Hi,
In the LiveIntervalAnalysis::runOnMachineFunction, there is a code to
compute the MBB2IdxMap, by remembering for each MBB its start and end
instruction numbers:
unsigned MIIndex = 0;
for (MachineFunction::iterator MBB = mf_->begin(), E = mf_->end();
MBB != E; ++MBB) {
unsigned StartIdx = MIIndex;
for (MachineBasicBlock::iterator I = MBB->begin(), E = MBB->end();
I != E; ++I) {
bool inserted = mi2iMap_.insert(std::make_pair(I,
MIIndex)).second;
assert(inserted && "multiple MachineInstr -> index
mappings");
i2miMap_.push_back(I);
MIIndex += InstrSlots::NUM;
}
// Set the MBB2IdxMap entry for this MBB.
MBB2IdxMap[MBB->getNumber()] = std::make_pair(StartIdx, MIIndex -
1);
For empty MBBs, the last line would create a pair, where the end
element is smaller than the start element, which is a bit
counter-intuitive. This may lead to some assertion failures in other
parts of the code generator, if they use MBB2IdxMap for checking is an
instruction belongs to a given MBB.
Question: Is this way of handling empty MBBs in the
LiveIntervalAnalysis a bug or a feature? Any special reason, why the
pair for an empty block does not look e.g. like (start, start)?
-Roman
__________________________________________________________
Gesendet von Yahoo! Mail.
Der Mailbox mit unbegrenztem Speicher.
http://de.overview.mail.yahoo.com
That's a bug. I'll fix it. Thanks. Evan On Apr 16, 2008, at 2:52 AM, Roman Levenstein wrote:> Hi, > > In the LiveIntervalAnalysis::runOnMachineFunction, there is a code to > compute the MBB2IdxMap, by remembering for each MBB its start and end > instruction numbers: > > unsigned MIIndex = 0; > for (MachineFunction::iterator MBB = mf_->begin(), E = mf_->end(); > MBB != E; ++MBB) { > unsigned StartIdx = MIIndex; > > for (MachineBasicBlock::iterator I = MBB->begin(), E = MBB->end(); > I != E; ++I) { > bool inserted = mi2iMap_.insert(std::make_pair(I, > MIIndex)).second; > assert(inserted && "multiple MachineInstr -> index mappings"); > i2miMap_.push_back(I); > MIIndex += InstrSlots::NUM; > } > > // Set the MBB2IdxMap entry for this MBB. > MBB2IdxMap[MBB->getNumber()] = std::make_pair(StartIdx, MIIndex - > 1); > > For empty MBBs, the last line would create a pair, where the end > element is smaller than the start element, which is a bit > counter-intuitive. This may lead to some assertion failures in other > parts of the code generator, if they use MBB2IdxMap for checking is an > instruction belongs to a given MBB. > > Question: Is this way of handling empty MBBs in the > LiveIntervalAnalysis a bug or a feature? Any special reason, why the > pair for an empty block does not look e.g. like (start, start)? > > -Roman > > > __________________________________________________________ > Gesendet von Yahoo! Mail. > Der Mailbox mit unbegrenztem Speicher. > http://de.overview.mail.yahoo.com > _______________________________________________ > LLVM Developers mailing list > LLVMdev at cs.uiuc.edu http://llvm.cs.uiuc.edu > http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev
Hi I'm seeing something probably related to this. I'm getting an assert from the lower_bound in LiveIntervals::findLiveinMBBs (from a checking std:: VS2005 implementation). Idx2MBBMap has two elements in it, both of which have a .first of 0. (I believe because of an empty MBB in the function below, so StartIndex doesn't advance). scott On Wed, Apr 16, 2008 at 2:52 AM, Roman Levenstein <romixlev at yahoo.com> wrote:> Hi, > > In the LiveIntervalAnalysis::runOnMachineFunction, there is a code to > compute the MBB2IdxMap, by remembering for each MBB its start and end > instruction numbers: > > unsigned MIIndex = 0; > for (MachineFunction::iterator MBB = mf_->begin(), E = mf_->end(); > MBB != E; ++MBB) { > unsigned