Java HotSpot(TM) Server VM warning: increase O_BUFLEN in ostream.hpp -- output truncated
Christian.Thalinger at Sun.COM
Fri Jan 15 00:29:12 PST 2010
On Thu, 2010-01-14 at 18:19 +0100, Ulf Zibis wrote:
> Am 14.01.2010 18:04, Christian Thalinger schrieb:
> > On Thu, 2010-01-14 at 17:41 +0100, Ulf Zibis wrote:
> >> Thanks for clarification. But I'm wondering why I can't see any
> >> difference, using it. The output is still without raw bytes, see:
> >> [snip]
> >> 000
> >> 00c B1: # B72 B2 <- BLOCK HEAD IS JUNK Freq: 1
> >> 00c # stack bang
> >> PUSHL EBP
> >> SUB ESP,104 # Create frame
> >> 01a MOV [ESP + #28],ECX
> >> 01e MOV EDI,ECX
> >> 020 MOV ESI,[ECX + #40] # int ! Field
> >> sun/nio/cs/FastCharset$Decoder.dstOff
> >> 023 MOV ECX,[ECX + #48] ! Field
> >> sun/nio/cs/FastCharset$Decoder.dstBuf
> >> [snap]
> > Because this is the output of PrintOptoAssembly. PrintAssemblyOptions
> > only applies (as the name suggests) to PrintAssembly.
> Thanks, you are right, it works after I had "enabled" PrintAssembly.
> Sometimes I rename hsdis-i386.dll to get the better annotated output
> from PrintOptoAssembly.
> Is there any chance to have the better annotations also for PrintAssembly?
> Anyway I'm noting, that the two instructions, PUSHL, SUB are lacking
> from there pc counter value a line start. Bug?
No. The method prolog is emitted as one instruction and that's why it
only has one offset.
More information about the hotspot-compiler-dev