Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1364232 > unrolled thread

[PATCH v7 00/17] Add ACPI support for Xen Dom0 on ARM64

Back to article view | Back to linux.kernel

Contents

  [PATCH v7 00/17] Add ACPI support for Xen Dom0 on ARM64 Shannon Zhao <shannon.zhao@linaro.org> - 2016-03-24 15:50 +0100
    [PATCH v7 08/17] Xen: public/hvm: sync changes of HVM_PARAM_CALLBACK_VIA ABI from Xen Shannon Zhao <shannon.zhao@linaro.org> - 2016-03-24 16:00 +0100
    [PATCH v7 06/17] Xen: ARM: Add support for mapping platform device mmio Shannon Zhao <shannon.zhao@linaro.org> - 2016-03-24 16:00 +0100
    [PATCH v7 02/17] xen/grant-table: Move xlated_setup_gnttab_pages to common place Shannon Zhao <shannon.zhao@linaro.org> - 2016-03-24 16:00 +0100
    [PATCH v7 10/17] arm/xen: Get event-channel irq through HVM_PARAM when booting with ACPI Shannon Zhao <shannon.zhao@linaro.org> - 2016-03-24 16:00 +0100
      Re: [PATCH v7 10/17] arm/xen: Get event-channel irq through HVM_PARAM  when booting with ACPI Julien Grall <julien.grall@arm.com> - 2016-03-29 18:40 +0200
        Re: [PATCH v7 10/17] arm/xen: Get event-channel irq through HVM_PARAM  when booting with ACPI Shannon Zhao <zhaoshenglong@huawei.com> - 2016-03-30 09:40 +0200
    [PATCH v7 05/17] xen: memory : Add new XENMAPSPACE type XENMAPSPACE_dev_mmio Shannon Zhao <shannon.zhao@linaro.org> - 2016-03-24 16:00 +0100
    [PATCH v7 03/17] Xen: xlate: Use page_to_xen_pfn instead of page_to_pfn Shannon Zhao <shannon.zhao@linaro.org> - 2016-03-24 16:00 +0100
      Re: [PATCH v7 03/17] Xen: xlate: Use page_to_xen_pfn instead of  page_to_pfn Julien Grall <julien.grall@arm.com> - 2016-03-29 18:30 +0200
        Re: [PATCH v7 03/17] Xen: xlate: Use page_to_xen_pfn instead of page_to_pfn Shannon Zhao <zhaoshenglong@huawei.com> - 2016-03-30 09:40 +0200
          Re: [PATCH v7 03/17] Xen: xlate: Use page_to_xen_pfn instead of  page_to_pfn Julien Grall <julien.grall@arm.com> - 2016-03-30 13:30 +0200
            Re: [PATCH v7 03/17] Xen: xlate: Use page_to_xen_pfn instead of  page_to_pfn Shannon Zhao <shannon.zhao@linaro.org> - 2016-03-30 14:50 +0200

#1364232 — [PATCH v7 00/17] Add ACPI support for Xen Dom0 on ARM64

This patch set adds ACPI support for Xen Dom0 on ARM64. The relevant Xen
ACPI on ARM64 design document could be found from [1].

This patch set adds a new FDT node "uefi" under /hypervisor to pass UEFI
information. Introduce a bus notifier of AMBA and Platform bus to map
the new added device's MMIO space. Make Xen domain use
xlated_setup_gnttab_pages to setup grant table and a new hypercall to
get event-channel irq.

Regarding the initialization flow of Linux kernel, it needs to move
xen_early_init() before efi_init(). Then xen_early_init() will check
whether it runs on Xen through the /hypervisor node and efi_init() will
call a new function fdt_find_xen_uefi_params(), to parse those
xen,uefi-* parameters just like the existing efi_get_fdt_params().

And in arm64_enable_runtime_services() it will check whether it runs on
Xen and call another new function xen_efi_runtime_setup() to setup
runtime service instead of efi_native_runtime_setup(). The
xen_efi_runtime_setup() will assign the runtime function pointers with
the functions of driver/xen/efi.c.

And since we pass a /hypervisor node and a /chosen node to Dom0, it
needs to check whether the DTS only contains a /hypervisor node and a
/chosen node in acpi_boot_table_init().

Patches are tested on FVP base model.

Thanks,
Shannon

[1] http://lists.xen.org/archives/html/xen-devel/2015-11/msg00488.html

Changes since v6:
* rebase on linux master
* refactor codes as acpi_get_spcr_uart_addr (PATCH 1)
* sync with Xen (patch 9)

Changes since v5:
* rebase on linux master
* use acpi_dev_resource_memory to parse the device memory info(patch 1)
* sync with Xen (patch 9)

Changes since v4:
* rebase on linux master
* move the check acpi_device_should_be_hidden into
  acpi_bus_type_and_status (patch 1)
* use existing function fdt_subnode_offset (patch 16)

Changes since v3:
* rebase on linux master
* print a warning when there is no SPCR table
* rephase the commit message of PATCH 3
* rephase the words of PATCH 13
* use strcmp and factor the function in PATCH 16
* Add several ACKs and RBs, thanks a lot


Changes since v2:
* Use 0 to check if it should ignore the UART
* Fix the use of page_to_xen_pfn
* Factor ACPI and DT parts in xen_guest_init
* Check "uefi" node by full path
* Fix the statement of Documentation/devicetree/bindings/arm/xen.txt

Changes since v1:
* Rebase on linux mainline and wallclock patch from Stefano
* Refactor AMBA and platform device MMIO map to one file
* Use EFI_PARAVIRT to check if it supports XEN EFI
* Refactor Xen EFI codes
* Address other comments

Shannon Zhao (17):
  Xen: ACPI: Hide UART used by Xen
  xen/grant-table: Move xlated_setup_gnttab_pages to common place
  Xen: xlate: Use page_to_xen_pfn instead of page_to_pfn
  arm/xen: Use xen_xlate_map_ballooned_pages to setup grant table
  xen: memory : Add new XENMAPSPACE type XENMAPSPACE_dev_mmio
  Xen: ARM: Add support for mapping platform device mmio
  Xen: ARM: Add support for mapping AMBA device mmio
  Xen: public/hvm: sync changes of HVM_PARAM_CALLBACK_VIA ABI from Xen
  xen/hvm/params: Add a new delivery type for event-channel in
    HVM_PARAM_CALLBACK_IRQ
  arm/xen: Get event-channel irq through HVM_PARAM when booting with
    ACPI
  ARM: XEN: Move xen_early_init() before efi_init()
  ARM64: ACPI: Check if it runs on Xen to enable or disable ACPI
  ARM: Xen: Document UEFI support on Xen ARM virtual platforms
  XEN: EFI: Move x86 specific codes to architecture directory
  ARM64: XEN: Add a function to initialize Xen specific UEFI runtime
    services
  FDT: Add a helper to get the subnode by given name
  Xen: EFI: Parse DT parameters for Xen specific UEFI

