RE: [PATCH net v2] net: qualcomm: rmnet: restore skb->dev on deaggregated frames

> -----Original Message-----
> From: Xiang Mei <[email protected]>
> Sent: Thursday, August 20, 2026 1:53 PM
> To: Subash Abhinov Kasiviswanathan
> <[email protected]>; Sean Tranchetti
> <[email protected]>; Andrew Lunn
> <[email protected]>; David S . Miller <[email protected]>; Eric
> Dumazet <[email protected]>; Jakub Kicinski <[email protected]>; Paolo
> Abeni <[email protected]>
> Cc: [email protected]; [email protected];
> [email protected]; Xiang Mei <[email protected]>
> Subject: [PATCH net v2] net: qualcomm: rmnet: restore skb->dev on
> deaggregated frames
> 
> rmnet_map_deaggregate() allocates each sub-frame with alloc_skb() and
> leaves skb->dev NULL. __rmnet_map_ingress_handler() assigns
> skb->dev = ep->egress_dev only on the data path, but a MAP command frame
> is dispatched to rmnet_map_command() before that, so
> rmnet_map_send_ack() runs netif_tx_lock(skb->dev) on a NULL device. An
> unprivileged user reaches this by unsharing a user+net namespace, creating
an
> rmnet link over a tap device with INGRESS_DEAGGREGATION and
> INGRESS_MAP_COMMANDS, and writing an aggregated frame carrying a
> flow-control command to the tap fd.
> 
> Restore the assignment dropped by 378e25357ac7, so every skb leaving
> rmnet_map_deaggregate() has a valid device.
> 
>   BUG: KASAN: null-ptr-deref in _raw_spin_lock
> (kernel/locking/spinlock.c:158)
>   Write of size 4 at addr 00000000000004b4 by task exploit/144
>   Call Trace:
>    _raw_spin_lock (kernel/locking/spinlock.c:158)
>    netif_tx_lock (net/sched/sch_generic.c:497)
>    rmnet_map_command
> (drivers/net/ethernet/qualcomm/rmnet/rmnet_map_command.c:67)
>    rmnet_rx_handler
> (drivers/net/ethernet/qualcomm/rmnet/rmnet_handlers.c:125)
>    __netif_receive_skb_core.constprop.0 (net/core/dev.c:6103)
>    ...
>    __netif_receive_skb_one_core (net/core/dev.c:6214)
>    netif_receive_skb (net/core/dev.c:6474)
>    tun_get_user (drivers/net/tun.c:1966)
>    tun_chr_write_iter (drivers/net/tun.c:2012)
>    vfs_write (fs/read_write.c:687)
>    ksys_write (fs/read_write.c:739)
>    do_syscall_64 (arch/x86/entry/syscall_64.c:94)
>    entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:121)
>   Kernel panic - not syncing: Fatal exception in interrupt
> 
> Fixes: 378e25357ac7 ("net: qualcomm: rmnet: Remove unnecessary device
> assignment")
> Reported-by: [email protected]
> Closes:
> https://lore.kernel.org/netdev/ijg79FFMfIvKJbivdJEKvTO90Q9dTvyBkJck@bu
> gs.sh/T/#u
> Assisted-by: Claude:claude-opus-4-8
> Signed-off-by: Xiang Mei <[email protected]>
> ---
> v2: assign in rmnet_map_deaggregate() instead of rmnet_map_command(),
>     thank @Eric Dumazet.
> 
>  drivers/net/ethernet/qualcomm/rmnet/rmnet_map_data.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/drivers/net/ethernet/qualcomm/rmnet/rmnet_map_data.c
> b/drivers/net/ethernet/qualcomm/rmnet/rmnet_map_data.c
> index 305ae15ae8f3..e6f48dc9fe64 100644
> --- a/drivers/net/ethernet/qualcomm/rmnet/rmnet_map_data.c
> +++ b/drivers/net/ethernet/qualcomm/rmnet/rmnet_map_data.c
> @@ -394,6 +394,7 @@ struct sk_buff *rmnet_map_deaggregate(struct
> sk_buff *skb,
>  	if (!skbn)
>  		return NULL;
> 
> +	skbn->dev = skb->dev;
>  	skb_reserve(skbn, RMNET_MAP_DEAGGR_HEADROOM);
>  	skb_put(skbn, packet_len);
>  	memcpy(skbn->data, skb->data, packet_len);
> --
> 2.43.0

Reviewed-by: Subash Abhinov Kasiviswanathan
<[email protected]>

A minor point here is that the skb->dev for data packets would be first set
to real_dev in rmnet_map_deaggregate() followed by the expected mux specific
rmnet dev in __rmnet_map_ingress_handler() with this patchset. The v1 patch
would be assigning the skb->dev for the command packets only which would
prevent the unnecessary reassignment of skb->dev in case of the data
packets.