LinuxLists.cc - [PATCH 1/2] RISC-V: kexec: Fix memory leak of fdt buffer

2022-11-04 10:45:56

Subject: [PATCH 1/2] RISC-V: kexec: Fix memory leak of fdt buffer

This is reported by kmemleak detector:

unreferenced object 0xff60000082864000 (size 9588):
comm "kexec", pid 146, jiffies 4294900634 (age 64.788s)
hex dump (first 32 bytes):
d0 0d fe ed 00 00 12 ed 00 00 00 48 00 00 11 40 ...........H...@
00 00 00 28 00 00 00 11 00 00 00 02 00 00 00 00 ...(............
backtrace:
[<00000000f95b17c4>] kmemleak_alloc+0x34/0x3e
[<00000000b9ec8e3e>] kmalloc_order+0x9c/0xc4
[<00000000a95cf02e>] kmalloc_order_trace+0x34/0xb6
[<00000000f01e68b4>] __kmalloc+0x5c2/0x62a
[<000000002bd497b2>] kvmalloc_node+0x66/0xd6
[<00000000906542fa>] of_kexec_alloc_and_setup_fdt+0xa6/0x6ea
[<00000000e1166bde>] elf_kexec_load+0x206/0x4ec
[<0000000036548e09>] kexec_image_load_default+0x40/0x4c
[<0000000079fbe1b4>] sys_kexec_file_load+0x1c4/0x322
[<0000000040c62c03>] ret_from_syscall+0x0/0x2

In elf_kexec_load(), a buffer is allocated via kvmalloc() to store fdt.
While it's not freed back to system when kexec kernel is reloaded or
unloaded. Then memory leak is caused. Fix it by introducing riscv
specific function arch_kimage_file_post_load_cleanup(), and freeing the
buffer there.

Fixes: 6261586e0c91 ("RISC-V: Add kexec_file support")
Signed-off-by: Li Huafei <[email protected]>
---
arch/riscv/include/asm/kexec.h | 5 +++++
arch/riscv/kernel/elf_kexec.c | 10 ++++++++++
2 files changed, 15 insertions(+)

diff --git a/arch/riscv/include/asm/kexec.h b/arch/riscv/include/asm/kexec.h
index eee260e8ab30..2b56769cb530 100644
--- a/arch/riscv/include/asm/kexec.h
+++ b/arch/riscv/include/asm/kexec.h
@@ -39,6 +39,7 @@ crash_setup_regs(struct pt_regs *newregs,
#define ARCH_HAS_KIMAGE_ARCH

struct kimage_arch {
+ void *fdt; /* For CONFIG_KEXEC_FILE */
unsigned long fdt_addr;
};

@@ -62,6 +63,10 @@ int arch_kexec_apply_relocations_add(struct purgatory_info *pi,
const Elf_Shdr *relsec,
const Elf_Shdr *symtab);
#define arch_kexec_apply_relocations_add arch_kexec_apply_relocations_add
+
+struct kimage;
+int arch_kimage_file_post_load_cleanup(struct kimage *image);
+#define arch_kimage_file_post_load_cleanup arch_kimage_file_post_load_cleanup
#endif

#endif
diff --git a/arch/riscv/kernel/elf_kexec.c b/arch/riscv/kernel/elf_kexec.c
index 0cb94992c15b..ff30fcb43f47 100644
--- a/arch/riscv/kernel/elf_kexec.c
+++ b/arch/riscv/kernel/elf_kexec.c
@@ -21,6 +21,14 @@
#include <linux/memblock.h>
#include <asm/setup.h>

+int arch_kimage_file_post_load_cleanup(struct kimage *image)
+{
+ kvfree(image->arch.fdt);
+ image->arch.fdt = NULL;
+
+ return kexec_image_post_load_cleanup_default(image);
+}
+
static int riscv_kexec_elf_load(struct kimage *image, struct elfhdr *ehdr,
struct kexec_elf_info *elf_info, unsigned long old_pbase,
unsigned long new_pbase)
@@ -298,6 +306,8 @@ static void *elf_kexec_load(struct kimage *image, char *kernel_buf,
pr_err("Error add DTB kbuf ret=%d\n", ret);
goto out_free_fdt;
}
+ /* Cache the fdt buffer address for memory cleanup */
+ image->arch.fdt = fdt;
pr_notice("Loaded device tree at 0x%lx\n", kbuf.mem);
goto out;

