[Bug 3845] parallel port access is slow due to high cpu usage (eprom burner )
http://bugs.winehq.org/show_bug.cgi?id=3845 Ondrej Zary <linux(a)rainbow-software.org> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |linux(a)rainbow-software.org --- Comment #7 from Ondrej Zary <linux(a)rainbow-software.org> 2008-05-04 05:58:42 --- I have the same problem with Willem EPROM programmer. No wonder it's so slow when you look at Wine parport code - for each port access, massive amount code is run: 1. DOSVM_outport() http://source.winehq.org/source/dlls/winedos/ioports.c#L501 this fortunately calls IO_pp_outp() early 2. IO_pp_outp() http://source.winehq.org/source/dlls/winedos/ppdev.c#L295 this loops over all parport devices, changes data direction (in some cases) and calls IO_pp_do_access() 3. IO_pp_do_access() http://source.winehq.org/source/dlls/winedos/ppdev.c#L234 this claims the port, does the in/out and releases it again I guess that claim/release in (3) can be eliminated (just claim it on first access and never release as I doubt that anyone wants other applications to mess with parallel port during EPROM programming). The for loop in 2 could be probably improved somehow. Changing the data direction might not be needed always, depending on previous access (but I don't know parallel port programming details in Linux). -- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.
participants (1)
-
wine-bugs@winehq.org