From: Brijesh Singh <brijesh.singh@amd.com>
To: Borislav Petkov <bp@suse.de>, Thomas Gleixner <tglx@linutronix.de>
Cc: brijesh.singh@amd.com, x86@kernel.org,
linux-kernel@vger.kernel.org, kvm@vger.kernel.org,
"Tom Lendacky" <thomas.lendacky@amd.com>,
"H. Peter Anvin" <hpa@zytor.com>,
"Paolo Bonzini" <pbonzini@redhat.com>,
"Sean Christopherson" <sean.j.christopherson@intel.com>,
"Radim Krčmář" <rkrcmar@redhat.com>
Subject: Re: [PATCH v8 1/2] x86/mm: add .bss..decrypted section to hold shared variables
Date: Fri, 14 Sep 2018 09:27:09 -0500 [thread overview]
Message-ID: <43895f0e-fea7-001e-d727-d962a20f7f6c@amd.com> (raw)
In-Reply-To: <20180914141216.GA29898@zn.tnic>
On 09/14/2018 09:12 AM, Borislav Petkov wrote:
> On Fri, Sep 14, 2018 at 02:17:05PM +0200, Thomas Gleixner wrote:
>>> The sme_encrypt_kernel() does not have access to pmd (after pointer
>>> fixup is applied). You can extend the sme_encrypt_kernel() to pass an
>>> additional arguments but then we start getting in include hell. The pmd
>>> is defined as "pmdval_t". If we extend the sme_encrypt_kernel() then
>>> asm/mem_encrypt.h need to include the header file which defines
>>> "pmdval_t". Adding the 'asm/pgtable_type.h' was causing all kind of
>>> compilation errors. I didn't spend much time on it. IMO, we really don't
>>> need to go in this path unless we see some value from doing this.
>>
>> Keep it here then.
>
> *For what is worth*, a simple forward declaration works. I've taken the
> 64-bit forward declaration of pmdval_t as SME is 64-bit only anyway.
>
> The below diff ontop passes the mandatory all*config smoke builds:
>
> ---
> arch/x86/include/asm/mem_encrypt.h | 6 ++++--
> arch/x86/kernel/head64.c | 18 +-----------------
> arch/x86/mm/mem_encrypt_identity.c | 18 +++++++++++++++++-
> 3 files changed, 22 insertions(+), 20 deletions(-)
>
> diff --git a/arch/x86/include/asm/mem_encrypt.h b/arch/x86/include/asm/mem_encrypt.h
> index 616f8e637bc3..67c0e6cfdfb3 100644
> --- a/arch/x86/include/asm/mem_encrypt.h
> +++ b/arch/x86/include/asm/mem_encrypt.h
> @@ -19,6 +19,8 @@
>
> #include <asm/bootparam.h>
>
> +typedef unsigned long pmdval_t;
> +
> #ifdef CONFIG_AMD_MEM_ENCRYPT
>
> extern u64 sme_me_mask;
> @@ -40,7 +42,7 @@ void __init sme_unmap_bootdata(char *real_mode_data);
>
> void __init sme_early_init(void);
>
> -void __init sme_encrypt_kernel(struct boot_params *bp);
> +void __init sme_encrypt_kernel(struct boot_params *bp, pmdval_t *pmd);
> void __init sme_enable(struct boot_params *bp);
>
> int __init early_set_memory_decrypted(unsigned long vaddr, unsigned long size);
> @@ -69,7 +71,7 @@ static inline void __init sme_unmap_bootdata(char *real_mode_data) { }
>
> static inline void __init sme_early_init(void) { }
>
> -static inline void __init sme_encrypt_kernel(struct boot_params *bp) { }
> +static inline void __init sme_encrypt_kernel(struct boot_params *bp, pmdval_t *pmd) { }
> static inline void __init sme_enable(struct boot_params *bp) { }
>
> static inline bool sme_active(void) { return false; }
> diff --git a/arch/x86/kernel/head64.c b/arch/x86/kernel/head64.c
> index c16af27eb23f..6f8e9b534e80 100644
> --- a/arch/x86/kernel/head64.c
> +++ b/arch/x86/kernel/head64.c
> @@ -112,7 +112,6 @@ static bool __head check_la57_support(unsigned long physaddr)
> unsigned long __head __startup_64(unsigned long physaddr,
> struct boot_params *bp)
> {
> - unsigned long vaddr, vaddr_end;
> unsigned long load_delta, *p;
> unsigned long pgtable_flags;
> pgdval_t *pgd;
> @@ -233,22 +232,7 @@ unsigned long __head __startup_64(unsigned long physaddr,
> *fixup_long(&phys_base, physaddr) += load_delta - sme_get_me_mask();
>
> /* Encrypt the kernel and related (if SME is active) */
> - sme_encrypt_kernel(bp);
> -
> - /*
> - * Clear the memory encryption mask from the .bss..decrypted section.
