[PATCH 00/11] ACPI, APEI, Patches for 3.3

[PATCH 00/11] ACPI, APEI, Patches for 3.3

[Huang Ying]

Most patches has been sent for 3.2 before.  The changes is as follow:

- Add [01/11]: Add 64bit read/write support to atomicio on i386
- Change [07/11]: Do not use kmap on IA64

For [10/11] and [11/11], there is some other idea to fix the issue,
if some other method is preferred, please drop these two patches.

[PATCH 01/11] ACPI, Add 64bit read/write support to atomicio on i386
[PATCH 02/11] ACPI, APEI, GHES: Add PCIe AER recovery support
[PATCH 03/11] ACPI, APEI, Print resource errors in conventional
[PATCH 04/11] ACPI, APEI, Remove table not found message
[PATCH 05/11] ACPI, APEI, GHES, Distinguish interleaved error report
[PATCH 06/11] ACPI, APEI, Printk queued error record before panic
[PATCH 07/11] ACPI, Add RAM mapping support to ACPI atomic IO
[PATCH 08/11] ACPI, APEI, EINJ, Fix resource conflict on some
[PATCH 09/11] ACPI, APEI, EINJ, Refine the fix of resource conflict
[PATCH 10/11] ACPI, Record ACPI NVS regions
[PATCH 11/11] ACPI, APEI, Resolve false conflict between ACPI NVS

Best Regards,
Huang Ying

[PATCH 01/11] ACPI, Add 64bit read/write support to atomicio on i386

[Huang Ying]

There is no 64bit read/write support in ACPI atomicio because
readq/writeq is used to implement 64bit read/write, but readq/writeq
is not available on i386.  This patch implement 64bit read/write
support in atomicio via two readl/writel.

Signed-off-by: Huang Ying <ying.huang@intel.com>
---
 drivers/acpi/atomicio.c |   36 ++++++++++++++++++++++++++++++------
 1 file changed, 30 insertions(+), 6 deletions(-)

--- a/drivers/acpi/atomicio.c
+++ b/drivers/acpi/atomicio.c
@@ -260,6 +260,21 @@ int acpi_post_unmap_gar(struct acpi_gene
 }
 EXPORT_SYMBOL_GPL(acpi_post_unmap_gar);
 
+#ifdef readq
+static inline u64 read64(const volatile void __iomem *addr)
+{
+	return readq(addr);
+}
+#else
+static inline u64 read64(const volatile void __iomem *addr)
+{
+	u64 l, h;
+	l = readl(addr);
+	h = readl(addr+4);
+	return l | (h << 32);
+}
+#endif
+
 /*
  * Can be used in atomic (including NMI) or process context. RCU read
  * lock can only be released after the IO memory area accessing.
@@ -280,11 +295,9 @@ static int acpi_atomic_read_mem(u64 padd
 	case 32:
 		*val = readl(addr);
 		break;
-#ifdef readq
 	case 64:
-		*val = readq(addr);
+		*val = read64(addr);
 		break;
-#endif
 	default:
 		return -EINVAL;
 	}
@@ -293,6 +306,19 @@ static int acpi_atomic_read_mem(u64 padd
 	return 0;
 }
 
+#ifdef writeq
+static inline void write64(u64 val, volatile void __iomem *addr)
+{
+	writeq(val, addr);
+}
+#else
+static inline void write64(u64 val, volatile void __iomem *addr)
+{
+	writel(val, addr);
+	writel(val>>32, addr+4);
+}
+#endif
+
 static int acpi_atomic_write_mem(u64 paddr, u64 val, u32 width)
 {
 	void __iomem *addr;
@@ -309,11 +335,9 @@ static int acpi_atomic_write_mem(u64 pad
 	case 32:
 		writel(val, addr);
 		break;
-#ifdef writeq
 	case 64:
-		writeq(val, addr);
+		write64(val, addr);
 		break;
-#endif
 	default:
 		return -EINVAL;
 	}

[PATCH 02/11] ACPI, APEI, GHES: Add PCIe AER recovery support

[Huang Ying]

aer_recover_queue() is called when recoverable PCIe AER errors are
notified by firmware to do the recovery work.

Signed-off-by: Huang Ying <ying.huang@intel.com>
---
 drivers/acpi/apei/ghes.c |   23 +++++++++++++++++++++++
 1 file changed, 23 insertions(+)

