Skip to contentSkip to conversation

vmcore: prevent PT_NOTE p_memsz overflow during header update

PATCH In review Greg Pearson 2 participants 5 messages Feb 4, 2014 – Feb 5, 2014

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH v3 1/1 vmcore: prevent PT_NOTE p_memsz overflow during header update

Currently, update_note_header_size_elf64() and update_note_header_size_elf32() will add the size of a PT_NOTE entry to real_sz even if that causes real_sz to exceeds max_sz. This patch corrects the while loop logic in those routines to ensure that does not happen and prints a warning if a PT_NOTE entry is dropped. If zero PT_NOTE entries are found or this condition is encountered because the only entry was dropped, a warning is printed and an error is returned.

One possible negative side effect of exceeding the max_sz limit is an allocation failure in merge_note_headers_elf64() or merge_note_headers_elf32() which would produce console output such as the following while booting the crash kernel.

vmalloc: allocation failure: 14076997632 bytes swapper/0: page allocation failure: order:0, mode:0x80d2

CPU: 0 PID: 1 Comm: swapper/0 Not tainted 3.10.0-gbp1 #7
ffffffff817dcc30 ffff88003025fc28 ffffffff815bdb0b ffff88003025fcb0
ffffffff8113b3d0 ffffffff817dcc30 ffff88003025fc48 ffffc90000000018
ffff88003025fcc0 ffff88003025fc60 ffff88003025fc80 ffff88002b5df980
Call Trace:
[<ffffffff815bdb0b>] dump_stack+0x19/0x1b
[<ffffffff8113b3d0>] warn_alloc_failed+0xf0/0x160
[<ffffffff81a1267d>] ? merge_note_headers_elf64.constprop.9+0x116/0x24a
[<ffffffff8116d34e>] __vmalloc_node_range+0x19e/0x250
[<ffffffff81210454>] ? read_from_oldmem.part.0+0xa4/0xe0
[<ffffffff8116d4ec>] vmalloc_user+0x4c/0x70
[<ffffffff81a1267d>] ? merge_note_headers_elf64.constprop.9+0x116/0x24a
[<ffffffff81a1267d>] merge_note_headers_elf64.constprop.9+0x116/0x24a
[<ffffffff81a12cce>] vmcore_init+0x2d4/0x76c
[<ffffffff8120f9f0>] ? kcore_update_ram+0x1f0/0x1f0
[<ffffffff81063b92>] ? walk_system_raange+0x112/0x130
[<ffffffff81a129fa>] ? merge_note_headers_elf32.constprop.8+0x249/0x249
[<ffffffff810020e2>] do_one_initcall+0xe2/0x190
[<ffffffff819e20c4>] kernel_init_freeable+0x17c/0x207
[<ffffffff819e18d0>] ? do_early_param+0x88/0x88
[<ffffffff815a0d20>] ? rest_init+0x80/0x80
[<ffffffff815a0d2e>] kernel_init+0xe/0x180
[<ffffffff815cd8ac>] ret_from_fork+0x7c/0xb0
[<ffffffff815a0d20>] ? rest_init+0x80/0x80

Kdump: vmcore not initialized

kdump: dump target is /dev/sda4
kdump: saving to /sysroot//var/crash/127.0.0.1-2014.01.28-13:58:52/
kdump: saving vmcore-dmesg.txt
Cannot open /proc/vmcore: No such file or directory
kdump: saving vmcore-dmesg.txt failed
kdump: saving vmcore
kdump: saving vmcore failed

This type of failure has been seen on a four socket prototype system with certain memory configurations. Most PT_NOTE sections have a single entry similar to:

n_namesz = 0x5
n_descsz = 0x150
n_type = 0x1

Occasionally, a second entry is encountered with very large n_namesz and n_descsz sizes:

n_namesz = 0x80000008
n_descsz = 0x510ae163
n_type = 0x80000008

Not yet sure of the source of these extra entries, they seem bogus, but they shouldn't cause crash dump to fail.

---
 fs/proc/vmcore.c | 26 ++++++++++++++++++++------
 1 file changed, 20 insertions(+), 6 deletions(-)
fs/proc/vmcore.c +20 -6
index 2ca7ba0..88d4585 100644
@@ -468,17 +468,24 @@ static int __init update_note_header_size_elf64(const Elf64_Ehdr *ehdr_ptr)
 			return rc;
 		}
 		nhdr_ptr = notes_section;
-		while (real_sz < max_sz) {
-			if (nhdr_ptr->n_namesz == 0)
-				break;
+		while (nhdr_ptr->n_namesz != 0) {
 			sz = sizeof(Elf64_Nhdr) +
 				((nhdr_ptr->n_namesz + 3) & ~3) +
 				((nhdr_ptr->n_descsz + 3) & ~3);
+			if ((real_sz + sz) > max_sz) {
+				pr_warn("Warning: Exceeded p_memsz, dropping PT_NOTE entry n_namesz=0x%x, n_descsz=0x%x\n",
+					nhdr_ptr->n_namesz, nhdr_ptr->n_descsz);
+				break;
+			}
 			real_sz += sz;
 			nhdr_ptr = (Elf64_Nhdr*)((char*)nhdr_ptr + sz);
 		}
 		kfree(notes_section);
 		phdr_ptr->p_memsz = real_sz;
+		if (real_sz == 0) {
+			pr_warn("Warning: Zero PT_NOTE entries found\n");
+			return -EINVAL;
+		}
 	}
 
 	return 0;
@@ -648,17 +655,24 @@ static int __init update_note_header_size_elf32(const Elf32_Ehdr *ehdr_ptr)
 			return rc;
 		}
 		nhdr_ptr = notes_section;
-		while (real_sz < max_sz) {
-			if (nhdr_ptr->n_namesz == 0)
-				break;
+		while (nhdr_ptr->n_namesz != 0) {
 			sz = sizeof(Elf32_Nhdr) +
 				((nhdr_ptr->n_namesz + 3) & ~3) +
 				((nhdr_ptr->n_descsz + 3) & ~3);
+			if ((real_sz + sz) > max_sz) {
+				pr_warn("Warning: Exceeded p_memsz, dropping PT_NOTE entry n_namesz=0x%x, n_descsz=0x%x\n",
+					nhdr_ptr->n_namesz, nhdr_ptr->n_descsz);
+				break;
+			}
 			real_sz += sz;
 			nhdr_ptr = (Elf32_Nhdr*)((char*)nhdr_ptr + sz);
 		}
 		kfree(notes_section);
 		phdr_ptr->p_memsz = real_sz;
+		if (real_sz == 0) {
+			pr_warn("Warning: Zero PT_NOTE entries found\n");
+			return -EINVAL;
+		}
 	}
 
 	return 0;
Signature · 1 line
1.8.3.2
#1 lore raw

On Tue, Feb 04, 2014 at 04:25:52PM -0700, Greg Pearson wrote:

[..]

Quoted 8 lines
fs/proc/vmcore.c +0 -0
index 2ca7ba0..88d4585 100644
@@ -468,17 +468,24 @@ static int __init update_note_header_size_elf64(const Elf64_Ehdr *ehdr_ptr)
 			return rc;
 		}
 		nhdr_ptr = notes_section;

fs/proc/vmcore.c static int __init update_note_header_size_elf64(const Elf64_Ehdr *ehdr_ptr) · line 477

Quoted 10 lines
fs/proc/vmcore.c +4 -3
-		while (real_sz < max_sz) {
-			if (nhdr_ptr->n_namesz == 0)
-				break;
+		while (nhdr_ptr->n_namesz != 0) {
 			sz = sizeof(Elf64_Nhdr) +
 				((nhdr_ptr->n_namesz + 3) & ~3) +
 				((nhdr_ptr->n_descsz + 3) & ~3);
+			if ((real_sz + sz) > max_sz) {
+				pr_warn("Warning: Exceeded p_memsz, dropping PT_NOTE entry n_namesz=0x%x, n_descsz=0x%x\n",
+					nhdr_ptr->n_namesz, nhdr_ptr->n_descsz);

