Skip to content

Commit

Permalink
cet/shstk_alloc.c: will not show content of switched ssp due to kerne…
Browse files Browse the repository at this point in the history
…l improvement

Since the v6.5 cycle latest cet userspace shstk patch improvements, switched
ssp content will be protected and it makes sense, if code wants to show the
switched ssp content, it will get fake segfault like this:
"
ssp0 & (-8): 7f5d4efffff8
Segmentation fault (core dumped)
"
Dmesg shows the segfault as following:
"
shstk_alloc[3080]: segfault at 7f5d4e7ff000 ip 0000000000401258 sp 00007ffcd214f180 error 4 in shstk_alloc[401000+1000] likely on CPU 15 (core 3, socket 0)
"
So code should not show worked and switched ssp content to avoid fake failure.

Signed-off-by: Pengfei Xu <[email protected]>
  • Loading branch information
xupengfe authored and ysun committed Aug 25, 2023
1 parent 1e98a72 commit 9d381e0
Showing 1 changed file with 3 additions and 6 deletions.
9 changes: 3 additions & 6 deletions cet/shstk_alloc.c
Original file line number Diff line number Diff line change
@@ -1,6 +1,3 @@
// SPDX-License-Identifier: GPL-2.0-only
// Copyright (c) 2022 Intel Corporation.

// SPDX-License-Identifier: GPL-2.0
/*
* shstk_alloc.c - allocate a new shadow stack buffer aligenment by instructions
Expand Down Expand Up @@ -109,7 +106,7 @@ void try_shstk(unsigned long new_ssp)
asm volatile("rstorssp (%0)\n":: "r" (new_ssp));
asm volatile("saveprevssp");
ssp1 = get_ssp();
printf("ssp is now %lx, *ssp:%lx\n", ssp1, *((unsigned long *)ssp1));
printf("ssp is now %lx\n", ssp1);

ssp0 -= 8;
asm volatile("rstorssp (%0)\n":: "r" (ssp0));
Expand Down Expand Up @@ -146,9 +143,9 @@ int main(int argc, char *argv[])
if (ARCH_PRCTL(ARCH_SHSTK_DISABLE, ARCH_SHSTK_SHSTK)) {
printf("[FAIL]\tParent process disable shadow stack failed.\n");
return 1;
} else {
printf("[PASS]\tParent process disable shadow stack successfully.\n");
}
printf("[PASS]\tParent process disable shadow stack successfully.\n");

return 0;
}
#endif

0 comments on commit 9d381e0

Please sign in to comment.