[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [PATCH 3/9] block/vpc: Don't abuse the footer buffer for dynamic hea
From: |
Markus Armbruster |
Subject: |
Re: [PATCH 3/9] block/vpc: Don't abuse the footer buffer for dynamic header |
Date: |
Fri, 18 Dec 2020 14:52:54 +0100 |
User-agent: |
Gnus/5.13 (Gnus v5.13) Emacs/27.1 (gnu/linux) |
Max Reitz <mreitz@redhat.com> writes:
> On 17.12.20 17:19, Markus Armbruster wrote:
>> create_dynamic_disk() takes a buffer holding the footer as first
>> argument. It writes out the footer (512 bytes), then reuses the
>> buffer to initialize and write out the dynamic header (1024 bytes).
>> Works, because the caller passes a buffer that is large enough for
>> both purposes. I hate that.
>> Use a separate buffer for the dynamic header, and adjust the
>> caller's
>> buffer.
>> Signed-off-by: Markus Armbruster <armbru@redhat.com>
>> ---
>> block/vpc.c | 22 ++++++++++++----------
>> 1 file changed, 12 insertions(+), 10 deletions(-)
>> diff --git a/block/vpc.c b/block/vpc.c
>> index d18ecc3da1..34186640ee 100644
>> --- a/block/vpc.c
>> +++ b/block/vpc.c
>> @@ -822,8 +822,9 @@ static int calculate_geometry(int64_t total_sectors,
>> uint16_t *cyls,
>> static int create_dynamic_disk(BlockBackend *blk, uint8_t *buf,
>> int64_t total_sectors)
>> {
>> + uint8_t dyndisk_header_buf[1024];
>> VHDDynDiskHeader *dyndisk_header =
>> - (VHDDynDiskHeader *) buf;
>> + (VHDDynDiskHeader *)dyndisk_header_buf;
>> uint8_t bat_sector[512];
>> size_t block_size, num_bat_entries;
>> int i;
>> @@ -858,7 +859,7 @@ static int create_dynamic_disk(BlockBackend *blk,
>> uint8_t *buf,
>> }
>> /* Prepare the Dynamic Disk Header */
>> - memset(buf, 0, 1024);
>> + memset(dyndisk_header_buf, 0, 1024);
>> memcpy(dyndisk_header->magic, "cxsparse", 8);
>
> I’d prefer something like
>
> *dyndisk_header = (VHDDynDiskHeader){
> ...
> };
>
> but that isn’t possible before patch 5.
Yes. And afterwards, using an initializer would be even simpler, I
guess.
> (And can be done on top now
> anyway, especially given that Kevin probably wants to send a pull
> request today.)
Yes, would be good to get this wrapped before the break.
> [...]
>
>> @@ -972,8 +974,8 @@ static int coroutine_fn
>> vpc_co_create(BlockdevCreateOptions *opts,
>> BlockBackend *blk = NULL;
>> BlockDriverState *bs = NULL;
>> - uint8_t buf[1024];
>> - VHDFooter *footer = (VHDFooter *) buf;
>> + uint8_t footer_buf[HEADER_SIZE];
>> + VHDFooter *footer = (VHDFooter *)footer_buf;
>> uint16_t cyls = 0;
>> uint8_t heads = 0;
>> uint8_t secs_per_cyl = 0;
>> @@ -1036,7 +1038,7 @@ static int coroutine_fn
>> vpc_co_create(BlockdevCreateOptions *opts,
>> }
>> /* Prepare the Hard Disk Footer */
>> - memset(buf, 0, 1024);
>> + memset(footer_buf, 0, HEADER_SIZE);
>> memcpy(footer->creator, "conectix", 8);
>
> Same here, except here it’s patch 7.
>
> Reviewed-by: Max Reitz <mreitz@redhat.com>
Thanks!
- [PATCH 0/9] block/vpc: Clean up some buffer abuse, Markus Armbruster, 2020/12/17
- [PATCH 3/9] block/vpc: Don't abuse the footer buffer for dynamic header, Markus Armbruster, 2020/12/17
- [PATCH 2/9] block/vpc: Don't abuse the footer buffer as BAT sector buffer, Markus Armbruster, 2020/12/17
- [PATCH 4/9] block/vpc: Make vpc_checksum() take void *, Markus Armbruster, 2020/12/17
- [PATCH 6/9] block/vpc: Use sizeof() instead of 1024 for dynamic header size, Markus Armbruster, 2020/12/17
- [PATCH 8/9] block/vpc: Pass footer buffers as VHDFooter * instead of uint8_t *, Markus Armbruster, 2020/12/17
- [PATCH 7/9] block/vpc: Pad VHDFooter, replace uint8_t[] buffers, Markus Armbruster, 2020/12/17