[Bug target/40236] New: Request a single option to turn off all instructions which can cause #TS

[Bug target/40236] New: Request a single option to turn off all instructions which can cause #TS

[hpa at zytor dot com]

In the Linux kernel, we use (except in certain very restricted contexts) only
the subset of the x86 instruction set that can never cause #TS faults and thus
doesn't use the FSAVE/FXSAVE/XSAVE state.  It appears that we need to add new
-mno- options (-mno-sse -mno-sse2 -mno-sse3...) as gcc starts using new subsets
of the x86 instruction set.  This is error prone, for obvious reasons.

We would really like a single option that would turn off all instructions
subsets that can cause a #TS fault.


-- 
           Summary: Request a single option to turn off all instructions
                    which can cause #TS
           Product: gcc
           Version: unknown
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: target
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: hpa at zytor dot com
 GCC build triplet: Any
  GCC host triplet: Any
GCC target triplet: i386-any


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40236

[Bug target/40236] i386: request a single option to turn off all instructions which can cause #TS

[pinskia at gcc dot gnu dot org]

------- Comment #1 from pinskia at gcc dot gnu dot org  2009-05-24 18:22 -------
I don't see why adding -mno-mmx is a problem (that disables all of the MMX and
SSE instructions). -mno-mmx will always disabling all of the future subsets
too.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40236

[Bug target/40236] i386: request a single option to turn off all instructions which can cause #TS

[hpa at zytor dot com]

------- Comment #2 from hpa at zytor dot com  2009-05-24 18:26 -------
If that's the official word, then that's good enough for us.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40236

[Bug target/40236] i386: request a single option to turn off all instructions which can cause #TS

[hpa at zytor dot com]

------- Comment #3 from hpa at zytor dot com  2009-05-24 18:26 -------
I guess I would request that this be added to the documentation.  Keep in mind,
though, that this should include x87, MMX, SSE, and AVX.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40236

[Bug target/40236] i386: request a single option to turn off all instructions which can cause #TS

[ubizjak at gmail dot com]

------- Comment #4 from ubizjak at gmail dot com  2009-05-24 19:19 -------
(In reply to comment #1)
> I don't see why adding -mno-mmx is a problem (that disables all of the MMX and
> SSE instructions). -mno-mmx will always disabling all of the future subsets
> too.

No, MMX and SSE are handled separatelly. -mno-mmx disables MMX regs. It is
-mno-sse that turns off SSE registers and everything SSE related, including
AVX.

-mno-80387 should disable x87 instructions.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40236

[Bug target/40236] i386: request a single option to turn off all instructions which can cause #TS

[joseph at codesourcery dot com]

------- Comment #5 from joseph at codesourcery dot com  2009-05-29 09:42 -------
Subject: Re:   New: Request a single option to turn off
 all instructions which can cause #TS

This sounds very much like the long-requested -fno-implicit-fp / 
-mno-implicit-fp option.  We have an implementation of this at 
CodeSourcery (target-independent parts and some target-dependent parts) 
but I'm not sure there was ever any consensus on the acceptability of some 
form of this option for FSF GCC.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40236

[Bug target/40236] i386: request a single option to turn off all instructions which can cause #TS

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=40236

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |DUPLICATE
             Status|UNCONFIRMED                 |RESOLVED

--- Comment #6 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Dup of bug 70738 which was fixed for GCC 7.

*** This bug has been marked as a duplicate of bug 70738 ***