Re: [PATCH] arm64: kernel: initialize missing kexec_buf->random field

[Baoquan He <bhe@redhat.com>]

On 11/28/25 at 02:01pm, Mark Rutland wrote:
> On Fri, Nov 28, 2025 at 05:55:05AM -0800, Breno Leitao wrote:
> > 
> > On Fri, Nov 28, 2025 at 08:17:21AM +0800, Baoquan He wrote:
> > > On 11/27/25 at 11:37am, Andrew Morton wrote:
> > > > On Thu, 27 Nov 2025 18:26:44 +0000 Yeoreum Yun <yeoreum.yun@arm.com> wrote:
> > > > 
> > > > > Commit bf454ec31add ("kexec_file: allow to place kexec_buf randomly")
> > > > > introduced the kexec_buf->random field to enable random placement of
> > > > > kexec_buf.
> > > > > 
> > > > > However, this field was never properly initialized for kexec images
> > > > > that do not need to be placed randomly, leading to the following UBSAN
> > > > > warning:
> > > > > 
> > > > > [  +0.364528] ------------[ cut here ]------------
> > > > > [  +0.000019] UBSAN: invalid-load in ./include/linux/kexec.h:210:12
> > > > > [  +0.000131] load of value 2 is not a valid value for type 'bool' (aka '_Bool')
> > > > > [  +0.000003] CPU: 4 UID: 0 PID: 927 Comm: kexec Not tainted 6.18.0-rc7+ #3 PREEMPT(full)
> > > > > [  +0.000002] Hardware name: QEMU QEMU Virtual Machine, BIOS 0.0.0 02/06/2015
> > > > > [  +0.000000] Call trace:
> > > > > [  +0.000001]  show_stack+0x24/0x40 (C)
> > > > > [  +0.000006]  __dump_stack+0x28/0x48
> > > > > [  +0.000002]  dump_stack_lvl+0x7c/0xb0
> > > > > [  +0.000002]  dump_stack+0x18/0x34
> > > > > [  +0.000001]  ubsan_epilogue+0x10/0x50
> > > > > [  +0.000002]  __ubsan_handle_load_invalid_value+0xc8/0xd0
> > > > > [  +0.000003]  locate_mem_hole_callback+0x28c/0x2a0
> > > > > [  +0.000003]  kexec_locate_mem_hole+0xf4/0x2f0
> > > > > [  +0.000001]  kexec_add_buffer+0xa8/0x178
> > > > > [  +0.000002]  image_load+0xf0/0x258
> > > > > [  +0.000001]  __arm64_sys_kexec_file_load+0x510/0x718
> > > > > [  +0.000002]  invoke_syscall+0x68/0xe8
> > > > > [  +0.000001]  el0_svc_common+0xb0/0xf8
> > > > > [  +0.000002]  do_el0_svc+0x28/0x48
> > > > > [  +0.000001]  el0_svc+0x40/0xe8
> > > > > [  +0.000002]  el0t_64_sync_handler+0x84/0x140
> > > > > [  +0.000002]  el0t_64_sync+0x1bc/0x1c0
> > > > > 
> > > > > To address this, initialise kexec_buf->random field properly.
> > > > > 
> > > > > Fixes: bf454ec31add ("kexec_file: allow to place kexec_buf randomly")
> > > > 
> > > > Thanks, I'll add a cc:stable to this.
> > > 
> > > This has been fixed in below series from Breno Leitao.
> > > 
> > > [PATCH 0/3] kexec: Fix invalid field access
> > > https://lore.kernel.org/all/20250827-kbuf_all-v1-0-1df9882bb01a@debian.org/T/#u
> > 
> > Right, these fixes are on 6.18 since day one.
> > 
> > Yeoreum is hitting another code path that I haven't fixed, it seems
> > (through image_load()).
> > 
> > That said, I think the fix should be similar to commit 04d3cd43700 ("arm64:
> > kexec: initialize kexec_buf struct in load_other_segments()"). I.e:
> > 
> > 	--- a/arch/arm64/kernel/kexec_image.c
> > 	+++ b/arch/arm64/kernel/kexec_image.c
> > 	@@ -41,7 +41,7 @@ static void *image_load(struct kimage *image,
> > 		struct arm64_image_header *h;
> > 		u64 flags, value;
> > 		bool be_image, be_kernel;
> > 	-       struct kexec_buf kbuf;
> > 	+       struct kexec_buf kbuf = {};
> > 		unsigned long text_offset, kernel_segment_number;
> > 		struct kexec_segment *kernel_segment;
> > 		int ret;
> 
> FWIW, I completey agree; your proposal is a much better solution.

Yes, maybe Yeoreum or Breno can post a new version to fix it on arm64.

> 
> From a quick scan, it looks like loongarch also has some more dodgy
> instances:

In the latest mainline kernel, the problem on loongarch has gone.

> 
> | % git grep 'struct kexec_buf\s[a-z_0-9]\+;'
> | arch/arm64/kernel/kexec_image.c:        struct kexec_buf kbuf;
> | arch/loongarch/kernel/kexec_efi.c:      struct kexec_buf kbuf;
> | arch/loongarch/kernel/kexec_elf.c:      struct kexec_buf kbuf;
> | arch/loongarch/kernel/machine_kexec_file.c:     struct kexec_buf kbuf;
> | kernel/kexec_handover.c:        struct kexec_buf scratch;
> 
> The 'scratch' case in kernel/kexec_handover.c gets overwritten with a
> struct assignment that'll happen to zero the 'random' field.
> 
> Mark.
>