pstore/zone,blk: Add ftrace frontend support

Support backend for ftrace. To enable ftrace backend, just make
ftrace_size be greater than 0 and a multiple of 4096.

Signed-off-by: WeiXiong Liao <[email protected]>
Link: https://lore.kernel.org/lkml/[email protected]/
Co-developed-by: Colin Ian King <[email protected]>
Signed-off-by: Colin Ian King <[email protected]>
Link: https://lore.kernel.org/lkml/[email protected]
Signed-off-by: Kees Cook <[email protected]>

Author: WeiXiong Liao

fs/pstore/Kconfig

@@ -248,3 +248,15 @@ config PSTORE_BLK_CONSOLE_SIZE
 
 	  NOTE that, both Kconfig and module parameters can configure
 	  pstore/blk, but module parameters have priority over Kconfig.
+
+config PSTORE_BLK_FTRACE_SIZE
+	int "Size in Kbytes of ftrace log to store"
+	depends on PSTORE_BLK
+	depends on PSTORE_FTRACE
+	default 64
+	help
+	  This just sets size of ftrace log (ftrace_size) for pstore/blk. The
+	  size is in KB and must be a multiple of 4.
+
+	  NOTE that, both Kconfig and module parameters can configure
+	  pstore/blk, but module parameters have priority over Kconfig.

fs/pstore/blk.c

@@ -43,6 +43,14 @@ static long console_size = -1;
 module_param(console_size, long, 0400);
 MODULE_PARM_DESC(console_size, "console size in kbytes");
 
+#if IS_ENABLED(CONFIG_PSTORE_FTRACE)
+static long ftrace_size = CONFIG_PSTORE_BLK_FTRACE_SIZE;
+#else
+static long ftrace_size = -1;
+#endif
+module_param(ftrace_size, long, 0400);
+MODULE_PARM_DESC(ftrace_size, "ftrace size in kbytes");
+
 /*
  * blkdev - the block device to use for pstore storage
  *
@@ -152,6 +160,7 @@ static int psblk_register_do(struct pstore_device_info *dev)
 	verify_size(kmsg_size, 4096, dev->flags & PSTORE_FLAGS_DMESG);
 	verify_size(pmsg_size, 4096, dev->flags & PSTORE_FLAGS_PMSG);
 	verify_size(console_size, 4096, dev->flags & PSTORE_FLAGS_CONSOLE);
+	verify_size(ftrace_size, 4096, dev->flags & PSTORE_FLAGS_FTRACE);
 #undef verify_size
 
 	pstore_zone_info->total_size = dev->total_size;

fs/pstore/zone.c

@@ -92,11 +92,14 @@ struct pstore_zone {
  * @kpszs: kmsg dump storage zones
  * @ppsz: pmsg storage zone
  * @cpsz: console storage zone
+ * @fpszs: ftrace storage zones
  * @kmsg_max_cnt: max count of @kpszs
  * @kmsg_read_cnt: counter of total read kmsg dumps
  * @kmsg_write_cnt: counter of total kmsg dump writes
  * @pmsg_read_cnt: counter of total read pmsg zone
  * @console_read_cnt: counter of total read console zone
+ * @ftrace_max_cnt: max count of @fpszs
+ * @ftrace_read_cnt: counter of max read ftrace zone
  * @oops_counter: counter of oops dumps
  * @panic_counter: counter of panic dumps
  * @recovered: whether finished recovering data from storage
@@ -109,11 +112,14 @@ struct psz_context {
 	struct pstore_zone **kpszs;
 	struct pstore_zone *ppsz;
 	struct pstore_zone *cpsz;
+	struct pstore_zone **fpszs;
 	unsigned int kmsg_max_cnt;
 	unsigned int kmsg_read_cnt;
 	unsigned int kmsg_write_cnt;
 	unsigned int pmsg_read_cnt;
 	unsigned int console_read_cnt;
+	unsigned int ftrace_max_cnt;
+	unsigned int ftrace_read_cnt;
 	/*
 	 * These counters should be calculated during recovery.
 	 * It records the oops/panic times after crashes rather than boots.
@@ -314,6 +320,8 @@ static void psz_flush_all_dirty_zones(struct work_struct *work)
 		ret |= psz_flush_dirty_zone(cxt->cpsz);
 	if (cxt->kpszs)
 		ret |= psz_flush_dirty_zones(cxt->kpszs, cxt->kmsg_max_cnt);
+	if (cxt->fpszs)
+		ret |= psz_flush_dirty_zones(cxt->fpszs, cxt->ftrace_max_cnt);
 	if (ret && cxt->pstore_zone_info)
 		schedule_delayed_work(&psz_cleaner, msecs_to_jiffies(1000));
 }
@@ -550,6 +558,31 @@ static int psz_recover_zone(struct psz_context *cxt, struct pstore_zone *zone)
 	return ret;
 }
 
