[parisc-linux] init and exit text problem

[parisc-linux] init and exit text problem

[Matthew Wilcox]

Our vmlinux.lds.S doesn't discard init and exit text segments.  Anyone
want to look into fixing this?

-- 
"Next the statesmen will invent cheap lies, putting the blame upon 
the nation that is attacked, and every man will be glad of those
conscience-soothing falsities, and will diligently study them, and refuse
to examine any refutations of them; and thus he will by and by convince 
himself that the war is just, and will thank God for the better sleep 
he enjoys after this process of grotesque self-deception." -- Mark Twain

RE: [parisc-linux] init and exit text problem

[Joel Soete]

>
>Our vmlinux.lds.S doesn't discard init and exit text segments.  Anyone
>want to look into fixing this?
>
is it what you look for:

=========><=========
--- linux-2.6.1-pa3.orig/arch/parisc/kernel/vmlinux.lds.S	2003-09-22 09:06:55.000000000
+0200
+++ linux-2.6.1-pa3/arch/parisc/kernel/vmlinux.lds.S	2004-01-16 18:25:15.175817000
+0100
@@ -102,6 +102,17 @@
 
   _end = . ;
 
+  /* When something in the kernel is NOT compiled as a module, the
+   * module cleanup code and data are put into these segments.  Both
+   * can then be thrown away, as cleanup code is never called unless
+   * it's a module.
+   */
+  /DISCARD/ : {
+	*(.exit.text)
+	*(.exit.data)
+	*(.exitcall.exit)
+	}
+
   /* Stabs debugging sections.  */
   .stab 0 : { *(.stab) }
   .stabstr 0 : { *(.stabstr) }
=========><=========