You will need line break in pr_warn(). Too long a line. Limit it 80 columns per line.

fs/proc/vmcore.c static int __init update_note_header_size_elf64(const Elf64_Ehdr *ehdr_ptr) · line 487

Quoted 10 lines
fs/proc/vmcore.c +5 -0
+				break;
+			}
 			real_sz += sz;
 			nhdr_ptr = (Elf64_Nhdr*)((char*)nhdr_ptr + sz);
 		}
 		kfree(notes_section);
 		phdr_ptr->p_memsz = real_sz;
+		if (real_sz == 0) {
+			pr_warn("Warning: Zero PT_NOTE entries found\n");
+			return -EINVAL;

Given the fact that this is the first time I have heard about a PT_NOTE being corrup, I will be fine with this patch too. If one encounters an empty PT_NOTE, error out and don't create vmcore.

So please repost this patch with line lenght fixed.

Thanks
Vivek

#2 lore raw
Quoted 40 lines · Vivek Goyal

On Tue, Feb 04, 2014 at 04:25:52PM -0700, Greg Pearson wrote:

[..]

Quoted 8 lines
fs/proc/vmcore.c +0 -0
index 2ca7ba0..88d4585 100644
@@ -468,17 +468,24 @@ static int __init update_note_header_size_elf64(const Elf64_Ehdr *ehdr_ptr)
 			return rc;
 		}
 		nhdr_ptr = notes_section;

fs/proc/vmcore.c static int __init update_note_header_size_elf64(const Elf64_Ehdr *ehdr_ptr) · line 477

Quoted 10 lines
fs/proc/vmcore.c +4 -3
-		while (real_sz < max_sz) {
-			if (nhdr_ptr->n_namesz == 0)
-				break;
+		while (nhdr_ptr->n_namesz != 0) {
 			sz = sizeof(Elf64_Nhdr) +
 				((nhdr_ptr->n_namesz + 3) & ~3) +
 				((nhdr_ptr->n_descsz + 3) & ~3);
+			if ((real_sz + sz) > max_sz) {
+				pr_warn("Warning: Exceeded p_memsz, dropping PT_NOTE entry n_namesz=0x%x, n_descsz=0x%x\n",
+					nhdr_ptr->n_namesz, nhdr_ptr->n_descsz);

You will need line break in pr_warn(). Too long a line. Limit it 80 columns per line.

fs/proc/vmcore.c static int __init update_note_header_size_elf64(const Elf64_Ehdr *ehdr_ptr) · line 487

Quoted 10 lines
fs/proc/vmcore.c +5 -0
+				break;
+			}
 			real_sz += sz;
 			nhdr_ptr = (Elf64_Nhdr*)((char*)nhdr_ptr + sz);
 		}
 		kfree(notes_section);
 		phdr_ptr->p_memsz = real_sz;
+		if (real_sz == 0) {
+			pr_warn("Warning: Zero PT_NOTE entries found\n");
+			return -EINVAL;

Given the fact that this is the first time I have heard about a PT_NOTE being corrup, I will be fine with this patch too. If one encounters an empty PT_NOTE, error out and don't create vmcore.

Quoted 1 line

So please repost this patch with line lenght fixed.

Hi Greg,

Is there any more debugging info on why a PT_NOTE is getting corrupt. Did you get a chance to debug it further. This patch is more of a band-aid and real problem is why note is getting corrupted and we need to chase that down and fix that.

Thanks
Vivek

#3 lore raw
In reply to Vivek Goyal
Quoted 15 lines · Vivek Goyal

On Tue, Feb 04, 2014 at 04:25:52PM -0700, Greg Pearson wrote:

[..]