+static int psz_recover_zones(struct psz_context *cxt,
+		struct pstore_zone **zones, unsigned int cnt)
+{
+	int ret;
+	unsigned int i;
+	struct pstore_zone *zone;
+
+	if (!zones)
+		return 0;
+
+	for (i = 0; i < cnt; i++) {
+		zone = zones[i];
+		if (unlikely(!zone))
+			continue;
+		ret = psz_recover_zone(cxt, zone);
+		if (ret)
+			goto recover_fail;
+	}
+
+	return 0;
+recover_fail:
+	pr_debug("recover %s[%u] failed\n", zone->name, i);
+	return ret;
+}
+
 /**
  * psz_recovery() - recover data from storage
  * @cxt: the context of pstore/zone
@@ -574,6 +607,10 @@ static inline int psz_recovery(struct psz_context *cxt)
 		goto out;
 
 	ret = psz_recover_zone(cxt, cxt->cpsz);
+	if (ret)
+		goto out;
+
+	ret = psz_recover_zones(cxt, cxt->fpszs, cxt->ftrace_max_cnt);
 
 out:
 	if (unlikely(ret))
@@ -592,6 +629,7 @@ static int psz_pstore_open(struct pstore_info *psi)
 	cxt->kmsg_read_cnt = 0;
 	cxt->pmsg_read_cnt = 0;
 	cxt->console_read_cnt = 0;
+	cxt->ftrace_read_cnt = 0;
 	return 0;
 }
 
@@ -658,6 +696,10 @@ static int psz_pstore_erase(struct pstore_record *record)
 		return psz_record_erase(cxt, cxt->ppsz);
 	case PSTORE_TYPE_CONSOLE:
 		return psz_record_erase(cxt, cxt->cpsz);
+	case PSTORE_TYPE_FTRACE:
+		if (record->id >= cxt->ftrace_max_cnt)
+			return -EINVAL;
+		return psz_record_erase(cxt, cxt->fpszs[record->id]);
 	default: return -EINVAL;
 	}
 }
@@ -802,6 +844,13 @@ static int notrace psz_pstore_write(struct pstore_record *record)
 		return psz_record_write(cxt->cpsz, record);
 	case PSTORE_TYPE_PMSG:
 		return psz_record_write(cxt->ppsz, record);
+	case PSTORE_TYPE_FTRACE: {
+		int zonenum = smp_processor_id();
+
+		if (!cxt->fpszs)
+			return -ENOSPC;
+		return psz_record_write(cxt->fpszs[zonenum], record);
+	}
 	default:
 		return -EINVAL;
 	}
@@ -817,6 +866,14 @@ static struct pstore_zone *psz_read_next_zone(struct psz_context *cxt)
 			return zone;
 	}
 
+	if (cxt->ftrace_read_cnt < cxt->ftrace_max_cnt)
+		/*
+		 * No need psz_old_ok(). Let psz_ftrace_read() do so for
+		 * combination. psz_ftrace_read() should traverse over
+		 * all zones in case of some zone without data.
+		 */
+		return cxt->fpszs[cxt->ftrace_read_cnt++];
+
 	if (cxt->pmsg_read_cnt == 0) {
 		cxt->pmsg_read_cnt++;
 		zone = cxt->ppsz;
@@ -891,6 +948,38 @@ static ssize_t psz_kmsg_read(struct pstore_zone *zone,
 	return size + hlen;
 }
 
+/* try to combine all ftrace zones */
+static ssize_t psz_ftrace_read(struct pstore_zone *zone,
+		struct pstore_record *record)
+{
+	struct psz_context *cxt;
+	struct psz_buffer *buf;
+	int ret;
+
+	if (!zone || !record)
+		return -ENOSPC;
+
+	if (!psz_old_ok(zone))
+		goto out;
+
+	buf = (struct psz_buffer *)zone->oldbuf;
+	if (!buf)
+		return -ENOMSG;
+
+	ret = pstore_ftrace_combine_log(&record->buf, &record->size,
+			(char *)buf->data, atomic_read(&buf->datalen));
+	if (unlikely(ret))
+		return ret;
+
+out:
+	cxt = record->psi->data;
+	if (cxt->ftrace_read_cnt < cxt->ftrace_max_cnt)
+		/* then, read next ftrace zone */
+		return -ENOMSG;
+	record->id = 0;
+	return record->size ? record->size : -ENOMSG;
+}
+
 static ssize_t psz_record_read(struct pstore_zone *zone,
 		struct pstore_record *record)
 {
@@ -941,6 +1030,9 @@ static ssize_t psz_pstore_read(struct pstore_record *record)
 		readop = psz_kmsg_read;
 		record->id = cxt->kmsg_read_cnt - 1;
 		break;
+	case PSTORE_TYPE_FTRACE:
+		readop = psz_ftrace_read;
+		break;
 	case PSTORE_TYPE_CONSOLE:
 		fallthrough;
 	case PSTORE_TYPE_PMSG:
@@ -1005,6 +1097,8 @@ static void psz_free_all_zones(struct psz_context *cxt)
 		psz_free_zone(&cxt->ppsz);
 	if (cxt->cpsz)
 		psz_free_zone(&cxt->cpsz);
+	if (cxt->fpszs)
+		psz_free_zones(&cxt->fpszs, &cxt->ftrace_max_cnt);
 }
 
