* [PATCH v3 0/8] static shared memory on dom0less system
@ 2022-05-12 9:11 Penny Zheng
2022-05-12 9:11 ` [PATCH v3 1/8] xen/arm: introduce static shared memory Penny Zheng
` (7 more replies)
0 siblings, 8 replies; 18+ messages in thread
From: Penny Zheng @ 2022-05-12 9:11 UTC (permalink / raw)
To: xen-devel
Cc: wei.chen, Penny Zheng, Stefano Stabellini, Julien Grall,
Bertrand Marquis, Volodymyr Babchuk, Andrew Cooper,
George Dunlap, Jan Beulich, Wei Liu
In safety-critical environment, it is not considered safe to
dynamically change important configurations at runtime. Everything
should be statically defined and statically verified.
In this case, if the system configuration knows a priori that there are
only 2 VMs and they need to communicate over shared memory, it is safer
to pre-configure the shared memory at build time rather than let the VMs
attempt to share memory at runtime. And it is faster too.
Furthermore, on dom0less system, the legacy way to build up communication
channels between domains, like grant table, are normally absent there.
So this patch serie introduces a set of static shared memory device tree nodes
to allow users to statically set up shared memory on dom0less system, enabling
domains to do shm-based communication.
The only way to trigger this static shared memory configuration should
be via device tree, which is at the same level as the XSM rules.
It was inspired by the patch serie of ["xl/libxl-based shared mem](
https://marc.info/?l=xen-devel&m=154404821731186ory").
Looking into related [design link](
https://lore.kernel.org/all/a50d9fde-1d06-7cda-2779-9eea9e1c0134@xen.org/T/)
for more details.
Penny Zheng (8):
xen/arm: introduce static shared memory
xen/arm: allocate static shared memory to the default owner dom_io
xen/arm: allocate static shared memory to a specific owner domain
xen/arm: introduce put_page_nr and get_page_nr
xen/arm: Add additional reference to owner domain when the owner is
allocated
xen/arm: set up shared memory foreign mapping for borrower domain
xen/arm: create shared memory nodes in guest device tree
xen/arm: enable statically shared memory on Dom0
docs/misc/arm/device-tree/booting.txt | 120 ++++++++
xen/arch/arm/Kconfig | 6 +
xen/arch/arm/bootfdt.c | 68 +++++
xen/arch/arm/domain_build.c | 378 +++++++++++++++++++++++++-
xen/arch/arm/include/asm/kernel.h | 1 +
xen/arch/arm/include/asm/mm.h | 4 +
xen/arch/arm/include/asm/setup.h | 4 +
xen/arch/arm/mm.c | 36 ++-
xen/common/domain.c | 5 +
9 files changed, 612 insertions(+), 10 deletions(-)
--
2.25.1
^ permalink raw reply [flat|nested] 18+ messages in thread
* [PATCH v3 1/8] xen/arm: introduce static shared memory
2022-05-12 9:11 [PATCH v3 0/8] static shared memory on dom0less system Penny Zheng
@ 2022-05-12 9:11 ` Penny Zheng
2022-05-12 23:47 ` Stefano Stabellini
2022-05-12 9:11 ` [PATCH v3 2/8] xen/arm: allocate static shared memory to the default owner dom_io Penny Zheng
` (6 subsequent siblings)
7 siblings, 1 reply; 18+ messages in thread
From: Penny Zheng @ 2022-05-12 9:11 UTC (permalink / raw)
To: xen-devel
Cc: wei.chen, Penny Zheng, Stefano Stabellini, Julien Grall,
Bertrand Marquis, Volodymyr Babchuk
From: Penny Zheng <penny.zheng@arm.com>
This patch serie introduces a new feature: setting up static
shared memory on a dom0less system, through device tree configuration.
This commit parses shared memory node at boot-time, and reserve it in
bootinfo.reserved_mem to avoid other use.
This commits proposes a new Kconfig CONFIG_STATIC_SHM to wrap
static-shm-related codes, and this option depends on static memory(
CONFIG_STATIC_MEMORY). That's because that later we want to reuse a few
helpers, guarded with CONFIG_STATIC_MEMORY, like acquire_staticmem_pages, etc,
on static shared memory.
Signed-off-by: Penny Zheng <penny.zheng@arm.com>
---
v3 change:
- make nr_shm_domain unsigned int
---
v2 change:
- document refinement
- remove bitmap and use the iteration to check
- add a new field nr_shm_domain to keep the number of shared domain
---
docs/misc/arm/device-tree/booting.txt | 120 ++++++++++++++++++++++++++
xen/arch/arm/Kconfig | 6 ++
xen/arch/arm/bootfdt.c | 68 +++++++++++++++
xen/arch/arm/include/asm/setup.h | 3 +
4 files changed, 197 insertions(+)
diff --git a/docs/misc/arm/device-tree/booting.txt b/docs/misc/arm/device-tree/booting.txt
index a94125394e..e63ce171fc 100644
--- a/docs/misc/arm/device-tree/booting.txt
+++ b/docs/misc/arm/device-tree/booting.txt
@@ -355,3 +355,123 @@ device-tree:
This will reserve a 512MB region starting at the host physical address
0x30000000 to be exclusively used by DomU1.
+
+Static Shared Memory
+====================
+
+The static shared memory device tree nodes allow users to statically set up
+shared memory on dom0less system, enabling domains to do shm-based
+communication.
+
+- compatible
+
+ "xen,domain-shared-memory-v1"
+
+- xen,shm-id
+
+ An u8 value represents the unique identifier of the shared memory region.
+ The maximum identifier shall be "xen,shm-id = <0xff>".
+
+- xen,shared-mem
+
+ An array takes a physical address, which is the base address of the
+ shared memory region in host physical address space, a size, and a guest
+ physical address, as the target address of the mapping. The number of cells
+ for the host address (and size) is the same as the guest pseudo-physical
+ address and they are inherited from the parent node.
+
+- role (Optional)
+
+ A string property specifying the ownership of a shared memory region,
+ the value must be one of the following: "owner", or "borrower"
+ A shared memory region could be explicitly backed by one domain, which is
+ called "owner domain", and all the other domains who are also sharing
+ this region are called "borrower domain".
+ If not specified, the default value is "borrower" and owner is
+ "dom_shared", a system domain.
+
+As an example:
+
+chosen {
+ #address-cells = <0x1>;
+ #size-cells = <0x1>;
+ xen,xen-bootargs = "console=dtuart dtuart=serial0 bootscrub=0";
+
+ ......
+
+ /* this is for Dom0 */
+ dom0-shared-mem@10000000 {
+ compatible = "xen,domain-shared-memory-v1";
+ role = "owner";
+ xen,shm-id = <0x0>;
+ xen,shared-mem = <0x10000000 0x10000000 0x10000000>;
+ }
+
+ domU1 {
+ compatible = "xen,domain";
+ #address-cells = <0x1>;
+ #size-cells = <0x1>;
+ memory = <0 131072>;
+ cpus = <2>;
+ vpl011;
+
+ /*
+ * shared memory region identified as 0x0(xen,shm-id = <0x0>)
+ * is shared between Dom0 and DomU1.
+ */
+ domU1-shared-mem@10000000 {
+ compatible = "xen,domain-shared-memory-v1";
+ role = "borrower";
+ xen,shm-id = <0x0>;
+ xen,shared-mem = <0x10000000 0x10000000 0x50000000>;
+ }
+
+ /*
+ * shared memory region identified as 0x1(xen,shm-id = <0x1>)
+ * is shared between DomU1 and DomU2.
+ */
+ domU1-shared-mem@50000000 {
+ compatible = "xen,domain-shared-memory-v1";
+ xen,shm-id = <0x1>;
+ xen,shared-mem = <0x50000000 0x20000000 0x60000000>;
+ }
+
+ ......
+
+ };
+
+ domU2 {
+ compatible = "xen,domain";
+ #address-cells = <0x1>;
+ #size-cells = <0x1>;
+ memory = <0 65536>;
+ cpus = <1>;
+
+ /*
+ * shared memory region identified as 0x1(xen,shm-id = <0x1>)
+ * is shared between domU1 and domU2.
+ */
+ domU2-shared-mem@50000000 {
+ compatible = "xen,domain-shared-memory-v1";
+ xen,shm-id = <0x1>;
+ xen,shared-mem = <0x50000000 0x20000000 0x70000000>;
+ }
+
+ ......
+ };
+};
+
+This is an example with two static shared memory regions.
+
+For the static shared memory region identified as 0x0, host physical
+address starting at 0x10000000 of 256MB will be reserved to be shared between
+Dom0 and DomU1. It will get mapped at 0x10000000 in Dom0 guest physical address
+space, and at 0x50000000 in DomU1 guest physical address space. Dom0 is
+explicitly defined as the owner domain, and DomU1 is the borrower domain.
+
+For the static shared memory region identified as 0x1, host physical
+address starting at 0x50000000 of 512MB will be reserved to be shared between
+DomU1 and DomU2. It will get mapped at 0x60000000 in DomU1 guest physical
+address space, and at 0x70000000 in DomU2 guest physical address space. DomU1
+and DomU2 are both the borrower domain, the owner domain is the default owner
+domain dom_shared.
diff --git a/xen/arch/arm/Kconfig b/xen/arch/arm/Kconfig
index ecfa6822e4..5ee9921f56 100644
--- a/xen/arch/arm/Kconfig
+++ b/xen/arch/arm/Kconfig
@@ -106,6 +106,12 @@ config TEE
source "arch/arm/tee/Kconfig"
+config STATIC_SHM
+ bool "Statically shared memory on a dom0less system" if UNSUPPORTED
+ depends on STATIC_MEMORY
+ help
+ This option enables statically shared memory on a dom0less system.
+
endmenu
menu "ARM errata workaround via the alternative framework"
diff --git a/xen/arch/arm/bootfdt.c b/xen/arch/arm/bootfdt.c
index 29671c8df0..c8e8747c3e 100644
--- a/xen/arch/arm/bootfdt.c
+++ b/xen/arch/arm/bootfdt.c
@@ -360,6 +360,70 @@ static int __init process_domain_node(const void *fdt, int node,
size_cells, &bootinfo.reserved_mem, true);
}
+#ifdef CONFIG_STATIC_SHM
+static int __init process_shm_node(const void *fdt, int node,
+ u32 address_cells, u32 size_cells)
+{
+ const struct fdt_property *prop;
+ const __be32 *cell;
+ paddr_t paddr, size;
+ struct meminfo *mem = &bootinfo.reserved_mem;
+ unsigned long i;
+
+ if ( address_cells < 1 || size_cells < 1 )
+ {
+ printk("fdt: invalid #address-cells or #size-cells for static shared memory node.\n");
+ return -EINVAL;
+ }
+
+ prop = fdt_get_property(fdt, node, "xen,shared-mem", NULL);
+ if ( !prop )
+ return -ENOENT;
+
+ /*
+ * xen,shared-mem = <paddr, size, gaddr>;
+ * Memory region starting from physical address #paddr of #size shall
+ * be mapped to guest physical address #gaddr as static shared memory
+ * region.
+ */
+ cell = (const __be32 *)prop->data;
+ device_tree_get_reg(&cell, address_cells, size_cells, &paddr, &size);
+ for ( i = 0; i < mem->nr_banks; i++ )
+ {
+ /*
+ * A static shared memory region could be shared between multiple
+ * domains.
+ */
+ if ( paddr == mem->bank[i].start && size == mem->bank[i].size )
+ break;
+ }
+
+ if ( i == mem->nr_banks )
+ {
+ if ( i < NR_MEM_BANKS )
+ {
+ /* Static shared memory shall be reserved from any other use. */
+ mem->bank[mem->nr_banks].start = paddr;
+ mem->bank[mem->nr_banks].size = size;
+ mem->bank[mem->nr_banks].xen_domain = true;
+ mem->nr_banks++;
+ }
+ else
+ {
+ printk("Warning: Max number of supported memory regions reached.\n");
+ return -ENOSPC;
+ }
+ }
+ /*
+ * keep a count of the number of domains, which later may be used to
+ * calculate the number of the reference count.
+ */
+ mem->bank[i].nr_shm_domain++;
+
+ return 0;
+}
+#endif
+
static int __init early_scan_node(const void *fdt,
int node, const char *name, int depth,
u32 address_cells, u32 size_cells,
@@ -380,6 +444,10 @@ static int __init early_scan_node(const void *fdt,
process_chosen_node(fdt, node, name, address_cells, size_cells);
else if ( depth == 2 && device_tree_node_compatible(fdt, node, "xen,domain") )
rc = process_domain_node(fdt, node, name, address_cells, size_cells);
+#ifdef CONFIG_STATIC_SHM
+ else if ( depth <= 3 && device_tree_node_compatible(fdt, node, "xen,domain-shared-memory-v1") )
+ rc = process_shm_node(fdt, node, address_cells, size_cells);
+#endif
if ( rc < 0 )
printk("fdt: node `%s': parsing failed\n", name);
diff --git a/xen/arch/arm/include/asm/setup.h b/xen/arch/arm/include/asm/setup.h
index 7a1e1d6798..35449bd7d8 100644
--- a/xen/arch/arm/include/asm/setup.h
+++ b/xen/arch/arm/include/asm/setup.h
@@ -27,6 +27,9 @@ struct membank {
paddr_t start;
paddr_t size;
bool xen_domain; /* whether the memory bank is bound to a Xen domain. */
+#ifdef CONFIG_STATIC_SHM
+ unsigned int nr_shm_domain;
+#endif
};
struct meminfo {
--
2.25.1
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH v3 2/8] xen/arm: allocate static shared memory to the default owner dom_io
2022-05-12 9:11 [PATCH v3 0/8] static shared memory on dom0less system Penny Zheng
2022-05-12 9:11 ` [PATCH v3 1/8] xen/arm: introduce static shared memory Penny Zheng
@ 2022-05-12 9:11 ` Penny Zheng
2022-05-12 23:47 ` Stefano Stabellini
2022-05-12 9:11 ` [PATCH v3 3/8] xen/arm: allocate static shared memory to a specific owner domain Penny Zheng
` (5 subsequent siblings)
7 siblings, 1 reply; 18+ messages in thread
From: Penny Zheng @ 2022-05-12 9:11 UTC (permalink / raw)
To: xen-devel
Cc: wei.chen, Penny Zheng, Stefano Stabellini, Julien Grall,
Bertrand Marquis, Volodymyr Babchuk, Andrew Cooper,
George Dunlap, Jan Beulich, Wei Liu
From: Penny Zheng <penny.zheng@arm.com>
This commit introduces process_shm to cope with static shared memory in
domain construction.
