Re: [webkit-dev] want to port JIT to MIPS - how patchOffset* constant determined?
Hi, this really helps! The two pointers may also help Mips. In Mips dest offset patches are in two instructions, no constant biase can be used as in X86. Currently I do check instruction syntax to determine the patch offset. thanks again! joe --- On Wed, 3/4/09, Zoltan Herczeg zherc...@inf.u-szeged.hu wrote: From: Zoltan Herczeg zherc...@inf.u-szeged.hu Subject: Re: [webkit-dev] want to port JIT to MIPS - how patchOffset* constant determined? To: webkit-dev@lists.webkit.org Date: Wednesday, March 4, 2009, 3:33 PM Hi, they generate instructions, which size is known in advance. Think about the following sequence: hotPathBegin: mov regX, 32bit_const - 6 bytes (*) (**) add regX, regY - 2 bytes jo 32bit_addr - 5 bytes (*) * (Note) : these instructions will be modified during runtime. ** (Note) : there is a short form for mov regX, 8bit_const, which length is only 3 bytes, but they force the longer version in such cases to keep the size of the instruction. As you can see, the address of jo is always (hotPathBegin + 6 + 2). They simply introduce a new constant: patchOffsetXXX = 8, and use this constant to access the jo instruction later. In ARM we can't rely on such constant, because the constant pool can be placed after any instruction. hotPathBegin: ldr rX, [pc + const_pool_addr] ; 32 bit const [...] - the const pool can be placed here add rX, rX, rY [...] - the const pool can be placed here hotPath2: ldr pc, [pc + const_pool_addr] ; 32 bit target address We need to store both pointers (hotPathBegin and hotPath2). Zoltan Zoltan, Thanks for reply, I'm trying to understand your example. But,X86 instruction size is from 1 to 17bytes, not constant. I may misunderstand your comments? Many X86 instruction can have imm32 at the end, thus this pointer can be used for patch as well as next address after call. Does Arm have similar things? or else you still need to figure out why patchOffsetOpCallCompareToJump = 9;? may be some instruction lengths relates to the 9? rgds joe --- On Wed, 3/4/09, Zoltan Herczeg zherc...@inf.u-szeged.hu wrote: From: Zoltan Herczeg zherc...@inf.u-szeged.hu Subject: Re: [webkit-dev] want to port JIT to MIPS - how patchOffset* constant determined? To: webkit-dev@lists.webkit.org Date: Wednesday, March 4, 2009, 3:45 AM On x86, the size of the instructions are fixed. If you want to access multiple instructions in the instruction stream, you only need to store the address of the first one, and can access the others by their relative address. This saves a little memory. Example (see JIT::linkCall): instruction at callLinkInfo-hotPathBegin: points to callee comparison instruction at callLinkInfo-hotPathBegin + patchOffsetOpCallCompareToJump: points to the slow case entry jump Zoltan in jit.h, for example: static const int patchOffsetOpCallCompareToJump = 9; static const int patchOffsetPutByIdStructure = 7; static const int patchOffsetPutByIdPropertyMapOffset = 22; static const int patchOffsetGetByIdBranchToSlowCase = 13; thanks for help, I'm stucked here now. joe --- On Sat, 2/28/09, Gavin Barraclough barraclo...@apple.com wrote: From: Gavin Barraclough barraclo...@apple.com Subject: Re: [webkit-dev] want to port JIT to MIPS - JIT reg usage clean up? To: WebKit Development webkit-dev@lists.webkit.org Date: Saturday, February 28, 2009, 12:19 PM On Feb 27, 2009, at 4:55 PM, x yz wrote: The regTx seems to be working registers here, yet their definition are regparm(3) registers for function arugments. Such usage would cause conflict on other platforms. May I suggest that we use individual defined set of regs for func i/o argument and working? First up, I think you're getting slightly confused about regparm(3). This is not used anywhere in the JS language JIT, only in WREC. In some configurations of the JIT we use fastcall semantics on x86... but none of this is really relevant to MIPS. Just ignore all this. Stick to the default MIPS ABI for stub functions. Reading between the lines, I'm guessing your concern here is that in setting up arguments for a JIT stub call you may trample the JIT's temporary registers? If so, I think you need to look at the argument passing more closely. The mechanisms to pass arguments to stub functions pass all arguments in memory – typically passing a single pointer to the stub functions, which can be used to retrieve the arguments. This pointer argument can be set up immediately prior to the call, so it does not interfere with the regT? temporaries. We follow this pattern on x86-64, where the ABI is typically to pass arguments in registers. You cannot
Re: [webkit-dev] want to port JIT to MIPS - how patchOffset* constant determined?
