请问：如何将CPU1上面所有任务全部移到CPU0上来运行？

jian9158101

回复 18# gaojl0728
这跟我遇到的问题并做的测试不一致，所以比较奇怪，我来详细描述下：
1. 板子起来后挂载某MTD为UBIFS，而UBIFS会有内核线程来协同工作，见如下Log
===
[   56.533388] UBI: attaching mtd7 to ubi7
[   56.641140] UBI: scanning is finished
[   56.644824] UBI: empty MTD device detected
[   56.700290] UBI: attached mtd7 (name "**", size 128 MiB) to ubi7
[   56.706369] UBI: PEB size: 4194304 bytes (4096 KiB), LEB size: 4161536 bytes
[   56.713453] UBI: min./max. I/O unit sizes: 16384/16384, sub-page size 16384
[   56.720446] UBI: VID header offset: 16384 (aligned 16384), data offset: 32768
[   56.727606] UBI: good PEBs: 32, bad PEBs: 0, corrupted PEBs: 0
[   56.733465] UBI: user volume: 0, internal volumes: 1, max. volumes count: 128
[   56.740634] UBI: max/mean erase counter: 0/0, WL threshold: 256, image sequence number: 1338532295
[   56.749623] UBI: available PEBs: 21, total reserved PEBs: 11, PEBs reserved for bad PEB handling: 7
[   56.758722] UBI: background thread "ubi_bgt7d" started, PID 1492
===
2. 此内核线程挂在CPU1上运行，当我将CPU1 down掉后，此内核线程还是睡在CPU1上
3. 当我去操作MTD7时，发生UBI读写错误，如下：
===
[   96.844922] UBIFS error : ubifs_read_node: bad node type (37 but expected 1)
[   96.852976] UBIFS error : ubifs_read_node: bad node at LEB 18:1133512, LEB mapping status 1
[   96.862284] UBIFS error : do_readpage: cannot read page 10413 of inode 65, error -22
[   99.123916] UBIFS error : ubifs_read_node: bad node type (21 but expected 9)
[   99.131945] UBIFS error : ubifs_read_node: bad node at LEB 9:802432, LEB mapping status 1
[   99.141085] UBIFS error : do_readpage: cannot read page 11309 of inode 65, error -22
[   99.152108] UBIFS error : ubifs_read_node: bad node type (21 but expected 9)
[   99.160156] UBIFS error : ubifs_read_node: bad node at LEB 9:802624, LEB mapping status 1
[   99.169299] UBIFS error : do_readpage: cannot read page 11317 of inode 65, error -22
[   99.178020] UBIFS error : ubifs_read_node: bad node type (21 but expected 9)
[   99.185996] UBIFS error : ubifs_read_node: bad node at LEB 9:802624, LEB mapping status 1
[   99.195145] UBIFS error : do_readpage: cannot read page 11318 of inode 65, error -22
===
4. 而如果CPU1在online状态下，对此MTD操作没有任何问题


   

jian9158101

回复 21# gaojl0728
对的，问题就在这边，当需要这个进程工作的时候，它看上去并没有被wake_up，而是继续跟CPU1一起睡大觉......所以在想是否有什么办法可以在CPU1 DOWN之前将所有TASK迁移至CPU0
尝试用TASKSET，失败......有点怀疑是否是我这版kernel-3.2.54本身关于cpu-hotplug有问题...


   

jian9158101

回复 13# 瀚海书香
请忽略。。。