* [PATCH v5] xen/arm: Allow balooning working with 1:1 memory mapping
@ 2013-12-16 17:00 Julien Grall
2013-12-17 8:57 ` Jan Beulich
0 siblings, 1 reply; 3+ messages in thread
From: Julien Grall @ 2013-12-16 17:00 UTC (permalink / raw)
To: xen-devel
Cc: Keir Fraser, ian.campbell, patches, Julien Grall, tim,
stefano.stabellini, Jan Beulich
With the lack of iommu, dom0 must have a 1:1 memory mapping for all
these guest physical address. When the ballon decides to give back a
page to the kernel, this page must have the same address as previously.
Otherwise, we will loose the 1:1 mapping and will break DMA-capable
devices.
Signed-off-by: Julien Grall <julien.grall@linaro.org>
Cc: Keir Fraser <keir@xen.org>
Cc: Jan Beulich <jbeulich@suse.com>
---
Release: This is a bug that prevents DMA-capable devices to work after
a guest has started.
Changes in v5:
- Export dom0_11_mapping and transform is_domain_direct_mapped in macro
- Add ifndef in common code to define is_domain_direct_mapped
Changes in v4:
- Fix typoes
- Update comment in the code
Changes in v3:
- Remove spurious page = NULL
- Rename is_dom0_mapped_11 to is_domain_direct_mapped
- Coding style
Changes in v2:
- Drop CONFIG_ARM and add is_dom0_mapped_11
---
xen/arch/arm/domain_build.c | 2 +-
xen/common/memory.c | 35 ++++++++++++++++++++++++++++++++++-
xen/include/asm-arm/domain.h | 3 +++
3 files changed, 38 insertions(+), 2 deletions(-)
diff --git a/xen/arch/arm/domain_build.c b/xen/arch/arm/domain_build.c
index faff88e..47b781b 100644
--- a/xen/arch/arm/domain_build.c
+++ b/xen/arch/arm/domain_build.c
@@ -22,7 +22,7 @@
static unsigned int __initdata opt_dom0_max_vcpus;
integer_param("dom0_max_vcpus", opt_dom0_max_vcpus);
-static int dom0_11_mapping = 1;
+int dom0_11_mapping = 1;
#define DOM0_MEM_DEFAULT 0x8000000 /* 128 MiB */
static u64 __initdata dom0_mem = DOM0_MEM_DEFAULT;
diff --git a/xen/common/memory.c b/xen/common/memory.c
index eb7b72b..1a93196 100644
--- a/xen/common/memory.c
+++ b/xen/common/memory.c
@@ -29,6 +29,10 @@
#include <xsm/xsm.h>
#include <xen/trace.h>
+#ifndef is_domain_direct_mapped
+# define is_domain_direct_mapped(d) ((void)(d), 0)
+#endif
+
struct memop_args {
/* INPUT */
struct domain *domain; /* Domain to be affected. */
@@ -122,7 +126,29 @@ static void populate_physmap(struct memop_args *a)
}
else
{
- page = alloc_domheap_pages(d, a->extent_order, a->memflags);
+ if ( is_domain_direct_mapped(d) )
+ {
+ mfn = gpfn;
+ if ( !mfn_valid(mfn) )
+ {
+ gdprintk(XENLOG_INFO, "Invalid mfn 0x%"PRI_xen_pfn"\n",
+ mfn);
+ goto out;
+ }
+
+ page = mfn_to_page(mfn);
+ if ( !get_page(page, d) )
+ {
+ gdprintk(XENLOG_INFO,
+ "mfn 0x%"PRI_xen_pfn" doesn't belong to the"
+ " domain\n", mfn);
+ goto out;
+ }
+ put_page(page);
+ }
+ else
+ page = alloc_domheap_pages(d, a->extent_order, a->memflags);
+
if ( unlikely(page == NULL) )
{
if ( !opt_tmem || (a->extent_order != 0) )
@@ -270,6 +296,13 @@ static void decrease_reservation(struct memop_args *a)
&& p2m_pod_decrease_reservation(a->domain, gmfn, a->extent_order) )
continue;
+ /* With the lack for iommu on some ARM platform, domain with DMA-capable
+ * device must retrieve the same pfn when the hypercall
+ * populate_physmap is called.
