Searched hist:2010 (Results 801 - 825 of 929) sorted by relevance
/gem5/src/arch/sparc/isa/formats/mem/ | ||
H A D | util.isa | diff 7741:340b6f01d69b Thu Nov 11 05:03:00 EST 2010 Gabe Black <gblack@eecs.umich.edu> SPARC: Clean up some historical style issues. diff 7712:7733c562e5e3 Fri Oct 22 03:23:00 EDT 2010 Gabe Black <gblack@eecs.umich.edu> ISA: Simplify various implementations of completeAcc. |
/gem5/configs/example/ | ||
H A D | memtest.py | diff 7656:dd4d229b716d Thu Aug 26 00:55:00 EDT 2010 Steve Reinhardt <steve.reinhardt@amd.com> memtest: scale associativity and mshrs according to config Use the actual fanouts in the tree specification to scale cache associativity and mshrs instead of dumb constants. diff 7525:722f2ad014a7 Tue Aug 17 08:06:00 EDT 2010 Steve Reinhardt <steve.reinhardt@amd.com> sim: make Python Root object a singleton Enforce that the Python Root SimObject is instantiated only once. The C++ Root object already panics if more than one is created. This change avoids the need to track what the root object is, since it's available from Root.getInstance() (if it exists). It's now redundant to have the user pass the root object to functions like instantiate(), checkpoint(), and restoreCheckpoint(), so that arg is gone. Users who use configs/common/Simulate.py should not notice. |
/gem5/src/arch/sparc/ | ||
H A D | faults.hh | diff 7741:340b6f01d69b Thu Nov 11 05:03:00 EST 2010 Gabe Black <gblack@eecs.umich.edu> SPARC: Clean up some historical style issues. diff 7678:f19b6a3a8cec Mon Sep 13 22:26:00 EDT 2010 Gabe Black <gblack@eecs.umich.edu> Faults: Pass the StaticInst involved, if any, to a Fault's invoke method. Also move the "Fault" reference counted pointer type into a separate file, sim/fault.hh. It would be better to name this less similarly to sim/faults.hh to reduce confusion, but fault.hh matches the name of the type. We could change Fault to FaultPtr to match other pointer types, and then changing the name of the file would make more sense. |
/gem5/src/cpu/checker/ | ||
H A D | cpu_impl.hh | diff 7679:f26cc2c68b48 Tue Sep 14 00:58:00 EDT 2010 Gabe Black <gblack@eecs.umich.edu> CPU: Get rid of the now unnecessary getInst/setInst family of functions. This code is no longer needed because of the preceeding change which adds a StaticInstPtr parameter to the fault's invoke method, obviating the only use for this pair of functions. diff 7678:f19b6a3a8cec Mon Sep 13 22:26:00 EDT 2010 Gabe Black <gblack@eecs.umich.edu> Faults: Pass the StaticInst involved, if any, to a Fault's invoke method. Also move the "Fault" reference counted pointer type into a separate file, sim/fault.hh. It would be better to name this less similarly to sim/faults.hh to reduce confusion, but fault.hh matches the name of the type. We could change Fault to FaultPtr to match other pointer types, and then changing the name of the file would make more sense. |
/gem5/src/mem/cache/tags/ | ||
H A D | base.cc | diff 6979:7732bca47f60 Wed Feb 24 16:46:00 EST 2010 Lisa Hsu <Lisa.Hsu@amd.com> cache stats: account for writebacks and/or device occupancy in the cache. Plus, a minor bugfix that neglects to update blk->contextSrc in certain cases on a cache insert. diff 6978:ab05e20dc4a7 Tue Feb 23 12:34:00 EST 2010 Lisa Hsu <Lisa.Hsu@amd.com> cache: Make caches sharing aware and add occupancy stats. On the config end, if a shared L2 is created for the system, it is parameterized to have n sharers as defined by option.num_cpus. In addition to making the cache sharing aware so that discriminating tag policies can make use of context_ids to make decisions, I added an occupancy AverageStat and an occ % stat to each cache so that you could know which contexts are occupying how much cache on average, both in terms of blocks and percentage. Note that since devices have context_id -1, having an array of occ stats that correspond to each context_id will break here, so in FS mode I add an extra bucket for device blocks. This bucket is explicitly not added in SE mode in order to not only avoid ugliness in the stats.txt file, but to avoid broken stats (some formulas break when a bucket is 0). |
H A D | base.hh | diff 7612:917946898102 Mon Aug 23 12:18:00 EDT 2010 Gene Wu <Gene.Wu@arm.com> MEM: Make CLREX a first class request operation and clear locks in caches when it in received diff 6978:ab05e20dc4a7 Tue Feb 23 12:34:00 EST 2010 Lisa Hsu <Lisa.Hsu@amd.com> cache: Make caches sharing aware and add occupancy stats. On the config end, if a shared L2 is created for the system, it is parameterized to have n sharers as defined by option.num_cpus. In addition to making the cache sharing aware so that discriminating tag policies can make use of context_ids to make decisions, I added an occupancy AverageStat and an occ % stat to each cache so that you could know which contexts are occupying how much cache on average, both in terms of blocks and percentage. Note that since devices have context_id -1, having an array of occ stats that correspond to each context_id will break here, so in FS mode I add an extra bucket for device blocks. This bucket is explicitly not added in SE mode in order to not only avoid ugliness in the stats.txt file, but to avoid broken stats (some formulas break when a bucket is 0). |
/gem5/src/cpu/o3/ | ||
H A D | rob_impl.hh | diff 7720:65d338a8dba4 Sun Oct 31 03:07:00 EDT 2010 Gabe Black <gblack@eecs.umich.edu> ISA,CPU,etc: Create an ISA defined PC type that abstracts out ISA behaviors. This change is a low level and pervasive reorganization of how PCs are managed in M5. Back when Alpha was the only ISA, there were only 2 PCs to worry about, the PC and the NPC, and the lsb of the PC signaled whether or not you were in PAL mode. As other ISAs were added, we had to add an NNPC, micro PC and next micropc, x86 and ARM introduced variable length instruction sets, and ARM started to keep track of mode bits in the PC. Each CPU model handled PCs in its own custom way that needed to be updated individually to handle the new dimensions of variability, or, in the case of ARMs mode-bit-in-the-pc hack, the complexity could be hidden in the ISA at the ISA implementation's expense. Areas like the branch predictor hadn't been updated to handle branch delay slots or micropcs, and it turns out that had introduced a significant (10s of percent) performance bug in SPARC and to a lesser extend MIPS. Rather than perpetuate the problem by reworking O3 again to handle the PC features needed by x86, this change was introduced to rework PC handling in a more modular, transparent, and hopefully efficient way. PC type: Rather than having the superset of all possible elements of PC state declared in each of the CPU models, each ISA defines its own PCState type which has exactly the elements it needs. A cross product of canned PCState classes are defined in the new "generic" ISA directory for ISAs with/without delay slots and microcode. These are either typedef-ed or subclassed by each ISA. To read or write this structure through a *Context, you use the new pcState() accessor which reads or writes depending on whether it has an argument. If you just want the address of the current or next instruction or the current micro PC, you can get those through read-only accessors on either the PCState type or the *Contexts. These are instAddr(), nextInstAddr(), and microPC(). Note the move away from readPC. That name is ambiguous since it's not clear whether or not it should be the actual address to fetch from, or if it should have extra bits in it like the PAL mode bit. Each class is free to define its own functions to get at whatever values it needs however it needs to to be used in ISA specific code. Eventually Alpha's PAL mode bit could be moved out of the PC and into a separate field like ARM. These types can be reset to a particular pc (where npc = pc + sizeof(MachInst), nnpc = npc + sizeof(MachInst), upc = 0, nupc = 1 as appropriate), printed, serialized, and compared. There is a branching() function which encapsulates code in the CPU models that checked if an instruction branched or not. Exactly what that means in the context of branch delay slots which can skip an instruction when not taken is ambiguous, and ideally this function and its uses can be eliminated. PCStates also generally know how to advance themselves in various ways depending on if they point at an instruction, a microop, or the last microop of a macroop. More on that later. Ideally, accessing all the PCs at once when setting them will improve performance of M5 even though more data needs to be moved around. This is because often all the PCs need to be manipulated together, and by getting them all at once you avoid multiple function calls. Also, the PCs of a particular thread will have spatial locality in the cache. Previously they were grouped by element in arrays which spread out accesses. Advancing the PC: The PCs were previously managed entirely by the CPU which had to know about PC semantics, try to figure out which dimension to increment the PC in, what to set NPC/NNPC, etc. These decisions are best left to the ISA in conjunction with the PC type itself. Because most of the information about how to increment the PC (mainly what type of instruction it refers to) is contained in the instruction object, a new advancePC virtual function was added to the StaticInst class. Subclasses provide an implementation that moves around the right element of the PC with a minimal amount of decision making. In ISAs like Alpha, the instructions always simply assign NPC to PC without having to worry about micropcs, nnpcs, etc. The added cost of a virtual function call should be outweighed by not having to figure out as much about what to do with the PCs and mucking around with the extra elements. One drawback of making the StaticInsts advance the PC is that you have to actually have one to advance the PC. This would, superficially, seem to require decoding an instruction before fetch could advance. This is, as far as I can tell, realistic. fetch would advance through memory addresses, not PCs, perhaps predicting new memory addresses using existing ones. More sophisticated decisions about control flow would be made later on, after the instruction was decoded, and handed back to fetch. If branching needs to happen, some amount of decoding needs to happen to see that it's a branch, what the target is, etc. This could get a little more complicated if that gets done by the predecoder, but I'm choosing to ignore that for now. Variable length instructions: To handle variable length instructions in x86 and ARM, the predecoder now takes in the current PC by reference to the getExtMachInst function. It can modify the PC however it needs to (by setting NPC to be the PC + instruction length, for instance). This could be improved since the CPU doesn't know if the PC was modified and always has to write it back. ISA parser: To support the new API, all PC related operand types were removed from the parser and replaced with a PCState type. There are two warts on this implementation. First, as with all the other operand types, the PCState still has to have a valid operand type even though it doesn't use it. Second, using syntax like PCS.npc(target) doesn't work for two reasons, this looks like the syntax for operand type overriding, and the parser can't figure out if you're reading or writing. Instructions that use the PCS operand (which I've consistently called it) need to first read it into a local variable, manipulate it, and then write it back out. Return address stack: The return address stack needed a little extra help because, in the presence of branch delay slots, it has to merge together elements of the return PC and the call PC. To handle that, a buildRetPC utility function was added. There are basically only two versions in all the ISAs, but it didn't seem short enough to put into the generic ISA directory. Also, the branch predictor code in O3 and InOrder were adjusted so that they always store the PC of the actual call instruction in the RAS, not the next PC. If the call instruction is a microop, the next PC refers to the next microop in the same macroop which is probably not desirable. The buildRetPC function advances the PC intelligently to the next macroop (in an ISA specific way) so that that case works. Change in stats: There were no change in stats except in MIPS and SPARC in the O3 model. MIPS runs in about 9% fewer ticks. SPARC runs with 30%-50% fewer ticks, which could likely be improved further by setting call/return instruction flags and taking advantage of the RAS. TODO: Add != operators to the PCState classes, defined trivially to be !(a==b). Smooth out places where PCs are split apart, passed around, and put back together later. I think this might happen in SPARC's fault code. Add ISA specific constructors that allow setting PC elements without calling a bunch of accessors. Try to eliminate the need for the branching() function. Factor out Alpha's PAL mode pc bit into a separate flag field, and eliminate places where it's blindly masked out or tested in the PC. diff 7717:f166f8bd8818 Sun Oct 24 03:43:00 EDT 2010 Gabe Black <gblack@eecs.umich.edu> O3: Get rid of a bunch of commented out lines. |
H A D | mem_dep_unit_impl.hh | diff 7720:65d338a8dba4 Sun Oct 31 03:07:00 EDT 2010 Gabe Black <gblack@eecs.umich.edu> ISA,CPU,etc: Create an ISA defined PC type that abstracts out ISA behaviors. This change is a low level and pervasive reorganization of how PCs are managed in M5. Back when Alpha was the only ISA, there were only 2 PCs to worry about, the PC and the NPC, and the lsb of the PC signaled whether or not you were in PAL mode. As other ISAs were added, we had to add an NNPC, micro PC and next micropc, x86 and ARM introduced variable length instruction sets, and ARM started to keep track of mode bits in the PC. Each CPU model handled PCs in its own custom way that needed to be updated individually to handle the new dimensions of variability, or, in the case of ARMs mode-bit-in-the-pc hack, the complexity could be hidden in the ISA at the ISA implementation's expense. Areas like the branch predictor hadn't been updated to handle branch delay slots or micropcs, and it turns out that had introduced a significant (10s of percent) performance bug in SPARC and to a lesser extend MIPS. Rather than perpetuate the problem by reworking O3 again to handle the PC features needed by x86, this change was introduced to rework PC handling in a more modular, transparent, and hopefully efficient way. PC type: Rather than having the superset of all possible elements of PC state declared in each of the CPU models, each ISA defines its own PCState type which has exactly the elements it needs. A cross product of canned PCState classes are defined in the new "generic" ISA directory for ISAs with/without delay slots and microcode. These are either typedef-ed or subclassed by each ISA. To read or write this structure through a *Context, you use the new pcState() accessor which reads or writes depending on whether it has an argument. If you just want the address of the current or next instruction or the current micro PC, you can get those through read-only accessors on either the PCState type or the *Contexts. These are instAddr(), nextInstAddr(), and microPC(). Note the move away from readPC. That name is ambiguous since it's not clear whether or not it should be the actual address to fetch from, or if it should have extra bits in it like the PAL mode bit. Each class is free to define its own functions to get at whatever values it needs however it needs to to be used in ISA specific code. Eventually Alpha's PAL mode bit could be moved out of the PC and into a separate field like ARM. These types can be reset to a particular pc (where npc = pc + sizeof(MachInst), nnpc = npc + sizeof(MachInst), upc = 0, nupc = 1 as appropriate), printed, serialized, and compared. There is a branching() function which encapsulates code in the CPU models that checked if an instruction branched or not. Exactly what that means in the context of branch delay slots which can skip an instruction when not taken is ambiguous, and ideally this function and its uses can be eliminated. PCStates also generally know how to advance themselves in various ways depending on if they point at an instruction, a microop, or the last microop of a macroop. More on that later. Ideally, accessing all the PCs at once when setting them will improve performance of M5 even though more data needs to be moved around. This is because often all the PCs need to be manipulated together, and by getting them all at once you avoid multiple function calls. Also, the PCs of a particular thread will have spatial locality in the cache. Previously they were grouped by element in arrays which spread out accesses. Advancing the PC: The PCs were previously managed entirely by the CPU which had to know about PC semantics, try to figure out which dimension to increment the PC in, what to set NPC/NNPC, etc. These decisions are best left to the ISA in conjunction with the PC type itself. Because most of the information about how to increment the PC (mainly what type of instruction it refers to) is contained in the instruction object, a new advancePC virtual function was added to the StaticInst class. Subclasses provide an implementation that moves around the right element of the PC with a minimal amount of decision making. In ISAs like Alpha, the instructions always simply assign NPC to PC without having to worry about micropcs, nnpcs, etc. The added cost of a virtual function call should be outweighed by not having to figure out as much about what to do with the PCs and mucking around with the extra elements. One drawback of making the StaticInsts advance the PC is that you have to actually have one to advance the PC. This would, superficially, seem to require decoding an instruction before fetch could advance. This is, as far as I can tell, realistic. fetch would advance through memory addresses, not PCs, perhaps predicting new memory addresses using existing ones. More sophisticated decisions about control flow would be made later on, after the instruction was decoded, and handed back to fetch. If branching needs to happen, some amount of decoding needs to happen to see that it's a branch, what the target is, etc. This could get a little more complicated if that gets done by the predecoder, but I'm choosing to ignore that for now. Variable length instructions: To handle variable length instructions in x86 and ARM, the predecoder now takes in the current PC by reference to the getExtMachInst function. It can modify the PC however it needs to (by setting NPC to be the PC + instruction length, for instance). This could be improved since the CPU doesn't know if the PC was modified and always has to write it back. ISA parser: To support the new API, all PC related operand types were removed from the parser and replaced with a PCState type. There are two warts on this implementation. First, as with all the other operand types, the PCState still has to have a valid operand type even though it doesn't use it. Second, using syntax like PCS.npc(target) doesn't work for two reasons, this looks like the syntax for operand type overriding, and the parser can't figure out if you're reading or writing. Instructions that use the PCS operand (which I've consistently called it) need to first read it into a local variable, manipulate it, and then write it back out. Return address stack: The return address stack needed a little extra help because, in the presence of branch delay slots, it has to merge together elements of the return PC and the call PC. To handle that, a buildRetPC utility function was added. There are basically only two versions in all the ISAs, but it didn't seem short enough to put into the generic ISA directory. Also, the branch predictor code in O3 and InOrder were adjusted so that they always store the PC of the actual call instruction in the RAS, not the next PC. If the call instruction is a microop, the next PC refers to the next microop in the same macroop which is probably not desirable. The buildRetPC function advances the PC intelligently to the next macroop (in an ISA specific way) so that that case works. Change in stats: There were no change in stats except in MIPS and SPARC in the O3 model. MIPS runs in about 9% fewer ticks. SPARC runs with 30%-50% fewer ticks, which could likely be improved further by setting call/return instruction flags and taking advantage of the RAS. TODO: Add != operators to the PCState classes, defined trivially to be !(a==b). Smooth out places where PCs are split apart, passed around, and put back together later. I think this might happen in SPARC's fault code. Add ISA specific constructors that allow setting PC elements without calling a bunch of accessors. Try to eliminate the need for the branching() function. Factor out Alpha's PAL mode pc bit into a separate flag field, and eliminate places where it's blindly masked out or tested in the PC. diff 7676:92274350b953 Fri Sep 10 17:58:00 EDT 2010 Nathan Binkert <nate@binkert.org> style: fix sorting of includes and whitespace in some files |
H A D | iew_impl.hh | diff 7782:9b87755cb699 Tue Dec 07 19:19:00 EST 2010 Min Kyu Jeong <minkyu.jeong@arm.com> O3: Support SWAP and predicated loads/store in ARM. diff 7720:65d338a8dba4 Sun Oct 31 03:07:00 EDT 2010 Gabe Black <gblack@eecs.umich.edu> ISA,CPU,etc: Create an ISA defined PC type that abstracts out ISA behaviors. This change is a low level and pervasive reorganization of how PCs are managed in M5. Back when Alpha was the only ISA, there were only 2 PCs to worry about, the PC and the NPC, and the lsb of the PC signaled whether or not you were in PAL mode. As other ISAs were added, we had to add an NNPC, micro PC and next micropc, x86 and ARM introduced variable length instruction sets, and ARM started to keep track of mode bits in the PC. Each CPU model handled PCs in its own custom way that needed to be updated individually to handle the new dimensions of variability, or, in the case of ARMs mode-bit-in-the-pc hack, the complexity could be hidden in the ISA at the ISA implementation's expense. Areas like the branch predictor hadn't been updated to handle branch delay slots or micropcs, and it turns out that had introduced a significant (10s of percent) performance bug in SPARC and to a lesser extend MIPS. Rather than perpetuate the problem by reworking O3 again to handle the PC features needed by x86, this change was introduced to rework PC handling in a more modular, transparent, and hopefully efficient way. PC type: Rather than having the superset of all possible elements of PC state declared in each of the CPU models, each ISA defines its own PCState type which has exactly the elements it needs. A cross product of canned PCState classes are defined in the new "generic" ISA directory for ISAs with/without delay slots and microcode. These are either typedef-ed or subclassed by each ISA. To read or write this structure through a *Context, you use the new pcState() accessor which reads or writes depending on whether it has an argument. If you just want the address of the current or next instruction or the current micro PC, you can get those through read-only accessors on either the PCState type or the *Contexts. These are instAddr(), nextInstAddr(), and microPC(). Note the move away from readPC. That name is ambiguous since it's not clear whether or not it should be the actual address to fetch from, or if it should have extra bits in it like the PAL mode bit. Each class is free to define its own functions to get at whatever values it needs however it needs to to be used in ISA specific code. Eventually Alpha's PAL mode bit could be moved out of the PC and into a separate field like ARM. These types can be reset to a particular pc (where npc = pc + sizeof(MachInst), nnpc = npc + sizeof(MachInst), upc = 0, nupc = 1 as appropriate), printed, serialized, and compared. There is a branching() function which encapsulates code in the CPU models that checked if an instruction branched or not. Exactly what that means in the context of branch delay slots which can skip an instruction when not taken is ambiguous, and ideally this function and its uses can be eliminated. PCStates also generally know how to advance themselves in various ways depending on if they point at an instruction, a microop, or the last microop of a macroop. More on that later. Ideally, accessing all the PCs at once when setting them will improve performance of M5 even though more data needs to be moved around. This is because often all the