Re: [PATCH] mm: page_alloc: fix page_poison=1 / INIT_ON_ALLOC_DEFAULT_ON interaction

[Andrew Morton]

On Mon, 12 Jul 2021 22:58:16 +0100 Sergei Trofimovich <slyfox@xxxxxxxxxx> wrote:

> To reproduce the failure we need the following system:
>   - kernel command: page_poison=1 init_on_free=0 init_on_alloc=0
>   - kernel config:
>     * CONFIG_INIT_ON_ALLOC_DEFAULT_ON=y
>     * CONFIG_INIT_ON_FREE_DEFAULT_ON=y
>     * CONFIG_PAGE_POISONING=y
> 
>     0000000085629bdd: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
>     0000000022861832: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
>     00000000c597f5b0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
>     CPU: 11 PID: 15195 Comm: bash Kdump: loaded Tainted: G     U     O      5.13.1-gentoo-x86_64 #1
>     Hardware name: System manufacturer System Product Name/PRIME Z370-A, BIOS 2801 01/13/2021
>     Call Trace:
>      dump_stack+0x64/0x7c
>      __kernel_unpoison_pages.cold+0x48/0x84
>      post_alloc_hook+0x60/0xa0
>      get_page_from_freelist+0xdb8/0x1000
>      __alloc_pages+0x163/0x2b0
>      __get_free_pages+0xc/0x30
>      pgd_alloc+0x2e/0x1a0
>      ? dup_mm+0x37/0x4f0
>      mm_init+0x185/0x270
>      dup_mm+0x6b/0x4f0
>      ? __lock_task_sighand+0x35/0x70
>      copy_process+0x190d/0x1b10
>      kernel_clone+0xba/0x3b0
>      __do_sys_clone+0x8f/0xb0
>      do_syscall_64+0x68/0x80
>      ? do_syscall_64+0x11/0x80
>      entry_SYSCALL_64_after_hwframe+0x44/0xae
> 
> Before the 51cba1eb ("init_on_alloc: Optimize static branches")
> init_on_alloc never enabled static branch by default. It could
> only be enabed explicitly by init_mem_debugging_and_hardening().
> 
> But after the 51cba1eb static branch could already be enabled
> by default. There was no code to ever disable it. That caused
> page_poison=1 / init_on_free=1 conflict.
> 
> This change extends init_mem_debugging_and_hardening() to also
> disable static branch disabling.
> 
> ...
>
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -840,18 +840,22 @@ void init_mem_debugging_and_hardening(void)
>  	}
>  #endif
>  
> -	if (_init_on_alloc_enabled_early) {
> -		if (page_poisoning_requested)
> +	if (_init_on_alloc_enabled_early ||
> +	    IS_ENABLED(CONFIG_INIT_ON_ALLOC_DEFAULT_ON)) {
> +		if (page_poisoning_requested) {
>  			pr_info("mem auto-init: CONFIG_PAGE_POISONING is on, "
>  				"will take precedence over init_on_alloc\n");
> -		else
> +			static_branch_disable(&init_on_alloc);
> +		} else
>  			static_branch_enable(&init_on_alloc);
>  	}
> -	if (_init_on_free_enabled_early) {
> -		if (page_poisoning_requested)
> +	if (_init_on_free_enabled_early ||
> +	    IS_ENABLED(CONFIG_INIT_ON_FREE_DEFAULT_ON)) {
> +		if (page_poisoning_requested) {
>  			pr_info("mem auto-init: CONFIG_PAGE_POISONING is on, "
>  				"will take precedence over init_on_free\n");
> -		else
> +			static_branch_disable(&init_on_free);
> +		} else
>  			static_branch_enable(&init_on_free);
>  	}
>  

I'm thinking this is sufficiently serious and sufficiently reported to
warrant a cc:stable backport.  Agree?