 static struct pstore_zone *psz_init_zone(enum pstore_type_id type,
@@ -1115,6 +1209,17 @@ static int psz_alloc_zones(struct psz_context *cxt)
 		goto free_out;
 	}
 
+	off_size += info->ftrace_size;
+	cxt->fpszs = psz_init_zones(PSTORE_TYPE_FTRACE, &off,
+			info->ftrace_size,
+			info->ftrace_size / nr_cpu_ids,
+			&cxt->ftrace_max_cnt);
+	if (IS_ERR(cxt->fpszs)) {
+		err = PTR_ERR(cxt->fpszs);
+		cxt->fpszs = NULL;
+		goto free_out;
+	}
+
 	cxt->kpszs = psz_init_zones(PSTORE_TYPE_DMESG, &off,
 			info->total_size - off_size,
 			info->kmsg_size, &cxt->kmsg_max_cnt);
@@ -1149,7 +1254,8 @@ int register_pstore_zone(struct pstore_zone_info *info)
 		return -EINVAL;
 	}
 
-	if (!info->kmsg_size && !info->pmsg_size && !info->console_size) {
+	if (!info->kmsg_size && !info->pmsg_size && !info->console_size &&
+	    !info->ftrace_size) {
 		pr_warn("at least one record size must be non-zero\n");
 		return -EINVAL;
 	}
@@ -1173,6 +1279,7 @@ int register_pstore_zone(struct pstore_zone_info *info)
 	check_size(kmsg_size, SECTOR_SIZE);
 	check_size(pmsg_size, SECTOR_SIZE);
 	check_size(console_size, SECTOR_SIZE);
+	check_size(ftrace_size, SECTOR_SIZE);
 
 #undef check_size
 
@@ -1200,6 +1307,7 @@ int register_pstore_zone(struct pstore_zone_info *info)
 	pr_debug("\tkmsg size : %ld Bytes\n", info->kmsg_size);
 	pr_debug("\tpmsg size : %ld Bytes\n", info->pmsg_size);
 	pr_debug("\tconsole size : %ld Bytes\n", info->console_size);
+	pr_debug("\tftrace size : %ld Bytes\n", info->ftrace_size);
 
 	err = psz_alloc_zones(cxt);
 	if (err) {
@@ -1237,6 +1345,10 @@ int register_pstore_zone(struct pstore_zone_info *info)
 		cxt->pstore.flags |= PSTORE_FLAGS_CONSOLE;
 		pr_cont(" console");
 	}
+	if (info->ftrace_size) {
+		cxt->pstore.flags |= PSTORE_FLAGS_FTRACE;
+		pr_cont(" ftrace");
+	}
 	pr_cont("\n");
 
 	err = pstore_register(&cxt->pstore);

include/linux/pstore_zone.h

@@ -19,6 +19,7 @@ typedef ssize_t (*pstore_zone_write_op)(const char *, size_t, loff_t);
  * @max_reason: Maximum kmsg dump reason to store.
  * @pmsg_size:	The size of pmsg zone which is the same as @kmsg_size.
  * @console_size:The size of console zone which is the same as @kmsg_size.
+ * @ftrace_size:The size of ftrace zone which is the same as @kmsg_size.
  * @read:	The general read operation. Both of the function parameters
  *		@size and @offset are relative value to storage.
  *		On success, the number of bytes should be returned, others
@@ -37,6 +38,7 @@ struct pstore_zone_info {
 	int max_reason;
 	unsigned long pmsg_size;
 	unsigned long console_size;
+	unsigned long ftrace_size;
 	pstore_zone_read_op read;
 	pstore_zone_write_op write;
 	pstore_zone_write_op panic_write;