 Documentation/devicetree/bindings/arm/xen.txt |  33 +++++
 arch/arm/include/asm/xen/xen-ops.h            |   6 +
 arch/arm/xen/Makefile                         |   1 +
 arch/arm/xen/efi.c                            |  40 ++++++
 arch/arm/xen/enlighten.c                      | 109 +++++++++++----
 arch/arm64/include/asm/xen/xen-ops.h          |   6 +
 arch/arm64/kernel/acpi.c                      |  12 +-
 arch/arm64/kernel/setup.c                     |   2 +-
 arch/arm64/xen/Makefile                       |   1 +
 arch/x86/xen/efi.c                            | 112 ++++++++++++++++
 arch/x86/xen/grant-table.c                    |  57 +-------
 drivers/acpi/scan.c                           |  68 ++++++++++
 drivers/firmware/efi/arm-runtime.c            |  17 ++-
 drivers/firmware/efi/efi.c                    |  45 ++++++-
 drivers/of/fdt.c                              |  13 ++
 drivers/xen/Kconfig                           |   2 +-
 drivers/xen/Makefile                          |   1 +
 drivers/xen/arm-device.c                      | 184 ++++++++++++++++++++++++++
 drivers/xen/efi.c                             | 174 +++++-------------------
 drivers/xen/xlate_mmu.c                       |  67 ++++++++++
 include/linux/of_fdt.h                        |   2 +
 include/xen/interface/hvm/params.h            |  40 +++++-
 include/xen/interface/memory.h                |   1 +
 include/xen/xen-ops.h                         |  32 +++--
 24 files changed, 778 insertions(+), 247 deletions(-)
 create mode 100644 arch/arm/include/asm/xen/xen-ops.h
 create mode 100644 arch/arm/xen/efi.c
 create mode 100644 arch/arm64/include/asm/xen/xen-ops.h
 create mode 100644 drivers/xen/arm-device.c

-- 
2.1.4

[toc] | [next] | [standalone]

#1364237 — [PATCH v7 08/17] Xen: public/hvm: sync changes of HVM_PARAM_CALLBACK_VIA ABI from Xen

Sync the changes of HVM_PARAM_CALLBACK_VIA ABI introduced by
Xen commit <ca5c54b6ff05> (public/hvm: export the HVM_PARAM_CALLBACK_VIA
ABI in the API).

Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
Acked-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
---
 include/xen/interface/hvm/params.h | 27 +++++++++++++++++++++------
 1 file changed, 21 insertions(+), 6 deletions(-)

diff --git a/include/xen/interface/hvm/params.h b/include/xen/interface/hvm/params.h
index a6c7991..70ad208 100644
--- a/include/xen/interface/hvm/params.h
+++ b/include/xen/interface/hvm/params.h
@@ -27,16 +27,31 @@
  * Parameter space for HVMOP_{set,get}_param.
  */
 
+#define HVM_PARAM_CALLBACK_IRQ 0
 /*
  * How should CPU0 event-channel notifications be delivered?
- * val[63:56] == 0: val[55:0] is a delivery GSI (Global System Interrupt).
- * val[63:56] == 1: val[55:0] is a delivery PCI INTx line, as follows:
- *                  Domain = val[47:32], Bus  = val[31:16],
- *                  DevFn  = val[15: 8], IntX = val[ 1: 0]
- * val[63:56] == 2: val[7:0] is a vector number.
+ *
  * If val == 0 then CPU0 event-channel notifications are not delivered.
+ * If val != 0, val[63:56] encodes the type, as follows:
+ */
+
+#define HVM_PARAM_CALLBACK_TYPE_GSI      0
+/*
+ * val[55:0] is a delivery GSI.  GSI 0 cannot be used, as it aliases val == 0,
+ * and disables all notifications.
+ */
+
+#define HVM_PARAM_CALLBACK_TYPE_PCI_INTX 1
+/*
+ * val[55:0] is a delivery PCI INTx line:
+ * Domain = val[47:32], Bus = val[31:16] DevFn = val[15:8], IntX = val[1:0]
+ */
+
+#define HVM_PARAM_CALLBACK_TYPE_VECTOR   2
+/*
+ * val[7:0] is a vector number.  Check for XENFEAT_hvm_callback_vector to know
+ * if this delivery method is available.
  */
-#define HVM_PARAM_CALLBACK_IRQ 0
 
 #define HVM_PARAM_STORE_PFN    1
 #define HVM_PARAM_STORE_EVTCHN 2
-- 
2.1.4

[toc] | [prev] | [next] | [standalone]

#1364242 — [PATCH v7 06/17] Xen: ARM: Add support for mapping platform device mmio

Add a bus_notifier for platform bus device in order to map the device
mmio regions when DOM0 booting with ACPI.

Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
Acked-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
---
 drivers/xen/Makefile     |   1 +
 drivers/xen/arm-device.c | 141 +++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 142 insertions(+)
 create mode 100644 drivers/xen/arm-device.c

diff --git a/drivers/xen/Makefile b/drivers/xen/Makefile
index 9b7a35c..415f286 100644
--- a/drivers/xen/Makefile
+++ b/drivers/xen/Makefile
@@ -9,6 +9,7 @@ CFLAGS_features.o			:= $(nostackp)
 