--
2.17.1

2022-11-04 13:22:34

by Conor Dooley

[permalink] [raw]

Subject: Re: [PATCH 1/2] RISC-V: kexec: Fix memory leak of fdt buffer

On Fri, Nov 04, 2022 at 05:56:57PM +0800, Li Huafei wrote:
> This is reported by kmemleak detector:
>
> unreferenced object 0xff60000082864000 (size 9588):
> comm "kexec", pid 146, jiffies 4294900634 (age 64.788s)
> hex dump (first 32 bytes):
> d0 0d fe ed 00 00 12 ed 00 00 00 48 00 00 11 40 ...........H...@
> 00 00 00 28 00 00 00 11 00 00 00 02 00 00 00 00 ...(............
> backtrace:
> [<00000000f95b17c4>] kmemleak_alloc+0x34/0x3e
> [<00000000b9ec8e3e>] kmalloc_order+0x9c/0xc4
> [<00000000a95cf02e>] kmalloc_order_trace+0x34/0xb6
> [<00000000f01e68b4>] __kmalloc+0x5c2/0x62a
> [<000000002bd497b2>] kvmalloc_node+0x66/0xd6
> [<00000000906542fa>] of_kexec_alloc_and_setup_fdt+0xa6/0x6ea
> [<00000000e1166bde>] elf_kexec_load+0x206/0x4ec
> [<0000000036548e09>] kexec_image_load_default+0x40/0x4c
> [<0000000079fbe1b4>] sys_kexec_file_load+0x1c4/0x322
> [<0000000040c62c03>] ret_from_syscall+0x0/0x2
>
> In elf_kexec_load(), a buffer is allocated via kvmalloc() to store fdt.
> While it's not freed back to system when kexec kernel is reloaded or
> unloaded. Then memory leak is caused. Fix it by introducing riscv
> specific function arch_kimage_file_post_load_cleanup(), and freeing the
> buffer there.
>
> Fixes: 6261586e0c91 ("RISC-V: Add kexec_file support")
> Signed-off-by: Li Huafei <[email protected]>

Both of these bits of cleanup seem to echo what's the case on arm64.
Seems reasonable to me..
Reviewed-by: Conor Dooley <[email protected]>

> ---
> arch/riscv/include/asm/kexec.h | 5 +++++
> arch/riscv/kernel/elf_kexec.c | 10 ++++++++++
> 2 files changed, 15 insertions(+)
>
> diff --git a/arch/riscv/include/asm/kexec.h b/arch/riscv/include/asm/kexec.h
> index eee260e8ab30..2b56769cb530 100644
> --- a/arch/riscv/include/asm/kexec.h
> +++ b/arch/riscv/include/asm/kexec.h
> @@ -39,6 +39,7 @@ crash_setup_regs(struct pt_regs *newregs,
> #define ARCH_HAS_KIMAGE_ARCH
>
> struct kimage_arch {
> + void *fdt; /* For CONFIG_KEXEC_FILE */
> unsigned long fdt_addr;
> };
>
> @@ -62,6 +63,10 @@ int arch_kexec_apply_relocations_add(struct purgatory_info *pi,
> const Elf_Shdr *relsec,
> const Elf_Shdr *symtab);
> #define arch_kexec_apply_relocations_add arch_kexec_apply_relocations_add
> +
> +struct kimage;
> +int arch_kimage_file_post_load_cleanup(struct kimage *image);
> +#define arch_kimage_file_post_load_cleanup arch_kimage_file_post_load_cleanup
> #endif
>
> #endif
> diff --git a/arch/riscv/kernel/elf_kexec.c b/arch/riscv/kernel/elf_kexec.c
> index 0cb94992c15b..ff30fcb43f47 100644
> --- a/arch/riscv/kernel/elf_kexec.c
> +++ b/arch/riscv/kernel/elf_kexec.c
> @@ -21,6 +21,14 @@
> #include <linux/memblock.h>
> #include <asm/setup.h>
>
> +int arch_kimage_file_post_load_cleanup(struct kimage *image)
> +{
> + kvfree(image->arch.fdt);
> + image->arch.fdt = NULL;
> +
> + return kexec_image_post_load_cleanup_default(image);
> +}
> +
> static int riscv_kexec_elf_load(struct kimage *image, struct elfhdr *ehdr,
> struct kexec_elf_info *elf_info, unsigned long old_pbase,
> unsigned long new_pbase)
> @@ -298,6 +306,8 @@ static void *elf_kexec_load(struct kimage *image, char *kernel_buf,
> pr_err("Error add DTB kbuf ret=%d\n", ret);
> goto out_free_fdt;
> }
> + /* Cache the fdt buffer address for memory cleanup */
> + image->arch.fdt = fdt;
> pr_notice("Loaded device tree at 0x%lx\n", kbuf.mem);
> goto out;
>
> --
> 2.17.1
>

