Skip to contentSkip to conversation

fdc: check for illegal dma length calculation

PATCH In review Jon Maloy 2 participants 3 messages Jan 14, 2022 – Jan 26, 2022

lore
Share & tools
lore

Jon Maloy

author
#0 lore raw

PATCH 1/1 fdc: check for illegal dma length calculation

The function fdctrl_start_transfer() calculates the dma data length wrongly when certain boundary conditions are fulfilled. We have noticed that the if ((fdctrl->fifo[5] - fdctrl->fifo[6]) > 1) we get a dma length that will be interpreted as negative by the next function in the chain, fdctrl_transfer_handler(). This leads to a crash.

Rather than trying to fix this obscure calculation, we just check if the harmful condition is fulfilled, and return without action if that is the case. Since this is a condition that can only be created by a malicious user we deem this solution safe.

This fix is intended to address CVE-2021-3507.

---
 hw/block/fdc.c | 5 +++++
 1 file changed, 5 insertions(+)
hw/block/fdc.c +5 -0
index 21d18ac2e3..80a1f1750a 100644
@@ -1532,6 +1532,11 @@ static void fdctrl_start_transfer(FDCtrl *fdctrl, int direction)
         if (fdctrl->fifo[0] & 0x80)
             tmp += fdctrl->fifo[6];
         fdctrl->data_len *= tmp;
+        if (tmp < 0) {
+            FLOPPY_DPRINTF("calculated illegal data_len=%u, tmp=%i\n",
+                           fdctrl->data_len, tmp);
+            return;
+        }
     }
     fdctrl->eot = fdctrl->fifo[6];
     if (fdctrl->dor & FD_DOR_DMAEN) {
Signature · 1 line
2.31.1
author
#1 lore raw
Quoted 24 lines · Jon Maloy

The function fdctrl_start_transfer() calculates the dma data length wrongly when certain boundary conditions are fulfilled. We have noticed that the if ((fdctrl->fifo[5] - fdctrl->fifo[6]) > 1) we get a dma length that will be interpreted as negative by the next function in the chain, fdctrl_transfer_handler(). This leads to a crash.

Rather than trying to fix this obscure calculation, we just check if the harmful condition is fulfilled, and return without action if that is the case. Since this is a condition that can only be created by a malicious user we deem this solution safe.

This fix is intended to address CVE-2021-3507.

---
  hw/block/fdc.c | 5 +++++
  1 file changed, 5 insertions(+)
hw/block/fdc.c +0 -0
index 21d18ac2e3..80a1f1750a 100644
@@ -1532,6 +1532,11 @@ static void fdctrl_start_transfer(FDCtrl *fdctrl, int direction)
          if (fdctrl->fifo[0] & 0x80)

hw/block/fdc.c static void fdctrl_start_transfer(FDCtrl *fdctrl, int direction) · line 1542

Quoted 10 lines
hw/block/fdc.c +5 -0
              tmp += fdctrl->fifo[6];
          fdctrl->data_len *= tmp;
+        if (tmp < 0) {
+            FLOPPY_DPRINTF("calculated illegal data_len=%u, tmp=%i\n",
+                           fdctrl->data_len, tmp);
+            return;
+        }
      }
      fdctrl->eot = fdctrl->fifo[6];
      if (fdctrl->dor & FD_DOR_DMAEN) {

I never received any feedback on this one.
Is there any?

///jon

#2 lore raw

Hi Jon,

Quoted 6 lines · Jon Maloy
Quoted 5 lines · Jon Maloy

The function fdctrl_start_transfer() calculates the dma data length wrongly when certain boundary conditions are fulfilled. We have noticed that the if ((fdctrl->fifo[5] - fdctrl->fifo[6]) > 1) we get a dma length that will be interpreted as negative by the next function in the chain, fdctrl_transfer_handler(). This leads to a crash.

If a security issue is reproducible (like the ones found by fuzzers), please include the reproducer along.

Quoted 6 lines
Quoted 6 lines

Rather than trying to fix this obscure calculation, we just check if the harmful condition is fulfilled, and return without action if that is the case. Since this is a condition that can only be created by a malicious user we deem this solution safe.

This fix is intended to address CVE-2021-3507.

Ah, this is similar to:
https://lore.kernel.org/qemu-devel/[email protected]/
which already contains the reproducer...
You might want to review it ;)

Quoted 16 lines
Quoted 16 lines

---
  hw/block/fdc.c | 5 +++++

  1 file changed, 5 insertions(+)
hw/block/fdc.c +2 -0
index 21d18ac2e3..80a1f1750a 100644
@@ -1532,6 +1532,11 @@ static void fdctrl_start_transfer(FDCtrl 
 *fdctrl, int direction)
           if (fdctrl->fifo[0] & 0x80)
               tmp += fdctrl->fifo[6];
           fdctrl->data_len *= tmp;
+        if (tmp < 0) {
+            FLOPPY_DPRINTF("calculated illegal data_len=%u, tmp=%i\n",
Quoted 8 lines
Quoted 6 lines
hw/block/fdc.c +3 -0
+                           fdctrl->data_len, tmp);
+            return;
+        }
       }
       fdctrl->eot = fdctrl->fifo[6];

      if (fdctrl->dor & FD_DOR_DMAEN) {

I never received any feedback on this one.
Is there any?

Probably none so far because you forgot to Cc the maintainers:

$ ./scripts/get_maintainer.pl -f hw/block/fdc.c
John Snow <[email protected]> (supporter:Floppy)
Kevin Wolf <[email protected]> (supporter:Block layer core)
Hanna Reitz <[email protected]> (supporter:Block layer core)
[email protected] (open list:Floppy)
[email protected] (open list:All patches CC here)

(now Cc'ed).