 CFLAGS_efi.o				+= -fshort-wchar
 
+dom0-$(CONFIG_ARM64) += arm-device.o
 dom0-$(CONFIG_PCI) += pci.o
 dom0-$(CONFIG_USB_SUPPORT) += dbgp.o
 dom0-$(CONFIG_XEN_ACPI) += acpi.o $(xen-pad-y)
diff --git a/drivers/xen/arm-device.c b/drivers/xen/arm-device.c
new file mode 100644
index 0000000..76e26e5
--- /dev/null
+++ b/drivers/xen/arm-device.c
@@ -0,0 +1,141 @@
+/*
+ * Copyright (c) 2015, Linaro Limited, Shannon Zhao
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/platform_device.h>
+#include <linux/acpi.h>
+#include <xen/xen.h>
+#include <xen/page.h>
+#include <xen/interface/memory.h>
+#include <asm/xen/hypervisor.h>
+#include <asm/xen/hypercall.h>
+
+static int xen_unmap_device_mmio(struct resource *resource, unsigned int count)
+{
+	unsigned int i, j, nr;
+	int rc = 0;
+	struct resource *r;
+	struct xen_remove_from_physmap xrp;
+
+	for (i = 0; i < count; i++) {
+		r = &resource[i];
+		nr = DIV_ROUND_UP(resource_size(r), XEN_PAGE_SIZE);
+		if ((resource_type(r) != IORESOURCE_MEM) || (nr == 0))
+			continue;
+
+		for (j = 0; j < nr; j++) {
+			xrp.domid = DOMID_SELF;
+			xrp.gpfn = XEN_PFN_DOWN(r->start) + j;
+			rc = HYPERVISOR_memory_op(XENMEM_remove_from_physmap,
+						  &xrp);
+			if (rc)
+				return rc;
+		}
+	}
+
+	return rc;
+}
+
+static int xen_map_device_mmio(struct resource *resource, unsigned int count)
+{
+	unsigned int i, j, nr;
+	int rc = 0;
+	struct resource *r;
+	xen_pfn_t *gpfns;
+	xen_ulong_t *idxs;
+	int *errs;
+	struct xen_add_to_physmap_range xatp;
+
+	for (i = 0; i < count; i++) {
+		r = &resource[i];
+		nr = DIV_ROUND_UP(resource_size(r), XEN_PAGE_SIZE);
+		if ((resource_type(r) != IORESOURCE_MEM) || (nr == 0))
+			continue;
+
+		gpfns = kzalloc(sizeof(xen_pfn_t) * nr, GFP_KERNEL);
+		idxs = kzalloc(sizeof(xen_ulong_t) * nr, GFP_KERNEL);
+		errs = kzalloc(sizeof(int) * nr, GFP_KERNEL);
+		if (!gpfns || !idxs || !errs) {
+			kfree(gpfns);
+			kfree(idxs);
+			kfree(errs);
+			return -ENOMEM;
+		}
+
+		for (j = 0; j < nr; j++) {
+			gpfns[j] = XEN_PFN_DOWN(r->start) + j;
+			idxs[j] = XEN_PFN_DOWN(r->start) + j;
+		}
+
+		xatp.domid = DOMID_SELF;
+		xatp.size = nr;
+		xatp.space = XENMAPSPACE_dev_mmio;
+
+		set_xen_guest_handle(xatp.gpfns, gpfns);
+		set_xen_guest_handle(xatp.idxs, idxs);
+		set_xen_guest_handle(xatp.errs, errs);
+
+		rc = HYPERVISOR_memory_op(XENMEM_add_to_physmap_range, &xatp);
+		kfree(gpfns);
+		kfree(idxs);
+		kfree(errs);
+		if (rc)
+			return rc;
+	}
+
+	return rc;
+}
+
+static int xen_platform_notifier(struct notifier_block *nb,
+				 unsigned long action, void *data)
+{
+	struct platform_device *pdev = to_platform_device(data);
+	int r = 0;
+
+	if (pdev->num_resources == 0 || pdev->resource == NULL)
+		return NOTIFY_OK;
+
+	switch (action) {
+	case BUS_NOTIFY_ADD_DEVICE:
+		r = xen_map_device_mmio(pdev->resource, pdev->num_resources);
+		break;
+	case BUS_NOTIFY_DEL_DEVICE:
+		r = xen_unmap_device_mmio(pdev->resource, pdev->num_resources);
+		break;
+	default:
+		return NOTIFY_DONE;
+	}
+	if (r)
+		dev_err(&pdev->dev, "Platform: Failed to %s device %s MMIO!\n",
+			action == BUS_NOTIFY_ADD_DEVICE ? "map" :
+			(action == BUS_NOTIFY_DEL_DEVICE ? "unmap" : "?"),
+			pdev->name);
+
+	return NOTIFY_OK;
+}
+
+static struct notifier_block platform_device_nb = {
+	.notifier_call = xen_platform_notifier,
+};
+
+static int __init register_xen_platform_notifier(void)
+{
+	if (!xen_initial_domain() || acpi_disabled)
+		return 0;
+
+	return bus_register_notifier(&platform_bus_type, &platform_device_nb);
+}
+
+arch_initcall(register_xen_platform_notifier);
-- 
2.1.4

[toc] | [prev] | [next] | [standalone]

#1364243 — [PATCH v7 02/17] xen/grant-table: Move xlated_setup_gnttab_pages to common place

Move xlated_setup_gnttab_pages to common place, so it can be reused by
ARM to setup grant table.

Rename it to xen_xlate_map_ballooned_pages.

Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
Reviewed-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
---
 arch/x86/xen/grant-table.c | 57 +++++--------------------------------------
 drivers/xen/xlate_mmu.c    | 61 ++++++++++++++++++++++++++++++++++++++++++++++
 include/xen/xen-ops.h      |  2 ++
 3 files changed, 69 insertions(+), 51 deletions(-)

diff --git a/arch/x86/xen/grant-table.c b/arch/x86/xen/grant-table.c
index e079500..de4144c 100644
--- a/arch/x86/xen/grant-table.c
+++ b/arch/x86/xen/grant-table.c
@@ -111,63 +111,18 @@ int arch_gnttab_init(unsigned long nr_shared)
 }
 
 #ifdef CONFIG_XEN_PVH
-#include <xen/balloon.h>
 #include <xen/events.h>
-#include <linux/slab.h>
-static int __init xlated_setup_gnttab_pages(void)
-{
-	struct page **pages;
-	xen_pfn_t *pfns;
-	void *vaddr;
-	int rc;
-	unsigned int i;
-	unsigned long nr_grant_frames = gnttab_max_grant_frames();
-
-	BUG_ON(nr_grant_frames == 0);
-	pages = kcalloc(nr_grant_frames, sizeof(pages[0]), GFP_KERNEL);
-	if (!pages)
-		return -ENOMEM;
-
-	pfns = kcalloc(nr_grant_frames, sizeof(pfns[0]), GFP_KERNEL);
-	if (!pfns) {
-		kfree(pages);
-		return -ENOMEM;
-	}
-	rc = alloc_xenballooned_pages(nr_grant_frames, pages);
-	if (rc) {
-		pr_warn("%s Couldn't balloon alloc %ld pfns rc:%d\n", __func__,
-			nr_grant_frames, rc);
-		kfree(pages);
-		kfree(pfns);
-		return rc;
-	}
-	for (i = 0; i < nr_grant_frames; i++)
-		pfns[i] = page_to_pfn(pages[i]);
-
-	vaddr = vmap(pages, nr_grant_frames, 0, PAGE_KERNEL);
-	if (!vaddr) {
-		pr_warn("%s Couldn't map %ld pfns rc:%d\n", __func__,
-			nr_grant_frames, rc);
-		free_xenballooned_pages(nr_grant_frames, pages);
-		kfree(pages);
-		kfree(pfns);
-		return -ENOMEM;
-	}
-	kfree(pages);
-
-	xen_auto_xlat_grant_frames.pfn = pfns;
-	xen_auto_xlat_grant_frames.count = nr_grant_frames;
-	xen_auto_xlat_grant_frames.vaddr = vaddr;
-
-	return 0;
-}
-
+#include <xen/xen-ops.h>
 static int __init xen_pvh_gnttab_setup(void)
 {
 	if (!xen_pvh_domain())
 		return -ENODEV;
 
