yes but the array is in 68k code, so it CAN be aligned in 2-byte
boundary. this is exactly what ben was talking about.
On 5/17/06, P. Douglas Reeder <[EMAIL PROTECTED]> wrote:
On May 17, 2006, at 1:20 PM, Tinnus wrote:
> Can't you just declare it as an array of UInt32?
>
Nope. See "Integer A
On May 17, 2006, at 1:20 PM, Tinnus wrote:
Can't you just declare it as an array of UInt32?
Nope. See "Integer Alignment" under
http://www.palmos.com/dev/support/docs/palmos/PalmOSCompanion/
PalmOS5ARM_Chapter.html#918090
"ARM processors require that 4-byte integers be aligned on a 4-b
no. in 68k uint32 can still be aligned on 2-byte boundary.
On 5/17/06, Tinnus <[EMAIL PROTECTED]> wrote:
Can't you just declare it as an array of UInt32?
-- For information on using the PalmSource Developer Forums, or to
unsubscribe, please see
http://www.palmos.com/dev/support/forums/
--
B
Can't you just declare it as an array of UInt32?
--
For information on using the PalmSource Developer Forums, or to unsubscribe, please see http://www.palmos.com/dev/support/forums/
hello,
good call, ben.
or if you use cw set alignment to 4bytes
On 5/16/06, Ben Combee <[EMAIL PROTECTED]> wrote:
Actually, that code won't work if the 68K compiler aligned the char
array on a 2-byte boundary, something that's quite possible because
the stack is only kept 2-byte aligned. Decl
Actually, that code won't work if the 68K compiler aligned the char
array on a 2-byte boundary, something that's quite possible because
the stack is only kept 2-byte aligned. Declaring that byte array as a
global would be the safest bet.
On 5/16/06, Dmitry Grinberg <[EMAIL PROTECTED]> wrote:
ok
ok then this
void enableWrites(Boolean enable){
char code[]={0x01, 0x00, 0xA0, 0xE1, 0x04 0x10, 0x19, 0xE5, 0x4C,
0xF1, 0x91, 0xE5};
PceNativeCall(code,enable);
}
call it with enable set to "true" to enable writes, "false" to disable
writes to data area.
this is GUARANTEED to work on all cu
I thought those calls didn't work on 100% of the OS5 devices?2006/5/16, Dmitry Grinberg <[EMAIL PROTECTED]>:
hello,memory semaphore disables task switching on PalmOS, as well as timers,so any sound being played back in the background will stop. all systemservices that run in a thread will stop (lik
hello,
memory semaphore disables task switching on PalmOS, as well as timers,
so any sound being played back in the background will stop. all system
services that run in a thread will stop (like NVFS!!!) so i would not
recomend this at all, instead try
HwrEnableDataWrites() and HwrDisableDataWr
[EMAIL PROTECTED] wrote:
> Now this app is running on the Palm T|X as well as on the Simulator 5.4.
> On the Emulator there was no problem all the time.
>
> But I haven't run Gremlins, etc.
>
> I wonder why are you asking?
Honestly, I think it's just curiosity. Plus, IMO many people
posting que
Hi Jay,
As C.A.R. Hoare used to say: Testing can only prove the presence of bugs but
never their absence;)
The app I needed to port to Garnet was crashing constantly on a Palm T|X and on
the Palm Simulators as well.
Well as I said, I removed the calls to the Memory Semaphores Functions around
[EMAIL PROTECTED] wrote:
> Greetings everyone,
>
> I wonder if someone can really answer this:
> In this forum and on the Usenet the opinion
> is spread that one cannot use
> MemSemaphoreReserve/MemSemaphoreRelease
> with PACE (for Garnet) any more.
> Unfortunately I
Greetings everyone,
I wonder if someone can really answer this:
In this forum and on the Usenet the opinion
is spread that one cannot use
MemSemaphoreReserve/MemSemaphoreRelease
with PACE (for Garnet) any more.
Unfortunately I had to port an app which makes
heavy use of these functions, so I
*> what are you doing that requires that much read/write RAM?
Well it's a port of UAE (amiga emulator) and it is already working,
I'm just improving it, trying to get more RAM for emulation.
*> you may need to seek an alternative algorithm that uses less
*> read/write
*> RAM that is designed wi
On 11/15/05, Sinisa Marovic (AC/EDD) <[EMAIL PROTECTED]> wrote:
> *> are you sure you cannot isolate the read/write RAM from read/only?
>
> I already did. As I said I have 2MB chunk in read only RAM. The rest should
> be read/write.
what are you doing that requires that much read/write RAM?
you m
*> are you sure its RAM you need? maybe you only need read only
*> RAM. you can store information in two ways; one on the dynamic heap
*> and one on the storage heap. the storage heap - which, is what you
*> have done right now - can be read without a semaphore. if you want
*> to write to it, you n
On 11/15/05, Sinisa Marovic (AC/EDD) <[EMAIL PROTECTED]> wrote:
> I know that it's not recommended to use it, still I know that some
> people were and are using it and I would like to have that knowledge,
> which is not in the documentation (and I read it thoroughly).
documentation, as everyone k
I know that it's not recommended to use it, still I know that some
people were and are using it and I would like to have that knowledge,
which is not in the documentation (and I read it thoroughly).
