From: Petr Mladek <pmladek@suse.com>
To: "Luis R. Rodriguez" <mcgrof@kernel.org>
Cc: shuah@kernel.org, jeyu@redhat.com, rusty@rustcorp.com.au,
ebiederm@xmission.com, dmitry.torokhov@gmail.com,
acme@redhat.com, corbet@lwn.net, martin.wilck@suse.com,
mmarek@suse.com, hare@suse.com, rwright@hpe.com, jeffm@suse.com,
DSterba@suse.com, fdmanana@suse.com, neilb@suse.com,
linux@roeck-us.net, rgoldwyn@suse.com, subashab@codeaurora.org,
xypron.glpk@gmx.de, keescook@chromium.org, atomlin@redhat.com,
mbenes@suse.cz, paulmck@linux.vnet.ibm.com,
dan.j.williams@intel.com, jpoimboe@redhat.com,
davem@davemloft.net, mingo@redhat.com, akpm@linux-foundation.org,
torvalds@linux-foundation.org, linux-kselftest@vger.kernel.org,
linux-doc@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [RFC 03/10] kmod: add dynamic max concurrent thread count
Date: Wed, 14 Dec 2016 16:38:27 +0100 [thread overview]
Message-ID: <20161214153827.GH16064@pathway.suse.cz> (raw)
In-Reply-To: <20161208194814.2485-1-mcgrof@kernel.org>
On Thu 2016-12-08 11:48:14, Luis R. Rodriguez wrote:
> We currently statically limit the number of modprobe threads which
> we allow to run concurrently to 50. As per Keith Owens, this was a
> completely arbitrary value, and it was set in the 2.3.38 days [0]
> over 16 years ago in year 2000.
>
> Although we haven't yet hit our lower limits, experimentation [1]
> shows that when and if we hit this limit in the worst case, will be
> fatal -- consider get_fs_type() failures upon mount on a system which
> has many partitions, some of which might even be with the same
> filesystem. Its best to be prudent and increase and set this
> value to something more sensible which ensures we're far from hitting
> the limit and also allows default build/user run time override.
>
> The worst case is fatal given that once a module fails to load there
> is a period of time during which subsequent request for the same module
> will fail, so in the case of partitions its not just one request that
> could fail, but whole series of partitions. This later issue of a
> module request failure domino effect can be addressed later, but
> increasing the limit to something more meaninful should at least give us
> enough cushion to avoid this for a while.
>
> Set this value up with a bit more meaninful modern limits:
>
> Bump this up to 64 max for small systems (CONFIG_BASE_SMALL)
> Bump this up to 128 max for larger systems (!CONFIG_BASE_SMALL)
>
> diff --git a/init/Kconfig b/init/Kconfig
> index 271692a352f1..da2c25746937 100644
> --- a/init/Kconfig
> +++ b/init/Kconfig
> @@ -2111,6 +2111,29 @@ config TRIM_UNUSED_KSYMS
>
> If unsure, or if you need to build out-of-tree modules, say N.
>
> +config MAX_KMOD_CONCURRENT
> + int "Max allowed concurrent request_module() calls (6=>64, 10=>1024)"
> + range 0 14
Would not too small range break loading module dependencies?
I am not sure how it is implemented but it might require having
some more module loads in progress.
I would give 6 as minimum. Nobody has troubles with the current limit.
> + default 6 if !BASE_SMALL
> + default 7 if BASE_SMALL
Aren't the conditions inversed?
> diff --git a/init/main.c b/init/main.c
> index 8161208d4ece..1fa441aa32c6 100644
> --- a/init/main.c
> +++ b/init/main.c
> @@ -638,6 +638,7 @@ asmlinkage __visible void __init start_kernel(void)
> thread_stack_cache_init();
> cred_init();
> fork_init();
> + init_kmod_umh();
> proc_caches_init();
> buffer_init();
> key_init();
> diff --git a/kernel/kmod.c b/kernel/kmod.c
> index 0277d1216f80..cb6f7ca7b8a5 100644
> --- a/kernel/kmod.c
> +++ b/kernel/kmod.c
> @@ -186,6 +174,31 @@ int __request_module(bool wait, const char *fmt, ...)