--- a/drivers/acpi/apei/ghes.c
+++ b/drivers/acpi/apei/ghes.c
@@ -45,6 +45,8 @@
 #include <linux/irq_work.h>
 #include <linux/llist.h>
 #include <linux/genalloc.h>
+#include <linux/pci.h>
+#include <linux/aer.h>
 #include <acpi/apei.h>
 #include <acpi/atomicio.h>
 #include <acpi/hed.h>
@@ -476,6 +478,27 @@ static void ghes_do_proc(const struct ac
 			}
 #endif
 		}
+#ifdef CONFIG_ACPI_APEI_PCIEAER
+		else if (!uuid_le_cmp(*(uuid_le *)gdata->section_type,
+				      CPER_SEC_PCIE)) {
+			struct cper_sec_pcie *pcie_err;
+			pcie_err = (struct cper_sec_pcie *)(gdata+1);
+			if (sev == GHES_SEV_RECOVERABLE &&
+			    sec_sev == GHES_SEV_RECOVERABLE &&
+			    pcie_err->validation_bits & CPER_PCIE_VALID_DEVICE_ID &&
+			    pcie_err->validation_bits & CPER_PCIE_VALID_AER_INFO) {
+				unsigned int devfn;
+				int aer_severity;
+				devfn = PCI_DEVFN(pcie_err->device_id.device,
+						  pcie_err->device_id.function);
+				aer_severity = cper_severity_to_aer(sev);
+				aer_recover_queue(pcie_err->device_id.segment,
+						  pcie_err->device_id.bus,
+						  devfn, aer_severity);
+			}
+
+		}
+#endif
 	}
 }
 

[PATCH 03/11] ACPI, APEI, Print resource errors in conventional format

[Huang Ying]

From: Bjorn Helgaas <bhelgaas@google.com>

Use the normal %pR-like format for MMIO and I/O port ranges.

Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Huang Ying <ying.huang@intel.com>
---
 drivers/acpi/apei/apei-base.c |    8 ++++----
 drivers/acpi/apei/einj.c      |   11 ++++++-----
 2 files changed, 10 insertions(+), 9 deletions(-)

--- a/drivers/acpi/apei/apei-base.c
+++ b/drivers/acpi/apei/apei-base.c
@@ -460,9 +460,9 @@ int apei_resources_request(struct apei_r
 				       desc);
 		if (!r) {
 			pr_err(APEI_PFX
-		"Can not request iomem region <%016llx-%016llx> for GARs.\n",
+		"Can not request [mem %#010llx-%#010llx] for %s registers\n",
 			       (unsigned long long)res->start,
-			       (unsigned long long)res->end);
+			       (unsigned long long)res->end - 1, desc);
 			res_bak = res;
 			goto err_unmap_iomem;
 		}
@@ -472,9 +472,9 @@ int apei_resources_request(struct apei_r
 		r = request_region(res->start, res->end - res->start, desc);
 		if (!r) {
 			pr_err(APEI_PFX
-		"Can not request ioport region <%016llx-%016llx> for GARs.\n",
+		"Can not request [io  %#06llx-%#06llx] for %s registers\n",
 			       (unsigned long long)res->start,
-			       (unsigned long long)res->end);
+			       (unsigned long long)res->end - 1, desc);
 			res_bak = res;
 			goto err_unmap_ioport;
 		}
--- a/drivers/acpi/apei/einj.c
+++ b/drivers/acpi/apei/einj.c
@@ -209,9 +209,10 @@ static int __einj_error_trigger(u64 trig
 			       "APEI EINJ Trigger Table");
 	if (!r) {
 		pr_err(EINJ_PFX
-	"Can not request iomem region <%016llx-%016llx> for Trigger table.\n",
+	"Can not request [mem %#010llx-%#010llx] for Trigger table\n",
 		       (unsigned long long)trigger_paddr,
-		       (unsigned long long)trigger_paddr+sizeof(*trigger_tab));
+		       (unsigned long long)trigger_paddr +
+			    sizeof(*trigger_tab) - 1);
 		goto out;
 	}
 	trigger_tab = ioremap_cache(trigger_paddr, sizeof(*trigger_tab));
@@ -232,9 +233,9 @@ static int __einj_error_trigger(u64 trig
 			       "APEI EINJ Trigger Table");
 	if (!r) {
 		pr_err(EINJ_PFX
-"Can not request iomem region <%016llx-%016llx> for Trigger Table Entry.\n",
-		       (unsigned long long)trigger_paddr+sizeof(*trigger_tab),
-		       (unsigned long long)trigger_paddr + table_size);
+"Can not request [mem %#010llx-%#010llx] for Trigger Table Entry\n",
+		       (unsigned long long)trigger_paddr + sizeof(*trigger_tab),
+		       (unsigned long long)trigger_paddr + table_size - 1);
 		goto out_rel_header;
 	}
 	iounmap(trigger_tab);

[PATCH 04/11] ACPI, APEI, Remove table not found message

[Huang Ying]

Because APEI tables are optional, these message may confuse users, for
example,

https://bugs.launchpad.net/ubuntu/+source/linux/+bug/599715

Reported-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Huang Ying <ying.huang@intel.com>
---
 drivers/acpi/apei/einj.c |    5 ++---
 drivers/acpi/apei/erst.c |    5 ++---
 drivers/acpi/apei/hest.c |    5 ++---
 3 files changed, 6 insertions(+), 9 deletions(-)

