Re: [Xen-devel] [PATCH RFC 3/9] libxl: introduce specific xenstore error codes

[Rob Hoes <Rob.Hoes@xxxxxxxxxx>]

> On 26 Jun 2015, at 15:36, Rob Hoes <rob.hoes@xxxxxxxxxx> wrote:
> 
>> On 24 Jun 2015, at 16:10, Ian Jackson <ian.jackson@xxxxxxxxxxxxx> wrote:
>> 
>> Rob Hoes writes ("[PATCH RFC 3/9] libxl: introduce specific xenstore error 
>> codes"):
>>> Signed-off-by: Rob Hoes <rob.hoes@xxxxxxxxxx>
>> 
>> I don't understand the distinctions you are trying to make here.  Is
>> it really useful for an application or a sysadmin to distinguish what
>> operation was being attempted against xenstore ?
>> 
>> I think the xenstore error code would be more informative.  Probably
>> they can be categorised into:
>> * xenstored communication failed
>> * xenstored seems to have done something mad
>> * permission denied reading xenstore
>> * permission denied writing xenstore
>> * some other unexpected error happened talking to xenstore
> 
> Besides the connection problems, the possible error codes from xenstored seem 
> to be:
> * EINVAL: invalid input
> * EEXIST: tried to create something that already exists
> * ENOENT: what you asked for does not exist
> * EACCES: permission denied
> * EQUOTA: quota exceeded
> * E2BIG: data too big
> * ENOSYS: unknown operation
> * EAGAIN and EBUSY: transaction problemsâ please retry (?)
> 
> Should we just pass these on as ERROR_XS_EINVAL, ERROR_XS_EEXIST, etc.?

Of course these errors may occur somewhere deep down inside libxl, and might be 
translated into an error code that makes more sense in the context of the 
higher level operation that the user requested from libxl.

Rob
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel