I've got a home weather station connected to my computer via a serial cable communicating at 9600 baud. I used to have this same setup on a different computer running Fedora 14 and it worked just fine. I am running this inside a VMware virtual machine running XP. If I use the native Linux version of the weather software (which isn't as functional as the Windows version, which is why I'm running it in a virtual), I still get the same problem.
The serial is a NetMos Technology PCI 9835 Multi-I/O Controller with two serial ports.
The problem is that it works for 10-15 minutes and then the kernel shuts down the serial port. This is what appears in the messages file:
Jun 17 09:16:09 gmc kernel: [42503.704991] irq 18: nobody cared (try booting with the "irqpoll" option) Jun 17 09:16:09 gmc kernel: [42503.704995] Pid: 2336, comm: vmware-vmx Tainted: P 2.6.38.7-30.fc15.x86_64 #1 Jun 17 09:16:09 gmc kernel: [42503.704996] Call Trace: Jun 17 09:16:09 gmc kernel: [42503.704997] <IRQ> [<ffffffff8146dfb8>] __report_bad_irq+0x38/0x87 Jun 17 09:16:09 gmc kernel: [42503.705007] [<ffffffff810adf7c>] note_interrupt+0x122/0x18e Jun 17 09:16:09 gmc kernel: [42503.705009] [<ffffffff810aeaa0>] handle_fasteoi_irq+0xab/0xd7 Jun 17 09:16:09 gmc kernel: [42503.705012] [<ffffffff8100c0b5>] handle_irq+0x88/0x8e Jun 17 09:16:09 gmc kernel: [42503.705015] [<ffffffff8147bdc5>] do_IRQ+0x4d/0xa5 Jun 17 09:16:09 gmc kernel: [42503.705018] [<ffffffff81475d13>] ret_from_intr+0x0/0x15 Jun 17 09:16:09 gmc kernel: [42503.705019] <EOI> [<ffffffffa0bc87e4>] ? Task_Switch+0x6a1/0xc09 [vmmon] Jun 17 09:16:09 gmc kernel: [42503.705032] [<ffffffffa0bc9518>] Vmx86_RunVM+0x40/0x28a [vmmon] Jun 17 09:16:09 gmc kernel: [42503.705034] [<ffffffff81474228>] ? _cond_resched+0xe/0x22 Jun 17 09:16:09 gmc kernel: [42503.705037] [<ffffffffa0bc1611>] LinuxDriver_Ioctl+0x2b9/0xf6a [vmmon] Jun 17 09:16:09 gmc kernel: [42503.705040] [<ffffffff811ea1b0>] ? avc_has_perm+0x51/0x63 Jun 17 09:16:09 gmc kernel: [42503.705042] [<ffffffff811eb273>] ? inode_has_perm+0x76/0x8c Jun 17 09:16:09 gmc kernel: [42503.705044] [<ffffffff811ea737>] ? current_has_perm+0x36/0x3b Jun 17 09:16:09 gmc kernel: [42503.705047] [<ffffffff811eabd0>] ? selinux_task_kill+0x4f/0x51 Jun 17 09:16:09 gmc kernel: [42503.705049] [<ffffffff811e85a8>] ? security_task_kill+0x16/0x18 Jun 17 09:16:09 gmc kernel: [42503.705053] [<ffffffff81062fa6>] ? check_kill_permission+0x128/0x143 Jun 17 09:16:09 gmc kernel: [42503.705055] [<ffffffff811eb32d>] ? file_has_perm+0xa4/0xc6 Jun 17 09:16:09 gmc kernel: [42503.705058] [<ffffffffa0bc22da>] LinuxDriver_UnlockedIoctl+0x18/0x1c [vmmon] Jun 17 09:16:09 gmc kernel: [42503.705060] [<ffffffff8112f51c>] do_vfs_ioctl+0x47e/0x4bf Jun 17 09:16:09 gmc kernel: [42503.705062] [<ffffffff8112f5b3>] sys_ioctl+0x56/0x7b Jun 17 09:16:09 gmc kernel: [42503.705064] [<ffffffff81009bc2>] system_call_fastpath+0x16/0x1b Jun 17 09:16:09 gmc kernel: [42503.705065] handlers: Jun 17 09:16:09 gmc kernel: [42503.705066] [<ffffffff812cdea6>] (serial8250_interrupt+0x0/0x106) Jun 17 09:16:09 gmc kernel: [42503.705070] Disabling IRQ #18
Here is what my system shows for interrupt 18:
CPU0 CPU1 CPU2 CPU3 18: 200061 0 0 0 IO-APIC-fasteoi serial
I've tried moving the board to another PCI slot. That assigns it interrupt 19 instead, but the problem is the same.
Any clues on the problem or how to fix it?