--- a/drivers/acpi/apei/einj.c
+++ b/drivers/acpi/apei/einj.c
@@ -466,10 +466,9 @@ static int __init einj_init(void)
 
 	status = acpi_get_table(ACPI_SIG_EINJ, 0,
 				(struct acpi_table_header **)&einj_tab);
-	if (status == AE_NOT_FOUND) {
-		pr_info(EINJ_PFX "Table is not found!\n");
+	if (status == AE_NOT_FOUND)
 		return -ENODEV;
-	} else if (ACPI_FAILURE(status)) {
+	else if (ACPI_FAILURE(status)) {
 		const char *msg = acpi_format_exception(status);
 		pr_err(EINJ_PFX "Failed to get table, %s\n", msg);
 		return -EINVAL;
--- a/drivers/acpi/apei/erst.c
+++ b/drivers/acpi/apei/erst.c
@@ -1125,10 +1125,9 @@ static int __init erst_init(void)
 
 	status = acpi_get_table(ACPI_SIG_ERST, 0,
 				(struct acpi_table_header **)&erst_tab);
-	if (status == AE_NOT_FOUND) {
-		pr_info(ERST_PFX "Table is not found!\n");
+	if (status == AE_NOT_FOUND)
 		goto err;
-	} else if (ACPI_FAILURE(status)) {
+	else if (ACPI_FAILURE(status)) {
 		const char *msg = acpi_format_exception(status);
 		pr_err(ERST_PFX "Failed to get table, %s\n", msg);
 		rc = -EINVAL;
--- a/drivers/acpi/apei/hest.c
+++ b/drivers/acpi/apei/hest.c
@@ -221,10 +221,9 @@ void __init acpi_hest_init(void)
 
 	status = acpi_get_table(ACPI_SIG_HEST, 0,
 				(struct acpi_table_header **)&hest_tab);
-	if (status == AE_NOT_FOUND) {
-		pr_info(HEST_PFX "Table not found.\n");
+	if (status == AE_NOT_FOUND)
 		goto err;
-	} else if (ACPI_FAILURE(status)) {
+	else if (ACPI_FAILURE(status)) {
 		const char *msg = acpi_format_exception(status);
 		pr_err(HEST_PFX "Failed to get table, %s\n", msg);
 		rc = -EINVAL;

[PATCH 05/11] ACPI, APEI, GHES, Distinguish interleaved error report in kernel log

[Huang Ying]

In most cases, printk only guarantees messages from different printk
calling will not be interleaved between each other.  But, one APEI
GHES hardware error report will involve multiple printk calling,
normally each for one line.  So it is possible that the hardware error
report comes from different generic hardware error source will be
interleaved.

In this patch, a sequence number is prefixed to each line of error
report.  So that, even if they are interleaved, they still can be
distinguished by the prefixed sequence number.

Signed-off-by: Huang Ying <ying.huang@intel.com>
---
 drivers/acpi/apei/ghes.c |   16 +++++++++++-----
 1 file changed, 11 insertions(+), 5 deletions(-)

--- a/drivers/acpi/apei/ghes.c
+++ b/drivers/acpi/apei/ghes.c
@@ -506,16 +506,22 @@ static void __ghes_print_estatus(const c
 				 const struct acpi_hest_generic *generic,
 				 const struct acpi_hest_generic_status *estatus)
 {
+	static atomic_t seqno;
+	unsigned int curr_seqno;
+	char pfx_seq[64];
+
 	if (pfx == NULL) {
 		if (ghes_severity(estatus->error_severity) <=
 		    GHES_SEV_CORRECTED)
-			pfx = KERN_WARNING HW_ERR;
+			pfx = KERN_WARNING;
 		else
-			pfx = KERN_ERR HW_ERR;
+			pfx = KERN_ERR;
 	}
+	curr_seqno = atomic_inc_return(&seqno);
+	snprintf(pfx_seq, sizeof(pfx_seq), "%s{%u}" HW_ERR, pfx, curr_seqno);
 	printk("%s""Hardware error from APEI Generic Hardware Error Source: %d\n",
-	       pfx, generic->header.source_id);
-	apei_estatus_print(pfx, estatus);
+	       pfx_seq, generic->header.source_id);
+	apei_estatus_print(pfx_seq, estatus);
 }
 
 static int ghes_print_estatus(const char *pfx,
@@ -798,7 +804,7 @@ static int ghes_notify_nmi(unsigned int
 
 	if (sev_global >= GHES_SEV_PANIC) {
 		oops_begin();
-		__ghes_print_estatus(KERN_EMERG HW_ERR, ghes_global->generic,
+		__ghes_print_estatus(KERN_EMERG, ghes_global->generic,
 				     ghes_global->estatus);
 		/* reboot to log the error! */
 		if (panic_timeout == 0)

[PATCH 06/11] ACPI, APEI, Printk queued error record before panic

[Huang Ying]

Because printk is not safe inside NMI handler, the recoverable error
records received in NMI handler will be queued to be printked in a
delayed IRQ context via irq_work.  If a fatal error occurs after the
recoverable error and before the irq_work processed, we lost a error
report.