StartIdx = MIIndex; > > for (MachineBasicBlock::iterator I = MBB->begin(), E = MBB->end(); > I != E; ++I) { > bool inserted = mi2iMap_.insert(std::make_pair(I, > MIIndex)).second; > assert(inserted && "multiple MachineInstr -> index mappings"); > i2miMap_.push_back(I); > MIIndex += InstrSlots::NUM; > } > > // Set the MBB2IdxMap entry for this MBB. > MBB2IdxMap[MBB->getNumber()] = std::make_pair(StartIdx, MIIndex - > 1); > > For empty MBBs, the last line would create a pair, where the end > element is smaller than the start element, which is a bit > counter-intuitive. This may lead to some assertion failures in other > parts of the code generator, if they use MBB2IdxMap for checking is an > instruction belongs to a given MBB. > > Question: Is this way of handling empty MBBs in the > LiveIntervalAnalysis a bug or a feature? Any special reason, why the > pair for an empty block does not look e.g. like (start, start)? > > -Roman > > > __________________________________________________________ > Gesendet von Yahoo! Mail. > Der Mailbox mit unbegrenztem Speicher. > http://de.overview.mail.yahoo.com > _______________________________________________ > LLVM Developers mailing list > LLVMdev at cs.uiuc.edu http://llvm.cs.uiuc.edu > http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev >
Can you file a bug so I don't forget? I'm a little occupied right now. But I'll take care of this soon. Evan On Apr 16, 2008, at 10:52 AM, Scott Graham wrote:> Hi > > I'm seeing something probably related to this. I'm getting an assert > from the lower_bound in LiveIntervals::findLiveinMBBs (from a checking > std:: VS2005 implementation). Idx2MBBMap has two elements in it, both > of which have a .first of 0. (I believe because of an empty MBB in the > function below, so StartIndex doesn't advance). > > scott > > On Wed, Apr 16, 2008 at 2:52 AM, Roman Levenstein > <romixlev at yahoo.com> wrote: >> Hi, >> >> In the LiveIntervalAnalysis::runOnMachineFunction, there is a code to >> compute the MBB2IdxMap, by remembering for each MBB its start and end >> instruction numbers: >> >> unsigned MIIndex = 0; >> for (MachineFunction::iterator MBB = mf_->begin(), E = mf_->end(); >> MBB != E; ++MBB) { >> unsigned StartIdx = MIIndex; >> >> for (MachineBasicBlock::iterator I = MBB->begin(), E = MBB->end(); >> I != E; ++I) { >> bool inserted = mi2iMap_.insert(std::make_pair(I, >> MIIndex)).second; >> assert(inserted && "multiple MachineInstr -> index mappings"); >> i2miMap_.push_back(I); >> MIIndex += InstrSlots::NUM; >> } >> >> // Set the MBB2IdxMap entry for this MBB. >> MBB2IdxMap[MBB->getNumber()] = std::make_pair(StartIdx, MIIndex - >> 1); >> >> For empty MBBs, the last line would create a pair, where the end >> element is smaller than the start element, which is a bit >> counter-intuitive. This may lead to some assertion failures in other >> parts of the code generator, if they use MBB2IdxMap for checking is >> an >> instruction belongs to a given MBB. >> >> Question: Is this way of handling empty MBBs in the >> LiveIntervalAnalysis a bug or a feature? Any special reason, why the >> pair for an empty block does not look e.g. like (start, start)? >> >> -Roman >> >> >> __________________________________________________________ >> Gesendet von Yahoo! Mail. >> Der Mailbox mit unbegrenztem Speicher. >> http://de.overview.mail.yahoo.com >> _______________________________________________ >> LLVM Developers mailing list >> LLVMdev at cs.uiuc.edu http://llvm.cs.uiuc.edu >> http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev >> > _______________________________________________ > LLVM Developers mailing list > LLVMdev at cs.uiuc.edu http://llvm.cs.uiuc.edu > http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev