* [daveh-devel:kss-2026-10-23 25/25] arch/x86/kernel/cet.c:330:6: error: redefinition of 'arch_kernel_stack_lock_current'
@ 2026-10-24 0:50 kernel test robot
0 siblings, 0 replies; 2+ messages in thread
From: kernel test robot @ 2026-10-24 0:50 UTC (permalink / raw)
To: Dave Hansen; +Cc: llvm, oe-kbuild-all
tree: https://git.kernel.org/pub/scm/linux/kernel/git/daveh/devel.git kss-2026-10-23
head: b6fa923e47fcecf329844fd91ed615de35478ea3
commit: b6fa923e47fcecf329844fd91ed615de35478ea3 [25/25] x86/boot: Add CPU shadow stack bootstrapping
config: x86_64-kexec (https://download.01.org/0day-ci/archive/20261024/202610240248.jnn1TnCR-lkp@intel.com/config)
compiler: clang version 22.1.8 (https://github.com/llvm/llvm-project ca7933e47d3a3451d81e72ac174dcb5aa28b59d1)
reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20261024/202610240248.jnn1TnCR-lkp@intel.com/reproduce)
If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <lkp@intel.com>
| Closes: https://lore.kernel.org/oe-kbuild-all/202610240248.jnn1TnCR-lkp@intel.com/
All errors (new ones prefixed by >>):
arch/x86/kernel/cet.c:277:17: warning: no previous prototype for function 'kshstk_build' [-Wmissing-prototypes]
277 | asmlinkage void kshstk_build(unsigned long *root_sp)
| ^
arch/x86/kernel/cet.c:277:12: note: declare 'static' if the function is not intended to be used outside of this translation unit
277 | asmlinkage void kshstk_build(unsigned long *root_sp)
| ^
| static
>> arch/x86/kernel/cet.c:330:6: error: redefinition of 'arch_kernel_stack_lock_current'
330 | void arch_kernel_stack_lock_current(void)
| ^
arch/x86/include/asm/shstk.h:49:20: note: previous definition is here
49 | static inline void arch_kernel_stack_lock_current(void) {}
| ^
>> arch/x86/kernel/cet.c:351:18: error: no member named 'kshstk_ssp' in 'struct thread_struct'
351 | current->thread.kshstk_ssp = kshstk_enable_current();
| ~~~~~~~~~~~~~~~ ^
>> arch/x86/kernel/cet.c:351:31: error: call to undeclared function 'kshstk_enable_current'; ISO C99 and later do not support implicit function declarations [-Wimplicit-function-declaration]
351 | current->thread.kshstk_ssp = kshstk_enable_current();
| ^
1 warning and 3 errors generated.
vim +/arch_kernel_stack_lock_current +330 arch/x86/kernel/cet.c
266
267 /*
268 * Use the unwinder infrastructure to walk back up the regular stack
269 * and mirror all the return addresses over to the shadow stack.
270 *
271 * root_sp is captured in assembly before calling here to prevent the
272 * compiler from messing with it. There may be additional frames or
273 * other gunk on the stack below root_sp by the time the unwinder
274 * starts, but those will ignored. All that matters is that the stack
275 * is back to "root_sp" when shadow stacks are enabled.
276 */
> 277 asmlinkage void kshstk_build(unsigned long *root_sp)
278 {
279 unsigned long ents[KSHSTK_MAX_ENTS];
280 struct unwind_state state;
281 unsigned long *ro_slot;
282 unsigned long *rw_slot;
283 int n = 0;
284
285 /*
286 * Use the unwinder to find the return addresses of callers.
287 * Note that ents[0] is the most recent call. The is the
288 * opposite of stacks. ents[] is tempoarary "upside down"
289 * storage that will be reversed when copied to the shadow
290 * stack.
291 */
292 for (unwind_start(&state, current, NULL, root_sp);
293 !unwind_done(&state) && n < KSHSTK_MAX_ENTS;
294 unwind_next_frame(&state))
295 ents[n++] = unwind_get_return_address(&state);
296
297 BUG_ON(n == KSHSTK_MAX_ENTS);
298
299 /* Transfer the return addresses to the shadow stack: */
300 ro_slot = task_shstk_slot(0);
301 rw_slot = shstk_alias(ro_slot);
302 for (int i = n - 1; i >= 0; i--) {
303 *rw_slot = ents[i];
304 rw_slot--;
305 ro_slot--;
306 }
307 /* Put a restore token in the last slot: */
308 shstk_create_restore_token(ro_slot);
309 }
310
311 static void kshstk_prepare_pages(void)
312 {
313 /*
314 * Necessary for the init_task shadow stack on the boot CPU,
315 * but not the secondary CPUs which have normal vmap'd stacks.
316 */
317 set_memory_shstk_noalias((unsigned long)task_shadow_stack(current),
318 shadow_stack_size() / PAGE_SIZE);
319 }
320
321 /*
322 * This "locks" the stack in place and does not allow any control
323 * flow manipulation to occur because of changes to the stack.
324 *
325 * CPU enters with shadow stacks disabled and leaves with them
326 * enabled.