PCs need to be manipulated together, and by getting them all at once you avoid multiple function calls. Also, the PCs of a particular thread will have spatial locality in the cache. Previously they were grouped by element in arrays which spread out accesses. Advancing the PC: The PCs were previously managed entirely by the CPU which had to know about PC semantics, try to figure out which dimension to increment the PC in, what to set NPC/NNPC, etc. These decisions are best left to the ISA in conjunction with the PC type itself. Because most of the information about how to increment the PC (mainly what type of instruction it refers to) is contained in the instruction object, a new advancePC virtual function was added to the StaticInst class. Subclasses provide an implementation that moves around the right element of the PC with a minimal amount of decision making. In ISAs like Alpha, the instructions always simply assign NPC to PC without having to worry about micropcs, nnpcs, etc. The added cost of a virtual function call should be outweighed by not having to figure out as much about what to do with the PCs and mucking around with the extra elements. One drawback of making the StaticInsts advance the PC is that you have to actually have one to advance the PC. This would, superficially, seem to require decoding an instruction before fetch could advance. This is, as far as I can tell, realistic. fetch would advance through memory addresses, not PCs, perhaps predicting new memory addresses using existing ones. More sophisticated decisions about control flow would be made later on, after the instruction was decoded, and handed back to fetch. If branching needs to happen, some amount of decoding needs to happen to see that it's a branch, what the target is, etc. This could get a little more complicated if that gets done by the predecoder, but I'm choosing to ignore that for now. Variable length instructions: To handle variable length instructions in x86 and ARM, the predecoder now takes in the current PC by reference to the getExtMachInst function. It can modify the PC however it needs to (by setting NPC to be the PC + instruction length, for instance). This could be improved since the CPU doesn't know if the PC was modified and always has to write it back. ISA parser: To support the new API, all PC related operand types were removed from the parser and replaced with a PCState type. There are two warts on this implementation. First, as with all the other operand types, the PCState still has to have a valid operand type even though it doesn't use it. Second, using syntax like PCS.npc(target) doesn't work for two reasons, this looks like the syntax for operand type overriding, and the parser can't figure out if you're reading or writing. Instructions that use the PCS operand (which I've consistently called it) need to first read it into a local variable, manipulate it, and then write it back out. Return address stack: The return address stack needed a little extra help because, in the presence of branch delay slots, it has to merge together elements of the return PC and the call PC. To handle that, a buildRetPC utility function was added. There are basically only two versions in all the ISAs, but it didn't seem short enough to put into the generic ISA directory. Also, the branch predictor code in O3 and InOrder were adjusted so that they always store the PC of the actual call instruction in the RAS, not the next PC. If the call instruction is a microop, the next PC refers to the next microop in the same macroop which is probably not desirable. The buildRetPC function advances the PC intelligently to the next macroop (in an ISA specific way) so that that case works. Change in stats: There were no change in stats except in MIPS and SPARC in the O3 model. MIPS runs in about 9% fewer ticks. SPARC runs with 30%-50% fewer ticks, which could likely be improved further by setting call/return instruction flags and taking advantage of the RAS. TODO: Add != operators to the PCState classes, defined trivially to be !(a==b). Smooth out places where PCs are split apart, passed around, and put back together later. I think this might happen in SPARC's fault code. Add ISA specific constructors that allow setting PC elements without calling a bunch of accessors. Try to eliminate the need for the branching() function. Factor out Alpha's PAL mode pc bit into a separate flag field, and eliminate places where it's blindly masked out or tested in the PC. diff 7616:1a0ab2308bbe Mon Aug 23 12:18:00 EDT 2010 Min Kyu Jeong <minkyu.jeong@arm.com> O3: Skipping mem-order violation check for uncachable loads. Uncachable load is not executed until it reaches the head of the ROB, hence cannot cause one. diff 7599:f6bbf266f2c8 Mon Aug 23 12:18:00 EDT 2010 Min Kyu Jeong <minkyu.jeong@arm.com> ARM: mark msr/mrs instructions as SerializeBefore/After Since miscellaneous registers bypass wakeup logic, force serialization to resolve data dependencies through them * * * ARM: adding non-speculative/serialize flags for instructions change CPSR diff 7598:c0ae58952ed0 Mon Aug 23 12:18:00 EDT 2010 Min Kyu Jeong <minkyu.jeong@arm.com> O3: Handle loads when the destination is the PC. For loads that PC is the destination, check if the load was mispredicted again when the value being loaded returns from memory |
H A D | lsq_unit_impl.hh | diff 7782:9b87755cb699 Tue Dec 07 19:19:00 EST 2010 Min Kyu Jeong <minkyu.jeong@arm.com> O3: Support SWAP and predicated loads/store in ARM. diff 7720:65d338a8dba4 Sun Oct 31 03:07:00 EDT 2010 Gabe Black <gblack@eecs.umich.edu> ISA,CPU,etc: Create an ISA defined PC type that abstracts out ISA behaviors. This change is a low level and pervasive reorganization of how PCs are managed in M5. Back when Alpha was the only ISA, there were only 2 PCs to worry about, the PC and the NPC, and the lsb of the PC signaled whether or not you were in PAL mode. As other ISAs were added, we had to add an NNPC, micro PC and next micropc, x86 and ARM introduced variable length instruction sets, and ARM started to keep track of mode bits in the PC. Each CPU model handled PCs in its own custom way that needed to be updated individually to handle the new dimensions of variability, or, in the case of ARMs mode-bit-in-the-pc hack, the complexity could be hidden in the ISA at the ISA implementation's expense. Areas like the branch predictor hadn't been updated to handle branch delay slots or micropcs, and it turns out that had introduced a significant (10s of percent) performance bug in SPARC and to a lesser extend MIPS. Rather than perpetuate the problem by reworking O3 again to handle the PC features needed by x86, this change was introduced to rework PC handling in a more modular, transparent, and hopefully efficient way. PC type: Rather than having the superset of all possible elements of PC state declared in each of the CPU models, each ISA defines its own PCState type which has exactly the elements it needs. A cross product of canned PCState classes are defined in the new "generic" ISA directory for ISAs with/without delay slots and microcode. These are either typedef-ed or subclassed by each ISA. To read or write this structure through a *Context, you use the new pcState() accessor which reads or writes depending on whether it has an argument. If you just want the address of the current or next instruction or the current micro PC, you can get those through read-only accessors on either the PCState type or the *Contexts. These are instAddr(), nextInstAddr(), and microPC(). Note the move away from readPC. That name is ambiguous since it's not clear whether or not it should be the actual address to fetch from, or if it should have extra bits in it like the PAL mode bit. Each class is free to define its own functions to get at whatever values it needs however it needs to to be used in ISA specific code. Eventually Alpha's PAL mode bit could be moved out of the PC and into a separate field like ARM. These types can be reset to a particular pc (where npc = pc + sizeof(MachInst), nnpc = npc + sizeof(MachInst), upc = 0, nupc = 1 as appropriate), printed, serialized, and compared. There is a branching() function which encapsulates code in the CPU models that checked if an instruction branched or not. Exactly what that means in the context of branch delay slots which can skip an instruction when not taken is ambiguous, and ideally this function and its uses can be eliminated. PCStates also generally know how to advance themselves in various ways depending on if they point at an instruction, a microop, or the last microop of a macroop. More on that later. Ideally, accessing all the PCs at once when setting them will improve performance of M5 even though more data needs to be moved around. This is because often all the PCs need to be manipulated together, and by getting them all at once you avoid multiple function calls. Also, the PCs of a particular thread will have spatial locality in the cache. Previously they were grouped by element in arrays which spread out accesses. Advancing the PC: The PCs were previously managed entirely by the CPU which had to know about PC semantics, try to figure out which dimension to increment the PC in, what to set NPC/NNPC, etc. These decisions are best left to the ISA in conjunction with the PC type itself. Because most of the information about how to increment the PC (mainly what type of instruction it refers to) is contained in the instruction object, a new advancePC virtual function was added to the StaticInst class. Subclasses provide an implementation that moves around the right element of the PC with a minimal amount of decision making. In ISAs like Alpha, the instructions always simply assign NPC to PC without having to worry about micropcs, nnpcs, etc. The added cost of a virtual function call should be outweighed by not having to figure out as much about what to do with the PCs and mucking around with the extra elements. One drawback of making the StaticInsts advance the PC is that you have to actually have one to advance the PC. This would, superficially, seem to require decoding an instruction before fetch could advance. This is, as far as I can tell, realistic. fetch would advance through memory addresses, not PCs, perhaps predicting new memory addresses using existing ones. More sophisticated decisions about control flow would be made later on, after the instruction was decoded, and handed back to fetch. If branching needs to happen, some amount of decoding needs to happen to see that it's a branch, what the target is, etc. This could get a little more complicated if that gets done by the predecoder, but I'm choosing to ignore that for now. Variable length instructions: To handle variable length instructions in x86 and ARM, the predecoder now takes in the current PC by reference to the getExtMachInst function. It can modify the PC however it needs to (by setting NPC to be the PC + instruction length, for instance). This could be improved since the CPU doesn't know if the PC was modified and always has to write it back. ISA parser: To support the new API, all PC related operand types were removed from the parser and replaced with a PCState type. There are two warts on this implementation. First, as with all the other operand types, the PCState still has to have a valid operand type even though it doesn't use it. Second, using syntax like PCS.npc(target) doesn't work for two reasons, this looks like the syntax for operand type overriding, and the parser can't figure out if you're reading or writing. Instructions that use the PCS operand (which I've consistently called it) need to first read it into a local variable, manipulate it, and then write it back out. Return address stack: The return address stack needed a little extra help because, in the presence of branch delay slots, it has to merge together elements of the return PC and the call PC. To handle that, a buildRetPC utility function was added. There are basically only two versions in all the ISAs, but it didn't seem short enough to put into the generic ISA directory. Also, the branch predictor code in O3 and InOrder were adjusted so that they always store the PC of the actual call instruction in the RAS, not the next PC. If the call instruction is a microop, the next PC refers to the next microop in the same macroop which is probably not desirable. The buildRetPC function advances the PC intelligently to the next macroop (in an ISA specific way) so that that case works. Change in stats: There were no change in stats except in MIPS and SPARC in the O3 model. MIPS runs in about 9% fewer ticks. SPARC runs with 30%-50% fewer ticks, which could likely be improved further by setting call/return instruction flags and taking advantage of the RAS. TODO: Add != operators to the PCState classes, defined trivially to be !