Login

Important information

This site uses cookies to store information on your computer. By continuing to use our site, you consent to our cookies.

ARM websites use two types of cookie: (1) those that enable the site to function and perform as required; and (2) analytical cookies which anonymously track visitors only while using the site. If you are not happy with this use of these cookies please review our Privacy Policy to learn how they can be disabled. By disabling cookies some features of the site will not work.

ARM Community: Exclusive Access in AXI - ARM Community

Jump to content

Page 1 of 1
  • You cannot start a new topic
  • You cannot reply to this topic

Exclusive Access in AXI Rate Topic: -----

#1 User is offline   paragsathe 

  • Member
  • Pip
  • Group: Members
  • Posts: 1
  • Joined: 18-January 12

Posted 18 January 2012 - 01:17 PM

Hi all,

I had a few doubts in case of exclusive access in AXI protocol.

1) Consider the following series of operation

Exclusive read to address A
Normal Write to address A (What will be the response here)
Exclusive write to address A (What will be the response here)

2) Is exclusive access compulsarily for exclusive read followed by exclusive write, where the exclusive write may occur at a later point of time without blocking the bus?

3) Does the monitoring of exclusive addresss end after a pair of exclusive read-write is done?

4) How many such exclusive addresses can be there in a single master and a multiple master system respectively?

All the AXI experts out there.....please answer my doubts...

Thanks
0

#2 User is offline   JD 

  • Regular Contributor
  • PipPipPip
  • Group: Members.
  • Posts: 260
  • Joined: 17-November 06

Posted 19 January 2012 - 09:48 AM

Hello paragsathe

1.

Response to the normal write would be OKAY (or SLVERR if the slave is unhappy with the write for any reason).

Response then to the exclusive write MUST also be OKAY because the monitored address has been written to since the exclusive read.

2.

Note sure I understand the question. Exclusive accesses are a sequence of a read transaction and then a write transaction to the same address range. The delay between the read and write parts of the sequence is undefined, but the longer you leave it, the more chance there is that another write will corrupt the monitored address range.

3.

The need for monitoring the address ends when a write occurs to the monitored address range.

4.

You can have one active exclusive access per ID, so as masters can use multiple IDs the relationship is not determined directly by the number of masters.

If you are trying to work out how many monitors you need, the AXI spec states in section 6.2.3 that a slave should have "a monitor unit for exclusive-capable master ID that can access it.

JD
0

#3 User is offline   AndreAndre 

  • Member
  • Pip
  • Group: Members
  • Posts: 1
  • Joined: 13-July 12

Posted 13 July 2012 - 10:26 AM

I have another question.

Why exclusive access need to check AWID match ARID for the complete operation in addition to the same address?
Why can not only check if the exclusive write address match exclusive read address?
What's the reason for the requirement of same ID?

Thanks!



0

Share this topic:


Page 1 of 1
  • You cannot start a new topic
  • You cannot reply to this topic