in jit.h, for example: static const int patchOffsetOpCallCompareToJump = 9; static const int patchOffsetPutByIdStructure = 7; static const int patchOffsetPutByIdPropertyMapOffset = 22; static const int patchOffsetGetByIdBranchToSlowCase = 13; thanks for help, I'm stucked here now. joe --- On Sat, 2/28/09, Gavin Barraclough barraclo...@apple.com wrote: From: Gavin Barraclough barraclo...@apple.com Subject: Re: [webkit-dev] want to port JIT to MIPS - JIT reg usage clean up? To: WebKit Development webkit-dev@lists.webkit.org Date: Saturday, February 28, 2009, 12:19 PM On Feb 27, 2009, at 4:55 PM, x yz wrote: The regTx seems to be working registers here, yet their definition are regparm(3) registers for function arugments. Such usage would cause conflict on other platforms. May I suggest that we use individual defined set of regs for func i/o argument and working? First up, I think you're getting slightly confused about regparm(3). This is not used anywhere in the JS language JIT, only in WREC. In some configurations of the JIT we use fastcall semantics on x86... but none of this is really relevant to MIPS. Just ignore all this. Stick to the default MIPS ABI for stub functions. Reading between the lines, I'm guessing your concern here is that in setting up arguments for a JIT stub call you may trample the JIT's temporary registers? If so, I think you need to look at the argument passing more closely. The mechanisms to pass arguments to stub functions pass all arguments in memory – typically passing a single pointer to the stub functions, which can be used to retrieve the arguments. This pointer argument can be set up immediately prior to the call, so it does not interfere with the regT? temporaries. We follow this pattern on x86-64, where the ABI is typically to pass arguments in registers. You cannot trivially change the way this works, since the argument pointer is used for other purposes too (e.g. retrieving the arguments passed into the JIT code from within the stubs). We strongly prefer small, simple, incremental changes. A patch that tried to both port the JIT to a new platform and to introduce a new argument passing interface to the JIT stub functions sounds unlikely to get anywhere (a patch porting the JIT to a new platform is on its own very likely to be too much more than we'd want to land in one chunk). I'd suggest that a port would be wise to engineer it's initial solution to fit one of the existing argument passing mechanisms (these are selected by JIT_STUB_ARGUMENT_* switches, to help find the relevant code). (Alternatively, you're welcome to attempt to port x86-64 to make use of an in-register argument passing solution, which could be hugely useful. With this landed first and separately, a port could then build on top of this.) As a more direct answer to your question, you could endeavour to make the set of hardware registers used as JIT temporaries non-overlapping with ABI function argument registers on MIPS, but this is unlikely to be a general solution to anything for all platforms, due to limited register availability on some architectures. we would put all these definition in a file named regMap.h, then we can remove all X86:: from other JIT files. I don't think we'll be keen on taking preemptive changes so far ahead in preparation of a port. The first logical step in porting to a new platform is still to start with WREC, and this requires no changes in the JIT directory. Any refactoring of the existing JIT would make more sense more directly prior to work in that area. cheers, G. I'd apperciate if sb can do it or help me to do it. rgds joe --- On Sat, 2/28/09, x yz last...@yahoo.com wrote: From: x yz last...@yahoo.com Subject: Re: [webkit-dev] want to port JIT to MIPS - which calling convention is used here? To: webkit-dev@lists.webkit.org, Zoltan Herczeg zherc...@inf.u-szeged.hu Date: Saturday, February 28, 2009, 7:40 AM Hi, Thanks for your help in advance:) in JITPropertyAccess.cpp: if (transitionWillNeedStorageRealloc(oldStructure, newStructure)) { pop(X86::ebx); ///pop return address, why? for exception? #if PLATFORM(X86_64)///which convention is this? move(Imm32(newStructure-propertyStorageCapacity()), regT1); //edx move(Imm32(oldStructure-propertyStorageCapacity()), X86::esi); move(regT0, X86::edi); callTarget = call(); #else ///__cdecl, yet how can I know resizePropertyStorage() use __cdecl? push(Imm32(newStructure-propertyStorageCapacity())); push(Imm32(oldStructure-propertyStorageCapacity())); push(regT0); callTarget = call(); addPtr(Imm32(3 * sizeof(void*)), X86::esp); ///clean stack #endif ___
Re: [webkit-dev] want to port JIT to MIPS - how patchOffset* constant determined?
