Skip to contentSkip to conversation

multiboot: check mh_load_end_addr address field

PATCH In review P J P 2 participants 3 messages Feb 27, 2018 – Mar 6, 2018

lore
Share & tools
lore

P J P

author
#0 lore raw

PATCH 1/1 multiboot: check mh_load_end_addr address field

From: Prasad J Pandit <[email protected]>

While loading kernel via multiboot-v1 image, (flags & 0x00010000) indicates that multiboot header contains valid addresses to load the kernel image. In that, end of the data segment address 'mh_load_end_addr' should be less than the bss segment address 'mh_bss_end_addr'. Add check to validate that.

---
 hw/i386/multiboot.c | 4 ++++
 1 file changed, 4 insertions(+)
hw/i386/multiboot.c +4 -0
index 46d9c68bf5..d16e32bf4a 100644
@@ -227,6 +227,10 @@ int load_multiboot(FWCfgState *fw_cfg,
             fprintf(stderr, "invalid mh_load_addr address\n");
             exit(1);
         }
+        if (mh_load_end_addr > mh_bss_end_addr) {
+            fprintf(stderr, "invalid mh_load_end_addr address\n");
+            exit(1);
+        }
 
         uint32_t mb_kernel_text_offset = i - (mh_header_addr - mh_load_addr);
         uint32_t mb_load_size = 0;
Signature · 1 line
2.14.3
author
#1 lore raw
+-- On Wed, 28 Feb 2018, P J P wrote --+
| While loading kernel via multiboot-v1 image, (flags & 0x00010000)
| indicates that multiboot header contains valid addresses to load
| the kernel image. In that, end of the data segment address
| 'mh_load_end_addr' should be less than the bss segment address
| 'mh_bss_end_addr'. Add check to validate that.
| 
| +        if (mh_load_end_addr > mh_bss_end_addr) {
| +            fprintf(stderr, "invalid mh_load_end_addr address\n");
| +            exit(1);
| +        }
|  

Ping...!
--
Prasad J Pandit / Red Hat Product Security Team
47AF CE69 3A90 54AA 9045 1053 DD13 3D32 FE5B 041F

#2 lore raw
In reply to P J P

Hi Prasad.

The Multiboot Spec will allow for a zero bss end address.  (Please see section 3.1.3 at https://www.gnu.org/software/grub/manual/multiboot/  . ) For a zero bss end address, this patch will not do the right thing.

I had proposed some patches to properly handle zero bss end address, but they got delayed.  However, I will be re-sending updated patches out later today.  Please stay tuned...

    Thanks,
    Jack

Quoted 29 lines · P J P

From: Prasad J Pandit <[email protected]>

While loading kernel via multiboot-v1 image, (flags & 0x00010000) indicates that multiboot header contains valid addresses to load the kernel image. In that, end of the data segment address 'mh_load_end_addr' should be less than the bss segment address 'mh_bss_end_addr'. Add check to validate that.

---
  hw/i386/multiboot.c | 4 ++++
  1 file changed, 4 insertions(+)
hw/i386/multiboot.c +4 -0
index 46d9c68bf5..d16e32bf4a 100644
@@ -227,6 +227,10 @@ int load_multiboot(FWCfgState *fw_cfg,
              fprintf(stderr, "invalid mh_load_addr address\n");
              exit(1);
          }
+        if (mh_load_end_addr > mh_bss_end_addr) {
+            fprintf(stderr, "invalid mh_load_end_addr address\n");
+            exit(1);
+        }
 
          uint32_t mb_kernel_text_offset = i - (mh_header_addr - mh_load_addr);
          uint32_t mb_load_size = 0;