Unfortunately failled to build kernel because:
[snip]
  	ld  -T arch/parisc/kernel/vmlinux.lds.s arch/parisc/kernel/head.o    init/built-in.o
--start-group  usr/built-in.o  arch/parisc/mm/built-in.o  arch/parisc/kernel/built-in.o
 arch/parisc/math-emu/built-in.o arch/parisc/kernel/init_task.o  kernel/built-in.o
 mm/built-in.o  fs/built-in.o  ipc/built-in.o  security/built-in.o  crypto/built-in.o
 lib/lib.a  arch/parisc/lib/lib.a `gcc -print-libgcc-file-name`  lib/built-in.o
 arch/parisc/lib/built-in.o `gcc -print-libgcc-file-name`  drivers/built-in.o
 sound/built-in.o  arch/parisc/oprofile/built-in.o  net/built-in.o --end-group
 -o .tmp_vmlinux1
local symbol 0: discarded in section `.exit.text' from kernel/built-in.o
local symbol 1: discarded in section `.exit.text' from kernel/built-in.o
local symbol 2: discarded in section `.exit.text' from fs/built-in.o
local symbol 3: discarded in section `.exit.text' from fs/built-in.o
local symbol 4: discarded in section `.exit.text' from fs/built-in.o
local symbol 5: discarded in section `.exit.text' from fs/built-in.o
local symbol 6: discarded in section `.exit.text' from fs/built-in.o
local symbol 7: discarded in section `.exit.text' from fs/built-in.o
local symbol 8: discarded in section `.exit.text' from fs/built-in.o
local symbol 9: discarded in section `.exit.text' from fs/built-in.o
local symbol 10: discarded in section `.exit.text' from fs/built-in.o
local symbol 11: discarded in section `.exit.text' from fs/built-in.o
local symbol 12: discarded in section `.exit.text' from fs/built-in.o
local symbol 13: discarded in section `.exit.text' from fs/built-in.o
local symbol 14: discarded in section `.exit.text' from fs/built-in.o
local symbol 15: discarded in section `.exit.text' from fs/built-in.o
local symbol 16: discarded in section `.exit.text' from fs/built-in.o
local symbol 17: discarded in section `.exit.text' from fs/built-in.o
local symbol 18: discarded in section `.exit.text' from fs/built-in.o
local symbol 19: discarded in section `.exit.text' from fs/built-in.o
local symbol 20: discarded in section `.exit.text' from fs/built-in.o
local symbol 21: discarded in section `.exit.text' from fs/built-in.o
local symbol 22: discarded in section `.exit.text' from fs/built-in.o
local symbol 23: discarded in section `.exit.text' from fs/built-in.o
local symbol 24: discarded in section `.exit.text' from fs/built-in.o
local symbol 25: discarded in section `.exit.text' from fs/built-in.o
local symbol 26: discarded in section `.exit.text' from fs/built-in.o
local symbol 27: discarded in section `.exit.text' from fs/built-in.o
local symbol 28: discarded in section `.exit.text' from drivers/built-in.o
local symbol 29: discarded in section `.exit.text' from drivers/built-in.o
local symbol 30: discarded in section `.exit.text' from drivers/built-in.o
local symbol 31: discarded in section `.exit.text' from drivers/built-in.o
local symbol 32: discarded in section `.exit.text' from drivers/built-in.o
local symbol 33: discarded in section `.exit.text' from drivers/built-in.o
local symbol 34: discarded in section `.exit.text' from drivers/built-in.o
local symbol 35: discarded in section `.exit.text' from drivers/built-in.o
local symbol 36: discarded in section `.exit.text' from drivers/built-in.o
local symbol 37: discarded in section `.exit.text' from drivers/built-in.o
local symbol 38: discarded in section `.exit.text' from drivers/built-in.o
local symbol 39: discarded in section `.exit.text' from drivers/built-in.o
local symbol 40: discarded in section `.exit.text' from drivers/built-in.o
local symbol 41: discarded in section `.exit.text' from drivers/built-in.o
local symbol 42: discarded in section `.exit.text' from drivers/built-in.o
local symbol 43: discarded in section `.exit.text' from drivers/built-in.o
local symbol 44: discarded in section `.exit.text' from drivers/built-in.o
local symbol 45: discarded in section `.exit.text' from drivers/built-in.o
local symbol 46: discarded in section `.exit.text' from drivers/built-in.o
local symbol 47: discarded in section `.exit.text' from drivers/built-in.o
local symbol 48: discarded in section `.exit.text' from drivers/built-in.o
local symbol 49: discarded in section `.exit.text' from drivers/built-in.o
local symbol 50: discarded in section `.exit.text' from drivers/built-in.o
local symbol 51: discarded in section `.exit.text' from drivers/built-in.o
local symbol 52: discarded in section `.exit.text' from drivers/built-in.o
local symbol 53: discarded in section `.exit.text' from drivers/built-in.o
local symbol 54: discarded in section `.exit.text' from drivers/built-in.o
local symbol 55: discarded in section `.exit.text' from drivers/built-in.o
local symbol 56: discarded in section `.exit.text' from drivers/built-in.o
local symbol 57: discarded in section `.exit.text' from drivers/built-in.o
local symbol 58: discarded in section `.exit.text' from drivers/built-in.o
local symbol 59: discarded in section `.exit.text' from drivers/built-in.o
local symbol 60: discarded in section `.exit.text' from drivers/built-in.o
local symbol 61: discarded in section `.exit.text' from drivers/built-in.o
local symbol 62: discarded in section `.exit.text' from drivers/built-in.o
local symbol 63: discarded in section `.exit.text' from drivers/built-in.o
local symbol 64: discarded in section `.exit.text' from drivers/built-in.o
local symbol 65: discarded in section `.exit.text' from drivers/built-in.o
local symbol 66: discarded in section `.exit.text' from drivers/built-in.o
local symbol 67: discarded in section `.exit.text' from drivers/built-in.o
local symbol 68: discarded in section `.exit.text' from drivers/built-in.o
local symbol 69: discarded in section `.exit.text' from drivers/built-in.o
local symbol 70: discarded in section `.exit.text' from drivers/built-in.o