Does that mean that if an app needs more heap RAM than it's available,
one should just abandon it?
On 11/11/05, Sinisa Marovic (AC/EDD) <[EMAIL PROTECTED]> wrote:
> solved with releasing the semaphore before "SndStreamStart".
> Thanks for (no) help.
if you take the time to read the documentation - using semaphores is not
endorsed by palm or palmsource. the fact that you ventured into that
terr
Hi,
solved with releasing the semaphore before "SndStreamStart".
Thanks for (no) help.
--
For information on using the PalmSource Developer Forums, or to unsubscribe,
please see http://www.palmos.com/dev/support/forums/
Hi,
I tried releasing the semaphore before creating the stream, also in
the callback itself, nothing helps. Before introducing semaphores the
application had sound. I also noticed there were problems when using
VFS functions, but that was easily solved by releasing the semaphore
just before usi
Hi!
I have the following problem:
1) I have a rather long-lasting function in a
MemSemaphoreReserve(true) / MemSemaphoreRelease(true) segment
(I need write access during it to the feature memory I allocated)
2) I am playing music using advanced sound in the background and I
have the
ourse), check out
and others up there...
Kevin
-Original Message-
From: steve [mailto:[EMAIL PROTECTED]
Sent: Tuesday, August 26, 2003 8:50 AM
To: Palm Developer Forum
Subject: RE: MemSemaphoreReserve
What about disabling the interrupts for a small amount of time?
Steve
-Ori
What about disabling the interrupts for a small amount of time?
Steve
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of
Douglas Handy
Sent: Tuesday, August 26, 2003 6:40 AM
To: Palm Developer Forum
Subject: Re: MemSemaphoreReserve
The semaphore is
Hey Alan,
> Is there any documentation on these undocumented features? :-)
Found this thread on MemSemaphoreXXX()
http://lists.squeakfoundation.org/pipermail/squeak-dev/1998-
August/018867.html
And of course, this again this one on Palms' site...:-P
http://www.palmos.com/dev/support/docs/p
"Viren" <[EMAIL PROTECTED]> wrote in message news:[EMAIL PROTECTED]
> Okay so that boils down to MemSemaphoreReserve API again. What
> options does Damir have if he decides not to use un-documented API ?
> I am very much interested in knowing answer to that.
I thought
Viren,
>Ok. Understood. He wanna put it in a "critical section" kinds stuff.
>Rite?
Correct.
>Okay so that boils down to MemSemaphoreReserve API again. What
>options does Damir have if he decides not to use un-documented API ?
The semaphore is used for controlli
> threads (of which there are precious few) cannot interrupt him as
they are now.
Okay so that boils down to MemSemaphoreReserve API again. What
options does Damir have if he decides not to use un-documented API ?
I am very much interested in knowing answer to that.
I know many people have wr
Viren,
>HSo you basically want to do "multitask". Is it?
No; it is the opposite problem. He wants to *prevent* his code from being
interrupted by another task.
>You could perhaps make use of alarm manager
>to simulate the multi-tasking stuff particularly if the task has soft-
>dead li
Damir,
> Does anyone know where can I find description of
> MemSemaphoreReserve/MemSemaphoreRelease API.
MemSemaphoreReserve/MemSemaphoreRelease are not the "documented" APIs
of Palm. So please dont expect to get any documented stuff on it.
Nevertheless, search for the archive f
Does anyone know where can I find description of
MemSemaphoreReserve/MemSemaphoreRelease API.
I am working on a UART driver and making some hardware registers
manipulation. I have a problem with PalmOS devices combined with
cell phone. It is seems, that while I am doing my task, system
interacts
> > > Is it a fact that MemSemaphoreReserve() and such will be history?
> >> 5.0
>
> So tell me, Aaron, how do you do, e.g., this:
> convert a record from upper case to lower case.
>
> Don't tell me to copy it to the dynamic memory, do it there, and wr
> > Is it a fact that MemSemaphoreReserve() and such will be history?
>> 5.0
So tell me, Aaron, how do you do, e.g., this:
convert a record from upper case to lower case.
Don't tell me to copy it to the dynamic memory, do it there, and write
it back. The performance loss is trem
> Is it a fact that MemSemaphoreReserve() and such will be history?
Effective
> with what OS #? /Paul N.
5.0
someone did say that it has been working on 5.0 - but, i know i removed
all references to this a long time ago (+12 months now) specific for OS5.
there is much more dynamic heap -
Is it a fact that MemSemaphoreReserve() and such will be history? Effective
with what OS #? /Paul N.
--
For information on using the Palm Developer Forums, or to unsubscribe, please see
http://www.palmos.com/dev/support/forums/
MemSemaphoreReserve(Boolean) causes the Simulator
(running PalmOS 5) to crash.
Have explicit calls to MemSemaphoreReserve(Boolean)
become illegal in Palm OS 5? Is there a way to
get around this? Or is this a problem with the
Simulator?
Please advise if you know anything about this, or
can
36 matches
Mail list logo
