linux-kernel - Re: [PATCH v2] x86/sgx: Free backing memory after faulting the enclave page

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] [day] [month] [year] [list]
Message-ID: <Yav+jLMu5aFRqWxJ@iki.fi>
Date:   Sun, 5 Dec 2021 01:49:32 +0200
From:   Jarkko Sakkinen <jarkko@...nel.org>
To:     Dave Hansen <dave.hansen@...ux.intel.com>,
        Thomas Gleixner <tglx@...utronix.de>,
        Ingo Molnar <mingo@...hat.com>, Borislav Petkov <bp@...en8.de>,
        x86@...nel.org, "H. Peter Anvin" <hpa@...or.com>,
        Sean Christopherson <seanjc@...gle.com>,
        Jethro Beekman <jethro@...tanix.com>
Cc:     reinette.chatre@...el.com, tony.luck@...el.com,
        nathaniel@...fian.com, stable@...r.kernel.org,
        Borislav Petkov <bp@...e.de>, linux-sgx@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2] x86/sgx: Free backing memory after faulting the
 enclave page

On Thu, Nov 11, 2021 at 07:44:01PM +0200, Jarkko Sakkinen wrote:
> There is a limited amount of SGX memory (EPC) on each system.  When that
> memory is used up, SGX has its own swapping mechanism which is similar
> in concept but totally separate from the core mm/* code.  Instead of
> swapping to disk, SGX swaps from EPC to normal RAM.  That normal RAM
> comes from a shared memory pseudo-file and can itself be swapped by the
> core mm code.  There is a hierarchy like this:
> 
> 	EPC <-> shmem <-> disk
> 
> After data is swapped back in from shmem to EPC, the shmem backing
> storage needs to be freed.  Currently, the backing shmem is not freed.
> This effectively wastes the shmem while the enclave is running.  The
> memory is recovered when the enclave is destroyed and the backing
> storage freed.
> 
> Sort this out by freeing memory with shmem_truncate_range(), as soon as
> a page is faulted back to the EPC.  In addition, free the memory for
> PCMD pages as soon as all PCMD's in a page have been marked as unused
> by zeroing its contents.
> 
> Reported-by: Dave Hansen <dave.hansen@...ux.intel.com>
> Cc: stable@...r.kernel.org
> Fixes: 1728ab54b4be ("x86/sgx: Add a page reclaimer")
> Signed-off-by: Jarkko Sakkinen <jarkko@...nel.org>
> ---
> v2:
> * Rewrite commit message as proposed by Dave.
> * Truncate PCMD pages (Dave).
> ---
>  arch/x86/kernel/cpu/sgx/encl.c | 48 +++++++++++++++++++++++++++++++---
>  1 file changed, 44 insertions(+), 4 deletions(-)
> 
> diff --git a/arch/x86/kernel/cpu/sgx/encl.c b/arch/x86/kernel/cpu/sgx/encl.c
> index 001808e3901c..ea43c10e5458 100644
> --- a/arch/x86/kernel/cpu/sgx/encl.c
> +++ b/arch/x86/kernel/cpu/sgx/encl.c
> @@ -12,6 +12,27 @@
>  #include "encls.h"
>  #include "sgx.h"
>  
> +
> +/*
> + * Get the page number of the page in the backing storage, which stores the PCMD
> + * of the enclave page in the given page index.  PCMD pages are located after
> + * the backing storage for the visible enclave pages and SECS.
> + */
> +static inline pgoff_t sgx_encl_get_backing_pcmd_nr(struct sgx_encl *encl, pgoff_t index)
> +{
> +	return PFN_DOWN(encl->size) + 1 + (index / sizeof(struct sgx_pcmd));
> +}
> +
> +/*
> + * Free a page from the backing storage in the given page index.
> + */
> +static inline void sgx_encl_truncate_backing_page(struct sgx_encl *encl, pgoff_t index)
> +{
> +	struct inode *inode = file_inode(encl->backing);
> +
> +	shmem_truncate_range(inode, PFN_PHYS(index), PFN_PHYS(index) + PAGE_SIZE - 1);
> +}
> +
>  /*
>   * ELDU: Load an EPC page as unblocked. For more info, see "OS Management of EPC
>   * Pages" in the SDM.
> @@ -24,7 +45,10 @@ static int __sgx_encl_eldu(struct sgx_encl_page *encl_page,
>  	struct sgx_encl *encl = encl_page->encl;
>  	struct sgx_pageinfo pginfo;
>  	struct sgx_backing b;
> +	bool pcmd_page_empty;
>  	pgoff_t page_index;
> +	pgoff_t pcmd_index;
> +	u8 *pcmd_page;
>  	int ret;
>  
>  	if (secs_page)
> @@ -38,8 +62,8 @@ static int __sgx_encl_eldu(struct sgx_encl_page *encl_page,
>  
>  	pginfo.addr = encl_page->desc & PAGE_MASK;
>  	pginfo.contents = (unsigned long)kmap_atomic(b.contents);
> -	pginfo.metadata = (unsigned long)kmap_atomic(b.pcmd) +
> -			  b.pcmd_offset;
> +	pcmd_page = kmap_atomic(b.pcmd);
> +	pginfo.metadata = (unsigned long)pcmd_page + b.pcmd_offset;
>  
>  	if (secs_page)
>  		pginfo.secs = (u64)sgx_get_epc_virt_addr(secs_page);
> @@ -55,11 +79,27 @@ static int __sgx_encl_eldu(struct sgx_encl_page *encl_page,
>  		ret = -EFAULT;
>  	}
>  
> -	kunmap_atomic((void *)(unsigned long)(pginfo.metadata - b.pcmd_offset));
> +	memset(pcmd_page + b.pcmd_offset, 0, sizeof(struct sgx_pcmd));
> +
> +	/*
> +	 * The area for the PCMD in the page was zeroed above.  Check if the
> +	 * whole page is now empty meaning that all PCMD's have been zeroed:
> +	 */
> +	pcmd_page_empty = !memchr_inv(pcmd_page, 0, PAGE_SIZE);
> +
> +	kunmap_atomic(pcmd_page);
>  	kunmap_atomic((void *)(unsigned long)pginfo.contents);
>  
>  	sgx_encl_put_backing(&b, false);
>  
> +	/* Free the backing memory. */
> +	sgx_encl_truncate_backing_page(encl, page_index);
> +
> +	if (pcmd_page_empty) {
> +		pcmd_index = sgx_encl_get_backing_pcmd_nr(encl, page_index);
> +		sgx_encl_truncate_backing_page(encl, pcmd_index);
> +	}
> +
>  	return ret;
>  }
>  
> @@ -577,7 +617,7 @@ static struct page *sgx_encl_get_backing_page(struct sgx_encl *encl,
>  int sgx_encl_get_backing(struct sgx_encl *encl, unsigned long page_index,
>  			 struct sgx_backing *backing)
>  {
> -	pgoff_t pcmd_index = PFN_DOWN(encl->size) + 1 + (page_index >> 5);
> +	pgoff_t pcmd_index = sgx_encl_get_backing_pcmd_nr(encl, page_index);
>  	struct page *contents;
>  	struct page *pcmd;
>  
> -- 
> 2.32.0
> 

Just a friendly remainder.

/Jarkko