DOMID_IO will be the default owner of memory pre-shared among multiple domains
at boot time, when no explicit owner is specified.
This commit only considers allocating static shared memory to dom_io
when owner domain is not explicitly defined in device tree, all the left,
including the "borrower" code path, the "explicit owner" code path, shall
be introduced later in the following patches.
Signed-off-by: Penny Zheng <penny.zheng@arm.com>
---
v3 change:
- refine in-code comment
---
v2 change:
- instead of introducing a new system domain, reuse the existing dom_io
- make dom_io a non-auto-translated domain, then no need to create P2M
for it
- change dom_io definition and make it wider to support static shm here too
- introduce is_shm_allocated_to_domio to check whether static shm is
allocated yet, instead of using shm_mask bitmap
- add in-code comment
---
xen/arch/arm/domain_build.c | 132 +++++++++++++++++++++++++++++++++++-
xen/common/domain.c | 5 ++
2 files changed, 136 insertions(+), 1 deletion(-)
diff --git a/xen/arch/arm/domain_build.c b/xen/arch/arm/domain_build.c
index 1472ca4972..5b41ec57f0 100644
--- a/xen/arch/arm/domain_build.c
+++ b/xen/arch/arm/domain_build.c
@@ -519,6 +519,10 @@ static bool __init append_static_memory_to_bank(struct domain *d,
return true;
}
+/*
+ * If cell is NULL, pbase and psize should hold valid values.
+ * Otherwise, cell will be populated together with pbase and psize.
+ */
static mfn_t __init acquire_static_memory_bank(struct domain *d,
const __be32 **cell,
u32 addr_cells, u32 size_cells,
@@ -527,7 +531,8 @@ static mfn_t __init acquire_static_memory_bank(struct domain *d,
mfn_t smfn;
int res;
- device_tree_get_reg(cell, addr_cells, size_cells, pbase, psize);
+ if ( cell )
+ device_tree_get_reg(cell, addr_cells, size_cells, pbase, psize);
ASSERT(IS_ALIGNED(*pbase, PAGE_SIZE) && IS_ALIGNED(*psize, PAGE_SIZE));
if ( PFN_DOWN(*psize) > UINT_MAX )
{
@@ -751,6 +756,125 @@ static void __init assign_static_memory_11(struct domain *d,
panic("Failed to assign requested static memory for direct-map domain %pd.",
d);
}
+
+#ifdef CONFIG_STATIC_SHM
+/*
+ * This function checks whether the static shared memory region is
+ * already allocated to dom_io.
+ */
+static bool __init is_shm_allocated_to_domio(paddr_t pbase)
+{
+ struct page_info *page;
+
+ page = maddr_to_page(pbase);
+ ASSERT(page);
+
+ if ( page_get_owner(page) == NULL )
+ return false;
+
+ ASSERT(page_get_owner(page) == dom_io);
+ return true;
+}
+
+static mfn_t __init acquire_shared_memory_bank(struct domain *d,
+ u32 addr_cells, u32 size_cells,
+ paddr_t *pbase, paddr_t *psize)
+{
+ /*
+ * Pages of statically shared memory shall be included
+ * in domain_tot_pages().
+ */
+ d->max_pages += PFN_DOWN(*psize);
+
+ return acquire_static_memory_bank(d, NULL, addr_cells, size_cells,
+ pbase, psize);
+
+}
+
+/*
+ * Func allocate_shared_memory is supposed to be only called
+ * from the owner.
+ */
+static int __init allocate_shared_memory(struct domain *d,
+ u32 addr_cells, u32 size_cells,
+ paddr_t pbase, paddr_t psize)
+{
+ mfn_t smfn;
+
+ dprintk(XENLOG_INFO,
+ "Allocate static shared memory BANK %#"PRIpaddr"-%#"PRIpaddr".\n",
+ pbase, pbase + psize);
+
+ smfn = acquire_shared_memory_bank(d, addr_cells, size_cells, &pbase,
+ &psize);
+ if ( mfn_eq(smfn, INVALID_MFN) )
+ return -EINVAL;
+
+ /*
+ * DOMID_IO is the domain, like DOMID_XEN, that is not auto-translated.
+ * It sees RAM 1:1 and we do not need to create P2M mapping for it
+ */
+ ASSERT(d == dom_io);
+ return 0;
+}
+
+static int __init process_shm(struct domain *d,
+ const struct dt_device_node *node)
+{
+ struct dt_device_node *shm_node;
+ int ret = 0;
+ const struct dt_property *prop;
+ const __be32 *cells;
+ u32 shm_id;
+ u32 addr_cells, size_cells;
+ paddr_t gbase, pbase, psize;
+
+ dt_for_each_child_node(node, shm_node)
+ {
+ if ( !dt_device_is_compatible(shm_node, "xen,domain-shared-memory-v1") )
+ continue;
+
+ if ( !dt_property_read_u32(shm_node, "xen,shm-id", &shm_id) )
+ {
+ printk("Shared memory node does not provide \"xen,shm-id\" property.\n");
+ return -ENOENT;
+ }
+
+ addr_cells = dt_n_addr_cells(shm_node);
+ size_cells = dt_n_size_cells(shm_node);
+ prop = dt_find_property(shm_node, "xen,shared-mem", NULL);
+ if ( !prop )
+ {
+ printk("Shared memory node does not provide \"xen,shared-mem\" property.\n");
+ return -ENOENT;
+ }
+ cells = (const __be32 *)prop->value;
+ /* xen,shared-mem = <pbase, psize, gbase>; */
+ device_tree_get_reg(&cells, addr_cells, size_cells, &pbase, &psize);
+ ASSERT(IS_ALIGNED(pbase, PAGE_SIZE) && IS_ALIGNED(psize, PAGE_SIZE));
+ gbase = dt_read_number(cells, addr_cells);
+
+ /* TODO: Consider owner domain is not the default dom_io. */
+ /*
+ * Per static shared memory region could be shared between multiple
+ * domains.
+ * In case re-allocating the same shared memory region, we check
+ * if it is already allocated to the default owner dom_io before
+ * the actual allocation.
+ */
+ if ( !is_shm_allocated_to_domio(pbase) )
+ {
+ /* Allocate statically shared pages to the default owner dom_io. */
+ ret = allocate_shared_memory(dom_io, addr_cells, size_cells,
+ pbase, psize);
+ if ( ret )
+ return ret;
+ }
+ }
+
+ return 0;
+}
+#endif /* CONFIG_STATIC_SHM */
#else
static void __init allocate_static_memory(struct domain *d,
struct kernel_info *kinfo,
@@ -3149,6 +3273,12 @@ static int __init construct_domU(struct domain *d,
else
assign_static_memory_11(d, &kinfo, node);
+#ifdef CONFIG_STATIC_SHM
+ rc = process_shm(d, node);
+ if ( rc < 0 )
+ return rc;
+#endif
+
/*
* Base address and irq number are needed when creating vpl011 device
* tree node in prepare_dtb_domU, so initialization on related variables
diff --git a/xen/common/domain.c b/xen/common/domain.c
index 8d2c2a9897..43237c287a 100644
--- a/xen/common/domain.c
+++ b/xen/common/domain.c
@@ -780,6 +780,11 @@ void __init setup_system_domains(void)
* This domain owns I/O pages that are within the range of the page_info
* array. Mappings occur at the priv of the caller.
* Quarantined PCI devices will be associated with this domain.
+ *
+ * DOMID_IO could also be used for mapping memory when no explicit
+ * domain is specified.
+ * For instance, DOMID_IO is the owner of memory pre-shared among
+ * multiple domains at boot time, when no explicit owner is specified.
*/
dom_io = domain_create(DOMID_IO, NULL, 0);
if ( IS_ERR(dom_io) )
--
2.25.1
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH v3 3/8] xen/arm: allocate static shared memory to a specific owner domain
2022-05-12 9:11 [PATCH v3 0/8] static shared memory on dom0less system Penny Zheng
2022-05-12 9:11 ` [PATCH v3 1/8] xen/arm: introduce static shared memory Penny Zheng
2022-05-12 9:11 ` [PATCH v3 2/8] xen/arm: allocate static shared memory to the default owner dom_io Penny Zheng
@ 2022-05-12 9:11 ` Penny Zheng
2022-05-12 23:47 ` Stefano Stabellini
2022-05-12 9:11 ` [PATCH v3 4/8] xen/arm: introduce put_page_nr and get_page_nr Penny Zheng
` (4 subsequent siblings)
7 siblings, 1 reply; 18+ messages in thread
From: Penny Zheng @ 2022-05-12 9:11 UTC (permalink / raw)
To: xen-devel
Cc: wei.chen, Penny Zheng, Stefano Stabellini, Julien Grall,
Bertrand Marquis, Volodymyr Babchuk, Penny Zheng
If owner property is defined, then owner domain of a static shared memory
region is not the default dom_io anymore, but a specific domain.
This commit implements allocating static shared memory to a specific domain
when owner property is defined.
Coding flow for dealing borrower domain will be introduced later in the
following commits.
Signed-off-by: Penny Zheng <penny.zheng@arm.com>
---
v3 change:
- simplify the code since o_gbase is not used if the domain is dom_io
---
v2 change:
- P2M mapping is restricted to normal domain
- in-code comment fix
---
xen/arch/arm/domain_build.c | 44 +++++++++++++++++++++++++++----------
1 file changed, 33 insertions(+), 11 deletions(-)
diff --git a/xen/arch/arm/domain_build.c b/xen/arch/arm/domain_build.c
index 5b41ec57f0..9c53b9fa7b 100644
--- a/xen/arch/arm/domain_build.c
+++ b/xen/arch/arm/domain_build.c
@@ -797,9 +797,11 @@ static mfn_t __init acquire_shared_memory_bank(struct domain *d,
*/
static int __init allocate_shared_memory(struct domain *d,
u32 addr_cells, u32 size_cells,
- paddr_t pbase, paddr_t psize)
+ paddr_t pbase, paddr_t psize,
+ paddr_t gbase)
{
mfn_t smfn;
+ int ret = 0;
dprintk(XENLOG_INFO,
"Allocate static shared memory BANK %#"PRIpaddr"-%#"PRIpaddr".\n",
@@ -814,8 +816,18 @@ static int __init allocate_shared_memory(struct domain *d,
* DOMID_IO is the domain, like DOMID_XEN, that is not auto-translated.
* It sees RAM 1:1 and we do not need to create P2M mapping for it
*/
- ASSERT(d == dom_io);
- return 0;
+ if ( d != dom_io )
+ {
+ ret = guest_physmap_add_pages(d, gaddr_to_gfn(gbase), smfn, PFN_DOWN(psize));
+ if ( ret )
+ {
+ printk(XENLOG_ERR
+ "Failed to map shared memory to %pd.\n", d);
+ return ret;
+ }
+ }
+
+ return ret;
}
static int __init process_shm(struct domain *d,
@@ -828,6 +840,8 @@ static int __init process_shm(struct domain *d,
u32 shm_id;
u32 addr_cells, size_cells;
paddr_t gbase, pbase, psize;
+ const char *role_str;
+ bool owner_dom_io = true;
dt_for_each_child_node(node, shm_node)
{
@@ -854,19 +868,27 @@ static int __init process_shm(struct domain *d,
ASSERT(IS_ALIGNED(pbase, PAGE_SIZE) && IS_ALIGNED(psize, PAGE_SIZE));
gbase = dt_read_number(cells, addr_cells);
- /* TODO: Consider owner domain is not the default dom_io. */
+ /*
+ * "role" property is optional and if it is defined explicitly,
+ * then the owner domain is not the default "dom_io" domain.
+ */
+ if ( dt_property_read_string(shm_node, "role", &role_str) == 0 )
+ owner_dom_io = false;
+
/*
* Per static shared memory region could be shared between multiple
* domains.
- * In case re-allocating the same shared memory region, we check
- * if it is already allocated to the default owner dom_io before
- * the actual allocation.
+ * So when owner domain is the default dom_io, in case re-allocating
+ * the same shared memory region, we check if it is already allocated
+ * to the default owner dom_io before the actual allocation.
*/
- if ( !is_shm_allocated_to_domio(pbase) )
+ if ( (owner_dom_io && !is_shm_allocated_to_domio(pbase)) ||
+ (!owner_dom_io && strcmp(role_str, "owner") == 0) )
{
- /* Allocate statically shared pages to the default owner dom_io. */
- ret = allocate_shared_memory(dom_io, addr_cells, size_cells,
- pbase, psize);
+ /* Allocate statically shared pages to the owner domain. */
+ ret = allocate_shared_memory(owner_dom_io ? dom_io : d,
+ addr_cells, size_cells,
+ pbase, psize, gbase);
if ( ret )
return ret;
}
--
2.25.1
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH v3 4/8] xen/arm: introduce put_page_nr and get_page_nr
2022-05-12 9:11 [PATCH v3 0/8] static shared memory on dom0less system Penny Zheng
` (2 preceding siblings ...)