local symbol 71: discarded in section `.exit.text' from drivers/built-in.o
local symbol 72: discarded in section `.exit.text' from drivers/built-in.o
local symbol 73: discarded in section `.exit.text' from drivers/built-in.o
local symbol 74: discarded in section `.exit.text' from drivers/built-in.o
local symbol 75: discarded in section `.exit.text' from drivers/built-in.o
local symbol 76: discarded in section `.exit.text' from drivers/built-in.o
local symbol 77: discarded in section `.exit.text' from drivers/built-in.o
local symbol 78: discarded in section `.exit.text' from drivers/built-in.o
local symbol 79: discarded in section `.exit.text' from drivers/built-in.o
local symbol 80: discarded in section `.exit.text' from drivers/built-in.o
local symbol 81: discarded in section `.exit.text' from drivers/built-in.o
local symbol 82: discarded in section `.exit.text' from drivers/built-in.o
local symbol 83: discarded in section `.exit.text' from drivers/built-in.o
local symbol 84: discarded in section `.exit.text' from drivers/built-in.o
local symbol 85: discarded in section `.exit.text' from drivers/built-in.o
local symbol 86: discarded in section `.exit.text' from drivers/built-in.o
local symbol 87: discarded in section `.exit.text' from drivers/built-in.o
fb_console_exit: discarded in section `.exit.text' from drivers/built-in.o
local symbol 88: discarded in section `.exit.text' from drivers/built-in.o
local symbol 89: discarded in section `.exit.text' from drivers/built-in.o
local symbol 90: discarded in section `.exit.text' from drivers/built-in.o
local symbol 91: discarded in section `.exit.text' from drivers/built-in.o
local symbol 92: discarded in section `.exit.text' from drivers/built-in.o
local symbol 93: discarded in section `.exit.text' from drivers/built-in.o
local symbol 94: discarded in section `.exit.text' from drivers/built-in.o
local symbol 95: discarded in section `.exit.text' from drivers/built-in.o
local symbol 96: discarded in section `.exit.text' from drivers/built-in.o
local symbol 97: discarded in section `.exit.text' from drivers/built-in.o
local symbol 98: discarded in section `.exit.text' from drivers/built-in.o
local symbol 99: discarded in section `.exit.text' from drivers/built-in.o
local symbol 100: discarded in section `.exit.text' from drivers/built-in.o
local symbol 101: discarded in section `.exit.text' from drivers/built-in.o
local symbol 102: discarded in section `.exit.text' from drivers/built-in.o
local symbol 103: discarded in section `.exit.text' from drivers/built-in.o
local symbol 104: discarded in section `.exit.text' from drivers/built-in.o
local symbol 105: discarded in section `.exit.text' from drivers/built-in.o
local symbol 106: discarded in section `.exit.text' from drivers/built-in.o
local symbol 107: discarded in section `.exit.text' from drivers/built-in.o
local symbol 108: discarded in section `.exit.text' from drivers/built-in.o
local symbol 109: discarded in section `.exit.text' from drivers/built-in.o
local symbol 110: discarded in section `.exit.text' from drivers/built-in.o
local symbol 111: discarded in section `.exit.text' from drivers/built-in.o
local symbol 112: discarded in section `.exit.text' from drivers/built-in.o
local symbol 113: discarded in section `.exit.text' from drivers/built-in.o
local symbol 114: discarded in section `.exit.text' from drivers/built-in.o
local symbol 115: discarded in section `.exit.text' from drivers/built-in.o
local symbol 116: discarded in section `.exit.text' from drivers/built-in.o
local symbol 117: discarded in section `.exit.text' from drivers/built-in.o
local symbol 118: discarded in section `.exit.text' from drivers/built-in.o
local symbol 119: discarded in section `.exit.text' from drivers/built-in.o
local symbol 120: discarded in section `.exit.text' from drivers/built-in.o
local symbol 121: discarded in section `.exit.text' from drivers/built-in.o
local symbol 122: discarded in section `.exit.text' from drivers/built-in.o
local symbol 123: discarded in section `.exit.text' from drivers/built-in.o
local symbol 124: discarded in section `.exit.text' from drivers/built-in.o
local symbol 125: discarded in section `.exit.text' from drivers/built-in.o
local symbol 126: discarded in section `.exit.text' from drivers/built-in.o
local symbol 127: discarded in section `.exit.text' from drivers/built-in.o
local symbol 128: discarded in section `.exit.text' from drivers/built-in.o
local symbol 129: discarded in section `.exit.text' from drivers/built-in.o
local symbol 130: discarded in section `.exit.text' from drivers/built-in.o
local symbol 131: discarded in section `.exit.text' from sound/built-in.o
local symbol 132: discarded in section `.exit.text' from sound/built-in.o
local symbol 133: discarded in section `.exit.text' from arch/parisc/oprofile/built-in.o
local symbol 134: discarded in section `.exit.text' from arch/parisc/oprofile/built-in.o
oprofilefs_unregister: discarded in section `.exit.text' from arch/parisc/oprofile/built-in.o
local symbol 135: discarded in section `.exit.text' from arch/parisc/oprofile/built-in.o
local symbol 136: discarded in section `.exit.text' from net/built-in.o
local symbol 137: discarded in section `.exit.text' from net/built-in.o
local symbol 138: discarded in section `.exit.text' from net/built-in.o
local symbol 139: discarded in section `.exit.text' from net/built-in.o
local symbol 140: discarded in section `.exit.text' from net/built-in.o
local symbol 141: discarded in section `.exit.text' from net/built-in.o
local symbol 142: discarded in section `.exit.text' from net/built-in.o
local symbol 143: discarded in section `.exit.text' from net/built-in.o
local symbol 144: discarded in section `.exit.text' from net/built-in.o
local symbol 145: discarded in section `.exit.text' from net/built-in.o
make: *** [.tmp_vmlinux1] Error 1

