Re: [Openocd-development] pointers modification within functions
On Thu, Nov 3, 2011 at 1:43 AM, Tomek CEDRO wrote:
> Hello Andreas :-)
>
> On Thu, Nov 3, 2011 at 12:36 AM, Andreas Fritiofson
> wrote:
> > This won't even compile. You pass a pointer-to-int, but swd_bus_read_ack
> > expects a pointer-to-pointer-to-char.
>
> naah this is only typo in mind-shortcut, code builds well, but i dont
> get it why i cannot use single pointer to pass back a memory location
> to function caller... this is what pointers exist..
>
>
Ah, well, but c passes parameters by value, to get out parameters you have
to go via a pointer.
Consider:
void foo(int x)
{
x++;
}
int main(void)
{
int a = 3;
foo(a);
return 0;
}
Did you expect the value of a to change after the call to foo? No, didn't
think so. You'd have to do this instead:
void foo(int *x)
{
*x++;
}
int main(void)
{
int a = 3;
foo(&a);
return 0;
}
Here a gets the value 4 after the call to foo.
Now change the variable type to a pointer instead of an integer:
void foo(int **x)
{
*x++;
}
int main(void)
{
int *a = &bar;
foo(&a);
return 0;
}
Here a points to the integer following bar after the call to foo. It's
exactly the same concept, just a change of variable type from int to int*.
Note that &a is now a double pointer, nothing magic here. Do the same
variable type change in the first, non-functional, example and convince
yourself that that wont work either.
___
Openocd-development mailing list
Openocd-development@lists.berlios.de
https://lists.berlios.de/mailman/listinfo/openocd-development
Re: [Openocd-development] pointers modification within functions
Hello Andreas :-)
On Thu, Nov 3, 2011 at 12:36 AM, Andreas Fritiofson
wrote:
> This won't even compile. You pass a pointer-to-int, but swd_bus_read_ack
> expects a pointer-to-pointer-to-char.
naah this is only typo in mind-shortcut, code builds well, but i dont
get it why i cannot use single pointer to pass back a memory location
to function caller... this is what pointers exist..
i have constructed this simple program to test if this can work - it
can but i need to pass a address of a variable and write value to that
variable, playing with pointers require double pointer :-( ill try
this on more complex application :-)
#include
#include
int f2(int *pf21, int *pf22){
int *stuff;
stuff=(int*)calloc(1,sizeof(int));
if (!stuff) exit(-1);
*stuff=0xDEADBEEF;
printf("stuff[@%X]=%X\n", stuff, *stuff);
*pf21=stuff;
*pf22=*stuff;
return 0;
}
int f1(int *pf11, int *pf12){
return f2(pf11, pf12);
}
int main(){
int a1=0,a2=0;
f1(&a1,&a2);
printf("a1[@%X]=%X\na2[@%X]=%X\n", &a1, a1, &a2, a2);
return 0;
}
--
CeDeROM, SQ7MHZ, http://www.tomek.cedro.info
___
Openocd-development mailing list
Openocd-development@lists.berlios.de
https://lists.berlios.de/mailman/listinfo/openocd-development
Re: [Openocd-development] pointers modification within functions
On Thu, Nov 3, 2011 at 12:41 AM, Tomek CEDRO wrote:
> Hey,
>
> I have some problems with pointers and need some support plz ;-)
>
I'm not sure I understand what the problem is, but I can give some general
hints.
int swd_bus_read_ack(swd_ctx_t *swdctx, swd_operation_t operation, char
> **ack){
> swd_cmd_enqueue_miso_ack(swdctx, ack);
>
You discard the return value and one of the parameters, but perhaps this is
not the complete function?
}
>
> int swd_cmd_enqueue_miso_ack(swd_ctx_t *swdctx, char **ack){
> if (swdctx==NULL) return SWD_ERROR_NULLCONTEXT;
> int res;
> swd_cmd_t *cmd;
> cmd=(swd_cmd_t *)calloc(1,sizeof(swd_cmd_t));
>
As a side note, cmd = calloc(1, sizeof(*cmd)); is preferred, in case the
type of cmd changes later.
if (cmd==NULL) return SWD_ERROR_OUTOFMEM;
> if (ack!=NULL) *ack=&cmd->ack;
>
You probably want to wait with setting *ack until after checking res below.
If swd_cmd_enqueue fails, ack will point to free'd memory, but you discard
the return value in swd_bus_read_ack so the caller won't be able to tell.
