Timed wait/sleep

Timed wait/sleep

[Alex Roman]

Hello,

Is there any mechanism to wait/sleep for a specified time in GRUB?

The reason I need something like this is that, for the ATA driver I am
working on, I need to be able to wait for a memory mapped register to
change, but I don't want to wait indefinitely. I'd like to be able to
have some sort of loop that will read the register, say, 10 times,
with 10 ms pause in between the reads...

I've searched for this and wasn't able to find anything in the source code...

If there isn't one, what would be the best way to implement something
like this? Perhaps using the RTC, or straight loops? Would there be
any use, besides my driver, in developing an RTC interface?


Thanks!

-- 
Alex Roman <alex.roman@gmail.com>

Re: Timed wait/sleep

[Haudy Kazemi]

On Jul 6 2007, Alex Roman wrote:

>Hello,
>
>Is there any mechanism to wait/sleep for a specified time in GRUB?
>
>The reason I need something like this is that, for the ATA driver I am
>working on, I need to be able to wait for a memory mapped register to
>change, but I don't want to wait indefinitely. I'd like to be able to
>have some sort of loop that will read the register, say, 10 times,
>with 10 ms pause in between the reads...
>
> I've searched for this and wasn't able to find anything in the source 
> code...
>
>If there isn't one, what would be the best way to implement something
>like this? Perhaps using the RTC, or straight loops? Would there be
>any use, besides my driver, in developing an RTC interface?

I'd avoid straight loops...with those you run into processor speed 
issues...runs too fast on new CPUs, runs too slow on old hardware. Of 
course I guess you could calibrate some straight loops to the RTC, and then 
use the loops rather than polling the RTC periodically, but I don't think 
the extra level of complication would provide any real benefit.

>Thanks!
>
>

Re: Timed wait/sleep

[Alex Roman]

On 06 Jul 2007 17:04:19 -0500, Haudy Kazemi <kaze0010@umn.edu> wrote:
> On Jul 6 2007, Alex Roman wrote:
> I'd avoid straight loops...with those you run into processor speed
> issues...runs too fast on new CPUs, runs too slow on old hardware. Of
> course I guess you could calibrate some straight loops to the RTC, and then
> use the loops rather than polling the RTC periodically, but I don't think
> the extra level of complication would provide any real benefit.

Yes, I am reluctant to use loops for that reason.

I could definitely do a simple read RTC in a loop until the time
difference from when I started the loop is what I want it to be...

I'll try to get that working, but I am open to other suggestions as well!


Thanks!

-- 
Alex Roman <alex.roman@gmail.com>

RE: Timed wait/sleep

[Amin Azez]

Donzt beat me for this, but grub2 sounds more like a kernel every day.

I almost want to suggest nicking things from hurd.

Grub2 could even be the init on a minimal hurd setup, surely?

Cough

Sam

-----Original Message-----
From: "Alex Roman" <alex.roman@gmail.com>
To: "The development of GRUB 2" <grub-devel@gnu.org>
Sent: 06/07/07 23:21
Subject: Re: Timed wait/sleep

On 06 Jul 2007 17:04:19 -0500, Haudy Kazemi <kaze0010@umn.edu> wrote:
> On Jul 6 2007, Alex Roman wrote:
> I'd avoid straight loops...with those you run into processor speed
> issues...runs too fast on new CPUs, runs too slow on old hardware. Of
> course I guess you could calibrate some straight loops to the RTC, and then
> use the loops rather than polling the RTC periodically, but I don't think
> the extra level of complication would provide any real benefit.

Yes, I am reluctant to use loops for that reason.

I could definitely do a simple read RTC in a loop until the time
difference from when I started the loop is what I want it to be...

I'll try to get that working, but I am open to other suggestions as well!


Thanks!

-- 
Alex Roman <alex.roman@gmail.com>


_______________________________________________
Grub-devel mailing list
Grub-devel@gnu.org
http://lists.gnu.org/mailman/listinfo/grub-devel

Re: Timed wait/sleep

[Jan C. Kleinsorge]

For i/o delays on x86, Linux uses a sequence of outp(0x80). 0x80 is
usually unsed.
One to three times should be quite sufficient.

Jan

ex Roman wrote:
> On 06 Jul 2007 17:04:19 -0500, Haudy Kazemi <kaze0010@umn.edu> wrote:
>> On Jul 6 2007, Alex Roman wrote:
>> I'd avoid straight loops...with those you run into processor speed
>> issues...runs too fast on new CPUs, runs too slow on old hardware. Of
>> course I guess you could calibrate some straight loops to the RTC,
>> and then
>> use the loops rather than polling the RTC periodically, but I don't
>> think
>> the extra level of complication would provide any real benefit.
>
> Yes, I am reluctant to use loops for that reason.
>
> I could definitely do a simple read RTC in a loop until the time
> difference from when I started the loop is what I want it to be...
>
> I'll try to get that working, but I am open to other suggestions as well!
>
>
> Thanks!
>

Re: Timed wait/sleep

[Marco Gerards]

"Alex Roman" <alex.roman@gmail.com> writes:

Hi,

> Is there any mechanism to wait/sleep for a specified time in GRUB?

Not yet.  I will also need this for networking.  Actually, we can do
still in this idle time.  I will get back on this.

You can read the time in milliseconds, IIRC.  You can use that to
check in a loop.  But please do not use busy waiting.  That will give
problems with qemu, etc.

--
Marco

Re: Timed wait/sleep

[Alex Roman]

- Original message -
"Alex Roman" <alex.roman@gmail.com> writes: Hi, Not yet...

Sorry, what do you mean by busy waiting?

Thanks


Alex

On 22/07/07, Marco Gerards <mgerards@xs4all.nl> wrote:
> "Alex Roman" <alex.roman@gmail.com> writes:
>
> Hi,
>
> > Is there any mechanism to wait/sleep for a specified time in GRUB?
>
> Not yet.  I will also need this for networking.  Actually, we can do
> still in this idle time.  I will get back on this.
>
> You can read the time in milliseconds, IIRC.  You can use that to
> check in a loop.  But please do not use busy waiting.  That will give
> problems with qemu, etc.
>
> --
> Marco
>
>
>
> _______________________________________________
> Grub-devel mailing list
> Grub-devel@gnu.org
> http://lists.gnu.org/mailman/listinfo/grub-devel
>


-- 
Alex Roman <alex.roman@gmail.com>

Re: Timed wait/sleep

[Jeroen Dekkers]

At Sun, 22 Jul 2007 17:38:14 -0400,
Alex Roman wrote:
> 
> Sorry, what do you mean by busy waiting?

Busy waiting is looping while checking whether a condition
changed. E.g.

while(condition_is_true());

It would be nicer to let the processor sleep for a while. I'm not sure
whether we can actually do that easily in GRUB however...


Jeroen Dekkers