3.18.113-stable review

This is the start of the stable review cycle for the 3.18.113 release. There are 21 patches in this series, all will be posted as a response to this one. If anyone has any issues with these being applied, please let me know.

Responses should be made by Thu Jun 14 16:48:15 UTC 2018. Anything received after that time might be too late.

The whole patch series can be found in one patch at:

	https://www.kernel.org/pub/linux/kernel/v3.x/stable-review/patch-3.18.113-rc1.gz

or in the git tree and branch at:
git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable-rc.git linux-3.18.y
and the diffstat can be found below.

thanks,

greg k-h

-------------
Pseudo-Shortlog of commits:

Greg Kroah-Hartman <[email protected]>

    Linux 3.18.113-rc1

Eric Dumazet <[email protected]>
rtnetlink: validate attributes in do_setlink()

Dan Carpenter <[email protected]>
team: use netdev_features_t instead of u32

Jack Morgenstein <[email protected]>
net/mlx4: Fix irq-unsafe spinlock usage

Daniele Palmas <[email protected]>
net: usb: cdc_mbim: add flag FLAG_SEND_ZLP

Eric Dumazet <[email protected]>
net/packet: refine check for priv area size

Wenwen Wang <[email protected]>
isdn: eicon: fix a missing-check bug

Sabrina Dubroca <[email protected]>
ip6mr: only set ip6mr_table from setsockopt when ip6mr_new_table succeeds

Govindarajulu Varadarajan <[email protected]>
enic: set DMA mask to 47 bit

Alexey Kodanev <[email protected]>
dccp: don't free ccid2_hc_tx_sock struct in dccp_disconnect()

Julia Lawall <[email protected]>
bnx2x: use the right constant

Dave Airlie <[email protected]>
drm: set FMODE_UNSIGNED_OFFSET for drm files

Linus Torvalds <[email protected]>
mmap: relax file size limit for regular files

Linus Torvalds <[email protected]>
mmap: introduce sane default mmap limits

Hugh Dickins <[email protected]>
mm: fix the NULL mapping case in __isolate_lru_page()

Al Viro <[email protected]>
fix io_destroy()/aio_complete() race

Ondrej Zary <[email protected]>
drm/i915: Disable LVDS on Radiant P845

Maciej W. Rozycki <[email protected]>
MIPS: ptrace: Fix PTRACE_PEEKUSR requests for 64-bit FGRs

Eric Dumazet <[email protected]>
tcp: avoid integer overflows in tcp_rcv_space_adjust()

Eric Biggers <[email protected]>
cfg80211: further limit wiphy names to 64 bytes

Sachin Grover <[email protected]>
selinux: KASAN: slab-out-of-bounds in xattr_getsecurity

Steven Rostedt (VMware) <[email protected]>
tracing: Fix crash when freeing instances with event triggers

-------------

Diffstat:

 Makefile                                         |  4 +--
 arch/mips/kernel/ptrace.c                        |  2 +-
 arch/mips/kernel/ptrace32.c                      |  2 +-
 drivers/gpu/drm/drm_fops.c                       |  1 +
 drivers/gpu/drm/i915/intel_lvds.c                |  8 ++++++
 drivers/isdn/hardware/eicon/diva.c               | 22 ++++++++++------
 drivers/isdn/hardware/eicon/diva.h               |  5 ++--
 drivers/isdn/hardware/eicon/divasmain.c          | 18 +++++++------
 drivers/net/ethernet/broadcom/bnx2x/bnx2x_link.c |  2 +-
 drivers/net/ethernet/cisco/enic/enic_main.c      |  8 +++---
 drivers/net/ethernet/mellanox/mlx4/qp.c          |  4 +--
 drivers/net/team/team.c                          |  3 ++-
 drivers/net/usb/cdc_mbim.c                       |  2 +-
 fs/aio.c                                         |  3 +--
 include/linux/tcp.h                              |  2 +-
 include/uapi/linux/nl80211.h                     |  2 +-
 kernel/trace/trace_events_trigger.c              |  5 ++--
 mm/mmap.c                                        | 32 ++++++++++++++++++++++++
 mm/vmscan.c                                      |  2 +-
 net/core/rtnetlink.c                             |  8 +++---
 net/dccp/proto.c                                 |  2 --
 net/ipv4/tcp_input.c                             | 10 +++++---
 net/ipv6/ip6mr.c                                 |  3 ++-
 net/packet/af_packet.c                           |  2 +-
 security/selinux/ss/services.c                   |  2 +-
 25 files changed, 105 insertions(+), 49 deletions(-)

3.18-stable review patch. If anyone has any objections, please let me know.

------------------

From: Steven Rostedt (VMware) <[email protected]>

commit 86b389ff22bd6ad8fd3cb98e41cd271886c6d023 upstream.

If a instance has an event trigger enabled when it is freed, it could cause an access of free memory. Here's the case that crashes:

 # cd /sys/kernel/tracing
 # mkdir instances/foo
 # echo snapshot > instances/foo/events/initcall/initcall_start/trigger
 # rmdir instances/foo

Would produce:

 general protection fault: 0000 [#1] PREEMPT SMP PTI
 Modules linked in: tun bridge ...
 CPU: 5 PID: 6203 Comm: rmdir Tainted: G        W         4.17.0-rc4-test+ #933
 Hardware name: Hewlett-Packard HP Compaq Pro 6300 SFF/339A, BIOS K01 v03.03 07/14/2016
 RIP: 0010:clear_event_triggers+0x3b/0x70
 RSP: 0018:ffffc90003783de0 EFLAGS: 00010286
 RAX: 0000000000000000 RBX: 6b6b6b6b6b6b6b2b RCX: 0000000000000000
 RDX: 0000000000000000 RSI: 0000000000000000 RDI: ffff8800c7130ba0
 RBP: ffffc90003783e00 R08: ffff8801131993f8 R09: 0000000100230016
 R10: ffffc90003783d80 R11: 0000000000000000 R12: ffff8800c7130ba0
 R13: ffff8800c7130bd8 R14: ffff8800cc093768 R15: 00000000ffffff9c
 FS:  00007f6f4aa86700(0000) GS:ffff88011eb40000(0000) knlGS:0000000000000000
 CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
 CR2: 00007f6f4a5aed60 CR3: 00000000cd552001 CR4: 00000000001606e0
 Call Trace:
  event_trace_del_tracer+0x2a/0xc5
  instance_rmdir+0x15c/0x200
  tracefs_syscall_rmdir+0x52/0x90
  vfs_rmdir+0xdb/0x160
  do_rmdir+0x16d/0x1c0
  __x64_sys_rmdir+0x17/0x20
  do_syscall_64+0x55/0x1a0
  entry_SYSCALL_64_after_hwframe+0x49/0xbe

This was due to the call the clears out the triggers when an instance is being deleted not removing the trigger from the link list.

---
 kernel/trace/trace_events_trigger.c |    5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

@@ -469,9 +469,10 @@ clear_event_triggers(struct trace_array
 	struct ftrace_event_file *file;
 
 	list_for_each_entry(file, &tr->events, list) {
-		struct event_trigger_data *data;
-		list_for_each_entry_rcu(data, &file->triggers, list) {
+		struct event_trigger_data *data, *n;
+		list_for_each_entry_safe(data, n, &file->triggers, list) {
 			trace_event_trigger_enable_disable(file, 0);
+			list_del_rcu(&data->list);
 			if (data->ops->free)
 				data->ops->free(data->ops, data);
 		}

3.18-stable review patch. If anyone has any objections, please let me know.

------------------

From: Sachin Grover <[email protected]>

commit efe3de79e0b52ca281ef6691480c8c68c82a4657 upstream.

