llvm dev - May 2016 - RFC: metadata attachments for global variables

On Fri, May 6, 2016 at 2:09 PM, David Blaikie <dblaikie at gmail.com>
wrote:
>
>
> On Fri, May 6, 2016 at 1:55 PM, Peter Collingbourne via llvm-dev <
> llvm-dev at lists.llvm.org> wrote:
>
>>
>>
>> On Fri, May 6, 2016 at 1:21 PM, Duncan P. N. Exon Smith <
>> dexonsmith at apple.com> wrote:
>>
>>>
>>> > On 2016-May-06, at 13:17, Peter Collingbourne <peter at
pcc.me.uk> wrote:
>>> >
>>> > Hi all,
>>> >
>>> > I'd like to add support for metadata attachments for
global variables
>>> in the same way as we did for functions.
>>> >
>>> > Syntax would be pretty simple:
>>> > @foo = global i32 0, !foo !0, !bar !1
>>> > (the extra commas are required to disambiguate from a named
metadata
>>> on the next line)
>>>
>>> SGTM.
>>>
>>> > Benefits:
>>> > 1) Lets us reverse the DIGlobalVariable -> GlobalVariable
edge, which
>>> should hopefully clear the way for removing the llvm.dbg.cu named
>>> metadata node.
>>>
>>> A little harder than it sounds (need to somehow support a
GlobalVariable
>>> that is RAUW'ed with a ConstantInt), but I think this is
important to do.
>>>
>>
>> How can a GlobalVariable be replaced with a ConstantInt? Wouldn't
that
>> just be an absolute address?
>>
>
> If the global variable has a known constant value & the address is
never
> needed, then we might optimize away the storage and still want debug info
> describing the constant.
>
I'm not sure under what conditions, if any, that currently works - so this
may be more of a "nice to have" or "be good to think about how
any future
design wouldn't preclude fixing this gap" sort of thing.

- David

>
>
>>
>> Thanks,
>> --
>> --
>> Peter
>>
>> _______________________________________________
>> LLVM Developers mailing list
>> llvm-dev at lists.llvm.org
>> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20160506/074835fa/attachment.html>

On Fri, May 6, 2016 at 2:10 PM, David Blaikie <dblaikie at gmail.com>
wrote:
>
>
> On Fri, May 6, 2016 at 2:09 PM, David Blaikie <dblaikie at gmail.com>
wrote:
>
>>
>>
>> On Fri, May 6, 2016 at 1:55 PM, Peter Collingbourne via llvm-dev <
>> llvm-dev at lists.llvm.org> wrote:
>>
>>>
>>>
>>> On Fri, May 6, 2016 at 1:21 PM, Duncan P. N. Exon Smith <
>>> dexonsmith at apple.com> wrote:
>>>
>>>>
>>>> > On 2016-May-06, at 13:17, Peter Collingbourne <peter at
pcc.me.uk>
>>>> wrote:
>>>> >
>>>> > Hi all,
>>>> >
>>>> > I'd like to add support for metadata attachments for
global variables
>>>> in the same way as we did for functions.
>>>> >
>>>> > Syntax would be pretty simple:
>>>> > @foo = global i32 0, !foo !0, !bar !1
>>>> > (the extra commas are required to disambiguate from a
named metadata
>>>> on the next line)
>>>>
>>>> SGTM.
>>>>
>>>> > Benefits:
>>>> > 1) Lets us reverse the DIGlobalVariable ->
GlobalVariable edge, which
>>>> should hopefully clear the way for removing the llvm.dbg.cu
named
>>>> metadata node.
>>>>
>>>> A little harder than it sounds (need to somehow support a
>>>> GlobalVariable that is RAUW'ed with a ConstantInt), but I
think this is
>>>> important to do.
>>>>
>>>
>>> How can a GlobalVariable be replaced with a ConstantInt?
Wouldn't that
>>> just be an absolute address?
>>>
>>
>> If the global variable has a known constant value & the address is
never
>> needed, then we might optimize away the storage and still want debug
info
>> describing the constant.
>>
>
> I'm not sure under what conditions, if any, that currently works - so
this
> may be more of a "nice to have" or "be good to think about
how any future
> design wouldn't preclude fixing this gap" sort of thing.
>
I see. This does seem a lot like something like macro expansions where the
entity doesn't have a direct correspondent in the IR. So perhaps any pass
that would want to preserve debug info in such a case could call some
utility function that would move the DIGlobalVariable to (I suppose it
would have to be) some named metadata node to which we'd attach things like
macro expansions.

