Hi Gary,
the options.c file is generated from the gcc/common.opt file. Report was a
keyword that could be given to optimization options but which was dropped
sometime in December I think. The patches I sent you should have the
keyword Report dropped. Are you applying your sources already? If not,
Hello,
I'm currently working on improving my understanding of the implementation
of the intraprocedural points-to analysis in GCC. I have already read the
papers by Daniel Berlin and have been looking at the source for some time,
but I understand that those references are a little bit old and migh
Hello,
I'm still trying to compare the solution generated from the
intraprocedural points-to analysis in GCC against an external solver.
Yesterday it was pointed out that "NULL is not conservatively
correctly represented in the constraints". Can someone expand on this?
To me this sounds like a co
Hi Richard, I think I misunderstood yesterday's answer and deviated a
little bit. But now I want to focus on this:
> > * the process in GCC that generates the constraints for NULL works
> > fine (i.e., feeding the constraints generated by GCC to an external
> > solver should yield a conservatively
Mm... ignore this for now please, I think I messed up the analysis by
hand. I will try again. Thanks!
On Wed, 17 Mar 2021 at 16:16, Erick Ochoa wrote:
>
> Hi Richard, I think I misunderstood yesterday's answer and deviated a
> little bit. But now I want to focus on this:
>
> > > * the process in
No, I think it is correct. Any help would be clearly appreciated.
Thanks! (Even if I'm wrong, of course I'd like to know.)
On Wed, 17 Mar 2021 at 17:21, Erick Ochoa wrote:
>
> Mm... ignore this for now please, I think I messed up the analysis by
> hand. I will try again. Thanks!
>
> On Wed, 17 Ma
Perfect, thank you! This is exactly the answer that I was looking for.
On Thu, 18 Mar 2021 at 13:27, Richard Biener wrote:
>
> On Wed, Mar 17, 2021 at 4:17 PM Erick Ochoa wrote:
> >
> > Hi Richard, I think I misunderstood yesterday's answer and deviated a
> > little bit. But now I want to focus
Hello Richard,
I have a related question. I have the following GCC program (which is
again an altered version of ipa-pta-16.c:
```
struct X
{
long l1;
struct Y
{
long l2;
int *p;
} y;
};
int i;
static int __attribute__((noinline))
foo (struct X *x)
{
struct Y y = x->y;
Hello,
I already have some experience developing SIMPLE_IPA_PASSes, but I am
looking to understand IPA_PASSes better. I have made a hello world ipa
pass that stores "hello world $FUNCTION_NAME" in the function
summaries; however, I am having trouble reading this information back.
Can someone help
Hello,
just trying again to increase visibility of this question. Many thanks
in advance!
On Fri, 26 Mar 2021 at 13:49, Erick Ochoa wrote:
>
> Hello,
>
> I already have some experience developing SIMPLE_IPA_PASSes, but I am
> looking to understand IPA_PASSes better. I have made a hello world ip
Hi,
I am looking at the points-to analysis in GCC and I found the
following comment in tree-ssa-structalias.c:
/* Collect field information. */
if (use_field_sensitive
&& var_can_have_subvars (decl)
/* ??? Force us to not use subfields for globals in IPA mode.
Else we'd
> If the global is module local we should initialize it with NULL, yes. If it
> is
> not module local it should be initialized with NONLOCAL (that's both what
> should currently happen correctly - it's needed for non-field-sensitive init
> as well).
>
Awesome, thanks Richard! One more question:
Hello,
just trying one more time. Any direction or hint is appreciated. Just
as a note, I looked at the ipa-devirt as an inspiration for these
small functions I wrote, but for some reason I can't read what I
assume have written to the summaries. (What I'm trying to do is simply
use LTO using a rea
Hi,
just a high level question. I know that IPA-PTA is a SIMPLE_IPA_PASS
and that ideally it would be better as an IPA_PASS. I understand that
one of the biggest challenges of changing IPA-PTA to an IPA_PASS is
that on the current LTO framework, the LTRANS stage can happen at the
same time for mul
>
> I don't think this would remove any problem that is present.
