Re: [Tinycc-devel] A double-to-float conversion bug
On 7/31/23 12:43, grischka wrote: On 31.07.2023 13:42, Herman ten Brugge via Tinycc-devel wrote: I agree with your comments above. The size is incorrect. I could change gfunc_sret in x86_64-gen.c and then calculate the size in tccgen.c as you suggested. But I am not sure regsize is set correctly all the time. I like this better: --- a/tccgen.c +++ b/tccgen.c @@ -6142,7 +6142,7 @@ special_math_val: space. Assume register size is power of 2. */ if (regsize > align) align = regsize; - loc &= -align; + size = (size + regsize - 1) & -regsize; loc = (loc - size) & -align; addr = loc; offset = 0; The size should be a multiple of regsize. What is your opinion? Hi, yes, that seems to make sense. Then again I think that align should additionally depend on just 'ret_align', as in if (ret_align > align) align = ret_align; I was working on a better fix for riscv and reverted the patch from Yao Zi by accident. I will reapply that change. Well, as I was trying to say, that change wasn't a good one either. Anyway, I just applied some more extensive fix for both. Hope it does work now. Also yet another pragma_once_normalize patch (which you may or may not want to comment ;) I tested all changes and all work on multiple targets. Great work. I have no comments on the pragma once patch. Sorry Herman ___ Tinycc-devel mailing list Tinycc-devel@nongnu.org https://lists.nongnu.org/mailman/listinfo/tinycc-devel
Re: [Tinycc-devel] A double-to-float conversion bug
On 31.07.2023 13:42, Herman ten Brugge via Tinycc-devel wrote: I agree with your comments above. The size is incorrect. I could change gfunc_sret in x86_64-gen.c and then calculate the size in tccgen.c as you suggested. But I am not sure regsize is set correctly all the time. I like this better: --- a/tccgen.c +++ b/tccgen.c @@ -6142,7 +6142,7 @@ special_math_val: space. Assume register size is power of 2. */ if (regsize > align) align = regsize; - loc &= -align; + size = (size + regsize - 1) & -regsize; loc = (loc - size) & -align; addr = loc; offset = 0; The size should be a multiple of regsize. What is your opinion? Hi, yes, that seems to make sense. Then again I think that align should additionally depend on just 'ret_align', as in if (ret_align > align) align = ret_align; I was working on a better fix for riscv and reverted the patch from Yao Zi by accident. I will reapply that change. Well, as I was trying to say, that change wasn't a good one either. Anyway, I just applied some more extensive fix for both. Hope it does work now. Also yet another pragma_once_normalize patch (which you may or may not want to comment ;) -- gr Herman ___ Tinycc-devel mailing list Tinycc-devel@nongnu.org https://lists.nongnu.org/mailman/listinfo/tinycc-devel
Re: [Tinycc-devel] A double-to-float conversion bug
On 7/31/23 10:08, grischka wrote:
On 31.07.2023 09:32, Herman ten Brugge via Tinycc-devel wrote:
On 7/30/23 18:27, Vincent Lefevre wrote:
On 2023-07-30 16:07:29 +0600, Viktor M wrote:
Hello everyone. So today I stumbled upon this bug when doing math
involving conversions between float and double. A minimal example:
[...]
Not really minimal. I could simplify it even further:
I fixed this on mob.
The problem was that stack was overwritten because stack was not
aligned correctly.
Hi Herman,
hope you don't mind some critic: At first, if we want to reserve
some space of 'size' on stack, while respecting alignment of 'align',
then there is nothing wrong with this line:
loc = (loc - size) & -align;
As such your addition
loc &= -align;
loc = (loc - size) & -align;
does not make sense, even if it may fix the problem.
So, what is the problem about really? The problem is that
sizeof (struct V { int x, y, z; })
is 12, but when returned in registers (on x86_64-linux), the size of
two registers is 2*8 = 16.
Therefor the problem is not wrong alignment, it is wrong size.
Digging further, it turns out that gfunc_sret() on x86_64-linux for
struct V { int x, y, z; }
returns: one register of type VT_QLONG with regsize 8.
That does not look right either. In fact, tcc handles VT_QLONG
as sort of pseudo register, using two processor registers (vtop->r/r2)
so I'd think that for VT_QLONG, it should pass 16 as the 'regsize'.
In the end, it seems that the space to be reserved on stack should be
calculated like this
size = ret_nregs * regsize;
rather than with 'size = type_size()'
Btw note that the other part or your patch
- || (align & (ret_align-1))) {
+ && (align & (ret_align-1))) {
exactly undoes a previous patch from Yao Zi
- && (align & (ret_align-1))) {
+ || (align & (ret_align-1))) {
As I tried to point out in an earlier email, this previous patch was
not the correct fix for the other problem, either.
That's why I think that our patches must strive for two things always:
1) to fix the problem and 2) in a way that logically does make sense ;)
I agree with your comments above. The size is incorrect.
I could change gfunc_sret in x86_64-gen.c and then calculate the size in
tccgen.c
as you suggested. But I am not sure regsize is set correctly all the time.
I like this better:
--- a/tccgen.c
+++ b/tccgen.c
@@ -6142,7 +6142,7 @@ special_math_val:
space. Assume register size is power of 2. */
if (regsize > align)
align = regsize;
- loc &= -align;
+ size = (size + regsize - 1) & -regsize;
loc = (loc - size) & -align;
addr = loc;
offset = 0;
The size should be a multiple of regsize.
