|
Environment: Windows 7, 64 bit and 32 bit, Visual Studio 2008/2012, C++, multi-threading
My searches to find a definition or description as to what memory operations are atomic have turned up nothing. When I include the search term +atomic most of the pages do not even have the word atomic in them. That is a bit irritating.
Question: What level of memory write is atomic: 8 bit, 16 bit, 32 bit, or maybe even 64 bit? Are there any special conditions? How do you know and how can that be verified?
Bonus point: Google flunks on this task. What search engines are good for this endeavor?
Thanks for your time
Thank you for your time
If you work with telemetry, please check this bulletin board: www.irigbb.com
|
|
|
|
|
OK, this is just a guess, but I can't restrain myself: I think the level of memory write that is atomic is the word size of the CPU.
I have nothing to back this up, but I just wanted to throw this out there.
The difficult we do right away...
...the impossible takes slightly longer.
|
|
|
|
|
|
Hello,
RE: Not according to https://www.google.com/search?q=memory+write+atomic+hardware[^].
When I entered "+atomic" and received multiple pages that did not contain the phrase atomic, I see that as an error by the search engine. If I am incorrect then please tell me how to put a requirement for that specific phrase into the search request.
I entered that phrase into a Google search and followed up the links to the first 30 (three pages worth) and did not find any thing that described what instructions are atomic.
I gladly fess up to not having the best of skills, but this is a difficult topic to research. That I why I requested suggestions for a search engine more attuned to this kind of a question.
Thank you for your time
If you work with telemetry, please check this bulletin board: www.irigbb.com
|
|
|
|
|
Depends on the CPU, obviously. Read the manual.
For example, see Intel System Programming Guide, section 8.1.1 "Guaranteed Atomic Operations":
The Intel486 processor (and newer processors since) guarantees that the following basic memory operations will
always be carried out atomically:
• Reading or writing a byte
• Reading or writing a word aligned on a 16-bit boundary
• Reading or writing a doubleword aligned on a 32-bit boundary
The Pentium processor (and newer processors since) guarantees that the following additional memory operations
will always be carried out atomically:
• Reading or writing a quadword aligned on a 64-bit boundary
• 16-bit accesses to uncached memory locations that fit within a 32-bit data bus
The P6 family processors (and newer processors since) guarantee that the following additional memory operation
will always be carried out atomically:
• Unaligned 16-, 32-, and 64-bit accesses to cached memory that fit within a cache line
The most relevant part is the last line, which says that almost load/store that that doesn't cross a cache line boundary is atomic (except anything wider than 8 bytes).
|
|
|
|
|
Hello,
I found the document and saved a copy. That is what I was looking for.
Thank you.
Thank you for your time
If you work with telemetry, please check this bulletin board: www.irigbb.com
|
|
|
|
|
Its not Google that flunks this task. Its your Google-Fu that does.
The information you seek is processor dependant, but in most cases it's the word size of the CPU.
|
|
|
|
|
It is an operation that is not interruptible. So, a single machine code read or write.
Why is it important? SO that if you are reading a chunk of memory bigger that this you make sure you don't get interrupted.
In the Linux kernel, the term 'atomic' (for which there is a func, in_atomic()) means 'is the code running at a privileged level' which is either at soft or hard interrupt level.
Thus timer expiry is soft interrupt. A card triggering an interrupt is hard.
So in the windows kernel DPC or DIRQL are the 'atomic' modes. Passive, is non atomic.
Harolds answer is spot on, just wanted to add a bit more colour.
One other feature of running 'in atomic', ie, at soft or hard interrupt, is you cant access paged out memory or functions. Else you get a page fault exception.
|
|
|
|
|
I am trying to access rtsp url using QuickTime player. Now if there is some authentication set, then it asks for it otherwise it just plays the stream. It is good till here. But if I add username and password in the url, then it shows message: Server Not Found. I am passing the url in the following manner.
rtsp://username:password@IP-Address:Port
How could I pass rtsp url in QuickTime player with the credentials?. Thanks for any input
This world is going to explode due to international politics, SOON.
|
|
|
|
|
|
Sharath C V wrote: Mobile Broadband
Is that the same as 'wireless WAN' ?
|
|
|
|
|
|
I cant see any mention up updates to it on that link.
