Insufficient memory for Bluetooth driver?

[Ketsugi]

On my new iPaq 2750, I often get this error after turning on my unit, sometimes even right after a soft reset: Could not activate Bluetooth radio due to insufficient memory for Bluetooth driver (not verbatim).

What on earth could be causing this?

Edit: here's the exact error given (the dialog title is BTTrayCE)

"The Bluetooth Radio failed to turn ON due to insufficient driver memory available. You must perform a normal reset before you can turn ON the Bluetooth Radio."

[Janak Parekh]

This is a problem that's existed with most Bluetooth stacks since the ancient iPAQ 3870. :| It's really annoying that they haven't been able to fix it yet.

A soft-reset should usually cure it, though. How much software do you have installed on the unit?

--janak

[Darius Wey]

Also, check to see that you actually have enough memory available.

[Sven Johannsen]

I find it interesting that I don't recall hearing about this before. Maybe folks mentioned it, and I just blew it off. I never saw it on my 2215, or my or my wife's 4155s, or my X30. Got an X50v and have seen it quite often. I doubt seriously that it actually is a low memory issue. Mine has 128M RAM (upgraded). It never gets all that low. Soft reset pretty much solves it at that moment. It's annoying though.

[Darius Wey]

I find it interesting that I don't recall hearing about this before. Maybe folks mentioned it, and I just blew it off. I never saw it on my 2215, or my or my wife's 4155s, or my X30. Got an X50v and have seen it quite often. I doubt seriously that it actually is a low memory issue. Mine has 128M RAM (upgraded). It never gets all that low. Soft reset pretty much solves it at that moment. It's annoying though.

I've had it occur on my h1940 a number of times, but that was before I hard reset it the other day. I had a meagre 4MB of memory left - enough to do absolutely nothing. ;) So the insufficient memory problem was understandable in that respect. :P

[Ketsugi]

It's never an actual memory issue; I usually have about 40mb free program memory when this error occurs; pushing it up to 60mb free doesn't help, either. A soft reset usually fixes it, but the error will pop up after turning on the unit a few times. Sometimes it occurs within a few minutes of a soft-reset. It's getting to be rather irritating.

[dewine]

I'm having this same problem and it's happening several times a day now.
It happens right after a soft reset and there is plenty of memory available. I haven't installed any software recently.

[Pat Logsdon]

I had the same problem with my PDA2k. BT was working fine one day, then I got the error. Installing the new iMate ROM update fixed it for me.

[theipaqlover]

Try going to the BT settings and disable the services. I found I jut turned them all off since activesync does not need any of them. The less the better, problem should stop.

George

[Zidane]

The problem is not available RAM, but available driver memory. The kernel only allocates so much memory to the device drivers and if that memory gets "eaten" up, then you get an error like this. Unfortunately, I don't know if it really is the Bluetooth driver's fault as it occurs on devices from multiple manufacturers that probably use different Bluetooth drivers. It may be that the OS is leaking driver memory, just like it leaks main RAM. That could be why the OEMs can't issue a patch to fix it.

[yankeejeep]

This is a situation similar to the system resources issue faced on Windows desktops from 3x through ME. The warning message is vague and doesn't actually tell you exactly what the problem is. And, like the old design of 16-bit Windows, leakage is what gets you there most of the time. The more drivers you have loaded, the quicker you will get to the error message. We can hope that the next WM version will drop this preset constraint in favor of a more dynamic allocation.

[Peter Traugot]

Well, I've got a 4700 with a fair amout of apps, but certanly a few meg of available ram, and I too can not get past this driver issue.

I have no services running other than my bluetooth keyboard.

What I have choosen to to for the time being is enable bluetooth on soft reset.

I'n mot sure how much juice it takes, but if I know I'm going to be away from a charging source for more than a day, I'll turn it off and reboot if needed.

what a drag. So much potential and stupid little failures like this just eat at the foundation.

[Peter Traugot]

O.K., snooped around the HP site and found this. . .

This Bluetooth issue is one that's plagued iPAQS from the beginning. The problem is a result of a number of decisions by various groups.

1) Microsoft decided to allocate only one block of memory for drivers. In Windows Mobile 2003, that changed to two blocks, one for system drivers and one for application drivers. That helped a bit, but does not completely solve the problem. Bluetooth is considered a system driver.
2) HP decided not to allocate the driver memory until the Bluetooth radio was turned on. That gives you access to that memory when the radio is off, but, if another program gets there first, you get this annoying message. Actually, HP put the message in there to warn you. Other PPCs just fail to load the Bluetooth stack and don't tell you. Yet other PPCs allocate the driver memory whether Bluetooth is ever used or not.
3) The Bluetooth consortium decided that memory for every profile that is supported in a machine will be allocated when the Bluetooth stack is loaded. That means that, even if you only use one Bluetooth profile, memory is allocated for every one supported.

As a result, if you turn on the Bluetooth radio right after a soft-reset, you'll always be able to use it. If you wait, or turn it off. Other drivers, or even programs, may use the memory that the Bluetooth stack needs. Since all the system drivers load into one specific block, even if you have tons of free memory, if drivers or programs have taken enough of the system driver block that the Bluetooth stack can't load, you get the warning message and the only thing you can do to make it work is to do a soft-reset.

HP could "solve" the problem by pre-allocating the Bluetooth stack when you do a soft-reset. Microsoft could truly solve the problem by allowing drivers to load anywhere in memory.

Let me know if it works for anyone.

I may not have been far off with my solution after all :wink: