Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:18 as permitted sender) client-ip=2620:137:e000::1:18;
From:   Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To:     linux-kernel@vger.kernel.org
Cc:     Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        stable@vger.kernel.org, Baoquan He <bhe@redhat.com>,
        Dave Young <dyoung@redhat.com>,
        Andrew Morton <akpm@linux-foundation.org>
Subject: [PATCH 5.17 036/772] x86/kexec: fix memory leak of elf header buffer
Date:   Tue,  7 Jun 2022 18:53:48 +0200
Message-Id: <20220607164950.087234411@linuxfoundation.org>
In-Reply-To: <20220607164948.980838585@linuxfoundation.org>
References: <20220607164948.980838585@linuxfoundation.org>
User-Agent: quilt/0.66
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Precedence: bulk

From: Baoquan He <bhe@redhat.com>

commit b3e34a47f98974d0844444c5121aaff123004e57 upstream.

This is reported by kmemleak detector:

unreferenced object 0xffffc900002a9000 (size 4096):
  comm "kexec", pid 14950, jiffies 4295110793 (age 373.951s)
  hex dump (first 32 bytes):
    7f 45 4c 46 02 01 01 00 00 00 00 00 00 00 00 00  .ELF............
    04 00 3e 00 01 00 00 00 00 00 00 00 00 00 00 00  ..>.............
  backtrace:
    [<0000000016a8ef9f>] __vmalloc_node_range+0x101/0x170
    [<000000002b66b6c0>] __vmalloc_node+0xb4/0x160
    [<00000000ad40107d>] crash_prepare_elf64_headers+0x8e/0xcd0
    [<0000000019afff23>] crash_load_segments+0x260/0x470
    [<0000000019ebe95c>] bzImage64_load+0x814/0xad0
    [<0000000093e16b05>] arch_kexec_kernel_image_load+0x1be/0x2a0
    [<000000009ef2fc88>] kimage_file_alloc_init+0x2ec/0x5a0
    [<0000000038f5a97a>] __do_sys_kexec_file_load+0x28d/0x530
    [<0000000087c19992>] do_syscall_64+0x3b/0x90
    [<0000000066e063a4>] entry_SYSCALL_64_after_hwframe+0x44/0xae

In crash_prepare_elf64_headers(), a buffer is allocated via vmalloc() to
store elf headers.  While it's not freed back to system correctly when
kdump kernel is reloaded or unloaded.  Then memory leak is caused.  Fix it
by introducing x86 specific function arch_kimage_file_post_load_cleanup(),
and freeing the buffer there.

And also remove the incorrect elf header buffer freeing code.  Before
calling arch specific kexec_file loading function, the image instance has
been initialized.  So 'image->elf_headers' must be NULL.  It doesn't make
sense to free the elf header buffer in the place.

Three different people have reported three bugs about the memory leak on
x86_64 inside Redhat.

Link: https://lkml.kernel.org/r/20220223113225.63106-2-bhe@redhat.com
Signed-off-by: Baoquan He <bhe@redhat.com>
Acked-by: Dave Young <dyoung@redhat.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
 arch/x86/kernel/machine_kexec_64.c |   12 +++++++++---
 1 file changed, 9 insertions(+), 3 deletions(-)

--- a/arch/x86/kernel/machine_kexec_64.c
+++ b/arch/x86/kernel/machine_kexec_64.c
@@ -374,9 +374,6 @@ void machine_kexec(struct kimage *image)
 #ifdef CONFIG_KEXEC_FILE
 void *arch_kexec_kernel_image_load(struct kimage *image)
 {
-	vfree(image->elf_headers);
-	image->elf_headers = NULL;
-
 	if (!image->fops || !image->fops->load)
 		return ERR_PTR(-ENOEXEC);
 
@@ -512,6 +509,15 @@ overflow:
 	       (int)ELF64_R_TYPE(rel[i].r_info), value);
 	return -ENOEXEC;
 }
+
+int arch_kimage_file_post_load_cleanup(struct kimage *image)
+{
+	vfree(image->elf_headers);
+	image->elf_headers = NULL;
+	image->elf_headers_sz = 0;
+
+	return kexec_image_post_load_cleanup_default(image);
+}
 #endif /* CONFIG_KEXEC_FILE */
 
 static int