Call trace:
 [<ffffff9203a8d7a8>] dump_backtrace+0x0/0x428
 [<ffffff9203a8dbf8>] show_stack+0x28/0x38
 [<ffffff920409bfb8>] dump_stack+0xd4/0x124
 [<ffffff9203d187e8>] print_address_description+0x68/0x258
 [<ffffff9203d18c00>] kasan_report.part.2+0x228/0x2f0
 [<ffffff9203d1927c>] kasan_report+0x5c/0x70
 [<ffffff9203d1776c>] check_memory_region+0x12c/0x1c0
 [<ffffff9203d17cdc>] memcpy+0x34/0x68
 [<ffffff9203d75348>] xattr_getsecurity+0xe0/0x160
 [<ffffff9203d75490>] vfs_getxattr+0xc8/0x120
 [<ffffff9203d75d68>] getxattr+0x100/0x2c8
 [<ffffff9203d76fb4>] SyS_fgetxattr+0x64/0xa0
 [<ffffff9203a83f70>] el0_svc_naked+0x24/0x28

If user get root access and calls security.selinux setxattr() with an embedded NUL on a file and then if some process performs a getxattr() on that file with a length greater than the actual length of the string, it would result in a panic.

To fix this, add the actual length of the string to the security context instead of the length passed by the userspace process.

---
 security/selinux/ss/services.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -1271,7 +1271,7 @@ static int security_context_to_sid_core(
 				      scontext_len, &context, def_sid);
 	if (rc == -EINVAL && force) {
 		context.str = str;
-		context.len = scontext_len;
+		context.len = strlen(str) + 1;
 		str = NULL;
 	} else if (rc)
 		goto out_unlock;

3.18-stable review patch. If anyone has any objections, please let me know.

------------------

From: Eric Biggers <[email protected]>

commit 814596495dd2b9d4aab92d8f89cf19060d25d2ea upstream.

wiphy names were recently limited to 128 bytes by commit a7cfebcb7594 ("cfg80211: limit wiphy names to 128 bytes"). As it turns out though, this isn't sufficient because dev_vprintk_emit() needs the syslog header string "SUBSYSTEM=ieee80211\0DEVICE=+ieee80211:$devname" to fit into 128 bytes. This triggered the "device/subsystem name too long" WARN when the device name was >= 90 bytes. As before, this was reproduced by syzbot by sending an HWSIM_CMD_NEW_RADIO command to the MAC80211_HWSIM generic netlink family.

Fix it by further limiting wiphy names to 64 bytes.

---
 include/uapi/linux/nl80211.h |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -2026,7 +2026,7 @@ enum nl80211_attrs {
 #define NL80211_ATTR_KEYS NL80211_ATTR_KEYS
 #define NL80211_ATTR_FEATURE_FLAGS NL80211_ATTR_FEATURE_FLAGS
 
-#define NL80211_WIPHY_NAME_MAXLEN		128
+#define NL80211_WIPHY_NAME_MAXLEN		64
 
 #define NL80211_MAX_SUPP_RATES			32
 #define NL80211_MAX_SUPP_HT_RATES		77

3.18-stable review patch. If anyone has any objections, please let me know.

------------------

From: Eric Dumazet <[email protected]>

commit 607065bad9931e72207b0cac365d7d4abc06bd99 upstream.

When using large tcp_rmem[2] values (I did tests with 500 MB), I noticed overflows while computing rcvwin.

Lets fix this before the following patch.

---
 include/linux/tcp.h  |    2 +-
 net/ipv4/tcp_input.c |   10 ++++++----
 2 files changed, 7 insertions(+), 5 deletions(-)

@@ -292,7 +292,7 @@ struct tcp_sock {
 
 /* Receiver queue space */
 	struct {
-		int	space;
+		u32	space;
 		u32	seq;
 		u32	time;
 	} rcvq_space;

@@ -550,8 +550,8 @@ static inline void tcp_rcv_rtt_measure_t
 void tcp_rcv_space_adjust(struct sock *sk)
 {
 	struct tcp_sock *tp = tcp_sk(sk);
+	u32 copied;
 	int time;
-	int copied;
 
 	time = tcp_time_stamp - tp->rcvq_space.time;
 	if (time < (tp->rcv_rtt_est.rtt >> 3) || tp->rcv_rtt_est.rtt == 0)

@@ -573,12 +573,13 @@ void tcp_rcv_space_adjust(struct sock *s
 
 	if (sysctl_tcp_moderate_rcvbuf &&
 	    !(sk->sk_userlocks & SOCK_RCVBUF_LOCK)) {
-		int rcvwin, rcvmem, rcvbuf;
+		int rcvmem, rcvbuf;
+		u64 rcvwin;
 
 		/* minimal window to cope with packet losses, assuming
 		 * steady state. Add some cushion because of small variations.
 		 */
-		rcvwin = (copied << 1) + 16 * tp->advmss;
+		rcvwin = ((u64)copied << 1) + 16 * tp->advmss;
 
 		/* If rate increased by 25%,
 		 *	assume slow start, rcvwin = 3 * copied

@@ -598,7 +599,8 @@ void tcp_rcv_space_adjust(struct sock *s
 		while (tcp_win_from_space(rcvmem) < tp->advmss)
 			rcvmem += 128;
 
-		rcvbuf = min(rcvwin / tp->advmss * rcvmem, sysctl_tcp_rmem[2]);
+		do_div(rcvwin, tp->advmss);
+		rcvbuf = min_t(u64, rcvwin * rcvmem, sysctl_tcp_rmem[2]);
 		if (rcvbuf > sk->sk_rcvbuf) {
 			sk->sk_rcvbuf = rcvbuf;
 

3.18-stable review patch. If anyone has any objections, please let me know.

------------------

From: Maciej W. Rozycki <[email protected]>

commit c7e814628df65f424fe197dde73bfc67e4a244d7 upstream.

Use 64-bit accesses for 64-bit floating-point general registers with PTRACE_PEEKUSR, removing the truncation of their upper halves in the FR=1 mode, caused by commit bbd426f542cb ("MIPS: Simplify FP context access"), which inadvertently switched them to using 32-bit accesses.

The PTRACE_POKEUSR side is fine as it's never been broken and continues using 64-bit accesses.

Patchwork: https://patchwork.linux-mips.org/patch/19334/

---
 arch/mips/kernel/ptrace.c   |    2 +-
 arch/mips/kernel/ptrace32.c |    2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

@@ -713,7 +713,7 @@ long arch_ptrace(struct task_struct *chi
 				break;
 			}
 #endif
-			tmp = get_fpr32(&fregs[addr - FPR_BASE], 0);
+			tmp = get_fpr64(&fregs[addr - FPR_BASE], 0);
 			break;
 		case PC:
 			tmp = regs->cp0_epc;

@@ -107,7 +107,7 @@ long compat_arch_ptrace(struct task_stru
 						addr & 1);
 				break;
 			}
-			tmp = get_fpr32(&fregs[addr - FPR_BASE], 0);
+			tmp = get_fpr64(&fregs[addr - FPR_BASE], 0);
 			break;
 		case PC:
 			tmp = regs->cp0_epc;

3.18-stable review patch. If anyone has any objections, please let me know.

------------------

From: Al Viro <[email protected]>

commit 4faa99965e027cc057c5145ce45fa772caa04e8d upstream.

If io_destroy() gets to cancelling everything that can be cancelled and gets to kiocb_cancel() calling the function driver has left in ->ki_cancel, it becomes vulnerable to a race with IO completion. At that point req is already taken off the list and aio_complete() does *NOT* spin until we (in free_ioctx_users()) releases ->ctx_lock. As the result, it proceeds to kiocb_free(), freing req just it gets passed to ->ki_cancel().

Fix is simple - remove from the list after the call of kiocb_cancel(). All instances of ->ki_cancel() already have to cope with the being called with iocb still on list - that's what happens in io_cancel(2).

