From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: linux-kernel@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
stable@vger.kernel.org, Oliver Zweigle <Oliver.Zweigle@faro.com>,
Bernd Eckstein <3ernd.Eckstein@gmail.com>,
"David S. Miller" <davem@davemloft.net>
Subject: [PATCH 4.9 18/50] usbnet: ipheth: fix potential recvmsg bug and recvmsg bug 2
Date: Tue, 4 Dec 2018 11:50:13 +0100 [thread overview]
Message-ID: <20181204103715.447520170@linuxfoundation.org> (raw)
In-Reply-To: <20181204103714.485546262@linuxfoundation.org>
4.9-stable review patch. If anyone has any objections, please let me know.
------------------
From: Bernd Eckstein <3erndeckstein@gmail.com>
[ Upstream commit 45611c61dd503454b2edae00aabe1e429ec49ebe ]
The bug is not easily reproducable, as it may occur very infrequently
(we had machines with 20minutes heavy downloading before it occurred)
However, on a virual machine (VMWare on Windows 10 host) it occurred
pretty frequently (1-2 seconds after a speedtest was started)
dev->tx_skb mab be freed via dev_kfree_skb_irq on a callback
before it is set.
This causes the following problems:
- double free of the skb or potential memory leak
- in dmesg: 'recvmsg bug' and 'recvmsg bug 2' and eventually
general protection fault
Example dmesg output:
[ 134.841986] ------------[ cut here ]------------
[ 134.841987] recvmsg bug: copied 9C24A555 seq 9C24B557 rcvnxt 9C25A6B3 fl 0
[ 134.841993] WARNING: CPU: 7 PID: 2629 at /build/linux-hwe-On9fm7/linux-hwe-4.15.0/net/ipv4/tcp.c:1865 tcp_recvmsg+0x44d/0xab0
[ 134.841994] Modules linked in: ipheth(OE) kvm_intel kvm irqbypass crct10dif_pclmul crc32_pclmul ghash_clmulni_intel pcbc aesni_intel aes_x86_64 crypto_simd glue_helper cryptd vmw_balloon intel_rapl_perf joydev input_leds serio_raw vmw_vsock_vmci_transport vsock shpchp i2c_piix4 mac_hid binfmt_misc vmw_vmci parport_pc ppdev lp parport autofs4 vmw_pvscsi vmxnet3 hid_generic usbhid hid vmwgfx ttm drm_kms_helper syscopyarea sysfillrect mptspi mptscsih sysimgblt ahci psmouse fb_sys_fops pata_acpi mptbase libahci e1000 drm scsi_transport_spi
[ 134.842046] CPU: 7 PID: 2629 Comm: python Tainted: G W OE 4.15.0-34-generic #37~16.04.1-Ubuntu
[ 134.842046] Hardware name: VMware, Inc. VMware Virtual Platform/440BX Desktop Reference Platform, BIOS 6.00 05/19/2017
[ 134.842048] RIP: 0010:tcp_recvmsg+0x44d/0xab0
[ 134.842048] RSP: 0018:ffffa6630422bcc8 EFLAGS: 00010286
[ 134.842049] RAX: 0000000000000000 RBX: ffff997616f4f200 RCX: 0000000000000006
[ 134.842049] RDX: 0000000000000007 RSI: 0000000000000082 RDI: ffff9976257d6490
[ 134.842050] RBP: ffffa6630422bd98 R08: 0000000000000001 R09: 000000000004bba4
[ 134.842050] R10: 0000000001e00c6f R11: 000000000004bba4 R12: ffff99760dee3000
[ 134.842051] R13: 0000000000000000 R14: ffff99760dee3514 R15: 0000000000000000
[ 134.842051] FS: 00007fe332347700(0000) GS:ffff9976257c0000(0000) knlGS:0000000000000000
[ 134.842052] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 134.842053] CR2: 0000000001e41000 CR3: 000000020e9b4006 CR4: 00000000003606e0
[ 134.842055] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[ 134.842055] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[ 134.842057] Call Trace:
[ 134.842060] ? aa_sk_perm+0x53/0x1a0
[ 134.842064] inet_recvmsg+0x51/0xc0
[ 134.842066] sock_recvmsg+0x43/0x50
[ 134.842070] SYSC_recvfrom+0xe4/0x160
[ 134.842072] ? __schedule+0x3de/0x8b0
[ 134.842075] ? ktime_get_ts64+0x4c/0xf0
[ 134.842079] SyS_recvfrom+0xe/0x10
[ 134.842082] do_syscall_64+0x73/0x130
[ 134.842086] entry_SYSCALL_64_after_hwframe+0x3d/0xa2
[ 134.842086] RIP: 0033:0x7fe331f5a81d
[ 134.842088] RSP: 002b:00007ffe8da98398 EFLAGS: 00000246 ORIG_RAX: 000000000000002d