To solve the issue, the queued error records are printked in NMI
handler if system will go panic.

Signed-off-by: Huang Ying <ying.huang@intel.com>
---
 drivers/acpi/apei/ghes.c |   53 +++++++++++++++++++++++++++++++++++++++--------
 1 file changed, 44 insertions(+), 9 deletions(-)

--- a/drivers/acpi/apei/ghes.c
+++ b/drivers/acpi/apei/ghes.c
@@ -740,26 +740,34 @@ static int ghes_notify_sci(struct notifi
 	return ret;
 }
 
+static struct llist_node *llist_nodes_reverse(struct llist_node *llnode)
+{
+	struct llist_node *next, *tail = NULL;
+
+	while (llnode) {
+		next = llnode->next;
+		llnode->next = tail;
+		tail = llnode;
+		llnode = next;
+	}
+
+	return tail;
+}
+
 static void ghes_proc_in_irq(struct irq_work *irq_work)
 {
-	struct llist_node *llnode, *next, *tail = NULL;
+	struct llist_node *llnode, *next;
 	struct ghes_estatus_node *estatus_node;
 	struct acpi_hest_generic *generic;
 	struct acpi_hest_generic_status *estatus;
 	u32 len, node_len;
 
+	llnode = llist_del_all(&ghes_estatus_llist);
 	/*
 	 * Because the time order of estatus in list is reversed,
 	 * revert it back to proper order.
 	 */
-	llnode = llist_del_all(&ghes_estatus_llist);
-	while (llnode) {
-		next = llnode->next;
-		llnode->next = tail;
-		tail = llnode;
-		llnode = next;
-	}
-	llnode = tail;
+	llnode = llist_nodes_reverse(llnode);
 	while (llnode) {
 		next = llnode->next;
 		estatus_node = llist_entry(llnode, struct ghes_estatus_node,
@@ -779,6 +787,32 @@ static void ghes_proc_in_irq(struct irq_
 	}
 }
 
+static void ghes_print_queued_estatus(void)
+{
+	struct llist_node *llnode;
+	struct ghes_estatus_node *estatus_node;
+	struct acpi_hest_generic *generic;
+	struct acpi_hest_generic_status *estatus;
+	u32 len, node_len;
+
+	llnode = llist_del_all(&ghes_estatus_llist);
+	/*
+	 * Because the time order of estatus in list is reversed,
+	 * revert it back to proper order.
+	 */
+	llnode = llist_nodes_reverse(llnode);
+	while (llnode) {
+		estatus_node = llist_entry(llnode, struct ghes_estatus_node,
+					   llnode);
+		estatus = GHES_ESTATUS_FROM_NODE(estatus_node);
+		len = apei_estatus_len(estatus);
+		node_len = GHES_ESTATUS_NODE_LEN(len);
+		generic = estatus_node->generic;
+		ghes_print_estatus(NULL, generic, estatus);
+		llnode = llnode->next;
+	}
+}
+
 static int ghes_notify_nmi(unsigned int cmd, struct pt_regs *regs)
 {
 	struct ghes *ghes, *ghes_global = NULL;
@@ -804,6 +838,7 @@ static int ghes_notify_nmi(unsigned int
 
 	if (sev_global >= GHES_SEV_PANIC) {
 		oops_begin();
+		ghes_print_queued_estatus();
 		__ghes_print_estatus(KERN_EMERG, ghes_global->generic,
 				     ghes_global->estatus);
 		/* reboot to log the error! */

[PATCH 07/11] ACPI, Add RAM mapping support to ACPI atomic IO support

[Huang Ying]

On one of our testing machine, the following EINJ command lines:

  # echo 0x10000000 > param1
  # echo 0xfffffffffffff000 > param2
  # echo 0x8 > error_type
  # echo 1 > error_inject

Will get:

  echo: write error: Input/output error

The EIO comes from:

    rc = apei_exec_pre_map_gars(&trigger_ctx);

The root cause is as follow.  Normally, ACPI atomic IO support is used
to access IO memory.  But in EINJ of that machine, it is used to
access RAM to trigger the injected error.  And the ioremap() called by
apei_exec_pre_map_gars() can not map the RAM.

This patch add RAM mapping support to ACPI atomic IO support to
satisfy EINJ requirement.

