Skip to contentSkip to conversation

mips: PMC MSP71xx mips common

PATCH Marc St-Jean 4 participants 5 messages Feb 27, 2007 – Feb 28, 2007 MIPS

lore
Share & tools
lore
#0 lore raw
Quoted 43 lines · Andrew Sharp
Quoted 26 lines · On Mon, 26 Feb 2007 18:12:55 -0600 Marc St-Jean <[email protected]> wrote:

[PATCH 2/5] mips: PMC MSP71xx mips common

Patch to add mips common support for the PMC-Sierra MSP71xx devices.

These 5 patches along with the previously posted serial patch will boot the PMC-Sierra MSP7120 Residential Gateway board.

Thanks,
Marc

---
Re-posting patch with two recommended changes:
1. Dropped the PMC_MSP_UNCACHED configuration item as this was already available in arch/mips/Kconfig.debug.
3. Dropped 'else' case to simplify patch to do_watch() in arch/mips/kernel/traps.c

 arch/mips/Kconfig           |   78 ++++++++++++++++++++
 arch/mips/Makefile          |    8 ++
 arch/mips/kernel/head.S     |    8 ++
 arch/mips/kernel/traps.c    |    6 +
 include/asm-mips/bootinfo.h |   12 +++
 include/asm-mips/mipsregs.h |   30 +++++++
 include/asm-mips/regops.h   |  168
Quoted 2 lines

++++++++++++++++++++++++++++++++++++++++++++
include/asm-mips/war.h | 11 ++ 8 files changed, 321 insertions(+)

My mailer kind of made a mess of things, hope I caught them all.

arch/mips/kernel/head.S line 133

Quoted 9 lines
arch/mips/kernel/head.S +1 -0
index 6f57ca4..d7451b1 100644
@@ -130,10 +130,18 @@
      .endm
 
      /*
+      * Reserverd space not required for PMC boards, although we 

need to

Quoted 15 lines

arch/mips/kernel/head.S line 144

Quoted 11 lines
arch/mips/kernel/head.S +7 -0
+      * jump to kernel start.
+      */
+#ifdef CONFIG_PMC_MSP
+     jal     kernel_entry
+#else
+     /*
       * Reserved space for exception handlers.
       * Necessary for machines which link their kernels at KSEG0.
       */
      .fill   0x400
+#endif /* CONFIG_PMC_MSP */

This is getting kind of ugly. There are a whole lot of config choices that need to use the 'j kernel_entry'. Do they all have to have their own? I'm not sure what the best way is to handle them all.

I agree but don't know the best way to handle this. I could introduce a SYS_NO_EXEPT_FILL or similar flag but this seems excessive.

Any other ideas from arch/mips folks?

I'm not sure why the fill is needed before _stext. The comment states kernels linked against kseg0 require this. We link our kernels to 0x8010_0000 and expect that to be the start of text. We aren't changing anything else in the startup code.

Quoted 25 lines

include/asm-mips/bootinfo.h line 218

Quoted 11 lines
include/asm-mips/bootinfo.h +4 -0
index c7c945b..ab29fd4 100644
@@ -213,6 +213,18 @@
 #define MACH_GROUP_NEC_EMMA2RH 25    /* NEC EMMA2RH (was 23)         */
 #define  MACH_NEC_MARKEINS    0       /* NEC EMMA2RH Mark-eins        */
+/*
+ * Valid machtype for group PMC-MSP
+ */
+#define MACH_GROUP_MSP         23    /* PMC-Sierra MSP 

boards/CPUs */

include/asm-mips/bootinfo.h line 219

Quoted 1 line
include/asm-mips/bootinfo.h +1 -0
+#define MACH_MSP4200_EVAL       0    /* PMC-Sierra MSP4200 