(a==b). Smooth out places where PCs are split apart, passed around, and put back together later. I think this might happen in SPARC's fault code. Add ISA specific constructors that allow setting PC elements without calling a bunch of accessors. Try to eliminate the need for the branching() function. Factor out Alpha's PAL mode pc bit into a separate flag field, and eliminate places where it's blindly masked out or tested in the PC. diff 7616:1a0ab2308bbe Mon Aug 23 12:18:00 EDT 2010 Min Kyu Jeong <minkyu.jeong@arm.com> O3: Skipping mem-order violation check for uncachable loads. Uncachable load is not executed until it reaches the head of the ROB, hence cannot cause one. diff 7600:eff7f79f7dfd Mon Aug 23 12:18:00 EDT 2010 Min Kyu Jeong <minkyu.jeong@arm.com> CPU: Make Exec trace to print predication result (if false) for memory instructions diff 7598:c0ae58952ed0 Mon Aug 23 12:18:00 EDT 2010 Min Kyu Jeong <minkyu.jeong@arm.com> O3: Handle loads when the destination is the PC. For loads that PC is the destination, check if the load was mispredicted again when the value being loaded returns from memory diff 7597:063f160e8b50 Mon Aug 23 12:18:00 EDT 2010 Min Kyu Jeong <minkyu.jeong@arm.com> ARM/O3: store the result of the predicate evaluation in DynInst or Threadstate. THis allows the CPU to handle predicated-false instructions accordingly. This particular patch makes loads that are predicated-false to be sent straight to the commit stage directly, not waiting for return of the data that was never requested since it was predicated-false. diff 6974:4d4903a3e7c5 Fri Feb 12 14:53:00 EST 2010 Timothy M. Jones <tjones1@inf.ed.ac.uk> O3PCU: Split loads and stores that cross cache line boundaries. When each load or store is sent to the LSQ, we check whether it will cross a cache line boundary and, if so, split it in two. This creates two TLB translations and two memory requests. Care has to be taken if the first packet of a split load is sent but the second blocks the cache. Similarly, for a store, if the first packet cannot be sent, we must store the second one somewhere to retry later. This modifies the LSQSenderState class to record both packets in a split load or store. Finally, a new const variable, HasUnalignedMemAcc, is added to each ISA to indicate whether unaligned memory accesses are allowed. This is used throughout the changed code so that compiler can optimise away code dealing with split requests for ISAs that don't need them. |
/gem5/src/arch/alpha/ | ||
H A D | stacktrace.cc | diff 7723:ee4ac00d0774 Mon Nov 08 14:58:00 EST 2010 Ali Saidi <Ali.Saidi@ARM.com> sim: Use forward declarations for ports. Virtual ports need TLB data which means anything touching a file in the arch directory rebuilds any file that includes system.hh which in everything. diff 7720:65d338a8dba4 Sun Oct 31 03:07:00 EDT 2010 Gabe Black <gblack@eecs.umich.edu> ISA,CPU,etc: Create an ISA defined PC type that abstracts out ISA behaviors. This change is a low level and pervasive reorganization of how PCs are managed in M5. Back when Alpha was the only ISA, there were only 2 PCs to worry about, the PC and the NPC, and the lsb of the PC signaled whether or not you were in PAL mode. As other ISAs were added, we had to add an NNPC, micro PC and next micropc, x86 and ARM introduced variable length instruction sets, and ARM started to keep track of mode bits in the PC. Each CPU model handled PCs in its own custom way that needed to be updated individually to handle the new dimensions of variability, or, in the case of ARMs mode-bit-in-the-pc hack, the complexity could be hidden in the ISA at the ISA implementation's expense. Areas like the branch predictor hadn't been updated to handle branch delay slots or micropcs, and it turns out that had introduced a significant (10s of percent) performance bug in SPARC and to a lesser extend MIPS. Rather than perpetuate the problem by reworking O3 again to handle the PC features needed by x86, this change was introduced to rework PC handling in a more modular, transparent, and hopefully efficient way. PC type: Rather than having the superset of all possible elements of PC state declared in each of the CPU models, each ISA defines its own PCState type which has exactly the elements it needs. A cross product of canned PCState classes are defined in the new "generic" ISA directory for ISAs with/without delay slots and microcode. These are either typedef-ed or subclassed by each ISA. To read or write this structure through a *Context, you use the new pcState() accessor which reads or writes depending on whether it has an argument. If you just want the address of the current or next instruction or the current micro PC, you can get those through read-only accessors on either the PCState type or the *Contexts. These are instAddr(), nextInstAddr(), and microPC(). Note the move away from readPC. That name is ambiguous since it's not clear whether or not it should be the actual address to fetch from, or if it should have extra bits in it like the PAL mode bit. Each class is free to define its own functions to get at whatever values it needs however it needs to to be used in ISA specific code. Eventually Alpha's PAL mode bit could be moved out of the PC and into a separate field like ARM. These types can be reset to a particular pc (where npc = pc + sizeof(MachInst), nnpc = npc + sizeof(MachInst), upc = 0, nupc = 1 as appropriate), printed, serialized, and compared. There is a branching() function which encapsulates code in the CPU models that checked if an instruction branched or not. Exactly what that means in the context of branch delay slots which can skip an instruction when not taken is ambiguous, and ideally this function and its uses can be eliminated. PCStates also generally know how to advance themselves in various ways depending on if they point at an instruction, a microop, or the last microop of a macroop. More on that later. Ideally, accessing all the PCs at once when setting them will improve performance of M5 even though more data needs to be moved around. This is because often all the PCs need to be manipulated together, and by getting them all at once you avoid multiple function calls. Also, the PCs of a particular thread will have spatial locality in the cache. Previously they were grouped by element in arrays which spread out accesses. Advancing the PC: The PCs were previously managed entirely by the CPU which had to know about PC semantics, try to figure out which dimension to increment the PC in, what to set NPC/NNPC, etc. These decisions are best left to the ISA in conjunction with the PC type itself. Because most of the information about how to increment the PC (mainly what type of instruction it refers to) is contained in the instruction object, a new advancePC virtual function was added to the StaticInst class. Subclasses provide an implementation that moves around the right element of the PC with a minimal amount of decision making. In ISAs like Alpha, the instructions always simply assign NPC to PC without having to worry about micropcs, nnpcs, etc. The added cost of a virtual function call should be outweighed by not having to figure out as much about what to do with the PCs and mucking around with the extra elements. One drawback of making the StaticInsts advance the PC is that you have to actually have one to advance the PC. This would, superficially, seem to require decoding an instruction before fetch could advance. This is, as far as I can tell, realistic. fetch would advance through memory addresses, not PCs, perhaps predicting new memory addresses using existing ones. More sophisticated decisions about control flow would be made later on, after the instruction was decoded, and handed back to fetch. If branching needs to happen, some amount of decoding needs to happen to see that it's a branch, what the target is, etc. This could get a little more complicated if that gets done by the predecoder, but I'm choosing to ignore that for now. Variable length instructions: To handle variable length instructions in x86 and ARM, the predecoder now takes in the current PC by reference to the getExtMachInst function. It can modify the PC however it needs to (by setting NPC to be the PC + instruction length, for instance). This could be improved since the CPU doesn't know if the PC was modified and always has to write it back. ISA parser: To support the new API, all PC related operand types were removed from the parser and replaced with a PCState type. There are two warts on this implementation. First, as with all the other operand types, the PCState still has to have a valid operand type even though it doesn't use it. Second, using syntax like PCS.npc(target) doesn't work for two reasons, this looks like the syntax for operand type overriding, and the parser can't figure out if you're reading or writing. Instructions that use the PCS operand (which I've consistently called it) need to first read it into a local variable, manipulate it, and then write it back out. Return address stack: The return address stack needed a little extra help because, in the presence of branch delay slots, it has to merge together elements of the return PC and the call PC. To handle that, a buildRetPC utility function was added. There are basically only two versions in all the ISAs, but it didn't seem short enough to put into the generic ISA directory. Also, the branch predictor code in O3 and InOrder were adjusted so that they always store the PC of the actual call instruction in the RAS, not the next PC. If the call instruction is a microop, the next PC refers to the next microop in the same macroop which is probably not desirable. The buildRetPC function advances the PC intelligently to the next macroop (in an ISA specific way) so that that case works. Change in stats: There were no change in stats except in MIPS and SPARC in the O3 model. MIPS runs in about 9% fewer ticks. SPARC runs with 30%-50% fewer ticks, which could likely be improved further by setting call/return instruction flags and taking advantage of the RAS. TODO: Add != operators to the PCState classes, defined trivially to be !(a==b). Smooth out places where PCs are split apart, passed around, and put back together later. I think this might happen in SPARC's fault code. Add ISA specific constructors that allow setting PC elements without calling a bunch of accessors. Try to eliminate the need for the branching() function. Factor out Alpha's PAL mode pc bit into a separate flag field, and eliminate places where it's blindly masked out or tested in the PC. |
/gem5/src/arch/alpha/linux/ | ||
H A D | system.cc | diff 7723:ee4ac00d0774 Mon Nov 08 14:58:00 EST 2010 Ali Saidi <Ali.Saidi@ARM.com> sim: Use forward declarations for ports. Virtual ports need TLB data which means anything touching a file in the arch directory rebuilds any file that includes system.hh which in everything. diff 7064:586b0e3a12b3 Thu Apr 15 19:24:00 EDT 2010 Nathan Binkert <nate@binkert.org> tick: rename Clock namespace to SimClock |
/gem5/src/mem/ | ||