Leave *ack unchanged unless the function succeeds.
cmd->bits=SWD_ACK_BITLEN;
> cmd->cmdtype=SWD_CMDTYPE_MISO_ACK;
> res=swd_cmd_enqueue(swdctx, cmd); //should be 1 on success
> if (res<1) free(cmd);
> return res;
> }
>
> main(){
> int *ack;
swd_bus_read_ack(swdctx, operation, ack);
>
This won't even compile. You pass a pointer-to-int, but swd_bus_read_ack
expects a pointer-to-pointer-to-char.
}
>
> The problem is:
> 1. I need to use double pointers to return back the address of the
> queue element (*ack=&cmd->ack).
>
Correct.
2. If I use single pointer *ack the value of the ack is only changed
> inside swd_cmd_enqueue_miso_ack() but after its return to
> swd_bus_read_ack() the ack value is taken back to the state as it was
> before swd_cmd_enqueue_miso_ack() call.
>
You have already concluded that you need a double pointer in 1, so I don't
know why you say this. A single pointer won't work.
3. I have tried to use single pointer *ack and call
> swd_bus_read_ack(swdctx, operation, &ack) but is changes nothing.
>
Ditto. You're reusing the same name (ack) for several things in your
description (two function parameters and a variable) so it's hard to
understand what you mean.
> This makes impossible to give back data on queue_dp_read(). There is
> something wrong with these pointers!!! Help! :-)
>
___
Openocd-development mailing list
Openocd-development@lists.berlios.de
https://lists.berlios.de/mailman/listinfo/openocd-development
[Openocd-development] pointers modification within functions
Hey,
I have some problems with pointers and need some support plz ;-)
int swd_bus_read_ack(swd_ctx_t *swdctx, swd_operation_t operation, char **ack){
swd_cmd_enqueue_miso_ack(swdctx, ack);
}
int swd_cmd_enqueue_miso_ack(swd_ctx_t *swdctx, char **ack){
if (swdctx==NULL) return SWD_ERROR_NULLCONTEXT;
int res;
swd_cmd_t *cmd;
cmd=(swd_cmd_t *)calloc(1,sizeof(swd_cmd_t));
if (cmd==NULL) return SWD_ERROR_OUTOFMEM;
if (ack!=NULL) *ack=&cmd->ack;
cmd->bits=SWD_ACK_BITLEN;
cmd->cmdtype=SWD_CMDTYPE_MISO_ACK;
res=swd_cmd_enqueue(swdctx, cmd); //should be 1 on success
if (res<1) free(cmd);
return res;
}
main(){
int *ack;
swd_bus_read_ack(swdctx, operation, ack);
}
The problem is:
1. I need to use double pointers to return back the address of the
queue element (*ack=&cmd->ack).
2. If I use single pointer *ack the value of the ack is only changed
inside swd_cmd_enqueue_miso_ack() but after its return to
swd_bus_read_ack() the ack value is taken back to the state as it was
before swd_cmd_enqueue_miso_ack() call.
3. I have tried to use single pointer *ack and call
swd_bus_read_ack(swdctx, operation, &ack) but is changes nothing.
This makes impossible to give back data on queue_dp_read(). There is
something wrong with these pointers!!! Help! :-)
Tomek
--
CeDeROM, SQ7MHZ, http://www.tomek.cedro.info
___
Openocd-development mailing list
Openocd-development@lists.berlios.de
https://lists.berlios.de/mailman/listinfo/openocd-development
Re: [Openocd-development] openocd patch: 96c3cbf cortex_m: add missing error checking
On Tue, Nov 1, 2011 at 11:16 PM, Tomek CEDRO wrote: > Hey does this code is related anyhow to arm_adi_v5 or this is totally > alternative solution? > > Yes, it *uses* arm_adi_v5. /Andreas ___ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
Re: [Openocd-development] include linux/scripts in OpenOCD project?
On Wed, Nov 2, 2011 at 8:09 PM, Øyvind Harboe wrote: > On Wed, Nov 2, 2011 at 9:06 PM, Tomek CEDRO wrote: >> On Wed, Nov 2, 2011 at 8:03 PM, Øyvind Harboe >> wrote: >>> Should we? >> >> What does these scripts do? Will they work on FreeBSD? :-P :-) > > They allow you to check patches for coding style, remove whitespace, > automatically indent, etc. Aaah, ok, I finally upgraded my hardware so I can run VM and test stuff on/for other operating systems, this might be good option to have everything in one place :-) -- CeDeROM, SQ7MHZ, http://www.tomek.cedro.info ___ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
Re: [Openocd-development] include linux/scripts in OpenOCD project?
