Skip to contentSkip to conversation

hw/adc: Fix out-of-bounds write in Aspeed ADC model

PATCH In review Cédric Le Goater 6 participants 8 messages Jan 26 – Feb 5 Arm/Aspeed machine support

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 hw/adc: Fix out-of-bounds write in Aspeed ADC model

The 'regs' array has ASPEED_ADC_NR_REGS (52) elements, while the memory region covers offsets 0x00-0xFC. The aspeed_adc_engine_write() function has an out-of-bounds write vulnerability when accessing unimplemented registers.

Fix this by using 'return' instead of 'break' in the default case, which prevents execution from reaching the s->regs[reg] assignment for unimplemented registers.

Fixes 5857974d5d11 ("hw/adc: Add basic Aspeed ADC model")
---
 hw/adc/aspeed_adc.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
hw/adc/aspeed_adc.c +2 -1
index fd3af308296e..3cc75bbcd6e2 100644
@@ -228,7 +228,8 @@ static void aspeed_adc_engine_write(void *opaque, hwaddr addr, uint64_t value,
         qemu_log_mask(LOG_UNIMP, "%s: engine[%u]: "
                       "0x%" HWADDR_PRIx " 0x%" PRIx64 "\n",
                       __func__, s->engine_id, addr, value);
-        break;
+        /* Do not update the regs[] array */
+        return;
     }
 
     s->regs[reg] = value;
Signature · 1 line
2.52.0
#1 lore raw
Quoted 6 lines · Cédric Le Goater

The 'regs' array has ASPEED_ADC_NR_REGS (52) elements, while the memory region covers offsets 0x00-0xFC. The aspeed_adc_engine_write() function has an out-of-bounds write vulnerability when accessing unimplemented registers.

Quoted 7 lines

Fix this by using 'return' instead of 'break' in the default case, which prevents execution from reaching the s->regs[reg] assignment for unimplemented registers.

Fixes 5857974d5d11 ("hw/adc: Add basic Aspeed ADC model")

Seems like a good candidate for Cc: [email protected] ?

thanks
-- PMM

#2 lore raw
Quoted 11 lines · Peter Maydell
Quoted 10 lines · Cédric Le Goater

The 'regs' array has ASPEED_ADC_NR_REGS (52) elements, while the memory region covers offsets 0x00-0xFC. The aspeed_adc_engine_write() function has an out-of-bounds write vulnerability when accessing unimplemented registers.

Fix this by using 'return' instead of 'break' in the default case, which prevents execution from reaching the s->regs[reg] assignment for unimplemented registers.

Quoted 5 lines
Quoted 3 lines
Fixes 5857974d5d11 ("hw/adc: Add basic Aspeed ADC model")

Seems like a good candidate for Cc: [email protected] ?

Yes.

Thanks,

C.

#3 lore raw

Hello, gentlemen,

Is there any possibility of making an exception and assigning a CVE for the issue ?? (It's my first ever found vulnerability :)

Thank you in advance.

Quoted 26 lines · Cédric Le Goater
Quoted 11 lines · Peter Maydell
Quoted 10 lines · Cédric Le Goater

The 'regs' array has ASPEED_ADC_NR_REGS (52) elements, while the memory region covers offsets 0x00-0xFC. The aspeed_adc_engine_write() function has an out-of-bounds write vulnerability when accessing unimplemented registers.

Fix this by using 'return' instead of 'break' in the default case, which prevents execution from reaching the s->regs[reg] assignment for unimplemented registers.

Quoted 5 lines
Quoted 3 lines
Fixes 5857974d5d11 ("hw/adc: Add basic Aspeed ADC model")

Seems like a good candidate for Cc: [email protected] ?

Yes.

Thanks,

C.

Signature · 5 lines
*Saad ELHARAJ*
Cybersecurity Specialist
📞 +212 6 80 51 11 85
✉️ [email protected]
🔗 LinkedIn <https://www.linkedin.com/in/saad-elharaj-1891371a0/>
#4 lore raw
Quoted 2 lines · Elhrj Saad

Is there any possibility of making an exception and assigning a CVE for the issue ?? (It's my first ever found vulnerability :)

Hi; this is not a security vulnerability according to QEMU's security policy, which you can find here:
https://www.qemu.org/docs/master/system/security.html

The board that uses this device is not one usable under the "virtualization" use case, and so we do not consider that bugs in it are security issues. So it's not a "vulnerability", just a normal bug, and we would not issue a CVE for it.

thanks
-- PMM

#5 lore raw
In reply to Elhrj Saad
Quoted 4 lines · Elhrj Saad

Hello, gentlemen,

Is there any possibility of making an exception and assigning a CVE for the issue ?? (It's my first ever found vulnerability :)

Thank you for your contribution to improving the quality of QEMU. In many projects this would indeed qualify for assignment of a CVE, but as Peter mentions, this falls outside the scope of QEMU's security policy.

Assigning CVEs has quite a ripple effect on those consuming QEMU, due to it triggering many security response processes. As such we do not wish to issue CVEs for things that the project considers to merely be hardening bug fixes in an area of code that is not designed to be providing a security boundary.

With regards,
Daniel

Signature · 3 lines
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|
Reviewed-by on v1 1/1
#6 lore raw
In reply to Cédric Le Goater
Quoted 9 lines · Cédric Le Goater

The 'regs' array has ASPEED_ADC_NR_REGS (52) elements, while the memory region covers offsets 0x00-0xFC. The aspeed_adc_engine_write() function has an out-of-bounds write vulnerability when accessing unimplemented registers.

Fix this by using 'return' instead of 'break' in the default case, which prevents execution from reaching the s->regs[reg] assignment for unimplemented registers.

Quoted 6 lines
Fixes 5857974d5d11 ("hw/adc: Add basic Aspeed ADC model")
---
  hw/adc/aspeed_adc.c | 3 ++-
  1 file changed, 2 insertions(+), 1 deletion(-)
#7 lore raw
In reply to Cédric Le Goater
Quoted 9 lines · Cédric Le Goater

The 'regs' array has ASPEED_ADC_NR_REGS (52) elements, while the memory region covers offsets 0x00-0xFC. The aspeed_adc_engine_write() function has an out-of-bounds write vulnerability when accessing unimplemented registers.

Fix this by using 'return' instead of 'break' in the default case, which prevents execution from reaching the s->regs[reg] assignment for unimplemented registers.

Quoted 6 lines
Fixes 5857974d5d11 ("hw/adc: Add basic Aspeed ADC model")
---
  hw/adc/aspeed_adc.c | 3 ++-
  1 file changed, 2 insertions(+), 1 deletion(-)

This feels like a qemu-stable material.

Please let me know if it isn't.

Thanks,

/mjt

Quoted 14 lines
hw/adc/aspeed_adc.c +2 -1
index fd3af308296e..3cc75bbcd6e2 100644
@@ -228,7 +228,8 @@ static void aspeed_adc_engine_write(void *opaque, hwaddr addr, uint64_t value,
          qemu_log_mask(LOG_UNIMP, "%s: engine[%u]: "
                        "0x%" HWADDR_PRIx " 0x%" PRIx64 "\n",
                        __func__, s->engine_id, addr, value);
-        break;
+        /* Do not update the regs[] array */
+        return;
      }
 
      s->regs[reg] = value;