KVM Archive on lore.kernel.org
 help / color / Atom feed
From: Christophe Leroy <christophe.leroy@csgroup.eu>
To: Sean Christopherson <seanjc@google.com>,
	Paolo Bonzini <pbonzini@redhat.com>,
	Brijesh Singh <brijesh.singh@amd.com>,
	Tom Lendacky <thomas.lendacky@amd.com>,
	John Allen <john.allen@amd.com>
Cc: Vitaly Kuznetsov <vkuznets@redhat.com>,
	Wanpeng Li <wanpengli@tencent.com>,
	Jim Mattson <jmattson@google.com>, Joerg Roedel <joro@8bytes.org>,
	kvm@vger.kernel.org, linux-crypto@vger.kernel.org,
	linux-kernel@vger.kernel.org, Borislav Petkov <bp@suse.de>
Subject: Re: [PATCH 3/5] crypto: ccp: Play nice with vmalloc'd memory for SEV command structs
Date: Sun, 4 Apr 2021 08:48:17 +0200
Message-ID: <8ea3744f-fdf7-1704-2860-40c2b8fb47e1@csgroup.eu> (raw)
In-Reply-To: <20210402233702.3291792-4-seanjc@google.com>



Le 03/04/2021 à 01:37, Sean Christopherson a écrit :
> Copy vmalloc'd data to an internal buffer instead of rejecting outright
> so that callers can put SEV command buffers on the stack without running
> afoul of CONFIG_VMAP_STACK=y.  Currently, the largest supported command
> takes a 68 byte buffer, i.e. pretty much every command can be put on the
> stack.  Because sev_cmd_mutex is held for the entirety of a transaction,
> only a single bounce buffer is required.
> 
> Use a flexible array for the buffer, sized to hold the largest known
> command.   Alternatively, the buffer could be a union of all known
> command structs, but that would incur a higher maintenance cost due to
> the need to update the union for every command in addition to updating
> the existing sev_cmd_buffer_len().
> 
> Align the buffer to an 8-byte boundary, mimicking the alignment that
> would be provided by the compiler if any of the structs were embedded
> directly.  Note, sizeof() correctly incorporates this alignment.
> 
> Cc: Brijesh Singh <brijesh.singh@amd.com>
> Cc: Borislav Petkov <bp@suse.de>
> Cc: Tom Lendacky <thomas.lendacky@amd.com>
> Signed-off-by: Sean Christopherson <seanjc@google.com>
> ---
>   drivers/crypto/ccp/sev-dev.c | 33 +++++++++++++++++++++++++++------
>   drivers/crypto/ccp/sev-dev.h |  7 +++++++
>   2 files changed, 34 insertions(+), 6 deletions(-)
> 
> diff --git a/drivers/crypto/ccp/sev-dev.c b/drivers/crypto/ccp/sev-dev.c
> index 4c513318f16a..6d5882290cfc 100644
> --- a/drivers/crypto/ccp/sev-dev.c
> +++ b/drivers/crypto/ccp/sev-dev.c
> @@ -135,13 +135,14 @@ static int sev_cmd_buffer_len(int cmd)
>   	return 0;
>   }
>   
> -static int __sev_do_cmd_locked(int cmd, void *data, int *psp_ret)
> +static int __sev_do_cmd_locked(int cmd, void *__data, int *psp_ret)
>   {
>   	struct psp_device *psp = psp_master;
>   	struct sev_device *sev;
>   	unsigned int phys_lsb, phys_msb;
>   	unsigned int reg, ret = 0;
>   	int buf_len;
> +	void *data;
>   
>   	if (!psp || !psp->sev_data)
>   		return -ENODEV;
> @@ -152,11 +153,21 @@ static int __sev_do_cmd_locked(int cmd, void *data, int *psp_ret)
>   	sev = psp->sev_data;
>   
>   	buf_len = sev_cmd_buffer_len(cmd);
> -	if (WARN_ON_ONCE(!!data != !!buf_len))
> +	if (WARN_ON_ONCE(!!__data != !!buf_len))
>   		return -EINVAL;
>   
> -	if (WARN_ON_ONCE(data && is_vmalloc_addr(data)))
> -		return -EINVAL;
> +	if (__data && is_vmalloc_addr(__data)) {
> +		/*
> +		 * If the incoming buffer is virtually allocated, copy it to
> +		 * the driver's scratch buffer as __pa() will not work for such
> +		 * addresses, vmalloc_to_page() is not guaranteed to succeed,
> +		 * and vmalloc'd data may not be physically contiguous.
> +		 */
> +		data = sev->cmd_buf;
> +		memcpy(data, __data, buf_len);
> +	} else {
> +		data = __data;
> +	}

I don't know how big commands are, but if they are small, it would probably be more efficient to 
inconditionnally copy them to the buffer rather then doing the test.

