[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [PATCH] migration: fix analyze-migration.py script
From: |
Alexander Graf |
Subject: |
Re: [Qemu-devel] [PATCH] migration: fix analyze-migration.py script |
Date: |
Sun, 6 Sep 2015 10:36:30 +0200 |
User-agent: |
Mozilla/5.0 (Macintosh; Intel Mac OS X 10.10; rv:38.0) Gecko/20100101 Thunderbird/38.2.0 |
On 05.09.15 21:51, Mark Cave-Ayland wrote:
> Commit 61964 "Add configuration section" broke the analyze-migration.py script
> which terminates due to the unrecognised section. Fix the script by parsing
> the contents of the configuration section directly into a new
> ConfigurationSection object (although nothing is done with it yet).
>
> Signed-off-by: Mark Cave-Ayland <address@hidden>
> ---
> scripts/analyze-migration.py | 13 +++++++++++++
> 1 file changed, 13 insertions(+)
>
> diff --git a/scripts/analyze-migration.py b/scripts/analyze-migration.py
> index f6894be..1455387 100755
> --- a/scripts/analyze-migration.py
> +++ b/scripts/analyze-migration.py
> @@ -252,6 +252,15 @@ class HTABSection(object):
> def getDict(self):
> return ""
>
> +
> +class ConfigurationSection(object):
> + def __init__(self, file):
> + self.file = file
> +
> + def read(self):
> + name_len = self.file.read32()
> + name = self.file.readstr(len = name_len)
> +
> class VMSDFieldGeneric(object):
> def __init__(self, desc, file):
> self.file = file
> @@ -474,6 +483,7 @@ class MigrationDump(object):
> QEMU_VM_SECTION_FULL = 0x04
> QEMU_VM_SUBSECTION = 0x05
> QEMU_VM_VMDESCRIPTION = 0x06
> + QEMU_VM_CONFIGURATION = 0x07
> QEMU_VM_SECTION_FOOTER= 0x7e
>
> def __init__(self, filename):
> @@ -514,6 +524,9 @@ class MigrationDump(object):
> section_type = file.read8()
> if section_type == self.QEMU_VM_EOF:
> break
> + elif section_type == self.QEMU_VM_CONFIGURATION:
> + section = ConfigurationSection(file)
> + section.read()
So since we don't have a normal section header, there is no version
field either. That in turn means that the format is determined by the
machine version only - bleks.
So if there ever has to be more in the configuration section than the
machine name, please move to a more detectable scheme. Ideally something
that contains
* version
* length of dynamically sized fields
* lenght of full blob
would be ideal, so that we have a chance to at least put code into the
analyze script to examine it.
For now, I think the hard coded solution in the analyze script is
reasonable.
However, I think we should print out the name if we find it. It should
be as simple as adding a special case for the configuration section in
MigrationDump.getDict().
Thanks a lot for the patch :),
Alex
> elif section_type == self.QEMU_VM_SECTION_START or section_type
> == self.QEMU_VM_SECTION_FULL:
> section_id = file.read32()
> name = file.readstr()
>