lupyuen/local-nuttx-update
1
0
Fork 0
forked from nuttx/nuttx-update
Code Pull requests Activity

A few early fixes in kernel build testing

Browse source 
git-svn-id: svn://svn.code.sf.net/p/nuttx/code/trunk@5734 42af7a65-404d-4744-a932-0658087f49c3
This commit is contained in:
patacongo 2013-03-12 16:02:48 +00:00
parent 23c8c591e9
commit 73ef58fda6
8 changed files with 125 additions and 49 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
arch/arm/src/armv7-m/up_mpu.c
View file

@ -117,7 +117,7 @@ uint8_t mpu_log2regionsize(size_t size)
/* The minimum permitted region size is 16 bytes (log2(16) = 4. */
for (l2size = 4; l2size < 32 && size > (1 << l2size); size++);
for (l2size = 4; l2size < 32 && size > (1 << l2size); l2size++);
return l2size;
}

40
configs/open1788/README.txt
View file

@ -296,16 +296,16 @@ CONFIGURATION
----
This is identical to the nsh configuration below except that NuttX
is built as a kernel-mode, monolithic module and the user applications
are built separately. This build requires a special make command; not
just 'make' but make with the following two arguments:
are built separately. It is recommends to use a special make command;
not just 'make' but make with the following two arguments:
make pass1 pass2
This is required because in the normal case (just 'make'), make will
create all dependencies then execute the pass1 and pass2 targets. But
this example, pass2 depends on auto-generatd files produced during pass1.
This special make command ('make pass1 pass2') will make the dependencies
separately for each pass.
In the normal case (just 'make'), make will attempt to build both user-
and kernel-mode blobs more or less interleaved. This actual works!
However, for me it is very confusing so I prefer the above make command:
Make the user-space binaries first (pass1), then make the the kernel-space
binaries (pass2)
NOTES:
@ -324,15 +324,23 @@ CONFIGURATION
CONFIG_ARMV7M_TOOLCHAIN_BUILDROOT=y : Buildroot toolchain
CONFIG_ARMV7M_OABI_TOOLCHAIN=y : Older, OABI toolchain
3. At the end of the build, there four files will top-level build
directory:
3. At the end of the build, there will be several files in the top-level
NuttX build directory:
nuttx_user.elf - The pass1 user-space ELF file
nuttx - The pass2 kernel-space ELF file
nuttx_user.hex - The pass1 Intel HEX format file
nuttx.hex - The pass2 Intel HEX file
System.map - Symbols in the kernel-space ELF file
User.map - Symbols in the user-space ELF file
PASS1:
nuttx_user.elf - The pass1 user-space ELF file
nuttx_user.hex - The pass1 Intel HEX format file (selected in defconfig)
User.map - Symbols in the user-space ELF file
PASS2:
nuttx - The pass2 kernel-space ELF file
nuttx.hex - The pass2 Intel HEX file (selected in defconfig)
System.map - Symbols in the kernel-space ELF file
Loading these .elf files with OpenOCD is tricky. It appears to me
that when nuttx_user.elf is loaded, it destroys the the nuttx image
in FLASH. But loading the nuttx ELF does not harm the nuttx_user.elf
in FLASH. Conclusion: Always load nuttx_user.elf before nuttx.
nsh
---
@ -345,7 +353,7 @@ CONFIGURATION
change this configuration using that tool, you should:
a. Build and install the kconfig-mconf tool. See nuttx/README.txt
and misc/tools/
and misc/tools//README.txt.
b. Execute 'make menuconfig' in nuttx/ in order to start the
reconfiguration process.