On x86, the size of the instructions are fixed. If you want to access multiple instructions in the instruction stream, you only need to store the address of the first one, and can access the others by their relative address. This saves a little memory. Example (see JIT::linkCall): instruction at callLinkInfo-hotPathBegin: points to callee comparison instruction at callLinkInfo-hotPathBegin + patchOffsetOpCallCompareToJump: points to the slow case entry jump Zoltan in jit.h, for example: static const int patchOffsetOpCallCompareToJump = 9; static const int patchOffsetPutByIdStructure = 7; static const int patchOffsetPutByIdPropertyMapOffset = 22; static const int patchOffsetGetByIdBranchToSlowCase = 13; thanks for help, I'm stucked here now. joe --- On Sat, 2/28/09, Gavin Barraclough barraclo...@apple.com wrote: From: Gavin Barraclough barraclo...@apple.com Subject: Re: [webkit-dev] want to port JIT to MIPS - JIT reg usage clean up? To: WebKit Development webkit-dev@lists.webkit.org Date: Saturday, February 28, 2009, 12:19 PM On Feb 27, 2009, at 4:55 PM, x yz wrote: The regTx seems to be working registers here, yet their definition are regparm(3) registers for function arugments. Such usage would cause conflict on other platforms. May I suggest that we use individual defined set of regs for func i/o argument and working? First up, I think you're getting slightly confused about regparm(3). This is not used anywhere in the JS language JIT, only in WREC. In some configurations of the JIT we use fastcall semantics on x86... but none of this is really relevant to MIPS. Just ignore all this. Stick to the default MIPS ABI for stub functions. Reading between the lines, I'm guessing your concern here is that in setting up arguments for a JIT stub call you may trample the JIT's temporary registers? If so, I think you need to look at the argument passing more closely. The mechanisms to pass arguments to stub functions pass all arguments in memory – typically passing a single pointer to the stub functions, which can be used to retrieve the arguments. This pointer argument can be set up immediately prior to the call, so it does not interfere with the regT? temporaries. We follow this pattern on x86-64, where the ABI is typically to pass arguments in registers. You cannot trivially change the way this works, since the argument pointer is used for other purposes too (e.g. retrieving the arguments passed into the JIT code from within the stubs). We strongly prefer small, simple, incremental changes. A patch that tried to both port the JIT to a new platform and to introduce a new argument passing interface to the JIT stub functions sounds unlikely to get anywhere (a patch porting the JIT to a new platform is on its own very likely to be too much more than we'd want to land in one chunk). I'd suggest that a port would be wise to engineer it's initial solution to fit one of the existing argument passing mechanisms (these are selected by JIT_STUB_ARGUMENT_* switches, to help find the relevant code). (Alternatively, you're welcome to attempt to port x86-64 to make use of an in-register argument passing solution, which could be hugely useful. With this landed first and separately, a port could then build on top of this.) As a more direct answer to your question, you could endeavour to make the set of hardware registers used as JIT temporaries non-overlapping with ABI function argument registers on MIPS, but this is unlikely to be a general solution to anything for all platforms, due to limited register availability on some architectures. we would put all these definition in a file named regMap.h, then we can remove all X86:: from other JIT files. I don't think we'll be keen on taking preemptive changes so far ahead in preparation of a port. The first logical step in porting to a new platform is still to start with WREC, and this requires no changes in the JIT directory. Any refactoring of the existing JIT would make more sense more directly prior to work in that area. cheers, G. I'd apperciate if sb can do it or help me to do it. rgds joe --- On Sat, 2/28/09, x yz last...@yahoo.com wrote: From: x yz last...@yahoo.com Subject: Re: [webkit-dev] want to port JIT to MIPS - which calling convention is used here? To: webkit-dev@lists.webkit.org, Zoltan Herczeg zherc...@inf.u-szeged.hu Date: Saturday, February 28, 2009, 7:40 AM Hi, Thanks for your help in advance:) in JITPropertyAccess.cpp: if (transitionWillNeedStorageRealloc(oldStructure, newStructure)) { pop(X86::ebx); ///pop return address, why? for exception? #if PLATFORM(X86_64)///which convention is this? move(Imm32(newStructure-propertyStorageCapacity()), regT1); //edx move(Imm32(oldStructure-propertyStorageCapacity()), X86::esi);
Re: [webkit-dev] want to port JIT to MIPS - how patchOffset* constant determined?
