Date:   Wed, 23 Nov 2022 06:44:15 -0800
From:   "Paul E. McKenney" <paulmck@...nel.org>
To:     Mirsad Goran Todorovac <mirsad.todorovac@....unizg.hr>
Cc:     Phillip Lougher <phillip@...ashfs.org.uk>,
        LKML <linux-kernel@...r.kernel.org>, phillip.lougher@...il.com,
        Thorsten Leemhuis <regressions@...mhuis.info>, elliott@....com
Subject: Re: BUG: BISECTED: in squashfs_xz_uncompress() (Was: RCU stalls in
 squashfs_readahead())

On Wed, Nov 23, 2022 at 10:14:07AM +0100, Mirsad Goran Todorovac wrote:
> On 22. 11. 2022. 03:07, Paul E. McKenney wrote:
> 
> > > I'm afraid that I would lose in Far Cry miserably if my cores
> > > decided to all lock up for 21 secs. :-(
> > 
> > Agreed, 21 seconds is an improvement over the earlier 60 seconds, but
> > still a very long time.  Me, I come from DYNIX/ptx, where the equivalent
> > to the RCU CPU stall warning was 1.5 seconds.  On the other hand, it
> > is also the case that DYNIX/ptx had nowhere near the variety of drivers
> > and subsystems, nor did it scale anywhere near as far as Linux does today.
> > 
> > But you only need one CPU to lock up for 21 seconds to get an RCU CPU
> > stall warning, not all of them.  ;-)
> 
> I can recall an occasion or a couple of them where the entire X Window system
> had been unresponsive for quite a number of seconds that sometimes made me reset
> the Ubuntu box.
> 
> I have the good news: the patches did not apply because they were already applied
> in the mainline tree:
> 
> mtodorov@...ac:~/linux/kernel/linux_stable_build_b$ git log | grep -C5 28b3ae426598
>     Signed-off-by: Borislav Petkov <bp@...e.de>
>     Reviewed-by: Kees Cook <keescook@...omium.org>
>     Acked-by: Florian Weimer <fweimer@...hat.com>
>     Link: https://lore.kernel.org/r/898932fe61db6a9d61bc2458fa2f6049f1ca9f5c.1652290558.git.luto@kernel.org
> 
> commit 28b3ae426598e722cf5d5ab9cc7038791b955a56
> Author: Uladzislau Rezki <uladzislau.rezki@...y.com>
> Date:   Wed Feb 16 14:52:09 2022 +0100
> 
>     rcu: Introduce CONFIG_RCU_EXP_CPU_STALL_TIMEOUT
> 
> mtodorov@...ac:~/linux/kernel/linux_stable_build_b$ git log | grep -C5 1045a06724f3
>     Somehow kernel-doc complains here about strong markup, but
>     we really don't need the [] so just remove that.
> 
>     Signed-off-by: Johannes Berg <johannes.berg@...el.com>
> 
> commit 1045a06724f322ed61f1ffb994427c7bdbe64647
> Author: Christoph Hellwig <hch@....de>
> Date:   Wed Jun 29 17:01:02 2022 +0200
> 
>     remove CONFIG_ANDROID
> 
> mtodorov@...ac:~/linux/kernel/linux_stable_build_b$

OK, very good.

But your CONFIG_RCU_EXP_CPU_STALL_TIMEOUT seems to be about 20.
As in 20 milliseconds.

Is that intentional?

> > > This is at present just the wishful thinking, as I lack your 30 years of
> > > experience with the kernel and RCU update system. I am only beginning to realise
> > > why it is more efficient than the traditional locking, and IMHO it should
> > > avoid locking up cores instead of increasing the number of complaints.
> > 
> > Just to set the record straight, RCU does not normally lock up any of
> > the cores.  Instead, RCU detects that cores have been locked up.
> > 
> > Give or take the occasional bug in RCU, of course!
> 
> Currently, I cannot be the judge of that, for I can't seem to understand how the
> magic of RCU works., how it is implemented. There's more homework to be done ;-)

The high-level specification is dead simple.  Calls synchronize_rcu()
must wait for any CPU or task that has already executed rcu_read_lock()
to reach the corresponding rcu_read_unlock().  For more requirements:

Documentation/RCU/Design/Requirements/Requirements.rst

> > > But even if the Linux kernel source is magically "memory mapped" into my
> > > mind, I still do not see how it could be done. My Linux kernel learning curve
> > > had not yet got that up, but I have no doubts that it is designed by
> > > Intelligent Designers who are very witty people, and not village idiots ;-)
> > 
> > There is the school of thought that claims that the Linux kernel is
> > driven by evolutionary forces rather than intelligent design.  And as
> > we all know, evolutionary forces are driven by random changes, which
> > absolutely anyone could make.
> 
> Give or take the rate of improbability where a bunch of monkeys randomly typing
> would produce a working Linux kernel source would be about a couple of working
> sources in a space of 96^30,000,00 (something like 10^300,000,000), it is comparable
> to the probability of random coming of the first simplest DNA into the existence
> from the amino acid primordial soup.
> 
> (Not that many atoms in the Universe - 10^82, you'd need an awful lot of wasted
> multiverses with no even single cell life and certainly no working Linux kernels.)

