[Xcb] [PATCH] proto: experimental Xi and SELinux extensions

Eamon Walsh ewalsh at tycho.nsa.gov
Wed Dec 19 14:46:53 PST 2007


Eamon Walsh wrote:
> Barton C Massey wrote:
>   
>> In message <4759BE79.3050709 at tycho.nsa.gov> you wrote:
>>   
>>     
>>> I pushed the autotools patch.  I would like to push XInput and SELinux 
>>> extensions upstream, disabled by default.  I am still testing XInput so 
>>> I have marked it experimental in the configure option and pc.in file.  
>>> Please consider for upstream, this is an important piece of the puzzle 
>>> for people experimenting with SELinux support for X.
>>>     
>>>       
>> We're excited to have them---thanks much!  I suspect Jamey
>> or someone will want to review them carefully, but I don't
>> see any reason why they shouldn't be in mainline soon.
>>     

I'm done with testing and I believe the XInput part is ready to go.  The 
revised xinput.xml is attached.  There are 2 requests and 5 replies that 
I could not figure out how to represent in the current schema.  A 
description of each is below.  What I mean by "variable-sized structure" 
is that there are several possible structures that may be present.  The 
structures all have a type code and length as the first two fields, but 
the contents after that vary.  It's basically the "struct sockaddr" model.

I have commented out the two offending requests.   So I think this can 
be upstreamed while a way is found to represent the troublesome items.  
I'm not sure how to do that but it's going to require the Python parser, 
if not hand-coded C.

Comments?


ListInputDevices:

The reply contains a list of fixed-size structures A, followed by a list 
of variable-sized structures B, followed by a list of NUL-terminated 
strings C.  Each element of A is associated with some of the items in B 
and one string in C; to find the length of B, each element of A must be 
examined and the total length calculated by adding up the individual 
lengths in each A structure.  The length of C is equivalent to the 
length of A.  It should be noted this differs from the wire protocol 
given in the spec.


GetDeviceMotionEvents:

The reply contains two length fields m and n, followed by a list of 
structures A of length n.  Each member of A has one CARD32 followed a 
list of CARD32's of length m.


GetFeedbackControl:
QueryDeviceState:

The reply contains a list of variable-sized structures.


ChangeFeedbackControl:
ChangeDeviceControl

The request contains a variable-sized structure.


GetDeviceControl:

The reply contains a variable-sized structure.


-- 
Eamon Walsh <ewalsh at tycho.nsa.gov>
National Security Agency

-------------- next part --------------
A non-text attachment was scrubbed...
Name: xinput.xml
Type: text/xml
Size: 29382 bytes
Desc: not available
Url : http://lists.freedesktop.org/archives/xcb/attachments/20071219/4191d8fb/attachment-0001.bin 


More information about the Xcb mailing list