---
 fs/aio.c |    3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

@@ -560,9 +560,8 @@ static void free_ioctx_users(struct perc
 	while (!list_empty(&ctx->active_reqs)) {
 		req = list_first_entry(&ctx->active_reqs,
 				       struct kiocb, ki_list);
-
-		list_del_init(&req->ki_list);
 		kiocb_cancel(req);
+		list_del_init(&req->ki_list);
 	}
 
 	spin_unlock_irq(&ctx->ctx_lock);

3.18-stable review patch. If anyone has any objections, please let me know.

------------------

From: Hugh Dickins <[email protected]>

commit 145e1a71e090575c74969e3daa8136d1e5b99fc8 upstream.

George Boole would have noticed a slight error in 4.16 commit 69d763fc6d3a ("mm: pin address_space before dereferencing it while isolating an LRU page"). Fix it, to match both the comment above it, and the original behaviour.

Although anonymous pages are not marked PageDirty at first, we have an old habit of calling SetPageDirty when a page is removed from swap
cache: so there's a category of ex-swap pages that are easily migratable, but were inadvertently excluded from compaction's async migration in 4.16.

---
 mm/vmscan.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -1234,7 +1234,7 @@ int __isolate_lru_page(struct page *page
 				return ret;
 
 			mapping = page_mapping(page);
-			migrate_dirty = mapping && mapping->a_ops->migratepage;
+			migrate_dirty = !mapping || mapping->a_ops->migratepage;
 			unlock_page(page);
 			if (!migrate_dirty)
 				return ret;

3.18-stable review patch. If anyone has any objections, please let me know.

------------------

From: Linus Torvalds <[email protected]>

commit be83bbf806822b1b89e0a0f23cd87cddc409e429 upstream.

The internal VM "mmap()" interfaces are based on the mmap target doing everything using page indexes rather than byte offsets, because traditionally (ie 32-bit) we had the situation that the byte offset didn't fit in a register. So while the mmap virtual address was limited by the word size of the architecture, the backing store was not.

So we're basically passing "pgoff" around as a page index, in order to be able to describe backing store locations that are much bigger than the word size (think files larger than 4GB etc).

But while this all makes a ton of sense conceptually, we've been dogged by various drivers that don't really understand this, and internally work with byte offsets, and then try to work with the page index by turning it into a byte offset with "pgoff << PAGE_SHIFT".

Which obviously can overflow.

Adding the size of the mapping to it to get the byte offset of the end of the backing store just exacerbates the problem, and if you then use this overflow-prone value to check various limits of your device driver mmap capability, you're just setting yourself up for problems.

The correct thing for drivers to do is to do their limit math in page indices, the way the interface is designed. Because the generic mmap code _does_ test that the index doesn't overflow, since that's what the mmap code really cares about.

HOWEVER.

Finding and fixing various random drivers is a sisyphean task, so let's just see if we can just make the core mmap() code do the limiting for us. Realistically, the only "big" backing stores we need to care about are regular files and block devices, both of which are known to do this properly, and which have nice well-defined limits for how much data they can access.

So let's special-case just those two known cases, and then limit other random mmap users to a backing store that still fits in "unsigned long". Realistically, that's not much of a limit at all on 64-bit, and on 32-bit architectures the only worry might be the GPU drivers, which can have big physical address spaces.

To make it possible for drivers like that to say that they are 64-bit clean, this patch does repurpose the "FMODE_UNSIGNED_OFFSET" bit in the file flags to allow drivers to mark their file descriptors as safe in the full 64-bit mmap address space.

[ The timing for doing this is less than optimal, and this should really
go in a merge window. But realistically, this needs wide testing more than it needs anything else, and being main-line is the only way to do that.

So the earlier the better, even if it's outside the proper development
cycle - Linus ]

---
 mm/mmap.c |   32 ++++++++++++++++++++++++++++++++
 1 file changed, 32 insertions(+)

@@ -1271,6 +1271,35 @@ static inline int mlock_future_check(str
 	return 0;
 }
 
+static inline u64 file_mmap_size_max(struct file *file, struct inode *inode)
+{
+	if (S_ISREG(inode->i_mode))
+		return inode->i_sb->s_maxbytes;
+
+	if (S_ISBLK(inode->i_mode))
+		return MAX_LFS_FILESIZE;
+
+	/* Special "we do even unsigned file positions" case */
+	if (file->f_mode & FMODE_UNSIGNED_OFFSET)
+		return 0;
+
+	/* Yes, random drivers might want more. But I'm tired of buggy drivers */
+	return ULONG_MAX;
+}
+
+static inline bool file_mmap_ok(struct file *file, struct inode *inode,
+				unsigned long pgoff, unsigned long len)
+{
+	u64 maxsize = file_mmap_size_max(file, inode);
+
+	if (maxsize && len > maxsize)
+		return false;
+	maxsize -= len;
+	if (pgoff > maxsize >> PAGE_SHIFT)
+		return false;
+	return true;
+}
+
 /*
  * The caller must hold down_write(&current->mm->mmap_sem).
  */

@@ -1338,6 +1367,9 @@ unsigned long do_mmap_pgoff(struct file
 	if (file) {
 		struct inode *inode = file_inode(file);
 
+		if (!file_mmap_ok(file, inode, pgoff, len))
+			return -EOVERFLOW;
+
 		switch (flags & MAP_TYPE) {
 		case MAP_SHARED:
 			if ((prot&PROT_WRITE) && !(file->f_mode&FMODE_WRITE))

3.18-stable review patch. If anyone has any objections, please let me know.

------------------

From: Linus Torvalds <[email protected]>

commit 423913ad4ae5b3e8fb8983f70969fb522261ba26 upstream.

Commit be83bbf80682 ("mmap: introduce sane default mmap limits") was introduced to catch problems in various ad-hoc character device drivers doing mmap and getting the size limits wrong. In the process, it used "known good" limits for the normal cases of mapping regular files and block device drivers.

It turns out that the "s_maxbytes" limit was less "known good" than I thought. In particular, /proc doesn't set it, but exposes one regular file to mmap: /proc/vmcore. As a result, that file got limited to the default MAX_INT s_maxbytes value.

This went unnoticed for a while, because apparently the only thing that needs it is the s390 kernel zfcpdump, but there might be other tools that use this too.

Vasily suggested just changing s_maxbytes for all of /proc, which isn't wrong, but makes me nervous at this stage. So instead, just make the new mmap limit always be MAX_LFS_FILESIZE for regular files, which won't affect anything else. It wasn't the regular file case I was worried about.

I'd really prefer for maxsize to have been per-inode, but that is not how things are today.

---
 mm/mmap.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -1274,7 +1274,7 @@ static inline int mlock_future_check(str
 static inline u64 file_mmap_size_max(struct file *file, struct inode *inode)
 {
 	if (S_ISREG(inode->i_mode))
-		return inode->i_sb->s_maxbytes;
+		return MAX_LFS_FILESIZE;
 
 	if (S_ISBLK(inode->i_mode))
 		return MAX_LFS_FILESIZE;

3.18-stable review patch. If anyone has any objections, please let me know.

------------------

From: Dave Airlie <[email protected]>

commit 76ef6b28ea4f81c3d511866a9b31392caa833126 upstream.

Since we have the ttm and gem vma managers using a subset of the file address space for objects, and these start at 0x100000000 they will overflow the new mmap checks.

I've checked all the mmap routines I could see for any bad behaviour but overall most people use GEM/TTM VMA managers even the legacy drivers have a hashtable.

---
 drivers/gpu/drm/drm_fops.c |    1 +
 1 file changed, 1 insertion(+)

@@ -155,6 +155,7 @@ static int drm_open_helper(struct file *
 		return -ENOMEM;
 
 	filp->private_data = priv;
+	filp->f_mode |= FMODE_UNSIGNED_OFFSET;
 	priv->filp = filp;
 	priv->uid = current_euid();
 	priv->pid = get_pid(task_pid(current));

3.18-stable review patch. If anyone has any objections, please let me know.

------------------

From: Julia Lawall <[email protected]>

[ Upstream commit dd612f18a49b63af8b3a5f572d999bdb197385bc ]

Nearby code that also tests port suggests that the P0 constant should be used when port is zero.

The semantic match that finds this problem is as follows:
(http://coccinelle.lip6.fr/)

// <smpl>
@@
expression e,e1;
@@

* e ? e1 : e1
// </smpl>

---
 drivers/net/ethernet/broadcom/bnx2x/bnx2x_link.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -588,7 +588,7 @@ static void bnx2x_ets_e3b0_nig_disabled(
 	 * slots for the highest priority.
 	 */
 	REG_WR(bp, (port) ? NIG_REG_P1_TX_ARB_NUM_STRICT_ARB_SLOTS :
-		   NIG_REG_P1_TX_ARB_NUM_STRICT_ARB_SLOTS, 0x100);
+		   NIG_REG_P0_TX_ARB_NUM_STRICT_ARB_SLOTS, 0x100);
 	/* Mapping between the CREDIT_WEIGHT registers and actual client
 	 * numbers
 	 */

3.18-stable review patch. If anyone has any objections, please let me know.

------------------

From: Alexey Kodanev <[email protected]>

[ Upstream commit 2677d20677314101293e6da0094ede7b5526d2b1 ]

Syzbot reported the use-after-free in timer_is_static_object() [1].

This can happen because the structure for the rto timer (ccid2_hc_tx_sock) is removed in dccp_disconnect(), and ccid2_hc_tx_rto_expire() can be called after that.

The report [1] is similar to the one in commit 120e9dabaf55 ("dccp:
defer ccid_hc_tx_delete() at dismantle time"). And the fix is the same, delay freeing ccid2_hc_tx_sock structure, so that it is freed in dccp_sk_destruct().

[1]

==================================================================

BUG: KASAN: use-after-free in timer_is_static_object+0x80/0x90
kernel/time/timer.c:607
Read of size 8 at addr ffff8801bebb5118 by task syz-executor2/25299

CPU: 1 PID: 25299 Comm: syz-executor2 Not tainted 4.17.0-rc5+ #54
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS
Google 01/01/2011
Call Trace:
  <IRQ>
  __dump_stack lib/dump_stack.c:77 [inline]
  dump_stack+0x1b9/0x294 lib/dump_stack.c:113
  print_address_description+0x6c/0x20b mm/kasan/report.c:256
  kasan_report_error mm/kasan/report.c:354 [inline]
  kasan_report.cold.7+0x242/0x2fe mm/kasan/report.c:412
  __asan_report_load8_noabort+0x14/0x20 mm/kasan/report.c:433
  timer_is_static_object+0x80/0x90 kernel/time/timer.c:607
  debug_object_activate+0x2d9/0x670 lib/debugobjects.c:508
  debug_timer_activate kernel/time/timer.c:709 [inline]
  debug_activate kernel/time/timer.c:764 [inline]
  __mod_timer kernel/time/timer.c:1041 [inline]
  mod_timer+0x4d3/0x13b0 kernel/time/timer.c:1102
  sk_reset_timer+0x22/0x60 net/core/sock.c:2742
  ccid2_hc_tx_rto_expire+0x587/0x680 net/dccp/ccids/ccid2.c:147
  call_timer_fn+0x230/0x940 kernel/time/timer.c:1326
  expire_timers kernel/time/timer.c:1363 [inline]
  __run_timers+0x79e/0xc50 kernel/time/timer.c:1666
  run_timer_softirq+0x4c/0x70 kernel/time/timer.c:1692
  __do_softirq+0x2e0/0xaf5 kernel/softirq.c:285
  invoke_softirq kernel/softirq.c:365 [inline]
  irq_exit+0x1d1/0x200 kernel/softirq.c:405
  exiting_irq arch/x86/include/asm/apic.h:525 [inline]
  smp_apic_timer_interrupt+0x17e/0x710 arch/x86/kernel/apic/apic.c:1052
  apic_timer_interrupt+0xf/0x20 arch/x86/entry/entry_64.S:863
  </IRQ>
...
Allocated by task 25374:
  save_stack+0x43/0xd0 mm/kasan/kasan.c:448
  set_track mm/kasan/kasan.c:460 [inline]
  kasan_kmalloc+0xc4/0xe0 mm/kasan/kasan.c:553
  kasan_slab_alloc+0x12/0x20 mm/kasan/kasan.c:490
  kmem_cache_alloc+0x12e/0x760 mm/slab.c:3554
  ccid_new+0x25b/0x3e0 net/dccp/ccid.c:151
  dccp_hdlr_ccid+0x27/0x150 net/dccp/feat.c:44
  __dccp_feat_activate+0x184/0x270 net/dccp/feat.c:344
  dccp_feat_activate_values+0x3a7/0x819 net/dccp/feat.c:1538
  dccp_create_openreq_child+0x472/0x610 net/dccp/minisocks.c:128
  dccp_v4_request_recv_sock+0x12c/0xca0 net/dccp/ipv4.c:408
  dccp_v6_request_recv_sock+0x125d/0x1f10 net/dccp/ipv6.c:415
  dccp_check_req+0x455/0x6a0 net/dccp/minisocks.c:197
  dccp_v4_rcv+0x7b8/0x1f3f net/dccp/ipv4.c:841
  ip_local_deliver_finish+0x2e3/0xd80 net/ipv4/ip_input.c:215
  NF_HOOK include/linux/netfilter.h:288 [inline]
  ip_local_deliver+0x1e1/0x720 net/ipv4/ip_input.c:256
  dst_input include/net/dst.h:450 [inline]
  ip_rcv_finish+0x81b/0x2200 net/ipv4/ip_input.c:396
  NF_HOOK include/linux/netfilter.h:288 [inline]
  ip_rcv+0xb70/0x143d net/ipv4/ip_input.c:492
  __netif_receive_skb_core+0x26f5/0x3630 net/core/dev.c:4592
  __netif_receive_skb+0x2c/0x1e0 net/core/dev.c:4657
  process_backlog+0x219/0x760 net/core/dev.c:5337
  napi_poll net/core/dev.c:5735 [inline]
  net_rx_action+0x7b7/0x1930 net/core/dev.c:5801
  __do_softirq+0x2e0/0xaf5 kernel/softirq.c:285

Freed by task 25374:
  save_stack+0x43/0xd0 mm/kasan/kasan.c:448
  set_track mm/kasan/kasan.c:460 [inline]
  __kasan_slab_free+0x11a/0x170 mm/kasan/kasan.c:521
  kasan_slab_free+0xe/0x10 mm/kasan/kasan.c:528
  __cache_free mm/slab.c:3498 [inline]
  kmem_cache_free+0x86/0x2d0 mm/slab.c:3756
  ccid_hc_tx_delete+0xc3/0x100 net/dccp/ccid.c:190
  dccp_disconnect+0x130/0xc66 net/dccp/proto.c:286
  dccp_close+0x3bc/0xe60 net/dccp/proto.c:1045
  inet_release+0x104/0x1f0 net/ipv4/af_inet.c:427
  inet6_release+0x50/0x70 net/ipv6/af_inet6.c:460
  sock_release+0x96/0x1b0 net/socket.c:594
  sock_close+0x16/0x20 net/socket.c:1149
  __fput+0x34d/0x890 fs/file_table.c:209
  ____fput+0x15/0x20 fs/file_table.c:243
  task_work_run+0x1e4/0x290 kernel/task_work.c:113
  tracehook_notify_resume include/linux/tracehook.h:191 [inline]
  exit_to_usermode_loop+0x2bd/0x310 arch/x86/entry/common.c:166
  prepare_exit_to_usermode arch/x86/entry/common.c:196 [inline]
  syscall_return_slowpath arch/x86/entry/common.c:265 [inline]
  do_syscall_64+0x6ac/0x800 arch/x86/entry/common.c:290
  entry_SYSCALL_64_after_hwframe+0x49/0xbe

