2022-07-04 08:45:27

by Jianglei Nie

[permalink] [raw]
Subject: [PATCH v2] proc/vmcore: fix potential memory leak in vmcore_init()

elfcorehdr_alloc() allocates a memory chunk for elfcorehdr_addr with
kzalloc(). If is_vmcore_usable() returns false, elfcorehdr_addr is a
predefined value. If parse_crash_elf_headers() occurs some error and
returns a negetive value, the elfcorehdr_addr should be released with
elfcorehdr_free().

We can fix by calling elfcorehdr_free() when parse_crash_elf_headers()
fails.

Signed-off-by: Jianglei Nie <[email protected]>
---
fs/proc/vmcore.c | 5 ++++-
1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/fs/proc/vmcore.c b/fs/proc/vmcore.c
index 4eaeb645e759..86887bd90263 100644
--- a/fs/proc/vmcore.c
+++ b/fs/proc/vmcore.c
@@ -1569,7 +1569,7 @@ static int __init vmcore_init(void)
rc = parse_crash_elf_headers();
if (rc) {
pr_warn("Kdump: vmcore not initialized\n");
- return rc;
+ goto fail;
}
elfcorehdr_free(elfcorehdr_addr);
elfcorehdr_addr = ELFCORE_ADDR_ERR;
@@ -1577,6 +1577,9 @@ static int __init vmcore_init(void)
proc_vmcore = proc_create("vmcore", S_IRUSR, NULL, &vmcore_proc_ops);
if (proc_vmcore)
proc_vmcore->size = vmcore_size;
+
+fail:
+ elfcorehdr_free(elfcorehdr_addr);
return 0;
}
fs_initcall(vmcore_init);
--
2.25.1


2022-07-08 08:40:35

by Baoquan He

[permalink] [raw]
Subject: Re: [PATCH v2] proc/vmcore: fix potential memory leak in vmcore_init()

On 07/04/22 at 04:18pm, Jianglei Nie wrote:
> elfcorehdr_alloc() allocates a memory chunk for elfcorehdr_addr with
> kzalloc(). If is_vmcore_usable() returns false, elfcorehdr_addr is a
> predefined value. If parse_crash_elf_headers() occurs some error and
> returns a negetive value, the elfcorehdr_addr should be released with
> elfcorehdr_free().
>
> We can fix by calling elfcorehdr_free() when parse_crash_elf_headers()
> fails.

LGTM,

Acked-by: Baoquan He <[email protected]>

>
> Signed-off-by: Jianglei Nie <[email protected]>
> ---
> fs/proc/vmcore.c | 5 ++++-
> 1 file changed, 4 insertions(+), 1 deletion(-)
>
> diff --git a/fs/proc/vmcore.c b/fs/proc/vmcore.c
> index 4eaeb645e759..86887bd90263 100644
> --- a/fs/proc/vmcore.c
> +++ b/fs/proc/vmcore.c
> @@ -1569,7 +1569,7 @@ static int __init vmcore_init(void)
> rc = parse_crash_elf_headers();
> if (rc) {
> pr_warn("Kdump: vmcore not initialized\n");
> - return rc;
> + goto fail;
> }
> elfcorehdr_free(elfcorehdr_addr);
> elfcorehdr_addr = ELFCORE_ADDR_ERR;
> @@ -1577,6 +1577,9 @@ static int __init vmcore_init(void)
> proc_vmcore = proc_create("vmcore", S_IRUSR, NULL, &vmcore_proc_ops);
> if (proc_vmcore)
> proc_vmcore->size = vmcore_size;
> +
> +fail:
> + elfcorehdr_free(elfcorehdr_addr);
> return 0;
> }
> fs_initcall(vmcore_init);
> --
> 2.25.1
>

2022-07-09 00:40:20

by Baoquan He

[permalink] [raw]
Subject: Re: [PATCH v2] proc/vmcore: fix potential memory leak in vmcore_init()

On 07/08/22 at 04:34pm, Baoquan He wrote:
> On 07/04/22 at 04:18pm, Jianglei Nie wrote:
> > elfcorehdr_alloc() allocates a memory chunk for elfcorehdr_addr with
> > kzalloc(). If is_vmcore_usable() returns false, elfcorehdr_addr is a
> > predefined value. If parse_crash_elf_headers() occurs some error and
> > returns a negetive value, the elfcorehdr_addr should be released with
> > elfcorehdr_free().
> >
> > We can fix by calling elfcorehdr_free() when parse_crash_elf_headers()
> > fails.
>
> LGTM,
>
> Acked-by: Baoquan He <[email protected]>

Sorry, I didn't check the code change carefully. This v2 is not right. I
thought Jianglei took my suggested code change directly. Seems he
mistakenly took part of them and caused error.

>
> >
> > Signed-off-by: Jianglei Nie <[email protected]>
> > ---
> > fs/proc/vmcore.c | 5 ++++-
> > 1 file changed, 4 insertions(+), 1 deletion(-)
> >
> > diff --git a/fs/proc/vmcore.c b/fs/proc/vmcore.c
> > index 4eaeb645e759..86887bd90263 100644
> > --- a/fs/proc/vmcore.c
> > +++ b/fs/proc/vmcore.c
> > @@ -1569,7 +1569,7 @@ static int __init vmcore_init(void)
> > rc = parse_crash_elf_headers();
> > if (rc) {
> > pr_warn("Kdump: vmcore not initialized\n");
> > - return rc;
> > + goto fail;
> > }
> > elfcorehdr_free(elfcorehdr_addr);
~~
this line should be removed.
> > elfcorehdr_addr = ELFCORE_ADDR_ERR;
> > @@ -1577,6 +1577,9 @@ static int __init vmcore_init(void)
> > proc_vmcore = proc_create("vmcore", S_IRUSR, NULL, &vmcore_proc_ops);
> > if (proc_vmcore)
> > proc_vmcore->size = vmcore_size;
> > +
> > +fail:
> > + elfcorehdr_free(elfcorehdr_addr);
> > return 0;
^
return rc; the returned value is wrong with v2.

What I suggested is as below.

Andrew, please drop the one merged into mm-unstable branch.

Hi Jianglei,

Can you post v3 with below correct change and add Andrew to CC?

diff --git a/fs/proc/vmcore.c b/fs/proc/vmcore.c
index 4eaeb645e759..390515c249dd 100644
--- a/fs/proc/vmcore.c
+++ b/fs/proc/vmcore.c
@@ -1569,15 +1569,16 @@ static int __init vmcore_init(void)
rc = parse_crash_elf_headers();
if (rc) {
pr_warn("Kdump: vmcore not initialized\n");
- return rc;
+ goto fail;
}
- elfcorehdr_free(elfcorehdr_addr);
elfcorehdr_addr = ELFCORE_ADDR_ERR;

proc_vmcore = proc_create("vmcore", S_IRUSR, NULL, &vmcore_proc_ops);
if (proc_vmcore)
proc_vmcore->size = vmcore_size;
- return 0;
+fail:
+ elfcorehdr_free(elfcorehdr_addr);
+ return rc;
}
fs_initcall(vmcore_init);