[Bug libgcc/105016] [libgcc, TARGET_HAS_NO_HW_DIVIDE] Incorrect result for __udivmodti4

[rguenth at gcc dot gnu.org via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105016

Richard Biener  changed:

   What|Removed |Added

 Ever confirmed|0   |1
   Last reconfirmed||2022-03-22
 Resolution|INVALID |---
 Status|RESOLVED|NEW

[Bug libgcc/105016] [libgcc, TARGET_HAS_NO_HW_DIVIDE] Incorrect result for __udivmodti4

[rguenth at gcc dot gnu.org via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105016

--- Comment #5 from Richard Biener  ---
The UDIV_NEEDS_NORMALIZATION path with !TARGET_HAS_NO_HW_DIVIDE might be
similarly affected.

[Bug libgcc/105016] [libgcc, TARGET_HAS_NO_HW_DIVIDE] Incorrect result for __udivmodti4

[rguenth at gcc dot gnu.org via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105016

--- Comment #4 from Richard Biener  ---
I see.  The function should probably use count_leading_zeros instead of
__builtin_clzll?  Though I fail to see how that would make a difference
since it doesn't have support for TImode either.

[Bug libgcc/105016] [libgcc, TARGET_HAS_NO_HW_DIVIDE] Incorrect result for __udivmodti4

[vries at gcc dot gnu.org via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105016

--- Comment #3 from Tom de Vries  ---
In libgcc.h, I see:
...
#define __udivmoddi4__NDW(udivmod,4)
...
and for LIBGCC2_UNITS_PER_WORD == 8 we have:
...
#define __NDW(a,b)  __ ## a ## ti ## b
...

So, AFAICT it's possible that __udivmoddi4 is mapped to __udivmodti4.

[Bug libgcc/105016] [libgcc, TARGET_HAS_NO_HW_DIVIDE] Incorrect result for __udivmodti4

[rguenth at gcc dot gnu.org via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105016

Richard Biener  changed:

   What|Removed |Added

 Status|UNCONFIRMED |RESOLVED
 Resolution|--- |INVALID

--- Comment #2 from Richard Biener  ---
since long long is DImode the function cannot work with TImode UDWtype.  It's
called __udivmoddi4 for a reason.

[Bug libgcc/105016] [libgcc, TARGET_HAS_NO_HW_DIVIDE] Incorrect result for __udivmodti4

[vries at gcc dot gnu.org via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105016

--- Comment #1 from Tom de Vries  ---
Created attachment 52662
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=52662&action=edit
test-case