RE: [XenPPC] Changeset with 3 files on Friday (I think) has brokenXen on Mambo
Hi Mark, you didn't mention this but I understand you are loading your
dom0 in an unusual way. Does this patch solve your problem?
diff -r ed5ee9dde0bd xen/arch/powerpc/boot_of.c
--- a/xen/arch/powerpc/boot_of.cSun Jan 21 08:17:46 2007 -0500
+++ b/xen/arch/powerpc/boot_of.cMon Jan 22 16:14:28 2007 -0600
@@ -488,6 +488,23 @@ static void boot_of_free(ulong addr, ulo
}
#endif
+static void boot_of_reserve(ulong addr, ulong size)
+{
+ulong bits;
+ulong pos;
+ulong i;
+
+size = ALIGN_UP(size, PAGE_SIZE);
+bits = size >> PAGE_SHIFT;
+pos = addr >> PAGE_SHIFT;
+
+for (i = 0; i < bits; i++) {
+if (test_bit(pos + i, mem_available_pages))
+of_panic("%s: pg :0x%lx already reserved.\n", __func__, pos + i);
+set_bit(pos + i, mem_available_pages);
+}
+}
+
static ulong boot_of_alloc(ulong size)
{
ulong bits;
@@ -1121,6 +1138,7 @@ static void * __init boot_of_module(ulon
/* was it handed to us in registers ? */
mod0_start = r3;
mod0_size = r4;
+boot_of_reserve(r3, r4);
of_printf("%s: Dom0 was loaded and found using r3/r4:"
"0x%lx[size 0x%lx]\n",
__func__, mod0_start, mod0_size);
--
Hollis Blanchard
IBM Linux Technology Center
On Mon, 2007-01-22 at 15:31 -0500, Mark F Mergen wrote:
>
> I cannot replicate Stuart's successful report. I'm sure my simulator
> is a much older build than his and Hollis' guess about device trees is
> another reason I should bring my simulator up to date. I apologize
> for raising a problem without doing that first. I might not get to it
> right away, but I will do it. Hollis, why don't you not waste your
> time with simulator until I test with a current one.
>
> Mark
>
___
Xen-ppc-devel mailing list
Xen-ppc-devel@lists.xensource.com
http://lists.xensource.com/xen-ppc-devel
RE: [XenPPC] Changeset with 3 files on Friday (I think) has brokenXen on Mambo
I cannot replicate Stuart's successful report. I'm sure my simulator is a much older build than his and Hollis' guess about device trees is another reason I should bring my simulator up to date. I apologize for raising a problem without doing that first. I might not get to it right away, but I will do it. Hollis, why don't you not waste your time with simulator until I test with a current one. Mark "Yoder Stuart-B08248" <[EMAIL PROTECTED]> 01/22/2007 01:58 PM To [EMAIL PROTECTED], Mark F Mergen/Watson/[EMAIL PROTECTED] cc Subject RE: [XenPPC] Changeset with 3 files on Friday (I think) has brokenXen on Mambo I just did a clean build of top of tree Xen and Linux and am able to boot my dom0 kernel on Xen with the simulator. So I'm not seeing the same issue Mark had. Mark-- I would try the very latest just to make sure. I built with linux changeset ab3b5849331d and Xen changeset ed5ee9dde0bd. It is taking a heck of a long time though. The simulator cycle count appears to have about doubled. Stuart > -Original Message- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] On Behalf > Of Hollis Blanchard > Sent: Monday, January 22, 2007 12:02 PM > To: Mark F Mergen > Cc: xen-ppc-devel@lists.xensource.com > Subject: Re: [XenPPC] Changeset with 3 files on Friday (I > think) has brokenXen on Mambo > > Hi Mark, it was probably changeset 5bcb155e5de5 that caused > this. Could > you open a bug (http://bugzilla.xensource.com/), assign it to me, and > attach the full boot log? (It looks like you trimmed some of the early > messages below.) > > In the meantime I'll try to install systemsim