On Wed, Nov 2, 2011 at 9:06 PM, Tomek CEDRO wrote: > On Wed, Nov 2, 2011 at 8:03 PM, Øyvind Harboe wrote: >> Should we? > > What does these scripts do? Will they work on FreeBSD? :-P :-) They allow you to check patches for coding style, remove whitespace, automatically indent, etc. Don't know about FreeBSD... -- Øyvind Harboe - Can Zylin Consulting help on your project? US toll free 1-866-980-3434 http://www.zylin.com/ ___ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
Re: [Openocd-development] include linux/scripts in OpenOCD project?
On Wed, Nov 2, 2011 at 8:03 PM, Øyvind Harboe wrote: > Should we? What does these scripts do? Will they work on FreeBSD? :-P :-) -- CeDeROM, SQ7MHZ, http://www.tomek.cedro.info ___ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
[Openocd-development] include linux/scripts in OpenOCD project?
Should we? -- Øyvind Harboe - Can Zylin Consulting help on your project? US toll free 1-866-980-3434 http://www.zylin.com/ ___ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
[Openocd-development] openocd patch: 2fb67e0 tx27stk5: add board init support for KaRo TX27 CPU module
This is an automated email from Gerrit.
Mark Vels (mark.v...@team-embedded.nl) just uploaded a new patch set to Gerrit,
which you can find at http://openocd.zylin.com/158
-- gerrit
commit 2fb67e0d442b3e00e81df5fadd4eedc930e3bd6d
Author: Mark Vels
Date: Tue Nov 1 13:29:11 2011 +0100
tx27stk5: add board init support for KaRo TX27 CPU module
This patch adds support for a KaRo TX27 CPU module on a StarterkitV base
board.
The register settings have been extracted from a RedBoot distribution
that is distributed along with the hardware by KaRo.
This setup has been tested with a JTAGKey. The testing has been focussed
on loading a program into memory and start execution.
Although the flash seems to be correctly detected, no effort has been put
in testing the NAND programming yet.
Change-Id: Ib17763f1e3ecacd0eb9b5fdc32f8cba7a5e59be5
Signed-off-by: Mark Vels
diff --git a/tcl/board/tx27_stk5.cfg b/tcl/board/tx27_stk5.cfg
new file mode 100644
index 000..2a1c8b8
--- /dev/null
+++ b/tcl/board/tx27_stk5.cfg
@@ -0,0 +1,60 @@
+source [find target/imx27.cfg]
+
+$_TARGETNAME configure -event gdb-attach { reset init }
+$_TARGETNAME configure -event reset-init { tx27_init }
+
+proc tx27_init { } {
+ # This setup puts RAM at 0xA000
+ # init_aipi (AIPI1.PSR0, AIPI2.PSR0, AIPI1.PSR1 and AIPI2.PSR1)
+ mww 0x1000 0x20040304
+ mww 0x1002 0x
+ mww 0x1004 0xDFFBFCFB
+ mww 0x10020004 0x
+
+ sleep 100
+
+ #init_max ( PORT0.MPR, #PORT0.AMPR, #PORT1.MPR, #PORT1.AMPR,
#PORT2.MPR, #PORT2.AMPR)
+ mww 0x1003F000 0x00302145
+ mww 0x1003F004 0x00302145
+ mww 0x1003F100 0x00302145
+ mww 0x1003F104 0x00302145
+ mww 0x1003F200 0x00302145
+ mww 0x1003F204 0x00302145
+
+ #init_drive_strength (#DSCR3, #DSCR5, #DSCR6, #DSCR7, #DSCR8 )
+ mww 0x10027828 0x
+ mww 0x10027830 0x
+ mww 0x10027834 0x
+ mww 0x10027838 0x5005
+ mww 0x1002783C 0x1555
+
+ #init_sdram_speed
+ #mww 0xD8001010 0x0004
+ mww 0xD8001010 0x0024
+
+ mww 0xD8001004 0x00395729
+
+ mww 0xD8001000 0x9212
+ mww 0xA400 0x0
+
+ mww 0xD8001000 0xA212
+ mww 0xA000 0x0
+ mww 0xA000 0x0
+
+ mww 0xD8001000 0xB212
+ mdb 0xA000
+ mdb 0xA033
+
+ mww 0xD8001000 0x82126485
+
+ # =
+ # Sync mode (AHB Clk = 133MHz ; BCLK = 44.3MHz)
+ # =
+ mww 0xD8002000 0x23524E80
+ mww 0xD8002004 0x1D03
+ mww 0xD8002008 0x00720900
+
+ nand probe 0
+}
+
+nand device tx27.nand imx27 $_TARGETNAME hwecc
--
___
Openocd-development mailing list
Openocd-development@lists.berlios.de
https://lists.berlios.de/mailman/listinfo/openocd-development
[Openocd-development] openocd patch: aa9bc05 HACKING: all you need is http access
This is an automated email from Gerrit.