-	return xlated_setup_gnttab_pages();
+	xen_auto_xlat_grant_frames.count = gnttab_max_grant_frames();
+
+	return xen_xlate_map_ballooned_pages(&xen_auto_xlat_grant_frames.pfn,
+					     &xen_auto_xlat_grant_frames.vaddr,
+					     xen_auto_xlat_grant_frames.count);
 }
 /* Call it _before_ __gnttab_init as we need to initialize the
  * xen_auto_xlat_grant_frames first. */
diff --git a/drivers/xen/xlate_mmu.c b/drivers/xen/xlate_mmu.c
index 5063c5e..9692656 100644
--- a/drivers/xen/xlate_mmu.c
+++ b/drivers/xen/xlate_mmu.c
@@ -29,6 +29,8 @@
  */
 #include <linux/kernel.h>
 #include <linux/mm.h>
+#include <linux/slab.h>
+#include <linux/vmalloc.h>
 
 #include <asm/xen/hypercall.h>
 #include <asm/xen/hypervisor.h>
@@ -37,6 +39,7 @@
 #include <xen/page.h>
 #include <xen/interface/xen.h>
 #include <xen/interface/memory.h>
+#include <xen/balloon.h>
 
 typedef void (*xen_gfn_fn_t)(unsigned long gfn, void *data);
 
@@ -185,3 +188,61 @@ int xen_xlate_unmap_gfn_range(struct vm_area_struct *vma,
 	return 0;
 }
 EXPORT_SYMBOL_GPL(xen_xlate_unmap_gfn_range);
+
+/**
+ * xen_xlate_map_ballooned_pages - map a new set of ballooned pages
+ * @gfns: returns the array of corresponding GFNs
+ * @virt: returns the virtual address of the mapped region
+ * @nr_grant_frames: number of GFNs
+ * @return 0 on success, error otherwise
+ *
+ * This allocates a set of ballooned pages and maps them into the
+ * kernel's address space.
+ */
+int __init xen_xlate_map_ballooned_pages(xen_pfn_t **gfns, void **virt,
+					 unsigned long nr_grant_frames)
+{
+	struct page **pages;
+	xen_pfn_t *pfns;
+	void *vaddr;
+	int rc;
+	unsigned int i;
+
+	BUG_ON(nr_grant_frames == 0);
+	pages = kcalloc(nr_grant_frames, sizeof(pages[0]), GFP_KERNEL);
+	if (!pages)
+		return -ENOMEM;
+
+	pfns = kcalloc(nr_grant_frames, sizeof(pfns[0]), GFP_KERNEL);
+	if (!pfns) {
+		kfree(pages);
+		return -ENOMEM;
+	}
+	rc = alloc_xenballooned_pages(nr_grant_frames, pages);
+	if (rc) {
+		pr_warn("%s Couldn't balloon alloc %ld pfns rc:%d\n", __func__,
+			nr_grant_frames, rc);
+		kfree(pages);
+		kfree(pfns);
+		return rc;
+	}
+	for (i = 0; i < nr_grant_frames; i++)
+		pfns[i] = page_to_pfn(pages[i]);
+
+	vaddr = vmap(pages, nr_grant_frames, 0, PAGE_KERNEL);
+	if (!vaddr) {
+		pr_warn("%s Couldn't map %ld pfns rc:%d\n", __func__,
+			nr_grant_frames, rc);
+		free_xenballooned_pages(nr_grant_frames, pages);
+		kfree(pages);
+		kfree(pfns);
+		return -ENOMEM;
+	}
+	kfree(pages);
+
+	*gfns = pfns;
+	*virt = vaddr;
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(xen_xlate_map_ballooned_pages);
diff --git a/include/xen/xen-ops.h b/include/xen/xen-ops.h
index 86abe07..072be1c 100644
--- a/include/xen/xen-ops.h
+++ b/include/xen/xen-ops.h
@@ -85,6 +85,8 @@ int xen_xlate_remap_gfn_array(struct vm_area_struct *vma,
 			      struct page **pages);
 int xen_xlate_unmap_gfn_range(struct vm_area_struct *vma,
 			      int nr, struct page **pages);
+int xen_xlate_map_ballooned_pages(xen_pfn_t **pfns, void **vaddr,
+				  unsigned long nr_grant_frames);
 
 bool xen_running_on_version_or_later(unsigned int major, unsigned int minor);
 
-- 
2.1.4

[toc] | [prev] | [next] | [standalone]

#1364247 — [PATCH v7 10/17] arm/xen: Get event-channel irq through HVM_PARAM when booting with ACPI

When booting with ACPI, it could get the event-channel irq through
HVM_PARAM_CALLBACK_IRQ.

Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
Reviewed-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
---
 arch/arm/xen/enlighten.c | 36 +++++++++++++++++++++++++++++++++++-
 1 file changed, 35 insertions(+), 1 deletion(-)

diff --git a/arch/arm/xen/enlighten.c b/arch/arm/xen/enlighten.c
index d94f726..680aae0 100644
--- a/arch/arm/xen/enlighten.c
+++ b/arch/arm/xen/enlighten.c
@@ -30,6 +30,7 @@
 #include <linux/time64.h>
 #include <linux/timekeeping.h>
 #include <linux/timekeeper_internal.h>
+#include <linux/acpi.h>
 
 #include <linux/mm.h>
 
@@ -278,6 +279,35 @@ void __init xen_early_init(void)
 		add_preferred_console("hvc", 0, NULL);
 }
 