>
I have a problem understanding what you mean here because later on you state:
> Now - the reason you think of is likely that IPA transform will instantiate
> IPA clones and do inlining and transfering the IPA PTA solution to the
>
pa_free_lang_data (gcc::context *ctxt);
extern simple_ipa_opt_pass *make_pass_ipa_free_fn_summary (gcc::context *ctxt);
+extern ipa_opt_pass_d *make_pass_ipa_hello_world (gcc::context *ctxt);
extern ipa_opt_pass_d *make_pass_ipa_cp (gcc::context *ctxt);
extern ipa_opt_pass_d *make_pass_ipa_sra
Hi,
I am looking for a small clarification. I understand that during late
SIMPLE_IPA_PASSes some statically initialized global variables might
have error_mark_node trees in their DECL_INITIAL field.
I believe that I read something similar in the past about how to get
the tree expressions in these
Hi,
I am having some trouble understanding the semantics of OBJ_TYPE_REF.
I understand that it is made of three operands:
1. OBJ_TYPE_REF_EXPR: An expression that evaluates the value to use.
2. OBJ_TYPE_REF_OBJECT: Is the object on whose behalf the lookup is
being performed
3. OBJ_TYPE_REF_TOKEN:
Cool, thanks! I think I understand now.
On Tue, 22 Jun 2021 at 23:58, Martin Jambor wrote:
>
> Hi,
>
> On Fri, Jun 18 2021, Erick Ochoa via Gcc wrote:
> > Hi,
> >
> > I am having some trouble understanding the semantics of OBJ_TYPE_REF.
> > I understa
Hello,
I know that some BUILT_IN functions are treated in a special way by
the points-to analysis. Those functions are those that take pointers
as arguments or return them but do not change their points-to set and
similar cases. (E.g. strcpy returns a pointer to the same object as
their first argu
> I guess that to assume otherwise, one would have to make sure the
> pointer does not correspond to a "%n" (or similar, perhaps even future)
> conversion specifier.
>
Oh, wow, I didn't know about the %n specifier.
Thanks Martin!
Hi Liu,
thanks, this also seems correct. I originally thought that (in your
example) ptr would be marked as pointing to { ESCAPE NONLOCAL } and
that would be enough to hinder some optimizations that might influence
variable value. Therefore, it wasn't necessary to mark "value" as
pointing to { ESC
Hello,
I'm trying to generate unique identifiers for some trees at link time.
I understand that there are already some unique identifiers in
declarations (DECL_UID) and perhaps others. Do all trees have unique
identifiers or only declarations?
Alternatively, if they don't have unique identifiers,
Hello,
I am wondering if there's a way to get the global namespace at LTO
during LGEN in each of the partitions being processed. I believe that
during parse time for c/c++ there is a global_namespace macro or
variable that can be used, but I don't think that it is possible to
use at link time.
Th
Hi,
I am still working on understanding the LTO framework and how the
gimple representation works across multiple partitions. I found myself
printing all global variables and printing their DECL_UID. I noticed
that for some variables, their DECL_UIDs were different across
different partitions. Tha
On Wed, 30 Jun 2021 at 17:06, Richard Biener wrote:
>
> On June 30, 2021 4:07:00 PM GMT+02:00, Erick Ochoa via Gcc
> wrote:
> >Hi,
> >
> >I am still working on understanding the LTO framework and how the
> >gimple representation works across multiple partition
Hello,
I have a function that looks at identifiers of formal parameters. I
found that when I ran this function on larger projects, the compiler
segfaulted. I looked into this and I found that there are some formal
parameters which have NULL in their DECL_NAME and that triggered a
NULL dereference.
Hi,
I am saving some tree declarations during LGEN that I will be later
analyzing at WPA time. I am able to read the decl from my summaries
and print it at WPA time. It corresponds to a global variable.
However, whenever I use symtab_node::get (decl) during WPA time I keep
getting NULL.
Does anyo
Hi, I noticed this is also happening also for local variables. Again,
storing tree declarations on a summary during LGEN and then at WPA
time reading from those summaries. I can print the declaration, but
when I try to look for its node in the symtab I get NULL as the return
value.