Zoltan, Thanks for reply, I'm trying to understand your example. But,X86 instruction size is from 1 to 17bytes, not constant. I may misunderstand your comments? Many X86 instruction can have imm32 at the end, thus this pointer can be used for patch as well as next address after call. Does Arm have similar things? or else you still need to figure out why patchOffsetOpCallCompareToJump = 9;? may be some instruction lengths relates to the 9? rgds joe --- On Wed, 3/4/09, Zoltan Herczeg zherc...@inf.u-szeged.hu wrote: From: Zoltan Herczeg zherc...@inf.u-szeged.hu Subject: Re: [webkit-dev] want to port JIT to MIPS - how patchOffset* constant determined? To: webkit-dev@lists.webkit.org Date: Wednesday, March 4, 2009, 3:45 AM On x86, the size of the instructions are fixed. If you want to access multiple instructions in the instruction stream, you only need to store the address of the first one, and can access the others by their relative address. This saves a little memory. Example (see JIT::linkCall): instruction at callLinkInfo-hotPathBegin: points to callee comparison instruction at callLinkInfo-hotPathBegin + patchOffsetOpCallCompareToJump: points to the slow case entry jump Zoltan in jit.h, for example: static const int patchOffsetOpCallCompareToJump = 9; static const int patchOffsetPutByIdStructure = 7; static const int patchOffsetPutByIdPropertyMapOffset = 22; static const int patchOffsetGetByIdBranchToSlowCase = 13; thanks for help, I'm stucked here now. joe --- On Sat, 2/28/09, Gavin Barraclough barraclo...@apple.com wrote: From: Gavin Barraclough barraclo...@apple.com Subject: Re: [webkit-dev] want to port JIT to MIPS - JIT reg usage clean up? To: WebKit Development webkit-dev@lists.webkit.org Date: Saturday, February 28, 2009, 12:19 PM On Feb 27, 2009, at 4:55 PM, x yz wrote: The regTx seems to be working registers here, yet their definition are regparm(3) registers for function arugments. Such usage would cause conflict on other platforms. May I suggest that we use individual defined set of regs for func i/o argument and working? First up, I think you're getting slightly confused about regparm(3). This is not used anywhere in the JS language JIT, only in WREC. In some configurations of the JIT we use fastcall semantics on x86... but none of this is really relevant to MIPS. Just ignore all this. Stick to the default MIPS ABI for stub functions. Reading between the lines, I'm guessing your concern here is that in setting up arguments for a JIT stub call you may trample the JIT's temporary registers? If so, I think you need to look at the argument passing more closely. The mechanisms to pass arguments to stub functions pass all arguments in memory – typically passing a single pointer to the stub functions, which can be used to retrieve the arguments. This pointer argument can be set up immediately prior to the call, so it does not interfere with the regT? temporaries. We follow this pattern on x86-64, where the ABI is typically to pass arguments in registers. You cannot trivially change the way this works, since the argument pointer is used for other purposes too (e.g. retrieving the arguments passed into the JIT code from within the stubs). We strongly prefer small, simple, incremental changes. A patch that tried to both port the JIT to a new platform and to introduce a new argument passing interface to the JIT stub functions sounds unlikely to get anywhere (a patch porting the JIT to a new platform is on its own very likely to be too much more than we'd want to land in one chunk). I'd suggest that a port would be wise to engineer it's initial solution to fit one of the existing argument passing mechanisms (these are selected by JIT_STUB_ARGUMENT_* switches, to help find the relevant code). (Alternatively, you're welcome to attempt to port x86-64 to make use of an in-register argument passing solution, which could be hugely useful. With this landed first and separately, a port could then build on top of this.) As a more direct answer to your question, you could endeavour to make the set of hardware registers used as JIT temporaries non-overlapping with ABI function argument registers on MIPS, but this is unlikely to be a general solution to anything for all platforms, due to limited register availability on some architectures. we would put all these definition in a file named regMap.h, then we can remove all X86:: from other JIT files. I don't think we'll be keen on taking preemptive changes so far ahead in preparation of a port. The first logical step in porting to a new platform is still to start with WREC, and this requires no changes in the JIT directory. Any refactoring of the existing JIT
Re: [webkit-dev] want to port JIT to MIPS - how patchOffset* constant determined?
