Swatantra-66 opened a new pull request, #21785:
URL: https://github.com/apache/nuttx/pull/21785

   ## Summary
   
   POSIX IEEE Std 1003.1 explicitly mandates for both `sendmsg()` and 
`recvmsg()`:
   > "[EMSGSIZE] The msg_iovlen member of the msghdr structure is less than or 
equal to 0 or is greater than {IOV_MAX}."
   
   In `net/netlink` (`netlink_sockif.c`), several vulnerabilities and POSIX 
non-compliances existed:
   
   1. `msg->msg_iov->iov_base` was dereferenced at the beginning of 
`netlink_sendmsg()` and `netlink_recvmsg()` before validating `msg_iovlen` or 
checking `msg_iov != NULL`. Passing `msg_iovlen == 0` or `msg_iov == NULL` 
resulted in immediate NULL pointer / out-of-bounds dereferences (CWE-476 / 
CWE-125).
   2. `msg_iovlen` was not validated against `<= 0 || > IOV_MAX` (returning 
`-EMSGSIZE` per POSIX).
   3. If `msg->msg_iov[0].iov_len < sizeof(struct nlmsghdr)`, or if 
`nlmsg->nlmsg_len < sizeof(struct nlmsghdr) || nlmsg->nlmsg_len > 
msg->msg_iov[0].iov_len`, `netlink_sendmsg()` would trigger assertion panics or 
allow protocol handlers to read out-of-bounds memory.
   4. `netlink_recvmsg()` did not clear `msg->msg_flags` or set `MSG_TRUNC` 
when the received Netlink response exceeded the user-provided buffer capacity, 
nor did it guard `memcpy()` against calling with a NULL buffer when `len == 0`.
   
   This patch addresses these issues:
   - Defers dereferencing `msg->msg_iov` until after validating `msg_iovlen`, 
`msg_iov != NULL`, and buffer pointers.
   - Validates `msg_iovlen == 0 || msg_iovlen > IOV_MAX` returning `-EMSGSIZE`.
   - Preserves `-ENOTSUP` when `msg_iovlen != 1`.
   - Validates `msg_iov == NULL` and `iov_len > 0 && iov_base == NULL` 
returning `-EINVAL`.
   - Validates that `msg_iov[0].iov_len >= sizeof(struct nlmsghdr)` and that 
`nlmsg->nlmsg_len >= sizeof(struct nlmsghdr) && nlmsg->nlmsg_len <= 
msg_iov[0].iov_len` before parsing in `netlink_sendmsg()`, returning `-EINVAL` 
on malformed headers.
   - Clears `msg->msg_flags = 0` and sets `MSG_TRUNC` when `len < 
entry->msg.nlmsg_len` in `netlink_recvmsg()`.
   - Guards `memcpy()` with `if (len > 0)` to prevent undefined behavior with 
0-length NULL user buffers.
   - Includes `<limits.h>` for `IOV_MAX`.
   
   ---
   
   ## Before vs After
   
   | Scenario | Before | After |
   |---|---|---|
   | `msg_iovlen == 0` or `msg_iov == NULL` | Immediate NULL/OOB dereference in 
local variable initialization (CWE-476 / CWE-125) | Safe validation; returns 
`-EMSGSIZE` or `-EINVAL` before any dereference |
   | `msg_iovlen > IOV_MAX` | Returned `-ENOTSUP` | Returns `-EMSGSIZE` per 
POSIX IEEE Std 1003.1 |
   | `iov_len > 0 && iov_base == NULL` | Unchecked, crashes later during buffer 
access | Rejects upfront with `-EINVAL` |
   | `sendmsg` with truncated Netlink header (`iov_len < sizeof(nlmsghdr)`) | 
Dereferences `nlmsg->nlmsg_len` on uninitialized memory; assertion failure or 
panic | Rejects upfront with `-EINVAL` |
   | `sendmsg` with invalid `nlmsg_len` (`nlmsg_len > iov_len`) | Assertion 
failure in debug mode, heap out-of-bounds read in release mode | Rejects 
upfront with `-EINVAL` |
   | `recvmsg` user buffer smaller than response | Truncated silently without 
notifying caller; `msg_flags` untouched | Sets `MSG_TRUNC` in `msg->msg_flags` 
per POSIX |
   | `recvmsg` with zero-length buffer (`len == 0`) | Called `memcpy(NULL, 
...)` leading to undefined behavior | Guarded with `if (len > 0)`; safely frees 
entry and returns 0 |
   
   ---
   
   ## Impact
   
   - Security: Eliminates potential CWE-476 NULL pointer dereferences and 
CWE-125 out-of-bounds reads.
   - POSIX Compliance: Conforms to POSIX IEEE Std 1003.1 `[EMSGSIZE]` and 
`MSG_TRUNC` specifications.
   - Compatibility: Preserves full backward compatibility for valid Netlink 
sockets.
   
   ---
   
   ## Testing
   
   ### 1. Code Style Check (`nxstyle`)
   ```bash
   $ ./tools/nxstyle net/netlink/netlink_sockif.c
   ```
   Output:
   ```text
   (0 errors, 0 warnings)
   ```
   
   ### 2. Full Build (`sim:dynconns` from `tools/ci/testlist/sim-02.dat`)
   `sim:dynconns` enables `CONFIG_NET_NETLINK=y` and `CONFIG_NETLINK_ROUTE=y`.
   
   ```bash
   $ ./tools/configure.sh sim:dynconns
   $ make -j8
   ```
   Output:
   ```text
   Create version.h
   ...
   CC:  netlink/netlink_sockif.c
   ...
   IN: libs/libc/libc.a -> staging/libc.a
   CC:  symtab_apps.c
   IN: /repos/apps/libapps.a -> staging/libapps.a
   CC:  sim/sim_head.c
   ...
   LD:  nuttx
   Pac SIM with dynamic libs..
   '/lib/x86_64-linux-gnu/libc.so.6' -> 'sim-pac/libs/libc.so.6'
   '/lib64/ld-linux-x86-64.so.2' -> 'sim-pac/ld-linux-x86-64.so.2'
   SIM elf with dynamic libs archive in nuttx.tgz
   ```
   
   Binary size:
   ```bash
   $ size nuttx
      text         data     bss     dec     hex filename
    599545         5244   46752  651541   9f115 nuttx
   ```
   
   ### 3. Runtime Boot & Execution in Simulator
   ```bash
   $ ./nuttx
   NuttShell (NSH) NuttX-12.8.0
   nsh> help
   help usage:  help [-v] [<cmd>]
   
       .            break        exec         mkdir        rm           true    
     
       [            cat          exit         mkfifo       rmdir        
truncate     
       ?            cd           expr         mkrd         rmmod        uname   
     
       addroute     cp           false        mount        set          umount  
     
       alias        cmp          help         mv           kill         unset   
     
       unalias      dirname      hexdump      nslookup     sleep        uptime  
     
       arp          date         insmod       poweroff     usleep       watch   
     
       base64dec    delroute     losetup      quit         source       xd      
     
       base64enc    dmesg        ls           printf       test         
       basename     echo         du           pwd          time         
   
   Builtin Apps:
       dd      init    nsh     sh      
   nsh> exit
   ```


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to