qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [PATCH 1/9] block/vpc: Make vpc_open() read the full dynamic header


From: Max Reitz
Subject: Re: [PATCH 1/9] block/vpc: Make vpc_open() read the full dynamic header
Date: Fri, 18 Dec 2020 11:19:46 +0100
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:78.0) Gecko/20100101 Thunderbird/78.5.0

On 17.12.20 17:19, Markus Armbruster wrote:
The dynamic header's size is 1024 bytes.

vpc_open() reads only the 512 bytes of the dynamic header into buf[].
Works, because it doesn't actually access the second half.  However, a
colleague told me that GCC 11 warns:

     ../block/vpc.c:358:51: error: array subscript 'struct VHDDynDiskHeader[0]' 
is partly outside array bounds of 'uint8_t[512]' [-Werror=array-bounds]

Clean up to read the full header.

Rename buf[] to dyndisk_header_buf[] while there.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
---
  block/vpc.c | 8 ++++----
  1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/block/vpc.c b/block/vpc.c
index 1ab55f9287..2fcf3f6283 100644
--- a/block/vpc.c
+++ b/block/vpc.c
@@ -220,7 +220,7 @@ static int vpc_open(BlockDriverState *bs, QDict *options, 
int flags,
      QemuOpts *opts = NULL;
      Error *local_err = NULL;
      bool use_chs;
-    uint8_t buf[HEADER_SIZE];
+    uint8_t dyndisk_header_buf[1024];

Perhaps this should be heap-allocated, but I don’t know whether qemu has ever established a (perhaps just inofficial) threshold on what goes on the stack and what goes on the heap.

      uint32_t checksum;
      uint64_t computed_size;
      uint64_t pagetable_size;
@@ -340,14 +340,14 @@ static int vpc_open(BlockDriverState *bs, QDict *options, 
int flags,
      }
if (disk_type == VHD_DYNAMIC) {
-        ret = bdrv_pread(bs->file, be64_to_cpu(footer->data_offset), buf,
-                         HEADER_SIZE);
+        ret = bdrv_pread(bs->file, be64_to_cpu(footer->data_offset),
+                         dyndisk_header_buf, 1024);

sizeof() would be better, but I see that’s what patch 6 is for.

          if (ret < 0) {
              error_setg(errp, "Error reading dynamic VHD header");
              goto fail;
          }
- dyndisk_header = (VHDDynDiskHeader *) buf;
+        dyndisk_header = (VHDDynDiskHeader *)dyndisk_header_buf;
if (strncmp(dyndisk_header->magic, "cxsparse", 8)) {
              error_setg(errp, "Invalid header magic");

Reviewed-by: Max Reitz <mreitz@redhat.com>




reply via email to

[Prev in Thread] Current Thread [Next in Thread]