The buggy address belongs to the object at ffff8801bebb4cc0
  which belongs to the cache ccid2_hc_tx_sock of size 1240
The buggy address is located 1112 bytes inside of
  1240-byte region [ffff8801bebb4cc0, ffff8801bebb5198)
The buggy address belongs to the page:
page:ffffea0006faed00 count:1 mapcount:0 mapping:ffff8801bebb41c0
index:0xffff8801bebb5240 compound_mapcount: 0
flags: 0x2fffc0000008100(slab|head)
raw: 02fffc0000008100 ffff8801bebb41c0 ffff8801bebb5240 0000000100000003
raw: ffff8801cdba3138 ffffea0007634120 ffff8801cdbaab40 0000000000000000
page dumped because: kasan: bad access detected

...
==================================================================

---
 net/dccp/proto.c |    2 --
 1 file changed, 2 deletions(-)

@@ -280,9 +280,7 @@ int dccp_disconnect(struct sock *sk, int
 
 	dccp_clear_xmit_timers(sk);
 	ccid_hc_rx_delete(dp->dccps_hc_rx_ccid, sk);
-	ccid_hc_tx_delete(dp->dccps_hc_tx_ccid, sk);
 	dp->dccps_hc_rx_ccid = NULL;
-	dp->dccps_hc_tx_ccid = NULL;
 
 	__skb_queue_purge(&sk->sk_receive_queue);
 	__skb_queue_purge(&sk->sk_write_queue);

3.18-stable review patch. If anyone has any objections, please let me know.

------------------

From: Govindarajulu Varadarajan <[email protected]>

[ Upstream commit 322eaa06d55ebc1402a4a8d140945cff536638b4 ]

In commit 624dbf55a359b ("driver/net: enic: Try DMA 64 first, then failover to DMA") DMA mask was changed from 40 bits to 64 bits. Hardware actually supports only 47 bits.

---
 drivers/net/ethernet/cisco/enic/enic_main.c |    8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

@@ -2362,11 +2362,11 @@ static int enic_probe(struct pci_dev *pd
 	pci_set_master(pdev);
 
 	/* Query PCI controller on system for DMA addressing
-	 * limitation for the device.  Try 64-bit first, and
+	 * limitation for the device.  Try 47-bit first, and
 	 * fail to 32-bit.
 	 */
 
-	err = pci_set_dma_mask(pdev, DMA_BIT_MASK(64));
+	err = pci_set_dma_mask(pdev, DMA_BIT_MASK(47));
 	if (err) {
 		err = pci_set_dma_mask(pdev, DMA_BIT_MASK(32));
 		if (err) {

@@ -2380,10 +2380,10 @@ static int enic_probe(struct pci_dev *pd
 			goto err_out_release_regions;
 		}
 	} else {
-		err = pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(64));
+		err = pci_set_consistent_dma_mask(pdev, DMA_BIT_MASK(47));
 		if (err) {
 			dev_err(dev, "Unable to obtain %u-bit DMA "
-				"for consistent allocations, aborting\n", 64);
+				"for consistent allocations, aborting\n", 47);
 			goto err_out_release_regions;
 		}
 		using_dac = 1;

3.18-stable review patch. If anyone has any objections, please let me know.

------------------

From: Sabrina Dubroca <[email protected]>

[ Upstream commit 848235edb5c93ed086700584c8ff64f6d7fc778d ]

Currently, raw6_sk(sk)->ip6mr_table is set unconditionally during ip6_mroute_setsockopt(MRT6_TABLE). A subsequent attempt at the same setsockopt will fail with -ENOENT, since we haven't actually created that table.

A similar fix for ipv4 was included in commit 5e1859fbcc3c ("ipv4: ipmr:
various fixes and cleanups").

---
 net/ipv6/ip6mr.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

@@ -1788,7 +1788,8 @@ int ip6_mroute_setsockopt(struct sock *s
 		ret = 0;
 		if (!ip6mr_new_table(net, v))
 			ret = -ENOMEM;
-		raw6_sk(sk)->ip6mr_table = v;
+		else
+			raw6_sk(sk)->ip6mr_table = v;
 		rtnl_unlock();
 		return ret;
 	}

3.18-stable review patch. If anyone has any objections, please let me know.

------------------

From: Wenwen Wang <[email protected]>

[ Upstream commit 6009d1fe6ba3bb2dab55921da60465329cc1cd89 ]

In divasmain.c, the function divas_write() firstly invokes the function diva_xdi_open_adapter() to open the adapter that matches with the adapter number provided by the user, and then invokes the function diva_xdi_write() to perform the write operation using the matched adapter. The two functions diva_xdi_open_adapter() and diva_xdi_write() are located in diva.c.

In diva_xdi_open_adapter(), the user command is copied to the object 'msg' from the userspace pointer 'src' through the function pointer 'cp_fn', which eventually calls copy_from_user() to do the copy. Then, the adapter number 'msg.adapter' is used to find out a matched adapter from the 'adapter_queue'. A matched adapter will be returned if it is found. Otherwise, NULL is returned to indicate the failure of the verification on the adapter number.

As mentioned above, if a matched adapter is returned, the function diva_xdi_write() is invoked to perform the write operation. In this function, the user command is copied once again from the userspace pointer 'src', which is the same as the 'src' pointer in diva_xdi_open_adapter() as both of them are from the 'buf' pointer in divas_write(). Similarly, the copy is achieved through the function pointer 'cp_fn', which finally calls copy_from_user(). After the successful copy, the corresponding command processing handler of the matched adapter is invoked to perform the write operation.

It is obvious that there are two copies here from userspace, one is in diva_xdi_open_adapter(), and one is in diva_xdi_write(). Plus, both of these two copies share the same source userspace pointer, i.e., the 'buf' pointer in divas_write(). Given that a malicious userspace process can race to change the content pointed by the 'buf' pointer, this can pose potential security issues. For example, in the first copy, the user provides a valid adapter number to pass the verification process and a valid adapter can be found. Then the user can modify the adapter number to an invalid number. This way, the user can bypass the verification process of the adapter number and inject inconsistent data.

This patch reuses the data copied in
diva_xdi_open_adapter() and passes it to diva_xdi_write(). This way, the above issues can be avoided.

---
 drivers/isdn/hardware/eicon/diva.c      |   22 +++++++++++++++-------
 drivers/isdn/hardware/eicon/diva.h      |    5 +++--
 drivers/isdn/hardware/eicon/divasmain.c |   18 +++++++++++-------
 3 files changed, 29 insertions(+), 16 deletions(-)

@@ -387,10 +387,10 @@ void divasa_xdi_driver_unload(void)
 **  Receive and process command from user mode utility
 */
 void *diva_xdi_open_adapter(void *os_handle, const void __user *src,
-			    int length,
+			    int length, void *mptr,
 			    divas_xdi_copy_from_user_fn_t cp_fn)
 {
-	diva_xdi_um_cfg_cmd_t msg;
+	diva_xdi_um_cfg_cmd_t *msg = (diva_xdi_um_cfg_cmd_t *)mptr;
 	diva_os_xdi_adapter_t *a = NULL;
 	diva_os_spin_lock_magic_t old_irql;
 	struct list_head *tmp;