> return ret;
> }
> EXPORT_SYMBOL(__request_module);
> +
> +/*
> + * If modprobe needs a service that is in a module, we get a recursive
> + * loop. Limit the number of running kmod threads to max_threads/2 or
> + * CONFIG_MAX_KMOD_CONCURRENT, whichever is the smaller. A cleaner method
> + * would be to run the parents of this process, counting how many times
> + * kmod was invoked. That would mean accessing the internals of the
> + * process tables to get the command line, proc_pid_cmdline is static
> + * and it is not worth changing the proc code just to handle this case.
> + *
> + * "trace the ppid" is simple, but will fail if someone's
> + * parent exits. I think this is as good as it gets.
> + *
> + * You can override with with a kernel parameter, for instance to allow
> + * 4096 concurrent modprobe instances:
> + *
> + * kmod.max_modprobes=4096
> + */
> +void __init init_kmod_umh(void)
> +{
> + if (!max_modprobes)
> + max_modprobes = min(max_threads/2,
> + 2 << CONFIG_MAX_KMOD_CONCURRENT);
This should be
1 << CONFIG_MAX_KMOD_CONCURRENT);
1 << 1 = 2;
Note that this calculation is mentioned also some comments and
documentation.
Best Regards,
Petr
next prev parent reply other threads:[~2016-12-14 15:54 UTC|newest]
Thread overview: 65+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-12-08 18:47 [RFC 00/10] kmod: stress test driver, few fixes and enhancements Luis R. Rodriguez
2016-12-08 18:47 ` [RFC 01/10] kmod: add test driver to stress test the module loader Luis R. Rodriguez
2016-12-08 20:24 ` Kees Cook
2016-12-13 21:10 ` Luis R. Rodriguez
2016-12-16 7:41 ` Luis R. Rodriguez
2016-12-08 19:48 ` [RFC 02/10] module: fix memory leak on early load_module() failures Luis R. Rodriguez
2016-12-08 20:30 ` Kees Cook
2016-12-08 21:10 ` Luis R. Rodriguez
2016-12-08 21:17 ` Kees Cook
2016-12-09 17:06 ` Miroslav Benes
2016-12-16 8:51 ` Luis R. Rodriguez
2016-12-15 18:46 ` Aaron Tomlin
2016-12-08 19:48 ` [RFC 03/10] kmod: add dynamic max concurrent thread count Luis R. Rodriguez
2016-12-08 20:28 ` Kees Cook
2016-12-08 21:00 ` Luis R. Rodriguez
2016-12-14 15:38 ` Petr Mladek [this message]
2016-12-16 8:39 ` Luis R. Rodriguez
2017-01-10 19:24 ` Luis R. Rodriguez
2016-12-08 19:48 ` [RFC 04/10] kmod: provide wrappers for kmod_concurrent inc/dec Luis R. Rodriguez
2016-12-08 20:29 ` Kees Cook
2016-12-08 21:08 ` Luis R. Rodriguez
2016-12-15 12:46 ` Petr Mladek
2016-12-16 8:05 ` Luis R. Rodriguez
2016-12-22 4:48 ` Jessica Yu
2017-01-06 20:54 ` Luis R. Rodriguez
2017-01-10 18:57 ` [RFC 04/10] " Luis R. Rodriguez
2017-01-11 20:08 ` Luis R. Rodriguez
2017-05-16 18:02 ` Luis R. Rodriguez
2017-05-18 2:37 ` Luis R. Rodriguez
2016-12-22 5:07 ` Jessica Yu
2017-01-10 20:28 ` Luis R. Rodriguez
