Hi all,
OBS1: i'm not a specialist in C.
running the ADC example we have this code in configuration function. I dont know if this lib package is distributed by keil or ST. I received it from ST people with some keil examples.
#else /* VECT_TAB_FLASH */ /* Set the Vector Table base location at 0x08000000*/ NVIC_SetVectorTable(NVIC_VectTab_FLASH, 0x0); #endif
the NVIC_SetVectorTable function is:
void NVIC_SetVectorTable(u32 NVIC_VectTab, u32 Offset) { /* Check the parameters */ assert(IS_NVIC_VECTTAB(NVIC_VectTab)); assert(IS_NVIC_OFFSET(Offset)); SCB->ExceptionTableOffset = NVIC_VectTab | (Offset & (u32)0x1FFFFF80); }
so, compiling this code with keil it is a 4798 bytes long. But if i hide the function call and write all the code inside the function...
#else /* VECT_TAB_FLASH */ /* Set the Vector Table base location at 0x08000000 */ //NVIC_SetVectorTable(NVIC_VectTab_FLASH, 0x0); assert(IS_NVIC_VECTTAB(NVIC_VectTab_FLASH)); assert(IS_NVIC_OFFSET(0x0)); SCB->ExceptionTableOffset = NVIC_VectTab_FLASH | (0x0 & (u32)0x1FFFFF80); #endif
... the code is 3682 bytes long. I think that 1k for this function is so much. Can anyone explain what is happening? (forget about the assert function because i disabled #DEBUG)
thanks!
Andy and Per,
maybe you dont understand the main job of a forum. This is not to put comments like "read the manual", it is for share experience, do you understand?? Manuals do not share any experience with anyone. Sorry keil group, but at least a moderator from keil should ban you both in my opinion. All your poor answers is "read the man..." Its a little bit arrogant. And a manual link is the easiest thing to find when you want one. So, i dismiss your useless posts!! And this is the last time that i reply you.
McNamara, i tried using optimization options like the "One ELF Section per Function". Thanks!! Result:
1 Without this option: a ) With function call (debug off) = 9974 b ) Just the command line (debig off) = 7706 2 With "One ELF..." a ) With function call (debug off) = 4332 b ) Just the command line (debig off) = 4228
Now its ok!! 4 bytes to write a simple reg!!
Thanks a lot McNamara!!
Per and Andy watch and learn!!
"This is not to put comments like 'read the manual'"
There is no point in posting stuff that just repeats what is already clearly stated in the manual, is there?
"it is for share experience, do you understand??"
Yes, I understand. Where a post indicates that the poster has made an effort, they will get good responses.
But where a post clearly indicates that the poster has made no effort at all, they will get "read the manual".
Such posts are just lazy - they don't need "shared experience"; they just need to do the basic groundwork! There is no point in just posting what is already written in the manual!
For example if a post just says, "How do I interface 'C' to assembler?" when there is a whole section in the manual clearly titled "Interfacing 'C' to Assmbler" does that post deserve anything other than a "Read The Manual" response?
However, if the post says, "I read about Interfacing 'C' to Assmebler, but don't understand XYZ..." then, again, they will get help.
In your case, you got the "shared experience" in pointing out the different types of listing files - you said yourself that this was, "very helpfull". You hadn't actually stated which toolset you are using, so I couldn't give you specific instructions on how to enable the Linker listing, could I? (there are 3 possible different ARM toolsets relevant to this forum).
So, if you pick up 100 of my posts and look at them - exactly how many references to "read the manual" do you think you will find?
When the OP has a question where there is a possibility to "share experience", my posts contains very specific information about what to think about. No, I do absolutely refuse to read other peoples data sheets, but my answers very specifically mentions what to think about in generic terms.
The only post you have received from me in this thread is a note that this isn't the official support channel, and that the people here are not payed. That must really have hurt you, since you immediately suggests banning...
A quick check of the threads currently on the first page: 10160: I wrote how I usually use common global memory to share between different seldom used functions. I also makes a not about a new alias being used. 10151: I mentioned that people here are not payed. 10155: I wrote that the code was not complete, and not readable. I also mentioned a test to deduce if the problem was in read or write code. 10161: I mentioned that code size was not the size of object files. 10166: I asked a couple of questions about how Hebrew characters were used in the project. 10117: Huge amounts of "sharing experience" about use of ADC.
That completes the list of threads that I - to my knowledge - have posted in for the first page of threads. Not a reference to "read the manual". In 10117, I have one or two posts telling how problems are normally solved, but mentioning that the specificS of the individual chip requires the OP to read the data sheet for that chip. I think not being paid gives me the right to not post debugged turn-key source code.
Obviously a reason for banning, since I'm not helpfull...