mirror of
https://github.com/llvm/llvm-project.git
synced 2025-05-03 18:26:05 +00:00
d3f3d932ee
- Changes bit-field access policy to try to use (aligned) register sized accesses. The idea here is that by using larger accesses we expose more coalescing potential to the backend when we have situations like adjacent bit-fields in the same structure (which is common), and that the backend should be smart enough to narrow the accesses down when no coalescing is done or when it is shown not to be profitable. -- $ clang -m32 -O3 -S -o - t.c _f0: ## @f0 pushl %ebp movl %esp, %ebp movl 8(%ebp), %eax movb (%eax), %cl andb $-128, %cl orb $1, %cl movb %cl, (%eax) movb 1(%eax), %cl andb $-128, %cl orb $1, %cl movb %cl, 1(%eax) movb 2(%eax), %cl andb $-128, %cl orb $1, %cl movb %cl, 2(%eax) movb 3(%eax), %cl andb $-128, %cl orb $1, %cl movb %cl, 3(%eax) popl %ebp ret $ clang -m32 -O3 -S -o - t.c -Xclang -fuse-register-sized-bitfield-access _f0: ## @f0 pushl %ebp movl %esp, %ebp movl 8(%ebp), %eax movl $-2139062144, %ecx ## imm = 0xFFFFFFFF80808080 andl (%eax), %ecx orl $16843009, %ecx ## imm = 0x1010101 movl %ecx, (%eax) popl %ebp ret -- llvm-svn: 133532
IRgen optimization opportunities. //===---------------------------------------------------------------------===// The common pattern of -- short x; // or char, etc (x == 10) -- generates an zext/sext of x which can easily be avoided. //===---------------------------------------------------------------------===// Bitfields accesses can be shifted to simplify masking and sign extension. For example, if the bitfield width is 8 and it is appropriately aligned then is is a lot shorter to just load the char directly. //===---------------------------------------------------------------------===// It may be worth avoiding creation of alloca's for formal arguments for the common situation where the argument is never written to or has its address taken. The idea would be to begin generating code by using the argument directly and if its address is taken or it is stored to then generate the alloca and patch up the existing code. In theory, the same optimization could be a win for block local variables as long as the declaration dominates all statements in the block. NOTE: The main case we care about this for is for -O0 -g compile time performance, and in that scenario we will need to emit the alloca anyway currently to emit proper debug info. So this is blocked by being able to emit debug information which refers to an LLVM temporary, not an alloca. //===---------------------------------------------------------------------===// We should try and avoid generating basic blocks which only contain jumps. At -O0, this penalizes us all the way from IRgen (malloc & instruction overhead), all the way down through code generation and assembly time. On 176.gcc:expr.ll, it looks like over 12% of basic blocks are just direct branches! //===---------------------------------------------------------------------===//