H A D | packet.hh | diff 7691:358c00c482f7 Thu Sep 30 10:35:00 EDT 2010 Ali Saidi <Ali.Saidi@ARM.com> CPU/Cache: Fix some errors exposed by valgrind diff 7687:d1ba390671ec Wed Sep 22 02:07:00 EDT 2010 Steve Reinhardt <steve.reinhardt@amd.com> cache: improve coherence handling of writebacks If we write back an exclusive copy, we now mark it as such, so the cache receiving the writeback can mark its copy as exclusive. This avoids some unnecessary upgrade requests when a cache later tries to re-acquire exclusive access to the block. diff 7669:cc222ba29079 Thu Sep 09 14:40:00 EDT 2010 Steve Reinhardt <steve.reinhardt@amd.com> cache: fail SC when invalidated while waiting for bus Corrects an oversight in cset f97b62be544f. The fix there only failed queued SCUpgradeReq packets that encountered an invalidation, which meant that the upgrade had to reach the L2 cache. To handle pending requests in the L1 we must similarly fail StoreCondReq packets too. diff 7636:59b6a1b5bb0c Wed Aug 25 17:08:00 EDT 2010 Steve Reinhardt <steve.reinhardt@amd.com> mem: fix dumb typo in copyrights diff 7550:7d97cec15818 Fri Aug 20 14:46:00 EDT 2010 Brad Beckmann <Brad.Beckmann@amd.com> ruby: fix ruby llsc support to sync sc outcomes Added support so that ruby can determine the outcome of store conditional operations and reflect that outcome to M5 physical memory and cpus. diff 7465:f97b62be544f Wed Jun 16 18:25:00 EDT 2010 Steve Reinhardt <steve.reinhardt@amd.com> cache: fail store conditionals when upgrade loses race Requires new "SCUpgradeReq" message that marks upgrades for store conditionals, so downstream caches can fail these when they run into invalidations. See http://www.m5sim.org/flyspray/task/197 diff 7464:8d92c2737ac8 Wed Jun 16 18:25:00 EDT 2010 Steve Reinhardt <steve.reinhardt@amd.com> cache: fix dirty bit setting Only set the dirty bit when we actually write to a block (not if we thought we might but didn't, as in a failed SC or CAS). This requires makeing sure the dirty bit stays set when we get an exclusive (writable) copy in a cache-to-cache transfer from another owner, which n turn requires copying the mem-inhibit flag from timing-mode requests to their associated responses. diff 7006:77c9c4d5007d Fri Mar 12 20:31:00 EST 2010 Nathan Binkert <nate@binkert.org> packet: add a method to set the size |
H A D | tport.cc | diff 7510:fb7fc9aca918 Thu Jul 22 13:54:00 EDT 2010 Timothy M. Jones <tjones1@inf.ed.ac.uk> Port: Only indicate that a SimpleTimingPort is drained if its send event is not scheduled, as well as the transmit list being empty. diff 7064:586b0e3a12b3 Thu Apr 15 19:24:00 EDT 2010 Nathan Binkert <nate@binkert.org> tick: rename Clock namespace to SimClock |
/gem5/src/sim/ | ||
H A D | System.py | diff 7673:b28bd1fa9a35 Thu Sep 09 17:15:00 EDT 2010 Nathan Binkert <nate@binkert.org> scons: use code_formatter wherever we can in the build system diff 7580:6f77f379a594 Mon Aug 23 12:18:00 EDT 2010 Ali Saidi <Ali.Saidi@arm.com> Loader: Make the load address mask be a parameter of the system rather than a constant. This allows one two different OS requirements for the same ISA to be handled. Some OSes are compiled for a virtual address and need to be loaded into physical memory that starts at address 0, while other bare metal tools generate images that start at address 0. |
H A D | system.cc | diff 7770:6286bb50127e Fri Nov 19 19:01:00 EST 2010 Ali Saidi <Ali.Saidi@ARM.com> SE: Fix simulating more than 4GB of RAM in SE mode This change removes some dead code in PhysicalMemory, uses a 64 bit type for the page pointer in System (instead of 32 bit) and cleans up some style. diff 7733:08d6a773d1b6 Mon Nov 08 14:58:00 EST 2010 Ali Saidi <Ali.Saidi@ARM.com> ARM: Add checkpointing support diff 7723:ee4ac00d0774 Mon Nov 08 14:58:00 EST 2010 Ali Saidi <Ali.Saidi@ARM.com> sim: Use forward declarations for ports. Virtual ports need TLB data which means anything touching a file in the arch directory rebuilds any file that includes system.hh which in everything. diff 7580:6f77f379a594 Mon Aug 23 12:18:00 EDT 2010 Ali Saidi <Ali.Saidi@arm.com> Loader: Make the load address mask be a parameter of the system rather than a constant. This allows one two different OS requirements for the same ISA to be handled. Some OSes are compiled for a virtual address and need to be loaded into physical memory that starts at address 0, while other bare metal tools generate images that start at address 0. diff 7445:dfd04ffc1773 Thu Jun 03 19:54:00 EDT 2010 Steve Reinhardt <steve.reinhardt@amd.com> Minor remote GDB cleanup. Expand the help text on the --remote-gdb-port option so people know you can use it to disable remote gdb without reading the source code, and thus don't waste any time trying to add a separate option to do that. Clean up some gdb-related cruft I found while looking for where one would add a gdb disable option, before I found the comment that told me that I didn't need to do that. |
H A D | process.cc | diff 7532:3f6413fc37a2 Tue Aug 17 08:17:00 EDT 2010 Steve Reinhardt <steve.reinhardt@amd.com> sim: revamp unserialization procedure Replace direct call to unserialize() on each SimObject with a pair of calls for better control over initialization in both ckpt and non-ckpt cases. If restoring from a checkpoint, loadState(ckpt) is called on each SimObject. The default implementation simply calls unserialize() if there is a corresponding checkpoint section, so we get backward compatibility for existing objects. However, objects can override loadState() to get other behaviors, e.g., doing other programmed initializations after unserialize(), or complaining if no checkpoint section is found. (Note that the default warning for a missing checkpoint section is now gone.) If not restoring from a checkpoint, we call the new initState() method on each SimObject instead. This provides a hook for state initializations that are only required when *not* restoring from a checkpoint. Given this new framework, do some cleanup of LiveProcess subclasses and X86System, which were (in some cases) emulating initState() behavior in startup via a local flag or (in other cases) erroneously doing initializations in startup() that clobbered state loaded earlier by unserialize(). diff 7488:cf7c4345ea82 Tue Jul 06 00:39:00 EDT 2010 Steve Reinhardt <steve.reinhardt@amd.com> process: get rid of some unused code & vars diff 7487:2a5e4070155e Tue Jul 06 00:39:00 EDT 2010 Steve Reinhardt <steve.reinhardt@amd.com> process: minor format/style cleanup diff 7447:3fc243687abb Thu Jun 03 22:41:00 EDT 2010 Steve Reinhardt <steve.reinhardt@amd.com> More minor gdb-related cleanup. Found several more stale includes and forward decls. diff 7096:e81026b9dfe0 Wed Jun 02 01:58:00 EDT 2010 Gabe Black <gblack@eecs.umich.edu> ARM: Allow ARM processes to start in Thumb mode. diff 7095:f7bb2d93cc49 Wed Jun 02 01:58:00 EDT 2010 Gabe Black <gblack@eecs.umich.edu> ARM: Detect thumb mode elf images. diff 6820:2980bd04e6df Wed Jan 20 01:03:00 EST 2010 Lisa Hsu <Lisa.Hsu@amd.com> util: do checkpoint aggregation more cleanly, fix last changeset. 1) Move alpha-specific code out of page_table.cc:serialize(). 2) Begin serializing M5_pid and unserializing it, but adding an function to do optional paramIn so that old checkpoints don't need to be fixed up. 3) Fix up alpha startup code so that the unserialized M5_pid value is properly written to DTB_IPR_ASN. 4) Fix the memory unserialize that I forgot somehow in the last changeset. 5) Add in an agg_se.py to handle aggregated checkpoints. --bench foo-bar plus positional arguments foo bar are the only changes in usage from se.py. Note this aggregation stuff has only been tested for Alpha and nothing else, though it should take a very minimal amount of work to get it to work with another ISA. |
/gem5/src/arch/arm/ | ||
H A D | ArmSystem.py | diff 7585:afbc40280b56 Mon Aug 23 12:18:00 EDT 2010 Ali Saidi <Ali.Saidi@arm.com> ARM: Add system for ARM/Linux and bootstrapping diff 7580:6f77f379a594 Mon Aug 23 12:18:00 EDT 2010 Ali Saidi <Ali.Saidi@arm.com> Loader: Make the load address mask be a parameter of the system rather than a constant. This allows one two different OS requirements for the same ISA to be handled. Some OSes are compiled for a virtual address and need to be loaded into physical memory that starts at address 0, while other bare metal tools generate images that start at address 0. |
/gem5/src/arch/x86/ | ||
H A D | interrupts.cc | diff 7629:0f0c231e3e97 Mon Aug 23 19:14:00 EDT 2010 Gabe Black <gblack@eecs.umich.edu> X86: Create a directory for files that define register indexes. This is to help tidy up arch/x86. These files should not be used external to the ISA. diff 7087:fb8d5786ff30 Mon May 24 01:44:00 EDT 2010 Nathan Binkert <nate@binkert.org> copyright: Change HP copyright on x86 code to be more friendly |
/gem5/src/base/ | ||
H A D | types.hh | diff 7720:65d338a8dba4 Sun Oct 31 03:07:00 EDT 2010 Gabe Black <gblack@eecs.umich.edu> ISA,CPU,etc: Create an ISA defined PC type that abstracts out ISA behaviors. This change is a low level and pervasive reorganization of how PCs are managed in M5. Back when Alpha was the only ISA, there were only 2 PCs to worry about, the PC and the NPC, and the lsb of the PC signaled whether or not you were in PAL mode. As other ISAs were added, we had to add an NNPC, micro PC and next micropc, x86 and ARM introduced variable length instruction sets, and ARM started to keep track of mode bits in the PC. Each CPU model handled PCs in its own custom way that needed to be updated individually to handle the new dimensions of variability, or, in the case of ARMs mode-bit-in-the-pc hack, the complexity could be hidden in the ISA at the ISA implementation's expense. Areas like the branch predictor hadn't been updated to handle branch delay slots or micropcs, and it turns out that had introduced a significant (10s of percent) performance bug in SPARC and to a lesser extend MIPS. Rather than perpetuate the problem by reworking O3 again to handle the PC features needed by x86, this change was introduced to rework PC handling in a more modular, transparent, and hopefully efficient way. PC type: Rather than having the superset of all possible elements of PC state declared in each of the CPU models, each ISA defines its own PCState type which has exactly the elements it needs. A cross product of canned PCState classes are defined in the new "generic" ISA directory for ISAs with/without delay slots and microcode. These are either typedef-ed or subclassed by each ISA. To read or write this structure through a *Context, you use the new pcState() accessor which reads or writes depending on whether it has an argument. If you just want the address of the current or next instruction or the current micro PC, you can get those through read-only accessors on either the PCState type or the *Contexts. These are instAddr(), nextInstAddr(), and microPC(). Note the move away from readPC. That name is ambiguous since it's not clear whether or not it should be the actual address to fetch from, or if it should have extra bits in it like the PAL mode bit. Each class is free to define its own functions to get at whatever values it needs however it needs to to be used in ISA specific code. Eventually Alpha's PAL mode bit could be moved out of the PC and into a separate field like ARM. These types can be reset to a particular pc (where npc = pc + sizeof(MachInst), nnpc = npc + sizeof(MachInst), upc = 0, nupc = 1 as appropriate), printed, serialized, and compared. There is a branching() function which encapsulates code in the CPU models that checked if an instruction branched or not. Exactly what that means in the context of branch delay slots which can skip an instruction when not taken is ambiguous, and ideally this function and its uses can be eliminated. PCStates also generally know how to advance themselves in various ways depending on if they point at an instruction, a microop, or the last microop of a macroop. More on that later. Ideally, accessing all the PCs at once when setting them will improve performance of M5 even though more data needs to be moved around. This is because often all the PCs need to be manipulated together, and by getting them all at once you avoid multiple function calls. Also, the PCs of a particular thread will have spatial locality in the cache. Previously they were grouped by element in arrays which spread out accesses. Advancing the PC: The PCs were previously managed entirely by the CPU which had to know about PC semantics, try to figure out which dimension to increment the PC in, what to set NPC/NNPC, etc. These decisions are best left to the ISA in conjunction with the PC type itself. Because most of the information about how to increment the PC (mainly what type of instruction it refers to) is contained in the instruction object, a new advancePC virtual function was added to the StaticInst class. Subclasses provide an implementation that moves around the right element of the PC with a minimal amount of decision making. In ISAs like Alpha, the instructions always simply assign NPC to PC without having to worry about micropcs, nnpcs, etc. The added cost of a virtual function call should be outweighed by not having to figure out as much about what to do with the PCs and mucking around with the extra elements. One drawback of making the StaticInsts advance the PC is that you have to actually have one to advance the PC. This would, superficially, seem to require decoding an instruction before fetch could advance. This is, as far as I can tell, realistic. fetch would advance through memory addresses, not PCs, perhaps predicting new memory addresses using existing ones. More sophisticated decisions about control flow would be made later on, after the instruction was decoded, and handed back to fetch. If branching needs to happen, some amount of decoding needs to happen to see that it's a branch, what the target is, etc. This could get a little more complicated if that gets done by the predecoder, but I'm choosing to ignore that for now. Variable length instructions: To handle variable length instructions in x86 and ARM, the predecoder now takes in the current PC by reference to the getExtMachInst function. It can modify the PC however it needs to (by setting NPC to be the PC + instruction length, for instance). This could be improved since the CPU doesn't know if the PC was modified and always has to write it back. ISA parser: To support the new API, all PC related operand types were removed from the parser and replaced with a PCState type. There are two warts on this implementation. First, as with all the other operand types, the PCState still has to have a valid operand type even though it doesn't use it. Second, using syntax like PCS.npc(target) doesn't work for two reasons, this looks like the syntax for operand type overriding, and the parser can't figure out if you're reading or writing. Instructions that use the PCS operand (which I've consistently called it) need to first read it into a local variable, manipulate it, and then write it back out. Return address stack: The return address stack needed a little extra help because, in the presence of branch delay slots, it has to merge together elements of the return PC and the call PC. To handle that, a buildRetPC utility function was added. There are basically only two versions in all the ISAs, but it didn't seem short enough to put into the generic ISA directory. Also, the branch predictor code in O3 and InOrder were adjusted so that they always store the PC of the actual call instruction in the RAS, not the next PC. If the call instruction is a microop, the next PC refers to the next microop in the same macroop which is probably not desirable. The buildRetPC function advances the PC intelligently to the next macroop (in an ISA specific way) so that that case works. Change in stats: There were no change in stats except in MIPS and SPARC in the O3 model. MIPS runs in about 9% fewer ticks. SPARC runs with 30%-50% fewer ticks, which could likely be improved further by setting call/return instruction flags and taking advantage of the RAS. TODO: Add != operators to the PCState classes, defined trivially to be !(a==b). Smooth out places where PCs are split apart, passed around, and put back together later. I think this might happen in SPARC's fault code. Add ISA specific constructors that allow setting PC elements without calling a bunch of accessors. Try to eliminate the need for the branching() function. Factor out Alpha's PAL mode pc bit into a separate flag field, and eliminate places where it's blindly masked out or tested in the PC. diff 7678:f19b6a3a8cec Mon Sep 13 22:26:00 EDT 2010 Gabe Black <gblack@eecs.umich.edu> Faults: Pass the StaticInst involved, if any, to a Fault's invoke method. Also move the "Fault" reference counted pointer type into a separate file, sim/fault.hh. It would be better to name this less similarly to sim/faults.hh to reduce confusion, but fault.hh matches the name of the type. We could change Fault to FaultPtr to match other pointer types, and then changing the name of the file would make more sense. |
/gem5/src/arch/sparc/isa/ | ||
H A D | decoder.isa | diff 7790:9df469679ac7 Wed Dec 08 03:27:00 EST 2010 Gabe Black <gblack@eecs.umich.edu> SPARC: Take advantage of new PCState syntax. diff 7784:e7649570ff3a Tue Dec 07 19:19:00 EST 2010 Ali Saidi <Ali.Saidi@ARM.com> O3: Support squashing all state after special instruction For SPARC ASIs are added to the ExtMachInst. If the ASI is changed simply marking the instruction as Serializing isn't enough beacuse that only stops rename. This provides a mechanism to squash all the instructions and refetch them diff 7741:340b6f01d69b Thu Nov 11 05:03:00 EST 2010 Gabe Black <gblack@eecs.umich.edu> SPARC: Clean up some historical style issues. diff 7720:65d338a8dba4 Sun Oct 31 03:07:00 EDT 2010 Gabe Black <gblack@eecs.umich.edu> ISA,CPU,etc: Create an ISA defined PC type that abstracts out ISA behaviors. This change is a low level and pervasive reorganization of how PCs are managed in M5. Back when Alpha was the only ISA, there were only 2 PCs to worry about, the PC and the NPC, and the lsb of the PC signaled whether or not you were in PAL mode. As other ISAs were added, we had to add an NNPC, micro PC and next micropc, x86 and ARM introduced variable length instruction sets, and ARM started to keep track of mode bits in the PC. Each CPU model handled PCs in its own custom way that needed to be updated individually to handle the new dimensions of variability, or, in the case of ARMs mode-bit-in-the-pc hack, the complexity could be hidden in the ISA at the ISA implementation's expense. Areas like the branch predictor hadn't been updated to handle branch delay slots or micropcs, and it turns out that had introduced a significant (10s of percent) performance bug in SPARC and to a lesser extend MIPS. Rather than perpetuate the problem by reworking O3 again to handle the PC features needed by x86, this change was introduced to rework PC handling in a more modular, transparent, and hopefully efficient way. PC type: Rather than having the superset of all possible elements of PC state declared in each of the CPU models, each ISA defines its own PCState type which has exactly the elements it needs. A cross product of canned PCState classes are defined in the new "generic" ISA directory for ISAs with/without delay slots and microcode. These are either typedef-ed or subclassed by each ISA. To read or write this structure through a *Context, you use the new pcState() accessor which reads or writes depending on whether it has an argument. If you just want the address of the current or next instruction or the current micro PC, you can get those through read-only accessors on either the PCState type or the *Contexts. These are instAddr(), nextInstAddr(), and microPC(). Note the move away from readPC. That name is ambiguous since it's not clear whether or not it should be the actual address to fetch from, or if it should have extra bits in it like the PAL mode bit. Each class is free to define its own functions to get at whatever values it needs however it needs to to be used in ISA specific code. Eventually Alpha's PAL mode bit could be moved out of the PC and into a separate field like ARM. These types can be reset to a particular pc (where npc = pc + sizeof(MachInst), nnpc = npc + sizeof(MachInst), upc = 0, nupc = 1 as appropriate), printed, serialized, and compared. There is a branching() function which encapsulates code in the CPU models that checked if an instruction branched or not. Exactly what that means in the context of branch delay slots which can skip an instruction when not taken is ambiguous, and ideally this function and its uses can be eliminated. PCStates also generally know how to advance themselves in various ways depending on if they point at an instruction, a microop, or the last microop of a macroop. More on that later. Ideally, accessing all the PCs at once when setting them will improve performance of M5 even though more data needs to be moved around. This is because often all the PCs need to be manipulated together, and by getting them all at once you avoid multiple function calls. Also, the PCs of a particular thread will have spatial locality in the cache. Previously they were grouped by element in arrays which spread out accesses. Advancing the PC: The PCs were previously managed entirely by the CPU which had to know about PC semantics, try to figure out which dimension to increment the PC in, what to set NPC/NNPC, etc. These decisions are best left to the ISA in conjunction with the PC type itself. Because most of the information about how to increment the PC (mainly what type of instruction it refers to) is contained in the instruction object, a new advancePC virtual function was added to the StaticInst class. Subclasses provide an implementation that moves around the right element of the PC with a minimal amount of decision making. In ISAs like Alpha, the instructions always simply assign NPC to PC without having to worry about micropcs, nnpcs, etc. The added cost of a virtual function call should be outweighed by not having to figure out as much about what to do with the PCs and mucking around with the extra elements. One drawback of making the StaticInsts advance the PC is that you have to actually have one to advance the PC. This would, superficially, seem to require decoding an instruction before fetch could advance. This is, as far as I can tell, realistic. fetch would advance through memory addresses, not PCs, perhaps predicting new memory addresses using existing ones. More sophisticated decisions about control flow would be made later on, after the instruction was decoded, and handed back to fetch. If branching needs to happen, some amount of decoding needs to happen to see that it's a branch, what the target is, etc. This could get a little more complicated if that gets done by the predecoder, but I'm choosing to ignore that for now. Variable length instructions: To handle variable length instructions in x86 and ARM, the predecoder now takes in the current PC by reference to the getExtMachInst function. It can modify the PC however it needs to (by setting NPC to be the PC + instruction length, for instance). This could be improved since the CPU doesn't know if the PC was modified and always has to write it back. ISA parser: To support the new API, all PC related operand types were removed from the parser and replaced with a PCState type. There are two warts on this implementation. First, as with all the other operand types, the PCState still has to have a valid operand type even though it doesn't use it. Second, using syntax like PCS.npc(target) doesn't work for two reasons, this looks like the syntax for operand type overriding, and the parser can't figure out if you're reading or writing. Instructions that use the PCS operand (which I've consistently called it) need to first read it into a local variable, manipulate it, and then write it back out. Return address stack: The return address stack needed a little extra help because, in the presence of branch delay slots, it has to merge together elements of the return PC and the call PC. To handle that, a buildRetPC utility function was added. There are basically only two versions in all the ISAs, but it didn't seem short enough to put into the generic ISA directory. Also, the branch predictor code in O3 and InOrder were adjusted so that they always store the PC of the actual call instruction in the RAS, not the next PC. If the call instruction is a microop, the next PC refers to the next microop in the same macroop which is probably not desirable. The buildRetPC function advances the PC intelligently to the next macroop (in an ISA specific way) so that that case works. Change in stats: There were no change in stats except in MIPS and SPARC in the O3 model. MIPS runs in about 9% fewer ticks. SPARC runs with 30%-50% fewer ticks, which could likely be improved further by setting call/return instruction flags and taking advantage of the RAS. TODO: Add != operators to the PCState classes, defined trivially to be !(a==b). Smooth out places where PCs are split apart, passed around, and put back together later. I think this might happen in SPARC's fault code. Add ISA specific constructors that allow setting PC elements without calling a bunch of accessors. Try to eliminate the need for the branching() function. Factor out Alpha's PAL mode pc bit into a separate flag field, and eliminate places where it's blindly masked out or tested in the PC. diff 7085:e4c5fbbc8633 Fri May 14 17:22:00 EDT 2010 Gabe Black <gblack@eecs.umich.edu> SPARC: Implement the version of movcc that uses the fp condition codes. |