(btw removing: "+	*(.exit.text)" in the above patch compile and boot (at
least with defconfig which afaik didn't already include module support).

Any idea on what I missed?

hth,
    Joel


-------------------------------------------------------------------------
Tiscali ADSL: 12 mois à 29,50 €/mois! L'Internet rapide, c'est pour tout
le monde.
http://reg.tiscali.be/default.asp?lg=fr

RE: [parisc-linux] init and exit text problem

[Joel Soete]

btw, I try to print a hard copy of ld.info (the only place where I found
some interesting info about /DISCARD/) with texi2dvi but this scipt failed.
Is somebody know a tips to help me?

Thanks in advance,
    Joel


-------------------------------------------------------------------------
Tiscali ADSL: 12 mois à 29,50 €/mois! L'Internet rapide, c'est pour tout
le monde.
http://reg.tiscali.be/default.asp?lg=fr

RE: [parisc-linux] init and exit text problem

[James Bottomley]

On Fri, 2004-01-16 at 11:50, Joel Soete wrote:
> +  /* When something in the kernel is NOT compiled as a module, the
> +   * module cleanup code and data are put into these segments.  Both
> +   * can then be thrown away, as cleanup code is never called unless
> +   * it's a module.
> +   */
> +  /DISCARD/ : {
> +	*(.exit.text)
> +	*(.exit.data)

This is wrong...exit.text and exit.data are runtime discards, not link
time ones, precisely because of the compile failures you get.

> +	*(.exitcall.exit)

This one should be OK.

I'm not sure we can actually do this, though...I seem to remember that
our 64 bit toolchain is not too happy doing discard sections.

James

Re: [parisc-linux] init and exit text problem

[Randolph Chung]

> I'm not sure we can actually do this, though...I seem to remember that
> our 64 bit toolchain is not too happy doing discard sections.

Bugs not withstanding, it does seem to work. We have this in the linker
script already for a while:

#ifdef CONFIG_PARISC64
  /* temporary hack until binutils is fixed to not emit these
     for static binaries */
  /DISCARD/ : {
    *(.dynsym)
    *(.dynstr)
    *(.dynamic)
    *(.hash)
  }
#endif

randolph
-- 
Randolph Chung
Debian GNU/Linux Developer, hppa/ia64 ports
http://www.tausq.org/

Re: [parisc-linux] init and exit text problem

[James Bottomley]

On Fri, 2004-01-16 at 15:14, Randolph Chung wrote:
> > I'm not sure we can actually do this, though...I seem to remember that
> > our 64 bit toolchain is not too happy doing discard sections.
> 
> Bugs not withstanding, it does seem to work. We have this in the linker
> script already for a while:

OK.

There look to be quite a few other differences between our linker script
and x86's.  I'll go over them and see if we need to change.