Thanks,
-- 
-- 
Peter
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20160506/27c3972d/attachment.html>

On Fri, May 6, 2016 at 2:14 PM, Peter Collingbourne <peter at pcc.me.uk>
wrote:
>
>
> On Fri, May 6, 2016 at 2:10 PM, David Blaikie <dblaikie at gmail.com>
wrote:
>
>>
>>
>> On Fri, May 6, 2016 at 2:09 PM, David Blaikie <dblaikie at
gmail.com> wrote:
>>
>>>
>>>
>>> On Fri, May 6, 2016 at 1:55 PM, Peter Collingbourne via llvm-dev
<
>>> llvm-dev at lists.llvm.org> wrote:
>>>
>>>>
>>>>
>>>> On Fri, May 6, 2016 at 1:21 PM, Duncan P. N. Exon Smith <
>>>> dexonsmith at apple.com> wrote:
>>>>
>>>>>
>>>>> > On 2016-May-06, at 13:17, Peter Collingbourne
<peter at pcc.me.uk>
>>>>> wrote:
>>>>> >
>>>>> > Hi all,
>>>>> >
>>>>> > I'd like to add support for metadata attachments
for global
>>>>> variables in the same way as we did for functions.
>>>>> >
>>>>> > Syntax would be pretty simple:
>>>>> > @foo = global i32 0, !foo !0, !bar !1
>>>>> > (the extra commas are required to disambiguate from a
named metadata
>>>>> on the next line)
>>>>>
>>>>> SGTM.
>>>>>
>>>>> > Benefits:
>>>>> > 1) Lets us reverse the DIGlobalVariable ->
GlobalVariable edge,
>>>>> which should hopefully clear the way for removing the
llvm.dbg.cu
>>>>> named metadata node.
>>>>>
>>>>> A little harder than it sounds (need to somehow support a
>>>>> GlobalVariable that is RAUW'ed with a ConstantInt), but
I think this is
>>>>> important to do.
>>>>>
>>>>
>>>> How can a GlobalVariable be replaced with a ConstantInt?
Wouldn't that
>>>> just be an absolute address?
>>>>
>>>
>>> If the global variable has a known constant value & the address
is never
>>> needed, then we might optimize away the storage and still want
debug info
>>> describing the constant.
>>>
>>
>> I'm not sure under what conditions, if any, that currently works -
so
>> this may be more of a "nice to have" or "be good to
think about how any
>> future design wouldn't preclude fixing this gap" sort of
thing.
>>
>
> I see. This does seem a lot like something like macro expansions where the
> entity doesn't have a direct correspondent in the IR.
>
Ish - macros disappear in the frontend, so, yes, when we're preserving them
we just attach them to the CU.

> So perhaps any pass that would want to preserve debug info in such a case
> could call some utility function that would move the DIGlobalVariable to (I
> suppose it would have to be) some named metadata node to which we'd
attach
> things like macro expansions.
>
Yep, something like that. But many passes might just blindly RAUW loads of
the pointer to the global with the constant & then the global just quietly
disappears due to lack of use (probably some cleanup path). But this seems
to already be happening today - Duncan, do you know any cases where we do
successfully describe a global via a constant value that would be the value
of the global, not its address? My simple cases just seemed to lose the
location entirely if the constant's storage was optimized away.

- David

>
> Thanks,
> --
> --
> Peter
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20160506/69c17834/attachment.html>

On Fri, May 6, 2016 at 2:10 PM, David Blaikie <dblaikie at gmail.com>
wrote:
>
>
> On Fri, May 6, 2016 at 2:09 PM, David Blaikie <dblaikie at gmail.com>
wrote:
>
>>
>>
>> On Fri, May 6, 2016 at 1:55 PM, Peter Collingbourne via llvm-dev <
>> llvm-dev at lists.llvm.org> wrote:
>>
>>>
>>>
>>> On Fri, May 6, 2016 at 1:21 PM, Duncan P. N. Exon Smith <
>>> dexonsmith at apple.com> wrote:
>>>
>>>>
>>>> > On 2016-May-06, at 13:17, Peter Collingbourne <peter at
pcc.me.uk>
>>>> wrote:
>>>> >
>>>> > Hi all,
>>>> >
>>>> > I'd like to add support for metadata attachments for
global variables
>>>> in the same way as we did for functions.
>>>> >
>>>> > Syntax would be pretty simple:
>>>> > @foo = global i32 0, !foo !0, !bar !1
>>>> > (the extra commas are required to disambiguate from a
named metadata
>>>> on the next line)
>>>>
>>>> SGTM.
>>>>
>>>> > Benefits:
>>>> > 1) Lets us reverse the DIGlobalVariable ->
GlobalVariable edge, which
>>>> should hopefully clear the way for removing the llvm.dbg.cu
named
>>>> metadata node.
>>>>
>>>> A little harder than it sounds (need to somehow support a
>>>> GlobalVariable that is RAUW'ed with a ConstantInt), but I
think this is
>>>> important to do.
>>>>
>>>
>>> How can a GlobalVariable be replaced with a ConstantInt?
Wouldn't that
>>> just be an absolute address?
>>>
>>
>> If the global variable has a known constant value & the address is
never
>> needed, then we might optimize away the storage and still want debug
info
>> describing the constant.
>>
>
> I'm not sure under what conditions, if any, that currently works - so
this
> may be more of a "nice to have" or "be good to think about
how any future
> design wouldn't preclude fixing this gap" sort of thing.
>
It looks like this is how debug info for static data members used to look
until r172591 landed back in 2013. There are a couple of tests that
exercise this path (DebugInfo/X86/multiple-at-const-val.ll
and DebugInfo/X86/pr12831.ll) but it seems likely that they were just
upgraded manually.