@@ -400,21 +400,21 @@ void *diva_xdi_open_adapter(void *os_han
 			 length, sizeof(diva_xdi_um_cfg_cmd_t)))
 			return NULL;
 	}
-	if ((*cp_fn) (os_handle, &msg, src, sizeof(msg)) <= 0) {
+	if ((*cp_fn) (os_handle, msg, src, sizeof(*msg)) <= 0) {
 		DBG_ERR(("A: A(?) open, write error"))
 			return NULL;
 	}
 	diva_os_enter_spin_lock(&adapter_lock, &old_irql, "open_adapter");
 	list_for_each(tmp, &adapter_queue) {
 		a = list_entry(tmp, diva_os_xdi_adapter_t, link);
-		if (a->controller == (int)msg.adapter)
+		if (a->controller == (int)msg->adapter)
 			break;
 		a = NULL;
 	}
 	diva_os_leave_spin_lock(&adapter_lock, &old_irql, "open_adapter");
 
 	if (!a) {
-		DBG_ERR(("A: A(%d) open, adapter not found", msg.adapter))
+		DBG_ERR(("A: A(%d) open, adapter not found", msg->adapter))
 			}
 
 	return (a);

@@ -436,8 +436,10 @@ void diva_xdi_close_adapter(void *adapte
 
 int
 diva_xdi_write(void *adapter, void *os_handle, const void __user *src,
-	       int length, divas_xdi_copy_from_user_fn_t cp_fn)
+	       int length, void *mptr,
+	       divas_xdi_copy_from_user_fn_t cp_fn)
 {
+	diva_xdi_um_cfg_cmd_t *msg = (diva_xdi_um_cfg_cmd_t *)mptr;
 	diva_os_xdi_adapter_t *a = (diva_os_xdi_adapter_t *) adapter;
 	void *data;
 

@@ -458,7 +460,13 @@ diva_xdi_write(void *adapter, void *os_h
 			return (-2);
 	}
 
-	length = (*cp_fn) (os_handle, data, src, length);
+	if (msg) {
+		*(diva_xdi_um_cfg_cmd_t *)data = *msg;
+		length = (*cp_fn) (os_handle, (char *)data + sizeof(*msg),
+				   src + sizeof(*msg), length - sizeof(*msg));
+	} else {
+		length = (*cp_fn) (os_handle, data, src, length);
+	}
 	if (length > 0) {
 		if ((*(a->interface.cmd_proc))
 		    (a, (diva_xdi_um_cfg_cmd_t *) data, length)) {

@@ -19,10 +19,11 @@ int diva_xdi_read(void *adapter, void *o
 		  int max_length, divas_xdi_copy_to_user_fn_t cp_fn);
 
 int diva_xdi_write(void *adapter, void *os_handle, const void __user *src,
-		   int length, divas_xdi_copy_from_user_fn_t cp_fn);
+		   int length, void *msg,
+		   divas_xdi_copy_from_user_fn_t cp_fn);
 
 void *diva_xdi_open_adapter(void *os_handle, const void __user *src,
-			    int length,
+			    int length, void *msg,
 			    divas_xdi_copy_from_user_fn_t cp_fn);
 
 void diva_xdi_close_adapter(void *adapter, void *os_handle);

@@ -591,19 +591,22 @@ static int divas_release(struct inode *i
 static ssize_t divas_write(struct file *file, const char __user *buf,
 			   size_t count, loff_t *ppos)
 {
+	diva_xdi_um_cfg_cmd_t msg;
 	int ret = -EINVAL;
 
 	if (!file->private_data) {
 		file->private_data = diva_xdi_open_adapter(file, buf,
-							   count,
+							   count, &msg,
 							   xdi_copy_from_user);
-	}
-	if (!file->private_data) {
-		return (-ENODEV);
+		if (!file->private_data)
+			return (-ENODEV);
+		ret = diva_xdi_write(file->private_data, file,
+				     buf, count, &msg, xdi_copy_from_user);
+	} else {
+		ret = diva_xdi_write(file->private_data, file,
+				     buf, count, NULL, xdi_copy_from_user);
 	}
 
-	ret = diva_xdi_write(file->private_data, file,
-			     buf, count, xdi_copy_from_user);
 	switch (ret) {
 	case -1:		/* Message should be removed from rx mailbox first */
 		ret = -EBUSY;

@@ -622,11 +625,12 @@ static ssize_t divas_write(struct file *
 static ssize_t divas_read(struct file *file, char __user *buf,
 			  size_t count, loff_t *ppos)
 {
+	diva_xdi_um_cfg_cmd_t msg;
 	int ret = -EINVAL;
 
 	if (!file->private_data) {
 		file->private_data = diva_xdi_open_adapter(file, buf,
-							   count,
+							   count, &msg,
 							   xdi_copy_from_user);
 	}
 	if (!file->private_data) {

3.18-stable review patch. If anyone has any objections, please let me know.

------------------

From: Eric Dumazet <[email protected]>

[ Upstream commit eb73190f4fbeedf762394e92d6a4ec9ace684c88 ]

syzbot was able to trick af_packet again [1]

Various commits tried to address the problem in the past, but failed to take into account V3 header size.

[1]

tpacket_rcv: packet too big, clamped from 72 to 4294967224. macoff=96

BUG: KASAN: use-after-free in prb_run_all_ft_ops net/packet/af_packet.c:1016 [inline]
BUG: KASAN: use-after-free in prb_fill_curr_block.isra.59+0x4e5/0x5c0 net/packet/af_packet.c:1039
Write of size 2 at addr ffff8801cb62000e by task kworker/1:2/2106

CPU: 1 PID: 2106 Comm: kworker/1:2 Not tainted 4.17.0-rc7+ #77
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
Workqueue: ipv6_addrconf addrconf_dad_work
Call Trace:
 __dump_stack lib/dump_stack.c:77 [inline]
 dump_stack+0x1b9/0x294 lib/dump_stack.c:113
 print_address_description+0x6c/0x20b mm/kasan/report.c:256
 kasan_report_error mm/kasan/report.c:354 [inline]
 kasan_report.cold.7+0x242/0x2fe mm/kasan/report.c:412
 __asan_report_store2_noabort+0x17/0x20 mm/kasan/report.c:436
 prb_run_all_ft_ops net/packet/af_packet.c:1016 [inline]
 prb_fill_curr_block.isra.59+0x4e5/0x5c0 net/packet/af_packet.c:1039
 __packet_lookup_frame_in_block net/packet/af_packet.c:1094 [inline]
 packet_current_rx_frame net/packet/af_packet.c:1117 [inline]
 tpacket_rcv+0x1866/0x3340 net/packet/af_packet.c:2282
 dev_queue_xmit_nit+0x891/0xb90 net/core/dev.c:2018
 xmit_one net/core/dev.c:3049 [inline]
 dev_hard_start_xmit+0x16b/0xc10 net/core/dev.c:3069
 __dev_queue_xmit+0x2724/0x34c0 net/core/dev.c:3584
 dev_queue_xmit+0x17/0x20 net/core/dev.c:3617
 neigh_resolve_output+0x679/0xad0 net/core/neighbour.c:1358
 neigh_output include/net/neighbour.h:482 [inline]
 ip6_finish_output2+0xc9c/0x2810 net/ipv6/ip6_output.c:120
 ip6_finish_output+0x5fe/0xbc0 net/ipv6/ip6_output.c:154
 NF_HOOK_COND include/linux/netfilter.h:277 [inline]
 ip6_output+0x227/0x9b0 net/ipv6/ip6_output.c:171
 dst_output include/net/dst.h:444 [inline]
 NF_HOOK include/linux/netfilter.h:288 [inline]
 ndisc_send_skb+0x100d/0x1570 net/ipv6/ndisc.c:491
 ndisc_send_ns+0x3c1/0x8d0 net/ipv6/ndisc.c:633
 addrconf_dad_work+0xbef/0x1340 net/ipv6/addrconf.c:4033
 process_one_work+0xc1e/0x1b50 kernel/workqueue.c:2145
 worker_thread+0x1cc/0x1440 kernel/workqueue.c:2279
 kthread+0x345/0x410 kernel/kthread.c:240
 ret_from_fork+0x3a/0x50 arch/x86/entry/entry_64.S:412

The buggy address belongs to the page:
page:ffffea00072d8800 count:0 mapcount:-127 mapping:0000000000000000 index:0xffff8801cb620e80
flags: 0x2fffc0000000000()
raw: 02fffc0000000000 0000000000000000 ffff8801cb620e80 00000000ffffff80
raw: ffffea00072e3820 ffffea0007132d20 0000000000000002 0000000000000000
page dumped because: kasan: bad access detected

Memory state around the buggy address:
 ffff8801cb61ff00: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
 ffff8801cb61ff80: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
>ffff8801cb620000: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
                      ^
 ffff8801cb620080: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
 ffff8801cb620100: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff

---
 net/packet/af_packet.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

@@ -3877,7 +3877,7 @@ static int packet_set_ring(struct sock *
 			goto out;
 		if (po->tp_version >= TPACKET_V3 &&
 		    req->tp_block_size <=
-			  BLK_PLUS_PRIV((u64)req_u->req3.tp_sizeof_priv))
+		    BLK_PLUS_PRIV((u64)req_u->req3.tp_sizeof_priv) + sizeof(struct tpacket3_hdr))
 			goto out;
 		if (unlikely(req->tp_frame_size < po->tp_hdrlen +
 					po->tp_reserve))

3.18-stable review patch. If anyone has any objections, please let me know.