2022-11-07 02:06:40

by Liao, Chang

[permalink] [raw]

Subject: Re: [PATCH 1/2] RISC-V: kexec: Fix memory leak of fdt buffer

在 2022/11/4 17:56, Li Huafei 写道:
> This is reported by kmemleak detector:
>
> unreferenced object 0xff60000082864000 (size 9588):
> comm "kexec", pid 146, jiffies 4294900634 (age 64.788s)
> hex dump (first 32 bytes):
> d0 0d fe ed 00 00 12 ed 00 00 00 48 00 00 11 40 ...........H...@
> 00 00 00 28 00 00 00 11 00 00 00 02 00 00 00 00 ...(............
> backtrace:
> [<00000000f95b17c4>] kmemleak_alloc+0x34/0x3e
> [<00000000b9ec8e3e>] kmalloc_order+0x9c/0xc4
> [<00000000a95cf02e>] kmalloc_order_trace+0x34/0xb6
> [<00000000f01e68b4>] __kmalloc+0x5c2/0x62a
> [<000000002bd497b2>] kvmalloc_node+0x66/0xd6
> [<00000000906542fa>] of_kexec_alloc_and_setup_fdt+0xa6/0x6ea
> [<00000000e1166bde>] elf_kexec_load+0x206/0x4ec
> [<0000000036548e09>] kexec_image_load_default+0x40/0x4c
> [<0000000079fbe1b4>] sys_kexec_file_load+0x1c4/0x322
> [<0000000040c62c03>] ret_from_syscall+0x0/0x2
>
> In elf_kexec_load(), a buffer is allocated via kvmalloc() to store fdt.
> While it's not freed back to system when kexec kernel is reloaded or
> unloaded. Then memory leak is caused. Fix it by introducing riscv
> specific function arch_kimage_file_post_load_cleanup(), and freeing the
> buffer there.
>
> Fixes: 6261586e0c91 ("RISC-V: Add kexec_file support")
> Signed-off-by: Li Huafei <[email protected]>
> ---
> arch/riscv/include/asm/kexec.h | 5 +++++
> arch/riscv/kernel/elf_kexec.c | 10 ++++++++++
> 2 files changed, 15 insertions(+)
>
> diff --git a/arch/riscv/include/asm/kexec.h b/arch/riscv/include/asm/kexec.h
> index eee260e8ab30..2b56769cb530 100644
> --- a/arch/riscv/include/asm/kexec.h
> +++ b/arch/riscv/include/asm/kexec.h
> @@ -39,6 +39,7 @@ crash_setup_regs(struct pt_regs *newregs,
> #define ARCH_HAS_KIMAGE_ARCH
>
> struct kimage_arch {
> + void *fdt; /* For CONFIG_KEXEC_FILE */
> unsigned long fdt_addr;
> };
>
> @@ -62,6 +63,10 @@ int arch_kexec_apply_relocations_add(struct purgatory_info *pi,
> const Elf_Shdr *relsec,
> const Elf_Shdr *symtab);
> #define arch_kexec_apply_relocations_add arch_kexec_apply_relocations_add
> +
> +struct kimage;
> +int arch_kimage_file_post_load_cleanup(struct kimage *image);
> +#define arch_kimage_file_post_load_cleanup arch_kimage_file_post_load_cleanup
> #endif
>
> #endif
> diff --git a/arch/riscv/kernel/elf_kexec.c b/arch/riscv/kernel/elf_kexec.c
> index 0cb94992c15b..ff30fcb43f47 100644
> --- a/arch/riscv/kernel/elf_kexec.c
> +++ b/arch/riscv/kernel/elf_kexec.c
> @@ -21,6 +21,14 @@
> #include <linux/memblock.h>
> #include <asm/setup.h>
>
> +int arch_kimage_file_post_load_cleanup(struct kimage *image)
> +{
> + kvfree(image->arch.fdt);
> + image->arch.fdt = NULL;
> +
> + return kexec_image_post_load_cleanup_default(image);
> +}

Good catch,LGTM

Reviewed-by: Liao Chang <[email protected]>

> +
> static int riscv_kexec_elf_load(struct kimage *image, struct elfhdr *ehdr,
> struct kexec_elf_info *elf_info, unsigned long old_pbase,
> unsigned long new_pbase)
> @@ -298,6 +306,8 @@ static void *elf_kexec_load(struct kimage *image, char *kernel_buf,
> pr_err("Error add DTB kbuf ret=%d\n", ret);
> goto out_free_fdt;
> }
> + /* Cache the fdt buffer address for memory cleanup */
> + image->arch.fdt = fdt;
> pr_notice("Loaded device tree at 0x%lx\n", kbuf.mem);
> goto out;
>

--
BR,
Liao, Chang

2022-12-05 22:50:07

by Palmer Dabbelt

[permalink] [raw]

Subject: Re: [PATCH 1/2] RISC-V: kexec: Fix memory leak of fdt buffer

On Fri, 4 Nov 2022 17:56:57 +0800, Li Huafei wrote:
> This is reported by kmemleak detector:
>
> unreferenced object 0xff60000082864000 (size 9588):
> comm "kexec", pid 146, jiffies 4294900634 (age 64.788s)
> hex dump (first 32 bytes):
> d0 0d fe ed 00 00 12 ed 00 00 00 48 00 00 11 40 ...........H...@
> 00 00 00 28 00 00 00 11 00 00 00 02 00 00 00 00 ...(............
> backtrace:
> [<00000000f95b17c4>] kmemleak_alloc+0x34/0x3e
> [<00000000b9ec8e3e>] kmalloc_order+0x9c/0xc4
> [<00000000a95cf02e>] kmalloc_order_trace+0x34/0xb6
> [<00000000f01e68b4>] __kmalloc+0x5c2/0x62a
> [<000000002bd497b2>] kvmalloc_node+0x66/0xd6
> [<00000000906542fa>] of_kexec_alloc_and_setup_fdt+0xa6/0x6ea
> [<00000000e1166bde>] elf_kexec_load+0x206/0x4ec
> [<0000000036548e09>] kexec_image_load_default+0x40/0x4c
> [<0000000079fbe1b4>] sys_kexec_file_load+0x1c4/0x322
> [<0000000040c62c03>] ret_from_syscall+0x0/0x2
>
> [...]

Applied, thanks!

[1/2] RISC-V: kexec: Fix memory leak of fdt buffer
https://git.kernel.org/palmer/c/96df59b1ae23
[2/2] RISC-V: kexec: Fix memory leak of elf header buffer
https://git.kernel.org/palmer/c/cbc32023ddbd

These are on for-next. They'd probably be fine fixes candidates had I gotten
to them in time, but it's pretty late and a leak during kexec doesn't seem like
a show-stopper. They'll get backported anyway, but this way they get an extra
week in linux-next just to see.