29
configs/open1788/scripts/kernel-space.ld
View file

@ -33,33 +33,34 @@
*
****************************************************************************/
/* The AT91SAM3U-4 has 256Kb of FLASH beginning at address 0x0008:0000,
* 32Kb of SRAM beginning at address 0x2000:0000, and 16Kb of SRAM beginning
* at address 0x2008:000 (used only for heap). When booting from FLASH,
* FLASH memory is aliased to address 0x0000:0000 where the code expects to
* begin execution by jumping to the entry point in the 0x0800:0000 address
* range.
/* The LPC1788 has 512Kb of FLASH beginning at address 0x0000:0000 and
* 96Kb of total SRAM: 64Kb of SRAM in the CPU block beginning at address
* 0x10000000 and 32Kb of Peripheral SRAM in two banks, 8Kb at addresses
* 0x20000000 bank0 first and 8kb at 0x20002000 at bank0 second. And 16Kb
* at 0x20004000 on bank1.
*
* For MPU support, the kernel-mode NuttX section is assumed to be 64Kb of
* FLASH and 4Kb of SRAM. That, of course, can be optimized as needed (See
* also configs/open1788/scripts/kernel.ld).
* also configs/open1788/scripts/kernel-space.ld).
*/
MEMORY
{
/* 256Kb FLASH */
kflash (rx) : ORIGIN = 0x00080000, LENGTH = 64K
uflash (rx) : ORIGIN = 0x00090000, LENGTH = 192K
kflash (rx) : ORIGIN = 0x00000000, LENGTH = 64K
uflash (rx) : ORIGIN = 0x00010000, LENGTH = 448K
/* 32Kb SRAM */
/* 64Kb of SRAM in the CPU block */
ksram (rwx) : ORIGIN = 0x20000000, LENGTH = 4K
usram (rwx) : ORIGIN = 0x20001000, LENGTH = 28K
ksram (rwx) : ORIGIN = 0x10000000, LENGTH = 4K
usram (rwx) : ORIGIN = 0x10001000, LENGTH = 60K
/* 16Kb SRAM */
/* Other peripheral memory (free, nothing is linked here) */
sram2 (rwx) : ORIGIN = 0x20080000, LENGTH = 16K
ahbram8_b0a(rwx) : ORIGIN = 0x20000000, LENGTH = 8K
ahbram8_b0b(rwx) : ORIGIN = 0x20002000, LENGTH = 8K
ahbram16(rwx) : ORIGIN = 0x20004000, LENGTH = 16K
}
OUTPUT_ARCH(arm)

4
configs/open1788/scripts/user-space.ld
View file

@ -53,8 +53,8 @@ MEMORY
/* 64Kb of SRAM in the CPU block */
ksram (rwx) : ORIGIN = 0x20000000, LENGTH = 4K
usram (rwx) : ORIGIN = 0x20001000, LENGTH = 60K
ksram (rwx) : ORIGIN = 0x10000000, LENGTH = 4K
usram (rwx) : ORIGIN = 0x10001000, LENGTH = 60K
/* Other peripheral memory (free, nothing is linked here) */

46
configs/open1788/tools/open1788.cfg
View file

@ -29,6 +29,7 @@ set CPUROMSIZE 0x80000
set CCLK 12000
### From /usr/local/share/openocd/scripts/targets/lpc17xx.cfg
### With additions to set the operating frequency to 120MHz
# Common LPC17xx logic
# LPC17xx chips support both JTAG and SWD transports.
@ -100,6 +101,49 @@ flash bank $_FLASHNAME lpc2000 0x0 $_CPUROMSIZE 0 0 $_TARGETNAME \
adapter_khz 1000
$_TARGETNAME configure -event reset-init {
echo "Going to 120MHz"
# PLL0CON: Disable PLL
mww 0x400FC080 0x00000000
# PLLFEED
mww 0x400FC08C 0x000000AA
# PLLFEED
mww 0x400FC08C 0x00000055
# CLKSEL: internal 12MHz RC oscillator Div 1.
mww 0x400FC104 0x00000001
# CLKSRCSEL: Clock source = internal 12MHz RC oscillator
mww 0x400FC10C 0x00000000
# PLL0CFG: M=10,P=1 -> PLL=240 MHz
mww 0x400FC084 0x00000009
# PLLFEED
mww 0x400FC08C 0x000000AA
# PLLFEED
mww 0x400FC08C 0x00000055
# PLL0CON: Enable PLL
mww 0x400FC080 0x00000001
# PLLFEED
mww 0x400FC08C 0x000000AA
# PLLFEED
mww 0x400FC08C 0x00000055
sleep 50
# PLL0CON: Connect PLL
# CCLKSEL=PLLED(240MHz)/2 (=120 MHz)
mww 0x400FC104 0x00000102
# PLLFEED
mww 0x400FC08C 0x000000AA
# PLLFEED
mww 0x400FC08C 0x00000055
# Dividing CPU clock by 8 should be pretty conservative
#
#
adapter_khz 1500
# Do not remap 0x0000-0x0020 to anything but the flash (i.e. select
# "User Flash Mode" where interrupt vectors are _not_ remapped,
# and reside in flash instead).
@ -114,7 +158,7 @@ $_TARGETNAME configure -event reset-init {
#
# http://ics.nxp.com/support/documents/microcontrollers/?scope=LPC1768&type=user
mww 0x4ba00477 0x01
mww 0x400FC040 0x01
}
# if srst is not fitted use VECTRESET to

50
configs/sam3u-ek/README.txt
View file

@ -378,26 +378,48 @@ must be is one of the following:
knsh:
This is identical to the nsh configuration below except that NuttX
is built as a kernel-mode, monolithic module and the user applications
are built separately. This build requires a special make command; not
just 'make' but make with the following two arguments:
are built separately. It is recommends to use a special make command;
not just 'make' but make with the following two arguments:
make pass1 pass2
This is required because in the normal case (just 'make'), make will
create all dependencies then execute the pass1 and pass2 targets. But
this example, pass2 depends on auto-generatd files produced during pass1.
This special make command ('make pass1 pass2') will make the dependencies
separately for each pass.
In the normal case (just 'make'), make will attempt to build both user-
and kernel-mode blobs more or less interleaved. This actual works!
However, for me it is very confusing so I prefer the above make command:
Make the user-space binaries first (pass1), then make the the kernel-space
binaries (pass2)
At the end of the build, there four files will top-level build
directory:
NOTES:
1. This configuration uses the mconf-based configuration tool. To
change this configuration using that tool, you should:
nuttx_user.elf - The pass1 ELF file
nuttx - The pass2 ELF file
nuttx_user.hex - The pass1 Intel HEX format file
nuttx.hex - The pass2 Intel HEX file
a. Build and install the kconfig-mconf tool. See nuttx/README.txt
and misc/tools/README.txt.
The J-Link program will except files in .hex, .mot, .srec, and .bin
b. Execute 'make menuconfig' in nuttx/ in order to start the
reconfiguration process.
2. Uses the older, OABI, buildroot toolchain. But that is easily
reconfigured:
CONFIG_ARMV7M_TOOLCHAIN_BUILDROOT=y : Buildroot toolchain
CONFIG_ARMV7M_OABI_TOOLCHAIN=y : Older, OABI toolchain
3. At the end of the build, there will be several files in the top-level
NuttX build directory:
PASS1:
nuttx_user.elf - The pass1 user-space ELF file
nuttx_user.hex - The pass1 Intel HEX format file (selected in defconfig)
User.map - Symbols in the user-space ELF file
PASS2:
nuttx - The pass2 kernel-space ELF file
nuttx.hex - The pass2 Intel HEX file (selected in defconfig)
System.map - Symbols in the kernel-space ELF file
The J-Link programmer will except files in .hex, .mot, .srec, and .bin
formats.
nsh:

1
include/nuttx/kmalloc.h
View file

@ -43,6 +43,7 @@
#include <nuttx/config.h>
#include <sys/types.h>
#include <stdbool.h>
#include <stdlib.h>
#include <nuttx/mm.h>

2
sched/group_leave.c
View file

@ -182,7 +182,7 @@ static inline void group_release(FAR struct task_group_s *group)
/* Free resource held by the stream list */
#if defined(CONFIG_NUTTX_KERNEL) && defined(CONFIG_MM_KERNEL_HEAP)
lib_releaselist(&group->tg_streamlist);
lib_releaselist(group->tg_streamlist);
#else
lib_releaselist(&group->tg_streamlist);
#endif