1 files changed, 196 insertions, 59 deletions

diff --git a/arch/powerpc/platforms/pseries/nvram.c b/arch/powerpc/platforms/pseries/nvram.c
index 7e828ba29bc..419707b0724 100644
--- a/arch/powerpc/platforms/pseries/nvram.c
+++ b/arch/powerpc/platforms/pseries/nvram.c
@@ -16,6 +16,8 @@
 #include <linux/errno.h>
 #include <linux/init.h>
 #include <linux/spinlock.h>
+#include <linux/slab.h>
+#include <linux/kmsg_dump.h>
 #include <asm/uaccess.h>
 #include <asm/nvram.h>
 #include <asm/rtas.h>
@@ -30,17 +32,54 @@ static int nvram_fetch, nvram_store;
 static char nvram_buf[NVRW_CNT];	/* assume this is in the first 4GB */
 static DEFINE_SPINLOCK(nvram_lock);
 
-static long nvram_error_log_index = -1;
-static long nvram_error_log_size = 0;
-
 struct err_log_info {
 	int error_type;
 	unsigned int seq_num;
 };
-#define NVRAM_MAX_REQ		2079
-#define NVRAM_MIN_REQ		1055
 
-#define NVRAM_LOG_PART_NAME	"ibm,rtas-log"
+struct nvram_os_partition {
+	const char *name;
+	int req_size;	/* desired size, in bytes */
+	int min_size;	/* minimum acceptable size (0 means req_size) */
+	long size;	/* size of data portion (excluding err_log_info) */
+	long index;	/* offset of data portion of partition */
+};
+
+static struct nvram_os_partition rtas_log_partition = {
+	.name = "ibm,rtas-log",
+	.req_size = 2079,
+	.min_size = 1055,
+	.index = -1
+};
+
+static struct nvram_os_partition oops_log_partition = {
+	.name = "lnx,oops-log",
+	.req_size = 4000,
+	.min_size = 2000,
+	.index = -1
+};
+
+static const char *pseries_nvram_os_partitions[] = {
+	"ibm,rtas-log",
+	"lnx,oops-log",
+	NULL
+};
+
+static void oops_to_nvram(struct kmsg_dumper *dumper,
+		enum kmsg_dump_reason reason,
+		const char *old_msgs, unsigned long old_len,
+		const char *new_msgs, unsigned long new_len);
+
+static struct kmsg_dumper nvram_kmsg_dumper = {
+	.dump = oops_to_nvram
+};
+
+/* See clobbering_unread_rtas_event() */
+#define NVRAM_RTAS_READ_TIMEOUT 5		/* seconds */
+static unsigned long last_unread_rtas_event;	/* timestamp */
+
+/* We preallocate oops_buf during init to avoid kmalloc during oops/panic. */
+static char *oops_buf;
 
 static ssize_t pSeries_nvram_read(char *buf, size_t count, loff_t *index)
 {
@@ -134,7 +173,7 @@ static ssize_t pSeries_nvram_get_size(void)
 }
 
 
-/* nvram_write_error_log
+/* nvram_write_os_partition, nvram_write_error_log
  *
  * We need to buffer the error logs into nvram to ensure that we have
  * the failure information to decode.  If we have a severe error there
@@ -156,48 +195,58 @@ static ssize_t pSeries_nvram_get_size(void)
  * The 'data' section would look like (in bytes):
  * +--------------+------------+-----------------------------------+
  * | event_logged | sequence # | error log                         |
- * |0            3|4          7|8            nvram_error_log_size-1|
+ * |0            3|4          7|8                  error_log_size-1|
  * +--------------+------------+-----------------------------------+
  *
  * event_logged: 0 if event has not been logged to syslog, 1 if it has
  * sequence #: The unique sequence # for each event. (until it wraps)
  * error log: The error log from event_scan
  */
-int nvram_write_error_log(char * buff, int length,
-                          unsigned int err_type, unsigned int error_log_cnt)
+int nvram_write_os_partition(struct nvram_os_partition *part, char * buff,
+		int length, unsigned int err_type, unsigned int error_log_cnt)
 {
 	int rc;
 	loff_t tmp_index;
 	struct err_log_info info;
 	
-	if (nvram_error_log_index == -1) {
+	if (part->index == -1) {
 		return -ESPIPE;
 	}
 
-	if (length > nvram_error_log_size) {
-		length = nvram_error_log_size;
+	if (length > part->size) {
+		length = part->size;
 	}
 
 	info.error_type = err_type;
 	info.seq_num = error_log_cnt;
 
