qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Anthony Liguori <anthony@codemonkey.ws>
To: qemu-devel@nongnu.org, "Daniel P. Berrange" <berrange@redhat.com>
Subject: Re: [Qemu-devel] [PATCH 4/4][RFC] Add logic to QEMU to read command line options from qcow2 images
Date: Sat, 11 Aug 2007 14:49:13 -0500	[thread overview]
Message-ID: <46BE12B9.1000203@codemonkey.ws> (raw)
In-Reply-To: <fb249edb0708111011x5abd6a3ehb4a0d3876c2a91c8@mail.gmail.com>

andrzej zaborowski wrote:
> On 08/08/07, Jorge Lucángeli Obes <t4m5yn@gmail.com> wrote:
>   
>> This patch makes QEMU check for command line options stored in qcow2 images.
>>
>> Signed-off-by: Laurent Vivier <Laurent.Vivier@bull.net>
>> Signed-off-by: Jorge Lucángeli Obes <t4m5yn@gmail.com>
>> ---
>> diff --git a/qemu/vl.c b/qemu/vl.c
>> index 4ad39f1..1d28794 100644
>> --- a/qemu/vl.c
>> +++ b/qemu/vl.c
>> @@ -184,6 +184,11 @@ int time_drift_fix = 0;
>>  const char *cpu_vendor_string;
>>
>>  /***********************************************************/
>> +/* Image annotation support */
>> +
>> +char *img_get_annot(char *filename);
>> +
>> +/***********************************************************/
>>  /* x86 ISA bus support */
>>
>>  target_phys_addr_t isa_mem_base = 0;
>> @@ -6917,6 +6922,13 @@ int main(int argc, char **argv)
>>      char usb_devices[MAX_USB_CMDLINE][128];
>>      int usb_devices_index;
>>      int fds[2];
>> +    char *tmpannot;
>> +    char annot[1024];
>> +    int done = 0;
>> +    unsigned int nbtoks = 0;
>> +    char *tok;
>> +    BlockDriver *drv;
>> +    BlockDriverState *bs;
>>
>>      saved_argc = argc;
>>      saved_argv = argv;
>> @@ -7000,6 +7012,58 @@ int main(int argc, char **argv)
>>      nb_nics = 0;
>>      /* default mac address of the first network interface */
>>
>> +    bdrv_init();
>> +
>> +    drv = bdrv_find_format("qcow2");
>> +
>> +    if (argc > 1 && argv[1][0] != '-') {
>> +        bs = bdrv_new("");
>> +        if (!bs) {
>> +            fprintf(stderr, "Not enough memory");
>> +            exit(1);
>> +        }
>> +        if (bdrv_open2(bs, argv[1], 0, drv) < 0) {
>> +            fprintf(stderr, "Could not open '%s'", argv[1]);
>> +            bdrv_delete(bs);
>> +            exit(1);
>> +        }
>> +
>> +        tmpannot = bdrv_get_annot(bs, "commandline_args");
>> +        if (tmpannot) {
>> +            pstrcpy(annot, 1024, tmpannot);
>>     
>
> I thought one of the arguments for storing config options was the
> commandline length limit, which is 32k or so characters on most
> systems.
>
>   
>> +
>> +            do {
>> +                tok = strtok(nbtoks == 0? tmpannot : NULL, " ");
>> +
>> +                if (tok != NULL)
>> +                    nbtoks++;
>> +                else
>> +                    done = 1;
>> +            } while (!done);
>> +
>> +            free(tmpannot);
>> +
>> +            if (nbtoks > 0) {
>> +                char **argvprime = malloc((nbtoks + argc) * sizeof(char*));
>> +
>> +                for (i = 0; i < argc; i++)
>> +                    argvprime[i] = argv[i];
>> +
>> +                for (i = 0; i < nbtoks; i++)
>> +                    argvprime[i + argc] = strtok(i == 0? annot : NULL, " ");
>> +
>> +                argv = argvprime;
>> +                argc = argc + nbtoks;
>> +
>> +                for (i = 0; i < nbtoks + 2; i++)
>> +                    printf("argv[%d] = %s\n", i, argv[i]);
>> +
>> +            }
>> +        }
>> +
>> +        bdrv_delete(bs);
>> +    }
>> +
>>      optind = 1;
>>      for(;;) {
>>          if (optind >= argc)
>> @@ -7558,7 +7622,6 @@ int main(int argc, char **argv)
>>  #endif
>>
>>      /* we always create the cdrom drive, even if no disk is there */
>> -    bdrv_init();
>>      if (cdrom_index >= 0) {
>>          bs_table[cdrom_index] = bdrv_new("cdrom");
>>          bdrv_set_type_hint(bs_table[cdrom_index], BDRV_TYPE_CDROM);
>> @@ -7764,5 +7827,8 @@ int main(int argc, char **argv)
>>
>>      main_loop();
>>      quit_timers();
>> +
>> +    /* was reassigned above so it needs free()ing */
>> +    free(argv);
>>      return 0;
>>  }
>>
>>
>>     
>
> Yes, the file format starting with "#! /path/to/qemu" is a much better
> idea, but the commandline length limitation has to be taken care of
> separately, perhaps the switches should follow on next lines in the
> file while the #! line shouldn't specify the switches.
>   

