Message ID | 20230724221326.384-1-andrew.kanner@gmail.com (mailing list archive) |
---|---|
State | Changes Requested |
Delegated to: | Netdev Maintainers |
Headers | show |
Series | [1/2] drivers: net: use xdp only inside tun_build_skb() | expand |
On Tue, Jul 25, 2023 at 6:14 AM Andrew Kanner <andrew.kanner@gmail.com> wrote: > > Syzkaller reported the following issue: > ======================================= > Too BIG xdp->frame_sz = 131072 > WARNING: CPU: 0 PID: 5020 at net/core/filter.c:4121 > ____bpf_xdp_adjust_tail net/core/filter.c:4121 [inline] > WARNING: CPU: 0 PID: 5020 at net/core/filter.c:4121 > bpf_xdp_adjust_tail+0x466/0xa10 net/core/filter.c:4103 > ... > Call Trace: > <TASK> > bpf_prog_4add87e5301a4105+0x1a/0x1c > __bpf_prog_run include/linux/filter.h:600 [inline] > bpf_prog_run_xdp include/linux/filter.h:775 [inline] > bpf_prog_run_generic_xdp+0x57e/0x11e0 net/core/dev.c:4721 > netif_receive_generic_xdp net/core/dev.c:4807 [inline] > do_xdp_generic+0x35c/0x770 net/core/dev.c:4866 > tun_get_user+0x2340/0x3ca0 drivers/net/tun.c:1919 > tun_chr_write_iter+0xe8/0x210 drivers/net/tun.c:2043 > call_write_iter include/linux/fs.h:1871 [inline] > new_sync_write fs/read_write.c:491 [inline] > vfs_write+0x650/0xe40 fs/read_write.c:584 > ksys_write+0x12f/0x250 fs/read_write.c:637 > do_syscall_x64 arch/x86/entry/common.c:50 [inline] > do_syscall_64+0x38/0xb0 arch/x86/entry/common.c:80 > entry_SYSCALL_64_after_hwframe+0x63/0xcd > > frame_sz < PAGE_SIZE check was introduced in commit c8741e2bfe87 > ("xdp: Allow bpf_xdp_adjust_tail() to grow packet size"). I wonder if this check makes sense for xdp generic when skb is linearized. > But > tun_get_user() still provides an execution path to exceed XDP limits > for packet size. I'd suggest being specific to the execution path, which is actually the generic XDP. > The way to mitigate this is to reduce xdp to be used > only in tun_build_skb(), without falling in tun_alloc_skb(), etc. > > Reported-and-tested-by: syzbot+f817490f5bd20541b90a@syzkaller.appspotmail.com > Closes: https://lore.kernel.org/all/000000000000774b9205f1d8a80d@google.com/T/ > Link: https://syzkaller.appspot.com/bug?id=5335c7c62bfff89bbb1c8f14cdabebe91909060f > Signed-off-by: Andrew Kanner <andrew.kanner@gmail.com> > --- > drivers/net/tun.c | 14 ++++++++------ > 1 file changed, 8 insertions(+), 6 deletions(-) > > diff --git a/drivers/net/tun.c b/drivers/net/tun.c > index d75456adc62a..18ccbbe9830a 100644 > --- a/drivers/net/tun.c > +++ b/drivers/net/tun.c > @@ -1580,8 +1580,14 @@ static void tun_rx_batched(struct tun_struct *tun, struct tun_file *tfile, > } > > static bool tun_can_build_skb(struct tun_struct *tun, struct tun_file *tfile, > - int len, int noblock, bool zerocopy) > + int len, int noblock, bool zerocopy, int *skb_xdp) > { > + if (SKB_DATA_ALIGN(len + TUN_RX_PAD) + > + SKB_DATA_ALIGN(sizeof(struct skb_shared_info)) > PAGE_SIZE) { > + *skb_xdp = 0; So if XDP is enabled, the packet will bypass the bpf program which is sub-optimal. I think we should at least drop the packet in this case. Thanks > + return false; > + } > + > if ((tun->flags & TUN_TYPE_MASK) != IFF_TAP) > return false; > > @@ -1594,10 +1600,6 @@ static bool tun_can_build_skb(struct tun_struct *tun, struct tun_file *tfile, > if (zerocopy) > return false; > > - if (SKB_DATA_ALIGN(len + TUN_RX_PAD) + > - SKB_DATA_ALIGN(sizeof(struct skb_shared_info)) > PAGE_SIZE) > - return false; > - > return true; > } > > @@ -1809,7 +1811,7 @@ static ssize_t tun_get_user(struct tun_struct *tun, struct tun_file *tfile, > zerocopy = true; > } > > - if (!frags && tun_can_build_skb(tun, tfile, len, noblock, zerocopy)) { > + if (tun_can_build_skb(tun, tfile, len, noblock, zerocopy, &skb_xdp) && !frags) { > /* For the packet that is not easy to be processed > * (e.g gso or jumbo packet), we will do it at after > * skb was created with generic XDP routine. > -- > 2.39.3 >