Evaluation board */

include/asm-mips/bootinfo.h line 220

Quoted 1 line
include/asm-mips/bootinfo.h +1 -0
+#define MACH_MSP4200_GW         1    /* PMC-Sierra MSP4200 Gateway 

demo board */

include/asm-mips/bootinfo.h line 221

Quoted 1 line
include/asm-mips/bootinfo.h +1 -0
+#define MACH_MSP4200_FPGA       2    /* PMC-Sierra MSP4200 Emulation 

board */

include/asm-mips/bootinfo.h line 222

Quoted 1 line
include/asm-mips/bootinfo.h +1 -0
+#define MACH_MSP7120_EVAL       3    /* PMC-Sierra MSP7120 

Evaluation board *
/

include/asm-mips/bootinfo.h line 223

Quoted 1 line
include/asm-mips/bootinfo.h +1 -0
+#define MACH_MSP7120_GW         4    /* PMC-Sierra MSP7120 

Residential Gateway board */

include/asm-mips/bootinfo.h line 224

Quoted 1 line
include/asm-mips/bootinfo.h +1 -0
+#define MACH_MSP7120_FPGA       5    /* PMC-Sierra MSP7120 Emulation 

board */

Quoted 8 lines

include/asm-mips/bootinfo.h line 225

Quoted 1 line
include/asm-mips/bootinfo.h +1 -0
+#define MACH_MSP_OTHER        255    /* PMC-Sierra unknown board 

type */

include/asm-mips/bootinfo.h line 226

Quoted 1 line
include/asm-mips/bootinfo.h +1 -0
+#define CL_SIZE                      COMMAND_LINE_SIZE

Really I would add MACH_GROUP_MSP after MACH_GROUP_NEC_EMMA2RH, perhaps 27 or 28, rather than an interior number. Especially if you are going to put it after MACH_GROUP_NEC_EMMA2RH in the file. ~:^)

Sure we aren't tied to this number. Looking at the numbering more closely, it looks like numbers aren't reused when they are dropped so it may be safer in case existing boards are still using 23.

Why do you recommend 27 or higher when apparently 26 hasn't been used?

Marc

#1 lore raw

Marc St-Jean wrote:
[snip]

Quoted 3 lines
Quoted 3 lines

Review

Quoted 3 lines
(quoted diff) +3 -0
+#ifdef CONFIG_PMC_MSP
+     jal     kernel_entry
+#else

Maybe introduce a CONFIG_BOOT_RAW (and use it in arch/mips/Makefile to objcopy the kernel to a raw binary).

Quoted 7 lines
Quoted 7 lines
Quoted 6 lines
(quoted diff) +2 -0
+     /*
       * Reserved space for exception handlers.
       * Necessary for machines which link their kernels at KSEG0.
       */
      .fill   0x400
+#endif /* CONFIG_PMC_MSP */
Quoted 8 lines
Quoted 3 lines

This is getting kind of ugly. There are a whole lot of config choices that need to use the 'j kernel_entry'. Do they all have to have their own? I'm not sure what the best way is to handle them all.

I agree but don't know the best way to handle this. I could introduce a SYS_NO_EXEPT_FILL or similar flag but this seems excessive.

Any other ideas from arch/mips folks?

Something like

#if LOADADDR == 0xffffffff80000000
	.fill	0x400

#endif

but by defining an appropriate name in arch/mips/Makefile instead of externalizing the load-y/LOADADDR there.

Thiemo

#2 lore raw
Quoted 8 lines · Thiemo Seufer

Something like

#if LOADADDR == 0xffffffff80000000
	.fill	0x400

#endif

but by defining an appropriate name in arch/mips/Makefile instead of externalizing the load-y/LOADADDR there.

Basically a good idea but it will fail for 64-bit kernels so the test would need to be extended to cover XKPHYS as well. Also R2 processors which have the c0_ebase registers do no need to reserve space for exception handlers as they can easily move them elsewhere.