Any help is app
> I'm not too familiar with it but I think you're supposed to stream encoded
> symtab references during LGEN/WPA,
Thanks Richard, this happened to be the solution. I am now using
lto_symtab_encoder_t to encode the declarations during LGEN and decode
them during WPA.
Are there any more limitations
Hi,
Just to clarify a similar question: I am using stream_write_tree and
looking at the comments it says that it is assumed that the tree T is
already in the encoder cache. Does this mean that I have to use
lto_symtab_encoder_t for all trees I want to store in summaries? I
thought the encoder only
> There are entities, like SSA names and STRING_CSTs which are specially
> encoded and if you stream those in your LGEN data you have to set up
> appropriate encoders. In general streaming arbitrary trees isn't the
> best thing to do, usually you're interested in specific pieces only. That's
> es
On Tue, 13 Jul 2021 at 11:41, Richard Biener wrote:
> There are entities, like SSA names and STRING_CSTs which are specially
> encoded and if you stream those in your LGEN data you have to set up
> appropriate encoders.
I forgot to ask, is there an example of these appropriate encoders
being use
> I guess the way to encode SSA trees would be to use sth like a
> , SSA-version tuple much like PTA internally
> uses the varinfo array index as identifier for the variables in the
> constraints. For local decls (as opposed to SSA names) it's a bit
> more difficult - you'd have to devise your own
Hi,
I was sorting SSA trees and DECL_P trees into different buckets. I
used DECL_P as a proxy for it being a local/global variable/function
(essentially a declaration). It seems that "probabilistically", I'm
kinda right. Normally there is no tree that is both an SSA_VAR_P and a
DECL_P at the same
Thanks! I'll have to see what conditions led to everything working
before and change them to meet reality.
On Thu, 15 Jul 2021 at 10:35, Richard Biener wrote:
>
> On Thu, Jul 15, 2021 at 10:22 AM Erick Ochoa via Gcc wrote:
> >
> > Hi,
> >
> > I was sort
Hello Richard, I need a little bit more help. In our previous messages
you mentioned ""
> >
> > > I guess the way to encode SSA trees would be to use sth like a
> > > , SSA-version tuple much like PTA internally
> > > uses the varinfo array index as identifier for the variables in the
> > > constr
> > If this is the case, I can indeed get the varpool node's at WPA time
> > (as shown above), but comparing their pointer addresses will be
> > distinct. How can one find out that two varpool nodes/cgraph nodes are
> > the same at WPA time? Is just looking at the assembler name enough? I
> > of co
>
> fopen $PID1 8 $ADDR1
> fopen $PID2 7 $ADDR2
>
Just to clarify a bit further. $PID is generated and stored during
LGEN. The encoding is obviously generated during LGEN.
These are read during WPA. And the encoding is decoded and dyn_casted
into a cgraph_node at WPA time.
All these are printed du
> > > 1. pid of lgen process that generated the encoding
> > > 2. index returned by lto_symtab_encoder_encode
> > > 3. varpool_node->name ()
> > > 4. the pointer address being pointed by varpool node
>
> Well, yes, during LGEN no WPA has run. Do you mean LTRANS after WPA?
> Sure, the encoder numbe
>
> But the addresses are at LGEN time?