James

Re: [parisc-linux] init and exit text problem

[James Bottomley]

The attached should be a first cut at this.

I shifted around or vmlinux.lds.S with a view to minimising the diffs
between ours and the one in i386 (so we can spot changes more easily).

I also:

- Implemented cacheline_aligned sections (this may save a few bytes, and
it mirrors i386)
- Corrected a misnaming of the init_task section
- updated our bss clearing algorithm
- removed the duplicate ".initcall.init" section from the head.S files
(it was confusing the linker script and we end up with it being called
.initcall.init.1)
- done the correct thing with regard to discard and init/exit sections
(although we still don't do runtime discards).
- added a few other sections that i386 has (but we don't) just in case
they'll be useful one day (also helps reduce the diff between us and
i386)

James

===== arch/parisc/kernel/head.S 1.4 vs edited =====
--- 1.4/arch/parisc/kernel/head.S	Wed Dec 17 23:48:38 2003
+++ edited/arch/parisc/kernel/head.S	Fri Jan 16 16:19:18 2004
@@ -22,17 +22,6 @@
 
 	.level 1.1
 
-	.section	.initcall.init
-	.align		4
-	.export __initcall_start
-__initcall_start:
-	.export __initcall_end
-__initcall_end:
-	.export __setup_start
-__setup_start:
-	.export __setup_end
-__setup_end:
-
 	.data
 
 	.export boot_args
@@ -64,13 +53,13 @@
 
 	/* Clear BSS (shouldn't the boot loader do this?) */
 
-	.import _edata,data
-	.import _end,data
+	.import __bss_start,data
+	.import __bss_stop,data
 
-	ldil            L%PA(_edata),%r3
-	ldo             R%PA(_edata)(%r3),%r3
-	ldil            L%PA(_end),%r4
-	ldo             R%PA(_end)(%r4),%r4
+	ldil            L%PA(__bss_start),%r3
+	ldo             R%PA(__bss_start)(%r3),%r3
+	ldil            L%PA(__bss_stop),%r4
+	ldo             R%PA(__bss_stop)(%r4),%r4
 $bss_loop:
 	cmpb,<<,n       %r3,%r4,$bss_loop
 	stw,ma          %r0,4(%r3)
===== arch/parisc/kernel/head64.S 1.3 vs edited =====
--- 1.3/arch/parisc/kernel/head64.S	Wed Dec 17 23:48:38 2003
+++ edited/arch/parisc/kernel/head64.S	Fri Jan 16 16:17:15 2004
@@ -26,17 +26,6 @@
 
 	.level 2.0w
 
-	.section	.initcall.init
-	.align		4
-	.export __initcall_start
-__initcall_start:
-	.export __initcall_end
-__initcall_end:
-	.export __setup_start
-__setup_start:
-	.export __setup_end
-__setup_end:
-
 	.data
 
 	.export boot_args
@@ -64,13 +53,13 @@
 
 	/* Clear BSS (shouldn't the boot loader do this?) */
 
-	.import _edata,data
-	.import _end,data
+	.import __bss_start,data
+	.import __bss_stop,data
 
-	ldil            L%PA(_edata),%r3
-	ldo             R%PA(_edata)(%r3),%r3
-	ldil            L%PA(_end),%r4
-	ldo             R%PA(_end)(%r4),%r4
+	ldil            L%PA(__bss_start),%r3
+	ldo             R%PA(__bss_start)(%r3),%r3
+	ldil            L%PA(__bss_stop),%r4
+	ldo             R%PA(__bss_stop)(%r4),%r4
 $bss_loop:
 	cmpb,<<,n       %r3,%r4,$bss_loop
 	stb,ma          %r0,1(%r3)
===== arch/parisc/kernel/vmlinux.lds.S 1.15 vs edited =====
--- 1.15/arch/parisc/kernel/vmlinux.lds.S	Mon Sep  8 17:00:21 2003
+++ edited/arch/parisc/kernel/vmlinux.lds.S	Fri Jan 16 16:34:28 2004
@@ -1,5 +1,7 @@
 #include <linux/config.h>
 #include <asm-generic/vmlinux.lds.h>
+/* needed for the processor specific cache alignment size */	
+#include <asm/cache.h>
 	