[ 134.842090] RAX: ffffffffffffffda RBX: ffffffffffffffff RCX: 00007fe331f5a81d
[ 134.842094] RDX: 00000000000003fb RSI: 0000000001e00874 RDI: 0000000000000003
[ 134.842095] RBP: 00007fe32f642c70 R08: 0000000000000000 R09: 0000000000000000
[ 134.842097] R10: 0000000000000000 R11: 0000000000000246 R12: 00007fe332347698
[ 134.842099] R13: 0000000001b7e0a0 R14: 0000000001e00874 R15: 0000000000000000
[ 134.842103] Code: 24 fd ff ff e9 cc fe ff ff 48 89 d8 41 8b 8c 24 10 05 00 00 44 8b 45 80 48 c7 c7 08 bd 59 8b 48 89 85 68 ff ff ff e8 b3 c4 7d ff <0f> 0b 48 8b 85 68 ff ff ff e9 e9 fe ff ff 41 8b 8c 24 10 05 00
[ 134.842126] ---[ end trace b7138fc08c83147f ]---
[ 134.842144] general protection fault: 0000 [#1] SMP PTI
[ 134.842145] Modules linked in: ipheth(OE) kvm_intel kvm irqbypass crct10dif_pclmul crc32_pclmul ghash_clmulni_intel pcbc aesni_intel aes_x86_64 crypto_simd glue_helper cryptd vmw_balloon intel_rapl_perf joydev input_leds serio_raw vmw_vsock_vmci_transport vsock shpchp i2c_piix4 mac_hid binfmt_misc vmw_vmci parport_pc ppdev lp parport autofs4 vmw_pvscsi vmxnet3 hid_generic usbhid hid vmwgfx ttm drm_kms_helper syscopyarea sysfillrect mptspi mptscsih sysimgblt ahci psmouse fb_sys_fops pata_acpi mptbase libahci e1000 drm scsi_transport_spi
[ 134.842161] CPU: 7 PID: 2629 Comm: python Tainted: G W OE 4.15.0-34-generic #37~16.04.1-Ubuntu
[ 134.842162] Hardware name: VMware, Inc. VMware Virtual Platform/440BX Desktop Reference Platform, BIOS 6.00 05/19/2017
[ 134.842164] RIP: 0010:tcp_close+0x2c6/0x440
[ 134.842165] RSP: 0018:ffffa6630422bde8 EFLAGS: 00010202
[ 134.842167] RAX: 0000000000000000 RBX: ffff99760dee3000 RCX: 0000000180400034
[ 134.842168] RDX: 5c4afd407207a6c4 RSI: ffffe868495bd300 RDI: ffff997616f4f200
[ 134.842169] RBP: ffffa6630422be08 R08: 0000000016f4d401 R09: 0000000180400034
[ 134.842169] R10: ffffa6630422bd98 R11: 0000000000000000 R12: 000000000000600c
[ 134.842170] R13: 0000000000000000 R14: ffff99760dee30c8 R15: ffff9975bd44fe00
[ 134.842171] FS: 00007fe332347700(0000) GS:ffff9976257c0000(0000) knlGS:0000000000000000
[ 134.842173] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 134.842174] CR2: 0000000001e41000 CR3: 000000020e9b4006 CR4: 00000000003606e0
[ 134.842177] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[ 134.842178] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[ 134.842179] Call Trace:
[ 134.842181] inet_release+0x42/0x70
[ 134.842183] __sock_release+0x42/0xb0
[ 134.842184] sock_close+0x15/0x20
[ 134.842187] __fput+0xea/0x220
[ 134.842189] ____fput+0xe/0x10
[ 134.842191] task_work_run+0x8a/0xb0
[ 134.842193] exit_to_usermode_loop+0xc4/0xd0
[ 134.842195] do_syscall_64+0xf4/0x130
[ 134.842197] entry_SYSCALL_64_after_hwframe+0x3d/0xa2
[ 134.842197] RIP: 0033:0x7fe331f5a560
[ 134.842198] RSP: 002b:00007ffe8da982e8 EFLAGS: 00000246 ORIG_RAX: 0000000000000003
[ 134.842200] RAX: 0000000000000000 RBX: 00007fe32f642c70 RCX: 00007fe331f5a560
[ 134.842201] RDX: 00000000008f5320 RSI: 0000000001cd4b50 RDI: 0000000000000003
[ 134.842202] RBP: 00007fe32f6500f8 R08: 000000000000003c R09: 00000000009343c0
[ 134.842203] R10: 0000000000000000 R11: 0000000000000246 R12: 00007fe32f6500d0
[ 134.842204] R13: 00000000008f5320 R14: 00000000008f5320 R15: 0000000001cd4770
[ 134.842205] Code: c8 00 00 00 45 31 e4 49 39 fe 75 4d eb 50 83 ab d8 00 00 00 01 48 8b 17 48 8b 47 08 48 c7 07 00 00 00 00 48 c7 47 08 00 00 00 00 <48> 89 42 08 48 89 10 0f b6 57 34 8b 47 2c 2b 47 28 83 e2 01 80
[ 134.842226] RIP: tcp_close+0x2c6/0x440 RSP: ffffa6630422bde8
[ 134.842227] ---[ end trace b7138fc08c831480 ]---
The proposed patch eliminates a potential racing condition.
