From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: linux-kernel@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
stable@vger.kernel.org, Thomas Gleixner <tglx@linutronix.de>,
Feng Tang <feng.tang@intel.com>,
"Paul E. McKenney" <paulmck@kernel.org>
Subject: [PATCH 5.4 60/70] x86/tsc: Disable clocksource watchdog for TSC on qualified platorms
Date: Mon, 6 Dec 2021 15:57:04 +0100 [thread overview]
Message-ID: <20211206145553.994509372@linuxfoundation.org> (raw)
In-Reply-To: <20211206145551.909846023@linuxfoundation.org>
From: Feng Tang <feng.tang@intel.com>
commit b50db7095fe002fa3e16605546cba66bf1b68a3e upstream.
There are cases that the TSC clocksource is wrongly judged as unstable by
the clocksource watchdog mechanism which tries to validate the TSC against
HPET, PM_TIMER or jiffies. While there is hardly a general reliable way to
check the validity of a watchdog, Thomas Gleixner proposed [1]:
"I'm inclined to lift that requirement when the CPU has:
1) X86_FEATURE_CONSTANT_TSC
2) X86_FEATURE_NONSTOP_TSC
3) X86_FEATURE_NONSTOP_TSC_S3
4) X86_FEATURE_TSC_ADJUST
5) At max. 4 sockets
After two decades of horrors we're finally at a point where TSC seems
to be halfway reliable and less abused by BIOS tinkerers. TSC_ADJUST
was really key as we can now detect even small modifications reliably
and the important point is that we can cure them as well (not pretty
but better than all other options)."
As feature #3 X86_FEATURE_NONSTOP_TSC_S3 only exists on several generations
of Atom processorz, and is always coupled with X86_FEATURE_CONSTANT_TSC
and X86_FEATURE_NONSTOP_TSC, skip checking it, and also be more defensive
to use maximal 2 sockets.
The check is done inside tsc_init() before registering 'tsc-early' and
'tsc' clocksources, as there were cases that both of them had been
wrongly judged as unreliable.
For more background of tsc/watchdog, there is a good summary in [2]
[tglx} Update vs. jiffies:
On systems where the only remaining clocksource aside of TSC is jiffies
there is no way to make this work because that creates a circular
dependency. Jiffies accuracy depends on not missing a periodic timer
interrupt, which is not guaranteed. That could be detected by TSC, but as
TSC is not trusted this cannot be compensated. The consequence is a
circulus vitiosus which results in shutting down TSC and falling back to
the jiffies clocksource which is even more unreliable.
[1]. https://lore.kernel.org/lkml/87eekfk8bd.fsf@nanos.tec.linutronix.de/
[2]. https://lore.kernel.org/lkml/87a6pimt1f.ffs@nanos.tec.linutronix.de/
[ tglx: Refine comment and amend changelog ]
Fixes: 6e3cd95234dc ("x86/hpet: Use another crystalball to evaluate HPET usability")
Suggested-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Feng Tang <feng.tang@intel.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: "Paul E. McKenney" <paulmck@kernel.org>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/20211117023751.24190-2-feng.tang@intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
arch/x86/kernel/tsc.c | 28 ++++++++++++++++++++++++----
1 file changed, 24 insertions(+), 4 deletions(-)
--- a/arch/x86/kernel/tsc.c
+++ b/arch/x86/kernel/tsc.c
@@ -1162,6 +1162,12 @@ void mark_tsc_unstable(char *reason)
EXPORT_SYMBOL_GPL(mark_tsc_unstable);
+static void __init tsc_disable_clocksource_watchdog(void)
+{
+ clocksource_tsc_early.flags &= ~CLOCK_SOURCE_MUST_VERIFY;
+ clocksource_tsc.flags &= ~CLOCK_SOURCE_MUST_VERIFY;
+}
+
static void __init check_system_tsc_reliable(void)
{
#if defined(CONFIG_MGEODEGX1) || defined(CONFIG_MGEODE_LX) || defined(CONFIG_X86_GENERIC)
@@ -1178,6 +1184,23 @@ static void __init check_system_tsc_reli
#endif
if (boot_cpu_has(X86_FEATURE_TSC_RELIABLE))
tsc_clocksource_reliable = 1;
+
+ /*
+ * Disable the clocksource watchdog when the system has:
+ * - TSC running at constant frequency
+ * - TSC which does not stop in C-States
+ * - the TSC_ADJUST register which allows to detect even minimal
+ * modifications
+ * - not more than two sockets. As the number of sockets cannot be
+ * evaluated at the early boot stage where this has to be
+ * invoked, check the number of online memory nodes as a
+ * fallback solution which is an reasonable estimate.
