summaryrefslogtreecommitdiffstats
path: root/results/classifier/zero-shot-user-mode/output/runtime/1910
diff options
context:
space:
mode:
authorChristian Krinitsin <mail@krinitsin.com>2025-07-08 13:28:15 +0200
committerChristian Krinitsin <mail@krinitsin.com>2025-07-08 13:28:28 +0200
commit5aa276efcbd67f4300ca1a7f809c6e00aadb03da (patch)
tree9b8f0e074014cda8d42f5a97a95bc25082d8b764 /results/classifier/zero-shot-user-mode/output/runtime/1910
parent1a3c4faf4e0a25ed0b86e8739d5319a634cb9112 (diff)
downloademulator-bug-study-5aa276efcbd67f4300ca1a7f809c6e00aadb03da.tar.gz
emulator-bug-study-5aa276efcbd67f4300ca1a7f809c6e00aadb03da.zip
restructure results
Diffstat (limited to 'results/classifier/zero-shot-user-mode/output/runtime/1910')
-rw-r--r--results/classifier/zero-shot-user-mode/output/runtime/191068
1 files changed, 0 insertions, 68 deletions
diff --git a/results/classifier/zero-shot-user-mode/output/runtime/1910 b/results/classifier/zero-shot-user-mode/output/runtime/1910
deleted file mode 100644
index 3518e738..00000000
--- a/results/classifier/zero-shot-user-mode/output/runtime/1910
+++ /dev/null
@@ -1,68 +0,0 @@
-runtime: 0.342
-instruction: 0.339
-syscall: 0.319
-
-
-
-Signal handlers in x86_64 userspace have wrongly aligned stack
-Description of problem:
-Various applications crash in signal handlers due to `movaps` getting a misaligned stack address. For some reason this is reported as a NULL deref, but `gdb` clearly shows the true cause.
-
-```plaintext
-> qemu-x86_64 /usr/bin/ruby -e '`true`'
--e:1: [BUG] Segmentation fault at 0x0000000000000000
-ruby 3.2.2 (2023-03-30 revision e51014f9c0) [x86_64-linux-gnu]
-
--- Control frame information -----------------------------------------------
-c:0003 p:---- s:0011 e:000010 CFUNC :`
-c:0002 p:0005 s:0006 e:000005 EVAL -e:1 [FINISH]
-c:0001 p:0000 s:0003 E:0015b0 DUMMY [FINISH]
-
--- Ruby level backtrace information ----------------------------------------
--e:1:in `<main>'
--e:1:in ``'
-
--- Machine register context ------------------------------------------------
- RIP: 0x00002aaaab50f98a RBP: 0x00002aaaabb136b8 RSP: 0x00002aaaab2a9c98
- RAX: 0x0000000000000000 RBX: 0x0000000000004946 RCX: 0x0000000000000000
- RDX: 0x00002aaaab2a9c98 RDI: 0x000000000caf0000 RSI: 0x0000000000000000
- R8: 0x00002aaaab2aaa50 R9: 0x0000000000000050 R10: 0x0000000000000008
- R11: 0x0000000000000000 R12: 0x0000000000000002 R13: 0x0000000000007310
- R14: 0x0000000000005e10 R15: 0x00002aaab0537f20 EFL: 0x0000000000000246
-
--- C level backtrace information -------------------------------------------
-```
-
-```plaintext
-(gdb) x/i $pc
-=> 0x2aaaab50f98a: movaps %xmm0,(%rsp)
-(gdb) p/x $rsp
-$3 = 0x2aaaab2a9998
-```
-Steps to reproduce:
-1. ```qemu-x86_64 /usr/bin/ruby -e '`true`'```
-Additional information:
-The x86_64 psABI says:
-
-> the value (%rsp − 8) is always a multiple of 16 when control is transferred to the function entry point.
-
-However, when QEMU jumps to the signal handler, $rsp is aligned to 16B, i.e. ends in `0x..0`.
-
-The relevant kernel code:
-
-https://elixir.bootlin.com/linux/v6.5.5/source/arch/x86/kernel/signal.c#L123
-
-```plaintext
- sp -= frame_size;
-
- if (ia32_frame)
- /*
- * Align the stack pointer according to the i386 ABI,
- * i.e. so that on function entry ((sp + 4) & 15) == 0.
- */
- sp = ((sp + 4) & -FRAME_ALIGNMENT) - 4;
- else
- sp = round_down(sp, FRAME_ALIGNMENT) - 8;
-```
-
-CC @lvivier @bonzini @rth7680