From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: linux-kernel@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
stable@vger.kernel.org, Paul Menzel <pmenzel@molgen.mpg.de>,
Josh Poimboeuf <jpoimboe@redhat.com>,
"Steven Rostedt (VMware)" <rostedt@goodmis.org>,
"Rafael J . Wysocki" <rjw@rjwysocki.net>,
linux-acpi@vger.kernel.org, Borislav Petkov <bp@alien8.de>,
Len Brown <lenb@kernel.org>, Thomas Gleixner <tglx@linutronix.de>
Subject: [PATCH 4.9 54/54] ftrace/x86: Fix triple fault with graph tracing and suspend-to-ram
Date: Mon, 1 May 2017 14:32:01 -0700 [thread overview]
Message-ID: <20170501212633.941060776@linuxfoundation.org> (raw)
In-Reply-To: <20170501212631.798128131@linuxfoundation.org>
4.9-stable review patch. If anyone has any objections, please let me know.
------------------
From: Josh Poimboeuf <jpoimboe@redhat.com>
commit 34a477e5297cbaa6ecc6e17c042a866e1cbe80d6 upstream.
On x86-32, with CONFIG_FIRMWARE and multiple CPUs, if you enable function
graph tracing and then suspend to RAM, it will triple fault and reboot when
it resumes.
The first fault happens when booting a secondary CPU:
startup_32_smp()
load_ucode_ap()
prepare_ftrace_return()
ftrace_graph_is_dead()
(accesses 'kill_ftrace_graph')
The early head_32.S code calls into load_ucode_ap(), which has an an
ftrace hook, so it calls prepare_ftrace_return(), which calls
ftrace_graph_is_dead(), which tries to access the global
'kill_ftrace_graph' variable with a virtual address, causing a fault
because the CPU is still in real mode.
The fix is to add a check in prepare_ftrace_return() to make sure it's
running in protected mode before continuing. The check makes sure the
stack pointer is a virtual kernel address. It's a bit of a hack, but
it's not very intrusive and it works well enough.
For reference, here are a few other (more difficult) ways this could
have potentially been fixed:
- Move startup_32_smp()'s call to load_ucode_ap() down to *after* paging
is enabled. (No idea what that would break.)
- Track down load_ucode_ap()'s entire callee tree and mark all the
functions 'notrace'. (Probably not realistic.)
- Pause graph tracing in ftrace_suspend_notifier_call() or bringup_cpu()
or __cpu_up(), and ensure that the pause facility can be queried from
real mode.
Reported-by: Paul Menzel <pmenzel@molgen.mpg.de>
Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com>
Tested-by: Paul Menzel <pmenzel@molgen.mpg.de>
Reviewed-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Cc: "Rafael J . Wysocki" <rjw@rjwysocki.net>
Cc: linux-acpi@vger.kernel.org
Cc: Borislav Petkov <bp@alien8.de>
Cc: Len Brown <lenb@kernel.org>
Link: http://lkml.kernel.org/r/5c1272269a580660703ed2eccf44308e790c7a98.1492123841.git.jpoimboe@redhat.com
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
arch/x86/kernel/ftrace.c | 12 ++++++++++++
1 file changed, 12 insertions(+)
--- a/arch/x86/kernel/ftrace.c
+++ b/arch/x86/kernel/ftrace.c
@@ -983,6 +983,18 @@ void prepare_ftrace_return(unsigned long
unsigned long return_hooker = (unsigned long)
&return_to_handler;
+ /*
+ * When resuming from suspend-to-ram, this function can be indirectly
+ * called from early CPU startup code while the CPU is in real mode,
+ * which would fail miserably. Make sure the stack pointer is a
+ * virtual address.
+ *
+ * This check isn't as accurate as virt_addr_valid(), but it should be
+ * good enough for this purpose, and it's fast.
