we can go on forever if you only want to give me minimum info. Pls send me
the entire expr
Sun
On Tue, Nov 29, 2011 at 10:16 AM, Gang Yu <yugang...@gmail.com> wrote:
> Before the change, the input expression is wrongly codemapped to
>
> >LDC I8 -2147483649 <u=1 cr17> flags:0x0 b=-1
>
> it should be rightly mapped to:
>
> > LDC I8 -2147483649 <u=1 cr17> flags:0x0 b=-1
> >U4I4CVT <u=1 cr18> isop_flags:0x0 flags:0x0 b=-1
>
> Regards
> Gang
>
>
> On Mon, Nov 28, 2011 at 8:22 PM, Sun Chan <sun.c...@gmail.com> wrote:
>
>> without your change, what is the output of this function for the same
>> input expr?
>> Sun
>>
>> On Mon, Nov 28, 2011 at 10:50 AM, Gang Yu <yugang...@gmail.com> wrote:
>>
>>> opt_htable.cxx
>>> 2244BOOL
>>> 2245CODEMAP::Canon_cvt(WN *wn,
>>> 2246 OPT_STAB *opt_stab,
>>> 2247 STMTREP *stmt,
>>> 2248 CANON_CR *ccr,
>>> 2249 CODEREP *cr,
>>> 2250 COPYPROP *copyprop)
>>> 2251{
>>> 2252 const OPCODE op = WN_opcode(wn);
>>> 2253 WN *kid = WN_kid(wn, 0);
>>> 2254 BOOL propagated = Add_expr(kid, opt_stab, stmt, ccr, copyprop);
>>> 2255#if defined(TARG_NVISA)
>>> 2256 if (ccr->Tree() == NULL) {
>>> 2257 // There is only a constant value in ccr
>>> 2258 return propagated;
>>> 2259 }
>>> 2260#endif
>>> 2261
>>> 2262#if defined (TARG_MIPS) && !defined (TARG_SL)
>>> 2263 // U8I4CVT and I8I4CVT are nops so return kid, MIPS III and above
>>> 2264 // since U8I4CVT is required to preserve the type of its type for
>>> 2265 // Fix_var_type at emitter time, we do not delete U8I4CVT #329096
>>> 2266 if (Is_Target_ISA_M3Plus() && (op == OPC_I8I4CVT || op ==
>>> OPC_U8I4CVT))
>>> 2267 return propagated;
>>> 2268#elif defined(TARG_X8664)
>>> 2269 if (! Is_Target_32bit() && op == OPC_U8U4CVT)
>>> 2270 return propagated;
>>> 2271#endif
>>> 2272 if ((Get_mtype_class(OPCODE_rtype(op)) &
>>> 2273 Get_mtype_class(OPCODE_desc(op))) != 0 &&
>>> 2274 MTYPE_size_min(OPCODE_rtype(op)) ==
>>> MTYPE_size_min(OPCODE_desc(op)) &&
>>> 2275 // bug912 open64.net. Do not delete U4I4CVT if his kid is a
>>> constant
>>> 2276 (!(OPCODE_rtype(op) == MTYPE_U4 &&
>>> 2277 OPCODE_desc(op) == MTYPE_I4 &&
>>> 2278 ccr->Tree() == NULL)))
>>> 2279 return propagated;
>>> 2280
>>> 2281 if ( WOPT_Enable_Cvt_Folding &&
>>> 2282#if defined(TARG_X8664) || defined(TARG_NVISA) || defined (TARG_SL)
>>> // bug 5851
>>> 2283 ! Is_Target_32bit() &&
>>> 2284#endif
>>> 2285 (op == OPC_I8U4CVT || op == OPC_U8U4CVT) &&
>>> 2286 WN_operator(kid) == OPR_LSHR ) {
>>> 2287 WN *bits = WN_kid1(kid);
>>> 2288 if ( WN_operator(bits) == OPR_INTCONST ) {
>>> 2289 INT bits_cnt = WN_const_val(bits);
>>> 2290 INT result_size = Actual_data_size(kid);
>>> 2291 if ( ( result_size <= MTYPE_size_min(MTYPE_I4) ) &&
>>> 2292 ( ( bits_cnt & 0x1F ) != 0 ) )
>>> 2293 {
>>> 2294 return propagated;
>>> 2295 }
>>> 2296 }
>>> 2297 }
>>> 2298
>>> 2299 CODEREP *retv;
>>> 2300 CODEREP *expr;
>>> 2301#if defined(TARG_X8664) || defined(TARG_NVISA)
>>> 2302 if (!Is_Target_32bit() && ccr->Tree() != NULL &&
>>> Allow_wrap_around_opt) {
>>> 2303 if (ccr->Tree()->Kind() == CK_OP && ccr->Tree()->Op() ==
>>> OPC_I8I4CVT &&
>>> 2304 op == OPC_I4I8CVT) { // bug 10707
>>> 2305 retv = ccr->Tree()->Opnd(0);
>>> 2306 ccr->Set_tree(retv);
>>> 2307 }
>>> 2308 else {
>>> 2309 cr->Set_opnd(0, ccr->Tree());
>>> 2310 retv = Hash_Op(cr);