2022-05-12 9:11 ` [PATCH v3 3/8] xen/arm: allocate static shared memory to a specific owner domain Penny Zheng
@ 2022-05-12 9:11 ` Penny Zheng
2022-05-12 10:14 ` Julien Grall
2022-05-12 9:11 ` [PATCH v3 5/8] xen/arm: Add additional reference to owner domain when the owner is allocated Penny Zheng
` (3 subsequent siblings)
7 siblings, 1 reply; 18+ messages in thread
From: Penny Zheng @ 2022-05-12 9:11 UTC (permalink / raw)
To: xen-devel
Cc: wei.chen, Penny Zheng, Stefano Stabellini, Julien Grall,
Bertrand Marquis, Volodymyr Babchuk, Penny Zheng
Later, we need to add the right amount of references, which should be
the number of borrower domains, to the owner domain. Since we only have
get_page() to increment the page reference by 1, a loop is needed per
page, which is inefficient and time-consuming.
To save the loop time, this commit introduces a set of new helpers
put_page_nr() and get_page_nr() to increment/drop the page reference by nr.
Signed-off-by: Penny Zheng <penny.zheng@arm.com>
---
v3 changes:
- check overflow with "n"
- remove spurious change
- bring back the check that we enter the loop only when count_info is
greater than 0
---
v2 change:
- new commit
---
xen/arch/arm/include/asm/mm.h | 4 ++++
xen/arch/arm/mm.c | 36 ++++++++++++++++++++++++++---------
2 files changed, 31 insertions(+), 9 deletions(-)
diff --git a/xen/arch/arm/include/asm/mm.h b/xen/arch/arm/include/asm/mm.h
index 424aaf2823..c737d51e4d 100644
--- a/xen/arch/arm/include/asm/mm.h
+++ b/xen/arch/arm/include/asm/mm.h
@@ -347,6 +347,10 @@ void free_init_memory(void);
int guest_physmap_mark_populate_on_demand(struct domain *d, unsigned long gfn,
unsigned int order);
+extern bool get_page_nr(struct page_info *page, const struct domain *domain,
+ unsigned long nr);
+extern void put_page_nr(struct page_info *page, unsigned long nr);
+
extern void put_page_type(struct page_info *page);
static inline void put_page_and_type(struct page_info *page)
{
diff --git a/xen/arch/arm/mm.c b/xen/arch/arm/mm.c
index 7b1f2f4906..a9461e07aa 100644
--- a/xen/arch/arm/mm.c
+++ b/xen/arch/arm/mm.c
@@ -1537,7 +1537,8 @@ long arch_memory_op(int op, XEN_GUEST_HANDLE_PARAM(void) arg)
return 0;
}
-struct domain *page_get_owner_and_reference(struct page_info *page)
+static struct domain *page_get_owner_and_nr_reference(struct page_info *page,
+ unsigned long nr)
{
unsigned long x, y = page->count_info;
struct domain *owner;
@@ -1548,10 +1549,10 @@ struct domain *page_get_owner_and_reference(struct page_info *page)
* Count == 0: Page is not allocated, so we cannot take a reference.
* Count == -1: Reference count would wrap, which is invalid.
*/
- if ( unlikely(((x + 1) & PGC_count_mask) <= 1) )
+ if ( unlikely(((x + nr) & PGC_count_mask) <= 1) )
return NULL;
}
- while ( (y = cmpxchg(&page->count_info, x, x + 1)) != x );
+ while ( (y = cmpxchg(&page->count_info, x, x + nr)) != x );
owner = page_get_owner(page);
ASSERT(owner);
@@ -1559,14 +1560,20 @@ struct domain *page_get_owner_and_reference(struct page_info *page)
return owner;
}
-void put_page(struct page_info *page)
+struct domain *page_get_owner_and_reference(struct page_info *page)
+{
+ return page_get_owner_and_nr_reference(page, 1);
+}
+
+void put_page_nr(struct page_info *page, unsigned long nr)
{
unsigned long nx, x, y = page->count_info;
do {
- ASSERT((y & PGC_count_mask) != 0);
+ ASSERT(((y & PGC_count_mask) != 0) &&
+ (((y - nr) & PGC_count_mask) >= 0));
x = y;
- nx = x - 1;
+ nx = x - nr;
}
while ( unlikely((y = cmpxchg(&page->count_info, x, nx)) != x) );
@@ -1576,19 +1583,30 @@ void put_page(struct page_info *page)
}
}
-bool get_page(struct page_info *page, const struct domain *domain)
+void put_page(struct page_info *page)
{
- const struct domain *owner = page_get_owner_and_reference(page);
+ put_page_nr(page, 1);
+}
+
+bool get_page_nr(struct page_info *page, const struct domain *domain,
+ unsigned long nr)
+{
+ const struct domain *owner = page_get_owner_and_nr_reference(page, nr);
if ( likely(owner == domain) )
return true;
if ( owner != NULL )
- put_page(page);
+ put_page_nr(page, nr);
return false;
}
+bool get_page(struct page_info *page, const struct domain *domain)
+{
+ return get_page_nr(page, domain, 1);
+}
+
/* Common code requires get_page_type and put_page_type.
* We don't care about typecounts so we just do the minimum to make it
* happy. */
--
2.25.1
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH v3 5/8] xen/arm: Add additional reference to owner domain when the owner is allocated
2022-05-12 9:11 [PATCH v3 0/8] static shared memory on dom0less system Penny Zheng
` (3 preceding siblings ...)
2022-05-12 9:11 ` [PATCH v3 4/8] xen/arm: introduce put_page_nr and get_page_nr Penny Zheng
@ 2022-05-12 9:11 ` Penny Zheng
2022-05-12 10:54 ` Julien Grall
2022-05-12 9:11 ` [PATCH v3 6/8] xen/arm: set up shared memory foreign mapping for borrower domain Penny Zheng
` (2 subsequent siblings)
7 siblings, 1 reply; 18+ messages in thread
From: Penny Zheng @ 2022-05-12 9:11 UTC (permalink / raw)
To: xen-devel
Cc: wei.chen, Penny Zheng, Stefano Stabellini, Julien Grall,
Bertrand Marquis, Volodymyr Babchuk, Penny Zheng
Borrower domain will fail to get a page ref using the owner domain
during allocation, when the owner is created after borrower.
So here, we decide to get and add the right amount of reference, which
is the number of borrowers, when the owner is allocated.
Signed-off-by: Penny Zheng <penny.zheng@arm.com>
---
v3 change:
- printk rather than dprintk since it is a serious error
---
v2 change:
- new commit
---
xen/arch/arm/domain_build.c | 62 +++++++++++++++++++++++++++++++++++++
1 file changed, 62 insertions(+)
diff --git a/xen/arch/arm/domain_build.c b/xen/arch/arm/domain_build.c
index 9c53b9fa7b..293e79e4e7 100644
--- a/xen/arch/arm/domain_build.c
+++ b/xen/arch/arm/domain_build.c
@@ -791,6 +791,34 @@ static mfn_t __init acquire_shared_memory_bank(struct domain *d,
}
+static int __init acquire_nr_borrower_domain(struct domain *d,
+ paddr_t pbase, paddr_t psize,
+ unsigned long *nr_borrowers)
+{
+ unsigned long bank;
+
+ /* Iterate reserved memory to find requested shm bank. */
+ for ( bank = 0 ; bank < bootinfo.reserved_mem.nr_banks; bank++ )
+ {
+ paddr_t bank_start = bootinfo.reserved_mem.bank[bank].start;
+ paddr_t bank_size = bootinfo.reserved_mem.bank[bank].size;
+
+ if ( pbase == bank_start && psize == bank_size )
+ break;
+ }
+
+ if ( bank == bootinfo.reserved_mem.nr_banks )
+ return -ENOENT;
+
+ if ( d == dom_io )
+ *nr_borrowers = bootinfo.reserved_mem.bank[bank].nr_shm_domain;
+ else
+ /* Exclude the owner domain itself. */
+ *nr_borrowers = bootinfo.reserved_mem.bank[bank].nr_shm_domain - 1;
+
+ return 0;
+}
+
/*
* Func allocate_shared_memory is supposed to be only called
* from the owner.
@@ -802,6 +830,8 @@ static int __init allocate_shared_memory(struct domain *d,
{
mfn_t smfn;
int ret = 0;
+ unsigned long nr_pages, nr_borrowers, i;
+ struct page_info *page;
dprintk(XENLOG_INFO,
"Allocate static shared memory BANK %#"PRIpaddr"-%#"PRIpaddr".\n",
@@ -816,6 +846,7 @@ static int __init allocate_shared_memory(struct domain *d,
* DOMID_IO is the domain, like DOMID_XEN, that is not auto-translated.
* It sees RAM 1:1 and we do not need to create P2M mapping for it
*/
+ nr_pages = PFN_DOWN(psize);
if ( d != dom_io )
{
ret = guest_physmap_add_pages(d, gaddr_to_gfn(gbase), smfn, PFN_DOWN(psize));
@@ -827,6 +858,37 @@ static int __init allocate_shared_memory(struct domain *d,
}
}
+ /*
+ * Get the right amount of references per page, which is the number of
+ * borrow domains.
+ */
+ ret = acquire_nr_borrower_domain(d, pbase, psize, &nr_borrowers);
+ if ( ret )
+ return ret;
+
+ /*
+ * Instead of let borrower domain get a page ref, we add as many
+ * additional reference as the number of borrowers when the owner
+ * is allocated, since there is a chance that owner is created
+ * after borrower.
+ */
+ page = mfn_to_page(smfn);
+ for ( i = 0; i < nr_pages; i++ )
+ {
+ if ( !get_page_nr(page + i, d, nr_borrowers) )
+ {
+ printk(XENLOG_ERR
+ "Failed to add %lu references to page %"PRI_mfn".\n",
+ nr_borrowers, mfn_x(smfn) + i);
+ goto fail;
+ }
+ }
+
+ return 0;
+
+ fail:
+ while ( --i >= 0 )
+ put_page_nr(page + i, nr_borrowers);
return ret;
}
--
2.25.1
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH v3 6/8] xen/arm: set up shared memory foreign mapping for borrower domain
2022-05-12 9:11 [PATCH v3 0/8] static shared memory on dom0less system Penny Zheng
` (4 preceding siblings ...)
2022-05-12 9:11 ` [PATCH v3 5/8] xen/arm: Add additional reference to owner domain when the owner is allocated Penny Zheng
@ 2022-05-12 9:11 ` Penny Zheng
2022-05-12 9:11 ` [PATCH v3 7/8] xen/arm: create shared memory nodes in guest device tree Penny Zheng
2022-05-12 9:11 ` [PATCH v3 8/8] xen/arm: enable statically shared memory on Dom0 Penny Zheng
7 siblings, 0 replies; 18+ messages in thread
From: Penny Zheng @ 2022-05-12 9:11 UTC (permalink / raw)
To: xen-devel
Cc: wei.chen, Penny Zheng, Stefano Stabellini, Julien Grall,
Bertrand Marquis, Volodymyr Babchuk, Penny Zheng
This commit sets up shared memory foreign mapping for borrower domain.
If owner domain is the default dom_io, all shared domain are treated as
borrower domain.
Signed-off-by: Penny Zheng <penny.zheng@arm.com>
Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
---
v3 change:
- use map_regions_p2mt instead
---
v2 change:
- remove guest_physmap_add_shm, since for borrower domain, we only
do P2M foreign memory mapping now.
---
xen/arch/arm/domain_build.c | 9 +++++++++
1 file changed, 9 insertions(+)
diff --git a/xen/arch/arm/domain_build.c b/xen/arch/arm/domain_build.c
index 293e79e4e7..aa7c264e23 100644
--- a/xen/arch/arm/domain_build.c
+++ b/xen/arch/arm/domain_build.c
@@ -954,6 +954,15 @@ static int __init process_shm(struct domain *d,
if ( ret )
return ret;
}
+
+ if ( owner_dom_io || (strcmp(role_str, "borrower") == 0) )
+ {
+ /* Set up P2M foreign mapping for borrower domain. */
+ ret = map_regions_p2mt(d, _gfn(PFN_UP(gbase)), PFN_DOWN(psize),
+ _mfn(PFN_UP(pbase)), p2m_map_foreign_rw);
+ if ( ret )
+ return ret;
+ }
}
return 0;
--
2.25.1
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH v3 7/8] xen/arm: create shared memory nodes in guest device tree
2022-05-12 9:11 [PATCH v3 0/8] static shared memory on dom0less system Penny Zheng
` (5 preceding siblings ...)
2022-05-12 9:11 ` [PATCH v3 6/8] xen/arm: set up shared memory foreign mapping for borrower domain Penny Zheng
@ 2022-05-12 9:11 ` Penny Zheng
2022-05-12 23:47 ` Stefano Stabellini
2022-05-12 9:11 ` [PATCH v3 8/8] xen/arm: enable statically shared memory on Dom0 Penny Zheng
7 siblings, 1 reply; 18+ messages in thread
From: Penny Zheng @ 2022-05-12 9:11 UTC (permalink / raw)
To: xen-devel
Cc: wei.chen, Penny Zheng, Stefano Stabellini, Julien Grall,
Bertrand Marquis, Volodymyr Babchuk, Penny Zheng
We expose the shared memory to the domU using the "xen,shared-memory-v1"
reserved-memory binding. See
Documentation/devicetree/bindings/reserved-memory/xen,shared-memory.txt
in Linux for the corresponding device tree binding.
To save the cost of re-parsing shared memory device tree configuration when
creating shared memory nodes in guest device tree, this commit adds new field
"shm_mem" to store shm-info per domain.
For each shared memory region, a range is exposed under
the /reserved-memory node as a child node. Each range sub-node is
named xen-shmem@<address> and has the following properties:
- compatible:
compatible = "xen,shared-memory-v1"
- reg:
the base guest physical address and size of the shared memory region
- xen,id:
a string that identifies the shared memory region.