I actually worked with Microsoft on the development of this API starting off on Vista. It didn't make the cut for that OS but came in on Win7.
Its complex. All that home network, roaming network, signal strength, PIN entry etc. I hope they haven't made it worse! 
|
|
|
|
|
Thanks very much for your reply Matt.
And great to know that you worked on WWAN APIs!
No, the APIs are nice. It is pretty nice as to how they have pulled it off. Finding it really helpful. Since Microsoft did that homework, our live is better and we can spend quality time on other work items.
-Sharath
|
|
|
|
|
Hi all:
I am new one in wdf. I got a problem.
I write a usb hid touch driver in xp(send one touch as mouse).
My problem is, when the usb touch device is reset(the reset command are not
sent by host), How can I send the reset command from usb touch device to host
and notify Os get descriptor again and re addressing?
My English is not good, please forgive me.
best regard, Victor
|
|
|
|
|
My friend recently formatted his pc (only C drive) and re-installed XP. Now he faces a weird problem. The usb ports are working fine but it's not reading any pen drives other than his own. At first I found it hard to beleive but that's what is happening. His pen drive works fine but none of the others do. Any help?
|
|
|
|
|
Open the device manager, go to Disk Management and lok for the drives (see also http://support.microsoft.com/kb/309000[^]). They might just lack a drive letter.
Otherwise, look in device manager for "unknown devices".
|
|
|
|
|
Only XP Service Pack 2 has built-in support for USB drives. Tell your friend to install Service Pack 2, and the drives will work.
The difficult we do right away...
...the impossible takes slightly longer.
|
|
|
|
|
What a pain in the but it is for getting a driver signed.
Why is it so dependent on hardware? If the hardware hasn't got a serial number, fail. Device descriptor is a bit shoddy at times, fail. Device goes unresponsive. BSOD and fail.
Not only that its also dependent on the PC its tested on. One of them has a failure as big as the titanic iceberg right in the ACPI/BIOS subsystem.
So the Windows kernel is a highly structured and layered environment, and the driver you are testing sits on at least two others, and is a long way from the hardware, so why does the HCK fail on hardware errors? Why doesn't it implement a virtual piece of hardware via a filter, which it has anyway, and test JUST your driver?
It really is a pain in the but at times. Especially if the hardware is a bit flaky.
"The whole idea that carbon dioxide is the main cause of the recent global warming is based on a guess that was proved false by empirical evidence during the 1990s." climate-models-go-cold
|
|
|
|
|
I have a HP Pavilion Elite with dual monitors. I'd like to add a third monitor. Does this require a new video card or some other hardware?
If it's not broken, fix it until it is
|
|
|
|
|
Never done this myself... but you can probably find video cards that support it. When you purchase a new video card however, make sure it supports three simultaneous video outputs (just having three outputs doesn't guarantee that it's support it). I have a desktop that has three outputs (two DVI and one HDMI) but you can only use two of them at a time. Not sure if this is a hardware limitation or just the driver software. Either way, make sure it's supported by the video card.
|
|
|
|
|
Hello. AFAIK, you can stream h264 live video from IP camera on windows media player. But when I enter URL (rtsp://xxx.xxx.xxx.xxx:7070), following message is shown
Windows Media Player can not play the file because the specified protocol is not supported.
Now the same url works fine on VLC player but NOT on windows media plyer. I have read the following knowledge base article but could not fix the issue. What could be wrong? Thanks for any pointer.
modified 14-Apr-14 2:23am.
|
|
|
|
|
|
I read most of them but could not find the solution. Most of them are repetition of the KB article I provided the link of. Just not sure what should I try now.
|
|
|
|
|
Why doesn't the Bluetooth for android allow hands free talk and text like the windows does?
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