>   
>   	/* Get the physical address of the command buffer */
>   	phys_lsb = data ? lower_32_bits(__psp_pa(data)) : 0;
> @@ -204,6 +215,13 @@ static int __sev_do_cmd_locked(int cmd, void *data, int *psp_ret)
>   	print_hex_dump_debug("(out): ", DUMP_PREFIX_OFFSET, 16, 2, data,
>   			     buf_len, false);
>   
> +	/*
> +	 * Copy potential output from the PSP back to __data.  Do this even on
> +	 * failure in case the caller wants to glean something from the error.
> +	 */
> +	if (__data && data != __data)
> +		memcpy(__data, data, buf_len);
> +
>   	return ret;
>   }
>   
> @@ -978,9 +996,12 @@ int sev_dev_init(struct psp_device *psp)
>   {
>   	struct device *dev = psp->dev;
>   	struct sev_device *sev;
> -	int ret = -ENOMEM;
> +	int ret = -ENOMEM, cmd_buf_size = 0, i;
>   
> -	sev = devm_kzalloc(dev, sizeof(*sev), GFP_KERNEL);
> +	for (i = 0; i < SEV_CMD_MAX; i++)
> +		cmd_buf_size = max(cmd_buf_size, sev_cmd_buffer_len(i));
> +
> +	sev = devm_kzalloc(dev, sizeof(*sev) + cmd_buf_size, GFP_KERNEL);
>   	if (!sev)
>   		goto e_err;
>   
> diff --git a/drivers/crypto/ccp/sev-dev.h b/drivers/crypto/ccp/sev-dev.h
> index dd5c4fe82914..b43283ce2d73 100644
> --- a/drivers/crypto/ccp/sev-dev.h
> +++ b/drivers/crypto/ccp/sev-dev.h
> @@ -52,6 +52,13 @@ struct sev_device {
>   	u8 api_major;
>   	u8 api_minor;
>   	u8 build;
> +
> +	/*
> +	 * Buffer used for incoming commands whose physical address cannot be
> +	 * resolved via __pa(), e.g. stack pointers when CONFIG_VMAP_STACK=y.
> +	 * Note, alignment isn't strictly required.
> +	 */
> +	u8 cmd_buf[] __aligned(8);
>   };
>   
>   int sev_dev_init(struct psp_device *psp);
> 

  parent reply index

Thread overview: 17+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-04-02 23:36 [PATCH 0/5] ccp: KVM: SVM: Use stack for SEV command buffers Sean Christopherson
2021-04-02 23:36 ` [PATCH 1/5] crypto: ccp: Detect and reject vmalloc addresses destined for PSP Sean Christopherson
2021-04-04  6:31   ` Christophe Leroy
2021-04-02 23:36 ` [PATCH 2/5] crypto: ccp: Reject SEV commands with mismatching command buffer Sean Christopherson
2021-04-03 17:02   ` Christophe Leroy
2021-04-05 16:26   ` Tom Lendacky
2021-04-05 16:33     ` Sean Christopherson
2021-04-05 16:37       ` Tom Lendacky
2021-04-02 23:37 ` [PATCH 3/5] crypto: ccp: Play nice with vmalloc'd memory for SEV command structs Sean Christopherson
2021-04-03 17:05   ` Christophe Leroy
2021-04-03 17:13   ` Christophe Leroy
2021-04-04  6:48   ` Christophe Leroy [this message]
2021-04-05 15:06     ` Sean Christopherson
2021-04-05 16:01       ` Brijesh Singh
2021-04-02 23:37 ` [PATCH 4/5] crypto: ccp: Use the stack for small SEV command buffers Sean Christopherson
2021-04-02 23:37 ` [PATCH 5/5] KVM: SVM: Allocate SEV command structures on local stack Sean Christopherson
2021-04-04 19:54 ` [PATCH 0/5] ccp: KVM: SVM: Use stack for SEV command buffers Brijesh Singh

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=8ea3744f-fdf7-1704-2860-40c2b8fb47e1@csgroup.eu \
    --to=christophe.leroy@csgroup.eu \
    --cc=bp@suse.de \
    --cc=brijesh.singh@amd.com \
    --cc=jmattson@google.com \
    --cc=john.allen@amd.com \
    --cc=joro@8bytes.org \
    --cc=kvm@vger.kernel.org \
    --cc=linux-crypto@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=pbonzini@redhat.com \
    --cc=seanjc@google.com \
    --cc=thomas.lendacky@amd.com \
    --cc=vkuznets@redhat.com \
    --cc=wanpengli@tencent.com \
    /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

KVM Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/kvm/0 kvm/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 kvm kvm/ https://lore.kernel.org/kvm \
		kvm@vger.kernel.org
	public-inbox-index kvm

Example config snippet for mirrors

Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.kernel.vger.kvm


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git