Signed-off-by: Penny Zheng <penny.zheng@arm.com>
---
v3 change:
- move field "shm_mem" to kernel_info
---
v2 change:
- using xzalloc
- shm_id should be uint8_t
- make reg a local variable
- add #address-cells and #size-cells properties
- fix alignment
---
xen/arch/arm/domain_build.c | 143 +++++++++++++++++++++++++++++-
xen/arch/arm/include/asm/kernel.h | 1 +
xen/arch/arm/include/asm/setup.h | 1 +
3 files changed, 143 insertions(+), 2 deletions(-)
diff --git a/xen/arch/arm/domain_build.c b/xen/arch/arm/domain_build.c
index aa7c264e23..fb9146b6e0 100644
--- a/xen/arch/arm/domain_build.c
+++ b/xen/arch/arm/domain_build.c
@@ -892,7 +892,22 @@ static int __init allocate_shared_memory(struct domain *d,
return ret;
}
-static int __init process_shm(struct domain *d,
+static int __init append_shm_bank_to_domain(struct kernel_info *kinfo,
+ paddr_t start, paddr_t size,
+ u32 shm_id)
+{
+ if ( (kinfo->shm_mem.nr_banks + 1) > NR_MEM_BANKS )
+ return -ENOMEM;
+
+ kinfo->shm_mem.bank[kinfo->shm_mem.nr_banks].start = start;
+ kinfo->shm_mem.bank[kinfo->shm_mem.nr_banks].size = size;
+ kinfo->shm_mem.bank[kinfo->shm_mem.nr_banks].shm_id = shm_id;
+ kinfo->shm_mem.nr_banks++;
+
+ return 0;
+}
+
+static int __init process_shm(struct domain *d, struct kernel_info *kinfo,
const struct dt_device_node *node)
{
struct dt_device_node *shm_node;
@@ -963,6 +978,14 @@ static int __init process_shm(struct domain *d,
if ( ret )
return ret;
}
+
+ /*
+ * Record static shared memory region info for later setting
+ * up shm-node in guest device tree.
+ */
+ ret = append_shm_bank_to_domain(kinfo, gbase, psize, shm_id);
+ if ( ret )
+ return ret;
}
return 0;
@@ -1293,6 +1316,117 @@ static int __init make_memory_node(const struct domain *d,
return res;
}
+#ifdef CONFIG_STATIC_SHM
+static int __init make_shm_memory_node(const struct domain *d,
+ void *fdt,
+ int addrcells, int sizecells,
+ struct meminfo *mem)
+{
+ unsigned long i = 0;
+ int res = 0;
+
+ if ( mem->nr_banks == 0 )
+ return -ENOENT;
+
+ /*
+ * For each shared memory region, a range is exposed under
+ * the /reserved-memory node as a child node. Each range sub-node is
+ * named xen-shmem@<address>.
+ */
+ dt_dprintk("Create xen-shmem node\n");
+
+ for ( ; i < mem->nr_banks; i++ )
+ {
+ uint64_t start = mem->bank[i].start;
+ uint64_t size = mem->bank[i].size;
+ uint8_t shm_id = mem->bank[i].shm_id;
+ /* Placeholder for xen-shmem@ + a 64-bit number + \0 */
+ char buf[27];
+ const char compat[] = "xen,shared-memory-v1";
+ __be32 reg[4];
+ __be32 *cells;
+ unsigned int len = (addrcells + sizecells) * sizeof(__be32);
+
+ snprintf(buf, sizeof(buf), "xen-shmem@%"PRIx64, mem->bank[i].start);
+ res = fdt_begin_node(fdt, buf);
+ if ( res )
+ return res;
+
+ res = fdt_property(fdt, "compatible", compat, sizeof(compat));
+ if ( res )
+ return res;
+
+ cells = reg;
+ dt_child_set_range(&cells, addrcells, sizecells, start, size);
+
+ res = fdt_property(fdt, "reg", reg, len);
+ if ( res )
+ return res;
+
+ dt_dprintk("Shared memory bank %lu: %#"PRIx64"->%#"PRIx64"\n",
+ i, start, start + size);
+
+ res = fdt_property_cell(fdt, "xen,id", shm_id);
+ if ( res )
+ return res;
+
+ res = fdt_end_node(fdt);
+ if ( res )
+ return res;
+ }
+
+ return res;
+}
+#else
+static int __init make_shm_memory_node(const struct domain *d,
+ void *fdt,
+ int addrcells, int sizecells,
+ struct meminfo *mem)
+{
+ ASSERT_UNREACHABLE();
+}
+#endif
+
+static int __init make_resv_memory_node(const struct domain *d,
+ void *fdt,
+ int addrcells, int sizecells,
+ struct meminfo *mem)
+{
+ int res = 0;
+ /* Placeholder for reserved-memory\0 */
+ char resvbuf[16] = "reserved-memory";
+
+ if ( mem->nr_banks == 0 )
+ /* No shared memory provided. */
+ return 0;
+
+ dt_dprintk("Create reserved-memory node\n");
+
+ res = fdt_begin_node(fdt, resvbuf);
+ if ( res )
+ return res;
+
+ res = fdt_property(fdt, "ranges", NULL, 0);
+ if ( res )
+ return res;
+
+ res = fdt_property_cell(fdt, "#address-cells", addrcells);
+ if ( res )
+ return res;
+
+ res = fdt_property_cell(fdt, "#size-cells", sizecells);
+ if ( res )
+ return res;
+
+ res = make_shm_memory_node(d, fdt, addrcells, sizecells, mem);
+ if ( res )
+ return res;
+
+ res = fdt_end_node(fdt);
+
+ return res;
+}
+
static int __init add_ext_regions(unsigned long s, unsigned long e, void *data)
{
struct meminfo *ext_regions = data;
@@ -3029,6 +3163,11 @@ static int __init prepare_dtb_domU(struct domain *d, struct kernel_info *kinfo)
if ( ret )
goto err;
+ ret = make_resv_memory_node(d, kinfo->fdt, addrcells, sizecells,
+ &kinfo->shm_mem);
+ if ( ret )
+ goto err;
+
/*
* domain_handle_dtb_bootmodule has to be called before the rest of
* the device tree is generated because it depends on the value of
@@ -3367,7 +3506,7 @@ static int __init construct_domU(struct domain *d,
assign_static_memory_11(d, &kinfo, node);
#ifdef CONFIG_STATIC_SHM
- rc = process_shm(d, node);
+ rc = process_shm(d, &kinfo, node);
if ( rc < 0 )
return rc;
#endif
diff --git a/xen/arch/arm/include/asm/kernel.h b/xen/arch/arm/include/asm/kernel.h
index 874aa108a7..e089980b87 100644
--- a/xen/arch/arm/include/asm/kernel.h
+++ b/xen/arch/arm/include/asm/kernel.h
@@ -19,6 +19,7 @@ struct kernel_info {
void *fdt; /* flat device tree */
paddr_t unassigned_mem; /* RAM not (yet) assigned to a bank */
struct meminfo mem;
+ struct meminfo shm_mem;
/* kernel entry point */
paddr_t entry;
diff --git a/xen/arch/arm/include/asm/setup.h b/xen/arch/arm/include/asm/setup.h
index 35449bd7d8..6cd45b6212 100644
--- a/xen/arch/arm/include/asm/setup.h
+++ b/xen/arch/arm/include/asm/setup.h
@@ -29,6 +29,7 @@ struct membank {
bool xen_domain; /* whether the memory bank is bound to a Xen domain. */
#ifdef CONFIG_STATIC_SHM
unsigned int nr_shm_domain;
+ uint8_t shm_id ; /* ID identifier of a static shared memory bank. */
#endif
};
--
2.25.1
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH v3 8/8] xen/arm: enable statically shared memory on Dom0
2022-05-12 9:11 [PATCH v3 0/8] static shared memory on dom0less system Penny Zheng
` (6 preceding siblings ...)
2022-05-12 9:11 ` [PATCH v3 7/8] xen/arm: create shared memory nodes in guest device tree Penny Zheng
@ 2022-05-12 9:11 ` Penny Zheng
7 siblings, 0 replies; 18+ messages in thread
From: Penny Zheng @ 2022-05-12 9:11 UTC (permalink / raw)
To: xen-devel
Cc: wei.chen, Penny Zheng, Stefano Stabellini, Julien Grall,
Bertrand Marquis, Volodymyr Babchuk
From: Penny Zheng <penny.zheng@arm.com>
To add statically shared memory nodes in Dom0, user shall put according
static shared memory configuration under /chosen node.
This commit adds shm-processing function process_shm in construct_dom0
to enable statically shared memory on Dom0.
Signed-off-by: Penny Zheng <penny.zheng@arm.com>
Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
---
v3 change:
- no change
---
v2 change:
- no change
---
xen/arch/arm/domain_build.c | 14 ++++++++++++++
1 file changed, 14 insertions(+)
diff --git a/xen/arch/arm/domain_build.c b/xen/arch/arm/domain_build.c
index fb9146b6e0..f4740f9085 100644
--- a/xen/arch/arm/domain_build.c
+++ b/xen/arch/arm/domain_build.c
@@ -2611,6 +2611,11 @@ static int __init handle_node(struct domain *d, struct kernel_info *kinfo,
if ( res )
return res;
}
+
+ res = make_resv_memory_node(d, kinfo->fdt, addrcells, sizecells,
+ &kinfo->shm_mem);
+ if ( res )
+ return res;
}
res = fdt_end_node(kinfo->fdt);
@@ -3622,6 +3627,9 @@ static int __init construct_dom0(struct domain *d)
{
struct kernel_info kinfo = {};
int rc;
+#ifdef CONFIG_STATIC_SHM
+ const struct dt_device_node *chosen = dt_find_node_by_path("/chosen");
+#endif
/* Sanity! */
BUG_ON(d->domain_id != 0);
@@ -3656,6 +3664,12 @@ static int __init construct_dom0(struct domain *d)
allocate_memory_11(d, &kinfo);
find_gnttab_region(d, &kinfo);
+#ifdef CONFIG_STATIC_SHM
+ rc = process_shm(d, &kinfo, chosen);
+ if ( rc < 0 )
+ return rc;
+#endif
+
/* Map extra GIC MMIO, irqs and other hw stuffs to dom0. */
rc = gic_map_hwdom_extra_mappings(d);
if ( rc < 0 )
--
2.25.1
^ permalink raw reply related [flat|nested] 18+ messages in thread
* Re: [PATCH v3 4/8] xen/arm: introduce put_page_nr and get_page_nr
2022-05-12 9:11 ` [PATCH v3 4/8] xen/arm: introduce put_page_nr and get_page_nr Penny Zheng
@ 2022-05-12 10:14 ` Julien Grall
2022-05-13 2:22 ` Penny Zheng
0 siblings, 1 reply; 18+ messages in thread
From: Julien Grall @ 2022-05-12 10:14 UTC (permalink / raw)
To: Penny Zheng, xen-devel
Cc: wei.chen, Stefano Stabellini, Bertrand Marquis, Volodymyr Babchuk
On 12/05/2022 10:11, Penny Zheng wrote:
> Later, we need to add the right amount of references, which should be
> the number of borrower domains, to the owner domain. Since we only have
> get_page() to increment the page reference by 1, a loop is needed per
> page, which is inefficient and time-consuming.
>
> To save the loop time, this commit introduces a set of new helpers
> put_page_nr() and get_page_nr() to increment/drop the page reference by nr.
>
> Signed-off-by: Penny Zheng <penny.zheng@arm.com>
> ---
> v3 changes:
> - check overflow with "n"
> - remove spurious change
> - bring back the check that we enter the loop only when count_info is
> greater than 0
> ---
> v2 change:
> - new commit
> ---
> xen/arch/arm/include/asm/mm.h | 4 ++++
> xen/arch/arm/mm.c | 36 ++++++++++++++++++++++++++---------
> 2 files changed, 31 insertions(+), 9 deletions(-)
>
> diff --git a/xen/arch/arm/include/asm/mm.h b/xen/arch/arm/include/asm/mm.h
> index 424aaf2823..c737d51e4d 100644
> --- a/xen/arch/arm/include/asm/mm.h
> +++ b/xen/arch/arm/include/asm/mm.h
> @@ -347,6 +347,10 @@ void free_init_memory(void);
> int guest_physmap_mark_populate_on_demand(struct domain *d, unsigned long gfn,
> unsigned int order);
>
> +extern bool get_page_nr(struct page_info *page, const struct domain *domain,
> + unsigned long nr);
> +extern void put_page_nr(struct page_info *page, unsigned long nr);
> +
> extern void put_page_type(struct page_info *page);
> static inline void put_page_and_type(struct page_info *page)
> {
> diff --git a/xen/arch/arm/mm.c b/xen/arch/arm/mm.c
> index 7b1f2f4906..a9461e07aa 100644
> --- a/xen/arch/arm/mm.c
> +++ b/xen/arch/arm/mm.c
> @@ -1537,7 +1537,8 @@ long arch_memory_op(int op, XEN_GUEST_HANDLE_PARAM(void) arg)
> return 0;
> }
>
> -struct domain *page_get_owner_and_reference(struct page_info *page)
> +static struct domain *page_get_owner_and_nr_reference(struct page_info *page,
> + unsigned long nr)
> {
> unsigned long x, y = page->count_info;
> struct domain *owner;
> @@ -1548,10 +1549,10 @@ struct domain *page_get_owner_and_reference(struct page_info *page)
> * Count == 0: Page is not allocated, so we cannot take a reference.
> * Count == -1: Reference count would wrap, which is invalid.
> */
> - if ( unlikely(((x + 1) & PGC_count_mask) <= 1) )
> + if ( unlikely(((x + nr) & PGC_count_mask) <= 1) )
This check looks wrong to me. You want to make sure that the right
equation return is at least equal to n otherwise.
Furthermore, I think we need to restrict 'nr' to PGC_count_mask to fully
catch any overflow.
Before the loop, the code would look like:
/* Restrict nr to avoid "double" overflow */
if ( nr >= PGC_count_mask )
{
ASSERT_UNREACHABLE();
return NULL;
}
The check in the loop would look like:
if ( unlikely((x + nr) & PGC_count_mask) <= n )
That said, it might be easier to read the overflow check if we do:
count = x & PGC_count_mask;
if ( !count || ((PGC_count_mask - count) <= n) )
I haven't measured and check which of the two options would result to
better code and performance (get_page() is often called).