Hi, they generate instructions, which size is known in advance. Think about the following sequence: hotPathBegin: mov regX, 32bit_const - 6 bytes (*) (**) add regX, regY - 2 bytes jo 32bit_addr - 5 bytes (*) * (Note) : these instructions will be modified during runtime. ** (Note) : there is a short form for mov regX, 8bit_const, which length is only 3 bytes, but they force the longer version in such cases to keep the size of the instruction. As you can see, the address of jo is always (hotPathBegin + 6 + 2). They simply introduce a new constant: patchOffsetXXX = 8, and use this constant to access the jo instruction later. In ARM we can't rely on such constant, because the constant pool can be placed after any instruction. hotPathBegin: ldr rX, [pc + const_pool_addr] ; 32 bit const [...] - the const pool can be placed here add rX, rX, rY [...] - the const pool can be placed here hotPath2: ldr pc, [pc + const_pool_addr] ; 32 bit target address We need to store both pointers (hotPathBegin and hotPath2). Zoltan Zoltan, Thanks for reply, I'm trying to understand your example. But,X86 instruction size is from 1 to 17bytes, not constant. I may misunderstand your comments? Many X86 instruction can have imm32 at the end, thus this pointer can be used for patch as well as next address after call. Does Arm have similar things? or else you still need to figure out why patchOffsetOpCallCompareToJump = 9;? may be some instruction lengths relates to the 9? rgds joe --- On Wed, 3/4/09, Zoltan Herczeg zherc...@inf.u-szeged.hu wrote: From: Zoltan Herczeg zherc...@inf.u-szeged.hu Subject: Re: [webkit-dev] want to port JIT to MIPS - how patchOffset* constant determined? To: webkit-dev@lists.webkit.org Date: Wednesday, March 4, 2009, 3:45 AM On x86, the size of the instructions are fixed. If you want to access multiple instructions in the instruction stream, you only need to store the address of the first one, and can access the others by their relative address. This saves a little memory. Example (see JIT::linkCall): instruction at callLinkInfo-hotPathBegin: points to callee comparison instruction at callLinkInfo-hotPathBegin + patchOffsetOpCallCompareToJump: points to the slow case entry jump Zoltan in jit.h, for example: static const int patchOffsetOpCallCompareToJump = 9; static const int patchOffsetPutByIdStructure = 7; static const int patchOffsetPutByIdPropertyMapOffset = 22; static const int patchOffsetGetByIdBranchToSlowCase = 13; thanks for help, I'm stucked here now. joe --- On Sat, 2/28/09, Gavin Barraclough barraclo...@apple.com wrote: From: Gavin Barraclough barraclo...@apple.com Subject: Re: [webkit-dev] want to port JIT to MIPS - JIT reg usage clean up? To: WebKit Development webkit-dev@lists.webkit.org Date: Saturday, February 28, 2009, 12:19 PM On Feb 27, 2009, at 4:55 PM, x yz wrote: The regTx seems to be working registers here, yet their definition are regparm(3) registers for function arugments. Such usage would cause conflict on other platforms. May I suggest that we use individual defined set of regs for func i/o argument and working? First up, I think you're getting slightly confused about regparm(3). This is not used anywhere in the JS language JIT, only in WREC. In some configurations of the JIT we use fastcall semantics on x86... but none of this is really relevant to MIPS. Just ignore all this. Stick to the default MIPS ABI for stub functions. Reading between the lines, I'm guessing your concern here is that in setting up arguments for a JIT stub call you may trample the JIT's temporary registers? If so, I think you need to look at the argument passing more closely. The mechanisms to pass arguments to stub functions pass all arguments in memory – typically passing a single pointer to the stub functions, which can be used to retrieve the arguments. This pointer argument can be set up immediately prior to the call, so it does not interfere with the regT? temporaries. We follow this pattern on x86-64, where the ABI is typically to pass arguments in registers. You cannot trivially change the way this works, since the argument pointer is used for other purposes too (e.g. retrieving the arguments passed into the JIT code from within the stubs). We strongly prefer small, simple, incremental changes. A patch that tried to both port the JIT to a new platform and to introduce a new argument passing interface to the JIT stub functions sounds unlikely to get anywhere (a patch porting the JIT to a new platform is on its own very likely to be too much more than we'd want to land in one chunk). I'd suggest that a port would be wise to engineer it's initial solution to fit one of the existing argument passing mechanisms (these are selected