The following is what runs at WPA time
unsigned long pid = streamer_read_uhwi (&ib);
unsigned long id = streamer_read_uhwi (&ib);
lto_symtab_encoder_t encoder = file_data->symtab_node_encoder;
cgraph_node *cnode =
dyn_cast(lto_symtab_encoder_deref(encoder,
Hello,
I've been working on an LTO points-to analysis pass for a little
while. Because of LTO's design, gimple bodies are inaccessible during
WPA. This essentially means that every LTO pass compiles down function
bodies into their own IR which gets stored in function summaries and
later read durin
Hello Richard,
I'm still working on the points-to analysis and I am happy to say that
after reviewing the ipa-cp code I was able to generate summaries for
local variables, ssa variables, heap variables, global variables and
functions. I am also using the callback hooks to find out if
cgraph_nodes
Hello,
I am refactoring some old code that runs as an IPA_PASS. This code is
intended to run at link-time using the LTO framework and so I am
always testing it that way. At the moment, I am trying to use
hash-maps but having some difficulties. I am adding some patches that
will help illustrate alo
Hi,
just as a note. This is also of interest to me. I have wanted to
compile a single pass that I wrote using ubsan/other sanitizers for
testing purposes. I was wondering if someone has already modified the
build system to use ubsan to test their passes and if they could
document the process for d
Hi,
I have an LTO pass which stores information collected during "generate
function summary" in a map which is symtab_node* -> data*. I know that
the symtab_node*s are encoded by an lto encoder and can be decoded
back during the "read function summary". I also am aware that other
optimizations mig
Hi,
My current understanding of LTO is that the callgraph is very dynamic
(i.e., optimizations might add or remove cgraph_nodes). A while back I
encountered an issue where I couldn't print the cgraph_node::name of a
function during the execute stage in LTO. I found that the only thing
different wa
Hello,
I have a SIMPLE_IPA_PASS that parses the program multiple times. As it
parses gimple, it builds a data structure with the information
collected that will provide some invariants to future iterations over
the program.
I was looking into adding a new feature that would take advantage of
devi
Hi,
I have been adding tests to the gcc/testsuite/gcc.dg/ipa folder
successfully for a while now. I am starting to add some tests into
gcc/testsuite/g++.dg/ipa now but I am having some issues.
1. Using `make check-g++` returns the following error message "No rule
to make target 'check-g++'".
2. W
Hi,
I am looking at tree-ssa-structalias.c looking at what makes a
function nonlocal during IPA-PTA. I am having some problems
understanding force_output and when it is set or unset.
1. What is the meaning of force_output? cgraph.h gives an example that
force output means that the symbol might be
Hi,
I would like to use ipa_ref in the PASS_LIST all_late_ipa_passes to query
the statement (ref->stmt) of where a global variable is used. However, I am
having some problems achieving this.
What I do is:
1. Check that ipa_ref->referring has a body and is not inlined.
2. get_body
3. try to print
On Mon, 14 Feb 2022 at 10:57, Jan Hubicka wrote:
> > Hi,
> >
> > I would like to use ipa_ref in the PASS_LIST all_late_ipa_passes to query
> > the statement (ref->stmt) of where a global variable is used. However, I
> am
> > having some problems achieving this.
> >
> > What I do is:
> >
> > 1. Ch
Hello,
I'm trying to understand ipa-bit-cp/ipa-cp and how the known bits are
propagated to the lattice in the case of a pointer_plus_expr.
I have a piece of code similar to the following (this being a simplified
example)
int main ()
{
// a = some pointer.
foo (a);
}
foo (void* a)
{
bar (a
at the static information available at
the argument and not the parameter?
Thanks!
On Thu, 17 Feb 2022 at 17:25, Martin Jambor wrote:
> Hi,
>
> On Thu, Feb 17 2022, Erick Ochoa via Gcc wrote:
> > Hello,
> >
> > I'm trying to understand ipa-bit-cp/ipa-cp and how the
> If I understand you correctly, that is indeed the jump function,
> obtainable through ipa_get_ith_jump_func (args, i) where args is a
> result of ipa_edge_args_sum->get and i is the index of the parameter.
>
Thanks Martin!
So then, am I correct in thinking that
callsite foo -> bar
param 0: PAS
Hi,
I am working on an analysis that is able to determine some static
information about a specific variable. At the moment, I would like to avoid
much of the transformation by taking advantage of other GCC's passes. So, I
can imagine something like, create an ASSERT_EXPR and let other passes
chang
Hi,
I have one function (F) that has been specialized for two different calling
contexts (F1 and F2) and two late SIMPLE_IPA_PASSes (A and B). Pass A
changes some MEM_REFs such that the type of MEM_REF is compatible with the
type of the first operand of the expression. Pass A changes both F1 and F
Hi Martin!
Thanks for replying, turns out that while I was trying to reply to you I
was able to get the answer. Turns out there is indeed one tree node which
is shared across the two functions. And that is
TREE_OPERAND (MEM_REF, 1).