Ralf

#3 lore raw
In reply to Marc St-Jean
Quoted 37 lines · Marc St-Jean
Quoted 26 lines · Andrew Sharp

include/asm-mips/bootinfo.h line 218

Quoted 11 lines · On Mon, 26 Feb 2007 18:12:55 -0600 Marc St-Jean <[email protected]> wrote:
include/asm-mips/bootinfo.h +4 -0
index c7c945b..ab29fd4 100644
@@ -213,6 +213,18 @@
 #define MACH_GROUP_NEC_EMMA2RH 25    /* NEC EMMA2RH (was 23)         */
 #define  MACH_NEC_MARKEINS    0       /* NEC EMMA2RH Mark-eins        */
+/*
+ * Valid machtype for group PMC-MSP
+ */
+#define MACH_GROUP_MSP         23    /* PMC-Sierra MSP 

boards/CPUs */

include/asm-mips/bootinfo.h line 219

Quoted 1 line
include/asm-mips/bootinfo.h +1 -0
+#define MACH_MSP4200_EVAL       0    /* PMC-Sierra MSP4200 

Evaluation board */

include/asm-mips/bootinfo.h line 220

Quoted 1 line
include/asm-mips/bootinfo.h +1 -0
+#define MACH_MSP4200_GW         1    /* PMC-Sierra MSP4200 Gateway 

demo board */

include/asm-mips/bootinfo.h line 221

Quoted 1 line
include/asm-mips/bootinfo.h +1 -0
+#define MACH_MSP4200_FPGA       2    /* PMC-Sierra MSP4200 Emulation 

board */

include/asm-mips/bootinfo.h line 222

Quoted 1 line
include/asm-mips/bootinfo.h +1 -0
+#define MACH_MSP7120_EVAL       3    /* PMC-Sierra MSP7120 

Evaluation board *
/

include/asm-mips/bootinfo.h line 223

Quoted 1 line
include/asm-mips/bootinfo.h +1 -0
+#define MACH_MSP7120_GW         4    /* PMC-Sierra MSP7120 

Residential Gateway board */

include/asm-mips/bootinfo.h line 224

Quoted 1 line
include/asm-mips/bootinfo.h +1 -0
+#define MACH_MSP7120_FPGA       5    /* PMC-Sierra MSP7120 Emulation 

board */

Quoted 8 lines

include/asm-mips/bootinfo.h line 225

Quoted 1 line
include/asm-mips/bootinfo.h +1 -0
+#define MACH_MSP_OTHER        255    /* PMC-Sierra unknown board 

type */

include/asm-mips/bootinfo.h line 226

Quoted 1 line
include/asm-mips/bootinfo.h +1 -0
+#define CL_SIZE                      COMMAND_LINE_SIZE

Really I would add MACH_GROUP_MSP after MACH_GROUP_NEC_EMMA2RH, perhaps 27 or 28, rather than an interior number. Especially if you are going to put it after MACH_GROUP_NEC_EMMA2RH in the file. ~:^)

Quoted 5 lines

Sure we aren't tied to this number. Looking at the numbering more closely, it looks like numbers aren't reused when they are dropped so it may be safer in case existing boards are still using 23.

Why do you recommend 27 or higher when apparently 26 hasn't been used?

Experience. ~:^) You never know when a little extra room for expansion might come in handy. For example, let's say next year you release a quad-core SOC (hint-hint), and a line of eval boards. You would have room to put them in the file next to your other boards. But just a suggestion.

Cheers,

a

#4 lore raw
Quoted 5 lines · Andrew Sharp

Experience. ~:^) You never know when a little extra room for expansion might come in handy. For example, let's say next year you release a quad-core SOC (hint-hint), and a line of eval boards. You would have room to put them in the file next to your other boards. But just a suggestion.

Since mips_machtype and mips_machgroup are mostly used to identify systems within Linux changing would not be the most terrible thing on earth. There is no requirement for these numbers to be contiguous either.

Ralf