why must windows pause when a dvd/cd is inserted?

CTho9305

Elite Member
Jul 26, 2000
9,214
1
81
I don't see that behavior with XP... if you're using XP, are your CD drives in PIO mode or DMA?
 

her34

Senior member
Dec 4, 2004
581
1
81
windows xp

dma mode

nec 2510


there's no pause at all for you? me it lasts like 7 seconds
 

bersl2

Golden Member
Aug 2, 2004
1,617
0
0
I personally think that it's because the kernel is blocking the whole system with I/O, when it should only be blocking one process. This phenomenon happens for me even for hard disk access, and it pisses me off to no end any time I use an XP machine.
 

QueZart

Member
May 27, 2005
165
0
0
AutoRun.........

Dont Like it, Right click on your drives in My Computer goto properties and turn it off
 

TonyRic

Golden Member
Nov 4, 1999
1,972
0
71
It is due to MS use of the synchronous I/O STILL. I don't understand why they haven't changed this behavior. OS/2 went async in 2.0 and Linux always has been.
 

Smilin

Diamond Member
Mar 4, 2002
7,357
0
0
What in the hell are all of you talking about?

Seriously.

I don't get any pause on any machine when I insert a CD with or without autoplay. Even this crap PII laptop here works fine. I think you guys need to fix your crap.

 

LTC8K6

Lifer
Mar 10, 2004
28,520
1,575
126
No pauses here either on several different pooters, p3's and p4's.
 

Link19

Senior member
Apr 22, 2003
971
0
0
It seems I get a pause only in the Explorer process. Like when I try to open My Computer, it will be stuck until the CD or DVD drive spins up the disc and reads the contents of it. I notice the same thing when I take a disc out.
 

Jeff7

Lifer
Jan 4, 2001
41,596
19
81
Same here - seems like forever until it's usable again. It's done this to me from Windows 95 all the way to WinXP, on every system.
Yes, DMA is enabled. Yes, autoplay is disabled. Those are two of the first tweaks I make when I install a new OS.

Put a disc in, and the cursor moves, but I can't change windows, access the taskbar, nothing, until the disc has spun up. This happens in my main system, secondary system, and my laptop. Always has.
 

imported_Phil

Diamond Member
Feb 10, 2001
9,837
0
0
Originally posted by: Jeff7
Same here - seems like forever until it's usable again. It's done this to me from Windows 95 all the way to WinXP, on every system.
Yes, DMA is enabled. Yes, autoplay is disabled. Those are two of the first tweaks I make when I install a new OS.

Put a disc in, and the cursor moves, but I can't change windows, access the taskbar, nothing, until the disc has spun up. This happens in my main system, secondary system, and my laptop. Always has.

Mine does roughly the same thing, but it's for about two seconds.
 

CTho9305

Elite Member
Jul 26, 2000
9,214
1
81
Originally posted by: Phil
Originally posted by: Jeff7
Same here - seems like forever until it's usable again. It's done this to me from Windows 95 all the way to WinXP, on every system.
Yes, DMA is enabled. Yes, autoplay is disabled. Those are two of the first tweaks I make when I install a new OS.

Put a disc in, and the cursor moves, but I can't change windows, access the taskbar, nothing, until the disc has spun up. This happens in my main system, secondary system, and my laptop. Always has.

Mine does roughly the same thing, but it's for about two seconds.

I did have that occur for a while when I was stuck with a dying CD drive. Of course, then it happened in Linux too.
 

Nothinman

Elite Member
Sep 14, 2001
30,672
0
0
I've seen it too, I just figured it was Explorer attempting to mount and read the disc.
 

bsobel

Moderator Emeritus<br>Elite Member
Dec 9, 2001
13,346
0
0
Originally posted by: TonyRic
It is due to MS use of the synchronous I/O STILL. I don't understand why they haven't changed this behavior. OS/2 went async in 2.0 and Linux always has been.

This is incorrect, I/O is most certainly not synchronous.