 /* ld script to make hppa Linux kernel */
 #ifndef CONFIG_PARISC64
@@ -22,7 +24,7 @@
   . = 0x10100000;
 
   _text = .;			/* Text and read-only data */
-  .text BLOCK(16) : {
+  .text ALIGN(16) : {
 	*(.text*)
 	*(.PARISC.unwind)
 	*(.fixup)
@@ -39,11 +41,34 @@
 
   RODATA
 
-  .data BLOCK(8192) : {			/* Data without special */
+  /* writeable */
+  .data : {			/* Data */
 	data_start = .;
 	*(.data)
+	CONSTRUCTORS
 	}
 
+  . = ALIGN(4096);
+  /* nosave data is really only used for software suspend...it's here
+   * just in case we ever implement it */
+  __nosave_begin = .;
+  .data_nosave : { *(.data.nosave) }
+  . = ALIGN(4096);
+  __nosave_end = .;
+
+  . = ALIGN(L1_CACHE_BYTES);
+  .data.cacheline_aligned : { *(.data.cacheline_aligned) }
+
+  _edata = .;			/* End of data section */
+
+  . = ALIGN(16384); 		/* init_task */
+  .data.init_task : { *(.data.init_task) }
+
+  /* The interrupt stack is currently partially coded, but not yet
+   * implemented */
+  . = ALIGN(16384);	
+  init_istack : { *(init_istack) }
+
 #ifdef CONFIG_PARISC64
   . = ALIGN(16);               /* Linkage tables */
   .opd : { *(.opd) } PROVIDE (__gp = .); 
@@ -63,7 +88,7 @@
   __setup_start = .;
   .init.setup : { *(.init.setup) }
   __setup_end = .;
-  __start___param =.; 
+  __start___param = .;
   __param : { *(__param) }
   __stop___param = .;
   __initcall_start = .;
@@ -81,6 +106,19 @@
   .con_initcall.init : { *(.con_initcall.init) }
   __con_initcall_end = .;
   SECURITY_INIT
+  /* alternate instruction replacement.  This is a mechanism x86 uses
+   * to detect the CPU type and replace generic instruction sequences
+   * with CPU specific ones.  We don't currently do this in PA, but
+   * it seems like a good idea... */
+  . = ALIGN(4);
+  __alt_instructions = .;
+  .altinstructions : { *(.altinstructions) } 
+  __alt_instructions_end = .; 
+ .altinstr_replacement : { *(.altinstr_replacement) } 
+  /* .exit.text is discard at runtime, not link time, to deal with references
+     from .altinstructions and .eh_frame */
+  .exit.text : { *(.exit.text) }
+  .exit.data : { *(.exit.data) }
   . = ALIGN(4096);
   __initramfs_start = .;
   .init.ramfs : { *(.init.ramfs) }
@@ -91,17 +129,27 @@
   __per_cpu_end = .;
   . = ALIGN(4096);
   __init_end = .;
-
-  init_task BLOCK(16384) : { *(init_task) }  /* The initial task and kernel stack */
-
-  _edata = .;			/* End of data section */
-
-
-  .bss : { *(.bss) *(COMMON) }		/* BSS */
-
+  /* freed after init ends here */
+	
+  __bss_start = .;		/* BSS */
+  .bss : { *(.bss) *(COMMON) }
+  __bss_stop = .; 
 
   _end = . ;
 
+  /* Sections to be discarded */
+  /DISCARD/ : {
+	*(.exitcall.exit)
+#ifdef CONFIG_PARISC64
+	/* temporary hack until binutils is fixed to not emit these
+	 for static binaries */
+	*(.dynsym)
+	*(.dynstr)
+	*(.dynamic)
+	*(.hash)
+#endif
+	}
+
   /* Stabs debugging sections.  */
   .stab 0 : { *(.stab) }
   .stabstr 0 : { *(.stabstr) }
@@ -112,14 +160,4 @@
   .comment 0 : { *(.comment) }
   .note 0 : { *(.note) }	
 
