contributed by Vladimir Merzliakov!

llvm-svn: 20084

llvm-svn: 20067

llvm-svn: 20066

llvm-svn: 20065

llvm-svn: 20064

llvm-svn: 20063

llvm-svn: 20062

llvm-svn: 20061

llvm-svn: 20060

llvm-svn: 20059

llvm-svn: 20058

llvm-svn: 20057

llvm-svn: 20056

llvm-svn: 20055

llvm-svn: 20053

llvm-svn: 20051

llvm-svn: 20050

llvm-svn: 20049

llvm-svn: 20048

field to hold it instead.  This shrinks memory usage for 176.gcc from
57628728 to 57598144 bytes, a small reduction of about 30K.

llvm-svn: 20047

LoadInst from 60 -> 56 bytes and StoreInst from 76 -> 72 bytes.

Note however, that this doesn't actually save any memory on common systems
where 'malloc' returns 8-byte aligned memory, as the saved space is replaced
by useless alignment padding.  :(

llvm-svn: 20046

llvm-svn: 20045

llvm-svn: 20044

play with the unused part.

llvm-svn: 20043

llvm-svn: 20042

llvm-svn: 20039

into a temporary graph, remember it for later, then inline the tmp graph into
the call site.

In the case where there are other call sites to the same set of functions, this
permits us to just inline the temporary graph instead of all of the callees.

This turns N*M inlining situations into an N+M inlining situation.

llvm-svn: 20036

llvm-svn: 20035

llvm-svn: 20034

llvm-svn: 20033

llvm-svn: 20031

targets.

llvm-svn: 20030

llvm-svn: 20028

llvm-svn: 20027

llvm-svn: 20026

* Remove duplicated CVS directions and refer reader to the CVS section
* Clean up directions to be brief

llvm-svn: 20023

llvm-svn: 20019

llvm-svn: 20018

llvm-svn: 20017

* llvmgcc and llvmg++ aliases are no longer needed (binaries have llvm- prefix)

llvm-svn: 20016