about
some problems from someone. No normal user is going to write them letter with
complains, let alone patches for hard-coded workarounds to artificial problems.
They need to redo the whole initialization logic to be more generic or at least
allow passing quirk-flags at runtime.

There is and will be myriad of devices with random IDs and crappy firmwares,
sometimes even circuitry, and kernel MUST make all that crap work at least
partially, not backdown on smallest of mislabelings. It is saddening to see
only recently created built-in Windows 10 BT stack to behave more sanely than
bluez.

-- 
You are receiving this mail because:
You are the assignee for the bug.