------------------

From: Jack Morgenstein <[email protected]>

[ Upstream commit d546b67cda015fb92bfee93d5dc0ceadb91deaee ]

spin_lock/unlock was used instead of spin_un/lock_irq in a procedure used in process space, on a spinlock which can be grabbed in an interrupt.

This caused the stack trace below to be displayed (on kernel 4.17.0-rc1 compiled with Lock Debugging enabled):

[  154.661474] WARNING: SOFTIRQ-safe -> SOFTIRQ-unsafe lock order detected
[  154.668909] 4.17.0-rc1-rdma_rc_mlx+ #3 Tainted: G          I
[  154.675856] -----------------------------------------------------
[  154.682706] modprobe/10159 [HC0[0]:SC0[0]:HE0:SE1] is trying to acquire:
[  154.690254] 00000000f3b0e495 (&(&qp_table->lock)->rlock){+.+.}, at: mlx4_qp_remove+0x20/0x50 [mlx4_core]
[  154.700927]
and this task is already holding:
[  154.707461] 0000000094373b5d (&(&cq->lock)->rlock/1){....}, at: destroy_qp_common+0x111/0x560 [mlx4_ib]
[  154.718028] which would create a new lock dependency:
[  154.723705]  (&(&cq->lock)->rlock/1){....} -> (&(&qp_table->lock)->rlock){+.+.}
[  154.731922]
but this new dependency connects a SOFTIRQ-irq-safe lock:
[  154.740798]  (&(&cq->lock)->rlock){..-.}
[  154.740800]
... which became SOFTIRQ-irq-safe at:
[  154.752163]   _raw_spin_lock_irqsave+0x3e/0x50
[  154.757163]   mlx4_ib_poll_cq+0x36/0x900 [mlx4_ib]
[  154.762554]   ipoib_tx_poll+0x4a/0xf0 [ib_ipoib]
...
to a SOFTIRQ-irq-unsafe lock:
[  154.815603]  (&(&qp_table->lock)->rlock){+.+.}
[  154.815604]
... which became SOFTIRQ-irq-unsafe at:
[  154.827718] ...
[  154.827720]   _raw_spin_lock+0x35/0x50
[  154.833912]   mlx4_qp_lookup+0x1e/0x50 [mlx4_core]
[  154.839302]   mlx4_flow_attach+0x3f/0x3d0 [mlx4_core]

Since mlx4_qp_lookup() is called only in process space, we can simply replace the spin_un/lock calls with spin_un/lock_irq calls.

---
 drivers/net/ethernet/mellanox/mlx4/qp.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

@@ -363,11 +363,11 @@ struct mlx4_qp *mlx4_qp_lookup(struct ml
 	struct mlx4_qp_table *qp_table = &mlx4_priv(dev)->qp_table;
 	struct mlx4_qp *qp;
 
-	spin_lock(&qp_table->lock);
+	spin_lock_irq(&qp_table->lock);
 
 	qp = __mlx4_qp_lookup(dev, qpn);
 
-	spin_unlock(&qp_table->lock);
+	spin_unlock_irq(&qp_table->lock);
 	return qp;
 }
 

3.18-stable review patch. If anyone has any objections, please let me know.

------------------

From: Dan Carpenter <[email protected]>

[ Upstream commit 25ea66544bfd1d9df1b7e1502f8717e85fa1e6e6 ]

This code was introduced in 2011 around the same time that we made netdev_features_t a u64 type. These days a u32 is not big enough to hold all the potential features.

---
 drivers/net/team/team.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

@@ -989,7 +989,8 @@ static void team_port_disable(struct tea
 static void __team_compute_features(struct team *team)
 {
 	struct team_port *port;
-	u32 vlan_features = TEAM_VLAN_FEATURES & NETIF_F_ALL_FOR_ALL;
+	netdev_features_t vlan_features = TEAM_VLAN_FEATURES &
+					  NETIF_F_ALL_FOR_ALL;
 	unsigned short max_hard_header_len = ETH_HLEN;
 	unsigned int dst_release_flag = IFF_XMIT_DST_RELEASE |
 					IFF_XMIT_DST_RELEASE_PERM;

3.18-stable review patch. If anyone has any objections, please let me know.

------------------

From: Eric Dumazet <[email protected]>

[ Upstream commit 644c7eebbfd59e72982d11ec6cc7d39af12450ae ]

It seems that rtnl_group_changelink() can call do_setlink while a prior call to validate_linkmsg(dev = NULL, ...) could not validate IFLA_ADDRESS / IFLA_BROADCAST

Make sure do_setlink() calls validate_linkmsg() instead of letting its callers having this responsibility.

With help from Dmitry Vyukov, thanks a lot !

BUG: KMSAN: uninit-value in is_valid_ether_addr include/linux/etherdevice.h:199 [inline]
BUG: KMSAN: uninit-value in eth_prepare_mac_addr_change net/ethernet/eth.c:275 [inline]
BUG: KMSAN: uninit-value in eth_mac_addr+0x203/0x2b0 net/ethernet/eth.c:308
CPU: 1 PID: 8695 Comm: syz-executor3 Not tainted 4.17.0-rc5+ #103
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
Call Trace:
 __dump_stack lib/dump_stack.c:77 [inline]
 dump_stack+0x185/0x1d0 lib/dump_stack.c:113
 kmsan_report+0x149/0x260 mm/kmsan/kmsan.c:1084
 __msan_warning_32+0x6e/0xc0 mm/kmsan/kmsan_instr.c:686
 is_valid_ether_addr include/linux/etherdevice.h:199 [inline]
 eth_prepare_mac_addr_change net/ethernet/eth.c:275 [inline]
 eth_mac_addr+0x203/0x2b0 net/ethernet/eth.c:308
 dev_set_mac_address+0x261/0x530 net/core/dev.c:7157
 do_setlink+0xbc3/0x5fc0 net/core/rtnetlink.c:2317
 rtnl_group_changelink net/core/rtnetlink.c:2824 [inline]
 rtnl_newlink+0x1fe9/0x37a0 net/core/rtnetlink.c:2976
 rtnetlink_rcv_msg+0xa32/0x1560 net/core/rtnetlink.c:4646
 netlink_rcv_skb+0x378/0x600 net/netlink/af_netlink.c:2448
 rtnetlink_rcv+0x50/0x60 net/core/rtnetlink.c:4664
 netlink_unicast_kernel net/netlink/af_netlink.c:1310 [inline]
 netlink_unicast+0x1678/0x1750 net/netlink/af_netlink.c:1336
 netlink_sendmsg+0x104f/0x1350 net/netlink/af_netlink.c:1901
 sock_sendmsg_nosec net/socket.c:629 [inline]
 sock_sendmsg net/socket.c:639 [inline]
 ___sys_sendmsg+0xec0/0x1310 net/socket.c:2117
 __sys_sendmsg net/socket.c:2155 [inline]
 __do_sys_sendmsg net/socket.c:2164 [inline]
 __se_sys_sendmsg net/socket.c:2162 [inline]
 __x64_sys_sendmsg+0x331/0x460 net/socket.c:2162
 do_syscall_64+0x152/0x230 arch/x86/entry/common.c:287
 entry_SYSCALL_64_after_hwframe+0x44/0xa9
RIP: 0033:0x455a09
RSP: 002b:00007fc07480ec68 EFLAGS: 00000246 ORIG_RAX: 000000000000002e
RAX: ffffffffffffffda RBX: 00007fc07480f6d4 RCX: 0000000000455a09
RDX: 0000000000000000 RSI: 00000000200003c0 RDI: 0000000000000014
RBP: 000000000072bea0 R08: 0000000000000000 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000246 R12: 00000000ffffffff
R13: 00000000000005d0 R14: 00000000006fdc20 R15: 0000000000000000