+static void __init xen_acpi_guest_init(void)
+{
+#ifdef CONFIG_ACPI
+	struct xen_hvm_param a;
+	int interrupt, trigger, polarity;
+
+	a.domid = DOMID_SELF;
+	a.index = HVM_PARAM_CALLBACK_IRQ;
+	xen_events_irq = 0;
+
+	if (!HYPERVISOR_hvm_op(HVMOP_get_param, &a)) {
+		if ((a.value >> 56) == HVM_PARAM_CALLBACK_TYPE_PPI) {
+			interrupt = a.value & 0xff;
+			trigger = ((a.value >> 8) & 0x1) ? ACPI_EDGE_SENSITIVE
+							 : ACPI_LEVEL_SENSITIVE;
+			polarity = ((a.value >> 8) & 0x2) ? ACPI_ACTIVE_LOW
+							  : ACPI_ACTIVE_HIGH;
+			xen_events_irq = acpi_register_gsi(NULL, interrupt,
+							   trigger, polarity);
+		}
+	}
+#endif
+}
+
+static void __init xen_dt_guest_init(void)
+{
+	xen_events_irq = irq_of_parse_and_map(xen_node, 0);
+}
+
 static int __init xen_guest_init(void)
 {
 	struct xen_add_to_physmap xatp;
@@ -286,7 +316,11 @@ static int __init xen_guest_init(void)
 	if (!xen_domain())
 		return 0;
 
-	xen_events_irq = irq_of_parse_and_map(xen_node, 0);
+	if (!acpi_disabled)
+		xen_acpi_guest_init();
+	else
+		xen_dt_guest_init();
+
 	if (!xen_events_irq) {
 		pr_err("Xen event channel interrupt not found\n");
 		return -ENODEV;
-- 
2.1.4

[toc] | [prev] | [next] | [standalone]

#1366454 — Re: [PATCH v7 10/17] arm/xen: Get event-channel irq through HVM_PARAM when booting with ACPI

Hi Shannon,

On 24/03/16 14:44, Shannon Zhao wrote:
> When booting with ACPI, it could get the event-channel irq through

The kernel will always get the event-channel IRQ through 
HVM_PARAM_CALLBACK_IRQ.

So I would say: ", the kernel will get the event-channel..."

> HVM_PARAM_CALLBACK_IRQ.
>
> Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
> Reviewed-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
> ---
>   arch/arm/xen/enlighten.c | 36 +++++++++++++++++++++++++++++++++++-
>   1 file changed, 35 insertions(+), 1 deletion(-)
>
> diff --git a/arch/arm/xen/enlighten.c b/arch/arm/xen/enlighten.c
> index d94f726..680aae0 100644
> --- a/arch/arm/xen/enlighten.c
> +++ b/arch/arm/xen/enlighten.c
> @@ -30,6 +30,7 @@
>   #include <linux/time64.h>
>   #include <linux/timekeeping.h>
>   #include <linux/timekeeper_internal.h>
> +#include <linux/acpi.h>
>
>   #include <linux/mm.h>
>
> @@ -278,6 +279,35 @@ void __init xen_early_init(void)
>   		add_preferred_console("hvc", 0, NULL);
>   }
>
> +static void __init xen_acpi_guest_init(void)
> +{
> +#ifdef CONFIG_ACPI
> +	struct xen_hvm_param a;
> +	int interrupt, trigger, polarity;
> +
> +	a.domid = DOMID_SELF;
> +	a.index = HVM_PARAM_CALLBACK_IRQ;
> +	xen_events_irq = 0;
> +
> +	if (!HYPERVISOR_hvm_op(HVMOP_get_param, &a)) {
> +		if ((a.value >> 56) == HVM_PARAM_CALLBACK_TYPE_PPI) {
> +			interrupt = a.value & 0xff;
> +			trigger = ((a.value >> 8) & 0x1) ? ACPI_EDGE_SENSITIVE
> +							 : ACPI_LEVEL_SENSITIVE;
> +			polarity = ((a.value >> 8) & 0x2) ? ACPI_ACTIVE_LOW
> +							  : ACPI_ACTIVE_HIGH;
> +			xen_events_irq = acpi_register_gsi(NULL, interrupt,
> +							   trigger, polarity);
> +		}
> +	}

Can you invert the condition to remove one layer of indentation?

Regards,

-- 
Julien Grall

[toc] | [prev] | [next] | [standalone]

#1366961 — Re: [PATCH v7 10/17] arm/xen: Get event-channel irq through HVM_PARAM when booting with ACPI

On 2016/3/30 0:35, Julien Grall wrote:
> Hi Shannon,
> 
> On 24/03/16 14:44, Shannon Zhao wrote:
>> When booting with ACPI, it could get the event-channel irq through
> 
> The kernel will always get the event-channel IRQ through
> HVM_PARAM_CALLBACK_IRQ.
> 
> So I would say: ", the kernel will get the event-channel..."
> 
>> HVM_PARAM_CALLBACK_IRQ.
>>
>> Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
>> Reviewed-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
>> ---
>>   arch/arm/xen/enlighten.c | 36 +++++++++++++++++++++++++++++++++++-
>>   1 file changed, 35 insertions(+), 1 deletion(-)
>>
>> diff --git a/arch/arm/xen/enlighten.c b/arch/arm/xen/enlighten.c
>> index d94f726..680aae0 100644
>> --- a/arch/arm/xen/enlighten.c
>> +++ b/arch/arm/xen/enlighten.c
>> @@ -30,6 +30,7 @@
>>   #include <linux/time64.h>
>>   #include <linux/timekeeping.h>
>>   #include <linux/timekeeper_internal.h>
>> +#include <linux/acpi.h>
>>
>>   #include <linux/mm.h>
>>
>> @@ -278,6 +279,35 @@ void __init xen_early_init(void)
>>           add_preferred_console("hvc", 0, NULL);
>>   }
>>
>> +static void __init xen_acpi_guest_init(void)
>> +{
>> +#ifdef CONFIG_ACPI
>> +    struct xen_hvm_param a;
>> +    int interrupt, trigger, polarity;
>> +
>> +    a.domid = DOMID_SELF;
>> +    a.index = HVM_PARAM_CALLBACK_IRQ;
>> +    xen_events_irq = 0;
>> +
>> +    if (!HYPERVISOR_hvm_op(HVMOP_get_param, &a)) {
>> +        if ((a.value >> 56) == HVM_PARAM_CALLBACK_TYPE_PPI) {
>> +            interrupt = a.value & 0xff;
>> +            trigger = ((a.value >> 8) & 0x1) ? ACPI_EDGE_SENSITIVE
>> +                             : ACPI_LEVEL_SENSITIVE;
>> +            polarity = ((a.value >> 8) & 0x2) ? ACPI_ACTIVE_LOW
>> +                              : ACPI_ACTIVE_HIGH;
>> +            xen_events_irq = acpi_register_gsi(NULL, interrupt,
>> +                               trigger, polarity);
>> +        }
>> +    }
> 
> Can you invert the condition to remove one layer of indentation?
Sure, will update.

