Regarding UVM_BACKDOOR Write

Hi ,

When performing a Backdoor-write I observe that the UVM Base Class Library internally does a Backdoor-Read ( via task backdoor_read ) first
followed by Backdoor-Write ( via task backdoor_write )

Logically I am confused why a Backdoor-Read is required prior to Backdoor-Write ? Why doesn’t it directly perform the Backdoor-Write ?

In reply to MICRO_91:

backdoor_read and backdoor_write are function defined by the user. Please look to the corresonding code.
And try peek and pooke for backdoor reading and writing.

In reply to chr_sue:

Class uvm_reg has tasks backdoor_read and backdoor_write which are called internally when user initiates a Backdoor Write .

Refer Lines 2224 and 2248 in uvm_reg

I am using UVM built-in backdoor and not user-defined Backdoor . As a result the above 2 tasks get called .

I was trying to understand the reason behind doing a backdoor_read ( although I initiate Backdoor Write transaction ) prior to backdoor_write.

Reason of using Backdoor write over peek is that the Write is done on basis of access policy of the field(s) unlike peek .