> return NULL;
> }
> - while ( (y = cmpxchg(&page->count_info, x, x + 1)) != x );
> + while ( (y = cmpxchg(&page->count_info, x, x + nr)) != x );
>
> owner = page_get_owner(page);
> ASSERT(owner);
> @@ -1559,14 +1560,20 @@ struct domain *page_get_owner_and_reference(struct page_info *page)
> return owner;
> }
>
> -void put_page(struct page_info *page)
> +struct domain *page_get_owner_and_reference(struct page_info *page)
> +{
> + return page_get_owner_and_nr_reference(page, 1);
> +}
> +
> +void put_page_nr(struct page_info *page, unsigned long nr)
> {
> unsigned long nx, x, y = page->count_info;
>
> do {
> - ASSERT((y & PGC_count_mask) != 0);
> + ASSERT(((y & PGC_count_mask) != 0) &&
> + (((y - nr) & PGC_count_mask) >= 0));
I think there are a potential underflow here if 'y' is smaller than
'nr'. But on v2, Stefano suggest to use ASSERT((y & PGC_count_mask) >=
nr); which I think is sufficient here.
> x = y;
> - nx = x - 1;
> + nx = x - nr;
> }
> while ( unlikely((y = cmpxchg(&page->count_info, x, nx)) != x) );
>
> @@ -1576,19 +1583,30 @@ void put_page(struct page_info *page)
> }
> }
>
> -bool get_page(struct page_info *page, const struct domain *domain)
> +void put_page(struct page_info *page)
> {
> - const struct domain *owner = page_get_owner_and_reference(page);
> + put_page_nr(page, 1);
> +}
> +
> +bool get_page_nr(struct page_info *page, const struct domain *domain,
> + unsigned long nr)
> +{
> + const struct domain *owner = page_get_owner_and_nr_reference(page, nr);
>
> if ( likely(owner == domain) )
> return true;
>
> if ( owner != NULL )
> - put_page(page);
> + put_page_nr(page, nr);
>
> return false;
> }
>
> +bool get_page(struct page_info *page, const struct domain *domain)
> +{
> + return get_page_nr(page, domain, 1);
> +}
> +
> /* Common code requires get_page_type and put_page_type.
> * We don't care about typecounts so we just do the minimum to make it
> * happy. */
Cheers,
--
Julien Grall
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [PATCH v3 5/8] xen/arm: Add additional reference to owner domain when the owner is allocated
2022-05-12 9:11 ` [PATCH v3 5/8] xen/arm: Add additional reference to owner domain when the owner is allocated Penny Zheng
@ 2022-05-12 10:54 ` Julien Grall
2022-05-13 2:42 ` Penny Zheng
0 siblings, 1 reply; 18+ messages in thread
From: Julien Grall @ 2022-05-12 10:54 UTC (permalink / raw)
To: Penny Zheng, xen-devel
Cc: wei.chen, Stefano Stabellini, Bertrand Marquis, Volodymyr Babchuk
On 12/05/2022 10:11, Penny Zheng wrote:
> @@ -827,6 +858,37 @@ static int __init allocate_shared_memory(struct domain *d,
> }
> }
>
> + /*
> + * Get the right amount of references per page, which is the number of
> + * borrow domains.
> + */
> + ret = acquire_nr_borrower_domain(d, pbase, psize, &nr_borrowers);
> + if ( ret )
> + return ret;
> +
> + /*
> + * Instead of let borrower domain get a page ref, we add as many
> + * additional reference as the number of borrowers when the owner
> + * is allocated, since there is a chance that owner is created
> + * after borrower.
> + */
> + page = mfn_to_page(smfn);
Where do you check that the range [smfn, smfn + nr_pages] is actual RAM?
If there are none, then you should use mfn_valid() for each mfn to
ensure there will be a struct page for it.
Cheers,
--
Julien Grall
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [PATCH v3 1/8] xen/arm: introduce static shared memory
2022-05-12 9:11 ` [PATCH v3 1/8] xen/arm: introduce static shared memory Penny Zheng
@ 2022-05-12 23:47 ` Stefano Stabellini
0 siblings, 0 replies; 18+ messages in thread
From: Stefano Stabellini @ 2022-05-12 23:47 UTC (permalink / raw)
To: Penny Zheng
Cc: xen-devel, wei.chen, Stefano Stabellini, Julien Grall,
Bertrand Marquis, Volodymyr Babchuk
On Thu, 12 May 2022, Penny Zheng wrote:
> From: Penny Zheng <penny.zheng@arm.com>
>
> This patch serie introduces a new feature: setting up static
> shared memory on a dom0less system, through device tree configuration.
>
> This commit parses shared memory node at boot-time, and reserve it in
> bootinfo.reserved_mem to avoid other use.
>
> This commits proposes a new Kconfig CONFIG_STATIC_SHM to wrap
> static-shm-related codes, and this option depends on static memory(
> CONFIG_STATIC_MEMORY). That's because that later we want to reuse a few
> helpers, guarded with CONFIG_STATIC_MEMORY, like acquire_staticmem_pages, etc,
> on static shared memory.
>
> Signed-off-by: Penny Zheng <penny.zheng@arm.com>
> ---
> v3 change:
> - make nr_shm_domain unsigned int
> ---
> v2 change:
> - document refinement
> - remove bitmap and use the iteration to check
> - add a new field nr_shm_domain to keep the number of shared domain
> ---
> docs/misc/arm/device-tree/booting.txt | 120 ++++++++++++++++++++++++++
> xen/arch/arm/Kconfig | 6 ++
> xen/arch/arm/bootfdt.c | 68 +++++++++++++++
> xen/arch/arm/include/asm/setup.h | 3 +
> 4 files changed, 197 insertions(+)
>
> diff --git a/docs/misc/arm/device-tree/booting.txt b/docs/misc/arm/device-tree/booting.txt
> index a94125394e..e63ce171fc 100644
> --- a/docs/misc/arm/device-tree/booting.txt
> +++ b/docs/misc/arm/device-tree/booting.txt
> @@ -355,3 +355,123 @@ device-tree:
>
> This will reserve a 512MB region starting at the host physical address
> 0x30000000 to be exclusively used by DomU1.
> +
> +Static Shared Memory
> +====================
> +
> +The static shared memory device tree nodes allow users to statically set up
> +shared memory on dom0less system, enabling domains to do shm-based
> +communication.
> +
> +- compatible
> +
> + "xen,domain-shared-memory-v1"
> +
> +- xen,shm-id
> +
> + An u8 value represents the unique identifier of the shared memory region.
> + The maximum identifier shall be "xen,shm-id = <0xff>".
NIT: a 8-bit integer that represent...
Can be fixed on commit.
Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
> +- xen,shared-mem
> +
> + An array takes a physical address, which is the base address of the
> + shared memory region in host physical address space, a size, and a guest
> + physical address, as the target address of the mapping. The number of cells
> + for the host address (and size) is the same as the guest pseudo-physical
> + address and they are inherited from the parent node.
> +
> +- role (Optional)
> +
> + A string property specifying the ownership of a shared memory region,
> + the value must be one of the following: "owner", or "borrower"
> + A shared memory region could be explicitly backed by one domain, which is
> + called "owner domain", and all the other domains who are also sharing
> + this region are called "borrower domain".
> + If not specified, the default value is "borrower" and owner is
> + "dom_shared", a system domain.
> +
> +As an example:
> +
> +chosen {
> + #address-cells = <0x1>;
> + #size-cells = <0x1>;
> + xen,xen-bootargs = "console=dtuart dtuart=serial0 bootscrub=0";
> +
> + ......
> +
> + /* this is for Dom0 */
> + dom0-shared-mem@10000000 {
> + compatible = "xen,domain-shared-memory-v1";
> + role = "owner";
> + xen,shm-id = <0x0>;
> + xen,shared-mem = <0x10000000 0x10000000 0x10000000>;
> + }
> +
> + domU1 {
> + compatible = "xen,domain";
> + #address-cells = <0x1>;
> + #size-cells = <0x1>;
> + memory = <0 131072>;
> + cpus = <2>;
> + vpl011;
> +
> + /*
> + * shared memory region identified as 0x0(xen,shm-id = <0x0>)
> + * is shared between Dom0 and DomU1.
> + */
> + domU1-shared-mem@10000000 {
> + compatible = "xen,domain-shared-memory-v1";
> + role = "borrower";
> + xen,shm-id = <0x0>;
> + xen,shared-mem = <0x10000000 0x10000000 0x50000000>;
> + }
> +
> + /*
> + * shared memory region identified as 0x1(xen,shm-id = <0x1>)
> + * is shared between DomU1 and DomU2.
> + */
> + domU1-shared-mem@50000000 {
> + compatible = "xen,domain-shared-memory-v1";
> + xen,shm-id = <0x1>;
> + xen,shared-mem = <0x50000000 0x20000000 0x60000000>;
> + }
> +
> + ......
> +
> + };
> +
> + domU2 {
> + compatible = "xen,domain";
> + #address-cells = <0x1>;
> + #size-cells = <0x1>;
> + memory = <0 65536>;
> + cpus = <1>;
> +
> + /*
> + * shared memory region identified as 0x1(xen,shm-id = <0x1>)
> + * is shared between domU1 and domU2.
> + */
> + domU2-shared-mem@50000000 {
> + compatible = "xen,domain-shared-memory-v1";
> + xen,shm-id = <0x1>;
> + xen,shared-mem = <0x50000000 0x20000000 0x70000000>;
> + }
> +
> + ......
> + };
> +};
> +
> +This is an example with two static shared memory regions.
> +
> +For the static shared memory region identified as 0x0, host physical
> +address starting at 0x10000000 of 256MB will be reserved to be shared between
> +Dom0 and DomU1. It will get mapped at 0x10000000 in Dom0 guest physical address
> +space, and at 0x50000000 in DomU1 guest physical address space. Dom0 is
> +explicitly defined as the owner domain, and DomU1 is the borrower domain.
> +
> +For the static shared memory region identified as 0x1, host physical
> +address starting at 0x50000000 of 512MB will be reserved to be shared between
> +DomU1 and DomU2. It will get mapped at 0x60000000 in DomU1 guest physical
> +address space, and at 0x70000000 in DomU2 guest physical address space. DomU1
> +and DomU2 are both the borrower domain, the owner domain is the default owner
> +domain dom_shared.
> diff --git a/xen/arch/arm/Kconfig b/xen/arch/arm/Kconfig
> index ecfa6822e4..5ee9921f56 100644
> --- a/xen/arch/arm/Kconfig
> +++ b/xen/arch/arm/Kconfig
> @@ -106,6 +106,12 @@ config TEE
>
> source "arch/arm/tee/Kconfig"
>
> +config STATIC_SHM
> + bool "Statically shared memory on a dom0less system" if UNSUPPORTED
> + depends on STATIC_MEMORY
> + help
> + This option enables statically shared memory on a dom0less system.
> +
> endmenu
>
> menu "ARM errata workaround via the alternative framework"
> diff --git a/xen/arch/arm/bootfdt.c b/xen/arch/arm/bootfdt.c
> index 29671c8df0..c8e8747c3e 100644
> --- a/xen/arch/arm/bootfdt.c
> +++ b/xen/arch/arm/bootfdt.c
> @@ -360,6 +360,70 @@ static int __init process_domain_node(const void *fdt, int node,
> size_cells, &bootinfo.reserved_mem, true);
> }
>
> +#ifdef CONFIG_STATIC_SHM
> +static int __init process_shm_node(const void *fdt, int node,
> + u32 address_cells, u32 size_cells)
> +{
> + const struct fdt_property *prop;
> + const __be32 *cell;
> + paddr_t paddr, size;
> + struct meminfo *mem = &bootinfo.reserved_mem;
> + unsigned long i;
> +
> + if ( address_cells < 1 || size_cells < 1 )
> + {
> + printk("fdt: invalid #address-cells or #size-cells for static shared memory node.\n");
> + return -EINVAL;
> + }
> +
> + prop = fdt_get_property(fdt, node, "xen,shared-mem", NULL);
> + if ( !prop )
> + return -ENOENT;
> +
> + /*
> + * xen,shared-mem = <paddr, size, gaddr>;
> + * Memory region starting from physical address #paddr of #size shall
> + * be mapped to guest physical address #gaddr as static shared memory
> + * region.
> + */
> + cell = (const __be32 *)prop->data;
> + device_tree_get_reg(&cell, address_cells, size_cells, &paddr, &size);
> + for ( i = 0; i < mem->nr_banks; i++ )
> + {
> + /*
> + * A static shared memory region could be shared between multiple
> + * domains.
> + */
> + if ( paddr == mem->bank[i].start && size == mem->bank[i].size )
> + break;
> + }
> +
> + if ( i == mem->nr_banks )
> + {
> + if ( i < NR_MEM_BANKS )
> + {
> + /* Static shared memory shall be reserved from any other use. */
> + mem->bank[mem->nr_banks].start = paddr;
> + mem->bank[mem->nr_banks].size = size;
> + mem->bank[mem->nr_banks].xen_domain = true;
> + mem->nr_banks++;
> + }
> + else
> + {
> + printk("Warning: Max number of supported memory regions reached.\n");
> + return -ENOSPC;
> + }
> + }
> + /*
> + * keep a count of the number of domains, which later may be used to
> + * calculate the number of the reference count.
> + */
> + mem->bank[i].nr_shm_domain++;
> +
> + return 0;
> +}
> +#endif
> +
> static int __init early_scan_node(const void *fdt,
> int node, const char *name, int depth,
> u32 address_cells, u32 size_cells,
> @@ -380,6 +444,10 @@ static int __init early_scan_node(const void *fdt,
> process_chosen_node(fdt, node, name, address_cells, size_cells);
> else if ( depth == 2 && device_tree_node_compatible(fdt, node, "xen,domain") )
> rc = process_domain_node(fdt, node, name, address_cells, size_cells);
> +#ifdef CONFIG_STATIC_SHM
> + else if ( depth <= 3 && device_tree_node_compatible(fdt, node, "xen,domain-shared-memory-v1") )
> + rc = process_shm_node(fdt, node, address_cells, size_cells);
> +#endif
>
> if ( rc < 0 )
> printk("fdt: node `%s': parsing failed\n", name);
> diff --git a/xen/arch/arm/include/asm/setup.h b/xen/arch/arm/include/asm/setup.h
> index 7a1e1d6798..35449bd7d8 100644
> --- a/xen/arch/arm/include/asm/setup.h
> +++ b/xen/arch/arm/include/asm/setup.h
> @@ -27,6 +27,9 @@ struct membank {
> paddr_t start;
> paddr_t size;
> bool xen_domain; /* whether the memory bank is bound to a Xen domain. */
> +#ifdef CONFIG_STATIC_SHM
> + unsigned int nr_shm_domain;
> +#endif
> };
>
> struct meminfo {
> --
> 2.25.1
>
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [PATCH v3 2/8] xen/arm: allocate static shared memory to the default owner dom_io
2022-05-12 9:11 ` [PATCH v3 2/8] xen/arm: allocate static shared memory to the default owner dom_io Penny Zheng
@ 2022-05-12 23:47 ` Stefano Stabellini
0 siblings, 0 replies; 18+ messages in thread
From: Stefano Stabellini @ 2022-05-12 23:47 UTC (permalink / raw)
To: Penny Zheng
Cc: xen-devel, wei.chen, Stefano Stabellini, Julien Grall,
Bertrand Marquis, Volodymyr Babchuk, Andrew Cooper,
George Dunlap, Jan Beulich, Wei Liu
On Thu, 12 May 2022, Penny Zheng wrote:
> From: Penny Zheng <penny.zheng@arm.com>
>
> This commit introduces process_shm to cope with static shared memory in
> domain construction.
>
> DOMID_IO will be the default owner of memory pre-shared among multiple domains
> at boot time, when no explicit owner is specified.
>
> This commit only considers allocating static shared memory to dom_io
> when owner domain is not explicitly defined in device tree, all the left,
> including the "borrower" code path, the "explicit owner" code path, shall
> be introduced later in the following patches.
>
> Signed-off-by: Penny Zheng <penny.zheng@arm.com>
Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
> ---
> v3 change:
> - refine in-code comment
> ---
> v2 change:
> - instead of introducing a new system domain, reuse the existing dom_io
> - make dom_io a non-auto-translated domain, then no need to create P2M
> for it
> - change dom_io definition and make it wider to support static shm here too
> - introduce is_shm_allocated_to_domio to check whether static shm is
> allocated yet, instead of using shm_mask bitmap
> - add in-code comment
> ---
> xen/arch/arm/domain_build.c | 132 +++++++++++++++++++++++++++++++++++-
> xen/common/domain.c | 5 ++
> 2 files changed, 136 insertions(+), 1 deletion(-)
>
> diff --git a/xen/arch/arm/domain_build.c b/xen/arch/arm/domain_build.c
> index 1472ca4972..5b41ec57f0 100644
> --- a/xen/arch/arm/domain_build.c
> +++ b/xen/arch/arm/domain_build.c
> @@ -519,6 +519,10 @@ static bool __init append_static_memory_to_bank(struct domain *d,
> return true;
> }
>
> +/*
> + * If cell is NULL, pbase and psize should hold valid values.
> + * Otherwise, cell will be populated together with pbase and psize.
> + */
> static mfn_t __init acquire_static_memory_bank(struct domain *d,
> const __be32 **cell,
> u32 addr_cells, u32 size_cells,
> @@ -527,7 +531,8 @@ static mfn_t __init acquire_static_memory_bank(struct domain *d,
> mfn_t smfn;
> int res;
>
> - device_tree_get_reg(cell, addr_cells, size_cells, pbase, psize);
> + if ( cell )
> + device_tree_get_reg(cell, addr_cells, size_cells, pbase, psize);
> ASSERT(IS_ALIGNED(*pbase, PAGE_SIZE) && IS_ALIGNED(*psize, PAGE_SIZE));
> if ( PFN_DOWN(*psize) > UINT_MAX )
> {
> @@ -751,6 +756,125 @@ static void __init assign_static_memory_11(struct domain *d,
> panic("Failed to assign requested static memory for direct-map domain %pd.",
> d);
> }
> +
> +#ifdef CONFIG_STATIC_SHM
> +/*
> + * This function checks whether the static shared memory region is
> + * already allocated to dom_io.
> + */
> +static bool __init is_shm_allocated_to_domio(paddr_t pbase)
> +{
> + struct page_info *page;
> +
> + page = maddr_to_page(pbase);
> + ASSERT(page);
> +
> + if ( page_get_owner(page) == NULL )
> + return false;
> +
> + ASSERT(page_get_owner(page) == dom_io);
> + return true;
> +}
> +
> +static mfn_t __init acquire_shared_memory_bank(struct domain *d,
> + u32 addr_cells, u32 size_cells,
> + paddr_t *pbase, paddr_t *psize)
> +{
> + /*
> + * Pages of statically shared memory shall be included
> + * in domain_tot_pages().
> + */
> + d->max_pages += PFN_DOWN(*psize);
> +
> + return acquire_static_memory_bank(d, NULL, addr_cells, size_cells,
> + pbase, psize);
> +
> +}
> +
> +/*
> + * Func allocate_shared_memory is supposed to be only called
> + * from the owner.
> + */
> +static int __init allocate_shared_memory(struct domain *d,
> + u32 addr_cells, u32 size_cells,
> + paddr_t pbase, paddr_t psize)
> +{
> + mfn_t smfn;
> +
> + dprintk(XENLOG_INFO,
> + "Allocate static shared memory BANK %#"PRIpaddr"-%#"PRIpaddr".\n",
> + pbase, pbase + psize);
> +
> + smfn = acquire_shared_memory_bank(d, addr_cells, size_cells, &pbase,
> + &psize);
> + if ( mfn_eq(smfn, INVALID_MFN) )
> + return -EINVAL;
> +
> + /*
> + * DOMID_IO is the domain, like DOMID_XEN, that is not auto-translated.
> + * It sees RAM 1:1 and we do not need to create P2M mapping for it
> + */
> + ASSERT(d == dom_io);
> + return 0;
> +}
> +
> +static int __init process_shm(struct domain *d,
> + const struct dt_device_node *node)
> +{
> + struct dt_device_node *shm_node;
> + int ret = 0;
> + const struct dt_property *prop;
> + const __be32 *cells;
> + u32 shm_id;
> + u32 addr_cells, size_cells;
> + paddr_t gbase, pbase, psize;
> +
> + dt_for_each_child_node(node, shm_node)
> + {
> + if ( !dt_device_is_compatible(shm_node, "xen,domain-shared-memory-v1") )
> + continue;
> +
> + if ( !dt_property_read_u32(shm_node, "xen,shm-id", &shm_id) )
> + {
> + printk("Shared memory node does not provide \"xen,shm-id\" property.\n");
> + return -ENOENT;
> + }
> +
> + addr_cells = dt_n_addr_cells(shm_node);
> + size_cells = dt_n_size_cells(shm_node);
> + prop = dt_find_property(shm_node, "xen,shared-mem", NULL);
> + if ( !prop )
> + {
> + printk("Shared memory node does not provide \"xen,shared-mem\" property.\n");
> + return -ENOENT;
> + }
> + cells = (const __be32 *)prop->value;
> + /* xen,shared-mem = <pbase, psize, gbase>; */
> + device_tree_get_reg(&cells, addr_cells, size_cells, &pbase, &psize);
> + ASSERT(IS_ALIGNED(pbase, PAGE_SIZE) && IS_ALIGNED(psize, PAGE_SIZE));
> + gbase = dt_read_number(cells, addr_cells);
> +
> + /* TODO: Consider owner domain is not the default dom_io. */
> + /*
> + * Per static shared memory region could be shared between multiple
> + * domains.
> + * In case re-allocating the same shared memory region, we check
> + * if it is already allocated to the default owner dom_io before
> + * the actual allocation.
> + */
> + if ( !is_shm_allocated_to_domio(pbase) )
> + {
> + /* Allocate statically shared pages to the default owner dom_io. */
> + ret = allocate_shared_memory(dom_io, addr_cells, size_cells,
> + pbase, psize);
> + if ( ret )
> + return ret;
> + }
> + }
> +
> + return 0;
> +}
> +#endif /* CONFIG_STATIC_SHM */
> #else
> static void __init allocate_static_memory(struct domain *d,
> struct kernel_info *kinfo,
> @@ -3149,6 +3273,12 @@ static int __init construct_domU(struct domain *d,
> else
> assign_static_memory_11(d, &kinfo, node);
>
> +#ifdef CONFIG_STATIC_SHM
> + rc = process_shm(d, node);
> + if ( rc < 0 )
> + return rc;
> +#endif
> +
> /*
> * Base address and irq number are needed when creating vpl011 device
> * tree node in prepare_dtb_domU, so initialization on related variables
> diff --git a/xen/common/domain.c b/xen/common/domain.c
> index 8d2c2a9897..43237c287a 100644
> --- a/xen/common/domain.c
> +++ b/xen/common/domain.c
> @@ -780,6 +780,11 @@ void __init setup_system_domains(void)
> * This domain owns I/O pages that are within the range of the page_info
> * array. Mappings occur at the priv of the caller.
> * Quarantined PCI devices will be associated with this domain.
> + *
> + * DOMID_IO could also be used for mapping memory when no explicit
> + * domain is specified.
> + * For instance, DOMID_IO is the owner of memory pre-shared among
> + * multiple domains at boot time, when no explicit owner is specified.
> */
> dom_io = domain_create(DOMID_IO, NULL, 0);
> if ( IS_ERR(dom_io) )
> --
> 2.25.1
>
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [PATCH v3 3/8] xen/arm: allocate static shared memory to a specific owner domain
2022-05-12 9:11 ` [PATCH v3 3/8] xen/arm: allocate static shared memory to a specific owner domain Penny Zheng
@ 2022-05-12 23:47 ` Stefano Stabellini
0 siblings, 0 replies; 18+ messages in thread
From: Stefano Stabellini @ 2022-05-12 23:47 UTC (permalink / raw)
To: Penny Zheng
Cc: xen-devel, wei.chen, Stefano Stabellini, Julien Grall,
Bertrand Marquis, Volodymyr Babchuk
On Thu, 12 May 2022, Penny Zheng wrote:
> If owner property is defined, then owner domain of a static shared memory
> region is not the default dom_io anymore, but a specific domain.
>
> This commit implements allocating static shared memory to a specific domain
> when owner property is defined.
>
> Coding flow for dealing borrower domain will be introduced later in the
> following commits.
>
> Signed-off-by: Penny Zheng <penny.zheng@arm.com>
Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
> ---
> v3 change:
> - simplify the code since o_gbase is not used if the domain is dom_io
> ---
> v2 change:
> - P2M mapping is restricted to normal domain
> - in-code comment fix
> ---
> xen/arch/arm/domain_build.c | 44 +++++++++++++++++++++++++++----------
> 1 file changed, 33 insertions(+), 11 deletions(-)
>
> diff --git a/xen/arch/arm/domain_build.c b/xen/arch/arm/domain_build.c
> index 5b41ec57f0..9c53b9fa7b 100644
> --- a/xen/arch/arm/domain_build.c
> +++ b/xen/arch/arm/domain_build.c
> @@ -797,9 +797,11 @@ static mfn_t __init acquire_shared_memory_bank(struct domain *d,
> */
> static int __init allocate_shared_memory(struct domain *d,
> u32 addr_cells, u32 size_cells,
> - paddr_t pbase, paddr_t psize)
> + paddr_t pbase, paddr_t psize,
> + paddr_t gbase)
> {
> mfn_t smfn;
> + int ret = 0;
>
> dprintk(XENLOG_INFO,
> "Allocate static shared memory BANK %#"PRIpaddr"-%#"PRIpaddr".\n",
> @@ -814,8 +816,18 @@ static int __init allocate_shared_memory(struct domain *d,
> * DOMID_IO is the domain, like DOMID_XEN, that is not auto-translated.
> * It sees RAM 1:1 and we do not need to create P2M mapping for it
> */
> - ASSERT(d == dom_io);
> - return 0;
> + if ( d != dom_io )
> + {
> + ret = guest_physmap_add_pages(d, gaddr_to_gfn(gbase), smfn, PFN_DOWN(psize));
> + if ( ret )
> + {
> + printk(XENLOG_ERR
> + "Failed to map shared memory to %pd.\n", d);
> + return ret;
> + }
> + }
> +
> + return ret;
> }
>
> static int __init process_shm(struct domain *d,
> @@ -828,6 +840,8 @@ static int __init process_shm(struct domain *d,
> u32 shm_id;
> u32 addr_cells, size_cells;
> paddr_t gbase, pbase, psize;
> + const char *role_str;
> + bool owner_dom_io = true;
>
> dt_for_each_child_node(node, shm_node)
> {
> @@ -854,19 +868,27 @@ static int __init process_shm(struct domain *d,
> ASSERT(IS_ALIGNED(pbase, PAGE_SIZE) && IS_ALIGNED(psize, PAGE_SIZE));
> gbase = dt_read_number(cells, addr_cells);
>
> - /* TODO: Consider owner domain is not the default dom_io. */
> + /*
> + * "role" property is optional and if it is defined explicitly,
> + * then the owner domain is not the default "dom_io" domain.
> + */
> + if ( dt_property_read_string(shm_node, "role", &role_str) == 0 )
> + owner_dom_io = false;
> +
> /*
> * Per static shared memory region could be shared between multiple
> * domains.
> - * In case re-allocating the same shared memory region, we check
> - * if it is already allocated to the default owner dom_io before
> - * the actual allocation.
> + * So when owner domain is the default dom_io, in case re-allocating
> + * the same shared memory region, we check if it is already allocated
> + * to the default owner dom_io before the actual allocation.
> */
> - if ( !is_shm_allocated_to_domio(pbase) )
> + if ( (owner_dom_io && !is_shm_allocated_to_domio(pbase)) ||
> + (!owner_dom_io && strcmp(role_str, "owner") == 0) )
> {
> - /* Allocate statically shared pages to the default owner dom_io. */
> - ret = allocate_shared_memory(dom_io, addr_cells, size_cells,
> - pbase, psize);
> + /* Allocate statically shared pages to the owner domain. */
> + ret = allocate_shared_memory(owner_dom_io ? dom_io : d,
> + addr_cells, size_cells,
> + pbase, psize, gbase);
> if ( ret )
> return ret;
> }
> --
> 2.25.1
>
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [PATCH v3 7/8] xen/arm: create shared memory nodes in guest device tree
2022-05-12 9:11 ` [PATCH v3 7/8] xen/arm: create shared memory nodes in guest device tree Penny Zheng
@ 2022-05-12 23:47 ` Stefano Stabellini
0 siblings, 0 replies; 18+ messages in thread
From: Stefano Stabellini @ 2022-05-12 23:47 UTC (permalink / raw)
To: Penny Zheng
Cc: xen-devel, wei.chen, Stefano Stabellini, Julien Grall,
Bertrand Marquis, Volodymyr Babchuk
On Thu, 12 May 2022, Penny Zheng wrote:
> We expose the shared memory to the domU using the "xen,shared-memory-v1"
> reserved-memory binding. See
> Documentation/devicetree/bindings/reserved-memory/xen,shared-memory.txt
> in Linux for the corresponding device tree binding.
>
> To save the cost of re-parsing shared memory device tree configuration when
> creating shared memory nodes in guest device tree, this commit adds new field
> "shm_mem" to store shm-info per domain.
>
> For each shared memory region, a range is exposed under
> the /reserved-memory node as a child node. Each range sub-node is
> named xen-shmem@<address> and has the following properties:
> - compatible:
> compatible = "xen,shared-memory-v1"
> - reg:
> the base guest physical address and size of the shared memory region
> - xen,id:
> a string that identifies the shared memory region.
>
> Signed-off-by: Penny Zheng <penny.zheng@arm.com>
Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
> ---
> v3 change:
> - move field "shm_mem" to kernel_info
> ---
> v2 change:
> - using xzalloc
> - shm_id should be uint8_t
> - make reg a local variable
> - add #address-cells and #size-cells properties
> - fix alignment
> ---
> xen/arch/arm/domain_build.c | 143 +++++++++++++++++++++++++++++-
> xen/arch/arm/include/asm/kernel.h | 1 +
> xen/arch/arm/include/asm/setup.h | 1 +
> 3 files changed, 143 insertions(+), 2 deletions(-)
>
> diff --git a/xen/arch/arm/domain_build.c b/xen/arch/arm/domain_build.c
> index aa7c264e23..fb9146b6e0 100644
> --- a/xen/arch/arm/domain_build.c
> +++ b/xen/arch/arm/domain_build.c
> @@ -892,7 +892,22 @@ static int __init allocate_shared_memory(struct domain *d,
> return ret;
> }
>
> -static int __init process_shm(struct domain *d,
> +static int __init append_shm_bank_to_domain(struct kernel_info *kinfo,
> + paddr_t start, paddr_t size,
> + u32 shm_id)
> +{
> + if ( (kinfo->shm_mem.nr_banks + 1) > NR_MEM_BANKS )
> + return -ENOMEM;
> +
> + kinfo->shm_mem.bank[kinfo->shm_mem.nr_banks].start = start;
> + kinfo->shm_mem.bank[kinfo->shm_mem.nr_banks].size = size;
> + kinfo->shm_mem.bank[kinfo->shm_mem.nr_banks].shm_id = shm_id;
> + kinfo->shm_mem.nr_banks++;
> +
> + return 0;
> +}
> +
> +static int __init process_shm(struct domain *d, struct kernel_info *kinfo,
> const struct dt_device_node *node)
> {
> struct dt_device_node *shm_node;
> @@ -963,6 +978,14 @@ static int __init process_shm(struct domain *d,
> if ( ret )
> return ret;
> }
> +
> + /*
> + * Record static shared memory region info for later setting
> + * up shm-node in guest device tree.
> + */
> + ret = append_shm_bank_to_domain(kinfo, gbase, psize, shm_id);
> + if ( ret )
> + return ret;
> }
>
> return 0;
> @@ -1293,6 +1316,117 @@ static int __init make_memory_node(const struct domain *d,
> return res;
> }
>
> +#ifdef CONFIG_STATIC_SHM
> +static int __init make_shm_memory_node(const struct domain *d,
> + void *fdt,
> + int addrcells, int sizecells,
> + struct meminfo *mem)
> +{
> + unsigned long i = 0;
> + int res = 0;
> +
> + if ( mem->nr_banks == 0 )
> + return -ENOENT;
> +
> + /*
> + * For each shared memory region, a range is exposed under
> + * the /reserved-memory node as a child node. Each range sub-node is
> + * named xen-shmem@<address>.
> + */
> + dt_dprintk("Create xen-shmem node\n");
> +
> + for ( ; i < mem->nr_banks; i++ )
> + {
> + uint64_t start = mem->bank[i].start;
> + uint64_t size = mem->bank[i].size;
> + uint8_t shm_id = mem->bank[i].shm_id;
> + /* Placeholder for xen-shmem@ + a 64-bit number + \0 */
> + char buf[27];
> + const char compat[] = "xen,shared-memory-v1";
> + __be32 reg[4];
> + __be32 *cells;
> + unsigned int len = (addrcells + sizecells) * sizeof(__be32);
> +
> + snprintf(buf, sizeof(buf), "xen-shmem@%"PRIx64, mem->bank[i].start);
> + res = fdt_begin_node(fdt, buf);
> + if ( res )
> + return res;
> +
> + res = fdt_property(fdt, "compatible", compat, sizeof(compat));
> + if ( res )
> + return res;
> +
> + cells = reg;
> + dt_child_set_range(&cells, addrcells, sizecells, start, size);
> +
> + res = fdt_property(fdt, "reg", reg, len);
> + if ( res )
> + return res;
> +
> + dt_dprintk("Shared memory bank %lu: %#"PRIx64"->%#"PRIx64"\n",
> + i, start, start + size);
> +
> + res = fdt_property_cell(fdt, "xen,id", shm_id);
> + if ( res )
> + return res;
> +
> + res = fdt_end_node(fdt);
> + if ( res )
> + return res;
> + }
> +
> + return res;
> +}
> +#else
> +static int __init make_shm_memory_node(const struct domain *d,
> + void *fdt,
> + int addrcells, int sizecells,
> + struct meminfo *mem)
> +{
> + ASSERT_UNREACHABLE();
> +}
> +#endif
> +
> +static int __init make_resv_memory_node(const struct domain *d,
> + void *fdt,
> + int addrcells, int sizecells,
> + struct meminfo *mem)
> +{
> + int res = 0;
> + /* Placeholder for reserved-memory\0 */
> + char resvbuf[16] = "reserved-memory";
> +
> + if ( mem->nr_banks == 0 )
> + /* No shared memory provided. */
> + return 0;
> +
> + dt_dprintk("Create reserved-memory node\n");
> +
> + res = fdt_begin_node(fdt, resvbuf);
> + if ( res )
> + return res;
> +
> + res = fdt_property(fdt, "ranges", NULL, 0);
> + if ( res )
> + return res;
> +
> + res = fdt_property_cell(fdt, "#address-cells", addrcells);
> + if ( res )
> + return res;
> +
> + res = fdt_property_cell(fdt, "#size-cells", sizecells);
> + if ( res )
> + return res;
> +
> + res = make_shm_memory_node(d, fdt, addrcells, sizecells, mem);
> + if ( res )
> + return res;
> +
> + res = fdt_end_node(fdt);
> +
> + return res;
> +}
> +
> static int __init add_ext_regions(unsigned long s, unsigned long e, void *data)
> {
> struct meminfo *ext_regions = data;
> @@ -3029,6 +3163,11 @@ static int __init prepare_dtb_domU(struct domain *d, struct kernel_info *kinfo)
> if ( ret )
> goto err;
>
> + ret = make_resv_memory_node(d, kinfo->fdt, addrcells, sizecells,
> + &kinfo->shm_mem);
> + if ( ret )
> + goto err;
> +
> /*
> * domain_handle_dtb_bootmodule has to be called before the rest of
> * the device tree is generated because it depends on the value of
> @@ -3367,7 +3506,7 @@ static int __init construct_domU(struct domain *d,
> assign_static_memory_11(d, &kinfo, node);
>
> #ifdef CONFIG_STATIC_SHM
> - rc = process_shm(d, node);
> + rc = process_shm(d, &kinfo, node);
> if ( rc < 0 )
> return rc;
> #endif
> diff --git a/xen/arch/arm/include/asm/kernel.h b/xen/arch/arm/include/asm/kernel.h
> index 874aa108a7..e089980b87 100644
> --- a/xen/arch/arm/include/asm/kernel.h
> +++ b/xen/arch/arm/include/asm/kernel.h
> @@ -19,6 +19,7 @@ struct kernel_info {
> void *fdt; /* flat device tree */
> paddr_t unassigned_mem; /* RAM not (yet) assigned to a bank */
> struct meminfo mem;
> + struct meminfo shm_mem;
>
> /* kernel entry point */
> paddr_t entry;
> diff --git a/xen/arch/arm/include/asm/setup.h b/xen/arch/arm/include/asm/setup.h
> index 35449bd7d8..6cd45b6212 100644
> --- a/xen/arch/arm/include/asm/setup.h
> +++ b/xen/arch/arm/include/asm/setup.h
> @@ -29,6 +29,7 @@ struct membank {
> bool xen_domain; /* whether the memory bank is bound to a Xen domain. */
> #ifdef CONFIG_STATIC_SHM
> unsigned int nr_shm_domain;
> + uint8_t shm_id ; /* ID identifier of a static shared memory bank. */
> #endif
> };
>
> --
> 2.25.1
>
^ permalink raw reply [flat|nested] 18+ messages in thread
* RE: [PATCH v3 4/8] xen/arm: introduce put_page_nr and get_page_nr
2022-05-12 10:14 ` Julien Grall
@ 2022-05-13 2:22 ` Penny Zheng
2022-05-13 9:20 ` Julien Grall
0 siblings, 1 reply; 18+ messages in thread
From: Penny Zheng @ 2022-05-13 2:22 UTC (permalink / raw)
To: Julien Grall, xen-devel
Cc: Wei Chen, Stefano Stabellini, Bertrand Marquis, Volodymyr Babchuk
Hi Julien
> -----Original Message-----
> From: Julien Grall <julien@xen.org>
> Sent: Thursday, May 12, 2022 6:14 PM
> To: Penny Zheng <Penny.Zheng@arm.com>; xen-devel@lists.xenproject.org
> Cc: Wei Chen <Wei.Chen@arm.com>; Stefano Stabellini
> <sstabellini@kernel.org>; Bertrand Marquis <Bertrand.Marquis@arm.com>;
> Volodymyr Babchuk <Volodymyr_Babchuk@epam.com>
> Subject: Re: [PATCH v3 4/8] xen/arm: introduce put_page_nr and get_page_nr
>
>
>
> On 12/05/2022 10:11, Penny Zheng wrote:
> > Later, we need to add the right amount of references, which should be
> > the number of borrower domains, to the owner domain. Since we only
> > have
> > get_page() to increment the page reference by 1, a loop is needed per
> > page, which is inefficient and time-consuming.
> >
> > To save the loop time, this commit introduces a set of new helpers
> > put_page_nr() and get_page_nr() to increment/drop the page reference by
> nr.
> >
> > Signed-off-by: Penny Zheng <penny.zheng@arm.com>
> > ---
> > v3 changes:
> > - check overflow with "n"
> > - remove spurious change
> > - bring back the check that we enter the loop only when count_info is
> > greater than 0
> > ---
> > v2 change:
> > - new commit
> > ---
> > xen/arch/arm/include/asm/mm.h | 4 ++++
> > xen/arch/arm/mm.c | 36 ++++++++++++++++++++++++++---------
> > 2 files changed, 31 insertions(+), 9 deletions(-)
> >
> > diff --git a/xen/arch/arm/include/asm/mm.h
> > b/xen/arch/arm/include/asm/mm.h index 424aaf2823..c737d51e4d 100644
> > --- a/xen/arch/arm/include/asm/mm.h
> > +++ b/xen/arch/arm/include/asm/mm.h
> > @@ -347,6 +347,10 @@ void free_init_memory(void);
> > int guest_physmap_mark_populate_on_demand(struct domain *d,
> unsigned long gfn,
> > unsigned int order);
> >
> > +extern bool get_page_nr(struct page_info *page, const struct domain
> *domain,
> > + unsigned long nr); extern void
> > +put_page_nr(struct page_info *page, unsigned long nr);
> > +
> > extern void put_page_type(struct page_info *page);
> > static inline void put_page_and_type(struct page_info *page)
> > {
> > diff --git a/xen/arch/arm/mm.c b/xen/arch/arm/mm.c index
> > 7b1f2f4906..a9461e07aa 100644
> > --- a/xen/arch/arm/mm.c
> > +++ b/xen/arch/arm/mm.c
> > @@ -1537,7 +1537,8 @@ long arch_memory_op(int op,
> XEN_GUEST_HANDLE_PARAM(void) arg)
> > return 0;
> > }
> >
> > -struct domain *page_get_owner_and_reference(struct page_info *page)
> > +static struct domain *page_get_owner_and_nr_reference(struct page_info
> *page,
> > + unsigned long
> > +nr)
> > {
> > unsigned long x, y = page->count_info;
> > struct domain *owner;
> > @@ -1548,10 +1549,10 @@ struct domain
> *page_get_owner_and_reference(struct page_info *page)
> > * Count == 0: Page is not allocated, so we cannot take a reference.