/gem5/src/cpu/ | ||
H A D | thread_context.hh | diff 7720:65d338a8dba4 Sun Oct 31 03:07:00 EDT 2010 Gabe Black <gblack@eecs.umich.edu> ISA,CPU,etc: Create an ISA defined PC type that abstracts out ISA behaviors. This change is a low level and pervasive reorganization of how PCs are managed in M5. Back when Alpha was the only ISA, there were only 2 PCs to worry about, the PC and the NPC, and the lsb of the PC signaled whether or not you were in PAL mode. As other ISAs were added, we had to add an NNPC, micro PC and next micropc, x86 and ARM introduced variable length instruction sets, and ARM started to keep track of mode bits in the PC. Each CPU model handled PCs in its own custom way that needed to be updated individually to handle the new dimensions of variability, or, in the case of ARMs mode-bit-in-the-pc hack, the complexity could be hidden in the ISA at the ISA implementation's expense. Areas like the branch predictor hadn't been updated to handle branch delay slots or micropcs, and it turns out that had introduced a significant (10s of percent) performance bug in SPARC and to a lesser extend MIPS. Rather than perpetuate the problem by reworking O3 again to handle the PC features needed by x86, this change was introduced to rework PC handling in a more modular, transparent, and hopefully efficient way. PC type: Rather than having the superset of all possible elements of PC state declared in each of the CPU models, each ISA defines its own PCState type which has exactly the elements it needs. A cross product of canned PCState classes are defined in the new "generic" ISA directory for ISAs with/without delay slots and microcode. These are either typedef-ed or subclassed by each ISA. To read or write this structure through a *Context, you use the new pcState() accessor which reads or writes depending on whether it has an argument. If you just want the address of the current or next instruction or the current micro PC, you can get those through read-only accessors on either the PCState type or the *Contexts. These are instAddr(), nextInstAddr(), and microPC(). Note the move away from readPC. That name is ambiguous since it's not clear whether or not it should be the actual address to fetch from, or if it should have extra bits in it like the PAL mode bit. Each class is free to define its own functions to get at whatever values it needs however it needs to to be used in ISA specific code. Eventually Alpha's PAL mode bit could be moved out of the PC and into a separate field like ARM. These types can be reset to a particular pc (where npc = pc + sizeof(MachInst), nnpc = npc + sizeof(MachInst), upc = 0, nupc = 1 as appropriate), printed, serialized, and compared. There is a branching() function which encapsulates code in the CPU models that checked if an instruction branched or not. Exactly what that means in the context of branch delay slots which can skip an instruction when not taken is ambiguous, and ideally this function and its uses can be eliminated. PCStates also generally know how to advance themselves in various ways depending on if they point at an instruction, a microop, or the last microop of a macroop. More on that later. Ideally, accessing all the PCs at once when setting them will improve performance of M5 even though more data needs to be moved around. This is because often all the PCs need to be manipulated together, and by getting them all at once you avoid multiple function calls. Also, the PCs of a particular thread will have spatial locality in the cache. Previously they were grouped by element in arrays which spread out accesses. Advancing the PC: The PCs were previously managed entirely by the CPU which had to know about PC semantics, try to figure out which dimension to increment the PC in, what to set NPC/NNPC, etc. These decisions are best left to the ISA in conjunction with the PC type itself. Because most of the information about how to increment the PC (mainly what type of instruction it refers to) is contained in the instruction object, a new advancePC virtual function was added to the StaticInst class. Subclasses provide an implementation that moves around the right element of the PC with a minimal amount of decision making. In ISAs like Alpha, the instructions always simply assign NPC to PC without having to worry about micropcs, nnpcs, etc. The added cost of a virtual function call should be outweighed by not having to figure out as much about what to do with the PCs and mucking around with the extra elements. One drawback of making the StaticInsts advance the PC is that you have to actually have one to advance the PC. This would, superficially, seem to require decoding an instruction before fetch could advance. This is, as far as I can tell, realistic. fetch would advance through memory addresses, not PCs, perhaps predicting new memory addresses using existing ones. More sophisticated decisions about control flow would be made later on, after the instruction was decoded, and handed back to fetch. If branching needs to happen, some amount of decoding needs to happen to see that it's a branch, what the target is, etc. This could get a little more complicated if that gets done by the predecoder, but I'm choosing to ignore that for now. Variable length instructions: To handle variable length instructions in x86 and ARM, the predecoder now takes in the current PC by reference to the getExtMachInst function. It can modify the PC however it needs to (by setting NPC to be the PC + instruction length, for instance). This could be improved since the CPU doesn't know if the PC was modified and always has to write it back. ISA parser: To support the new API, all PC related operand types were removed from the parser and replaced with a PCState type. There are two warts on this implementation. First, as with all the other operand types, the PCState still has to have a valid operand type even though it doesn't use it. Second, using syntax like PCS.npc(target) doesn't work for two reasons, this looks like the syntax for operand type overriding, and the parser can't figure out if you're reading or writing. Instructions that use the PCS operand (which I've consistently called it) need to first read it into a local variable, manipulate it, and then write it back out. Return address stack: The return address stack needed a little extra help because, in the presence of branch delay slots, it has to merge together elements of the return PC and the call PC. To handle that, a buildRetPC utility function was added. There are basically only two versions in all the ISAs, but it didn't seem short enough to put into the generic ISA directory. Also, the branch predictor code in O3 and InOrder were adjusted so that they always store the PC of the actual call instruction in the RAS, not the next PC. If the call instruction is a microop, the next PC refers to the next microop in the same macroop which is probably not desirable. The buildRetPC function advances the PC intelligently to the next macroop (in an ISA specific way) so that that case works. Change in stats: There were no change in stats except in MIPS and SPARC in the O3 model. MIPS runs in about 9% fewer ticks. SPARC runs with 30%-50% fewer ticks, which could likely be improved further by setting call/return instruction flags and taking advantage of the RAS. TODO: Add != operators to the PCState classes, defined trivially to be !(a==b). Smooth out places where PCs are split apart, passed around, and put back together later. I think this might happen in SPARC's fault code. Add ISA specific constructors that allow setting PC elements without calling a bunch of accessors. Try to eliminate the need for the branching() function. Factor out Alpha's PAL mode pc bit into a separate flag field, and eliminate places where it's blindly masked out or tested in the PC. diff 7680:f4eda002333b Tue Sep 14 03:29:00 EDT 2010 Gabe Black <gblack@eecs.umich.edu> CPU: Trim unnecessary includes from some common files. This reduces the scope of those includes and makes it less likely for there to be a dependency loop. This also moves the hashing functions associated with ExtMachInst objects to be with the ExtMachInst definitions and out of utility.hh. diff 7679:f26cc2c68b48 Tue Sep 14 00:58:00 EDT 2010 Gabe Black <gblack@eecs.umich.edu> CPU: Get rid of the now unnecessary getInst/setInst family of functions. This code is no longer needed because of the preceeding change which adds a StaticInstPtr parameter to the fault's invoke method, obviating the only use for this pair of functions. diff 7678:f19b6a3a8cec Mon Sep 13 22:26:00 EDT 2010 Gabe Black <gblack@eecs.umich.edu> Faults: Pass the StaticInst involved, if any, to a Fault's invoke method. Also move the "Fault" reference counted pointer type into a separate file, sim/fault.hh. It would be better to name this less similarly to sim/faults.hh to reduce confusion, but fault.hh matches the name of the type. We could change Fault to FaultPtr to match other pointer types, and then changing the name of the file would make more sense. diff 7597:063f160e8b50 Mon Aug 23 12:18:00 EDT 2010 Min Kyu Jeong <minkyu.jeong@arm.com> ARM/O3: store the result of the predicate evaluation in DynInst or Threadstate. THis allows the CPU to handle predicated-false instructions accordingly. This particular patch makes loads that are predicated-false to be sent straight to the commit stage directly, not waiting for return of the data that was never requested since it was predicated-false. |
/gem5/configs/common/ | ||
H A D | FSConfig.py | diff 7750:0731d632db76 Mon Nov 15 15:04:00 EST 2010 Ali Saidi <Ali.Saidi@ARM.com> ARM: Add support for a dumb IDE controller diff 7730:982b4c6c1470 Mon Nov 08 14:58:00 EST 2010 Ali Saidi <Ali.Saidi@ARM.com> Mem: Finish half-baked support for mmaping file in physmem. Physmem has a parameter to be able to mem map a file, however it isn't actually used. This changeset utilizes the parameter so a file can be mmapped. diff 7633:d8112aa18a1b Tue Aug 24 16:20:00 EDT 2010 Brad Beckmann <Brad.Beckmann@amd.com> config: fixed ruby dma device connections diff 7586:da93206873dc Mon Aug 23 12:18:00 EDT 2010 Ali Saidi <Ali.Saidi@arm.com> ARM: Add configuration for Linux/Full System diff 7069:edde97a6ea7c Mon Apr 19 00:33:00 EDT 2010 Nathan Binkert <nate@binkert.org> config: fix assertion for x86 in FSConfig.py diff 7014:441317194b08 Mon Mar 22 00:22:00 EDT 2010 Brad Beckmann <Brad.Beckmann@amd.com> ruby: fixed how ruby_fs creates phsyical memory Now ruby_fs creates physical memory of the right size. diff 6893:9cdf9b65d946 Fri Jan 29 23:29:00 EST 2010 Brad Beckmann <Brad.Beckmann@amd.com> ruby: FS support using the new configuration system |
/gem5/src/cpu/simple/ | ||