Uninit was stored to memory at:

 kmsan_save_stack_with_flags mm/kmsan/kmsan.c:279 [inline]
 kmsan_save_stack mm/kmsan/kmsan.c:294 [inline]
 kmsan_internal_chain_origin+0x12b/0x210 mm/kmsan/kmsan.c:685
 kmsan_memcpy_origins+0x11d/0x170 mm/kmsan/kmsan.c:527
 __msan_memcpy+0x109/0x160 mm/kmsan/kmsan_instr.c:478
 do_setlink+0xb84/0x5fc0 net/core/rtnetlink.c:2315
 rtnl_group_changelink net/core/rtnetlink.c:2824 [inline]
 rtnl_newlink+0x1fe9/0x37a0 net/core/rtnetlink.c:2976
 rtnetlink_rcv_msg+0xa32/0x1560 net/core/rtnetlink.c:4646
 netlink_rcv_skb+0x378/0x600 net/netlink/af_netlink.c:2448
 rtnetlink_rcv+0x50/0x60 net/core/rtnetlink.c:4664
 netlink_unicast_kernel net/netlink/af_netlink.c:1310 [inline]
 netlink_unicast+0x1678/0x1750 net/netlink/af_netlink.c:1336
 netlink_sendmsg+0x104f/0x1350 net/netlink/af_netlink.c:1901
 sock_sendmsg_nosec net/socket.c:629 [inline]
 sock_sendmsg net/socket.c:639 [inline]
 ___sys_sendmsg+0xec0/0x1310 net/socket.c:2117
 __sys_sendmsg net/socket.c:2155 [inline]
 __do_sys_sendmsg net/socket.c:2164 [inline]
 __se_sys_sendmsg net/socket.c:2162 [inline]
 __x64_sys_sendmsg+0x331/0x460 net/socket.c:2162
 do_syscall_64+0x152/0x230 arch/x86/entry/common.c:287
 entry_SYSCALL_64_after_hwframe+0x44/0xa9
Uninit was created at:
 kmsan_save_stack_with_flags mm/kmsan/kmsan.c:279 [inline]
 kmsan_internal_poison_shadow+0xb8/0x1b0 mm/kmsan/kmsan.c:189
 kmsan_kmalloc+0x94/0x100 mm/kmsan/kmsan.c:315
 kmsan_slab_alloc+0x10/0x20 mm/kmsan/kmsan.c:322
 slab_post_alloc_hook mm/slab.h:446 [inline]
 slab_alloc_node mm/slub.c:2753 [inline]
 __kmalloc_node_track_caller+0xb32/0x11b0 mm/slub.c:4395
 __kmalloc_reserve net/core/skbuff.c:138 [inline]
 __alloc_skb+0x2cb/0x9e0 net/core/skbuff.c:206
 alloc_skb include/linux/skbuff.h:988 [inline]
 netlink_alloc_large_skb net/netlink/af_netlink.c:1182 [inline]
 netlink_sendmsg+0x76e/0x1350 net/netlink/af_netlink.c:1876
 sock_sendmsg_nosec net/socket.c:629 [inline]
 sock_sendmsg net/socket.c:639 [inline]
 ___sys_sendmsg+0xec0/0x1310 net/socket.c:2117
 __sys_sendmsg net/socket.c:2155 [inline]
 __do_sys_sendmsg net/socket.c:2164 [inline]
 __se_sys_sendmsg net/socket.c:2162 [inline]
 __x64_sys_sendmsg+0x331/0x460 net/socket.c:2162
 do_syscall_64+0x152/0x230 arch/x86/entry/common.c:287
 entry_SYSCALL_64_after_hwframe+0x44/0xa9

---
 net/core/rtnetlink.c |    8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

@@ -1503,6 +1503,10 @@ static int do_setlink(const struct sk_bu
 	const struct net_device_ops *ops = dev->netdev_ops;
 	int err;
 
+	err = validate_linkmsg(dev, tb);
+	if (err < 0)
+		return err;
+
 	if (tb[IFLA_NET_NS_PID] || tb[IFLA_NET_NS_FD]) {
 		struct net *net = rtnl_link_get_net(dev_net(dev), tb);
 		if (IS_ERR(net)) {

@@ -1783,10 +1787,6 @@ static int rtnl_setlink(struct sk_buff *
 		goto errout;
 	}
 
-	err = validate_linkmsg(dev, tb);
-	if (err < 0)
-		goto errout;
-
 	err = do_setlink(skb, dev, ifm, tb, ifname, 0);
 errout:
 	return err;

Merged, compiled with -Werror, and installed on my Pixel XL.

No initial issues noticed in dmesg or general usage.

Thanks!
Nathan

Merged, compiled and flashed on the OnePlus 3T, no immediate regressions noticed. Thanks for the update!

Harsh Shandilya, PRJKT Development LLC

Great, thanks for testing and letting me know.

greg k-h

Compiled and booted on my test system. No dmesg regressions.

thanks,
-- Shuah

That is a lot of new failures, did the whole lab fail, or is this really a problem in v3.18.112 here?

thanks,

greg k-h

[...]

Whole lab failure (more precisely, lab operator failure) ;)

gak, I updated the rootfs images to the latest buildroot, which forced me to upgrade the kernel headers used to build the rootfs from v3.10 to v4.4. So I guess it's no surprise that every single board panic'd as soon as it hit userspace.

I downgraded the rootfs, and re-ran all those boot tests, and now things are 100% passing in my lab.

Sorry for the noise,

Kevin

I build my root file systems with buildroot, and had a similar problem. My fix was to patch buildroot to let me use older linux headers.

Guenter

Could you share the patch?

I was lazy and just rolled back the buildroot release.

Thanks,

Kevin

Try [email protected]:groeck/buildroot.git branch local-2018.05.

Guenter

Build results:

	total: 138 pass: 136 fail: 2
Failed builds:
	i386:tools/perf
	x86_64:tools/perf
Qemu test results:
	total: 127 pass: 127 fail: 0

The perf test only fails because I run "make tools/perf" instead of "(cd tools/perf; make)". Fixing this requires the following two patches.

379a9a28a1b4 tools build: No need to make libapi for perf explicitly 16671c1e1cac tools build: Fix Makefile(s) to properly invoke tools build

Details are available at http://kerneltests.org/builders/.

Guenter

Thanks for this, and the other 3 results. I'll just add these commits to the next kernel as this is not a new failure.

greg k-h

I've queued these up now, but I still can't build perf for 3.18. Given that no one has reported a problem with it so far, I'm guessing that no one cares, so I'll just leave it as-is for now.

thanks,

greg k-h

Ah, I should have waited for the test build to complete after applying the patches. Sorry for that. Apparently there are additional dependencies. (cd tools/perf; make) does work. I'll switch to use that for v3.16 (which has the same problem) and v3.18.

Guenter