Thanks,
-- 
Shannon

[toc] | [prev] | [next] | [standalone]

#1364248 — [PATCH v7 05/17] xen: memory : Add new XENMAPSPACE type XENMAPSPACE_dev_mmio

Add a new type of Xen map space for Dom0 to map device's MMIO region.

Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
---
 include/xen/interface/memory.h | 1 +
 1 file changed, 1 insertion(+)

diff --git a/include/xen/interface/memory.h b/include/xen/interface/memory.h
index 2ecfe4f..9aa8988 100644
--- a/include/xen/interface/memory.h
+++ b/include/xen/interface/memory.h
@@ -160,6 +160,7 @@ DEFINE_GUEST_HANDLE_STRUCT(xen_machphys_mapping_t);
 #define XENMAPSPACE_gmfn_foreign 4 /* GMFN from another dom,
 				    * XENMEM_add_to_physmap_range only.
 				    */
+#define XENMAPSPACE_dev_mmio     5 /* device mmio region */
 
 /*
  * Sets the GPFN at which a particular page appears in the specified guest's
-- 
2.1.4

[toc] | [prev] | [next] | [standalone]

#1364251 — [PATCH v7 03/17] Xen: xlate: Use page_to_xen_pfn instead of page_to_pfn

Make xen_xlate_map_ballooned_pages work with 64K pages. In that case
Kernel pages are 64K in size but Xen pages remain 4K in size. Xen pfns
refer to 4K pages.

Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
Reviewed-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
---
 drivers/xen/xlate_mmu.c | 26 ++++++++++++++++----------
 1 file changed, 16 insertions(+), 10 deletions(-)

diff --git a/drivers/xen/xlate_mmu.c b/drivers/xen/xlate_mmu.c
index 9692656..28f728b 100644
--- a/drivers/xen/xlate_mmu.c
+++ b/drivers/xen/xlate_mmu.c
@@ -207,9 +207,12 @@ int __init xen_xlate_map_ballooned_pages(xen_pfn_t **gfns, void **virt,
 	void *vaddr;
 	int rc;
 	unsigned int i;
+	unsigned long nr_pages;
+	xen_pfn_t xen_pfn = 0;
 
 	BUG_ON(nr_grant_frames == 0);
-	pages = kcalloc(nr_grant_frames, sizeof(pages[0]), GFP_KERNEL);
+	nr_pages = DIV_ROUND_UP(nr_grant_frames, XEN_PFN_PER_PAGE);
+	pages = kcalloc(nr_pages, sizeof(pages[0]), GFP_KERNEL);
 	if (!pages)
 		return -ENOMEM;
 
@@ -218,22 +221,25 @@ int __init xen_xlate_map_ballooned_pages(xen_pfn_t **gfns, void **virt,
 		kfree(pages);
 		return -ENOMEM;
 	}
-	rc = alloc_xenballooned_pages(nr_grant_frames, pages);
+	rc = alloc_xenballooned_pages(nr_pages, pages);
 	if (rc) {
-		pr_warn("%s Couldn't balloon alloc %ld pfns rc:%d\n", __func__,
-			nr_grant_frames, rc);
+		pr_warn("%s Couldn't balloon alloc %ld pages rc:%d\n", __func__,
+			nr_pages, rc);
 		kfree(pages);
 		kfree(pfns);
 		return rc;
 	}
-	for (i = 0; i < nr_grant_frames; i++)
-		pfns[i] = page_to_pfn(pages[i]);
+	for (i = 0; i < nr_grant_frames; i++) {
+		if ((i % XEN_PFN_PER_PAGE) == 0)
+			xen_pfn = page_to_xen_pfn(pages[i / XEN_PFN_PER_PAGE]);
+		pfns[i] = pfn_to_gfn(xen_pfn++);
+	}
 
-	vaddr = vmap(pages, nr_grant_frames, 0, PAGE_KERNEL);
+	vaddr = vmap(pages, nr_pages, 0, PAGE_KERNEL);
 	if (!vaddr) {
-		pr_warn("%s Couldn't map %ld pfns rc:%d\n", __func__,
-			nr_grant_frames, rc);
-		free_xenballooned_pages(nr_grant_frames, pages);
+		pr_warn("%s Couldn't map %ld pages rc:%d\n", __func__,
+			nr_pages, rc);
+		free_xenballooned_pages(nr_pages, pages);
 		kfree(pages);
 		kfree(pfns);
 		return -ENOMEM;
-- 
2.1.4

[toc] | [prev] | [next] | [standalone]

#1366441 — Re: [PATCH v7 03/17] Xen: xlate: Use page_to_xen_pfn instead of page_to_pfn

Hi Shannon,

On 24/03/16 14:44, Shannon Zhao wrote:
> Make xen_xlate_map_ballooned_pages work with 64K pages. In that case
> Kernel pages are 64K in size but Xen pages remain 4K in size. Xen pfns
> refer to 4K pages.
>
> Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
> Reviewed-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
> ---
>   drivers/xen/xlate_mmu.c | 26 ++++++++++++++++----------
>   1 file changed, 16 insertions(+), 10 deletions(-)
>
> diff --git a/drivers/xen/xlate_mmu.c b/drivers/xen/xlate_mmu.c
> index 9692656..28f728b 100644
> --- a/drivers/xen/xlate_mmu.c
> +++ b/drivers/xen/xlate_mmu.c
> @@ -207,9 +207,12 @@ int __init xen_xlate_map_ballooned_pages(xen_pfn_t **gfns, void **virt,
>   	void *vaddr;
>   	int rc;
>   	unsigned int i;
> +	unsigned long nr_pages;
> +	xen_pfn_t xen_pfn = 0;
>
>   	BUG_ON(nr_grant_frames == 0);
> -	pages = kcalloc(nr_grant_frames, sizeof(pages[0]), GFP_KERNEL);
> +	nr_pages = DIV_ROUND_UP(nr_grant_frames, XEN_PFN_PER_PAGE);
> +	pages = kcalloc(nr_pages, sizeof(pages[0]), GFP_KERNEL);
>   	if (!pages)
>   		return -ENOMEM;
>
> @@ -218,22 +221,25 @@ int __init xen_xlate_map_ballooned_pages(xen_pfn_t **gfns, void **virt,
>   		kfree(pages);
>   		return -ENOMEM;
>   	}
> -	rc = alloc_xenballooned_pages(nr_grant_frames, pages);
> +	rc = alloc_xenballooned_pages(nr_pages, pages);
>   	if (rc) {
> -		pr_warn("%s Couldn't balloon alloc %ld pfns rc:%d\n", __func__,
> -			nr_grant_frames, rc);
> +		pr_warn("%s Couldn't balloon alloc %ld pages rc:%d\n", __func__,
> +			nr_pages, rc);
>   		kfree(pages);
>   		kfree(pfns);
>   		return rc;
>   	}
> -	for (i = 0; i < nr_grant_frames; i++)
> -		pfns[i] = page_to_pfn(pages[i]);
> +	for (i = 0; i < nr_grant_frames; i++) {
> +		if ((i % XEN_PFN_PER_PAGE) == 0)
> +			xen_pfn = page_to_xen_pfn(pages[i / XEN_PFN_PER_PAGE]);
> +		pfns[i] = pfn_to_gfn(xen_pfn++);
> +	}