327 *
328 * Must be called before the CPU does its first context switch.
329 */
> 330 void arch_kernel_stack_lock_current(void)
331 {
332 unsigned long flags;
333
334 if (!cpu_feature_enabled(X86_FEATURE_KERN_SHSTK))
335 return;
336
337 kshstk_prepare_pages();
338
339 /*
340 * Regular interrupts are delivered on stack level 0. There
341 * is a window within the enabling where CET_SHSTK_EN=1 but
342 * SSP is not yet set. If an interrupt is delivered in that
343 * window, it would try to push on to this stack. Disable
344 * regular interrupts over this window.
345 *
346 * Interrupts that use other stack levels (like NMI) are OK
347 * because FRED switches stacks for them and does not push
348 * on stack level 0.
349 */
350 local_irq_save(flags);
> 351 current->thread.kshstk_ssp = kshstk_enable_current();
--
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki
^ permalink raw reply [flat|nested] 2+ messages in thread
* [daveh-devel:kss-2026-10-23 25/25] arch/x86/kernel/cet.c:330:6: error: redefinition of 'arch_kernel_stack_lock_current'
@ 2026-10-25 2:24 kernel test robot
0 siblings, 0 replies; 2+ messages in thread
From: kernel test robot @ 2026-10-25 2:24 UTC (permalink / raw)
To: Dave Hansen; +Cc: llvm, oe-kbuild-all
tree: https://git.kernel.org/pub/scm/linux/kernel/git/daveh/devel.git kss-2026-10-23
head: b6fa923e47fcecf329844fd91ed615de35478ea3
commit: b6fa923e47fcecf329844fd91ed615de35478ea3 [25/25] x86/boot: Add CPU shadow stack bootstrapping
config: x86_64-kexec (https://download.01.org/0day-ci/archive/20261025/202610251021.TiL9FTBK-lkp@intel.com/config)
compiler: clang version 22.1.8 (https://github.com/llvm/llvm-project ca7933e47d3a3451d81e72ac174dcb5aa28b59d1)
reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20261025/202610251021.TiL9FTBK-lkp@intel.com/reproduce)
If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <lkp@intel.com>
| Closes: https://lore.kernel.org/oe-kbuild-all/202610251021.TiL9FTBK-lkp@intel.com/
All errors (new ones prefixed by >>):
arch/x86/kernel/cet.c:277:17: warning: no previous prototype for function 'kshstk_build' [-Wmissing-prototypes]
277 | asmlinkage void kshstk_build(unsigned long *root_sp)
| ^
arch/x86/kernel/cet.c:277:12: note: declare 'static' if the function is not intended to be used outside of this translation unit
277 | asmlinkage void kshstk_build(unsigned long *root_sp)
| ^
| static
>> arch/x86/kernel/cet.c:330:6: error: redefinition of 'arch_kernel_stack_lock_current'
330 | void arch_kernel_stack_lock_current(void)
| ^
arch/x86/include/asm/shstk.h:49:20: note: previous definition is here
49 | static inline void arch_kernel_stack_lock_current(void) {}
| ^
>> arch/x86/kernel/cet.c:351:18: error: no member named 'kshstk_ssp' in 'struct thread_struct'
351 | current->thread.kshstk_ssp = kshstk_enable_current();
| ~~~~~~~~~~~~~~~ ^
>> arch/x86/kernel/cet.c:351:31: error: call to undeclared function 'kshstk_enable_current'; ISO C99 and later do not support implicit function declarations [-Wimplicit-function-declaration]
351 | current->thread.kshstk_ssp = kshstk_enable_current();
| ^
1 warning and 3 errors generated.
vim +/arch_kernel_stack_lock_current +330 arch/x86/kernel/cet.c
320
321 /*
322 * This "locks" the stack in place and does not allow any control
323 * flow manipulation to occur because of changes to the stack.
324 *
325 * CPU enters with shadow stacks disabled and leaves with them
326 * enabled.
327 *
328 * Must be called before the CPU does its first context switch.
329 */
> 330 void arch_kernel_stack_lock_current(void)
331 {
332 unsigned long flags;
333
334 if (!cpu_feature_enabled(X86_FEATURE_KERN_SHSTK))
335 return;
336
337 kshstk_prepare_pages();
338
339 /*
340 * Regular interrupts are delivered on stack level 0. There
341 * is a window within the enabling where CET_SHSTK_EN=1 but
342 * SSP is not yet set. If an interrupt is delivered in that
343 * window, it would try to push on to this stack. Disable
344 * regular interrupts over this window.
345 *
346 * Interrupts that use other stack levels (like NMI) are OK
347 * because FRED switches stacks for them and does not push
348 * on stack level 0.
349 */
350 local_irq_save(flags);
> 351 current->thread.kshstk_ssp = kshstk_enable_current();
--
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2026-10-25 2:25 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-10-25 2:24 [daveh-devel:kss-2026-10-23 25/25] arch/x86/kernel/cet.c:330:6: error: redefinition of 'arch_kernel_stack_lock_current' kernel test robot
-- strict thread matches above, loose matches on Subject: below --
2026-10-24 0:50 kernel test robot
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox