RE: off topic: Multiprocessor safety in the User Mode

If in a Multi processor system, i released the mutex, then can the thread
from different process running on different processor acquire the same mutex
or not. Or how can we notify another process that some event has occured in
multiprocessor systems

If criticalSection is multiprocessor safe then i feel that all the user mode
threads corresponding to a single process will run on same processor on
multiprocessor machine

Can you please confirm wether this assumption is right or not

On 12/13/06, Rohit wrote:
>
> If in a Multi processor system, i released the mutex, then can the thread
> from different process running on different processor acquire the same mutex
> or not. Or how can we notify another process that some event has occured in
> multiprocessor systems
>
>
>

> If in a Multi processor system, i released the mutex, then can the thread

from different process running on different processor acquire the same mutex
or not.

Yes it can.

Or how can we notify another process that some event has occured in
multiprocessor systems

Lots of ways, event objects is one, completing the pending IRPs is another.

Maxim Shatskih, Windows DDK MVP
StorageCraft Corporation
xxxxx@storagecraft.com
http://www.storagecraft.com

these lines are from file system internals by rajeev nagar

Typically, spin locks and other higher level (Executive) synchronization
mechanisms must be used consistently and correctly in Windows NT drivers to
ensure correct functionality on multiprocessor systems

Is this indicates that mutex are not multiprocessor safe in kernel drivers,
then how can it be possible that user level mutex are multi processor safe

On 12/13/06, Maxim S. Shatskih < xxxxx@storagecraft.com> wrote:

> If in a Multi processor system, i released the mutex, then can the
thread
> from different process running on different processor acquire the same
mutex
> or not.

Yes it can.

>Or how can we notify another process that some event has occured in
> multiprocessor systems

Lots of ways, event objects is one, completing the pending IRPs is
another.

Maxim Shatskih, Windows DDK MVP
StorageCraft Corporation
xxxxx@storagecraft.com
http://www.storagecraft.com


Questions? First check the IFS FAQ at https://www.osronline.com/article.cfm?id=17

You are currently subscribed to ntfsd as: xxxxx@gmail.com
To unsubscribe send a blank email to xxxxx@lists.osr.com

Since you are in the file systems newsgroup you will find spinlocks are very seldom needed because of eresources. They can do most of what you need to do in this arena. There are no hardware interrupts in file systems.
“Rohit” wrote in message news:xxxxx@ntfsd…
these lines are from file system internals by rajeev nagar

Typically, spin locks and other higher level (Executive) synchronization mechanisms must be used consistently and correctly in Windows NT drivers to ensure correct functionality on multiprocessor systems

Is this indicates that mutex are not multiprocessor safe in kernel drivers, then how can it be possible that user level mutex are multi processor safe

On 12/13/06, Maxim S. Shatskih < xxxxx@storagecraft.com> wrote:
> If in a Multi processor system, i released the mutex, then can the thread
> from different process running on different processor acquire the same mutex
> or not.

Yes it can.

>Or how can we notify another process that some event has occured in
> multiprocessor systems

Lots of ways, event objects is one, completing the pending IRPs is another.

Maxim Shatskih, Windows DDK MVP
StorageCraft Corporation
xxxxx@storagecraft.com
http://www.storagecraft.com


Questions? First check the IFS FAQ at https://www.osronline.com/article.cfm?id=17

You are currently subscribed to ntfsd as: xxxxx@gmail.com
To unsubscribe send a blank email to xxxxx@lists.osr.com

>> Typically, spin locks and other higher level (Executive) synchronization

> mechanisms must be used consistently and correctly in Windows NT
> drivers to ensure correct functionality on multiprocessor systems

Is this indicates that mutex are not multiprocessor safe in kernel drivers,

Really ? I guess it just indicates they must be used correctly :slight_smile:
All synchro objects are of course able to handle multiprocessor issues.

L.

Why are you commenting on a 14 year old post? And you are incorrect, Windows runs on NUMA systems, on systems where not all processors can receive all interrupts. Your data is as out of data as the post.