+ */
+ if (boot_cpu_has(X86_FEATURE_CONSTANT_TSC) &&
+ boot_cpu_has(X86_FEATURE_NONSTOP_TSC) &&
+ boot_cpu_has(X86_FEATURE_TSC_ADJUST) &&
+ nr_online_nodes <= 2)
+ tsc_disable_clocksource_watchdog();
}
/*
@@ -1369,9 +1392,6 @@ static int __init init_tsc_clocksource(v
if (tsc_unstable)
goto unreg;
- if (tsc_clocksource_reliable || no_tsc_watchdog)
- clocksource_tsc.flags &= ~CLOCK_SOURCE_MUST_VERIFY;
-
if (boot_cpu_has(X86_FEATURE_NONSTOP_TSC_S3))
clocksource_tsc.flags |= CLOCK_SOURCE_SUSPEND_NONSTOP;
@@ -1506,7 +1526,7 @@ void __init tsc_init(void)
}
if (tsc_clocksource_reliable || no_tsc_watchdog)
- clocksource_tsc_early.flags &= ~CLOCK_SOURCE_MUST_VERIFY;
+ tsc_disable_clocksource_watchdog();
clocksource_register_khz(&clocksource_tsc_early, tsc_khz);
detect_art();
next prev parent reply other threads:[~2021-12-06 15:24 UTC|newest]
Thread overview: 77+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-12-06 14:56 [PATCH 5.4 00/70] 5.4.164-rc1 review Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 01/70] NFSv42: Fix pagecache invalidation after COPY/CLONE Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 02/70] of: clk: Make <linux/of_clk.h> self-contained Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 03/70] arm64: dts: mcbin: support 2W SFP modules Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 04/70] can: j1939: j1939_tp_cmd_recv(): check the dst address of TP.CM_BAM Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 05/70] gfs2: Fix length of holes reported at end-of-file Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 06/70] drm/sun4i: fix unmet dependency on RESET_CONTROLLER for PHY_SUN6I_MIPI_DPHY Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 07/70] mac80211: do not access the IV when it was stripped Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 08/70] net/smc: Transfer remaining wait queue entries during fallback Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 09/70] atlantic: Fix OOB read and write in hw_atl_utils_fw_rpc_wait Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 10/70] net: return correct error code Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 11/70] platform/x86: thinkpad_acpi: Fix WWAN device disabled issue after S3 deep Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 12/70] s390/setup: avoid using memblock_enforce_memory_limit Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 13/70] btrfs: check-integrity: fix a warning on write caching disabled disk Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 14/70] thermal: core: Reset previous low and high trip during thermal zone init Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 15/70] scsi: iscsi: Unblock session then wake up error handler Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 16/70] ata: ahci: Add Green Sardine vendor ID as board_ahci_mobile Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 17/70] ethernet: hisilicon: hns: hns_dsaf_misc: fix a possible array overflow in hns_dsaf_ge_srst_by_port() Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 18/70] net: tulip: de4x5: fix the problem that the array lp->phy[8] may be out of bound Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 19/70] net: ethernet: dec: tulip: de4x5: fix possible array overflows in type3_infoblock() Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 20/70] perf hist: Fix memory leak of a perf_hpp_fmt Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 21/70] perf report: Fix memory leaks around perf_tip() Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 22/70] net/smc: Avoid warning of possible recursive locking Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 23/70] vrf: Reset IPCB/IP6CB when processing outbound pkts in vrf dev xmit Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 24/70] kprobes: Limit max data_size of the kretprobe instances Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 25/70] rt2x00: do not mark device gone on EPROTO errors during start Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 26/70] ipmi: Move remove_work to dedicated workqueue Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 27/70] cpufreq: Fix get_cpu_device() failure in add_cpu_dev_symlink() Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 28/70] s390/pci: move pseudo-MMIO to prevent MIO overlap Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 29/70] fget: check that the fd still exists after getting a ref to it Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 30/70] sata_fsl: fix UAF in sata_fsl_port_stop when rmmod sata_fsl Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 31/70] sata_fsl: fix warning in remove_proc_entry " Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 32/70] i2c: stm32f7: flush TX FIFO upon