When I was assigning to
TREE_TYPE ( TREE_OPERAND (MEM_REF, 1)
Hi,
I am trying to find a map between cgraph_edge*s before ipa-cp and after
ipa-cp has specialized nodes. Does anyone know if such a map is available?
Or an equivalent? I think technically it should be a map between:
(cgraph_node* caller, cgraph_edge* e, cgraph_node *callee) X (cgraph_node*
calle
Hi,
I am interested in annotating INTEGER_CSTs. I have added a field to
typed_trees and set the value in some cases. However, INTEGER_CSTs can be
shared and copied across the function and even copied to other functions. I
don't want all INTEGER_CSTs with the same value to be shared as now they
hav
Hi,
I am trying to understand what path is executed in GCC from parsing a C
expression (in a global variable declaration) to the value in DECL_INITIAL.
At the moment, I have annotated a tree during parsing. I have another
debugging pass that looks for this tree in subsequent passes. The
annotation
> I'm not sure I understand but no pass walks 'global variables', so you're
> not
> going to "see" the annotation from passes (whatever that means).
>
>
What I mean by walking global variables is that I have a GIMPLE_PASS that
ignores the function sent as an argument and instead just uses a
FOR_EAC
Hi,
I'm working on a pass that looks into the estimated values during ipa-cp
and stores them for a later analyses/pass. I would like to store the real
arguments' estimates in a cgraph_edge::call_stmt or somewhere else that
makes similar sense. (Note, this is different from the formal parameters'
e
o update the cgraph_edge that disappears when dst->caller is
inlined to its caller?
* something else?
Any direction is greatly appreciated!
Many thanks!
-Erick
On Sat, 21 May 2022 at 00:13, Martin Jambor wrote:
> Hello,
>
> On Fri, May 20 2022, Erick Ochoa via Gcc wrote:
> > Hi
Hi,
I understand some differences between IPA_PASSes and SIMPLE_IPA_PASSes.
However, I have questions about the cleanup processes that IPA_PASSes might
have.
* Would it be correct to assume that all node and edge summaries are
deleted after the last IPA_PASS? And would it also be correct to assum
Hi,
I have a pass that is able to speculate the target of indirect function
calls. This pass is an IPA_PASS. It:
1. generates summaries with the possible targets.
2. writes analysis summary
3. reads analysis summary
4. combines the results from multiple partitions and if needed fixes the
targets
On Fri, 1 Jul 2022 at 14:48, Martin Jambor wrote:
> Why so late, why not as part of number 4?
>
Hi Martin,
Thanks for the feedback. The original reason why the call to
make_edge_direct_to_target was done so late is because of the lack of
function bodies during WPA, summaries with insufficient i
Hello,
I'm looking for some help in how to create a new function at compile time /
link time. The idea is an alternative form of constant propagation.
The current implementation of ipa-cp, may specialize functions for which
arguments may be known at compile time. Call graph edges from the caller
Hi Martin,
thanks a lot for your help! You were right! I am now able to call
make_edge_direct_to_target during WPA.
-Erick
Hi Richard,
>
> > So instead of wrapping the function why not transform the original
> function
> > to have a prologue doing a runtime check for the compile-time specialized
> > versions and perform tail-calls to them?
> >
> > What I'm missing is who would call mul_test_param in your case?
>
The
On Thu, 14 Jul 2022 at 15:51, Richard Biener
wrote:
> With a value-profile it would be per call site and IPA-CP can use that
> to direct the cloning. I'm not sure how many
> values a value histogram can track reliably here.
>
Last time I checked, value profiling can only track a single value pe
On Thu, 14 Jul 2022 at 16:10, Martin Liška wrote:
> On 7/14/22 16:08, Erick Ochoa via Gcc wrote:
> > Last time I checked, value profiling can only track a single value per
> > statement.
>
> Hi.
>
> Take a look at HIST_TYPE_INDIR_CALL which we use for tracki
Awesome! Thanks!
Let's go a little bit into the transformation mechanics itself. I am
somewhat familiar with LTO but I am always confused about the
transformations scheduled during WPA.
Let's assume that:
1. we have the profiling pass, which profiles each argument in each
callsite.
2. we are now
73 matches
Mail list logo
