Bugzilla – Bug 130722
Add/removing cdrom to ultra bay causes issues
Last modified: 2006-06-07 16:26:34 UTC
On a thinkpad T42p if I insert my cdrom into the ultra bay while linux is running, hal does not see it (hal-device-manager does not report it). If I boot my machine with the cdrom inserted and then remove it, the machine locks up. I'm inserting/removing in order to add my ultra bay battery, I'll file a separate report for that.
Created attachment 55513 [details] hwinfo output of my machine
Do you know if that device works for anybody else with Linux? Usually it's not fixable, cause of the broken ide subsystem in the kernel. Only very few devices work, some devices work with with weird hdparm settings sometimes, but most of them don't.
I don't know for sure, but this link indicates it might be possible: http://theorie.physik.uni-wuerzburg.de/~arwagner/computer/T41p/
From the website: if you do: echo eject > /proc/acpi/ibm/bay Can you remove the drive safely and does it work if you plug it back?
No - the bay light goes off (not the dvd activity light, but a light on the machine for when something is inserted) but as soon as i yank it out. After reboot in /var/log/messages I see from the time I pulled it out: Nov 1 10:37:13 knight kernel: hdc: ATAPI reset complete Nov 1 10:37:13 knight kernel: hdc: status error: status=0x00 { } Nov 1 10:37:13 knight kernel: ide: failed opcode was: unknown Nov 1 10:37:13 knight kernel: hdc: status error: status=0x00 { } Nov 1 10:37:13 knight kernel: ide: failed opcode was: unknown Nov 1 10:37:13 knight kernel: hdc: status error: status=0x00 { } Nov 1 10:37:13 knight kernel: ide: failed opcode was: unknown Nov 1 10:37:13 knight kernel: hdc: status error: status=0x00 { }
As the hotswap tool seems to work, anything else we need, or can I close the bug?
Hotswap is not in autobuild yet, and we'll need a UI method of triggering it. Also, maybe this is simplistic, but if this utility can handle it the IDE hotswapping, why can't the kernel? Is it the surprise factor?
Cause the hotswap tool disables the ide controller _before_ the device is removed. The kernel can't handle a physical remove of an active controller. There is no solution currently. Windows probably gets an event for unlocking the latch to remove the device, but we have no idea how that works or where to get such a event from.
I get these ACPI events on a ThinkPad T43p when I unlock the latch to remove the drive. Theorethically it should be possible to use these events to prepare the removal of the device in the slot. This is a unlock/lock sequence, _without_ removing the drive: pim:~ # acpi_listen ibm/bay MSTR 00000003 00000000 ibm/bay MSTR 00000001 00000000 How does that look at your box?
I don't seem to be getting these with 2.6.15, and the hotswap tool isn't working for re-insertion either.
Stefan, care to put this bug in your magic queue. :) The current kernel can't handle this properly. I can't do anything at the moment.
Yes, I feared so. JP: To handle this properly we need to rewrite the ide-system of the kernel. A workaround might be for you to use the hotswap.rpm, but I don't think this will work for you. I'll create a feature entry for that.
I don't think we can fix this currently. I've created FATE 300648: Support IDE Hotplug for Laptops