[fc7bc51] | 1 | # |
---|
| 2 | # README,v 1.3 1995/12/19 20:13:47 joel Exp |
---|
| 3 | # |
---|
| 4 | |
---|
| 5 | This directory contains a stack bounds checker. It provides two |
---|
| 6 | primary features: |
---|
| 7 | |
---|
| 8 | + check for stack overflow at each context switch |
---|
| 9 | + provides an educated guess at each task's stack usage |
---|
| 10 | |
---|
| 11 | The stack overflow check at context switch works by looking for |
---|
| 12 | a 16 byte pattern at the logical end of the stack to be corrupted. |
---|
| 13 | The "guesser" assumes that the entire stack was prefilled with a known |
---|
| 14 | pattern and assumes that the pattern is still in place if the memory |
---|
| 15 | has not been used as a stack. |
---|
| 16 | |
---|
| 17 | Both of these can be fooled by pushing large holes onto the stack |
---|
| 18 | and not writing to them... or (much more unlikely) writing the |
---|
| 19 | magic patterns into memory. |
---|
| 20 | |
---|
| 21 | This code has not been extensively tested. It is provided as a tool |
---|
| 22 | for RTEMS users to catch the most common mistake in multitasking |
---|
| 23 | systems ... too little stack space. Suggestions and comments are appreciated. |
---|
| 24 | |
---|
| 25 | NOTES: |
---|
| 26 | |
---|
| 27 | 1. Stack usage information is questionable on CPUs which push |
---|
| 28 | large holes on stack. |
---|
| 29 | |
---|
| 30 | 2. The stack checker has a tendency to generate a fault when |
---|
| 31 | trying to print the helpful diagnostic message. If it comes |
---|
| 32 | out, congratulations. If not, then the variable Stack_check_Blown_task |
---|
| 33 | contains a pointer to the TCB of the offending task. This |
---|
| 34 | is usually enough to go on. |
---|
| 35 | |
---|
| 36 | FUTURE: |
---|
| 37 | |
---|
| 38 | 1. Determine how/if gcc will generate stack probe calls and support that. |
---|
| 39 | |
---|
| 40 | 2. Get accurate stack usage numbers on i960.. it pushes very large |
---|
| 41 | holes on the stack. |
---|