在FLASHDB 中建立多分区多DB运行出现问题

[tianzhilianwo]

朱工：您好！再用FLASH DB做测试时，发现一下问题，通过FAL建立多分多数据库时，KV出现不GC的情况，跟踪了一下，在最后一个GC时 会改变part_falsh_cache的初始化值，导致无法寻找到对应的flashdev 但是，采用两个分区时可以正常运行的；请教一下是什么问题？如何解决？

#define FAL_PART_TABLE
{
{FAL_PART_MAGIC_WROD, "fdb_tsdb0", "norflash0", 0L, 2L1024L1024L, 0},
{FAL_PART_MAGIC_WROD, "fdb_tsdb1", "norflash0", 2L1024L1024L, 2L1024L1024L, 0},
{FAL_PART_MAGIC_WROD, "fdb_tsdb2", "norflash0", 4L1024L1024L, 2L1024L1024L, 0},
{FAL_PART_MAGIC_WROD, "fdb_tsdb3", "norflash0", 6L1024L1024L, 2L1024L1024L, 0},
{FAL_PART_MAGIC_WROD, "fdb_tsdb4", "norflash0", 8L1024L1024L, 2L1024L1024L, 0},
{FAL_PART_MAGIC_WROD, "fdb_tsdb5", "norflash0", 10L1024L1024L, 2L1024L1024L, 0},
{FAL_PART_MAGIC_WROD, "fdb_tsdb6", "norflash0", 12L1024L1024L, 2L1024L1024L, 0},
{FAL_PART_MAGIC_WROD, "fdb_tsdb7", "norflash0", 14L1024L1024L, 2L1024L1024L, 0},
{FAL_PART_MAGIC_WORD, "fdb_Safty", "GD32_onchip", 6001024, 401024, 0},
{FAL_PART_MAGIC_WORD, "fdb_Device", "GD32_onchip", 6401024, 81024, 0},
{FAL_PART_MAGIC_WORD, "fdb_HD00", "GD32_onchip", 6481024, 321024, 0},
{FAL_PART_MAGIC_WORD, "fdb_HD01", "GD32_onchip", 6801024, 321024, 0},
{FAL_PART_MAGIC_WORD, "fdb_ACUM", "GD32_onchip", 7121024, 161024, 0},
{FAL_PART_MAGIC_WORD, "fdb_Systm", "GD32_onchip", 7281024, 161024, 0},
{FAL_PART_MAGIC_WORD, "fdb_Event", "GD32_onchip", 7441024, 401024, 0},
{FAL_PART_MAGIC_WORD, "fdb_CurntE", "GD32_onchip", 7841024, 81024, 0},
}
[10:40:05.641]收←◆[FlashDB][SYTIME] set SysTime mKey=179,0X180A--0X1210--0X001E stamp=1729267230
[FlashDB][SYTIME] Get SysTime mKey=179, 24-10-18 16:0:30

[10:40:05.822]收←◆[FlashDB][SYTIME] set SysTime mKey=180,0X180A--0X1210--0X001E stamp=1729267230
[FlashDB][SYTIME] Get SysTime mKey=180, 24-10-18 16:0:30

[10:40:06.002]收←◆[FlashDB][SYTIME] set SysTime mKey=181,0X180A--0X1210--0X001E stamp=1729267230
[FlashDB][SYTIME] Get SysTime mKey=181, 24-10-18 16:0:30

[10:40:06.182]收←◆[FlashDB][kv][SYTM][fdb_Systm] (..\module\F00302_Fdb\src\fdb_kvdb.c:929) Trigger a GC check after alloc KV failed.
[FlashDB][kv][SYTM][fdb_Systm] (..\module\F00302_Fdb\src\fdb_kvdb.c:1075) Alloc an KV (size 64) failed when new KV. Now will GC then retry.
[FlashDB][kv][SYTM][fdb_Systm] (..\module\F00302_Fdb\src\fdb_kvdb.c:1161) The remain empty sector is 1, GC threshold is 1.
[FlashDB][kv][SYTM][fdb_Systm] (..\module\F00302_Fdb\src\fdb_kvdb.c:1130) KV (SYSTFS) is garbage NOT need move, collect it.
[FlashDB][kv][SYTM][fdb_Systm] (..\module\F00302_Fdb\src\fdb_kvdb.c:1130) KV (SysTime) is garbage NOT need move, collect it.

在KV不断写入到最后一个分区时 ，不进行最后一个分区的GC ？出现以上问题后导致CPU进入硬件故障；

只保留一下两个分区：可以正常进行 GC
#define FAL_PART_TABLE
{
{FAL_PART_MAGIC_WORD, "fdb_Systm", "GD32_onchip", 7281024, 81024, 0},
{FAL_PART_MAGIC_WORD, "fdb_CurntE", "GD32_onchip", 7841024, 81024, 0},
}

[FlashDB][SYTIME] Get SysTime mKey=176, 24-10-18 16:0:30

[10:53:59.948]收←◆[FlashDB][SYTIME] set SysTime mKey=177,0X180A--0X1210--0X001E stamp=1729267230
[FlashDB][SYTIME] Get SysTime mKey=177, 24-10-18 16:0:30

[10:54:00.127]收←◆[FlashDB][kv][SYTM][fdb_Systm] (..\module\F00302_Fdb\src\fdb_kvdb.c:929) Trigger a GC check after alloc KV failed.
[FlashDB][kv][SYTM][fdb_Systm] (..\module\F00302_Fdb\src\fdb_kvdb.c:1075) Alloc an KV (size 64) failed when new KV. Now will GC then retry.
[FlashDB][kv][SYTM][fdb_Systm] (..\module\F00302_Fdb\src\fdb_kvdb.c:1161) The remain empty sector is 1, GC threshold is 1.
[FlashDB][kv][SYTM][fdb_Systm] (..\module\F00302_Fdb\src\fdb_kvdb.c:1058) Moved the KV (SYSTFS) from 0x00000024 to 0x00001024.
[FlashDB][kv][SYTM][fdb_Systm] (..\module\F00302_Fdb\src\fdb_kvdb.c:1130) KV (SysTime) is garbage NOT need move, collect it.
[FlashDB][kv][SYTM][fdb_Systm] (..\module\F00302_Fdb\src\fdb_kvdb.c:1130) KV (SysTime) is garbage NOT need move, collect it.
[FlashDB][kv][SYTM][fdb_Systm] (..\module\F00302_Fdb\src\fdb_kvdb.c:1130) KV (SysTime) is garbage NOT need move, collect it.
[FlashDB][kv][SYTM][fdb_Systm] (..\module\F00302_Fdb\src\fdb_kvdb.c:1130) KV (SysTime) is garbage NOT need move, collect it.
[FlashDB][kv][SYTM][fdb_Systm] (..\module\F00302_Fdb\src\fdb_kvdb.c:1130) KV (SysTime) is garbage NOT need move, collect it.

这问题很奇怪，请协助解答，非常感谢

[tianzhilianwo]

static const struct fal_flash_dev *flash_device_find_by_part(const struct fal_partition *part)
{
assert(part >= partition_table);
assert(part <= &partition_table[partition_table_len - 1]);

return part_flash_cache[part - partition_table].flash_dev;

}
出现以上BUG是在调用该函数时出现 异常，part_flash_cache 列表的内容被修改，不知道是flashDB传过来的异常还是这个计算方式有问题？

[armink]

你再深入抓一下，看看哪里把内存给踩踏了？