transfer errors Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 33/70] i2c: stm32f7: recover the bus on access timeout Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 34/70] i2c: stm32f7: stop dma transfer in case of NACK Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 35/70] i2c: cbus-gpio: set atomic transfer callback Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 36/70] natsemi: xtensa: fix section mismatch warnings Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 37/70] net: qlogic: qlcnic: Fix a NULL pointer dereference in qlcnic_83xx_add_rings() Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 38/70] net: mpls: Fix notifications when deleting a device Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 39/70] siphash: use _unaligned version by default Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 40/70] net/mlx4_en: Fix an use-after-free bug in mlx4_en_try_alloc_resources() Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 41/70] selftests: net: Correct case name Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 42/70] rxrpc: Fix rxrpc_local leak in rxrpc_lookup_peer() Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 43/70] net: usb: lan78xx: lan78xx_phy_init(): use PHY_POLL instead of "0" if no IRQ is available Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 44/70] net: marvell: mvpp2: Fix the computation of shared CPUs Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 45/70] net: annotate data-races on txq->xmit_lock_owner Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 46/70] ipv4: convert fib_num_tclassid_users to atomic_t Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 47/70] net/rds: correct socket tunable error in rds_tcp_tune() Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 48/70] net/smc: Keep smc_close_final rc during active close Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 49/70] drm/msm: Do hw_init() before capturing GPU state Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 50/70] ipv6: fix memory leak in fib6_rule_suppress Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 51/70] KVM: x86/pmu: Fix reserved bits for AMD PerfEvtSeln register Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 52/70] sched/uclamp: Fix rq->uclamp_max not set on first enqueue Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 53/70] parisc: Fix KBUILD_IMAGE for self-extracting kernel Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 54/70] parisc: Fix "make install" on newer debian releases Greg Kroah-Hartman
2021-12-06 14:56 ` [PATCH 5.4 55/70] vgacon: Propagate console boot parameters before calling `vc_resize Greg Kroah-Hartman
2021-12-06 14:57 ` [PATCH 5.4 56/70] xhci: Fix commad ring abort, write all 64 bits to CRCR register Greg Kroah-Hartman
2021-12-06 14:57 ` [PATCH 5.4 57/70] USB: NO_LPM quirk Lenovo Powered USB-C Travel Hub Greg Kroah-Hartman
2021-12-06 14:57 ` [PATCH 5.4 58/70] usb: typec: tcpm: Wait in SNK_DEBOUNCED until disconnect Greg Kroah-Hartman
2021-12-06 14:57 ` [PATCH 5.4 59/70] x86/tsc: Add a timer to make sure TSC_adjust is always checked Greg Kroah-Hartman
2021-12-06 14:57 ` Greg Kroah-Hartman [this message]
2021-12-06 14:57 ` [PATCH 5.4 61/70] x86/64/mm: Map all kernel memory into trampoline_pgd Greg Kroah-Hartman
2021-12-06 14:57 ` [PATCH 5.4 62/70] tty: serial: msm_serial: Deactivate RX DMA for polling support Greg Kroah-Hartman
2021-12-06 14:57 ` [PATCH 5.4 63/70] serial: pl011: Add ACPI SBSA UART match id Greg Kroah-Hartman
2021-12-06 14:57 ` [PATCH 5.4 64/70] serial: core: fix transmit-buffer reset and memleak Greg Kroah-Hartman
2021-12-06 14:57 ` [PATCH 5.4 65/70] serial: 8250_pci: Fix ACCES entries in pci_serial_quirks array Greg Kroah-Hartman
2021-12-06 14:57 ` [PATCH 5.4 66/70] serial: 8250_pci: rewrite pericom_do_set_divisor() Greg Kroah-Hartman
2021-12-06 14:57 ` [PATCH 5.4 67/70] iwlwifi: mvm: retry init flow if failed Greg Kroah-Hartman
2021-12-06 14:57 ` [PATCH 5.4 68/70] parisc: Mark cr16 CPU clocksource unstable on all SMP machines Greg Kroah-Hartman
2021-12-06 14:57 ` [PATCH 5.4 69/70] net/tls: Fix authentication failure in CCM mode Greg Kroah-Hartman
2021-12-06 14:57 ` [PATCH 5.4 70/70] ipmi: msghandler: Make symbol remove_work_wq static Greg Kroah-Hartman
2021-12-06 18:57 ` [PATCH 5.4 00/70] 5.4.164-rc1 review Florian Fainelli
2021-12-06 21:57 ` Shuah Khan
2021-12-07 2:19 ` Samuel Zou
2021-12-07 9:16 ` Naresh Kamboju
2021-12-07 20:41 ` Guenter Roeck
2021-12-08 10:29 ` Sudip Mukherjee
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=20211206145553.994509372@linuxfoundation.org \
--to=gregkh@linuxfoundation.org \
--cc=feng.tang@intel.com \
--cc=linux-kernel@vger.kernel.org \
--cc=paulmck@kernel.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