-	tmp_index = nvram_error_log_index;
+	tmp_index = part->index;
 
 	rc = ppc_md.nvram_write((char *)&info, sizeof(struct err_log_info), &tmp_index);
 	if (rc <= 0) {
-		printk(KERN_ERR "nvram_write_error_log: Failed nvram_write (%d)\n", rc);
+		pr_err("%s: Failed nvram_write (%d)\n", __FUNCTION__, rc);
 		return rc;
 	}
 
 	rc = ppc_md.nvram_write(buff, length, &tmp_index);
 	if (rc <= 0) {
-		printk(KERN_ERR "nvram_write_error_log: Failed nvram_write (%d)\n", rc);
+		pr_err("%s: Failed nvram_write (%d)\n", __FUNCTION__, rc);
 		return rc;
 	}
 	
 	return 0;
 }
 
+int nvram_write_error_log(char * buff, int length,
+                          unsigned int err_type, unsigned int error_log_cnt)
+{
+	int rc = nvram_write_os_partition(&rtas_log_partition, buff, length,
+						err_type, error_log_cnt);
+	if (!rc)
+		last_unread_rtas_event = get_seconds();
+	return rc;
+}
+
 /* nvram_read_error_log
  *
  * Reads nvram for error log for at most 'length'
@@ -209,13 +258,13 @@ int nvram_read_error_log(char * buff, int length,
 	loff_t tmp_index;
 	struct err_log_info info;
 	
-	if (nvram_error_log_index == -1)
+	if (rtas_log_partition.index == -1)
 		return -1;
 
-	if (length > nvram_error_log_size)
-		length = nvram_error_log_size;
+	if (length > rtas_log_partition.size)
+		length = rtas_log_partition.size;
 
-	tmp_index = nvram_error_log_index;
+	tmp_index = rtas_log_partition.index;
 
 	rc = ppc_md.nvram_read((char *)&info, sizeof(struct err_log_info), &tmp_index);
 	if (rc <= 0) {
@@ -244,37 +293,40 @@ int nvram_clear_error_log(void)
 	int clear_word = ERR_FLAG_ALREADY_LOGGED;
 	int rc;
 
-	if (nvram_error_log_index == -1)
+	if (rtas_log_partition.index == -1)
 		return -1;
 
-	tmp_index = nvram_error_log_index;
+	tmp_index = rtas_log_partition.index;
 	
 	rc = ppc_md.nvram_write((char *)&clear_word, sizeof(int), &tmp_index);
 	if (rc <= 0) {
 		printk(KERN_ERR "nvram_clear_error_log: Failed nvram_write (%d)\n", rc);
 		return rc;
 	}
+	last_unread_rtas_event = 0;
 
 	return 0;
 }
 
-/* pseries_nvram_init_log_partition
+/* pseries_nvram_init_os_partition
  *
- * This will setup the partition we need for buffering the
- * error logs and cleanup partitions if needed.
+ * This sets up a partition with an "OS" signature.
  *
  * The general strategy is the following:
- * 1.) If there is log partition large enough then use it.
- * 2.) If there is none large enough, search
- * for a free partition that is large enough.
- * 3.) If there is not a free partition large enough remove 
- * _all_ OS partitions and consolidate the space.
- * 4.) Will first try getting a chunk that will satisfy the maximum
- * error log size (NVRAM_MAX_REQ).
- * 5.) If the max chunk cannot be allocated then try finding a chunk
- * that will satisfy the minum needed (NVRAM_MIN_REQ).
+ * 1.) If a partition with the indicated name already exists...
+ *	- If it's large enough, use it.
+ *	- Otherwise, recycle it and keep going.
+ * 2.) Search for a free partition that is large enough.
+ * 3.) If there's not a free partition large enough, recycle any obsolete
+ * OS partitions and try again.
+ * 4.) Will first try getting a chunk that will satisfy the requested size.
+ * 5.) If a chunk of the requested size cannot be allocated, then try finding
+ * a chunk that will satisfy the minum needed.
+ *
+ * Returns 0 on success, else -1.
  */
-static int __init pseries_nvram_init_log_partition(void)
+static int __init pseries_nvram_init_os_partition(struct nvram_os_partition
+									*part)
 {
 	loff_t p;
 	int size;
@@ -282,47 +334,76 @@ static int __init pseries_nvram_init_log_partition(void)
 	/* Scan nvram for partitions */
 	nvram_scan_partitions();
 
-	/* Lookg for ours */
-	p = nvram_find_partition(NVRAM_LOG_PART_NAME, NVRAM_SIG_OS, &size);
+	/* Look for ours */
+	p = nvram_find_partition(part->name, NVRAM_SIG_OS, &size);
 
 	/* Found one but too small, remove it */
-	if (p && size < NVRAM_MIN_REQ) {
-		pr_info("nvram: Found too small "NVRAM_LOG_PART_NAME" partition"
-			",removing it...");
-		nvram_remove_partition(NVRAM_LOG_PART_NAME, NVRAM_SIG_OS);
+	if (p && size < part->min_size) {
+		pr_info("nvram: Found too small %s partition,"
+					" removing it...\n", part->name);
+		nvram_remove_partition(part->name, NVRAM_SIG_OS, NULL);
 		p = 0;
 	}
 
