linux-efi.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Ard Biesheuvel <ardb@kernel.org>
To: linux-efi@vger.kernel.org
Cc: Ard Biesheuvel <ardb@kernel.org>,
	Hans de Goede <hdegoede@redhat.com>,
	Arvind Sankar <nivedita@alum.mit.edu>,
	Andy Lutomirski <luto@amacapital.net>
Subject: [PATCH 3/3] efi/x86: Handle by-ref arguments covering multiple pages in mixed mode
Date: Thu, 13 Feb 2020 11:21:02 +0100	[thread overview]
Message-ID: <20200213102102.30170-4-ardb@kernel.org> (raw)
In-Reply-To: <20200213102102.30170-1-ardb@kernel.org>

We currently apply some stringent checks on the variable name and
buffer arguments passed to the EFI runtime services, and WARN()
if they are allocated in the vmalloc space and are not aligned to
their size - which must be a power of two - since in that case, it
is not guaranteed that they will not cross a page boundary.

However, we then simply proceed with the call, which could cause data
corruption if the next physical page belongs to a mapping that is
entirely unrelated.

Given that with vmap'ed stacks, this condition is much more likely
to trigger, let's relax the condition a bit, but fail the runtime
service call if it triggers.

Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
---
 arch/x86/platform/efi/efi_64.c | 45 +++++++++++---------
 1 file changed, 26 insertions(+), 19 deletions(-)

diff --git a/arch/x86/platform/efi/efi_64.c b/arch/x86/platform/efi/efi_64.c
index ae398587f264..d19a2edd63cb 100644
--- a/arch/x86/platform/efi/efi_64.c
+++ b/arch/x86/platform/efi/efi_64.c
@@ -180,7 +180,7 @@ void efi_sync_low_kernel_mappings(void)
 static inline phys_addr_t
 virt_to_phys_or_null_size(void *va, unsigned long size)
 {
-	bool bad_size;
+	phys_addr_t pa;
 
 	if (!va)
 		return 0;
@@ -188,16 +188,13 @@ virt_to_phys_or_null_size(void *va, unsigned long size)
 	if (virt_addr_valid(va))
 		return virt_to_phys(va);
 
-	/*
-	 * A fully aligned variable on the stack is guaranteed not to
-	 * cross a page bounary. Try to catch strings on the stack by
-	 * checking that 'size' is a power of two.
-	 */
-	bad_size = size > PAGE_SIZE || !is_power_of_2(size);
+	pa = slow_virt_to_phys(va);
 
-	WARN_ON(!IS_ALIGNED((unsigned long)va, size) || bad_size);
+	/* check if the object crosses a page boundary */
+	if (WARN_ON((pa ^ (pa + size - 1)) & PAGE_MASK))
+		return 0;
 
-	return slow_virt_to_phys(va);
+	return pa;
 }
 
 #define virt_to_phys_or_null(addr)				\
@@ -615,8 +612,11 @@ efi_thunk_get_variable(efi_char16_t *name, efi_guid_t *vendor,
 	phys_attr = virt_to_phys_or_null(attr);
 	phys_data = virt_to_phys_or_null_size(data, *data_size);
 
-	status = efi_thunk(get_variable, phys_name, phys_vendor,
-			   phys_attr, phys_data_size, phys_data);
+	if (!phys_name || (data && !phys_data))
+		status = EFI_INVALID_PARAMETER;
+	else
+		status = efi_thunk(get_variable, phys_name, phys_vendor,
+				   phys_attr, phys_data_size, phys_data);
 
 	spin_unlock_irqrestore(&efi_runtime_lock, flags);
 
@@ -641,9 +641,11 @@ efi_thunk_set_variable(efi_char16_t *name, efi_guid_t *vendor,
 	phys_vendor = virt_to_phys_or_null(vnd);
 	phys_data = virt_to_phys_or_null_size(data, data_size);
 
-	/* If data_size is > sizeof(u32) we've got problems */
-	status = efi_thunk(set_variable, phys_name, phys_vendor,
-			   attr, data_size, phys_data);
+	if (!phys_name || !phys_data)
+		status = EFI_INVALID_PARAMETER;
+	else
+		status = efi_thunk(set_variable, phys_name, phys_vendor,
+				   attr, data_size, phys_data);
 
 	spin_unlock_irqrestore(&efi_runtime_lock, flags);
 
@@ -670,9 +672,11 @@ efi_thunk_set_variable_nonblocking(efi_char16_t *name, efi_guid_t *vendor,
 	phys_vendor = virt_to_phys_or_null(vnd);
 	phys_data = virt_to_phys_or_null_size(data, data_size);
 
-	/* If data_size is > sizeof(u32) we've got problems */
-	status = efi_thunk(set_variable, phys_name, phys_vendor,
-			   attr, data_size, phys_data);
+	if (!phys_name || !phys_data)
+		status = EFI_INVALID_PARAMETER;
+	else
+		status = efi_thunk(set_variable, phys_name, phys_vendor,
+				   attr, data_size, phys_data);
 
 	spin_unlock_irqrestore(&efi_runtime_lock, flags);
 
@@ -698,8 +702,11 @@ efi_thunk_get_next_variable(unsigned long *name_size,
 	phys_vendor = virt_to_phys_or_null(vnd);
 	phys_name = virt_to_phys_or_null_size(name, *name_size);
 
-	status = efi_thunk(get_next_variable, phys_name_size,
-			   phys_name, phys_vendor);
+	if (!phys_name)
+		status = EFI_INVALID_PARAMETER;
+	else
+		status = efi_thunk(get_next_variable, phys_name_size,
+				   phys_name, phys_vendor);
 
 	spin_unlock_irqrestore(&efi_runtime_lock, flags);
 
-- 
2.17.1


      parent reply	other threads:[~2020-02-13 10:21 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-02-13 10:20 [PATCH 0/3] efi/x86: Some mixed mode fixes Ard Biesheuvel
2020-02-13 10:21 ` [PATCH 1/3] efi/x86: align GUIDs to their size in the mixed mode runtime wrapper Ard Biesheuvel
2020-02-14 12:55   ` Hans de Goede
2020-02-14 13:35     ` Ard Biesheuvel
2020-02-13 10:21 ` [PATCH 2/3] efi/x86: remove support for EFI time and counter services in mixed mode Ard Biesheuvel
2020-02-13 10:21 ` Ard Biesheuvel [this message]

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=20200213102102.30170-4-ardb@kernel.org \
    --to=ardb@kernel.org \
    --cc=hdegoede@redhat.com \
    --cc=linux-efi@vger.kernel.org \
    --cc=luto@amacapital.net \
    --cc=nivedita@alum.mit.edu \
    /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).