Before, usb_submit_urb was called and _after_ that, the skb was attached
(dev->tx_skb). So, on a callback it was possible, however unlikely that the
skb was freed before it was set. That way (because dev->tx_skb was not set
to NULL after it was freed), it could happen that a skb from a earlier
transmission was freed a second time (and the skb we should have freed did
not get freed at all)
Now we free the skb directly in ipheth_tx(). It is not passed to the
callback anymore, eliminating the posibility of a double free of the same
skb. Depending on the retval of usb_submit_urb() we use dev_kfree_skb_any()
respectively dev_consume_skb_any() to free the skb.
Signed-off-by: Oliver Zweigle <Oliver.Zweigle@faro.com>
Signed-off-by: Bernd Eckstein <3ernd.Eckstein@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
drivers/net/usb/ipheth.c | 10 ++++------
1 file changed, 4 insertions(+), 6 deletions(-)
--- a/drivers/net/usb/ipheth.c
+++ b/drivers/net/usb/ipheth.c
@@ -140,7 +140,6 @@ struct ipheth_device {
struct usb_device *udev;
struct usb_interface *intf;
struct net_device *net;
- struct sk_buff *tx_skb;
struct urb *tx_urb;
struct urb *rx_urb;
unsigned char *tx_buf;
@@ -229,6 +228,7 @@ static void ipheth_rcvbulk_callback(stru
case -ENOENT:
case -ECONNRESET:
case -ESHUTDOWN:
+ case -EPROTO:
return;
case 0:
break;
@@ -280,7 +280,6 @@ static void ipheth_sndbulk_callback(stru
dev_err(&dev->intf->dev, "%s: urb status: %d\n",
__func__, status);
- dev_kfree_skb_irq(dev->tx_skb);
netif_wake_queue(dev->net);
}
@@ -410,7 +409,7 @@ static int ipheth_tx(struct sk_buff *skb
if (skb->len > IPHETH_BUF_SIZE) {
WARN(1, "%s: skb too large: %d bytes\n", __func__, skb->len);
dev->net->stats.tx_dropped++;
- dev_kfree_skb_irq(skb);
+ dev_kfree_skb_any(skb);
return NETDEV_TX_OK;
}
@@ -430,12 +429,11 @@ static int ipheth_tx(struct sk_buff *skb
dev_err(&dev->intf->dev, "%s: usb_submit_urb: %d\n",
__func__, retval);
dev->net->stats.tx_errors++;
- dev_kfree_skb_irq(skb);
+ dev_kfree_skb_any(skb);
} else {
- dev->tx_skb = skb;
-
dev->net->stats.tx_packets++;
dev->net->stats.tx_bytes += skb->len;
+ dev_consume_skb_any(skb);
netif_stop_queue(net);
}
next prev parent reply other threads:[~2018-12-04 11:08 UTC|newest]
Thread overview: 58+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-12-04 10:49 [PATCH 4.9 00/50] 4.9.143-stable review Greg Kroah-Hartman
2018-12-04 10:49 ` [PATCH 4.9 01/50] mm/huge_memory: rename freeze_page() to unmap_page() Greg Kroah-Hartman
2018-12-04 10:49 ` [PATCH 4.9 02/50] mm/huge_memory.c: reorder operations in __split_huge_page_tail() Greg Kroah-Hartman
2018-12-04 10:49 ` [PATCH 4.9 03/50] mm/huge_memory: splitting set mapping+index before unfreeze Greg Kroah-Hartman
2018-12-04 10:49 ` [PATCH 4.9 04/50] mm/huge_memory: fix lockdep complaint on 32-bit i_size_read() Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 05/50] mm/khugepaged: collapse_shmem() stop if punched or truncated Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 06/50] shmem: shmem_charge: verify max_block is not exceeded before inode update Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 07/50] shmem: introduce shmem_inode_acct_block Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 08/50] mm/khugepaged: fix crashes due to misaccounted holes Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 09/50] mm/khugepaged: collapse_shmem() remember to clear holes Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 10/50] mm/khugepaged: minor reorderings in collapse_shmem() Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 11/50] mm/khugepaged: collapse_shmem() without freezing new_page Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 12/50] mm/khugepaged: collapse_shmem() do not crash on Compound Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 13/50] media: em28xx: Fix use-after-free when disconnecting Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 14/50] Revert "wlcore: Add missing PM call for wlcore_cmd_wait_for_event_or_timeout()" Greg Kroah-Hartman
2018-12-04 10:50 ` Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 15/50] net: skb_scrub_packet(): Scrub offload_fwd_mark Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 16/50] rapidio/rionet: do not free skb before reading its length Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 17/50] s390/qeth: fix length check in SNMP processing Greg Kroah-Hartman