?yvind Harboe (oyvindhar...@gmail.com) just uploaded a new patch set to Gerrit,
which you can find at http://openocd.zylin.com/157
-- gerrit
commit aa9bc05890f74e51599e38d5ad195bb7af848261
Author: Ãyvind Harboe
Date: Wed Nov 2 10:48:44 2011 +0100
HACKING: all you need is http access
Change-Id: I191c1da5126c4c9ea1ff8826576b6b24feaf9881
Signed-off-by: Ãyvind Harboe
diff --git a/HACKING b/HACKING
index dc06b45..353e72f 100644
--- a/HACKING
+++ b/HACKING
@@ -1,3 +1,6 @@
+NB! If you're behind a corporate wall with http only access to the
+world, you can still use these instructions!
+
Submitting patches to the OpenOCD Gerrit server:
OpenOCD is to some extent a "self service" open source project, so to
@@ -10,7 +13,7 @@ The procedure to create a patch is essentially:
- create a commit
- send the changes to the Gerrit server for review
- correct the patch and re-send it according to review feedback
-
+
0. Create a Gerrit account at:
@@ -43,7 +46,7 @@ you should scroll down the page till you get to the section:
'Next: Set Up SSH Keys', and follow the steps described.
1. Clone the git repository, rather than just
-download the source.
+download the source.
git clone git://openocd.git.sourceforge.net/gitroot/openocd/openocd
@@ -60,24 +63,35 @@ Add a new remote to git using Gerrit username:
git remote add review ssh://usern...@openocd.zylin.com:29418/openocd.git
git config remote.review.push HEAD:refs/for/master
+Or with http only:
+
+git remote add review http://openocd.zylin.com/p/openocd.git
+git config remote.review.push HEAD:refs/for/master
+
You will need to install this hook, we will look into a better
solution:
scp -p -P 29418 usern...@openocd.zylin.com:hooks/commit-msg .git/hooks/
+Or with http only:
+
+wget http://openocd.zylin.com/tools/hooks/commit-msg
+mv commit-msg .git/hooks
+chmod +x .git/hooks/commit-msg
+
3. Set up git with your name and email:
git config --global user.name "John Smith"
git config --global user.email "j...@smith.org"
-4. Work on your patches. Split the work into
+4. Work on your patches. Split the work into
multiple small patches that can be reviewed and
applied seperately and safely to the OpenOCD
repository.
while(!done) {
work - edit files using your favorite editor.
- run "git commit -s -a" to commit all changes.
+ run "git commit -s -a" to commit all changes.
}
TIP! use "git add ." before commit to add new files.
--
___
Openocd-development mailing list
Openocd-development@lists.berlios.de
https://lists.berlios.de/mailman/listinfo/openocd-development
Re: [Openocd-development] Gerrit with http only
On 2 November 2011 09:24, Øyvind Harboe wrote: > Setting up the commit-msg hook isn't possible using http only: > > scp -p -P 29418 usern...@openocd.zylin.com:hooks/commit-msg .git/hooks/ > > http://repo.or.cz/w/openocd.git/blame?f=HACKING > > Any suggestions? > http://openocd.zylin.com/tools/hooks/commit-msg Spen ___ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
[Openocd-development] Gerrit with http only
Setting up the commit-msg hook isn't possible using http only: scp -p -P 29418 usern...@openocd.zylin.com:hooks/commit-msg .git/hooks/ http://repo.or.cz/w/openocd.git/blame?f=HACKING Any suggestions? -- Øyvind Harboe - Can Zylin Consulting help on your project? US toll free 1-866-980-3434 http://www.zylin.com/ ___ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