> > * Count == -1: Reference count would wrap, which is invalid.
> > */
> > - if ( unlikely(((x + 1) & PGC_count_mask) <= 1) )
> > + if ( unlikely(((x + nr) & PGC_count_mask) <= 1) )
>
> This check looks wrong to me. You want to make sure that the right equation
> return is at least equal to n otherwise.
>
Right, right, I haven't considered thoroughly! A thousand thanks for the following
detailed explanation~
> Furthermore, I think we need to restrict 'nr' to PGC_count_mask to fully catch
> any overflow.
>
> Before the loop, the code would look like:
>
> /* Restrict nr to avoid "double" overflow */ if ( nr >= PGC_count_mask ) {
> ASSERT_UNREACHABLE();
> return NULL;
> }
>
> The check in the loop would look like:
>
> if ( unlikely((x + nr) & PGC_count_mask) <= n )
>
> That said, it might be easier to read the overflow check if we do:
>
> count = x & PGC_count_mask;
> if ( !count || ((PGC_count_mask - count) <= n) )
>
> I haven't measured and check which of the two options would result to better
> code and performance (get_page() is often called).
>
Correct me if I understand wrongly:
IMO, only option two is actually catching any overflow? Let (PGC_count_mask - count) <= nr
stay in the loop, not before the loop like option 1, to cover the changeable page->count_info.
> > return NULL;
> > }
> > - while ( (y = cmpxchg(&page->count_info, x, x + 1)) != x );
> > + while ( (y = cmpxchg(&page->count_info, x, x + nr)) != x );
> >
> > owner = page_get_owner(page);
> > ASSERT(owner);
> > @@ -1559,14 +1560,20 @@ struct domain
> *page_get_owner_and_reference(struct page_info *page)
> > return owner;
> > }
> >
> > -void put_page(struct page_info *page)
> > +struct domain *page_get_owner_and_reference(struct page_info *page) {
> > + return page_get_owner_and_nr_reference(page, 1); }
> > +
> > +void put_page_nr(struct page_info *page, unsigned long nr)
> > {
> > unsigned long nx, x, y = page->count_info;
> >
> > do {
> > - ASSERT((y & PGC_count_mask) != 0);
> > + ASSERT(((y & PGC_count_mask) != 0) &&
> > + (((y - nr) & PGC_count_mask) >= 0));
>
> I think there are a potential underflow here if 'y' is smaller than 'nr'. But on v2,
> Stefano suggest to use ASSERT((y & PGC_count_mask) >= nr); which I think is
> sufficient here.
>
Oh, understood. thanks for clarification~
> > x = y;
> > - nx = x - 1;
> > + nx = x - nr;
> > }
> > while ( unlikely((y = cmpxchg(&page->count_info, x, nx)) != x)
> > );
> >
> > @@ -1576,19 +1583,30 @@ void put_page(struct page_info *page)
> > }
> > }
> >
> > -bool get_page(struct page_info *page, const struct domain *domain)
> > +void put_page(struct page_info *page)
> > {
> > - const struct domain *owner = page_get_owner_and_reference(page);
> > + put_page_nr(page, 1);
> > +}
> > +
> > +bool get_page_nr(struct page_info *page, const struct domain *domain,
> > + unsigned long nr)
> > +{
> > + const struct domain *owner =
> > +page_get_owner_and_nr_reference(page, nr);
> >
> > if ( likely(owner == domain) )
> > return true;
> >
> > if ( owner != NULL )
> > - put_page(page);
> > + put_page_nr(page, nr);
> >
> > return false;
> > }
> >
> > +bool get_page(struct page_info *page, const struct domain *domain) {
> > + return get_page_nr(page, domain, 1); }
> > +
> > /* Common code requires get_page_type and put_page_type.
> > * We don't care about typecounts so we just do the minimum to make it
> > * happy. */
>
> Cheers,
>
> --
> Julien Grall
^ permalink raw reply [flat|nested] 18+ messages in thread
* RE: [PATCH v3 5/8] xen/arm: Add additional reference to owner domain when the owner is allocated
2022-05-12 10:54 ` Julien Grall
@ 2022-05-13 2:42 ` Penny Zheng
0 siblings, 0 replies; 18+ messages in thread
From: Penny Zheng @ 2022-05-13 2:42 UTC (permalink / raw)
To: Julien Grall, xen-devel
Cc: Wei Chen, Stefano Stabellini, Bertrand Marquis, Volodymyr Babchuk
Hi Julien
> -----Original Message-----
> From: Julien Grall <julien@xen.org>
> Sent: Thursday, May 12, 2022 6:54 PM
> To: Penny Zheng <Penny.Zheng@arm.com>; xen-devel@lists.xenproject.org
> Cc: Wei Chen <Wei.Chen@arm.com>; Stefano Stabellini
> <sstabellini@kernel.org>; Bertrand Marquis <Bertrand.Marquis@arm.com>;
> Volodymyr Babchuk <Volodymyr_Babchuk@epam.com>
> Subject: Re: [PATCH v3 5/8] xen/arm: Add additional reference to owner
> domain when the owner is allocated
>
>
>
> On 12/05/2022 10:11, Penny Zheng wrote:
> > @@ -827,6 +858,37 @@ static int __init allocate_shared_memory(struct
> domain *d,
> > }
> > }
> >
> > + /*
> > + * Get the right amount of references per page, which is the number of
> > + * borrow domains.
> > + */
> > + ret = acquire_nr_borrower_domain(d, pbase, psize, &nr_borrowers);
> > + if ( ret )
> > + return ret;
> > +
> > + /*
> > + * Instead of let borrower domain get a page ref, we add as many
> > + * additional reference as the number of borrowers when the owner
> > + * is allocated, since there is a chance that owner is created
> > + * after borrower.
> > + */
> > + page = mfn_to_page(smfn);
> Where do you check that the range [smfn, smfn + nr_pages] is actual RAM?
> If there are none, then you should use mfn_valid() for each mfn to ensure
> there will be a struct page for it.
>
Actually, before this, in acquire_shared_memory_bank, we will finally call into
acquire_staticmem_pages to acquire [smfn, smfn + nr_pages], and in there, we
will do the check. See xen/common/page_alloc.c:acquire_staticmem_pages:2676
> Cheers,
>
> --
> Julien Grall
^ permalink raw reply [flat|nested] 18+ messages in thread
* Re: [PATCH v3 4/8] xen/arm: introduce put_page_nr and get_page_nr
2022-05-13 2:22 ` Penny Zheng
@ 2022-05-13 9:20 ` Julien Grall
0 siblings, 0 replies; 18+ messages in thread
From: Julien Grall @ 2022-05-13 9:20 UTC (permalink / raw)
To: Penny Zheng, xen-devel
Cc: Wei Chen, Stefano Stabellini, Bertrand Marquis, Volodymyr Babchuk
On 13/05/2022 03:22, Penny Zheng wrote:
> Hi Julien
Hi Penny,
>> -----Original Message-----
>> From: Julien Grall <julien@xen.org>
>> Sent: Thursday, May 12, 2022 6:14 PM
>> To: Penny Zheng <Penny.Zheng@arm.com>; xen-devel@lists.xenproject.org
>> Cc: Wei Chen <Wei.Chen@arm.com>; Stefano Stabellini
>> <sstabellini@kernel.org>; Bertrand Marquis <Bertrand.Marquis@arm.com>;
>> Volodymyr Babchuk <Volodymyr_Babchuk@epam.com>
>> Subject: Re: [PATCH v3 4/8] xen/arm: introduce put_page_nr and get_page_nr
>>
>>
>>
>> On 12/05/2022 10:11, Penny Zheng wrote:
>>> Later, we need to add the right amount of references, which should be
>>> the number of borrower domains, to the owner domain. Since we only
>>> have
>>> get_page() to increment the page reference by 1, a loop is needed per
>>> page, which is inefficient and time-consuming.
>>>
>>> To save the loop time, this commit introduces a set of new helpers
>>> put_page_nr() and get_page_nr() to increment/drop the page reference by
>> nr.
>>>
>>> Signed-off-by: Penny Zheng <penny.zheng@arm.com>
>>> ---
>>> v3 changes:
>>> - check overflow with "n"
>>> - remove spurious change
>>> - bring back the check that we enter the loop only when count_info is
>>> greater than 0
>>> ---
>>> v2 change:
>>> - new commit
>>> ---
>>> xen/arch/arm/include/asm/mm.h | 4 ++++
>>> xen/arch/arm/mm.c | 36 ++++++++++++++++++++++++++---------
>>> 2 files changed, 31 insertions(+), 9 deletions(-)
>>>
>>> diff --git a/xen/arch/arm/include/asm/mm.h
>>> b/xen/arch/arm/include/asm/mm.h index 424aaf2823..c737d51e4d 100644
>>> --- a/xen/arch/arm/include/asm/mm.h
>>> +++ b/xen/arch/arm/include/asm/mm.h
>>> @@ -347,6 +347,10 @@ void free_init_memory(void);
>>> int guest_physmap_mark_populate_on_demand(struct domain *d,
>> unsigned long gfn,
>>> unsigned int order);
>>>
>>> +extern bool get_page_nr(struct page_info *page, const struct domain
>> *domain,
>>> + unsigned long nr); extern void
>>> +put_page_nr(struct page_info *page, unsigned long nr);
>>> +
>>> extern void put_page_type(struct page_info *page);
>>> static inline void put_page_and_type(struct page_info *page)
>>> {
>>> diff --git a/xen/arch/arm/mm.c b/xen/arch/arm/mm.c index
>>> 7b1f2f4906..a9461e07aa 100644
>>> --- a/xen/arch/arm/mm.c
>>> +++ b/xen/arch/arm/mm.c
>>> @@ -1537,7 +1537,8 @@ long arch_memory_op(int op,
>> XEN_GUEST_HANDLE_PARAM(void) arg)
>>> return 0;
>>> }
>>>
>>> -struct domain *page_get_owner_and_reference(struct page_info *page)
>>> +static struct domain *page_get_owner_and_nr_reference(struct page_info
>> *page,
>>> + unsigned long
>>> +nr)
>>> {
>>> unsigned long x, y = page->count_info;
>>> struct domain *owner;
>>> @@ -1548,10 +1549,10 @@ struct domain
>> *page_get_owner_and_reference(struct page_info *page)
>>> * Count == 0: Page is not allocated, so we cannot take a reference.
>>> * Count == -1: Reference count would wrap, which is invalid.
>>> */
>>> - if ( unlikely(((x + 1) & PGC_count_mask) <= 1) )
>>> + if ( unlikely(((x + nr) & PGC_count_mask) <= 1) )
>>
>> This check looks wrong to me. You want to make sure that the right equation
>> return is at least equal to n otherwise.
>>
>
> Right, right, I haven't considered thoroughly! A thousand thanks for the following
> detailed explanation~
>
>> Furthermore, I think we need to restrict 'nr' to PGC_count_mask to fully catch
>> any overflow.
>>
>> Before the loop, the code would look like:
>>
>> /* Restrict nr to avoid "double" overflow */ if ( nr >= PGC_count_mask ) {
>> ASSERT_UNREACHABLE();
>> return NULL;
>> }
>>
>> The check in the loop would look like:
>>
>> if ( unlikely((x + nr) & PGC_count_mask) <= n )
>>
>> That said, it might be easier to read the overflow check if we do:
>>
>> count = x & PGC_count_mask;
>> if ( !count || ((PGC_count_mask - count) <= n) )
>>
>> I haven't measured and check which of the two options would result to better
>> code and performance (get_page() is often called).
>>
>
> Correct me if I understand wrongly:
> IMO, only option two is actually catching any overflow? Let (PGC_count_mask - count) <= nr
> stay in the loop, not before the loop like option 1, to cover the changeable page->count_info.
Both option should catch the overflow. In option 1, this was 2 part check:
if ( nr >= PGC_count_mask )
...
do
{
if ( unlikely((x + nr) & PGC_count_mask <= n )
return NULL;
...
} while (...);
Cheers,
--
Julien Grall
^ permalink raw reply [flat|nested] 18+ messages in thread
end of thread, other threads:[~2022-05-13 9:20 UTC | newest]
Thread overview: 18+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-05-12 9:11 [PATCH v3 0/8] static shared memory on dom0less system Penny Zheng
2022-05-12 9:11 ` [PATCH v3 1/8] xen/arm: introduce static shared memory Penny Zheng
2022-05-12 23:47 ` Stefano Stabellini
2022-05-12 9:11 ` [PATCH v3 2/8] xen/arm: allocate static shared memory to the default owner dom_io Penny Zheng
2022-05-12 23:47 ` Stefano Stabellini
2022-05-12 9:11 ` [PATCH v3 3/8] xen/arm: allocate static shared memory to a specific owner domain Penny Zheng
2022-05-12 23:47 ` Stefano Stabellini
2022-05-12 9:11 ` [PATCH v3 4/8] xen/arm: introduce put_page_nr and get_page_nr Penny Zheng
2022-05-12 10:14 ` Julien Grall
2022-05-13 2:22 ` Penny Zheng
2022-05-13 9:20 ` Julien Grall
2022-05-12 9:11 ` [PATCH v3 5/8] xen/arm: Add additional reference to owner domain when the owner is allocated Penny Zheng
2022-05-12 10:54 ` Julien Grall
2022-05-13 2:42 ` Penny Zheng
2022-05-12 9:11 ` [PATCH v3 6/8] xen/arm: set up shared memory foreign mapping for borrower domain Penny Zheng
2022-05-12 9:11 ` [PATCH v3 7/8] xen/arm: create shared memory nodes in guest device tree Penny Zheng
2022-05-12 23:47 ` Stefano Stabellini
2022-05-12 9:11 ` [PATCH v3 8/8] xen/arm: enable statically shared memory on Dom0 Penny Zheng
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.