Date: Wed, 23 Dec 1998 04:55:53 +1100 From: Bruce Evans <bde@zeta.org.au> To: crossd@cs.rpi.edu, freebsd-current@FreeBSD.ORG, viren@rstcorp.com Subject: re: error: configured irq 5 not in bitmap of probed irqs 0 Message-ID: <199812221755.EAA04557@godzilla.zeta.org.au>
next in thread | raw e-mail | index | archive | help
>total, fits nicely as a bitmask). The code then does a logical and with the >configured IRQ (that was supplied either via config(8) or '-c'), if the result >is true then it assigns the IRQ as specified, if false then it gives you that >error. Actually, if false it gives you that warning and then assigns the IRQ as specified (unless there was a more serious error). The specified IRQ is unlikely to work if this warning is printed, but always setting it gives you control. In 2.2.7, this error was fatal and that was important for preventing IRQ conflicts for misconfigured kernels (ones that don't specify _all_ the serial ports with their correct IRQs). Now the probe believes the configuration but prints the warning to help debug misconfigurations. Booting in verbose mode gives more details. >I did not dig through the code enough to notice if it traps ALL irqs during >this test, or just the un-assigned ones. If it only traps unassigned ones check >dmesg output for things prior to sio that are bound to IRQ 5. It checks all IRQs without trapping, but the check (and normal operation) may be broken by almost any device physically bound to the same irq, independently of the probe order (exception: sio devices on the same irq but not yet probed are specially handled so that they don't interfere with previous probes). Bruce To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199812221755.EAA04557>