[ltp] Suspend and Hibernate w/ TP600 & RH7.1

[Byeong-ryeol Kim]

On Thu, 26 Apr 2001, Brian Lalor wrote:

> Is anyone running RedHat 7.1 on their ThinkPad 600?  I previously had 7.0
> up and running, but upgraded, and now I'm finding that *sometimes* when I
> suspend or hibernate and then resume (usually a few hours later), the
> machine just locks up.  This is with kernel 2.4.2.  There's nothing in the
> logs to speak of...
....

I'm not convinced the kernel-2.4.2 of Red Hat 7.1 is mostly stable, but
as far as APM stuff is concerned, I think as follows:

It is due to kernel configuration not adequate to ThinkPad models.
Red Hat have been configuring APM part in kernel like this:

CONFIG_PM=y
# CONFIG_ACPI is not set
CONFIG_APM=y
# CONFIG_APM_IGNORE_USER_SUSPEND is not set
# CONFIG_APM_DO_ENABLE is not set
# CONFIG_APM_CPU_IDLE is not set
# CONFIG_APM_DISPLAY_BLANK is not set
CONFIG_APM_RTC_IS_GMT=y
CONFIG_APM_ALLOW_INTS=y
CONFIG_APM_REAL_MODE_POWER_OFF=y

It seems that Red Hat changes this configuration in their Rawhide
kernel(2.4.3-*, added an option 'apm=allowints' to use at boot prompt
for ThinkPad users, pleas read APM part of
/usr/src/linux/Documentaion/Configure.help),
but, unless using 2.4.3 kernel package of Rawhide, ThinkPad users should
change kernel configuration like lines below and recompile theire kernels.

CONFIG_PM=y
# CONFIG_ACPI is not set
CONFIG_APM=y
# CONFIG_APM_IGNORE_USER_SUSPEND is not set
CONFIG_APM_DO_ENABLE=y
CONFIG_APM_CPU_IDLE=y
CONFIG_APM_DISPLAY_BLANK=y
# CONFIG_APM_RTC_IS_GMT is not set  <-- this is variable
CONFIG_APM_ALLOW_INTS=y
# CONFIG_APM_REAL_MODE_POWER_OFF is not set


> One weird thing I am seeing (that was also present with 7.0) is some odd
> timestamps coming from syslog when coming out of a suspend.  For example,
> these consecutive entries:
....

It is related with Red Hat's kernel configuration above, ie.,
'CONFIG_APM_RTC_IS_GMT is set' instead of '#CONFIG_APM_RTC_IS_GMT is not
set'.
It might be necessary to check and setup your system time with 'timeconfig'.

But, this is not a problem generally, since it is dependent on users'
conditions and their decisions.

>
> Also, if I've got the hotplug package installed, do I still need
> kernel-pcmcia-cs?  I'm seeing cardmgr complaining about an unsupported
> card, but it still works:
>
> Apr 26 07:30:57 localhost cardmgr[581]: initializing socket 1
> Apr 26 07:30:57 localhost cardmgr[581]: unsupported card in socket 1
> Apr 26 07:30:57 localhost kernel: 8139too Fast Ethernet driver 0.9.15
> loaded
> Apr 26 07:30:57 localhost kernel: PCI: Setting latency timer of device
> 04:00.0 to 64
> Apr 26 07:30:57 localhost kernel: eth0: RealTek RTL8139B PCI/CardBus at
> 0xd28b3000, 00:40:33:af:18:62, IRQ 11
> Apr 26 07:30:57 localhost kernel: eth0:  Identified 8139 chip type
> 'RTL-8139B'
> Apr 26 07:30:57 localhost cardmgr[581]:   product info: "ADDTRON", "AF-260
> 10/100Mbps Fast Ethernet CardBus PC Card", "01", "01"
> Apr 26 07:30:57 localhost cardmgr[581]:   manfid: 0x4602, 0x0004
> function: 6 (network)
> Apr 26 07:30:58 localhost /etc/hotplug/net.agent: register event not
> handled
>
>
> Clearly, the kernel sees the card, but is cardmgr just writing it off?
....

In 2.4.x pcmcia stuffs were incorporated into the kernel(the driver
part), but still needs utilities from pcmcia-cs package. Red Hat
repackagd this as kernel-pcmcia-cs. So kernel-pcmcia-cs is necessary
besides hotplug package(hotplug is mainly to handle the events while
plugging or unplugging pci hot-swap devices such as pcmcia, usb, etc.,
and kernel-pcmcia-cs package is to perform real opertion of pc, or
cardbus cards). At present, tt is known that hotplug and pcmcia drivers
of 2.4.x does not work perfectly, and as far as pc cards or cardbus
cards are concerned, hotplug util is not so necessary.

And in 2.4.x pcmcia, cardbus cards uses yenta_socket, so you need to set
it in /etc/sysconfig/pcmcia file.(to replace i82365 with yenta_socket).
I do not know well about the status of your cardbus(with 8139 chipset)
driver.
You should inspect more subtle option about youre 8139 related
driver(to use 'modinfo -p ...' could be a method).
The information about the status of pcmcia drivers of 2.4.x could be
obtained from the README-2.4 file of David Hinds' site, the maintainer
of pcmcia-cs package, ie., <ftp://pcmcia-cs.sourceforge.net/pub/pcmcia-cs/>.

Last, you should select only one method(the one is kernel-2.4.3,
kenrel-pcmcia-cs package, and the other is compiling kernel-2.4.x without
pcmcia support and using David Hinds' pcmcia-cs package).



----- The Linux ThinkPad mailing list -----
The linux-thinkpad mailing list home page is at:
http://www.bm-soft.com/~bm/tp_mailing.html