>>> 2311 ccr->Set_tree(retv); // move the CVT to the operand
>>> 2312 }
>>> 2313 }
>>> 2314 else {
>>> 2315#endif
>>> 2316 expr = ccr->Convert2cr(WN_kid0(wn), this, propagated);
>>> 2317 cr->Set_opnd(0, expr);
>>> 2318 retv = Hash_Op(cr);
>>> 2319 ccr->Set_tree(retv);
>>> 2320 ccr->Set_scale(0);
>>> 2321#if defined(TARG_X8664) || defined(TARG_NVISA)
>>> 2322 }
>>> 2323#endif
>>> 2324 return propagated;
>>> 2325}
>>>
>>> With this fix, controls goes to 2316,
>>> expr is
>>> >LDC I8 -2147483649 <u=1 cr17> flags:0x0 b=-1
>>> retv is
>>> > LDC I8 -2147483649 <u=1 cr17> flags:0x0 b=-1
>>> >U4I4CVT <u=1 cr18> isop_flags:0x0 flags:0x0 b=-1
>>> the input is still:
>>> I4INTCONST 1 (0x1)
>>> I4I4LDID 0 <st 20> T<4,.predef_I4,4>
>>> I4SHL
>>> I4INTCONST -1 (0xffffffffffffffff)
>>> I4ADD
>>> U4I4CVT
>>>
>>>
>>> Regards
>>> Gang
>>>
>>>
>>> On Mon, Nov 28, 2011 at 10:40 AM, Sun Chan <sun.c...@gmail.com> wrote:
>>>
>>>> Sorry, pls send the corresponding function so I get a bigger picture of
>>>> what the deal is. It will be good to give me the expression tree before and
>>>> after this function call
>>>> Sun
>>>>
>>>> On Mon, Nov 28, 2011 at 10:30 AM, Gang Yu <yugang...@gmail.com> wrote:
>>>>
>>>>> Inspired by your comments, I get a more reasonable fix.
>>>>>
>>>>> osprey/be/opt/opt_htable.cxx -- 1fc05d0..0b7a14a 100644
>>>>> --- a/osprey/be/opt/opt_htable.cxx
>>>>> +++ b/osprey/be/opt/opt_htable.cxx
>>>>> @@ -2271,7 +2271,11 @@ CODEMAP::Canon_cvt(WN *wn,
>>>>> #endif
>>>>> if ((Get_mtype_class(OPCODE_rtype(op)) &
>>>>> Get_mtype_class(OPCODE_desc(op))) != 0 &&
>>>>> - MTYPE_size_min(OPCODE_rtype(op)) ==
>>>>> MTYPE_size_min(OPCODE_desc(op)))
>>>>> + MTYPE_size_min(OPCODE_rtype(op)) ==
>>>>> MTYPE_size_min(OPCODE_desc(op)) &&
>>>>> + // bug912 open64.net. Do not delete U4I4CVT if his kid is a
>>>>> constant
>>>>> + (!(OPCODE_rtype(op) == MTYPE_U4 &&
>>>>> + OPCODE_desc(op) == MTYPE_I4 &&
>>>>> + ccr->Tree() == NULL)))
>>>>> return propagated;
>>>>> if ( WOPT_Enable_Cvt_Folding &&
>>>>> This new patch does not lost the chances to U4I4CVT optimisation,
>>>>> while it makes the case of U4I4CVT const right.
>>>>>
>>>>> Thanks a lot.
>>>>>
>>>>> Regards
>>>>> Gang
>>>>>
>>>>>
>>>>> On Mon, Nov 28, 2011 at 6:23 AM, Sun Chan <sun.c...@gmail.com> wrote:
>>>>>
>>>>>> so the const is of the right type to start out with. I'm not sure
>>>>>> htable is the right place to fix this (you could well be inhibiting
>>>>>> perfectly good optimization with your change).
>>>>>> I will need much more detail dump (e.g.when BAND first appear, before
>>>>>> and after) to know more about that.
>>>>>> Sun
>>>>>>
>>>>>> On Sun, Nov 27, 2011 at 9:48 PM, Gang Yu <yugang...@gmail.com>wrote:
>>>>>>
>>>>>>>
>>>>>>> On 11/27/2011 03:02 PM, Sun Chan wrote:
>>>>>>>
>>>>>>> you only showed the output, can you show input to htable?
>>>>>>> Sun
>>>>>>>
>>>>>>> Sun, Input of this expression:
>>>>>>>
>>>>>>> I4INTCONST 1 (0x1)
>>>>>>> I4I4LDID 0 <st 20> T<4,.predef_I4,4>
>>>>>>> I4SHL
>>>>>>> I4INTCONST -1 (0xffffffffffffffff)
>>>>>>> I4ADD
>>>>>>> U4I4CVT
>>>>>>>
>>>>>>> st 20 is var i ( copyproped to 31)
>>>>>>>
>>>>>>> regards
>>>>>>> Gang
>>>>>>>
>>>>>>> On Sun, Nov 27, 2011 at 11:24 AM, Gang Yu <yugang...@gmail.com>wrote:
>>>>>>>
>>>>>>>> Hi,
>>>>>>>>
>>>>>>>> Could a gatekeeper please help review the fix for bug912?