locally, but it > looks like > it doesn't play nicely with current distributions. My first guess is > that systemsim's device tree memory nodes are wrong somehow. > > It looks like you could simply revert the changeset in > question, but to > avoid merge problems in the future, it would be better to fall back to > the tree before that changeset went in (hg update -C $(hg parents -qr > 5bcb155e5de5)). > > Thanks! > > -- > Hollis Blanchard > IBM Linux Technology Center > > > On Sun, 2007-01-21 at 19:24 -0500, Mark F Mergen wrote: > > > > From the standpoint of us Lone Rangers using Mambo (Stuart and me), > > things are broken again. Below is the result of trying to > boot (this > > is an unpatched system, doesn't contain any of my mods) > after applying > > a single changeset with 3 files that I think was pushed on > Friday. I > > pulled it, it built OK, I ran out of time to test then, and I > > discarded the email that described the changeset. I'm > guessing, but I > > think it may have been the same changeset that was sent > again Saturday > > with the subject line "[XenPPC] [xenppc-unstable] [XEN][POWERPC] > > merge" and affected these files: > > xen/arch/powerpc/boot_of.c | 39 +++ > > xen/arch/powerpc/memory.c| 150 > > --- > > xen/include/asm-powerpc/mm.h |3 > > 3 files changed, 98 insertions(+), 94 deletions(-) > > > > The problem appears to be at the lines highlighted in red > below. Any > > help or ideas in fixing this will be much appreciated. > > > > 211611123: (211859452): (XEN) Trying Dom0 as 64bit ELF > > 211648608: (211896989): (XEN) *** LOADING DOMAIN 0 *** > > 22000: [0]: (PC:0x0041B654) : 2224.2 Kilo-Inst/Sec : > > .2 Kilo-Cycles/Sec > > 227352925: (227633054): (XEN) xen_start_info: 07ffe000 > > 227371602: (227651740): (XEN) shared_info: > 0x3fff000,07fff000 > > 227390712: (227670857): (XEN) loading OFH: 0x600, RMA: > 0x200 > > 227506947: (227787248): (XEN) loading OFD: 0x6024000 RMA: 0x2024000, > > 0x3 > > 227590077: (227870488): (XEN) loading 64-bit Dom0: 0x6054000, in > > RMA:0x2054000 > > 230086980: (230371430): (XEN) no initrd > > 230096126: (230380580): (XEN) DOM: pc = 0x2054000, r2 = 0x0 > > 230110154: (230394612): (XEN) Add /openprom props > > 230127650: (230412122): (XEN) Add /options props > > 230141413: (230425891): (XEN) Add /cpus props > > 230171687: (230456207): (XEN) Add /chosen props > > 230185151: (230469678): (XEN) DOM0 bootargs: init=/bin/bash > iommu=off > > lpj=0x26 > > 230211575: (230496117): (XEN) Remove /chosen/mmu, stub will replace > > 230229344: (230513891): (XEN) fix /memory props > > 230275249: (230559867): (XEN) Dom0: /[EMAIL PROTECTED]: , > > 0400 > > 230304736: (230589372): (XEN) Dom0:
RE: [XenPPC] Changeset with 3 files on Friday (I think) has brokenXen on Mambo
I just did a clean build of top of tree Xen and Linux and am able to boot my dom0 kernel on Xen with the simulator. So I'm not seeing the same issue Mark had. Mark-- I would try the very latest just to make sure. I built with linux changeset ab3b5849331d and Xen changeset ed5ee9dde0bd. It is taking a heck of a long time though. The simulator cycle count appears to have about doubled. Stuart > -Original Message- > From: [EMAIL PROTECTED] > [mailto:[EMAIL PROTECTED] On Behalf > Of Hollis Blanchard > Sent: Monday, January 22, 2007 12:02 PM > To: Mark F Mergen > Cc: xen-ppc-devel@lists.xensource.com > Subject: Re: [XenPPC] Changeset with 3 files on Friday (I > think) has brokenXen on Mambo > > Hi Mark, it was