----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: http://reviews.gem5.org/r/3002/#review6937 -----------------------------------------------------------
Recently we discarded patch 2786 because it optionally allowed SLICC programs to specify whether an object is a pointer or not. Our patch was motivated because PerfectCacheMemory differed from CacheMemory and was a pretty special case. We discarded that patch because it was correctly argued that SLICC has a long-standing standard not to expose pointers to the SLICC programmer. This patch would require all SLICC programs to consider pointers. That seems far more dangerous than our previous patch. If we are going to keep to a standard, then we must keep to it and this patch should be discarded. src/mem/protocol/MOESI_hammer-dir.sm (line 197) <http://reviews.gem5.org/r/3002/#comment5963> Is moving the global fwd_set variable to a local variable orthogonal to the rest of this patch? - Brad Beckmann On Aug. 3, 2015, 12:15 a.m., Nilay Vaish wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > http://reviews.gem5.org/r/3002/ > ----------------------------------------------------------- > > (Updated Aug. 3, 2015, 12:15 a.m.) > > > Review request for Default. > > > Repository: gem5 > > > Description > ------- > > Changeset 11017:b37918d5063f > --------------------------- > ruby: slicc: fix naming of variables > > The way SLICC compiler has been written, it was difficult to use a variable > that appears in the base class AbstractController without adding some special > code to the file src/mem/slicc/ast/ObjDeclAST.py. This eliminates the need > for > any such special code. The problem ultimately turned out to be the way SLICC > handles pointer and non-pointer variables. Till now, SLICC tried to handle > everything using pointers, unless some special rules were specified, as was > being done for three variables: version, machineID and clusterID. Now the Var > class will have a field that indicates whether a variable is pointer or not. > This field is used to differentiate between pointer variables and other > variables. Pointer-based variables are allocated in the init() function of > the > controller, as was the case before. Other variables now need not be pointer > based. Variables that need the constructor be called on them will now > necessarily be declared as pointers. The protocols currently in gem5 have > been > changed accordingly. > > > Diffs > ----- > > src/mem/slicc/ast/InPortDeclAST.py a618349a7953 > src/mem/slicc/ast/LocalVariableAST.py a618349a7953 > src/mem/protocol/MOESI_CMP_directory-dma.sm a618349a7953 > src/mem/protocol/MOESI_CMP_token-L1cache.sm a618349a7953 > src/mem/protocol/MOESI_CMP_token-L2cache.sm a618349a7953 > src/mem/protocol/MOESI_CMP_token-dir.sm a618349a7953 > src/mem/protocol/MOESI_hammer-cache.sm a618349a7953 > src/mem/protocol/MOESI_hammer-dir.sm a618349a7953 > src/mem/slicc/ast/ActionDeclAST.py a618349a7953 > src/mem/slicc/ast/EnqueueStatementAST.py a618349a7953 > src/mem/slicc/ast/FormalParamAST.py a618349a7953 > src/mem/protocol/MESI_Three_Level-L0cache.sm a618349a7953 > src/mem/protocol/MESI_Three_Level-L1cache.sm a618349a7953 > src/mem/protocol/MESI_Two_Level-L1cache.sm a618349a7953 > src/mem/protocol/MESI_Two_Level-L2cache.sm a618349a7953 > src/mem/protocol/MESI_Two_Level-dir.sm a618349a7953 > src/mem/protocol/MI_example-cache.sm a618349a7953 > src/mem/protocol/MI_example-dir.sm a618349a7953 > src/mem/protocol/MOESI_CMP_directory-L1cache.sm a618349a7953 > src/mem/protocol/MOESI_CMP_directory-L2cache.sm a618349a7953 > src/mem/protocol/MOESI_CMP_directory-dir.sm a618349a7953 > src/mem/slicc/ast/ObjDeclAST.py a618349a7953 > src/mem/slicc/ast/OutPortDeclAST.py a618349a7953 > src/mem/slicc/ast/PeekStatementAST.py a618349a7953 > src/mem/slicc/symbols/StateMachine.py a618349a7953 > src/mem/slicc/symbols/Type.py a618349a7953 > src/mem/slicc/symbols/Var.py a618349a7953 > > Diff: http://reviews.gem5.org/r/3002/diff/ > > > Testing > ------- > > > Thanks, > > Nilay Vaish > > _______________________________________________ gem5-dev mailing list [email protected] http://m5sim.org/mailman/listinfo/gem5-dev