>>>>>>>> http://bugs.open64.net/show_bug.cgi?id=912
>>>>>>>>
>>>>>>>> The cut down bug case:
>>>>>>>>
>>>>>>>> struct go7007 {
>>>>>>>> int sensor_framerate;
>>>>>>>> int fps_scale;
>>>>>>>> };
>>>>>>>> int vti_bitlen(struct go7007 *go)
>>>>>>>> {
>>>>>>>> unsigned int i, max_time_incr = go->sensor_framerate /
>>>>>>>> go->fps_scale;
>>>>>>>> for (i = 31; (max_time_incr & ((1 << i) - 1)) == max_time_incr;
>>>>>>>> --i);
>>>>>>>> return i + 1;
>>>>>>>> }
>>>>>>>> ### Assertion failure at line 2109 of
>>>>>>>> /fc/proj/ctires/open64/o64guru/src/Sat/trunk/osprey/be/cg/cgemit.cxx:
>>>>>>>> ### Compiler Error in file bug912-cutdown.c during Assembly phase:
>>>>>>>> ### literal for operand 1 is not in range
>>>>>>>> Analysis:
>>>>>>>> WHIRL dump snippet after WOPT:
>>>>>>>> ......
>>>>>>>> U4STID 50 <1,8,.preg_U4> T<8,.predef_U4,4> # max_time_incr {line:
>>>>>>>> 1/8}
>>>>>>>> U4U4LDID 50 <1,8,.preg_U4> T<8,.predef_U4,4> # max_time_incr
>>>>>>>> U4U4LDID 50 <1,8,.preg_U4> T<8,.predef_U4,4> # max_time_incr
>>>>>>>> I8INTCONST -2147483649 (0xffffffff7fffffff)
>>>>>>>> U4BAND
>>>>>>>> I4U4EQ
>>>>>>>> FALSEBR L2562 {line: 0/0}
>>>>>>>>
>>>>>>>> This is already wrong, U4BAND should not accept I8INTCONST kids.
>>>>>>>>
>>>>>>>> ((1 << 31) - 1) in such context should be codemapped to U4INTCONST
>>>>>>>> 2147483647 (0x7fffffff)
>>>>>>>>
>>>>>>>> Suggested patch:
>>>>>>>>
>>>>>>>> --- a/osprey/be/opt/opt_htable.cxx
>>>>>>>> +++ b/osprey/be/opt/opt_htable.cxx
>>>>>>>> @@ -3679,7 +3679,10 @@ CODEMAP::Add_expr(WN *wn, OPT_STAB
>>>>>>>> *opt_stab, STMTREP *stmt, CANON_CR *ccr,
>>>>>>>> if ((MTYPE_type_class(OPCODE_rtype(op)) &
>>>>>>>> MTYPE_CLASS_INTEGER) != 0
>>>>>>>> #ifdef TARG_X8664 // bug 7733
>>>>>>>> && ! MTYPE_is_vector(OPCODE_rtype(op))
>>>>>>>> - && ! MTYPE_is_vector(OPCODE_desc(op))
>>>>>>>> + && ! MTYPE_is_vector(OPCODE_desc(op))
>>>>>>>> + // bug912 open64.net. do not Canonicalize U4I4CVT
>>>>>>>> + && ! (OPCODE_rtype(op) == MTYPE_U4 &&
>>>>>>>> + OPCODE_desc(op) == MTYPE_I4)
>>>>>>>> #endif
>>>>>>>> )
>>>>>>>> return Canon_cvt(wn, opt_stab, stmt, ccr, cr, copyprop);
>>>>>>>> Could a gatekeeper please help review ? Thanks
>>>>>>>>
>>>>>>>> Regards
>>>>>>>> Gang
>>>>>>>>
>>>>>>>>
>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>> All the data continuously generated in your IT infrastructure
>>>>>>>> contains a definitive record of customers, application performance,
>>>>>>>> security threats, fraudulent activity, and more. Splunk takes this
>>>>>>>> data and makes sense of it. IT sense. And common sense.
>>>>>>>> http://p.sf.net/sfu/splunk-novd2d
>>>>>>>> _______________________________________________
>>>>>>>> Open64-devel mailing list
>>>>>>>> Open64-devel@lists.sourceforge.net
>>>>>>>> https://lists.sourceforge.net/lists/listinfo/open64-devel
>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>
>>>>>
>>>>
>>>
>>
>
------------------------------------------------------------------------------
All the data continuously generated in your IT infrastructure
contains a definitive record of customers, application performance,
security threats, fraudulent activity, and more. Splunk takes this
data and makes sense of it. IT sense. And common sense.
http://p.sf.net/sfu/splunk-novd2d
_______________________________________________
Open64-devel mailing list
Open64-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/open64-devel