Quoted 12 lines
fs/proc/vmcore.c +1 -3
index 2ca7ba0..88d4585 100644
@@ -468,17 +468,24 @@ static int __init update_note_header_size_elf64(const Elf64_Ehdr *ehdr_ptr)
  			return rc;
  		}
  		nhdr_ptr = notes_section;
-		while (real_sz < max_sz) {
-			if (nhdr_ptr->n_namesz == 0)
-				break;
+		while (nhdr_ptr->n_namesz != 0) {
Quoted 8 lines

fs/proc/vmcore.c static int __init update_note_header_size_elf64(const Elf64_Ehdr *ehdr_ptr) · line 477

Quoted 6 lines
fs/proc/vmcore.c +3 -0
  			sz = sizeof(Elf64_Nhdr) +
  				((nhdr_ptr->n_namesz + 3) & ~3) +
  				((nhdr_ptr->n_descsz + 3) & ~3);
+			if ((real_sz + sz) > max_sz) {
+				pr_warn("Warning: Exceeded p_memsz, dropping PT_NOTE entry n_namesz=0x%x, n_descsz=0x%x\n",
+					nhdr_ptr->n_namesz, nhdr_ptr->n_descsz);

You will need line break in pr_warn(). Too long a line. Limit it 80 columns per line.

The checkpatch.pl script issues a warning when you break quoted strings, I have no personal preference. Just let me know if you want me to ignore the checkpatch warning and conform to the 80 column per line limit.

Also, I'm still debugging the reason why the corrupt entries show up, I'll continue to work on that.

Thanks

--
Greg

Quoted 19 lines

fs/proc/vmcore.c static int __init update_note_header_size_elf64(const Elf64_Ehdr *ehdr_ptr) · line 487

Quoted 11 lines
fs/proc/vmcore.c +5 -0
+				break;
+			}
  			real_sz += sz;
  			nhdr_ptr = (Elf64_Nhdr*)((char*)nhdr_ptr + sz);
  		}
  		kfree(notes_section);
  		phdr_ptr->p_memsz = real_sz;
+		if (real_sz == 0) {
+			pr_warn("Warning: Zero PT_NOTE entries found\n");
+			return -EINVAL;

Given the fact that this is the first time I have heard about a PT_NOTE being corrup, I will be fine with this patch too. If one encounters an empty PT_NOTE, error out and don't create vmcore.

So please repost this patch with line lenght fixed.

Thanks
Vivek

Acked-by
#4 lore raw
Quoted 25 lines · Greg
Quoted 15 lines · Vivek Goyal

On Tue, Feb 04, 2014 at 04:25:52PM -0700, Greg Pearson wrote:

[..]

Quoted 12 lines
fs/proc/vmcore.c +1 -3
index 2ca7ba0..88d4585 100644
@@ -468,17 +468,24 @@ static int __init update_note_header_size_elf64(const Elf64_Ehdr *ehdr_ptr)
  			return rc;
  		}
  		nhdr_ptr = notes_section;
-		while (real_sz < max_sz) {
-			if (nhdr_ptr->n_namesz == 0)
-				break;
+		while (nhdr_ptr->n_namesz != 0) {
Quoted 8 lines

fs/proc/vmcore.c static int __init update_note_header_size_elf64(const Elf64_Ehdr *ehdr_ptr) · line 477

Quoted 6 lines
fs/proc/vmcore.c +3 -0
  			sz = sizeof(Elf64_Nhdr) +
  				((nhdr_ptr->n_namesz + 3) & ~3) +
  				((nhdr_ptr->n_descsz + 3) & ~3);
+			if ((real_sz + sz) > max_sz) {
+				pr_warn("Warning: Exceeded p_memsz, dropping PT_NOTE entry n_namesz=0x%x, n_descsz=0x%x\n",
+					nhdr_ptr->n_namesz, nhdr_ptr->n_descsz);

You will need line break in pr_warn(). Too long a line. Limit it 80 columns per line.

Quoted 3 lines

The checkpatch.pl script issues a warning when you break quoted strings, I have no personal preference. Just let me know if you want me to ignore the checkpatch warning and conform to the 80 column per line limit.

Hmm.., I am reading CodingStyle and it does say that do not break user visible strings as it breaks ability to grep these.

"However, never break user-visible strings such as printk messages, because that breaks the ability to grep for them."

Ok, so I am fine with this patch.

Thanks
Vivek