|
From: <sv...@va...> - 2011-10-20 20:42:31
|
Author: florian Date: 2011-10-20 21:37:46 +0100 (Thu, 20 Oct 2011) New Revision: 12180 Log: Filter out more warnings (Dwarf reader and debug info). Modified: trunk/tests/filter_stderr_basic Modified: trunk/tests/filter_stderr_basic =================================================================== --- trunk/tests/filter_stderr_basic 2011-10-20 16:18:50 UTC (rev 12179) +++ trunk/tests/filter_stderr_basic 2011-10-20 20:37:46 UTC (rev 12180) @@ -52,5 +52,10 @@ sed "s/\(signal [0-9]* (SIG[A-Z]*)\): dumping core/\1/" | # Remove the size in "The main thread stack size..." message. -sed "s/The main thread stack size used in this run was [0-9]*/The main thread stack size used in this run was .../" +sed "s/The main thread stack size used in this run was [0-9]*/The main thread stack size used in this run was .../" | +# Suppress warnings from incompatible debug info +sed '/warning: the debug information found in "[^"]*" does not match/d' | + +# Suppress warnings from Dwarf reader +sed '/warning: evaluate_Dwarf3_Expr: unhandled DW_OP_/d' |
|
From: Christian B. <bor...@de...> - 2011-10-20 20:46:08
|
> +# Suppress warnings from Dwarf reader > +sed '/warning: evaluate_Dwarf3_Expr: unhandled DW_OP_/d' Hmmm, isnt this actually a bug, if gcc generated dwarf Ops that valgrind doesnt handle? |
|
From: Florian K. <br...@ac...> - 2011-10-21 12:51:57
|
On 10/21/2011 12:13 AM, Julian Seward wrote: > On Thursday, October 20, 2011, Florian Krohm wrote: >> On 10/20/2011 04:45 PM, Christian Borntraeger wrote: >>>> +# Suppress warnings from Dwarf reader >>>> +sed '/warning: evaluate_Dwarf3_Expr: unhandled DW_OP_/d' >>> > > IIRC it was failing on DW_OP_piece (0x93 I think). This is related to > variables that are partly in memory and partly in registers, eg a 64-bit > value that has been partially spilled on a 32-bit CPU. The current > infrastructure cannot represent them. For sure, hiding error messages > with other DW_OP_ values isn't good, since we'll need to support them > at some point (as Christian implies). > I'll open a bugzilla to remind us of improving the dwarf machinery in this respect and mention the suppression there, so it can be undone and won't be forgotten. Florian |
|
From: Christian B. <bor...@de...> - 2011-10-21 19:42:42
|
> I'll open a bugzilla to remind us of improving the dwarf machinery in > this respect and mention the suppression there, so it can be undone and > won't be forgotten. Please change the suppression to only suppress that specific opcode. We want to see all others. Really. |
|
From: Florian K. <br...@ac...> - 2011-10-22 14:02:16
|
On 10/21/2011 08:51 AM, Florian Krohm wrote: > On 10/21/2011 12:13 AM, Julian Seward wrote: >> On Thursday, October 20, 2011, Florian Krohm wrote: >>> On 10/20/2011 04:45 PM, Christian Borntraeger wrote: >>>>> +# Suppress warnings from Dwarf reader >>>>> +sed '/warning: evaluate_Dwarf3_Expr: unhandled DW_OP_/d' >>>> >> >> IIRC it was failing on DW_OP_piece (0x93 I think). This is related to >> variables that are partly in memory and partly in registers, eg a 64-bit >> value that has been partially spilled on a 32-bit CPU. The current >> infrastructure cannot represent them. For sure, hiding error messages >> with other DW_OP_ values isn't good, since we'll need to support them >> at some point (as Christian implies). >> > > I'll open a bugzilla to remind us of improving the dwarf machinery in > this respect and mention the suppression there, so it can be undone and > won't be forgotten. > He he, we already have a bug for that: #185801 opened 2009-02-28 I've added a comment there. Florian |
|
From: Florian K. <br...@ac...> - 2011-10-20 21:25:41
|
On 10/20/2011 04:45 PM, Christian Borntraeger wrote: >> +# Suppress warnings from Dwarf reader >> +sed '/warning: evaluate_Dwarf3_Expr: unhandled DW_OP_/d' > > Hmmm, isnt this actually a bug, if gcc generated dwarf Ops that valgrind doesnt handle? > Yes, it is. However, Julian looked at it and concluded that this would be difficult to fix given the current dwarf infrastructure. Suppressing these warnings is clearly less than ideal, but helps filtering the regression that matter out of the noise. Florian |
|
From: Julian S. <js...@ac...> - 2011-10-21 04:15:33
|
On Thursday, October 20, 2011, Florian Krohm wrote: > On 10/20/2011 04:45 PM, Christian Borntraeger wrote: > >> +# Suppress warnings from Dwarf reader > >> +sed '/warning: evaluate_Dwarf3_Expr: unhandled DW_OP_/d' > > > > Hmmm, isnt this actually a bug, if gcc generated dwarf Ops that valgrind > > doesnt handle? > > Yes, it is. However, Julian looked at it and concluded that this would > be difficult to fix given the current dwarf infrastructure. > Suppressing these warnings is clearly less than ideal, but helps > filtering the regression that matter out of the noise. IIRC it was failing on DW_OP_piece (0x93 I think). This is related to variables that are partly in memory and partly in registers, eg a 64-bit value that has been partially spilled on a 32-bit CPU. The current infrastructure cannot represent them. For sure, hiding error messages with other DW_OP_ values isn't good, since we'll need to support them at some point (as Christian implies). J |