Re: [PATCH 1/1] mm: sysctl: add panic_on_mm_error sysctl

[Aaron Tomlin <atomlin@redhat.com>]

On Mon 2020-01-27 11:11 +0100, Grzegorz Halat wrote:
> Memory management subsystem performs various checks at runtime,
> if an inconsistency is detected then such event is being logged and kernel
> continues to run. While debugging such problems it is helpful to collect
> memory dump as early as possible. Currently, there is no easy way to panic
> kernel when such error is detected.
> 
> It was proposed[1] to panic the kernel if panic_on_oops is set but this
> approach was not accepted. One of alternative proposals was introduction of
> a new sysctl.
> 
> The patch adds panic_on_mm_error sysctl. If the sysctl is set then the
> kernel will be crashed when an inconsistency is detected by memory
> management. This currently means panic when bad page or bad PTE
> is detected(this may be extended to other places in MM).
> 
> Another use case of this sysctl may be in security-wise environments,
> it may be more desired to crash machine than continue to run with
> potentially damaged data structures.
> 
> [1] https://marc.info/?l=linux-mm&m=142649500728327&w=2
> 
> Signed-off-by: Grzegorz Halat <ghalat@redhat.com>
> ---
>  Documentation/admin-guide/sysctl/kernel.rst | 12 ++++++++++++
>  include/linux/kernel.h                      |  1 +
>  kernel/sysctl.c                             |  9 +++++++++
>  mm/memory.c                                 |  7 +++++++
>  mm/page_alloc.c                             |  4 +++-
>  5 files changed, 32 insertions(+), 1 deletion(-)
> 
> diff --git a/Documentation/admin-guide/sysctl/kernel.rst b/Documentation/admin-guide/sysctl/kernel.rst
> index def074807cee..2fecd6b2547e 100644
> --- a/Documentation/admin-guide/sysctl/kernel.rst
> +++ b/Documentation/admin-guide/sysctl/kernel.rst
> @@ -61,6 +61,7 @@ show up in /proc/sys/kernel:
>  - overflowgid
>  - overflowuid
>  - panic
> +- panic_on_mm_error
>  - panic_on_oops
>  - panic_on_stackoverflow
>  - panic_on_unrecovered_nmi
> @@ -611,6 +612,17 @@ an IO error.
>     and you can use this option to take a crash dump.
>  
>  
> +panic_on_mm_error:
> +==============
> +
> +Controls the kernel's behaviour when inconsistency is detected
> +by memory management code, for example bad page state or bad PTE.
> +
> +0: try to continue operation.
> +
> +1: panic immediately.
> +
> +
>  panic_on_oops:
>  ==============
>  
> diff --git a/include/linux/kernel.h b/include/linux/kernel.h
> index 0d9db2a14f44..5f9d408512ff 100644
> --- a/include/linux/kernel.h
> +++ b/include/linux/kernel.h
> @@ -518,6 +518,7 @@ extern int oops_in_progress;		/* If set, an oops, panic(), BUG() or die() is in
>  extern int panic_timeout;
>  extern unsigned long panic_print;
>  extern int panic_on_oops;
> +extern int panic_on_mm_error;
>  extern int panic_on_unrecovered_nmi;
>  extern int panic_on_io_nmi;
>  extern int panic_on_warn;
> diff --git a/kernel/sysctl.c b/kernel/sysctl.c
> index 70665934d53e..6477e1cce28b 100644
> --- a/kernel/sysctl.c
> +++ b/kernel/sysctl.c
> @@ -1238,6 +1238,15 @@ static struct ctl_table kern_table[] = {
>  		.extra1		= SYSCTL_ZERO,
>  		.extra2		= SYSCTL_ONE,
>  	},
> +	{
> +		.procname	= "panic_on_mm_error",
> +		.data		= &panic_on_mm_error,
> +		.maxlen		= sizeof(int),
> +		.mode		= 0644,
> +		.proc_handler	= proc_dointvec_minmax,
> +		.extra1		= SYSCTL_ZERO,
> +		.extra2		= SYSCTL_ONE,
> +	},
>  #if defined(CONFIG_SMP) && defined(CONFIG_NO_HZ_COMMON)
>  	{
>  		.procname	= "timer_migration",
> diff --git a/mm/memory.c b/mm/memory.c
> index 45442d9a4f52..cce74ff39447 100644
> --- a/mm/memory.c
> +++ b/mm/memory.c
> @@ -71,6 +71,7 @@
>  #include <linux/dax.h>
>  #include <linux/oom.h>
>  #include <linux/numa.h>
> +#include <linux/module.h>
>  
>  #include <trace/events/kmem.h>
>  
> @@ -88,6 +89,8 @@
>  #warning Unfortunate NUMA and NUMA Balancing config, growing page-frame for last_cpupid.
>  #endif
>  
> +int panic_on_mm_error __read_mostly;
> +
>  #ifndef CONFIG_NEED_MULTIPLE_NODES
>  /* use the per-pgdat data instead for discontigmem - mbligh */
>  unsigned long max_mapnr;
> @@ -543,6 +546,10 @@ static void print_bad_pte(struct vm_area_struct *vma, unsigned long addr,
>  		 vma->vm_ops ? vma->vm_ops->fault : NULL,
>  		 vma->vm_file ? vma->vm_file->f_op->mmap : NULL,
>  		 mapping ? mapping->a_ops->readpage : NULL);
> +
> +	print_modules();
> +	if (panic_on_mm_error)
> +		panic("Bad page map detected");
>  	dump_stack();
>  	add_taint(TAINT_BAD_PAGE, LOCKDEP_NOW_UNRELIABLE);
>  }
> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> index d047bf7d8fd4..2ea6a65ba011 100644
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -643,9 +643,11 @@ static void bad_page(struct page *page, const char *reason,
>  	if (bad_flags)
>  		pr_alert("bad because of flags: %#lx(%pGp)\n",
>  						bad_flags, &bad_flags);
> -	dump_page_owner(page);
>  
> +	dump_page_owner(page);
>  	print_modules();
> +	if (panic_on_mm_error)
> +		panic("Bad page state detected");
>  	dump_stack();
>  out:
>  	/* Leave bad fields for debug, except PageBuddy could make trouble */

Reviewed-by: Aaron Tomlin <atomlin@redhat.com>

-- 
Aaron Tomlin