rt_malloc时分配的内存地址与另一任务地址相同

发布于 2022-02-14 17:59:24 浏览：1131 订阅该版

我在STM32F103中使用了RTTHREAD，同时运行多个任务， 然后在任务A中malloc一块内存用于存储字符串数据，使用完成后立刻释放，结果发现释放时有时有发生hard fault问题，定位在r_free函数处。
再通过查看发生错误时的log显示：该数据的内存地址与正在运行任务B地址重叠的，导致释放时就发生了hard fault错误。 
  问题在于为什么rt_malloc申请的内存地址会与正在运行的任务B重叠？

LOG如下：allocate memory at 0x20003eb0 第一次申请了512大小的内存是任务B本身的，第二次就是任务A内的函数申请的。
```
[02/14 17:17:55.331]allocate memory at 0x20003e2c, size: 132
[02/14 17:17:55.331]malloc size 500

[02/14 17:17:55.331]allocate memory at 0x20003eb0, size: 512
[02/14 17:17:55.331]D/app      short thread run...

[02/14 17:17:55.956]malloc size 32
[02/14 17:17:55.956]allocate memory at 0x20003eb0, size: 44

[02/14 17:17:55.987]release memory 0x20003eb0, size: 44

[02/14 17:17:55.987]psr: 0x21000000
[02/14 17:17:55.987]r00: 0x20001278
[02/14 17:17:55.987]r01: 0x28029de7
[02/14 17:17:55.987]r02: 0x00002c2c
[02/14 17:17:55.987]r03: 0x000063f4
[02/14 17:17:55.987]r04: 0x20003ea4
[02/14 17:17:55.987]r05: 0x20000024
[02/14 17:17:55.987]r06: 0x20000024
[02/14 17:17:55.987]r07: 0xdeadbeef
[02/14 17:17:55.987]r08: 0xdeadbeef
[02/14 17:17:55.987]r09: 0xdeadbeef
[02/14 17:17:55.987]r10: 0xdeadbeef
[02/14 17:17:56.034]r11: 0xdeadbeef
[02/14 17:17:56.034]r12: 0xfffffffd
[02/14 17:17:56.034] lr: 0x08035901
[02/14 17:17:56.035] pc: 0x08034de2
[02/14 17:17:56.036]hard fault on thread: lcd thread
```