Signed-off-by: Huang Ying <ying.huang@intel.com>
Tested-by: Tony Luck <tony.luck@intel.com>
---
 drivers/acpi/atomicio.c |   41 +++++++++++++++++++++++++++++++++++++----
 1 file changed, 37 insertions(+), 4 deletions(-)

--- a/drivers/acpi/atomicio.c
+++ b/drivers/acpi/atomicio.c
@@ -32,6 +32,8 @@
 #include <linux/rculist.h>
 #include <linux/interrupt.h>
 #include <linux/slab.h>
+#include <linux/mm.h>
+#include <linux/highmem.h>
 #include <acpi/atomicio.h>
 
 #define ACPI_PFX "ACPI: "
@@ -97,6 +99,37 @@ static void __iomem *__acpi_try_ioremap(
 		return NULL;
 }
 
+#ifndef CONFIG_IA64
+#define should_use_kmap(pfn)	page_is_ram(pfn)
+#else
+/* ioremap will take care of cache attributes */
+#define should_use_kmap(pfn)	0
+#endif
+
+static void __iomem *acpi_map(phys_addr_t pg_off, unsigned long pg_sz)
+{
+	unsigned long pfn;
+
+	pfn = pg_off >> PAGE_SHIFT;
+	if (should_use_kmap(pfn)) {
+		if (pg_sz > PAGE_SIZE)
+			return NULL;
+		return (void __iomem __force *)kmap(pfn_to_page(pfn));
+	} else
+		return ioremap(pg_off, pg_sz);
+}
+
+static void acpi_unmap(phys_addr_t pg_off, void __iomem *vaddr)
+{
+	unsigned long pfn;
+
+	pfn = pg_off >> PAGE_SHIFT;
+	if (page_is_ram(pfn))
+		kunmap(pfn_to_page(pfn));
+	else
+		iounmap(vaddr);
+}
+
 /*
  * Used to pre-map the specified IO memory area. First try to find
  * whether the area is already pre-mapped, if it is, increase the
@@ -119,7 +152,7 @@ static void __iomem *acpi_pre_map(phys_a
 
 	pg_off = paddr & PAGE_MASK;
 	pg_sz = ((paddr + size + PAGE_SIZE - 1) & PAGE_MASK) - pg_off;
-	vaddr = ioremap(pg_off, pg_sz);
+	vaddr = acpi_map(pg_off, pg_sz);
 	if (!vaddr)
 		return NULL;
 	map = kmalloc(sizeof(*map), GFP_KERNEL);
@@ -135,7 +168,7 @@ static void __iomem *acpi_pre_map(phys_a
 	vaddr = __acpi_try_ioremap(paddr, size);
 	if (vaddr) {
 		spin_unlock_irqrestore(&acpi_iomaps_lock, flags);
-		iounmap(map->vaddr);
+		acpi_unmap(pg_off, map->vaddr);
 		kfree(map);
 		return vaddr;
 	}
@@ -144,7 +177,7 @@ static void __iomem *acpi_pre_map(phys_a
 
 	return map->vaddr + (paddr - map->paddr);
 err_unmap:
-	iounmap(vaddr);
+	acpi_unmap(pg_off, vaddr);
 	return NULL;
 }
 
@@ -177,7 +210,7 @@ static void acpi_post_unmap(phys_addr_t
 		return;
 
 	synchronize_rcu();
-	iounmap(map->vaddr);
+	acpi_unmap(map->paddr, map->vaddr);
 	kfree(map);
 }
 

[PATCH 08/11] ACPI, APEI, EINJ, Fix resource conflict on some machine

[Huang Ying]

Some APEI firmware implementation will access injected address
specified in param1 to trigger the error when injecting memory error.
This will cause resource conflict with RAM.

On one of our testing machine, if injecting at memory address
0x10000000, the following error will be reported in dmesg:

  APEI: Can not request iomem region <0000000010000000-0000000010000008> for GARs.

This patch removes the injecting memory address range from trigger
table resources to avoid conflict.

Signed-off-by: Huang Ying <ying.huang@intel.com>
Tested-by: Tony Luck <tony.luck@intel.com>
---
 drivers/acpi/apei/apei-base.c     |   11 +++++++++++
 drivers/acpi/apei/apei-internal.h |    3 +++
 drivers/acpi/apei/einj.c          |   24 ++++++++++++++++++++++--
 3 files changed, 36 insertions(+), 2 deletions(-)

--- a/drivers/acpi/apei/apei-base.c
+++ b/drivers/acpi/apei/apei-base.c
@@ -421,6 +421,17 @@ static int apei_resources_merge(struct a
 	return 0;
 }
 
+int apei_resources_add(struct apei_resources *resources,
+		       unsigned long start, unsigned long size,
+		       bool iomem)
+{
+	if (iomem)
+		return apei_res_add(&resources->iomem, start, size);
+	else
+		return apei_res_add(&resources->ioport, start, size);
+}
+EXPORT_SYMBOL_GPL(apei_resources_add);
+
 /*
  * EINJ has two groups of GARs (EINJ table entry and trigger table
  * entry), so common resources are subtracted from the trigger table
--- a/drivers/acpi/apei/apei-internal.h
+++ b/drivers/acpi/apei/apei-internal.h
@@ -95,6 +95,9 @@ static inline void apei_resources_init(s
 }
 
 void apei_resources_fini(struct apei_resources *resources);
+int apei_resources_add(struct apei_resources *resources,
+		       unsigned long start, unsigned long size,
+		       bool iomem);
 int apei_resources_sub(struct apei_resources *resources1,
 		       struct apei_resources *resources2);
 int apei_resources_request(struct apei_resources *resources,
--- a/drivers/acpi/apei/einj.c
+++ b/drivers/acpi/apei/einj.c
@@ -195,7 +195,8 @@ static int einj_check_trigger_header(str
 }
 
 /* Execute instructions in trigger error action table */
-static int __einj_error_trigger(u64 trigger_paddr)
+static int __einj_error_trigger(u64 trigger_paddr, u32 type,
+				u64 param1, u64 param2)
 {
 	struct acpi_einj_trigger *trigger_tab = NULL;
 	struct apei_exec_context trigger_ctx;
@@ -256,6 +257,25 @@ static int __einj_error_trigger(u64 trig
 	rc = apei_resources_sub(&trigger_resources, &einj_resources);
 	if (rc)
 		goto out_fini;
+	/*
+	 * Some firmware will access target address specified in
+	 * param1 to trigger the error when injecting memory error.
+	 * This will cause resource conflict with regular memory.  So
+	 * remove it from trigger table resources.
+	 */
+	if (param_extension && (type & 0x0038) && param2) {
+		struct apei_resources addr_resources;
+		apei_resources_init(&addr_resources);
+		rc = apei_resources_add(&addr_resources,
+					param1 & param2,
+					~param2 + 1, true);
+		if (rc)
+			goto out_fini;
+		rc = apei_resources_sub(&trigger_resources, &addr_resources);
+		apei_resources_fini(&addr_resources);
+		if (rc)
+			goto out_fini;
+	}
 	rc = apei_resources_request(&trigger_resources, "APEI EINJ Trigger");
 	if (rc)
 		goto out_fini;
@@ -325,7 +345,7 @@ static int __einj_error_inject(u32 type,
 	if (rc)
 		return rc;
 	trigger_paddr = apei_exec_ctx_get_output(&ctx);
-	rc = __einj_error_trigger(trigger_paddr);
+	rc = __einj_error_trigger(trigger_paddr, type, param1, param2);
 	if (rc)
 		return rc;
 	rc = apei_exec_run_optional(&ctx, ACPI_EINJ_END_OPERATION);

[PATCH 09/11] ACPI, APEI, EINJ, Refine the fix of resource conflict

[Huang Ying]

From: "Xiao, Hui" <hui.xiao@linux.intel.com>

Current fix for resource conflict is to remove the address region <param1 &
param2, ~param2+1> from trigger resource, which is highly relies on valid user
input. This patch is trying to avoid such potential issues by fetching the
exact address region from trigger action table entry.

Signed-off-by: Xiao, Hui <hui.xiao@linux.intel.com>
Signed-off-by: Huang Ying <ying.huang@intel.com>
---
 drivers/acpi/apei/einj.c |   38 ++++++++++++++++++++++++++++++++------
 1 file changed, 32 insertions(+), 6 deletions(-)

--- a/drivers/acpi/apei/einj.c
+++ b/drivers/acpi/apei/einj.c
@@ -194,6 +194,26 @@ static int einj_check_trigger_header(str
 	return 0;
 }
 
+static struct acpi_generic_address *einj_get_trigger_parameter_region(
+	struct acpi_einj_trigger *trigger_tab, u64 param1, u64 param2)
+{
+	int i;
+	struct acpi_whea_header *entry;
+
+	entry = (struct acpi_whea_header *)
+		((char *)trigger_tab + sizeof(struct acpi_einj_trigger));
+	for (i = 0; i < trigger_tab->entry_count; i++) {
+		if (entry->action == ACPI_EINJ_TRIGGER_ERROR &&
+		entry->instruction == ACPI_EINJ_WRITE_REGISTER_VALUE &&
+		entry->register_region.space_id ==
+			ACPI_ADR_SPACE_SYSTEM_MEMORY &&
+		(entry->register_region.address & param2) == (param1 & param2))
+			return &entry->register_region;
+		entry++;
+	}
+
+	return NULL;
+}
 /* Execute instructions in trigger error action table */
 static int __einj_error_trigger(u64 trigger_paddr, u32 type,
 				u64 param1, u64 param2)
@@ -205,6 +225,7 @@ static int __einj_error_trigger(u64 trig
 	struct resource *r;
 	u32 table_size;
 	int rc = -EIO;
+	struct acpi_generic_address *trigger_param_region = NULL;
 
 	r = request_mem_region(trigger_paddr, sizeof(*trigger_tab),
 			       "APEI EINJ Trigger Table");
@@ -266,12 +287,17 @@ static int __einj_error_trigger(u64 trig
 	if (param_extension && (type & 0x0038) && param2) {
 		struct apei_resources addr_resources;
 		apei_resources_init(&addr_resources);
-		rc = apei_resources_add(&addr_resources,
-					param1 & param2,
-					~param2 + 1, true);
-		if (rc)
-			goto out_fini;
-		rc = apei_resources_sub(&trigger_resources, &addr_resources);
+		trigger_param_region = einj_get_trigger_parameter_region(
+			trigger_tab, param1, param2);
+		if (trigger_param_region) {
+			rc = apei_resources_add(&addr_resources,
+				trigger_param_region->address,
+				trigger_param_region->bit_width/8, true);
+			if (rc)
+				goto out_fini;
+			rc = apei_resources_sub(&trigger_resources,
+					&addr_resources);
+		}
 		apei_resources_fini(&addr_resources);
 		if (rc)
 			goto out_fini;

[PATCH 10/11] ACPI, Record ACPI NVS regions

[Huang Ying]

Some firmware will access memory in ACPI NVS region via APEI.  That
is, instructions in APEI ERST/EINJ table will read/write ACPI NVS
region.  The original resource conflict checking in APEI code will
check memory/ioport accessed by APEI via general resource management
mechanism.  But ACPI NVS region is marked as busy already, so that the
false resource conflict will prevent APEI ERST/EINJ to work.

To fix this, this patch record ACPI NVS regions, so that we can avoid
request resources for memory region inside it.

Signed-off-by: Huang Ying <ying.huang@intel.com>
---
 arch/x86/kernel/e820.c |    4 +--
 drivers/acpi/Makefile  |    3 +-
 drivers/acpi/nvs.c     |   53 ++++++++++++++++++++++++++++++++++++++++++++++++-
 include/linux/acpi.h   |   20 ++++++++++++------
 4 files changed, 70 insertions(+), 10 deletions(-)

--- a/arch/x86/kernel/e820.c
+++ b/arch/x86/kernel/e820.c
@@ -714,7 +714,7 @@ void __init e820_mark_nosave_regions(uns
 }
 #endif
 
-#ifdef CONFIG_HIBERNATION
+#ifdef CONFIG_ACPI
 /**
  * Mark ACPI NVS memory region, so that we can save/restore it during
  * hibernation and the subsequent resume.
@@ -727,7 +727,7 @@ static int __init e820_mark_nvs_memory(v
 		struct e820entry *ei = &e820.map[i];
 
 		if (ei->type == E820_NVS)
-			suspend_nvs_register(ei->addr, ei->size);
+			acpi_nvs_register(ei->addr, ei->size);
 	}
 
 	return 0;
--- a/drivers/acpi/Makefile
+++ b/drivers/acpi/Makefile
@@ -20,11 +20,12 @@ obj-y				+= acpi.o \
 # All the builtin files are in the "acpi." module_param namespace.
 acpi-y				+= osl.o utils.o reboot.o
 acpi-y				+= atomicio.o
+acpi-y				+= nvs.o
 
 # sleep related files
 acpi-y				+= wakeup.o
 acpi-y				+= sleep.o
-acpi-$(CONFIG_ACPI_SLEEP)	+= proc.o nvs.o
+acpi-$(CONFIG_ACPI_SLEEP)	+= proc.o
 
 
 #
--- a/drivers/acpi/nvs.c
+++ b/drivers/acpi/nvs.c
@@ -15,6 +15,56 @@
 #include <linux/acpi_io.h>
 #include <acpi/acpiosxf.h>
 
+/* ACPI NVS regions, APEI may use it */
+
+struct nvs_region {
+	__u64 phys_start;
+	__u64 size;
+	struct list_head node;
+};
+
+static LIST_HEAD(nvs_region_list);
+
+#ifdef CONFIG_ACPI_SLEEP
+static int suspend_nvs_register(unsigned long start, unsigned long size);
+#else
+static inline int suspend_nvs_register(unsigned long a, unsigned long b)
+{
+	return 0;
+}
+#endif
+
+int acpi_nvs_register(__u64 start, __u64 size)
+{
+	struct nvs_region *region;
+
+	region = kmalloc(sizeof(*region), GFP_KERNEL);
+	if (!region)
+		return -ENOMEM;
+	region->phys_start = start;
+	region->size = size;
+	list_add_tail(&region->node, &nvs_region_list);
+
+	return suspend_nvs_register(start, size);
+}
+
+int acpi_nvs_for_each_region(int (*func)(__u64 start, __u64 size, void *data),
+			     void *data)
+{
+	int rc;
+	struct nvs_region *region;
+
+	list_for_each_entry(region, &nvs_region_list, node) {
+		rc = func(region->phys_start, region->size, data);
+		if (rc)
+			return rc;
+	}
+
+	return 0;
+}
+
+
+#ifdef CONFIG_ACPI_SLEEP
 /*
  * Platforms, like ACPI, may want us to save some memory used by them during
  * suspend and to restore the contents of this memory during the subsequent
@@ -41,7 +91,7 @@ static LIST_HEAD(nvs_list);
  *	things so that the data from page-aligned addresses in this region will
  *	be copied into separate RAM pages.
  */
-int suspend_nvs_register(unsigned long start, unsigned long size)
+static int suspend_nvs_register(unsigned long start, unsigned long size)
 {
 	struct nvs_page *entry, *next;
 
