selftests/x86: Minor fixes
PATCH 2/2 selftests/x86: Add missing .note.GNU-stack sections
In reply to Andy LutomirskiSeveral of the x86 selftests end up with executable stacks because the asm was missing the annotation that says that they are modern and don't need executable stacks. Add the annotations.
--- tools/testing/selftests/x86/raw_syscall_helper_32.S | 2 ++ tools/testing/selftests/x86/thunks.S | 2 ++ 2 files changed, 4 insertions(+)
tools/testing/selftests/x86/raw_syscall_helper_32.S +2 -0
@@ -45,3 +45,5 @@ int80_and_ret: .type int80_and_ret, @function .size int80_and_ret, .-int80_and_ret + +.section .note.GNU-stack,"",%progbits
tools/testing/selftests/x86/thunks.S +2 -0
@@ -57,3 +57,5 @@ call32_from_64: ret .size call32_from_64, .-call32_from_64 + +.section .note.GNU-stack,"",%progbits
Signature · 1 line
2.28.0
PATCH 1/2 selftests/x86/fsgsbase: Fix GS == 1, 2, and 3 tests
In reply to Andy LutomirskiSetting GS to 1, 2, or 3 causes a nonsensical part of the IRET microcode to change GS back to zero on a return from kernel mode to user mode. The result is that these tests fail randomly depending on when interrupts happen. Detect when this happens and let the test pass.
--- tools/testing/selftests/x86/fsgsbase.c | 12 ++++++++++-- 1 file changed, 10 insertions(+), 2 deletions(-)
tools/testing/selftests/x86/fsgsbase.c +10 -2
@@ -392,8 +392,8 @@ static void set_gs_and_switch_to(unsigned long local, local =read_base (GS );/* -* Signal delivery seems to mess up weird selectors. Put it -* back. +* Signal delivery is quite likely to change a selector +* of 1, 2, or 3 back to 0 due to IRET being defective. */ asmvolatile ("mov %0, %%gs" : :"rm" (force_sel)); }else {
@@ -411,6 +411,14 @@ static void set_gs_and_switch_to(unsigned long local,if (base == local && sel_pre_sched == sel_post_sched) {printf ("[OK]\tGS/BASE remained 0x%hx/0x%lx\n" , sel_pre_sched, local); + }else if (base == local && sel_pre_sched >=1 && sel_pre_sched <=3 && + sel_post_sched ==0 ) { +/* +* IRET is misdesigned and will squash selectors 1, 2, or 3 +* to zero. Don't fail the test just because this happened. +*/ +printf ("[OK]\tGS/BASE changed from 0x%hx/0x%lx to 0x%hx/0x%lx because IRET is defective\n" , + sel_pre_sched, local, sel_post_sched, base); }else { nerrs++; printf("[FAIL]\tGS/BASE changed from 0x%hx/0x%lx to 0x%hx/0x%lx\n",
Signature · 1 line
2.28.0
cover selftests/x86: Minor fixes
Two bugs fixed.
Andy Lutomirski (2):
selftests/x86/fsgsbase: Fix GS == 1, 2, and 3 tests selftests/x86: Add missing .note.GNU-stack sections
tools/testing/selftests/x86/fsgsbase.c | 12 ++++++++++-- tools/testing/selftests/x86/raw_syscall_helper_32.S | 2 ++ tools/testing/selftests/x86/thunks.S | 2 ++ 3 files changed, 14 insertions(+), 2 deletions(-)
Signature · 1 line
2.28.0
[tip: x86/misc] selftests/x86/fsgsbase: Fix GS == 1, 2, and 3 tests
In reply to Andy LutomirskiThe following commit has been merged into the x86/misc branch of tip:
The following commit has been merged into the x86/misc branch of tip:
Commit-ID: 716572b0003ef67a4889bd7d85baf5099c5a0248 Gitweb: https://git.kernel.org/tip/716572b0003ef67a4889bd7d85baf5099c5a0248 Author: Andy Lutomirski <[email protected]> AuthorDate: Mon, 02 Nov 2020 11:51:10 -08:00 Committer: Borislav Petkov <[email protected]> CommitterDate: Tue, 24 Nov 2020 13:46:16 +01:00
selftests/x86/fsgsbase: Fix GS == 1, 2, and 3 tests
Setting GS to 1, 2, or 3 causes a nonsensical part of the IRET microcode to change GS back to zero on a return from kernel mode to user mode. The result is that these tests fail randomly depending on when interrupts happen. Detect when this happens and let the test pass.
--- tools/testing/selftests/x86/fsgsbase.c | 12 ++++++++++-- 1 file changed, 10 insertions(+), 2 deletions(-)
tools/testing/selftests/x86/fsgsbase.c +10 -2
@@ -392,8 +392,8 @@ static void set_gs_and_switch_to(unsigned long local, local =read_base (GS );/* -* Signal delivery seems to mess up weird selectors. Put it -* back. +* Signal delivery is quite likely to change a selector +* of 1, 2, or 3 back to 0 due to IRET being defective. */ asmvolatile ("mov %0, %%gs" : :"rm" (force_sel)); }else {
@@ -411,6 +411,14 @@ static void set_gs_and_switch_to(unsigned long local,if (base == local && sel_pre_sched == sel_post_sched) {printf ("[OK]\tGS/BASE remained 0x%hx/0x%lx\n" , sel_pre_sched, local); + }else if (base == local && sel_pre_sched >=1 && sel_pre_sched <=3 && + sel_post_sched ==0 ) { +/* +* IRET is misdesigned and will squash selectors 1, 2, or 3 +* to zero. Don't fail the test just because this happened. +*/ +printf ("[OK]\tGS/BASE changed from 0x%hx/0x%lx to 0x%hx/0x%lx because IRET is defective\n" , + sel_pre_sched, local, sel_post_sched, base); }else { nerrs++; printf("[FAIL]\tGS/BASE changed from 0x%hx/0x%lx to 0x%hx/0x%lx\n",
[tip: x86/misc] selftests/x86: Add missing .note.GNU-stack sections
In reply to Andy LutomirskiThe following commit has been merged into the x86/misc branch of tip:
The following commit has been merged into the x86/misc branch of tip:
Commit-ID: aeaaf005da1de075929e56562dced4a58238efc4 Gitweb: https://git.kernel.org/tip/aeaaf005da1de075929e56562dced4a58238efc4 Author: Andy Lutomirski <[email protected]> AuthorDate: Mon, 02 Nov 2020 11:51:11 -08:00 Committer: Borislav Petkov <[email protected]> CommitterDate: Tue, 24 Nov 2020 13:55:39 +01:00
selftests/x86: Add missing .note.GNU-stack sections
Several of the x86 selftests end up with executable stacks because the asm was missing the annotation that says that they are modern and don't need executable stacks. Add the annotations.
--- tools/testing/selftests/x86/raw_syscall_helper_32.S | 2 ++ tools/testing/selftests/x86/thunks.S | 2 ++ 2 files changed, 4 insertions(+)
tools/testing/selftests/x86/raw_syscall_helper_32.S +2 -0
@@ -45,3 +45,5 @@ int80_and_ret: .type int80_and_ret, @function .size int80_and_ret, .-int80_and_ret + +.section .note.GNU-stack,"",%progbits
tools/testing/selftests/x86/thunks.S +2 -0
@@ -57,3 +57,5 @@ call32_from_64: ret .size call32_from_64, .-call32_from_64 + +.section .note.GNU-stack,"",%progbits