 	/* Create one if we didn't find */
 	if (!p) {
-		p = nvram_create_partition(NVRAM_LOG_PART_NAME, NVRAM_SIG_OS,
-					   NVRAM_MAX_REQ, NVRAM_MIN_REQ);
-		/* No room for it, try to get rid of any OS partition
-		 * and try again
-		 */
+		p = nvram_create_partition(part->name, NVRAM_SIG_OS,
+					part->req_size, part->min_size);
 		if (p == -ENOSPC) {
-			pr_info("nvram: No room to create "NVRAM_LOG_PART_NAME
-				" partition, deleting all OS partitions...");
-			nvram_remove_partition(NULL, NVRAM_SIG_OS);
-			p = nvram_create_partition(NVRAM_LOG_PART_NAME,
-						   NVRAM_SIG_OS, NVRAM_MAX_REQ,
-						   NVRAM_MIN_REQ);
+			pr_info("nvram: No room to create %s partition, "
+				"deleting any obsolete OS partitions...\n",
+				part->name);
+			nvram_remove_partition(NULL, NVRAM_SIG_OS,
+						pseries_nvram_os_partitions);
+			p = nvram_create_partition(part->name, NVRAM_SIG_OS,
+					part->req_size, part->min_size);
 		}
 	}
 
 	if (p <= 0) {
-		pr_err("nvram: Failed to find or create "NVRAM_LOG_PART_NAME
-		       " partition, err %d\n", (int)p);
-		return 0;
+		pr_err("nvram: Failed to find or create %s"
+		       " partition, err %d\n", part->name, (int)p);
+		return -1;
 	}
 
-	nvram_error_log_index = p;
-	nvram_error_log_size = nvram_get_partition_size(p) -
-		sizeof(struct err_log_info);
+	part->index = p;
+	part->size = nvram_get_partition_size(p) - sizeof(struct err_log_info);
 	
 	return 0;
 }
-machine_arch_initcall(pseries, pseries_nvram_init_log_partition);
+
+static void __init nvram_init_oops_partition(int rtas_partition_exists)
+{
+	int rc;
+
+	rc = pseries_nvram_init_os_partition(&oops_log_partition);
+	if (rc != 0) {
+		if (!rtas_partition_exists)
+			return;
+		pr_notice("nvram: Using %s partition to log both"
+			" RTAS errors and oops/panic reports\n",
+			rtas_log_partition.name);
+		memcpy(&oops_log_partition, &rtas_log_partition,
+						sizeof(rtas_log_partition));
+	}
+	oops_buf = kmalloc(oops_log_partition.size, GFP_KERNEL);
+	rc = kmsg_dump_register(&nvram_kmsg_dumper);
+	if (rc != 0) {
+		pr_err("nvram: kmsg_dump_register() failed; returned %d\n", rc);
+		kfree(oops_buf);
+		return;
+	}
+}
+
+static int __init pseries_nvram_init_log_partitions(void)
+{
+	int rc;
+
+	rc = pseries_nvram_init_os_partition(&rtas_log_partition);
+	nvram_init_oops_partition(rc == 0);
+	return 0;
+}
+machine_arch_initcall(pseries, pseries_nvram_init_log_partitions);
 
 int __init pSeries_nvram_init(void)
 {
@@ -353,3 +434,59 @@ int __init pSeries_nvram_init(void)
 
 	return 0;
 }
+
+/*
+ * Try to capture the last capture_len bytes of the printk buffer.  Return
+ * the amount actually captured.
+ */
+static size_t capture_last_msgs(const char *old_msgs, size_t old_len,
+				const char *new_msgs, size_t new_len,
+				char *captured, size_t capture_len)
+{
+	if (new_len >= capture_len) {
+		memcpy(captured, new_msgs + (new_len - capture_len),
+								capture_len);
+		return capture_len;
+	} else {
+		/* Grab the end of old_msgs. */
+		size_t old_tail_len = min(old_len, capture_len - new_len);
+		memcpy(captured, old_msgs + (old_len - old_tail_len),
+								old_tail_len);
+		memcpy(captured + old_tail_len, new_msgs, new_len);
+		return old_tail_len + new_len;
+	}
+}
+
+/*
+ * Are we using the ibm,rtas-log for oops/panic reports?  And if so,
+ * would logging this oops/panic overwrite an RTAS event that rtas_errd
+ * hasn't had a chance to read and process?  Return 1 if so, else 0.
+ *
+ * We assume that if rtas_errd hasn't read the RTAS event in
+ * NVRAM_RTAS_READ_TIMEOUT seconds, it's probably not going to.
+ */
+static int clobbering_unread_rtas_event(void)
+{
+	return (oops_log_partition.index == rtas_log_partition.index
+		&& last_unread_rtas_event
+		&& get_seconds() - last_unread_rtas_event <=
+						NVRAM_RTAS_READ_TIMEOUT);
+}
+
+/* our kmsg_dump callback */
+static void oops_to_nvram(struct kmsg_dumper *dumper,
+		enum kmsg_dump_reason reason,
+		const char *old_msgs, unsigned long old_len,
+		const char *new_msgs, unsigned long new_len)
+{
+	static unsigned int oops_count = 0;
+	size_t text_len;
+
+	if (clobbering_unread_rtas_event())
+		return;
+
+	text_len = capture_last_msgs(old_msgs, old_len, new_msgs, new_len,
+					oops_buf, oops_log_partition.size);
+	(void) nvram_write_os_partition(&oops_log_partition, oops_buf,
+		(int) text_len, ERR_TYPE_KERNEL_PANIC, ++oops_count);
+}