Unverified Commit 05db3a6e authored by openeuler-ci-bot's avatar openeuler-ci-bot Committed by Gitee
Browse files

!10976 CVE-2024-42126 

Merge Pull Request from: @ci-robot 
 
PR sync from: Jinjie Ruan <ruanjinjie@huawei.com>
https://mailweb.openeuler.org/hyperkitty/list/kernel@openeuler.org/message/GXQP4WE55CJS4ZHQWNGVO2F6VNR4EMNZ/ 
CVE-2024-42126

Jinjie Ruan (1):

Mahesh Salgaonkar (1):


-- 
2.34.1
 
https://gitee.com/src-openeuler/kernel/issues/IAGPSI 
 
Link:https://gitee.com/openeuler/kernel/pulls/10976

 

Reviewed-by: default avatarZhang Jianhua <chris.zjh@huawei.com>
Signed-off-by: default avatarYang Yingliang <yangyingliang@huawei.com>
parents da4af443 cb8be6d4

arch/powerpc/kernel/mce.c

+1 −1
Original line number Diff line number Diff line
@@ -595,7 +595,7 @@ long notrace machine_check_early(struct pt_regs *regs) 


	this_cpu_set_ftrace_enabled(0);

	/*

	 * Do not use nmi_enter/exit for pseries hpte guest.

	 * Do not use nmi_enter/exit for pseries hpte guest

	 *

	 * Likewise, do not use it in real mode if percpu first chunk is not

	 * embedded. With CONFIG_NEED_PER_CPU_PAGE_FIRST_CHUNK enabled there

arch/powerpc/kernel/traps.c

+7 −1
Original line number Diff line number Diff line
@@ -838,8 +838,14 @@ void machine_check_exception(struct pt_regs *regs) 
	 * This is silly. The BOOK3S_64 should just call a different function

	 * rather than expecting semantics to magically change. Something

	 * like 'non_nmi_machine_check_exception()', perhaps?

	 *

	 * Do not use nmi_enter/exit in real mode if percpu first chunk is

	 * not embedded. With CONFIG_NEED_PER_CPU_PAGE_FIRST_CHUNK enabled

	 * there are chances where percpu allocation can come from

	 * vmalloc area.

	 */

	const bool nmi = !IS_ENABLED(CONFIG_PPC_BOOK3S_64);

	const bool nmi = !IS_ENABLED(CONFIG_PPC_BOOK3S_64) &&

			 !percpu_first_chunk_is_paged;



	if (nmi) nmi_enter();