+ */
+ if (unlikely((long)__builtin_frame_address(0) >= 0))
+ return;
+
if (unlikely(ftrace_graph_is_dead()))
return;
next prev parent reply other threads:[~2017-05-01 21:35 UTC|newest]
Thread overview: 53+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-05-01 21:31 [PATCH 4.9 00/54] 4.9.26-stable review Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 01/54] [PATCH] Revert "mmc: sdhci-msm: Enable few quirks" Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 02/54] ping: implement proper locking Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 03/54] sparc64: kern_addr_valid regression Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 04/54] sparc64: Fix kernel panic due to erroneous #ifdef surrounding pmd_write() Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 05/54] net: neigh: guard against NULL solicit() method Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 06/54] net: phy: handle state correctly in phy_stop_machine Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 07/54] kcm: return immediately after copy_from_user() failure Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 08/54] bpf: improve verifier packet range checks Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 09/54] net/mlx5: Avoid dereferencing uninitialized pointer Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 10/54] l2tp: hold tunnel socket when handling control frames in l2tp_ip and l2tp_ip6 Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 11/54] l2tp: purge socket queues in the .destruct() callback Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 12/54] net/packet: fix overflow in check for tp_frame_nr Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 13/54] net/packet: fix overflow in check for tp_reserve Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 14/54] l2tp: take reference on sessions being dumped Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 15/54] l2tp: fix PPP pseudo-wire auto-loading Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 16/54] net: ipv4: fix multipath RTM_GETROUTE behavior when iif is given Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 17/54] sctp: listen on the sock only when its state is listening or closed Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 18/54] tcp: clear saved_syn in tcp_disconnect() Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 19/54] ipv6: Fix idev->addr_list corruption Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 20/54] net-timestamp: avoid use-after-free in ip_recv_error Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 21/54] net: vrf: Fix setting NLM_F_EXCL flag when adding l3mdev rule Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 22/54] sh_eth: unmap DMA buffers when freeing rings Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 24/54] gso: Validate assumption of frag_list segementation Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 25/54] net: ipv6: RTF_PCPU should not be settable from userspace Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 26/54] netpoll: Check for skb->queue_mapping Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 27/54] ip6mr: fix notification device destruction Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 28/54] net/mlx5: Fix driver load bad flow when having fw initializing timeout Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 29/54] net/mlx5e: Fix small packet threshold Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 30/54] net/mlx5e: Fix ETHTOOL_GRXCLSRLALL handling Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 31/54] macvlan: Fix device ref leak when purging bc_queue Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 32/54] net: ipv6: regenerate host route if moved to gc list Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 33/54] net: phy: fix auto-negotiation stall due to unavailable interrupt Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 34/54] ipv6: check skb->protocol before lookup for nexthop Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 35/54] tcp: memset ca_priv data to 0 properly Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 36/54] ipv6: check raw payload size correctly in ioctl Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 37/54] ALSA: oxfw: fix regression to handle Stanton SCS.1m/1d Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 38/54] ALSA: firewire-lib: fix inappropriate assignment between signed/unsigned type Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 39/54] ALSA: seq: Dont break snd_use_lock_sync() loop by timeout Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 40/54] ARC: [plat-eznps] Fix build error Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 41/54] MIPS: KGDB: Use kernel context for sleeping threads Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 42/54] MIPS: cevt-r4k: Fix out-of-bounds array access Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 43/54] MIPS: Avoid BUG warning in arch_check_elf Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 44/54] p9_client_readdir() fix Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 45/54] ASoC: intel: Fix PM and non-atomic crash in bytcr drivers Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 48/54] nfsd4: minor NFSv2/v3 write decoding cleanup Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 50/54] ceph: fix recursion between ceph_set_acl() and __ceph_setattr() Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 51/54] macsec: avoid heap overflow in skb_to_sgvec Greg Kroah-Hartman
2017-05-01 21:31 ` [PATCH 4.9 52/54] net: can: usb: gs_usb: Fix buffer on stack Greg Kroah-Hartman
2017-05-01 21:32 ` [PATCH 4.9 53/54] ARCv2: save r30 on kernel entry as gcc uses it for code-gen Greg Kroah-Hartman
2017-05-01 21:32 ` Greg Kroah-Hartman [this message]
[not found] ` <590808bb.a121ed0a.b040f.045c@mx.google.com>
2017-05-02 13:53 ` [PATCH 4.9 00/54] 4.9.26-stable review Shuah Khan
2017-05-02 17:36 ` Guenter Roeck
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=20170501212633.941060776@linuxfoundation.org \
--to=gregkh@linuxfoundation.org \
--cc=bp@alien8.de \
--cc=jpoimboe@redhat.com \
--cc=lenb@kernel.org \
--cc=linux-acpi@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=pmenzel@molgen.mpg.de \
--cc=rjw@rjwysocki.net \
--cc=rostedt@goodmis.org \
--cc=stable@vger.kernel.org \
--cc=tglx@linutronix.de \
/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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).