Just apply continuous integration to each monkey's output.  Just like
was done to the strands of amino acids.

(Sorry, couldn't resist!)

> > And one approach is to take a less aggressive RCU CPU stall timeout,
> > say reducing from 21 seconds to (say) 15 seconds instead of all the
> > way down to 20 milliseconds.  This could allow you to ease into the
> > latency-reduction work.
> > 
> > Alternatively, consider that response time is a property of the
> > entire system plus the environment that it runs in.  So I suspect that
> > the Android folks are accompanying that 20-millisecond timeout with
> > some restrictions on what the on-phone workloads are permitted to do.
> > Maybe ask the Android guys what those restrictions are and loosen them
> > slightly, again allowing you to ease into the latency-reduction work.
> 
> Good point.
> > Sometimes an NMI does get the CPUs back on track.  Sometimes the RCU CPU
> > stall warning is a symptom of the CPU having gotten too old and failing.
> > Most often, though, it is a sign of some sort of lockup, a too-long
> > RCU read-side critical section, or as Robert Elliot noted, the lack of
> > a cond_resched().
> > 
> > But please keep in mind that cond_resched() helps only in kernels built
> > with CONFIG_PREEMPTION=n.
> 
> I have bad news that 6.1-r6 is still affected with squashfs_xz_uncompress bug, despite having both of your fixes
> (as visible in above command's output -- double checked):
> 
> [   91.065659] rcu: INFO: rcu_preempt detected expedited stalls on CPUs/tasks: { 3-.... } 6 jiffies s: 621 root: 0x8/.

If you build with (say) CONFIG_RCU_EXP_CPU_STALL_TIMEOUT=200, does
this still happen?

