qemu-devel
[Top][All Lists]
Advanced

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

Re: [Qemu-devel] [PATCH v7 RFC] block/vxhs: Initial commit to add Verita


From: Ketan Nilangekar
Subject: Re: [Qemu-devel] [PATCH v7 RFC] block/vxhs: Initial commit to add Veritas HyperScale VxHS block device support
Date: Wed, 19 Oct 2016 20:01:36 +0000
User-agent: Microsoft-MacOutlook/0.0.0.160109

Hi Jeff,

Please see my comments inline.

Thanks,
Ketan.





On 10/18/16, 12:10 PM, "Jeff Cody" <address@hidden> wrote:

>On Tue, Oct 11, 2016 at 12:56:06AM -0700, ashish mittal wrote:
>> Checked in a test server to libqnio that allows to open, read, write
>> to a vxhs vdisk using the qemu-io binary.
>> 
>> Steps to run the test server:
>> (1) Touch a file in /tmp with the base-name of the vdisk ID to be opened.
>> touch /tmp/\{98f48eef-b62f-46ee-b6e3-ad48ffd9ad0a\}
>> (2) Start the test server. It listens on port 9999 by default.
>> /path/to/qnio_server
>> (3) In another terminal, start the qemu-io binary and open the vdisk
>> qemu-io> open 
>> vxhs://127.0.0.1:9999/%7B98f48eef-b62f-46ee-b6e3-ad48ffd9ad0a%7D
>> (4) Now you can write and read data from the vdisk that is backed by a file.
>> qemu-io> writev -P 81 0 1k
>> qemu-io> read -v 0 1k
>> 
>> Following change would be needed to my last patch to allow opening of the 
>> vdisk:
>> $ git diff
>> diff --git a/block/vxhs.c b/block/vxhs.c
>> index 90a4343..d849a9b 100644
>> --- a/block/vxhs.c
>> +++ b/block/vxhs.c
>> @@ -1215,7 +1215,7 @@ static int vxhs_qemu_init(QDict *options,
>> BDRVVXHSState *s,
>>      }
>> 
>>      ret = vxhs_qnio_iio_open(cfd, of_vsa_addr, rfd, file_name);
>> -    if (!ret) {
>> +    if (ret) {
>>          error_setg(&local_err, "Failed qnio_iio_open");
>>          ret = -EIO;
>>      }
>> 
>> Will work on the qemu-iotests test suite next.
>> 
>> Regards,
>> Ashish
>> 
>
>Hi Ashish,
>
>I wanted to check and see how the qemu-iotest suite was coming along.
>
>A couple of things on the test server:
>
>It is hardcoded to only accept 4MB images.  Is it possible to relax that
>restriction (is a restriction even needed)?

We are removing this restriction from the QEMU driver and qnio. 

>
>Also, the test server is hardcoded to use the path "/tmp".   This needs to
>be passed in as an argument, as qemu-iotests will want to run everything
>from its scratch test directory.

This is being worked upon and will be part of the next checkin.

>
>A suggestion: it would also be a good idea, if possible, to enable all the
>supported ioctl commands in the test server.

We will look into this.

>
>Jeff

reply via email to

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