qemu-block
[Top][All Lists]
Advanced

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

Re: [Qemu-block] [PATCH] scripts/dump-qcow2.pl: Script to dump qcow2 met


From: John Snow
Subject: Re: [Qemu-block] [PATCH] scripts/dump-qcow2.pl: Script to dump qcow2 metadata
Date: Thu, 19 Apr 2018 15:55:16 -0400
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101 Thunderbird/52.7.0


On 04/13/2018 10:45 AM, Max Reitz wrote:
> On 2018-03-28 15:38, Alberto Garcia wrote:
>> This script takes a qcow2 image and dumps its metadata: header,
>> snapshot table and some extensions (although not all qcow2 features
>> are supported yet).
>>
>> It can also display a list of all host clusters and the guest -> host
>> address mappings, so it's useful to debug allocations.
>>
>> The image is assumed not to be corrupted, and this script does not do
>> consistency checks (yet).
>>
>> Signed-off-by: Alberto Garcia <address@hidden>
>> ---
>>  scripts/dump-qcow2.pl | 425 
>> ++++++++++++++++++++++++++++++++++++++++++++++++++
>>  1 file changed, 425 insertions(+)
>>  create mode 100755 scripts/dump-qcow2.pl
> 
> First of all: The main reason I haven't reviewed this so far is because
> I don't like Perl.
> 

You know, I hadn't realized it was perl... I sat down to review it just
now and I'm probably not qualified to review anything in perl.

> OK, now that that's out of the way...  I'm wondering why you want to add
> this to the qemu tree?  If you'd written an iotest that would make use
> of it, sure.  But if it's just for debugging, then I'd personally think
> it would be better to just add it to a private repository.  That would
> give you more freedom to extend it, too.
> 
> (I know John has an own script for debugging qcow2 images.)
> 

(Very famously broken when trying to read compressed clusters!)

> Of course the nice thing about putting it into the qemu repository would
> be that people would be able to find it.  But then again, we could
> probably achieve the same with a Wiki page.
> 

Yeah, having debug tools in the tree is nice sometimes; but perhaps a
document with lists of pointers to tools we use to debug these things
would be as good for now.

> Max
> 



reply via email to

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