 

ProviaFan

Lifer
Mar 17, 2001
14,993
1
0
Originally posted by: Nothinman
I've seen it too, I just figured it was Explorer attempting to mount and read the disc.
Happens to me as well... DMA, Autoplay disabled, etc.

I've heard that it's an IDE thing, and that SCSI systems don't have this issue. Now, can anyone find me a SCSI DVD+/-RW? :confused:
 

Nothinman

Elite Member
Sep 14, 2001
30,672
0
0
I've heard that it's an IDE thing, and that SCSI systems don't have this issue. Now, can anyone find me a SCSI DVD+/-RW?

Um, no. SCSI has a number of features over IDE, but instantaneous spinup and reads aren't among them.
 

n7

Elite Member
Jan 4, 2004
21,281
4
81
With every computer i have ever used at home, & also likely 30+ different PCs at work, there is always a few seconds pause as the DVD/CD ROM/RW reads the disc, or at least the My Computer window freezes for a couple seconds.

What the heck are you guys doing that prevents that?
 

QueBert

Lifer
Jan 6, 2002
22,828
1,042
126
the people who say they don't get a pause are lying. I've had easy 10 different PC's, with god knows how many different brand of CD/DVD drives. And ALL of them have had a small pause. I don't buy anyone who says they don't get the pause.

 

Fullmetal Chocobo

Moderator<br>Distributed Computing
Moderator
May 13, 2003
13,704
7
81
I agree. I was getting that on my P4 2.4, and I thought it would go away when I built my dual Xeon, but it's still there. Not that I really care, but it is there nonetheless...
Tas.
 

lansalot

Senior member
Jan 25, 2005
298
0
0
Get a dual-core CPU, your responsiveness will appear like magic.

That was the first thing I noticed when moving from my BP6 (dual-466) to Athlon1700.
 

bersl2

Golden Member
Aug 2, 2004
1,617
0
0
I think there's an option that will spawn a new process for each explorer window. Why this isn't enabled by default, I'll never know. In theory, it should help, but if the kernel-land won't yield (and do realize that Windows has always been a kernel-heavy environment---I mean, it took forever for them to even stop using BIOS routines mostly (up to 9x series) (at least, I think that's the case...)), then the userland will freeze.

Also, because I'm curious, are Windows drivers usually SMP-safe? Or is this just not an issue due to the way it's designed?
 

ProviaFan

Lifer
Mar 17, 2001
14,993
1
0
Originally posted by: Nothinman
I've heard that it's an IDE thing, and that SCSI systems don't have this issue. Now, can anyone find me a SCSI DVD+/-RW?
Um, no. SCSI has a number of features over IDE, but instantaneous spinup and reads aren't among them.
The issue is not that a SCSI-only system would spin up the CDROM immediately (which does not happen, as you say), but rather it probably has something to do with the way the HBA works, or how the drivers for it are written, or something. Again, I really wish that I had a way to test this, but unfortunately I haven't a single SCSI system in the house. :(
 

Nothinman

Elite Member
Sep 14, 2001
30,672
0
0
The issue is not that a SCSI-only system would spin up the CDROM immediately (which does not happen, as you say), but rather it probably has something to do with the way the HBA works, or how the drivers for it are written, or something. Again, I really wish that I had a way to test this, but unfortunately I haven't a single SCSI system in the house

From what I've noticed, it's related to how long it takes the drive to spinup and read whatever it is Explorer looks for. Explorer blocks on that read request and since Explorer also controls the desktop and such everything blocks on that read request. As bersl2 said, enabling the option to have each Explorer window start a new process might help, I haven't tried it though.

As for the driver differences, In NT I believe the IDE driver used to have one central lock, so that when a read on one IDE controller was happening all of the rest would wait on that lock before being able to do any I/O serializing all of your IDE devices. I believe that was done because high-performance IDE is sort of an oxymoron, if you want lots of concurrent access to lots of disks you use SCSI. For SCSI there's at least one lock per controller, it might even be more fine-grained that that, probably down to the port level. Whether that's still true or not, I don't know. MS could have fixed that anywhere along the line and 99% of their users wouldn't have noticed at all.