Best regards,
--
Palmer Dabbelt <[email protected]>

2022-12-05 23:05:24

by patchwork-bot+linux-riscv

[permalink] [raw]

Subject: Re: [PATCH 1/2] RISC-V: kexec: Fix memory leak of fdt buffer

Hello:

This series was applied to riscv/linux.git (for-next)
by Palmer Dabbelt <[email protected]>:

On Fri, 4 Nov 2022 17:56:57 +0800 you wrote:
> This is reported by kmemleak detector:
>
> unreferenced object 0xff60000082864000 (size 9588):
> comm "kexec", pid 146, jiffies 4294900634 (age 64.788s)
> hex dump (first 32 bytes):
> d0 0d fe ed 00 00 12 ed 00 00 00 48 00 00 11 40 ...........H...@
> 00 00 00 28 00 00 00 11 00 00 00 02 00 00 00 00 ...(............
> backtrace:
> [<00000000f95b17c4>] kmemleak_alloc+0x34/0x3e
> [<00000000b9ec8e3e>] kmalloc_order+0x9c/0xc4
> [<00000000a95cf02e>] kmalloc_order_trace+0x34/0xb6
> [<00000000f01e68b4>] __kmalloc+0x5c2/0x62a
> [<000000002bd497b2>] kvmalloc_node+0x66/0xd6
> [<00000000906542fa>] of_kexec_alloc_and_setup_fdt+0xa6/0x6ea
> [<00000000e1166bde>] elf_kexec_load+0x206/0x4ec
> [<0000000036548e09>] kexec_image_load_default+0x40/0x4c
> [<0000000079fbe1b4>] sys_kexec_file_load+0x1c4/0x322
> [<0000000040c62c03>] ret_from_syscall+0x0/0x2
>
> [...]

Here is the summary with links:
- [1/2] RISC-V: kexec: Fix memory leak of fdt buffer
https://git.kernel.org/riscv/c/96df59b1ae23
- [2/2] RISC-V: kexec: Fix memory leak of elf header buffer
https://git.kernel.org/riscv/c/cbc32023ddbd

You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html

2022-12-07 01:59:06

by Li Huafei

[permalink] [raw]

Subject: Re: [PATCH 1/2] RISC-V: kexec: Fix memory leak of fdt buffer

On 2022/12/6 6:28, Palmer Dabbelt wrote:
> On Fri, 4 Nov 2022 17:56:57 +0800, Li Huafei wrote:
>> This is reported by kmemleak detector:
>>
>> unreferenced object 0xff60000082864000 (size 9588):
>> comm "kexec", pid 146, jiffies 4294900634 (age 64.788s)
>> hex dump (first 32 bytes):
>> d0 0d fe ed 00 00 12 ed 00 00 00 48 00 00 11 40 ...........H...@
>> 00 00 00 28 00 00 00 11 00 00 00 02 00 00 00 00 ...(............
>> backtrace:
>> [<00000000f95b17c4>] kmemleak_alloc+0x34/0x3e
>> [<00000000b9ec8e3e>] kmalloc_order+0x9c/0xc4
>> [<00000000a95cf02e>] kmalloc_order_trace+0x34/0xb6
>> [<00000000f01e68b4>] __kmalloc+0x5c2/0x62a
>> [<000000002bd497b2>] kvmalloc_node+0x66/0xd6
>> [<00000000906542fa>] of_kexec_alloc_and_setup_fdt+0xa6/0x6ea
>> [<00000000e1166bde>] elf_kexec_load+0x206/0x4ec
>> [<0000000036548e09>] kexec_image_load_default+0x40/0x4c
>> [<0000000079fbe1b4>] sys_kexec_file_load+0x1c4/0x322
>> [<0000000040c62c03>] ret_from_syscall+0x0/0x2
>>
>> [...]
>
> Applied, thanks!
>
> [1/2] RISC-V: kexec: Fix memory leak of fdt buffer
> https://git.kernel.org/palmer/c/96df59b1ae23
> [2/2] RISC-V: kexec: Fix memory leak of elf header buffer
> https://git.kernel.org/palmer/c/cbc32023ddbd
>
> These are on for-next. They'd probably be fine fixes candidates had I gotten
> to them in time, but it's pretty late and a leak during kexec doesn't seem like
> a show-stopper. They'll get backported anyway, but this way they get an extra
> week in linux-next just to see.
>

I have no problem with it being merged into the -next branch. Thanks Palmer!

> Best regards,
>

2022-12-07 02:06:04

by Li Huafei

[permalink] [raw]