Date: Thu, 8 May 1997 10:42:23 +1000 From: Bruce Evans <bde@zeta.org.au> To: bde@zeta.org.au, skynyrd@opus.cts.cwu.edu Cc: current@freebsd.org Subject: Re: Problems using remote gdb Message-ID: <199705080042.KAA27367@godzilla.zeta.org.au>
next in thread | raw e-mail | index | archive | help
>Actually the comment for flags=0x40 seems like it might have a typo. >Should it read "Do not reserve ...." or is there a complete other sentence >missing that starts "Do not"? > > "Do not reserve this unit for low level console operations. (Makes > kernel debugging easier at the expense of normal serial console > support.)" ??? Yes, the sentence somehow got truncated. It should say something like: "Do not probe the hardware for the device. (This makes debugging of kernel startup easier by avoiding state changes in the device being used to talk to the debugger. The context switching in i/o routines attempts to hide the state changes, but doesn't work in all cases.)" "Reserve" implies that other things aren't done. The device isn't attached, since it can't be used for normal i/o. This makes debugging of running kernels easier by avoiding state changes... >Would a low-level console operations include system messages written to >/dev/console? No. The console driver doesn't know about low level consoles. There are now 3 layers of consoles: /dev/console - an ordinary device, written to by syslogd, etc. low level part of console driver - used for kernel printfs low level consoles - lower than the console driver, called directly by i386_gdbstub. Bruce
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199705080042.KAA27367>