-#ifdef CONFIG_PARISC64
-  /* temporary hack until binutils is fixed to not emit these
-     for static binaries */
-  /DISCARD/ : {
-    *(.dynsym)
-    *(.dynstr)
-    *(.dynamic)
-    *(.hash)
-  }
-#endif
 }
===== include/asm-parisc/cache.h 1.4 vs edited =====
--- 1.4/include/asm-parisc/cache.h	Sun Jan  5 05:22:53 2003
+++ edited/include/asm-parisc/cache.h	Fri Jan 16 13:13:23 2004
@@ -7,7 +7,6 @@
 
 #include <linux/config.h>
 
-#ifndef __ASSEMBLY__
 /*
  * PA 2.0 processors have 64-byte cachelines; PA 1.1 processors have
  * 32-byte cachelines.  The default configuration is not for SMP anyway,
@@ -24,12 +23,12 @@
 #define L1_CACHE_SHIFT 5
 #endif
 
+#ifndef __ASSEMBLY__
+
 #define L1_CACHE_ALIGN(x)       (((x)+(L1_CACHE_BYTES-1))&~(L1_CACHE_BYTES-1))
 
 #define SMP_CACHE_BYTES L1_CACHE_BYTES
 #define L1_CACHE_SHIFT_MAX 5	/* largest L1 which this arch supports */
-
-#define __cacheline_aligned __attribute__((__aligned__(L1_CACHE_BYTES)))
 
 extern void flush_data_cache_local(void);  /* flushes local data-cache only */
 extern void flush_instruction_cache_local(void); /* flushes local code-cache only */

Re: [parisc-linux] init and exit text problem

[Grant Grundler]

On Fri, Jan 16, 2004 at 07:12:41PM -0500, James Bottomley wrote:
> The attached should be a first cut at this.
> 
> I shifted around or vmlinux.lds.S with a view to minimising the diffs
> between ours and the one in i386 (so we can spot changes more easily).

This boots and works on the A500 (ios).

2.6.1 kernel with this patch and "bridge_ctl" PCI patch from
Russell King are uploaded to ftp.p-l.o:kernels/a500/2.6.1-pa4.tgz
(No, it's not a -pa4 kernel but that's what it happens to use)

thanks!
grant

Re: [parisc-linux] init and exit text problem

[Joel Soete]

Grant,

A great news: your kernel (2.6.1-pa4) boot well on the N4000-55 :)
palx4000:/var/logs# dmesg | more
Linux version 2.6.1-pa4 (grundler@ios) (gcc version 3.0.4) #10 Sat Jan 17
00:04:07 PST 2004
FP[0] enabled: Rev 1 Model 16
The 64-bit Kernel has started...
Determining PDC firmware type: 64 bit PAT.
model 00005d30 00000491 00000000 00000002 27988e5d 100000f1 00000008 000000b2
000000b2
vers  00000301
CPUID vers 18 rev 11 (0x0000024b)
capabilities 0x1
model 9000/800/N4000-55
Total Memory: 2048 Mb
On node 0 totalpages: 524288
  DMA zone: 524288 pages, LIFO batch:16
  Normal zone: 0 pages, LIFO batch:1
  HighMem zone: 0 pages, LIFO batch:1
Building zonelist for node : 0
Kernel command line: root=/dev/sda5 HOME=/ console=ttyS0 TERM=vt102 palo_kernel=3/vmlinux-2.6.1-pa4
PID hash table entries: 16 (order 4: 256 bytes)
[snip]

:))

>2.6.1 kernel with this patch and "bridge_ctl" PCI patch from
>Russell King are uploaded to ftp.p-l.o:kernels/a500/2.6.1-pa4.tgz

If i well find back your mail and references to this "bridge_ctl" patch otc
I don't reach to apply Russel's patch like this (against 2.6.1) to test on
my B2k (which you know don't work yet with pdc_cons). Can you make your "bridge_ctl"
patch available on ftp.p-l.o?

Thanks again for all,
    Joel

PS: an additional question regarding pdc_cons: in 2.4 there was a device
driver in devices/char/pdc_console.[ch] which allowing to use ttyB device.
But in 2.6 there isn't anymore such driver? any idea?

