Skip to contentSkip to conversation

qcow2 spec: Describe string header extensions

PATCH In review Kevin Wolf 3 participants 4 messages Mar 7, 2019 – Mar 8, 2019

lore
Share & tools
lore

Kevin Wolf

author
#0 lore raw

PATCH 1/1 qcow2 spec: Describe string header extensions

Be more specific about the string representation in header extensions.

---
 docs/interop/qcow2.txt | 14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)
docs/interop/qcow2.txt +12 -2
index bfb97cfde3..8c3098d8d9 100644
@@ -170,11 +170,11 @@ be stored. Each extension has a structure like the following:
 
     Byte  0 -  3:   Header extension type:
                         0x00000000 - End of the header extension area
-                        0xE2792ACA - Backing file format name
+                        0xE2792ACA - Backing file format name string
                         0x6803f857 - Feature name table
                         0x23852875 - Bitmaps extension
                         0x0537be77 - Full disk encryption header pointer
-                        0x44415441 - External data file name
+                        0x44415441 - External data file name string
                         other      - Unknown header extension, can be safely
                                      ignored
 
@@ -196,6 +196,16 @@ data of compatible features that it doesn't support. Compatible features that
 need space for additional data can use a header extension.
 
 
+== String header extensions ==
+
+Some header extensions (such as the backing file format name and the external
+data file name) are just a single string. In this case, the header extension
+length is the string length and the string is not '\0' terminated. (The header
+extension padding can make it look like a string is '\0' terminated, but
+neither is padding always necessary nor is there a guarantee that zero bytes
+are used for padding.)
+
+
 == Feature name table ==
 
 The feature name table is an optional header extension that contains the name
Signature · 1 line
2.20.1
Reviewed-by on v1 1/1
#1 lore raw
Quoted 8 lines · Kevin Wolf

Be more specific about the string representation in header extensions.

---
 docs/interop/qcow2.txt | 14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)

Based-on: <[email protected]>

docs/interop/qcow2.txt data of compatible features that it doesn't support. Compatible features that · line 206

Quoted 10 lines

+== String header extensions ==

docs/interop/qcow2.txt +7 -0
+
+Some header extensions (such as the backing file format name and the external
+data file name) are just a single string. In this case, the header extension
+length is the string length and the string is not '\0' terminated. (The header
+extension padding can make it look like a string is '\0' terminated, but
+neither is padding always necessary nor is there a guarantee that zero bytes
+are used for padding.)

We didn't require 0 padding? (goes and re-reads) - oops, yes that's correct. It makes it harder to extend a struct by making use of that padding if you can't guarantee what the padding had to be prior to the extension, and means that you have to consider whether there are any potential security risks of the padding being used as a side channel to leak information while still being a well-formed file. But changing the standard to require zero padding is different than documenting existing practice, so your patch is correct.

Signature · 3 lines
Eric Blake, Principal Software Engineer
Red Hat, Inc.           +1-919-301-3226
Virtualization:  qemu.org | libvirt.org
author
#2 lore raw

Am 07.03.2019 um 18:40 hat Eric Blake geschrieben:

Quoted 14 lines · Kevin Wolf
Quoted 8 lines

Be more specific about the string representation in header extensions.

---
 docs/interop/qcow2.txt | 14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)

Based-on: <[email protected]>

Quoted 2 lines
Quoted 11 lines

docs/interop/qcow2.txt data of compatible features that it doesn't support. Compatible features that · line 206

Quoted 8 lines

+== String header extensions ==

docs/interop/qcow2.txt +7 -0
+
+Some header extensions (such as the backing file format name and the external
+data file name) are just a single string. In this case, the header extension
+length is the string length and the string is not '\0' terminated. (The header
+extension padding can make it look like a string is '\0' terminated, but
+neither is padding always necessary nor is there a guarantee that zero bytes
+are used for padding.)

We didn't require 0 padding? (goes and re-reads) - oops, yes that's correct.

Yes. QEMU does use zeroes, but the spec doesn't mandate it and changing it would be an incompatible change. And the worst that could happen is that someone leaves a hidden message in the padding, which doesn't really hurt.

Quoted 5 lines

It makes it harder to extend a struct by making use of that padding if you can't guarantee what the padding had to be prior to the extension, and means that you have to consider whether there are any potential security risks of the padding being used as a side channel to leak information while still being a well-formed file.

It doesn't actually make it harder because we have the length field which tells us the exact length of the valid data, so we shouldn't even try to use any of the padding.

Quoted 4 lines

But changing the standard to require zero padding is different than documenting existing practice, so your patch is correct.

Thanks!

Kevin