public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
* [Bug c++/100884] New: Comparing unsigned 32 bit integer values generates 64 bit compare instructions when optimized
@ 2021-06-02 20:47 mfarazma.ext at gmail dot com
2021-06-17 18:50 ` [Bug target/100884] " mfarazma.ext at gmail dot com
2021-06-17 18:51 ` mfarazma.ext at gmail dot com
0 siblings, 2 replies; 3+ messages in thread
From: mfarazma.ext at gmail dot com @ 2021-06-02 20:47 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100884
Bug ID: 100884
Summary: Comparing unsigned 32 bit integer values generates 64
bit compare instructions when optimized
Product: gcc
Version: 10.2.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: c++
Assignee: unassigned at gcc dot gnu.org
Reporter: mfarazma.ext at gmail dot com
Target Milestone: ---
**0.c**
```
#include <iostream>
uint32_t codegen();
int main(){
uint32_t expected = (codegen()) != 0 ? -1 : -2;
bool check_eq = expected == codegen();
if (!(check_eq)){
std::cout << "Check failed" << std::endl;
}
return 0;
}
```
**1.cc**
```
#include <iostream>
uint32_t codegen(){
volatile int64_t a = -1;
return a;
}
```
Compile it with optimization enabled:
g++ -O3 0.cc 1.cc
Checking the generate code shows this instruction is generated for ` if
(!(check_eq)){...}`:
```
cmpd r31,r3
```
Compiling the same code with gcc version 8.4.0 emits this instead which is
correct:
```
cmpw cr7,r3,r31
```
This issue is causing failures as the returned value from `codegen()` may not
have its upper 32 bits cleared and using `cmpd` on it will create the wrong
output.
^ permalink raw reply [flat|nested] 3+ messages in thread
* [Bug target/100884] Comparing unsigned 32 bit integer values generates 64 bit compare instructions when optimized
2021-06-02 20:47 [Bug c++/100884] New: Comparing unsigned 32 bit integer values generates 64 bit compare instructions when optimized mfarazma.ext at gmail dot com
@ 2021-06-17 18:50 ` mfarazma.ext at gmail dot com
2021-06-17 18:51 ` mfarazma.ext at gmail dot com
1 sibling, 0 replies; 3+ messages in thread
From: mfarazma.ext at gmail dot com @ 2021-06-17 18:50 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100884
--- Comment #1 from m farazma <mfarazma.ext at gmail dot com> ---
The development team has mentioned this is correct behaviour, according to PPC
ABI the returning function needs to sign or zero extend the result when
necessary:
https://refspecs.linuxfoundation.org/ELF/ppc64/PPC-elf64abi.html#RETURN-VAL
Will be closing this issue.
^ permalink raw reply [flat|nested] 3+ messages in thread
* [Bug target/100884] Comparing unsigned 32 bit integer values generates 64 bit compare instructions when optimized
2021-06-02 20:47 [Bug c++/100884] New: Comparing unsigned 32 bit integer values generates 64 bit compare instructions when optimized mfarazma.ext at gmail dot com
2021-06-17 18:50 ` [Bug target/100884] " mfarazma.ext at gmail dot com
@ 2021-06-17 18:51 ` mfarazma.ext at gmail dot com
1 sibling, 0 replies; 3+ messages in thread
From: mfarazma.ext at gmail dot com @ 2021-06-17 18:51 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100884
m farazma <mfarazma.ext at gmail dot com> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|UNCONFIRMED |RESOLVED
Resolution|--- |INVALID
--- Comment #2 from m farazma <mfarazma.ext at gmail dot com> ---
Details are mentioned in the above comment.
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2021-06-17 18:51 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-06-02 20:47 [Bug c++/100884] New: Comparing unsigned 32 bit integer values generates 64 bit compare instructions when optimized mfarazma.ext at gmail dot com
2021-06-17 18:50 ` [Bug target/100884] " mfarazma.ext at gmail dot com
2021-06-17 18:51 ` mfarazma.ext at gmail dot com
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).