[Bug 297380] netmap_ring_reinit() can run off the end of the slot[] array

https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=297380

            Bug ID: 297380
           Summary: netmap_ring_reinit() can run off the end of the slot[]
                    array
           Product: Base System
           Version: CURRENT
          Hardware: Any
                OS: Any
            Status: New
          Severity: Affects Some People
          Priority: ---
         Component: kern
          Assignee: [email protected]
          Reporter: [email protected]
 Attachment #273559 text/plain
         mime type:

Created attachment 273559
  --> https://bugs.freebsd.org/bugzilla/attachment.cgi?id=273559&action=edit
cause netmap_ring_reinit() to index off the end of slot[]

The attached program asks netmap for nr_mode=7 (NR_REG_NULL),
nr_rx_rings=1, and nr_rx_slots=0. The NR_REG_NULL causes the netmap to
be initialized with netmap_get_null_na(), which accepts rings=1 and
slots=0 as is. Then the program calls poll(), which results in
netmap_ring_reinit() being called, which says

        u_int i, lim = kring->nkr_num_slots - 1;
        ...;
        for (i = 0; i <= lim; i++) {

But num_slots is zero, so lim is 0xffffffff, so the for loop indexes
off the end of ring->slot[].

# cc netmap7b.c 
# uname -a
FreeBSD xxx 16.0-CURRENT FreeBSD 16.0-CURRENT #37 main-n275529-da69298e9825:
Fri Jul 17 07:08:15 AST 2026    
root@xxx:/usr/obj/usr/src/amd64.amd64/sys/GENERIC amd64
# ./a.out
...
panic: vm_fault_lookup: fault on nofault entry, addr: 0xfffffe0160765000
panic() at panic+0x43/frame 0xfffffe00d75456e0
vm_fault() at vm_fault+0x1db6/frame 0xfffffe00d7545870
vm_fault_trap() at vm_fault_trap+0x65/frame 0xfffffe00d75458b0
trap_pfault() at trap_pfault+0x295/frame 0xfffffe00d7545920
calltrap() at calltrap+0x8/frame 0xfffffe00d7545920
--- trap 0xc, rip = 0xffffffff807ef3ce, rsp = 0xfffffe00d75459f0, rbp =
0xfffffe00d7545a60 ---
netmap_ring_reinit() at netmap_ring_reinit+0x18e/frame 0xfffffe00d7545a60
netmap_poll() at netmap_poll+0x571/frame 0xfffffe00d7545b40
freebsd_netmap_poll() at freebsd_netmap_poll+0x3f/frame 0xfffffe00d7545b70
devfs_poll_f() at devfs_poll_f+0x8a/frame 0xfffffe00d7545bc0
kern_poll_kfds() at kern_poll_kfds+0x3f6/frame 0xfffffe00d7545c60
kern_poll() at kern_poll+0x111/frame 0xfffffe00d7545dc0
sys_poll() at sys_poll+0x5b/frame 0xfffffe00d7545df0
amd64_syscall() at amd64_syscall+0x17c/frame 0xfffffe00d7545f30
fast_syscall_common() at fast_syscall_common+0xf8/frame 0xfffffe00d7545f30
--- syscall (209, FreeBSD ELF64, poll), rip = 0x82366748a, rsp = 0x821023868,
rbp = 0x821023910 ---

-- 
You are receiving this mail because:
You are the assignee for the bug.