[PATCH] OpenCL: Add new types for OpenCL 2.0

Pedro Ferreira arkangath at gmail.com
Thu Jun 11 03:45:53 PDT 2015

Actually, I spoke too soon - I found a test with -cl-std=CL2.0. I missed

On Thu, 11 Jun 2015 at 11:40 Pedro Ferreira <arkangath at gmail.com> wrote:

> The codegen test would imply adding a -cl-std=2.0 option to Clang, which
> it currently does not have. This is because the types should only be
> recognised if the CL 2.0 standard is explicitly asked for (the default is
> to operate on 1.2 mode). Adding that option is a peripheral issue. I've
> added the types on the header test under the appropriate "#if defined" but
> when I tried to do the same on the .cl file, I found out that the test
> parser does not recognise the preprocessor macro and therefore was causing
> the test to (incorrectly) fail. As such, I reverted the test.
> As for the AS for the other types, I copy-pasted the code from event_t.
> That's the reason why I'm actually using the "0". Are you suggesting I
> should change event_t to use something else, and by consequence the new
> types too? That would be a separate issue.
> My guess is that these types are allocated on the stack, which by llvm
> convention will always be 0.
> The new types are used by new builtins. I don't think there are any other
> special semantics to it.
> I've added extension checks on the MSAA types, but I'm not sure if this is
> the right place. New patch attached.
> Pedro
> On Thu, 11 Jun 2015 at 10:33 Anastasia Stulova <Anastasia.Stulova at arm.com>
> wrote:
>> Hi Pedro,
>> Could we also add a Codegen test? Also it would be better not to use
>> constant directly as address space as the mapping could ideally be changed.
>> Is there any reason why you generate pointers to private AS?
>> Are there any operations allowed on new types? Any semantical checks
>> needed?
>> If MSAA types are part of an extension and not a part of the general
>> standard we should ideally diagnose that extension is enabled when they are
>> being used.
>> Regards,
>> Anastasia
>> ________________________________________
>> From: cfe-commits-bounces at cs.uiuc.edu [cfe-commits-bounces at cs.uiuc.edu]
>> On Behalf Of Pedro Ferreira [arkangath at gmail.com]
>> Sent: Thursday, June 11, 2015 8:18 AM
>> To: cfe-commits at cs.uiuc.edu
>> Subject: [PATCH] OpenCL: Add new types for OpenCL 2.0
>> Hi all,
>> This patch adds the new OpenCL types for 2.0 described at
>> https://www.khronos.org/registry/cl/sdk/2.0/docs/man/xhtml/otherDataTypes.html
>> <
>> https://urldefense.proofpoint.com/v2/url?u=https-3A__www.khronos.org_registry_cl_sdk_2.0_docs_man_xhtml_otherDataTypes.html&d=AwMFaQ&c=8hUWFZcy2Z-Za5rBPlktOQ&r=BSqEv9KvKMW_Ob8SyngJ70KdZISM_ASROnREeq0cCxk&m=42YnWExwxwpeU6GPDY2_3RFxCqQakUbj_CXZsMsQ2jU&s=REOBNoaDio7qDyIDCqmXhxFvZYjMOK6vuXAttjOVsNI&e=
>> >
>> I also opened https://llvm.org/bugs/show_bug.cgi?id=23794<
>> https://urldefense.proofpoint.com/v2/url?u=https-3A__llvm.org_bugs_show-5Fbug.cgi-3Fid-3D23794&d=AwMFaQ&c=8hUWFZcy2Z-Za5rBPlktOQ&r=BSqEv9KvKMW_Ob8SyngJ70KdZISM_ASROnREeq0cCxk&m=42YnWExwxwpeU6GPDY2_3RFxCqQakUbj_CXZsMsQ2jU&s=TAV4suAMaHgdIPA83Da3pQl7c68On7bAFWtnrUbt_Uk&e=>
>> for this. I keep forgetting you prefer patches sent to this mailing list.
>> This also adds lldb entries (fixes switch warnings).
>> The types are:
>> image2d_depth_t
>> image2d_array_depth_t
>> image2d_msaa_t
>> image2d_array_msaa_t
>> image2d_msaa_depth_t
>> image2d_array_msaa_depth_t
>> queue_t
>> ndrange_t
>> clk_event_t
>> reserve_id_t
>> let me know if something looks wrong,
>> Pedro
>> -- IMPORTANT NOTICE: The contents of this email and any attachments are
>> confidential and may also be privileged. If you are not the intended
>> recipient, please notify the sender immediately and do not disclose the
>> contents to any other person, use it for any purpose, or store or copy the
>> information in any medium.  Thank you.
>> ARM Limited, Registered office 110 Fulbourn Road, Cambridge CB1 9NJ,
>> Registered in England & Wales, Company No:  2557590
>> ARM Holdings plc, Registered office 110 Fulbourn Road, Cambridge CB1 9NJ,
>> Registered in England & Wales, Company No:  2548782
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20150611/21df6c7d/attachment.html>

More information about the cfe-commits mailing list