| Message ID | 1474730674-4017-1-git-send-email-pandit.parav@gmail.com (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
On Sat, Sep 24, 2016 at 03:24:34PM +0000, Parav Pandit wrote: > This patch fixes below kernel crash when CONFIG_IPV6 is disabled by > avoiding ip6 socket initialization. > > Additionally it has below minor related cleanup around recv_sockets elements. > > 1. Avoid extern for structure which is locally used in file > 2. Define recv_sockets as static as its used by single file > 3. Follow reverse deallocation order than allocation order for > recv_sockets elements Thanks Parav, This crash was already fixed upstream. https://git.kernel.org/cgit/linux/kernel/git/dledford/rdma.git/commit/drivers/infiniband/sw/rxe?id=dfdd6158ca2c8838313470c5ab79243d04da70b2
On Sun, Sep 25, 2016 at 10:06 AM, Leon Romanovsky <leon@kernel.org> wrote: > On Sat, Sep 24, 2016 at 03:24:34PM +0000, Parav Pandit wrote: >> This patch fixes below kernel crash when CONFIG_IPV6 is disabled by >> avoiding ip6 socket initialization. >> >> Additionally it has below minor related cleanup around recv_sockets elements. >> >> 1. Avoid extern for structure which is locally used in file >> 2. Define recv_sockets as static as its used by single file >> 3. Follow reverse deallocation order than allocation order for >> recv_sockets elements > > Thanks Parav, > This crash was already fixed upstream. > https://git.kernel.org/cgit/linux/kernel/git/dledford/rdma.git/commit/drivers/infiniband/sw/rxe?id=dfdd6158ca2c8838313470c5ab79243d04da70b2 ok. Thanks. I was in 4.8-rc branch and I searched for rdma_rxe tag for commits based on Stephen's rdma_rxe tag. This one was done using IB/rxe. -- To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/drivers/infiniband/sw/rxe/rxe_net.c b/drivers/infiniband/sw/rxe/rxe_net.c index 0b8d2ea..dc9851f 100644 --- a/drivers/infiniband/sw/rxe/rxe_net.c +++ b/drivers/infiniband/sw/rxe/rxe_net.c @@ -48,6 +48,8 @@ static LIST_HEAD(rxe_dev_list); static spinlock_t dev_list_lock; /* spinlock for device list */ +static struct rxe_recv_sockets recv_sockets; + struct rxe_dev *net_to_rxe(struct net_device *ndev) { struct rxe_dev *rxe; @@ -81,9 +83,6 @@ struct rxe_dev *get_rxe_by_name(const char* name) return found; } - -struct rxe_recv_sockets recv_sockets; - static __be64 rxe_mac_to_eui64(struct net_device *ndev) { unsigned char *mac_addr = ndev->dev_addr; @@ -277,7 +276,8 @@ static struct socket *rxe_setup_udp_tunnel(struct net *net, __be16 port, static void rxe_release_udp_tunnel(struct socket *sk) { - udp_tunnel_sock_release(sk); + if (sk) + udp_tunnel_sock_release(sk); } static void prepare_udp_hdr(struct sk_buff *skb, __be16 src_port, @@ -668,6 +668,7 @@ int rxe_net_init(void) spin_lock_init(&dev_list_lock); +#if IS_ENABLED(CONFIG_IPV6) recv_sockets.sk6 = rxe_setup_udp_tunnel(&init_net, htons(ROCE_V2_UDP_DPORT), true); if (IS_ERR(recv_sockets.sk6)) { @@ -675,7 +676,7 @@ int rxe_net_init(void) pr_err("rxe: Failed to create IPv6 UDP tunnel\n"); return -1; } - +#endif recv_sockets.sk4 = rxe_setup_udp_tunnel(&init_net, htons(ROCE_V2_UDP_DPORT), false); if (IS_ERR(recv_sockets.sk4)) { @@ -688,8 +689,8 @@ int rxe_net_init(void) err = register_netdevice_notifier(&rxe_net_notifier); if (err) { - rxe_release_udp_tunnel(recv_sockets.sk6); rxe_release_udp_tunnel(recv_sockets.sk4); + rxe_release_udp_tunnel(recv_sockets.sk6); pr_err("rxe: Failed to rigister netdev notifier\n"); } @@ -698,11 +699,8 @@ int rxe_net_init(void) void rxe_net_exit(void) { - if (recv_sockets.sk6) - rxe_release_udp_tunnel(recv_sockets.sk6); - - if (recv_sockets.sk4) - rxe_release_udp_tunnel(recv_sockets.sk4); + rxe_release_udp_tunnel(recv_sockets.sk4); + rxe_release_udp_tunnel(recv_sockets.sk6); unregister_netdevice_notifier(&rxe_net_notifier); } diff --git a/drivers/infiniband/sw/rxe/rxe_net.h b/drivers/infiniband/sw/rxe/rxe_net.h index 7b06f76..106c586 100644 --- a/drivers/infiniband/sw/rxe/rxe_net.h +++ b/drivers/infiniband/sw/rxe/rxe_net.h @@ -43,8 +43,6 @@ struct rxe_recv_sockets { struct socket *sk6; }; -extern struct rxe_recv_sockets recv_sockets; - struct rxe_dev *rxe_net_add(struct net_device *ndev); int rxe_net_init(void);
This patch fixes below kernel crash when CONFIG_IPV6 is disabled by avoiding ip6 socket initialization. Additionally it has below minor related cleanup around recv_sockets elements. 1. Avoid extern for structure which is locally used in file 2. Define recv_sockets as static as its used by single file 3. Follow reverse deallocation order than allocation order for recv_sockets elements kernel: [ 348.527616] general protection fault: 0000 [#1] SMP kernel: [ 348.528052] Modules linked in: rdma_rxe(+) udp_tunnel ib_core binfmt_misc dm_mirror dm_region_hash dm_log dm_mod mousedev psmouse sg pcspkr evdev floppy button acpi_cpufreq ext4(E) crc16(E) jbd2(E) mbcache(E) sd_mod(E) ata_piix(E) libata(E) scsi_mod(E) kernel: [ 348.528052] CPU: 0 PID: 2561 Comm: modprobe Tainted: G E 4.8.0-rc6+ #8 kernel: [ 348.528052] task: ffff880037af1d80 task.stack: ffff88003df50000 kernel: [ 348.528052] RIP: 0010:[<ffffffffa03212bd>] [<ffffffffa03212bd>] setup_udp_tunnel_sock+0xd/0x70 [udp_tunnel] kernel: [ 348.528052] RSP: 0018:ffff88003df53c38 EFLAGS: 00010202 kernel: [ 348.528052] RAX: 840fc28949c08548 RBX: 0000000000000000 RCX: 00000000685eba97 kernel: [ 348.528052] RDX: ffff88003df53c68 RSI: ffffffff81197185 RDI: ffffffff81acc800 kernel: [ 348.528052] RBP: ffff88003df53c38 R08: 000000000001b1c0 R09: ffffffff812ef300 kernel: [ 348.528052] R10: ffff88003e21b1c0 R11: ffffea0000f36000 R12: ffffffffa033e000 kernel: [ 348.528052] R13: ffffffffa0339c10 R14: ffff88003d17b240 R15: ffffffffa0339bc0 kernel: [ 348.528052] FS: 00007f910f913740(0000) GS:ffff88003e200000(0000) knlGS:0000000000000000 kernel: [ 348.528052] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 kernel: [ 348.528052] CR2: 00007f910eac0000 CR3: 0000000037a07000 CR4: 00000000001406f0 kernel: [ 348.528052] Stack: kernel: [ 348.528052] ffff88003df53ca0 ffffffffa0331fbb ffffffff81197185 0000000a00000000 kernel: [ 348.528052] 0000000000000000 000000080000b712 0000000000000000 ffffffffa0335701 kernel: [ 348.528052] ffffffffa0332320 0000000000000000 ffff88003d17b240 ffffffffa0339bc0 kernel: [ 348.528052] Call Trace: kernel: [ 348.528052] [<ffffffffa0331fbb>] rxe_setup_udp_tunnel.constprop.14+0x6b/0xc0 [rdma_rxe] kernel: [ 348.528052] [<ffffffff81197185>] ? kmem_cache_create+0x85/0x1c0 kernel: [ 348.528052] [<ffffffffa0332320>] ? net_to_rxe+0x80/0x80 [rdma_rxe] kernel: [ 348.528052] [<ffffffffa03326de>] rxe_net_init+0x1e/0xe0 [rdma_rxe] kernel: [ 348.528052] [<ffffffffa033e027>] rxe_module_init+0x27/0x57 [rdma_rxe] kernel: [ 348.528052] [<ffffffff8100214d>] do_one_initcall+0x3d/0x160 kernel: [ 348.528052] [<ffffffff811b4145>] ? __vunmap+0x85/0xd0 kernel: [ 348.528052] [<ffffffff811b41fe>] ? vfree+0x2e/0x70 kernel: [ 348.528052] [<ffffffff8116ce63>] do_init_module+0x60/0x1e0 kernel: [ 348.528052] [<ffffffff810ec622>] load_module+0x1dc2/0x2650 kernel: [ 348.528052] [<ffffffff810e9970>] ? __symbol_put+0x40/0x40 kernel: [ 348.528052] [<ffffffff810ecfae>] SYSC_init_module+0xfe/0x110 kernel: [ 348.528052] [<ffffffff810ed07e>] SyS_init_module+0xe/0x10 kernel: [ 348.528052] [<ffffffff81502df2>] entry_SYSCALL_64_fastpath+0x1a/0xa4 kernel: [ 348.528052] Code: ca 0e 10 e1 48 89 df e8 32 22 10 e1 5b 5d c3 66 66 66 66 66 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 48 8b 46 20 55 48 89 e5 <80> a0 d0 02 00 00 ef 80 80 d3 02 00 00 01 48 8b 0a 48 89 88 28 kernel: [ 348.528052] RIP [<ffffffffa03212bd>] setup_udp_tunnel_sock+0xd/0x70 [udp_tunnel] kernel: [ 348.528052] RSP <ffff88003df53c38> kernel: [ 348.772092] ---[ end trace 432dfe51deb7b58e ]--- Signed-off-by: Parav Pandit <pandit.parav@gmail.com> --- drivers/infiniband/sw/rxe/rxe_net.c | 20 +++++++++----------- drivers/infiniband/sw/rxe/rxe_net.h | 2 -- 2 files changed, 9 insertions(+), 13 deletions(-)