Would it be possible to re-use xen_for_each_gfn? This will avoid 
open-coding the loop to break down the Linux page.

Regards,

-- 
Julien Grall

[toc] | [prev] | [next] | [standalone]

#1366959 — Re: [PATCH v7 03/17] Xen: xlate: Use page_to_xen_pfn instead of page_to_pfn

On 2016/3/30 0:28, Julien Grall wrote:
> Hi Shannon,
> 
> On 24/03/16 14:44, Shannon Zhao wrote:
>> Make xen_xlate_map_ballooned_pages work with 64K pages. In that case
>> Kernel pages are 64K in size but Xen pages remain 4K in size. Xen pfns
>> refer to 4K pages.
>>
>> Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
>> Reviewed-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
>> ---
>>   drivers/xen/xlate_mmu.c | 26 ++++++++++++++++----------
>>   1 file changed, 16 insertions(+), 10 deletions(-)
>>
>> diff --git a/drivers/xen/xlate_mmu.c b/drivers/xen/xlate_mmu.c
>> index 9692656..28f728b 100644
>> --- a/drivers/xen/xlate_mmu.c
>> +++ b/drivers/xen/xlate_mmu.c
>> @@ -207,9 +207,12 @@ int __init
>> xen_xlate_map_ballooned_pages(xen_pfn_t **gfns, void **virt,
>>       void *vaddr;
>>       int rc;
>>       unsigned int i;
>> +    unsigned long nr_pages;
>> +    xen_pfn_t xen_pfn = 0;
>>
>>       BUG_ON(nr_grant_frames == 0);
>> -    pages = kcalloc(nr_grant_frames, sizeof(pages[0]), GFP_KERNEL);
>> +    nr_pages = DIV_ROUND_UP(nr_grant_frames, XEN_PFN_PER_PAGE);
>> +    pages = kcalloc(nr_pages, sizeof(pages[0]), GFP_KERNEL);
>>       if (!pages)
>>           return -ENOMEM;
>>
>> @@ -218,22 +221,25 @@ int __init
>> xen_xlate_map_ballooned_pages(xen_pfn_t **gfns, void **virt,
>>           kfree(pages);
>>           return -ENOMEM;
>>       }
>> -    rc = alloc_xenballooned_pages(nr_grant_frames, pages);
>> +    rc = alloc_xenballooned_pages(nr_pages, pages);
>>       if (rc) {
>> -        pr_warn("%s Couldn't balloon alloc %ld pfns rc:%d\n", __func__,
>> -            nr_grant_frames, rc);
>> +        pr_warn("%s Couldn't balloon alloc %ld pages rc:%d\n", __func__,
>> +            nr_pages, rc);
>>           kfree(pages);
>>           kfree(pfns);
>>           return rc;
>>       }
>> -    for (i = 0; i < nr_grant_frames; i++)
>> -        pfns[i] = page_to_pfn(pages[i]);
>> +    for (i = 0; i < nr_grant_frames; i++) {
>> +        if ((i % XEN_PFN_PER_PAGE) == 0)
>> +            xen_pfn = page_to_xen_pfn(pages[i / XEN_PFN_PER_PAGE]);
>> +        pfns[i] = pfn_to_gfn(xen_pfn++);
>> +    }
> 
> Would it be possible to re-use xen_for_each_gfn? This will avoid
> open-coding the loop to break down the Linux page.
I don't think so. Using xen_acpi_guest_init will require factoring
"pfns[i] = pfn_to_gfn(xen_pfn++)" to a function with parameter pfns[i].
How can we pass pfns[i]?

Thanks,
-- 
Shannon

[toc] | [prev] | [next] | [standalone]

#1367113 — Re: [PATCH v7 03/17] Xen: xlate: Use page_to_xen_pfn instead of page_to_pfn

Hi Shannon,

On 30/03/16 08:38, Shannon Zhao wrote:
>
>
> On 2016/3/30 0:28, Julien Grall wrote:
>> On 24/03/16 14:44, Shannon Zhao wrote:
>>> Make xen_xlate_map_ballooned_pages work with 64K pages. In that case
>>> Kernel pages are 64K in size but Xen pages remain 4K in size. Xen pfns
>>> refer to 4K pages.
>>>
>>> Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
>>> Reviewed-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
>>> ---
>>>    drivers/xen/xlate_mmu.c | 26 ++++++++++++++++----------
>>>    1 file changed, 16 insertions(+), 10 deletions(-)
>>>
>>> diff --git a/drivers/xen/xlate_mmu.c b/drivers/xen/xlate_mmu.c
>>> index 9692656..28f728b 100644
>>> --- a/drivers/xen/xlate_mmu.c
>>> +++ b/drivers/xen/xlate_mmu.c
>>> @@ -207,9 +207,12 @@ int __init
>>> xen_xlate_map_ballooned_pages(xen_pfn_t **gfns, void **virt,
>>>        void *vaddr;
>>>        int rc;
>>>        unsigned int i;
>>> +    unsigned long nr_pages;
>>> +    xen_pfn_t xen_pfn = 0;
>>>
>>>        BUG_ON(nr_grant_frames == 0);
>>> -    pages = kcalloc(nr_grant_frames, sizeof(pages[0]), GFP_KERNEL);
>>> +    nr_pages = DIV_ROUND_UP(nr_grant_frames, XEN_PFN_PER_PAGE);
>>> +    pages = kcalloc(nr_pages, sizeof(pages[0]), GFP_KERNEL);
>>>        if (!pages)
>>>            return -ENOMEM;
>>>
>>> @@ -218,22 +221,25 @@ int __init
>>> xen_xlate_map_ballooned_pages(xen_pfn_t **gfns, void **virt,
>>>            kfree(pages);
>>>            return -ENOMEM;
>>>        }
>>> -    rc = alloc_xenballooned_pages(nr_grant_frames, pages);
>>> +    rc = alloc_xenballooned_pages(nr_pages, pages);
>>>        if (rc) {
>>> -        pr_warn("%s Couldn't balloon alloc %ld pfns rc:%d\n", __func__,
>>> -            nr_grant_frames, rc);
>>> +        pr_warn("%s Couldn't balloon alloc %ld pages rc:%d\n", __func__,
>>> +            nr_pages, rc);
>>>            kfree(pages);
>>>            kfree(pfns);
>>>            return rc;
>>>        }
>>> -    for (i = 0; i < nr_grant_frames; i++)
>>> -        pfns[i] = page_to_pfn(pages[i]);
>>> +    for (i = 0; i < nr_grant_frames; i++) {
>>> +        if ((i % XEN_PFN_PER_PAGE) == 0)
>>> +            xen_pfn = page_to_xen_pfn(pages[i / XEN_PFN_PER_PAGE]);
>>> +        pfns[i] = pfn_to_gfn(xen_pfn++);
>>> +    }
>>
>> Would it be possible to re-use xen_for_each_gfn? This will avoid
>> open-coding the loop to break down the Linux page.
> I don't think so. Using xen_acpi_guest_init will require factoring
> "pfns[i] = pfn_to_gfn(xen_pfn++)" to a function with parameter pfns[i].
> How can we pass pfns[i]?

