When creating dom0 with grant table support disabled in Xen and no IOMMU,
the following assert is triggered (debug build):
"Assertion 's <= e' failed at common/rangeset.c:189"
This is because find_unallocated_memory() (used to find memory holes
for extended regions) calls rangeset_remove_range() for an empty
grant table region. Fix it by checking if the size of region is not 0.
Signed-off-by: Michal Orzel <michal.orzel@amd.com>
---
xen/arch/arm/domain_build.c | 18 +++++++++++-------
1 file changed, 11 insertions(+), 7 deletions(-)
diff --git a/xen/arch/arm/domain_build.c b/xen/arch/arm/domain_build.c
index 54bf5623c889..2e899458acdf 100644
--- a/xen/arch/arm/domain_build.c
+++ b/xen/arch/arm/domain_build.c
@@ -1633,14 +1633,18 @@ static int __init find_unallocated_memory(const struct kernel_info *kinfo,
}
/* Remove grant table region */
- start = kinfo->gnttab_start;
- end = kinfo->gnttab_start + kinfo->gnttab_size;
- res = rangeset_remove_range(unalloc_mem, PFN_DOWN(start), PFN_DOWN(end - 1));
- if ( res )
+ if ( kinfo->gnttab_size )
{
- printk(XENLOG_ERR "Failed to remove: %#"PRIpaddr"->%#"PRIpaddr"\n",
- start, end);
- goto out;
+ start = kinfo->gnttab_start;
+ end = kinfo->gnttab_start + kinfo->gnttab_size;
+ res = rangeset_remove_range(unalloc_mem, PFN_DOWN(start),
+ PFN_DOWN(end - 1));
+ if ( res )
+ {
+ printk(XENLOG_ERR "Failed to remove: %#"PRIpaddr"->%#"PRIpaddr"\n",
+ start, end);
+ goto out;
+ }
}
start = 0;
--
2.25.1
Hi, On 24/08/2023 10:06, Michal Orzel wrote: > When creating dom0 with grant table support disabled in Xen and no IOMMU, > the following assert is triggered (debug build): > "Assertion 's <= e' failed at common/rangeset.c:189" A partial stack trace would have been handy. This help the reader to understand how you came to the conclusion that the issue was in find_unallocated_memory(). > > This is because find_unallocated_memory() (used to find memory holes > for extended regions) calls rangeset_remove_range() for an empty > grant table region. Fix it by checking if the size of region is not 0. > > Signed-off-by: Michal Orzel <michal.orzel@amd.com> The change itself looks fine. So with the stack trace add: Reviewed-by: Julien Grall <jgrall@amazon.com> > --- > xen/arch/arm/domain_build.c | 18 +++++++++++------- > 1 file changed, 11 insertions(+), 7 deletions(-) > > diff --git a/xen/arch/arm/domain_build.c b/xen/arch/arm/domain_build.c > index 54bf5623c889..2e899458acdf 100644 > --- a/xen/arch/arm/domain_build.c > +++ b/xen/arch/arm/domain_build.c > @@ -1633,14 +1633,18 @@ static int __init find_unallocated_memory(const struct kernel_info *kinfo, > } > > /* Remove grant table region */ > - start = kinfo->gnttab_start; > - end = kinfo->gnttab_start + kinfo->gnttab_size; > - res = rangeset_remove_range(unalloc_mem, PFN_DOWN(start), PFN_DOWN(end - 1)); > - if ( res ) > + if ( kinfo->gnttab_size ) > { > - printk(XENLOG_ERR "Failed to remove: %#"PRIpaddr"->%#"PRIpaddr"\n", > - start, end); > - goto out; > + start = kinfo->gnttab_start; > + end = kinfo->gnttab_start + kinfo->gnttab_size; > + res = rangeset_remove_range(unalloc_mem, PFN_DOWN(start), > + PFN_DOWN(end - 1)); > + if ( res ) > + { > + printk(XENLOG_ERR "Failed to remove: %#"PRIpaddr"->%#"PRIpaddr"\n", > + start, end); > + goto out; > + } > } > > start = 0; Cheers, -- Julien Grall
Hi Julien, On 24/08/2023 11:10, Julien Grall wrote: > > > Hi, > > On 24/08/2023 10:06, Michal Orzel wrote: >> When creating dom0 with grant table support disabled in Xen and no IOMMU, >> the following assert is triggered (debug build): >> "Assertion 's <= e' failed at common/rangeset.c:189" > > A partial stack trace would have been handy. This help the reader to > understand how you came to the conclusion that the issue was in > find_unallocated_memory(). Here you go: (XEN) Xen call trace: (XEN) [<0000020000218568>] rangeset_remove_range+0xbc/0x2cc (PC) (XEN) [<00000200002c76bc>] domain_build.c#make_hypervisor_node+0x294/0x7c4 (LR) (XEN) [<00000200002ca240>] domain_build.c#handle_node+0x7ec/0x924 (XEN) [<00000200002ca7ac>] domain_build.c#construct_dom0+0x434/0x4d8 Can you append this to the commit msg while committing or do you want a respin? ~Michal
On 24/08/2023 10:17, Michal Orzel wrote: > On 24/08/2023 11:10, Julien Grall wrote: >> On 24/08/2023 10:06, Michal Orzel wrote: >>> When creating dom0 with grant table support disabled in Xen and no IOMMU, >>> the following assert is triggered (debug build): >>> "Assertion 's <= e' failed at common/rangeset.c:189" >> >> A partial stack trace would have been handy. This help the reader to >> understand how you came to the conclusion that the issue was in >> find_unallocated_memory(). > Here you go: > (XEN) Xen call trace: > (XEN) [<0000020000218568>] rangeset_remove_range+0xbc/0x2cc (PC) > (XEN) [<00000200002c76bc>] domain_build.c#make_hypervisor_node+0x294/0x7c4 (LR) > (XEN) [<00000200002ca240>] domain_build.c#handle_node+0x7ec/0x924 > (XEN) [<00000200002ca7ac>] domain_build.c#construct_dom0+0x434/0x4d8 > > Can you append this to the commit msg while committing or do you want a respin? Thanks. It can be done on commit. Cheers, -- Julien Grall
On 24/08/2023 11:04, Julien Grall wrote: > > > On 24/08/2023 10:17, Michal Orzel wrote: >> On 24/08/2023 11:10, Julien Grall wrote: >>> On 24/08/2023 10:06, Michal Orzel wrote: >>>> When creating dom0 with grant table support disabled in Xen and no >>>> IOMMU, >>>> the following assert is triggered (debug build): >>>> "Assertion 's <= e' failed at common/rangeset.c:189" >>> >>> A partial stack trace would have been handy. This help the reader to >>> understand how you came to the conclusion that the issue was in >>> find_unallocated_memory(). >> Here you go: >> (XEN) Xen call trace: >> (XEN) [<0000020000218568>] rangeset_remove_range+0xbc/0x2cc (PC) >> (XEN) [<00000200002c76bc>] >> domain_build.c#make_hypervisor_node+0x294/0x7c4 (LR) >> (XEN) [<00000200002ca240>] domain_build.c#handle_node+0x7ec/0x924 >> (XEN) [<00000200002ca7ac>] domain_build.c#construct_dom0+0x434/0x4d8 >> >> Can you append this to the commit msg while committing or do you want >> a respin? > > Thanks. It can be done on commit. I have now committed the patch. Sorry for the delay. Cheers, -- Julien Grall
© 2016 - 2024 Red Hat, Inc.