> - * The bss section will be memset to zero later in the initialization so
> - * there is no need to zero it after changing the memory encryption
> - * attribute.
> - */
> - if (mem_encrypt_active()) {
> - vaddr = (unsigned long)__start_bss_decrypted;
> - vaddr_end = (unsigned long)__end_bss_decrypted;
> - for (; vaddr < vaddr_end; vaddr += PMD_SIZE) {
> - i = pmd_index(vaddr);
> - pmd[i] -= sme_get_me_mask();
> - }
> - }
> + sme_encrypt_kernel(bp, pmd);
>
> /*
> * Return the SME encryption mask (if SME is active) to be used as a
> diff --git a/arch/x86/mm/mem_encrypt_identity.c b/arch/x86/mm/mem_encrypt_identity.c
> index a19ef1a416ff..9dbc145d10f8 100644
> --- a/arch/x86/mm/mem_encrypt_identity.c
> +++ b/arch/x86/mm/mem_encrypt_identity.c
> @@ -267,15 +267,17 @@ static unsigned long __init sme_pgtable_calc(unsigned long len)
> return entries + tables;
> }
>
> -void __init sme_encrypt_kernel(struct boot_params *bp)
> +void __init sme_encrypt_kernel(struct boot_params *bp, pmdval_t *pmd)
> {
> unsigned long workarea_start, workarea_end, workarea_len;
> unsigned long execute_start, execute_end, execute_len;
> unsigned long kernel_start, kernel_end, kernel_len;
> unsigned long initrd_start, initrd_end, initrd_len;
> struct sme_populate_pgd_data ppd;
> + unsigned long vaddr, vaddr_end;
> unsigned long pgtable_area_len;
> unsigned long decrypted_base;
> + int i;
>
> if (!sme_active())
> return;
> @@ -467,6 +469,20 @@ void __init sme_encrypt_kernel(struct boot_params *bp)
>
> /* Flush the TLB - no globals so cr3 is enough */
> native_write_cr3(__native_read_cr3());
> +
> + /*
> + * Clear the memory encryption mask from the .bss..decrypted section.
> + * The bss section will be memset to zero later in the initialization so
> + * there is no need to zero it after changing the memory encryption
> + * attribute.
> + */
> + vaddr = (unsigned long)__start_bss_decrypted;
> + vaddr_end = (unsigned long)__end_bss_decrypted;
> +
> + for (; vaddr < vaddr_end; vaddr += PMD_SIZE) {
> + i = pmd_index(vaddr);
> + pmd[i] -= sme_get_me_mask();
> + }
> }
The above code will never get executed for the SEV case.
See if (!sme_active()) check in the start of function.
If we decide to go on this patch, then we have to do something like
this:
sme_encrypt_kernel(...)
{
if (!mem_encrypt_active())
return;
if (sev_active())
goto out;
/* Do kernel and initrd in-place encrypts for SME only case */
.....
.....
out:
/* Clear the C-bit from .bss..decrypted section */
...
...
}
>
> void __init sme_enable(struct boot_params *bp)
>
next prev parent reply other threads:[~2018-09-14 14:27 UTC|newest]
Thread overview: 15+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-09-13 21:51 [PATCH v8 0/2] x86: Fix SEV guest regression Brijesh Singh
2018-09-13 21:51 ` [PATCH v8 1/2] x86/mm: add .bss..decrypted section to hold shared variables Brijesh Singh
2018-09-13 23:24 ` Thomas Gleixner
2018-09-14 2:14 ` Brijesh Singh
2018-09-14 7:10 ` Borislav Petkov
2018-09-14 11:58 ` Brijesh Singh
2018-09-14 12:17 ` Thomas Gleixner
2018-09-14 14:12 ` Borislav Petkov
2018-09-14 14:27 ` Brijesh Singh [this message]
2018-09-14 14:45 ` Borislav Petkov
2018-09-14 14:50 ` Tom Lendacky
2018-09-14 14:55 ` Thomas Gleixner
2018-09-13 21:51 ` [PATCH v8 2/2] x86/kvm: use __bss_decrypted attribute in " Brijesh Singh
2018-09-13 23:31 ` Thomas Gleixner
2018-09-14 0:25 ` kbuild test robot
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=43895f0e-fea7-001e-d727-d962a20f7f6c@amd.com \
--to=brijesh.singh@amd.com \
--cc=bp@suse.de \
--cc=hpa@zytor.com \
--cc=kvm@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=pbonzini@redhat.com \
--cc=rkrcmar@redhat.com \
--cc=sean.j.christopherson@intel.com \
--cc=tglx@linutronix.de \
--cc=thomas.lendacky@amd.com \
--cc=x86@kernel.org \
/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).