2018-12-04 10:50 ` Greg Kroah-Hartman [this message]
2018-12-04 10:50 ` [PATCH 4.9 19/50] kvm: mmu: Fix race in emulated page table writes Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 20/50] kvm: svm: Ensure an IBPB on all affected CPUs when freeing a vmcb Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 21/50] KVM: X86: Fix scan ioapic use-before-initialization Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 22/50] xtensa: enable coprocessors that are being flushed Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 23/50] xtensa: fix coprocessor context offset definitions Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 24/50] Btrfs: ensure path name is null terminated at btrfs_control_ioctl Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 25/50] perf/x86/intel: Move branch tracing setup to the Intel-specific source file Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 26/50] perf/x86/intel: Add generic branch tracing check to intel_pmu_has_bts() Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 27/50] fs: fix lost error code in dio_complete Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 28/50] ALSA: wss: Fix invalid snd_free_pages() at error path Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 29/50] ALSA: ac97: Fix incorrect bit shift at AC97-SPSA control write Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 30/50] ALSA: control: Fix race between adding and removing a user element Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 31/50] ALSA: sparc: Fix invalid snd_free_pages() at error path Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 32/50] ext2: fix potential use after free Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 33/50] dmaengine: at_hdmac: fix memory leak in at_dma_xlate() Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 34/50] dmaengine: at_hdmac: fix module unloading Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 35/50] btrfs: release metadata before running delayed refs Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 36/50] USB: usb-storage: Add new IDs to ums-realtek Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 37/50] usb: core: quirks: add RESET_RESUME quirk for Cherry G230 Stream series Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 38/50] Revert "usb: dwc3: gadget: skip Set/Clear Halt when invalid" Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 39/50] iio:st_magn: Fix enable device after trigger Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 40/50] mm: use swp_offset as key in shmem_replace_page() Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 41/50] Drivers: hv: vmbus: check the creation_status in vmbus_establish_gpadl() Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 42/50] misc: mic/scif: fix copy-paste error in scif_create_remote_lookup Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 43/50] efi/libstub: arm: support building with clang Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 44/50] ARM: 8766/1: drop no-thumb-interwork in EABI mode Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 45/50] ARM: 8767/1: add support for building ARM kernel with clang Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 46/50] bus: arm-cci: remove unnecessary unreachable() Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 47/50] ARM: trusted_foundations: do not use naked function Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 48/50] workqueue: avoid clang warning Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 49/50] efi/libstub: Make file I/O chunking x86-specific Greg Kroah-Hartman
2018-12-04 10:50 ` [PATCH 4.9 50/50] kbuild: Set KBUILD_CFLAGS before incl. arch Makefile Greg Kroah-Hartman
2018-12-04 15:52 ` [PATCH 4.9 00/50] 4.9.143-stable review kernelci.org bot
2018-12-04 21:40 ` Guenter Roeck
2018-12-05 5:12 ` Naresh Kamboju
2018-12-05 9:30 ` Jon Hunter
2018-12-05 9:30 ` Jon Hunter
2018-12-05 23:54 ` shuah
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20181204103715.447520170@linuxfoundation.org \
--to=gregkh@linuxfoundation.org \
--cc=3ernd.Eckstein@gmail.com \
--cc=Oliver.Zweigle@faro.com \
--cc=davem@davemloft.net \
--cc=linux-kernel@vger.kernel.org \
--cc=stable@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.