diff --git a/drivers/net/tun.c b/drivers/net/tun.c index d75456adc62a..18ccbbe9830a 100644 --- a/drivers/net/tun.c +++ b/drivers/net/tun.c @@ -1580,8 +1580,14 @@ static void tun_rx_batched(struct tun_struct *tun, struct tun_file *tfile, } static bool tun_can_build_skb(struct tun_struct *tun, struct tun_file *tfile, - int len, int noblock, bool zerocopy) + int len, int noblock, bool zerocopy, int *skb_xdp) { + if (SKB_DATA_ALIGN(len + TUN_RX_PAD) + + SKB_DATA_ALIGN(sizeof(struct skb_shared_info)) > PAGE_SIZE) { + *skb_xdp = 0; + return false; + } + if ((tun->flags & TUN_TYPE_MASK) != IFF_TAP) return false; @@ -1594,10 +1600,6 @@ static bool tun_can_build_skb(struct tun_struct *tun, struct tun_file *tfile, if (zerocopy) return false; - if (SKB_DATA_ALIGN(len + TUN_RX_PAD) + - SKB_DATA_ALIGN(sizeof(struct skb_shared_info)) > PAGE_SIZE) - return false; - return true; } @@ -1809,7 +1811,7 @@ static ssize_t tun_get_user(struct tun_struct *tun, struct tun_file *tfile, zerocopy = true; } - if (!frags && tun_can_build_skb(tun, tfile, len, noblock, zerocopy)) { + if (tun_can_build_skb(tun, tfile, len, noblock, zerocopy, &skb_xdp) && !frags) { /* For the packet that is not easy to be processed * (e.g gso or jumbo packet), we will do it at after * skb was created with generic XDP routine.
Syzkaller reported the following issue: ======================================= Too BIG xdp->frame_sz = 131072 WARNING: CPU: 0 PID: 5020 at net/core/filter.c:4121 ____bpf_xdp_adjust_tail net/core/filter.c:4121 [inline] WARNING: CPU: 0 PID: 5020 at net/core/filter.c:4121 bpf_xdp_adjust_tail+0x466/0xa10 net/core/filter.c:4103 ... Call Trace: <TASK> bpf_prog_4add87e5301a4105+0x1a/0x1c __bpf_prog_run include/linux/filter.h:600 [inline] bpf_prog_run_xdp include/linux/filter.h:775 [inline] bpf_prog_run_generic_xdp+0x57e/0x11e0 net/core/dev.c:4721 netif_receive_generic_xdp net/core/dev.c:4807 [inline] do_xdp_generic+0x35c/0x770 net/core/dev.c:4866 tun_get_user+0x2340/0x3ca0 drivers/net/tun.c:1919 tun_chr_write_iter+0xe8/0x210 drivers/net/tun.c:2043 call_write_iter include/linux/fs.h:1871 [inline] new_sync_write fs/read_write.c:491 [inline] vfs_write+0x650/0xe40 fs/read_write.c:584 ksys_write+0x12f/0x250 fs/read_write.c:637 do_syscall_x64 arch/x86/entry/common.c:50 [inline] do_syscall_64+0x38/0xb0 arch/x86/entry/common.c:80 entry_SYSCALL_64_after_hwframe+0x63/0xcd frame_sz < PAGE_SIZE check was introduced in commit c8741e2bfe87 ("xdp: Allow bpf_xdp_adjust_tail() to grow packet size"). But tun_get_user() still provides an execution path to exceed XDP limits for packet size. The way to mitigate this is to reduce xdp to be used only in tun_build_skb(), without falling in tun_alloc_skb(), etc. Reported-and-tested-by: syzbot+f817490f5bd20541b90a@syzkaller.appspotmail.com Closes: https://lore.kernel.org/all/000000000000774b9205f1d8a80d@google.com/T/ Link: https://syzkaller.appspot.com/bug?id=5335c7c62bfff89bbb1c8f14cdabebe91909060f Signed-off-by: Andrew Kanner <andrew.kanner@gmail.com> --- drivers/net/tun.c | 14 ++++++++------ 1 file changed, 8 insertions(+), 6 deletions(-)