|
Message
From: Uwe Bonnes <bon@e...>
Date: Tue, 10 Jun 2003 09:49:37 +0200
Subject: [pci] Icarus showing more problemetic code
Hallo,
this is the answer of Stephen Williams about a question for the code in
wb_bus_mon.v:332:
address <= {1'b1, (ADDR_O + `WB_SEL_WIDTH)} ;
> results in wb_bus_mon.v:332: error: operand of concatenation has indefinite
> width: (ADDR_O[31:0])+('sd4)
>
> Yes, this is an error in your code. This should be a FAQ, but it is
> certainly described in the ieee1364-notes.txt file that comes with
> Icarus Verilog distributions.
>
> You get away with it with most compilers, because most "XL bug
> compatible" compilers expand unsized expressions to 32 bits inside
> concatenations, but the expression is already 32bits wide. Icarus
> Verilog does not use or enforce a 32bit restriction on unsized
> values, so does report this error.
>
> Just be glad you're getting an error message instead of a chip that
> loses high bits with no warning:-)
Probably a wraparound is supposed to happen. Should I submit a patch?
Bye
--
Uwe Bonnes bon@e...
Institut fuer Kernphysik Schlossgartenstrasse 9 64289 Darmstadt
--------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------
|
 |
Browse