probably changeset 5bcb155e5de5 that caused > this. Could > you open a bug (http://bugzilla.xensource.com/), assign it to me, and > attach the full boot log? (It looks like you trimmed some of the early > messages below.) > > In the meantime I'll try to install systemsim locally, but it > looks like > it doesn't play nicely with current distributions. My first guess is > that systemsim's device tree memory nodes are wrong somehow. > > It looks like you could simply revert the changeset in > question, but to > avoid merge problems in the future, it would be better to fall back to > the tree before that changeset went in (hg update -C $(hg parents -qr > 5bcb155e5de5)). > > Thanks! > > -- > Hollis Blanchard > IBM Linux Technology Center > > > On Sun, 2007-01-21 at 19:24 -0500, Mark F Mergen wrote: > > > > From the standpoint of us Lone Rangers using Mambo (Stuart and me), > > things are broken again. Below is the result of trying to > boot (this > > is an unpatched system, doesn't contain any of my mods) > after applying > > a single changeset with 3 files that I think was pushed on > Friday. I > > pulled it, it built OK, I ran out of time to test then, and I > > discarded the email that described the changeset. I'm > guessing, but I > > think it may have been the same changeset that was sent > again Saturday > > with the subject line "[XenPPC] [xenppc-unstable] [XEN][POWERPC] > > merge" and affected these files: > > xen/arch/powerpc/boot_of.c | 39 +++ > > xen/arch/powerpc/memory.c| 150 > > --- > > xen/include/asm-powerpc/mm.h |3 > > 3 files changed, 98 insertions(+), 94 deletions(-) > > > > The problem appears to be at the lines highlighted in red > below. Any > > help or ideas in fixing this will be much appreciated. > > > > 211611123: (211859452): (XEN) Trying Dom0 as 64bit ELF > > 211648608: (211896989): (XEN) *** LOADING DOMAIN 0 *** > > 22000: [0]: (PC:0x0041B654) : 2224.2 Kilo-Inst/Sec : > > .2 Kilo-Cycles/Sec > > 227352925: (227633054): (XEN) xen_start_info: 07ffe000 > > 227371602: (227651740): (XEN) shared_info: > 0x3fff000,07fff000 > > 227390712: (227670857): (XEN) loading OFH: 0x600, RMA: > 0x200 > > 227506947: (227787248): (XEN) loading OFD: 0x6024000 RMA: 0x2024000, > > 0x3 > > 227590077: (227870488): (XEN) loading 64-bit Dom0: 0x6054000, in > > RMA:0x2054000 > > 230086980: (230371430): (XEN) no initrd > > 230096126: (230380580): (XEN) DOM: pc = 0x2054000, r2 = 0x0 > > 230110154: (230394612): (XEN) Add /openprom props > > 230127650: (230412122): (XEN) Add /options props > > 230141413: (230425891): (XEN) Add /cpus props > > 230171687: (230456207): (XEN) Add /chosen props > > 230185151: (230469678): (XEN) DOM0 bootargs: init=/bin/bash > iommu=off > > lpj=0x26 > > 230211575: (230496117): (XEN) Remove /chosen/mmu, stub will replace > > 230229344: (230513891): (XEN) fix /memory props > > 230275249: (230559867): (XEN) Dom0: /[EMAIL PROTECTED]: , > > 0400 > > 230304736: (230589372): (XEN) Dom0: /[EMAIL PROTECTED]: > 0400, > > 0800 > > 230327662: (230612304): (XEN) fix /xen props > > 230345978: (230630636): (XEN) Remove original /dart > > 230365995: (230650671): (XEN) Remove original /rtas > > 230413320: (230698061): (XEN) Assertion > 'allocated_in_map(first_page + > > i)' failed, line 147, file page_alloc.c > > 230445056: (230729810): (XEN) BUG at page_alloc.c:147 > > 230458933: (230743693): (XEN) [ Xen-3.0-unstable ] > > 230475250: (230760017): (XEN) CPU: DOMID: 7fff > > 230494869: (230779651): (XEN) pc msr >