For that 

> [   91.065694] rcu: blocking rcu_node structures (internal RCU debug):
> [   91.065704] Sending NMI from CPU 5 to CPUs 3:
> [   91.065721] NMI backtrace for cpu 3
> [   91.065730] CPU: 3 PID: 2829 Comm: snap-store Not tainted 6.1.0-rc6 #1
> [   91.065741] Hardware name: LENOVO 82H8/LNVNB161216, BIOS GGCN49WW 07/21/2022
> [   91.065746] RIP: 0010:__asan_load4+0x0/0xa0
> [   91.065764] Code: 9e c0 84 c0 75 e1 5d c3 cc cc cc cc 48 c1 e8 03 80 3c
> 10 00 75 e9 5d c3 cc cc cc cc 66 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 00 <55>
> 48 89 e5 48 8b 4d 08 48 83 ff fb 77 64 eb 0f 0f 1f 00 48 b8 00
> [   91.065771] RSP: 0000:ffff8881388ef140 EFLAGS: 00000246
> [   91.065779] RAX: 0000000000000000 RBX: 0000000000000003 RCX: ffffffff9be992fd
> [   91.065785] RDX: 0000000000000003 RSI: dffffc0000000000 RDI: ffff888125500004
> [   91.065789] RBP: ffff8881388ef1e0 R08: 0000000000000001 R09: ffffed1024aa0de8
> [   91.065794] R10: ffff888125506f39 R11: ffffed1024aa0de7 R12: 0000000001067db0
> [   91.065799] R13: ffff888125500000 R14: 00000000014fe803 R15: ffff888125502112
> [   91.065804] FS:  00007fdec50ab180(0000) GS:ffff888257180000(0000) knlGS:0000000000000000
> [   91.065810] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> [   91.065815] CR2: 00007fdeb7cb6260 CR3: 000000011a436005 CR4: 0000000000770ee0
> [   91.065820] PKRU: 55555554
> [   91.065823] Call Trace:
> [   91.065826]  <TASK>
> [   91.065829]  ? lzma_main+0x37a/0x1260
> [   91.065845]  lzma2_lzma+0x2b9/0x430
> [   91.065857]  xz_dec_lzma2_run+0x11f/0xb90
> [   91.065867]  ? __asan_load4+0x55/0xa0
> [   91.065880]  xz_dec_run+0x346/0x11f0
> [   91.065892]  squashfs_xz_uncompress+0x196/0x370
> [   91.065905]  ? lzo_uncompress+0x400/0x400
> [   91.065913]  squashfs_decompress+0x88/0xd0
> [   91.065923]  squashfs_read_data+0x1e5/0x900
> [   91.065930]  ? __create_object+0x4ae/0x560
> [   91.065942]  ? squashfs_bio_read.isra.3+0x230/0x230
> [   91.065951]  ? __kasan_kmalloc+0xb6/0xc0
> [   91.065961]  ? squashfs_page_actor_init_special+0x1a6/0x210
> [   91.065972]  squashfs_readahead+0xaa3/0xe80
> [   91.065985]  ? squashfs_fill_page+0x190/0x190
> [   91.065993]  ? __filemap_add_folio+0x3a1/0x680
> [   91.066003]  ? dio_warn_stale_pagecache.part.67+0x90/0x90
> [   91.066012]  read_pages+0x122/0x540
> [   91.066023]  ? file_ra_state_init+0x60/0x60
> [   91.066032]  ? filemap_add_folio+0xd4/0x140
> [   91.066040]  ? folio_alloc+0x1b/0x50
> [   91.066051]  page_cache_ra_unbounded+0x1e6/0x280
> [   91.066064]  do_page_cache_ra+0x7c/0x90
> [   91.066074]  page_cache_ra_order+0x393/0x400
> [   91.066087]  ondemand_readahead+0x2f1/0x4e0
> [   91.066098]  page_cache_async_ra+0x8b/0xa0
> [   91.066106]  filemap_fault+0x742/0x1490
> [   91.066113]  ? __folio_memcg_unlock+0x35/0x80
> [   91.066124]  ? read_cache_page_gfp+0x90/0x90
> [   91.066132]  ? filemap_map_pages+0x28e/0xc60
> [   91.066145]  __do_fault+0x76/0x1b0
> [   91.066154]  do_fault+0x1c6/0x680
> [   91.066163]  __handle_mm_fault+0x89a/0x1310
> [   91.066173]  ? copy_page_range+0x1b20/0x1b20
> [   91.066181]  ? mt_find+0x189/0x330
> [   91.066190]  ? mas_next_entry+0xa80/0xa80
> [   91.066204]  handle_mm_fault+0x11b/0x390
> [   91.066213]  do_user_addr_fault+0x258/0x860
> [   91.066225]  exc_page_fault+0x64/0xf0
> [   91.066235]  asm_exc_page_fault+0x27/0x30
> [   91.066245] RIP: 0033:0x7fdeb7a1e541
> [   91.066252] Code: 11 44 b8 e0 0f 11 44 b8 f0 0f 11 04 b8 48 83 c7 40 48
> 83 c6 f8 75 92 48 85 d2 74 2d 4c 01 d7 49 8d 04 b9 48 83 c0 10 48 f7 da <0f>
> 28 05 18 7d 29 00 0f 1f 84 00 00 00 00 00 0f 11 40 f0 0f 11 00
> [   91.066259] RSP: 002b:00007fff46f77b60 EFLAGS: 00010293
> [   91.066265] RAX: 000055fbe1a57c30 RBX: 00007fff46f77d18 RCX: 000055fbe1a57c20
> [   91.066270] RDX: fffffffffffffffb RSI: 0000000000000005 RDI: 0000000000000000
> [   91.066274] RBP: 000055fbe16563b0 R08: 0000000000000028 R09: 000055fbe1a57c20
> [   91.066279] R10: 0000000000000000 R11: 0000000000000000 R12: 000000000000002b
> [   91.066283] R13: 000055fbe1656408 R14: 00007fff46f77b80 R15: 0000000000000000
> [   91.066292]  </TASK>
> 
> (This is apparently only visible in CONFIG_KASAN=y build.)

And KASAN does add some overhead.

So maybe the value of CONFIG_RCU_EXP_CPU_STALL_TIMEOUT needs to be a
function of CONFIG_KASAN in your build.

> > > Yes, you guys do an amasing job of keeping 30 million lines of code organised
> > > and making some sense. I will cut the smalltalk as I know you are a busy man.
> > > If I make a progress to actually produce any patches fixing these lockups and
> > > stalls, I will be sure to include you into CC: as you requested.
> > 
> > Looking forward to seeing what you come up with!
> 
> There will have to be a lot of homework to catch up with to before I'd be able to do anything
> sensible. :)

Doing the homework up front can be a good thing!  ;-)

							Thanx, Paul

> Thanks,
> Mirsad
> 
> --
> Mirsad Goran Todorovac
> Sistem inženjer
> Grafički fakultet | Akademija likovnih umjetnosti
> Sveučilište u Zagrebu
> -- 
> System engineer
> Faculty of Graphic Arts | Academy of Fine Arts
> University of Zagreb, Republic of Croatia
> The European Union
> 

Powered by blists - more mailing lists