By using the variable data. Something along those lines:

struct map_balloon_pages
{
       xen_pfn_t *pfns;
       unsigned int idx;
};

static void setup_balloon_gfn(unsigned long gfn, void *data)
{
       struct map_balloon_pages *info = data;


       data->pfns[data->idx] = gfn;
       data->idx++;
}

And then in xen_xlate_map_ballooned_pages

xen_for_each_gfn(pages, nr_grant_frames, setup_balloon_gfn, &data);

Regards,

-- 
Julien Grall

[toc] | [prev] | [next] | [standalone]

#1367179 — Re: [PATCH v7 03/17] Xen: xlate: Use page_to_xen_pfn instead of page_to_pfn

On 2016年03月30日 19:22, Julien Grall wrote:
> Hi Shannon,
> 
> On 30/03/16 08:38, Shannon Zhao wrote:
>>
>>
>> On 2016/3/30 0:28, Julien Grall wrote:
>>> On 24/03/16 14:44, Shannon Zhao wrote:
>>>> Make xen_xlate_map_ballooned_pages work with 64K pages. In that case
>>>> Kernel pages are 64K in size but Xen pages remain 4K in size. Xen pfns
>>>> refer to 4K pages.
>>>>
>>>> Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
>>>> Reviewed-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
>>>> ---
>>>>    drivers/xen/xlate_mmu.c | 26 ++++++++++++++++----------
>>>>    1 file changed, 16 insertions(+), 10 deletions(-)
>>>>
>>>> diff --git a/drivers/xen/xlate_mmu.c b/drivers/xen/xlate_mmu.c
>>>> index 9692656..28f728b 100644
>>>> --- a/drivers/xen/xlate_mmu.c
>>>> +++ b/drivers/xen/xlate_mmu.c
>>>> @@ -207,9 +207,12 @@ int __init
>>>> xen_xlate_map_ballooned_pages(xen_pfn_t **gfns, void **virt,
>>>>        void *vaddr;
>>>>        int rc;
>>>>        unsigned int i;
>>>> +    unsigned long nr_pages;
>>>> +    xen_pfn_t xen_pfn = 0;
>>>>
>>>>        BUG_ON(nr_grant_frames == 0);
>>>> -    pages = kcalloc(nr_grant_frames, sizeof(pages[0]), GFP_KERNEL);
>>>> +    nr_pages = DIV_ROUND_UP(nr_grant_frames, XEN_PFN_PER_PAGE);
>>>> +    pages = kcalloc(nr_pages, sizeof(pages[0]), GFP_KERNEL);
>>>>        if (!pages)
>>>>            return -ENOMEM;
>>>>
>>>> @@ -218,22 +221,25 @@ int __init
>>>> xen_xlate_map_ballooned_pages(xen_pfn_t **gfns, void **virt,
>>>>            kfree(pages);
>>>>            return -ENOMEM;
>>>>        }
>>>> -    rc = alloc_xenballooned_pages(nr_grant_frames, pages);
>>>> +    rc = alloc_xenballooned_pages(nr_pages, pages);
>>>>        if (rc) {
>>>> -        pr_warn("%s Couldn't balloon alloc %ld pfns rc:%d\n",
>>>> __func__,
>>>> -            nr_grant_frames, rc);
>>>> +        pr_warn("%s Couldn't balloon alloc %ld pages rc:%d\n",
>>>> __func__,
>>>> +            nr_pages, rc);
>>>>            kfree(pages);
>>>>            kfree(pfns);
>>>>            return rc;
>>>>        }
>>>> -    for (i = 0; i < nr_grant_frames; i++)
>>>> -        pfns[i] = page_to_pfn(pages[i]);
>>>> +    for (i = 0; i < nr_grant_frames; i++) {
>>>> +        if ((i % XEN_PFN_PER_PAGE) == 0)
>>>> +            xen_pfn = page_to_xen_pfn(pages[i / XEN_PFN_PER_PAGE]);
>>>> +        pfns[i] = pfn_to_gfn(xen_pfn++);
>>>> +    }
>>>
>>> Would it be possible to re-use xen_for_each_gfn? This will avoid
>>> open-coding the loop to break down the Linux page.
>> I don't think so. Using xen_acpi_guest_init will require factoring
>> "pfns[i] = pfn_to_gfn(xen_pfn++)" to a function with parameter pfns[i].
>> How can we pass pfns[i]?
> 
> By using the variable data. Something along those lines:
> 
> struct map_balloon_pages
> {
>       xen_pfn_t *pfns;
>       unsigned int idx;
> };
> 
> static void setup_balloon_gfn(unsigned long gfn, void *data)
> {
>       struct map_balloon_pages *info = data;
> 
> 
>       data->pfns[data->idx] = gfn;
>       data->idx++;
> }
> 
> And then in xen_xlate_map_ballooned_pages
> 
> xen_for_each_gfn(pages, nr_grant_frames, setup_balloon_gfn, &data);
I think this looks like less direct. Anyway I will update this patch as
you said.

Thanks,
-- 
Shannon

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web