2016-12-08 19:48 ` [RFC 05/10] kmod: return -EBUSY if modprobe limit is reached Luis R. Rodriguez
2016-12-08 19:48 ` [RFC 06/10] kmod: provide sanity check on kmod_concurrent access Luis R. Rodriguez
2016-12-14 16:08 ` Petr Mladek
2016-12-14 17:12 ` Luis R. Rodriguez
2016-12-15 12:57 ` Petr Mladek
2017-01-10 20:00 ` Luis R. Rodriguez
2016-12-08 19:49 ` [RFC 07/10] kmod: use simplified rate limit printk Luis R. Rodriguez
2016-12-14 16:23 ` Petr Mladek
2016-12-14 16:41 ` Joe Perches
2016-12-16 8:44 ` Luis R. Rodriguez
2016-12-08 19:49 ` [RFC 08/10] sysctl: add support for unsigned int properly Luis R. Rodriguez
2016-12-08 19:49 ` [RFC 09/10] kmod: add helpers for getting kmod count and limit Luis R. Rodriguez
2016-12-15 16:56 ` Petr Mladek
2016-12-16 7:57 ` Luis R. Rodriguez
2017-01-11 18:27 ` Luis R. Rodriguez
2016-12-08 19:49 ` [RFC 10/10] kmod: add a sanity check on module loading Luis R. Rodriguez
2016-12-09 20:03 ` Martin Wilck
2016-12-09 20:56 ` Linus Torvalds
2016-12-15 18:08 ` Luis R. Rodriguez
2016-12-15 0:27 ` Rusty Russell
2016-12-16 8:31 ` Luis R. Rodriguez
2016-12-17 3:54 ` Rusty Russell
[not found] ` <CAB=NE6VvuA9a6hf6yoopGfUxVJQM5HyV5bNzUdsEtUV0UhbG-g@mail.gmail.com>
2016-12-20 0:53 ` Rusty Russell
2016-12-20 18:52 ` Luis R. Rodriguez
2016-12-21 2:21 ` Rusty Russell
2016-12-21 13:08 ` Luis R. Rodriguez
2017-01-03 0:04 ` Rusty Russell
2017-01-06 20:36 ` Luis R. Rodriguez
2017-01-06 21:53 ` Jessica Yu
2017-01-09 20:27 ` Luis R. Rodriguez
[not found] ` <87bmvgax51.fsf@rustcorp.com.au>
2017-01-09 19:56 ` [RFC 10/10] " Luis R. Rodriguez
2017-01-06 21:03 ` Jessica Yu
2017-01-04 2:47 ` Jessica Yu
2017-01-11 19:10 ` [RFC 00/10] kmod: stress test driver, few fixes and enhancements Luis R. Rodriguez
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=20161214153827.GH16064@pathway.suse.cz \
--to=pmladek@suse.com \
--cc=DSterba@suse.com \
--cc=acme@redhat.com \
--cc=akpm@linux-foundation.org \
--cc=atomlin@redhat.com \
--cc=corbet@lwn.net \
--cc=dan.j.williams@intel.com \
--cc=davem@davemloft.net \
--cc=dmitry.torokhov@gmail.com \
--cc=ebiederm@xmission.com \
--cc=fdmanana@suse.com \
--cc=hare@suse.com \
--cc=jeffm@suse.com \
--cc=jeyu@redhat.com \
--cc=jpoimboe@redhat.com \
--cc=keescook@chromium.org \
--cc=linux-doc@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-kselftest@vger.kernel.org \
--cc=linux@roeck-us.net \
--cc=martin.wilck@suse.com \
--cc=mbenes@suse.cz \
--cc=mcgrof@kernel.org \
--cc=mingo@redhat.com \
--cc=mmarek@suse.com \
--cc=neilb@suse.com \
--cc=paulmck@linux.vnet.ibm.com \
--cc=rgoldwyn@suse.com \
--cc=rusty@rustcorp.com.au \
--cc=rwright@hpe.com \
--cc=shuah@kernel.org \
--cc=subashab@codeaurora.org \
--cc=torvalds@linux-foundation.org \
--cc=xypron.glpk@gmx.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).