Setting up the signal stack in the handle_segv.c. #1
Conversation
Signed-off-by: Jagadeesh Pagadala <[email protected]>
Signed-off-by: Jagadeesh Pagadala <[email protected]>
Signed-off-by: Jagadeesh Pagadala <[email protected]>
|
Addressing the review comments. |
Signed-off-by: Jagadeesh Pagadala <[email protected]>
Signed-off-by: Jagadeesh Pagadala <[email protected]>
| @@ -197,9 +200,17 @@ int main(int argc, char **argv) | |||
| exit(1); | |||
| } | |||
|
|
|||
| /*Setting up the alternate stack to properly handle the SIGSEGV*/ | |||
There was a problem hiding this comment.
Nice!
just fix code style as per Linux kernel style guidelines
Tip- use 'indent -linux <fname.c>' to do so ..
|
Marking this PR as open again.. |
There was a problem hiding this comment.
Hi Jagadeesh,
It looks good as such, But...
When i run the original ver, the faulting addr dtl etc is all fine, but with the new ver using alt stack it's not:
Eg. run: w/ orig ver:
ch12 $ ./handle_segv u r
Attempting to read contents of arbitrary usermode va uptr = 0xdeadface:
handle_segv:
------------------- FATAL signal ---------------------------
myfault: received signal 11. errno=0
Cause/Origin: (si_code=1): SEGV_MAPERR: address not mapped to object
Faulting instr or address = 0xdeadface
--- Register Dump [x86_64] ---
RAX = 0x00000000deadface RBX = 0x0000000000000000 RCX = 0x0000000000000000
RDX = 0x0000000000000000 RSI = 0x000055bdcda50260 RDI = 0x0000000000000001
RBP = 0x00007ffc4093cdb0 R8 = 0x0000000000000000 R9 = 0x0000000000000000
R10 = 0x0000000000000000 R11 = 0x0000000000000246 R12 = 0x000055bdcd19fc50
R13 = 0x00007ffc4093ce90 R14 = 0x0000000000000000 R15 = 0x0000000000000000
RSP = 0x00007ffc4093cce0
RIP = 0x000055bdcd1a03cb EFLAGS = 0x0000000000010206
TRAP# = 14 ERROR = 04
CR2 = 0x00000000deadface
psiginfo helper: Segmentation fault (Address not mapped to object [0xdeadface])
With the new ver w/ alt stack:
ch12 $ ./handle_segv u r
Alt signal stack size = 8388608
Attempting to read contents of arbitrary usermode va uptr = 0xdeadface:
handle_segv:
------------------- FATAL signal ---------------------------
myfault: received signal 11. errno=4
Cause/Origin: (si_code=8): -none-
Faulting instr or address = 0x7ffe4eff0b30
--- Register Dump [x86_64] ---
RAX = 0x00000000deadface RBX = 0x0000000000000000 RCX = 0x0000000000000000
RDX = 0x0000000000000000 RSI = 0x000055a5d4e7a260 RDI = 0x0000000000000001
RBP = 0x00007ffe4eff0c40 R8 = 0x0000000000000000 R9 = 0x0000000000000000
R10 = 0x0000000000000000 R11 = 0x0000000000000246 R12 = 0x000055a5d2fe7ca0
R13 = 0x00007ffe4eff0d20 R14 = 0x0000000000000000 R15 = 0x0000000000000000
RSP = 0x00007ffe4eff0b70
RIP = 0x000055a5d2fe852e EFLAGS = 0x0000000000010202
TRAP# = 14 ERROR = 04
CR2 = 0x00000000deadface
psiginfo helper: Unknown signal 1685382480
? Unsure why... but we can't merge this until it's fixed and works correctly.
|
Pl ignore the extra bold fonts from "RIP" onward... |
Signed-off-by: Jagadeesh Pagadala <[email protected]>
|
It is working fine on my CentOS box, and even on seawolf. commit 4af72d5 RIP = 0x0000561af22cd510 EFLAGS = 0x0000000000010206
|
Hi Kaiwan, I am your student from HPE Bangalore, thought of providing some improvement with your permission.
Signed-off-by: Jagadeesh Pagadala [email protected]