-------------------------------------------------------------------------
Tiscali ADSL: 12 mois à 29,50 €/mois! L'Internet rapide, c'est pour tout
le monde.
http://reg.tiscali.be/default.asp?lg=fr

Re: [parisc-linux] init and exit text problem

[Grant Grundler]

On Mon, Jan 19, 2004 at 12:16:40PM +0100, Joel Soete wrote:
> Grant,
> 
> A great news: your kernel (2.6.1-pa4) boot well on the N4000-55 :)

awesome! thanks for trying it.

> If i well find back your mail and references to this "bridge_ctl" patch otc
> I don't reach to apply Russel's patch like this (against 2.6.1) to test on
> my B2k (which you know don't work yet with pdc_cons). Can you make your "bridge_ctl"
> patch available on ftp.p-l.o?

yes - certainly ftp://ftp.parisc-linux.org/patches/diff-2.6.1-bridge_ctl

> PS: an additional question regarding pdc_cons: in 2.4 there was a device
> driver in devices/char/pdc_console.[ch] which allowing to use ttyB device.
> But in 2.6 there isn't anymore such driver? any idea?

sorry - I don't know.

grant

Re: [parisc-linux] init and exit text problem

[Joel Soete]

Hi Grant,

> yes - certainly ftp://ftp.parisc-linux.org/patches/diff-2.6.1-bridge_ctl

Ok I applied it against 2.6.1-pa8 (+ some already accepted patch of Matthew
and James related to sym2 driver ;) ).

Compiled with unstable debian gcc-3.3.3 and defconfig, it compiles and boots
well on the b2k ;) (and I am on going to recompile toolchain with it and
all seems to works fine ;) ).

Thanks a lot,
    Joel

ps: I also drive severall test:

Unfortunately, it doesn't help to boot with 64bit kernel on this system
(neither build with my gcc-hppa64_3.3.2 nor with unofficial-deb gcc-hppa64_3.0.4ds3-7.1_hppa.deb
which should the one to use iirc)

And with your config file it always do an hpmc with the 3 compilers I used
(debian gcc-3.3.3, gcc-hppa64 (from gcc-hppa64_3.0.4d
3-7.1_hppa.deb), my own gcc-hppa64_3.3.2), so I would have to find what breaks
for b2k even in 32bit kernel (that will take a long time).

The bad thing is that I don't reach to get any relevant iaoq address from
pim to start investigation :( (just know that it failed very early after
'-- search the FAQ and mailing list...')

hmm just noticed the following diff:
Command line for kernel: 'root=/dev/sda5 HOME=/ console=ttyS0 TERM=vt102
palo_kernel=3/vmlinux-2.6.1-pa7'
Selected kernel: /vmlinux-2.6.1-pa7 from partition 3
ELF32 executable
Entry 001002f0 first 00100000 n 3
Segment 0 load 00100000 size 2586936 mediaptr 0x1000
Segment 1 load 00378000 size 368640 mediaptr 0x279000
Segment 2 load 003d4000 size 553093 mediaptr 0x2d3000
Branching to kernel entry point 0x001002f0.  If this is the last
message you see, you may need to switch your console.  This is
a common symptom -- search the FAQ and mailing list at parisc-linux.org

[snip] normal boot


Command line for kernel: 'roo
=/dev/sda5 HOME=/ console=ttyS0 TERM=vt102
palo_kernel=3/vmlinux-26-up64-gcc30'
Selected kernel: /vmlinux-26-up64-gcc30 from partition 3
ELF64 executable
Entry 00100000 first 00100000 n 2
Segment 0 load 00100000 size 3529136 mediaptr 0x1000
Segment 1 load 00460000 size 237702 mediaptr 0x35f000
Branching to kernel entry point 0x00100000.  If this is the last
message you see, you may need to switch your console.  This is
a common symptom -- search the FAQ and mailing list at parisc-linux.org


HPMC

So 3 segment loaded in 32bit and only 2 in 64bit, is it normal (i lost info
on the N4k boot :( ).
And also seems to "Branching to" a wrong entry point?



-------------------------------------------------------------------------
Tiscali ADSL: 12 mois à 29,50 €/mois! L'Internet rapide, c'est pour tout
le monde.
http://reg.tiscali.be/default.asp?lg=fr