Do we care about backwards compatibility for debug info from that far back?
I assume we don't, since this predates the last debug info version bump in
2015 (r231082 which introduced the debug info hierarchy). If that's the
case, I intend to "upgrade" this metadata by dropping it, and we can
probably think of some other representation later for a constant folded
global.

Thanks,
-- 
-- 
Peter
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20160509/d528725d/attachment.html>

On Mon, May 9, 2016 at 1:53 PM, Peter Collingbourne <peter at pcc.me.uk>
wrote:
>
>
> On Fri, May 6, 2016 at 2:10 PM, David Blaikie <dblaikie at gmail.com>
wrote:
>
>>
>>
>> On Fri, May 6, 2016 at 2:09 PM, David Blaikie <dblaikie at
gmail.com> wrote:
>>
>>>
>>>
>>> On Fri, May 6, 2016 at 1:55 PM, Peter Collingbourne via llvm-dev
<
>>> llvm-dev at lists.llvm.org> wrote:
>>>
>>>>
>>>>
>>>> On Fri, May 6, 2016 at 1:21 PM, Duncan P. N. Exon Smith <
>>>> dexonsmith at apple.com> wrote:
>>>>
>>>>>
>>>>> > On 2016-May-06, at 13:17, Peter Collingbourne
<peter at pcc.me.uk>
>>>>> wrote:
>>>>> >
>>>>> > Hi all,
>>>>> >
>>>>> > I'd like to add support for metadata attachments
for global
>>>>> variables in the same way as we did for functions.
>>>>> >
>>>>> > Syntax would be pretty simple:
>>>>> > @foo = global i32 0, !foo !0, !bar !1
>>>>> > (the extra commas are required to disambiguate from a
named metadata
>>>>> on the next line)
>>>>>
>>>>> SGTM.
>>>>>
>>>>> > Benefits:
>>>>> > 1) Lets us reverse the DIGlobalVariable ->
GlobalVariable edge,
>>>>> which should hopefully clear the way for removing the
llvm.dbg.cu
>>>>> named metadata node.
>>>>>
>>>>> A little harder than it sounds (need to somehow support a
>>>>> GlobalVariable that is RAUW'ed with a ConstantInt), but
I think this is
>>>>> important to do.
>>>>>
>>>>
>>>> How can a GlobalVariable be replaced with a ConstantInt?
Wouldn't that
>>>> just be an absolute address?
>>>>
>>>
>>> If the global variable has a known constant value & the address
is never
>>> needed, then we might optimize away the storage and still want
debug info
>>> describing the constant.
>>>
>>
>> I'm not sure under what conditions, if any, that currently works -
so
>> this may be more of a "nice to have" or "be good to
think about how any
>> future design wouldn't preclude fixing this gap" sort of
thing.
>>
>
> It looks like this is how debug info for static data members used to look
> until r172591 landed back in 2013.
>
Could you explain more what you mean by "<this> is how it worked
until" -
'this' being how it used to work and how you're proposing to make it
work?
I'm not quite following.

> There are a couple of tests that exercise this path
> (DebugInfo/X86/multiple-at-const-val.ll and DebugInfo/X86/pr12831.ll) but
> it seems likely that they were just upgraded manually.
>
> Do we care about backwards compatibility for debug info from that far back?
>
Yeah, pretty sure we don't care.

> I assume we don't, since this predates the last debug info version bump
in
> 2015 (r231082 which introduced the debug info hierarchy). If that's the
> case, I intend to "upgrade" this metadata by dropping it, and we
can
> probably think of some other representation later for a constant folded
> global.
>
Upgrades to the verifier to fail when it sees this would be good too, most
likely. (maybe? maybe not... not sure)

>
> Thanks,
> --
> --
> Peter
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20160509/1c940685/attachment.html>