What is your opinion?
I was working on a better fix for riscv and reverted the patch from Yao
Zi by accident.
I will reapply that change.
Herman
___
Tinycc-devel mailing list
Tinycc-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/tinycc-devel
Re: [Tinycc-devel] A double-to-float conversion bug
On 7/30/23 18:27, Vincent Lefevre wrote: On 2023-07-30 16:07:29 +0600, Viktor M wrote: Hello everyone. So today I stumbled upon this bug when doing math involving conversions between float and double. A minimal example: [...] Not really minimal. I could simplify it even further: I fixed this on mob. The problem was that stack was overwritten because stack was not aligned correctly. Herman ___ Tinycc-devel mailing list Tinycc-devel@nongnu.org https://lists.nongnu.org/mailman/listinfo/tinycc-devel
Re: [Tinycc-devel] A double-to-float conversion bug
On 2023-07-30 16:07:29 +0600, Viktor M wrote:
> Hello everyone. So today I stumbled upon this bug when doing math
> involving conversions between float and double. A minimal example:
[...]
Not really minimal. I could simplify it even further:
#include
struct V {
int x, y, z;
};
struct V vec(void)
{
return (struct V) { 0, 0, 1 };
}
void func(float f, struct V v)
{
printf("%f\n", f);
}
int main(void)
{
float d = 5.0f;
func(d, vec());
return 0;
}
I get successively on x86_64:
-213.969727
-0.01
-0.00
44308718004014403453773152256.00
Putting the 5.0f in the argument instead of d, or reducing the
struct size, or changing float to double makes the bug disappear.
--
Vincent Lefèvre - Web:
Re: [Tinycc-devel] A double-to-float conversion bug
Sorry. Ignore that.
30 juli 2023 16:18 av c...@voidptr.se:
> Use "%lf" for doubles.
>
> /cjw
>
> 30 juli 2023 15:34 av dayllen...@gmail.com:
>
>> Hello everyone. So today I stumbled upon this bug when doing math
>> involving conversions between float and double. A minimal example:
>>
>> ---
>> #include
>>
>> struct V {
>> int x, y, z;
>> };
>>
>> struct V vec(void)
>> {
>> return (struct V) { 0, 0, 1 };
>> }
>>
>> void func(float f, struct V v)
>> {
>> printf("%f %d %d %d\n", f, v.x, v.y, v.z);
>> }
>>
>> int main(void)
>> {
>> float f = 5;
>> double d = f;
>> func(d, vec());
>> return 0;
>> }
>> ---
>>
>> The code should print this:
>> 5.00 0 0 1
>>
>> but it prints garbage in the first value like this:
>> -85964.625000 0 0 1
>>
>> I've tried Linux x86_64 and both mob branch and 0.9.27 release.
>> tcc -run also fails.
>>
>> I hope there is an easy fix for this. Thanks.
>>
>> ___
>> Tinycc-devel mailing list
>> Tinycc-devel@nongnu.org
>> https://lists.nongnu.org/mailman/listinfo/tinycc-devel
>>
>
> ___
> Tinycc-devel mailing list
> Tinycc-devel@nongnu.org
> https://lists.nongnu.org/mailman/listinfo/tinycc-devel
>
___
Tinycc-devel mailing list
Tinycc-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/tinycc-devel
Re: [Tinycc-devel] A double-to-float conversion bug
Use "%lf" for doubles.
/cjw
30 juli 2023 15:34 av dayllen...@gmail.com:
> Hello everyone. So today I stumbled upon this bug when doing math
> involving conversions between float and double. A minimal example:
>
> ---
> #include
>
> struct V {
> int x, y, z;
> };
>
> struct V vec(void)
> {
> return (struct V) { 0, 0, 1 };
> }
>
> void func(float f, struct V v)
> {
> printf("%f %d %d %d\n", f, v.x, v.y, v.z);
> }
>
> int main(void)
> {
> float f = 5;
> double d = f;
> func(d, vec());
> return 0;
> }
> ---
>
> The code should print this:
> 5.00 0 0 1
>
> but it prints garbage in the first value like this:
> -85964.625000 0 0 1
>
> I've tried Linux x86_64 and both mob branch and 0.9.27 release.
> tcc -run also fails.
>
> I hope there is an easy fix for this. Thanks.
>
> ___
> Tinycc-devel mailing list
> Tinycc-devel@nongnu.org
> https://lists.nongnu.org/mailman/listinfo/tinycc-devel
>
___
Tinycc-devel mailing list
Tinycc-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/tinycc-devel
[Tinycc-devel] A double-to-float conversion bug
Hello everyone. So today I stumbled upon this bug when doing math
involving conversions between float and double. A minimal example:
---
#include
struct V {
int x, y, z;
};
struct V vec(void)
{
return (struct V) { 0, 0, 1 };
}
void func(float f, struct V v)
{
printf("%f %d %d %d\n", f, v.x, v.y, v.z);
}
int main(void)
{
float f = 5;
double d = f;
func(d, vec());
return 0;
}
---
The code should print this:
5.00 0 0 1
but it prints garbage in the first value like this:
-85964.625000 0 0 1
I've tried Linux x86_64 and both mob branch and 0.9.27 release.
tcc -run also fails.
I hope there is an easy fix for this. Thanks.
___
Tinycc-devel mailing list
Tinycc-devel@nongnu.org
https://lists.nongnu.org/mailman/listinfo/tinycc-devel