H A D | atomic.cc | diff 7720:65d338a8dba4 Sun Oct 31 03:07:00 EDT 2010 Gabe Black <gblack@eecs.umich.edu> ISA,CPU,etc: Create an ISA defined PC type that abstracts out ISA behaviors. This change is a low level and pervasive reorganization of how PCs are managed in M5. Back when Alpha was the only ISA, there were only 2 PCs to worry about, the PC and the NPC, and the lsb of the PC signaled whether or not you were in PAL mode. As other ISAs were added, we had to add an NNPC, micro PC and next micropc, x86 and ARM introduced variable length instruction sets, and ARM started to keep track of mode bits in the PC. Each CPU model handled PCs in its own custom way that needed to be updated individually to handle the new dimensions of variability, or, in the case of ARMs mode-bit-in-the-pc hack, the complexity could be hidden in the ISA at the ISA implementation's expense. Areas like the branch predictor hadn't been updated to handle branch delay slots or micropcs, and it turns out that had introduced a significant (10s of percent) performance bug in SPARC and to a lesser extend MIPS. Rather than perpetuate the problem by reworking O3 again to handle the PC features needed by x86, this change was introduced to rework PC handling in a more modular, transparent, and hopefully efficient way. PC type: Rather than having the superset of all possible elements of PC state declared in each of the CPU models, each ISA defines its own PCState type which has exactly the elements it needs. A cross product of canned PCState classes are defined in the new "generic" ISA directory for ISAs with/without delay slots and microcode. These are either typedef-ed or subclassed by each ISA. To read or write this structure through a *Context, you use the new pcState() accessor which reads or writes depending on whether it has an argument. If you just want the address of the current or next instruction or the current micro PC, you can get those through read-only accessors on either the PCState type or the *Contexts. These are instAddr(), nextInstAddr(), and microPC(). Note the move away from readPC. That name is ambiguous since it's not clear whether or not it should be the actual address to fetch from, or if it should have extra bits in it like the PAL mode bit. Each class is free to define its own functions to get at whatever values it needs however it needs to to be used in ISA specific code. Eventually Alpha's PAL mode bit could be moved out of the PC and into a separate field like ARM. These types can be reset to a particular pc (where npc = pc + sizeof(MachInst), nnpc = npc + sizeof(MachInst), upc = 0, nupc = 1 as appropriate), printed, serialized, and compared. There is a branching() function which encapsulates code in the CPU models that checked if an instruction branched or not. Exactly what that means in the context of branch delay slots which can skip an instruction when not taken is ambiguous, and ideally this function and its uses can be eliminated. PCStates also generally know how to advance themselves in various ways depending on if they point at an instruction, a microop, or the last microop of a macroop. More on that later. Ideally, accessing all the PCs at once when setting them will improve performance of M5 even though more data needs to be moved around. This is because often all the PCs need to be manipulated together, and by getting them all at once you avoid multiple function calls. Also, the PCs of a particular thread will have spatial locality in the cache. Previously they were grouped by element in arrays which spread out accesses. Advancing the PC: The PCs were previously managed entirely by the CPU which had to know about PC semantics, try to figure out which dimension to increment the PC in, what to set NPC/NNPC, etc. These decisions are best left to the ISA in conjunction with the PC type itself. Because most of the information about how to increment the PC (mainly what type of instruction it refers to) is contained in the instruction object, a new advancePC virtual function was added to the StaticInst class. Subclasses provide an implementation that moves around the right element of the PC with a minimal amount of decision making. In ISAs like Alpha, the instructions always simply assign NPC to PC without having to worry about micropcs, nnpcs, etc. The added cost of a virtual function call should be outweighed by not having to figure out as much about what to do with the PCs and mucking around with the extra elements. One drawback of making the StaticInsts advance the PC is that you have to actually have one to advance the PC. This would, superficially, seem to require decoding an instruction before fetch could advance. This is, as far as I can tell, realistic. fetch would advance through memory addresses, not PCs, perhaps predicting new memory addresses using existing ones. More sophisticated decisions about control flow would be made later on, after the instruction was decoded, and handed back to fetch. If branching needs to happen, some amount of decoding needs to happen to see that it's a branch, what the target is, etc. This could get a little more complicated if that gets done by the predecoder, but I'm choosing to ignore that for now. Variable length instructions: To handle variable length instructions in x86 and ARM, the predecoder now takes in the current PC by reference to the getExtMachInst function. It can modify the PC however it needs to (by setting NPC to be the PC + instruction length, for instance). This could be improved since the CPU doesn't know if the PC was modified and always has to write it back. ISA parser: To support the new API, all PC related operand types were removed from the parser and replaced with a PCState type. There are two warts on this implementation. First, as with all the other operand types, the PCState still has to have a valid operand type even though it doesn't use it. Second, using syntax like PCS.npc(target) doesn't work for two reasons, this looks like the syntax for operand type overriding, and the parser can't figure out if you're reading or writing. Instructions that use the PCS operand (which I've consistently called it) need to first read it into a local variable, manipulate it, and then write it back out. Return address stack: The return address stack needed a little extra help because, in the presence of branch delay slots, it has to merge together elements of the return PC and the call PC. To handle that, a buildRetPC utility function was added. There are basically only two versions in all the ISAs, but it didn't seem short enough to put into the generic ISA directory. Also, the branch predictor code in O3 and InOrder were adjusted so that they always store the PC of the actual call instruction in the RAS, not the next PC. If the call instruction is a microop, the next PC refers to the next microop in the same macroop which is probably not desirable. The buildRetPC function advances the PC intelligently to the next macroop (in an ISA specific way) so that that case works. Change in stats: There were no change in stats except in MIPS and SPARC in the O3 model. MIPS runs in about 9% fewer ticks. SPARC runs with 30%-50% fewer ticks, which could likely be improved further by setting call/return instruction flags and taking advantage of the RAS. TODO: Add != operators to the PCState classes, defined trivially to be !(a==b). Smooth out places where PCs are split apart, passed around, and put back together later. I think this might happen in SPARC's fault code. Add ISA specific constructors that allow setting PC elements without calling a bunch of accessors. Try to eliminate the need for the branching() function. Factor out Alpha's PAL mode pc bit into a separate flag field, and eliminate places where it's blindly masked out or tested in the PC. diff 7678:f19b6a3a8cec Mon Sep 13 22:26:00 EDT 2010 Gabe Black <gblack@eecs.umich.edu> Faults: Pass the StaticInst involved, if any, to a Fault's invoke method. Also move the "Fault" reference counted pointer type into a separate file, sim/fault.hh. It would be better to name this less similarly to sim/faults.hh to reduce confusion, but fault.hh matches the name of the type. We could change Fault to FaultPtr to match other pointer types, and then changing the name of the file would make more sense. diff 7655:8bce423f2075 Wed Aug 25 20:10:00 EDT 2010 Ali Saidi <ali.saidi@arm.com> CPU: Print out traces for faluting inst when the flag ExecFaulting is set diff 7520:67c670459d01 Fri Aug 13 09:16:00 EDT 2010 Gabe Black <gblack@eecs.umich.edu> CPU: Add readBytes and writeBytes functions to the exec contexts. diff 7518:917208416d2a Fri Aug 13 09:10:00 EDT 2010 Gabe Black <gblack@eecs.umich.edu> CPU: Tidy up endianness handling for mmapped "IPR"s. diff 7046:d21d575a6f99 Tue Mar 23 11:50:00 EDT 2010 Steve Reinhardt <steve.reinhardt@amd.com> cpu: get rid of uncached access "events" These recordEvent() calls could cause crashes since they access the req pointer after it's potentially been deleted during a failed translation call. (Similar problem to the traceData bug fixed in the previous cset.) Moving them above the translation call (as was done recentlyi in cset 8b2b8e5e7d35) avoids the crash but doesn't work, since at that point we don't know if the access is uncached or not. It's not clear why these calls are there, and no one seems to use them, so we'll just delete them. If they are needed, they should be moved to somewhere that's guaranteed to be after the translation completes but before the request is possibly deleted, e.g., in finishTranslation(). diff 7045:e21fe6a62b1c Tue Mar 23 11:50:00 EDT 2010 Steve Reinhardt <steve.reinhardt@amd.com> cpu: fix exec tracing memory corruption bug Accessing traceData (to call setAddress() and/or setData()) after initiating a timing translation was causing crashes, since a failed translation could delete the traceData object before returning. It turns out that there was never a need to access traceData after initiating the translation, as the traced data was always available earlier; this ordering was merely historical. Furthermore, traceData->setAddress() and traceData->setData() were being called both from the CPU model and the ISA definition, often redundantly. This patch standardizes all setAddress and setData calls for memory instructions to be in the CPU models and not in the ISA definition. It also moves those calls above the translation calls to eliminate the crashes. |
/gem5/src/ | ||
H A D | SConscript | diff 7800:318eca813e2b Tue Dec 21 11:02:00 EST 2010 Nathan Binkert <nate@binkert.org> scons: remove extra dependencies diff 7756:846fb3ffe0dc Mon Nov 15 15:04:00 EST 2010 Ali Saidi <Ali.Saidi@ARM.com> SCons: Cleanup SCons output during compile diff 7722:c10bc8ad3f97 Sat Nov 06 20:48:00 EDT 2010 Gabe Black <gblack@eecs.umich.edu> scons: Replace the build_dir parameter to SConscript with variant_dir. The build_dir parameter name has been deprecated and replaced with variant_dir. This change switches us over to avoid warning spew in newer versions of scons. diff 7677:c6e283904437 Sun Sep 12 18:41:00 EDT 2010 Nathan Binkert <nate@binkert.org> swig: make all generated files go into the m5.internal package This is necessary because versions of swig older than 1.3.39 fail to do the right thing and try to do relative imports for everything (even with the package= option to %module). Instead of putting params in the m5.internal.params package, put params in the m5.internal package and make all param modules start with param_. Same thing for m5.internal.enums. Also, stop importing all generated params into m5.objects. They are not necessary and now with everything using relative imports we wound up with pollution of the namespace (where builtin-range got overridden). diff 7675:2221ec64132f Thu Sep 09 17:26:00 EDT 2010 Nathan Binkert <nate@binkert.org> scons: Stop building the big monolithic swigged params module kill params.i and create a separate .i for each object (param, enums, etc.) diff 7674:8e3734851770 Thu Sep 09 17:15:00 EDT 2010 Nathan Binkert <nate@binkert.org> init: don't build files that centralize python and swig code Instead of putting all object files into m5/object/__init__.py, interrogate the importer to find out what should be imported. Instead of creating a single file that lists all of the embedded python modules, use static object construction to put those objects onto a list. Do something similar for embedded swig (C++) code. diff 7673:b28bd1fa9a35 Thu Sep 09 17:15:00 EDT 2010 Nathan Binkert <nate@binkert.org> scons: use code_formatter wherever we can in the build system diff 7502:3ef7ff12c788 Wed Jul 21 18:53:00 EDT 2010 Nathan Binkert <nate@binkert.org> python: Add mechanism to override code compiled into the exectuable If the user sets the environment variable M5_OVERRIDE_PY_SOURCE to True, then imports that would normally find python code compiled into the executable will instead first check in the absolute location where the code was found during the build of the executable. This only works for files in the src (or extras) directories, not automatically generated files. This is a developer feature! diff 7065:cf0422cb8555 Thu Apr 15 19:25:00 EDT 2010 Nathan Binkert <nate@binkert.org> scons: don't maintain files in sorted order This causes builds to happen in sorted order rather than in declaration order. This gets annoying when you make a global change and then you notice that the files that are being compiled are jumping around the directory hierarchy. |
Completed in 478 milliseconds