Re: [PATCH - RFC] number of Solaris slices

Re: [PATCH - RFC] number of Solaris slices

[Andries.Brouwer]

    From: Mikael Pettersson <mikpe@csd.uu.se>

    >People tell me that SOLARIS_X86_NUMSLICE should be 16 instead of 8.
    >And it seems there is some truth in that.
    >
    >On the other hand, there have certainly been times that 8 was the
    >right number. Instead of using a define for the number of slices
    >(partitions, if you prefer), is it OK for all Solaris versions to
    >use v->v_nparts?

    Your patch didn't break my dual boot Linux + Sol8 x86 box.
    It has about 8 slices in the Solaris partition.

Good - thanks!

I just tested it myself. Burned Solaris 8 install CDs,
[using ide-scsi, which worked fine for me under 2.6.0-test11 :-)],
installed and booted to see

 hda2: <solaris: [s0] hda5 [s1] hda6 [s2] hda7 [s6] hda8 [s8] hda9 [s9] hda10 >

so that this particular install indeed uses 16 and not 8 slices.
Expect to submit this patch when development thaws again.

Andries

Re: [PATCH - RFC] number of Solaris slices

[Mikael Pettersson]

On Sat, 6 Dec 2003 23:17:22 +0100 (MET), Andries.Brouwer@cwi.nl wrote:
>People tell me that SOLARIS_X86_NUMSLICE should be 16 instead of 8.
>And it seems there is some truth in that.
>
>On the other hand, there have certainly been times that 8 was the
>right number. Instead of using a define for the number of slices
>(partitions, if you prefer), is it OK for all Solaris versions to
>use v->v_nparts?

Your patch didn't break my dual boot Linux + Sol8 x86 box.
It has about 8 slices in the Solaris partition.

/Mikael

[PATCH - RFC] number of Solaris slices

[Andries.Brouwer]

People tell me that SOLARIS_X86_NUMSLICE should be 16 instead of 8.
And it seems there is some truth in that.

On the other hand, there have certainly been times that 8 was the
right number. Instead of using a define for the number of slices
(partitions, if you prefer), is it OK for all Solaris versions to
use v->v_nparts?

Andries

--- /linux/2.6/linux-2.6.0-test11/linux/fs/partitions/msdos.c	Wed Nov 26 21:44:30 2003
+++ ./msdos.c	Sat Dec  6 22:55:58 2003
@@ -170,7 +170,7 @@
 #ifdef CONFIG_SOLARIS_X86_PARTITION
 	Sector sect;
 	struct solaris_x86_vtoc *v;
-	int i;
+	int i, numslices;
 
 	v = (struct solaris_x86_vtoc *)read_dev_sector(bdev, offset+1, &sect);
 	if (!v)
@@ -186,7 +186,12 @@
 		put_dev_sector(sect);
 		return;
 	}
-	for (i=0; i<SOLARIS_X86_NUMSLICE && state->next<state->limit; i++) {
+
+	numslices = v->v_nparts;
+	if (numslices > SOLARIS_X86_NUMSLICE)
+		numslices = SOLARIS_X86_NUMSLICE;
+
+	for (i=0; i<numslices && state->next<state->limit; i++) {
 		struct solaris_x86_slice *s = &v->v_slice[i];
 		if (s->s_size == 0)
 			continue;
--- /linux/2.6/linux-2.6.0-test11/linux/include/linux/genhd.h	Wed Nov 26 21:43:08 2003
+++ ./genhd.h	Sat Dec  6 22:53:06 2003
@@ -212,7 +212,7 @@
 
 #ifdef CONFIG_SOLARIS_X86_PARTITION
 
-#define SOLARIS_X86_NUMSLICE	8
+#define SOLARIS_X86_NUMSLICE	16
 #define SOLARIS_X86_VTOC_SANE	(0x600DDEEEUL)
 
 struct solaris_x86_slice {