Yes, I think that's a very good idea.  I think it should be extended 
slightly to incorporate Dan's original suggestion of using an option to 
specify that the command line ought to be read from the image.  So you'd 
end up with:

#!/path/to/qemu -read-args-from-image <nl>
-m 512 -net tap -net nic,model=rtl8139 <padding><nl>

Such that this whole section was sector aligned.  The "#!" is the magic 
for the disk image type.  On windows, you could completely omit 
/path/to/qemu and just set up an extension association that would 
associate .qemu with "/path/to/qemu -read-args-from-image %s" and then 
it would work equally well with Windows.

Regards,

Anthony Liguori

> Regards,
> Andrzej (away until mid-September(?))
>   

      parent reply	other threads:[~2007-08-11 19:49 UTC|newest]

Thread overview: 25+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <59abf66e0708081124g14901b01i841b70d17ae1e097@mail.gmail.com>
2007-08-08 19:52 ` [Qemu-devel] [PATCH 4/4][RFC] Add logic to QEMU to read command line options from qcow2 images Jorge Lucángeli Obes
2007-08-08 20:24   ` Daniel P. Berrange
2007-08-09 14:54     ` Anthony Liguori
2007-08-09 15:07       ` Daniel P. Berrange
2007-08-09 20:16       ` Avi Kivity
2007-08-09 20:25         ` Anthony Liguori
2007-08-09 20:30           ` Avi Kivity
2007-08-09 20:32             ` Anthony Liguori
2007-08-09 20:39               ` Avi Kivity
2007-08-09 20:44               ` Brian Wheeler
2007-08-09 20:49                 ` Anthony Liguori
2007-08-10  3:51                 ` dmc
2007-08-10 13:26                   ` Carlos A. M. dos Santos
2007-08-09 20:55         ` Brian Wheeler
2007-08-10  0:48           ` Anthony Liguori
2007-08-11 17:11   ` andrzej zaborowski
2007-08-11 18:06     ` Philip Boulain
2007-08-11 19:08       ` Christian Brunschen
2007-08-11 19:53         ` Anthony Liguori
2007-08-11 19:52       ` Anthony Liguori
2007-08-11 21:28         ` Philip Boulain
2007-08-11 23:17           ` Anthony Liguori
2007-08-13  5:34             ` Jorge Lucángeli Obes
2007-08-13 15:15               ` [kvm-devel] " Anthony Liguori
2007-08-11 19:49     ` Anthony Liguori [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=46BE12B9.1000203@codemonkey.ws \
    --to=anthony@codemonkey.ws \
    --cc=berrange@redhat.com \
    --cc=qemu-devel@nongnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).