linux-kernel - Re: [PATCH v2 05/15] ia64: convert to legacy_timer

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAK8P3a0s5Wt=9sZXAV1JzebZFXVy4qn3KNvygEfPeH3eVMyO_w@mail.gmail.com>
Date:   Fri, 18 Dec 2020 23:07:26 +0100
From:   Arnd Bergmann <arnd@...nel.org>
To:     John Paul Adrian Glaubitz <glaubitz@...sik.fu-berlin.de>
Cc:     "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        Thomas Gleixner <tglx@...utronix.de>,
        Arnd Bergmann <arnd@...db.de>,
        Russell King <linux@...linux.org.uk>,
        Tony Luck <tony.luck@...el.com>,
        Fenghua Yu <fenghua.yu@...el.com>,
        Greg Ungerer <gerg@...ux-m68k.org>,
        Finn Thain <fthain@...egraphics.com.au>,
        Philip Blundell <philb@....org>,
        Joshua Thompson <funaho@...ai.org>,
        Sam Creasey <sammy@...my.net>,
        "James E.J. Bottomley" <James.Bottomley@...senpartnership.com>,
        Helge Deller <deller@....de>,
        Daniel Lezcano <daniel.lezcano@...aro.org>,
        John Stultz <john.stultz@...aro.org>,
        Stephen Boyd <sboyd@...nel.org>,
        Linus Walleij <linus.walleij@...aro.org>,
        linux-ia64@...r.kernel.org,
        Parisc List <linux-parisc@...r.kernel.org>,
        linux-m68k <linux-m68k@...ts.linux-m68k.org>,
        Linux ARM <linux-arm-kernel@...ts.infradead.org>,
        Mike Rapoport <rppt@...ux.ibm.com>,
        Anatoly Pugachev <matorola@...il.com>
Subject: Re: [PATCH v2 05/15] ia64: convert to legacy_timer_tick

On Fri, Dec 18, 2020 at 10:29 PM John Paul Adrian Glaubitz
<glaubitz@...sik.fu-berlin.de> wrote:
>
> Hi Arnd!
>
> On 10/30/20 4:17 PM, Arnd Bergmann wrote:
> > ia64 is the only architecture that calls xtime_update() in a loop,
> > once for each jiffie that has passed since the last event.
> >
> > Before commit 3171a0305d62 ("[PATCH] simplify update_times (avoid
> > jiffies/jiffies_64 aliasing problem)") in 2006, it could not actually do
> > this any differently, but now it seems simpler to just pass the number
> > of jiffies that passed in the meantime.
> >
> > While this loses the ability process interrupts in the middle of
> > the timer tick by calling local_irq_enable(), doing so is fairly
> > peculiar anyway and it seems better to just do what everyone
> > else does here.
>
> This broke the hpsa block device driver on my HP RX2600.

Hi Adrian,

Sorry for causing this bug, and thank you for bisecting it
down to my patch.

Do you see any other strange behavior with that patch, or is
this the only symptom you are aware of?

>
> I'm seeing this backtrace now:
>
> [  905.883273] usb 1-2: SerialNumber: A60020000001
> [  905.918170]  sda: sda1 sda2 sda3
> [  905.920107] sd 0:1:0:0: [sda] Attached SCSI disk
> [  905.944102] usb-storage 1-2:1.0: USB Mass Storage device detected
> [  905.944102] scsi host1: usb-storage 1-2:1.0
> [  905.944102] usbcore: registered new interface driver usb-storage
> [  905.944117] usbcore: registered new interface driver uas

The timestamps show that time is moving forward, which is at least
something. Do you have a feeling for whether the timestamps are
counting in (roughly) the correct speed, or is it going much faster
or slower than it should?

To clarify: the [905.944117] numbers are seconds/microseconds
since boot, so message would be 906 seconds after the kernel
started.

> Begin: Loading essential drivers ... done.                                                                                                                        > Begin: Running /scripts/init-premount ... done.                                                                                                                 > Begin: Mounting root file system ... Begin: Running /scripts/local-top ... done.

Ok, so it gets into user space. Is this initramfs or the actual read-only root?

> [  906.666923] hpsa 0000:05:00.0: scsi 0:1:0:0: resetting logical  Direct-Access     HP       LOGICAL VOLUME   RAID-0 SSDSmartPathCap- En- Exp=1
> [  906.670923] hpsa 0000:05:00.0: device is ready.
> [  906.670923] hpsa 0000:05:00.0: scsi 0:1:0:0: reset logical  completed successfully Direct-Access     HP       LOGICAL VOLUME   RAID-0 SSDSmartPathCap- En- Exp=1
> done.
> [  906.722166] rcu: INFO: rcu_sched detected stalls on CPUs/tasks:
> [  906.722166] rcu:     2-....: (3 ticks this GP) idle=fe6/1/0x4000000000000000 softirq=693/698 fqs=4
> [  906.722166]  (detected by 0, t=6115 jiffies, g=465, q=80)
This appears to be an 'rcu stall' warning, from print_cpu_stall_info(),
indicating that timer ticks are missing.

> [  909.360108] INFO: task systemd-sysv-ge:200 blocked for more than 127 seconds.
> [  909.360108]       Not tainted 5.10.0+ #130
> [  909.360108] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
> [  909.360108] task:systemd-sysv-ge state:D stack:    0 pid:  200 ppid:   189 flags:0x00000000
> [  909.364108]
> [  909.364108] Call Trace:
> [  909.364423]  [<a00000010109b210>] __schedule+0x890/0x21e0
> [  909.364423]                                 sp=e0000100487d7b70 bsp=e0000100487d1748
> [  909.368423]  [<a00000010109cc00>] schedule+0xa0/0x240
> [  909.368423]                                 sp=e0000100487d7b90 bsp=e0000100487d16e0
> [  909.368558]  [<a00000010109ce70>] io_schedule+0x70/0xa0
> [  909.368558]                                 sp=e0000100487d7b90 bsp=e0000100487d16c0
> [  909.372290]  [<a00000010109e1c0>] bit_wait_io+0x20/0xe0
> [  909.372290]                                 sp=e0000100487d7b90 bsp=e0000100487d1698
> [  909.374168] rcu: INFO: rcu_sched detected stalls on CPUs/tasks:
> [  909.376290]  [<a00000010109d860>] __wait_on_bit+0xc0/0x1c0
> [  909.376290]                                 sp=e0000100487d7b90 bsp=e0000100487d1648
> [  909.374168] rcu:     3-....: (2 ticks this GP) idle=19e/1/0x4000000000000002 softirq=1581/1581 fqs=2
> [  909.374168]  (detected by 0, t=5661 jiffies, g=1089, q=3)
> [  909.376290]  [<a00000010109da80>] out_of_line_wait_on_bit+0x120/0x140
> [  909.376290]                                 sp=e0000100487d7b90 bsp=e0000100487d1610
> [  909.374168] Task dump for CPU 3:
> [  909.374168] task:khungtaskd      state:R  running task

and this seems to be another instance of the same. I would assume that this
is completely unrelated to the block driver and just happened to trigger during
the same time the driver was doing something.

Can you see in your full logs if the "Oops: timer tick before it's due" warning
triggered at any point?

I've attached a patch for a partial revert of my original change, this
should still work with the final cleanup on top, but restore the loop
plus the local_irq_enable()/local_irq_disable() that I dropped from
the original code. Does this make a difference?

        Arnd

View attachment "ia64-timer-partial-revert.diff" of type "text/x-patch" (1319 bytes)