|
From: | Joel Sherrill |
Subject: | Re: [Simulavr-devel] anacomp getting reserved access messages |
Date: | Thu, 26 Mar 2009 11:35:49 -0500 |
User-agent: | Thunderbird 2.0.0.21 (X11/20090320) |
FYI the avr .h file has this: io4433.h:/* 0x3D..0x3E SP */ Is this by convention? Or is the simulator missing something? I am guessing gcc, is doing frame pointer management. --joel Joel Sherrill wrote:
Schwichtenberg, Knut wrote:RTFM helps :-)! The gcc-version you are using is inserting code that chages reserved cells! Reading the datasheet makes it clear 0x5e is Reserved! 0x5f is SREG and 0x5d is SP (not SPL/SPH) - so this depricated CPU shows cells that are marked reserved, gcc inserts code that changes them and simulavrxx detects it. Everything is fine!OK. But any idea why this instruction is hitting a reserved address 0x5e? I must admit I don't know the CPU 58: cd b7 in r28, 0x3d ; 61 5a: de b7 in r29, 0x3e ; 62 And what causes this code to be generated? I don't see anything that corresponds to it in main.c. Does this then qualify as a gcc bug? $ avr-gcc --version avr-gcc (GCC) 4.3.3 $ rpm -q avr-gcc avr-gcc-4.3.3-1.fc10.i386Fast work-around: Change the CPU to M128 and ensure you also change the analog comparator pin accordingly.You guys assume I know more about the AVR than I do. LOL! This is my first real AVR project. RTEMS experiences are just making the build and Tcl work go quickly.No simulavrxx error!I am glad to hear that we don't have a simulavr bug here or in something I changed.-----Original Message-----From: Joel Sherrill [mailto:address@hidden Sent: Thursday, March 26, 2009 2:12 PMTo: Schwichtenberg, Knut Cc: address@hidden Subject: Re: [Simulavr-devel] anacomp getting reserved access messages I captured a trace and deleted all of the stuff below the error. It appears to be happening very early in main.c. Look for RESERVED. --joel Joel Sherrill wrote:RAM-Area of the AVR is identicall to 0x00 in the IO-Space which is forbidden by most AVR's. This loop puts normally 10.000 "*" on you terminal. Maybe it's in there to delay or only to confuse the maintainer ;-).Schwichtenberg, Knut wrote:Joel,here you find a usage for "-W 0x20,-". 0x20 in theI thought of that but there was no way to turn that on via Tcluntil I did my work. So I guess this is just a left overhunk of code.BTW: The comment block at the top of the source is fullynonsense in conjunction with the source code.:)The 0x5e - I'm guessing - might have to do with compilingthe source for CPU A and using it on CPU B, is it?The reserved message can also come out when the CPU .cpp file has temporarily marked it as such. set dev1 [new_AvrDevice_at90s4433] And this code is in that cpu rw[0x5e]= new RWReserved(this); Is this a known register name?Knut----------------------------------------------------------------------This mail is made from 100% recycled electrons._______________________________________________ Simulavr-devel mailing list address@hidden http://lists.nongnu.org/mailman/listinfo/simulavr-devel-- Joel Sherrill, Ph.D. Director of Research & Development address@hidden On-Line Applications Research Ask me about RTEMS: a free RTOS Huntsville AL 35805 Support Available (256) 722-9985
-- Joel Sherrill, Ph.D. Director of Research & Development address@hidden On-Line Applications Research Ask me about RTEMS: a free RTOS Huntsville AL 35805 Support Available (256) 722-9985
[Prev in Thread] | Current Thread | [Next in Thread] |