+ */
+ if ( is_domain_direct_mapped(a->domain) )
+ continue;
+
for ( j = 0; j < (1 << a->extent_order); j++ )
if ( !guest_remove_page(a->domain, gmfn + j) )
goto out;
diff --git a/xen/include/asm-arm/domain.h b/xen/include/asm-arm/domain.h
index 28d39a0..3419c31 100644
--- a/xen/include/asm-arm/domain.h
+++ b/xen/include/asm-arm/domain.h
@@ -86,6 +86,9 @@ enum domain_type {
#define is_pv64_domain(d) (0)
#endif
+extern int dom0_11_mapping;
+#define is_domain_direct_mapped(d) (dom0_11_mapping && (d) == dom0)
+
struct vtimer {
struct vcpu *v;
int irq;
--
1.7.10.4
^ permalink raw reply related [flat|nested] 3+ messages in thread
* Re: [PATCH v5] xen/arm: Allow balooning working with 1:1 memory mapping
2013-12-16 17:00 [PATCH v5] xen/arm: Allow balooning working with 1:1 memory mapping Julien Grall
@ 2013-12-17 8:57 ` Jan Beulich
2013-12-17 14:11 ` Julien Grall
0 siblings, 1 reply; 3+ messages in thread
From: Jan Beulich @ 2013-12-17 8:57 UTC (permalink / raw)
To: Julien Grall
Cc: Keir Fraser, ian.campbell, patches, tim, stefano.stabellini, xen-devel
>>> On 16.12.13 at 18:00, Julien Grall <julien.grall@linaro.org> wrote:
> With the lack of iommu, dom0 must have a 1:1 memory mapping for all
> these guest physical address. When the ballon decides to give back a
> page to the kernel, this page must have the same address as previously.
> Otherwise, we will loose the 1:1 mapping and will break DMA-capable
> devices.
>
> Signed-off-by: Julien Grall <julien.grall@linaro.org>
> Cc: Keir Fraser <keir@xen.org>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
if you do one more change (and a little bit of further cleanup -
sorry I didn't notice this before):
> @@ -122,7 +126,29 @@ static void populate_physmap(struct memop_args *a)
> }
> else
> {
> - page = alloc_domheap_pages(d, a->extent_order, a->memflags);
> + if ( is_domain_direct_mapped(d) )
> + {
> + mfn = gpfn;
> + if ( !mfn_valid(mfn) )
> + {
> + gdprintk(XENLOG_INFO, "Invalid mfn 0x%"PRI_xen_pfn"\n",
"... %#"PRI_xen_pfn"\n" please.
> + mfn);
> + goto out;
> + }
> +
> + page = mfn_to_page(mfn);
> + if ( !get_page(page, d) )
> + {
> + gdprintk(XENLOG_INFO,
> + "mfn 0x%"PRI_xen_pfn" doesn't belong to the"
> + " domain\n", mfn);
Same here.
> --- a/xen/include/asm-arm/domain.h
> +++ b/xen/include/asm-arm/domain.h
> @@ -86,6 +86,9 @@ enum domain_type {
> #define is_pv64_domain(d) (0)
> #endif
>
> +extern int dom0_11_mapping;
> +#define is_domain_direct_mapped(d) (dom0_11_mapping && (d) == dom0)
This needs to be
#define is_domain_direct_mapped(d) ((d) == dom0 && dom0_11_mapping)
in order to make sure d is evaluated exactly once.
Jan
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCH v5] xen/arm: Allow balooning working with 1:1 memory mapping
2013-12-17 8:57 ` Jan Beulich
@ 2013-12-17 14:11 ` Julien Grall
0 siblings, 0 replies; 3+ messages in thread
From: Julien Grall @ 2013-12-17 14:11 UTC (permalink / raw)
To: Jan Beulich
Cc: Keir Fraser, ian.campbell, patches, tim, stefano.stabellini, xen-devel
On 12/17/2013 08:57 AM, Jan Beulich wrote:
>>>> On 16.12.13 at 18:00, Julien Grall <julien.grall@linaro.org> wrote:
>> With the lack of iommu, dom0 must have a 1:1 memory mapping for all
>> these guest physical address. When the ballon decides to give back a
>> page to the kernel, this page must have the same address as previously.
>> Otherwise, we will loose the 1:1 mapping and will break DMA-capable
>> devices.
>>
>> Signed-off-by: Julien Grall <julien.grall@linaro.org>
>> Cc: Keir Fraser <keir@xen.org>
>
> Reviewed-by: Jan Beulich <jbeulich@suse.com>
> if you do one more change (and a little bit of further cleanup -
> sorry I didn't notice this before):
Thanks, I will do all the changes.
--
Julien Grall
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2013-12-17 14:11 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2013-12-16 17:00 [PATCH v5] xen/arm: Allow balooning working with 1:1 memory mapping Julien Grall
2013-12-17 8:57 ` Jan Beulich
2013-12-17 14:11 ` Julien Grall
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.