@@ -159,3 +209,4 @@ void suspend_nvs_restore(void)
 		if (entry->data)
 			memcpy(entry->kaddr, entry->data, entry->size);
 }
+#endif
--- a/include/linux/acpi.h
+++ b/include/linux/acpi.h
@@ -306,6 +306,11 @@ extern acpi_status acpi_pci_osc_control_
 					     u32 *mask, u32 req);
 extern void acpi_early_init(void);
 
+extern int acpi_nvs_register(__u64 start, __u64 size);
+
+extern int acpi_nvs_for_each_region(int (*func)(__u64, __u64, void *),
+				    void *data);
+
 #else	/* !CONFIG_ACPI */
 
 #define acpi_disabled 1
@@ -348,15 +353,18 @@ static inline int acpi_table_parse(char
 {
 	return -1;
 }
-#endif	/* !CONFIG_ACPI */
 
-#ifdef CONFIG_ACPI_SLEEP
-int suspend_nvs_register(unsigned long start, unsigned long size);
-#else
-static inline int suspend_nvs_register(unsigned long a, unsigned long b)
+static inline int acpi_nvs_register(__u64 start, __u64 size)
 {
 	return 0;
 }
-#endif
+
+static inline int acpi_nvs_for_each_region(int (*func)(__u64, __u64, void *),
+					   void *data)
+{
+	return 0;
+}
+
+#endif	/* !CONFIG_ACPI */
 
 #endif	/*_LINUX_ACPI_H*/

[PATCH 11/11] ACPI, APEI, Resolve false conflict between ACPI NVS and APEI

[Huang Ying]

Some firmware will access memory in ACPI NVS region via APEI.  That
is, instructions in APEI ERST/EINJ table will read/write ACPI NVS
region.  The original resource conflict checking in APEI code will
check memory/ioport accessed by APEI via general resource management
mech.  But ACPI NVS region is marked as busy already, so that the
false resource conflict will prevent APEI ERST/EINJ to work.

To fix this, this patch excludes ACPI NVS regions when APEI components
request resources.  So that they will not conflict with ACPI NVS
regions.

Reported-and-tested-by: Pavel Ivanov <paivanof@gmail.com>
Signed-off-by: Huang Ying <ying.huang@intel.com>
---
 drivers/acpi/apei/apei-base.c |   29 ++++++++++++++++++++++++++++-
 1 file changed, 28 insertions(+), 1 deletion(-)

--- a/drivers/acpi/apei/apei-base.c
+++ b/drivers/acpi/apei/apei-base.c
@@ -449,8 +449,19 @@ int apei_resources_sub(struct apei_resou
 }
 EXPORT_SYMBOL_GPL(apei_resources_sub);
 
+static int apei_get_nvs_callback(__u64 start, __u64 size, void *data)
+{
+	struct apei_resources *resources = data;
+	return apei_res_add(&resources->iomem, start, size);
+}
+
+static int apei_get_nvs_resources(struct apei_resources *resources)
+{
+	return acpi_nvs_for_each_region(apei_get_nvs_callback, resources);
+}
+
 /*
- * IO memory/port rersource management mechanism is used to check
+ * IO memory/port resource management mechanism is used to check
  * whether memory/port area used by GARs conflicts with normal memory
  * or IO memory/port of devices.
  */
@@ -459,12 +470,26 @@ int apei_resources_request(struct apei_r
 {
 	struct apei_res *res, *res_bak = NULL;
 	struct resource *r;
+	struct apei_resources nvs_resources;
 	int rc;
 
 	rc = apei_resources_sub(resources, &apei_resources_all);
 	if (rc)
 		return rc;
 
+	/*
+	 * Some firmware uses ACPI NVS region, that has been marked as
+	 * busy, so exclude it from APEI resources to avoid false
+	 * conflict.
+	 */
+	apei_resources_init(&nvs_resources);
+	rc = apei_get_nvs_resources(&nvs_resources);
+	if (rc)
+		goto res_fini;
+	rc = apei_resources_sub(resources, &nvs_resources);
+	if (rc)
+		goto res_fini;
+
 	rc = -EINVAL;
 	list_for_each_entry(res, &resources->iomem, list) {
 		r = request_mem_region(res->start, res->end - res->start,
@@ -511,6 +536,8 @@ err_unmap_iomem:
 			break;
 		release_mem_region(res->start, res->end - res->start);
 	}
+res_fini:
+	apei_resources_fini(&nvs_resources);
 	return rc;
 }
 EXPORT_SYMBOL_GPL(apei_resources_request);