2003-08-15 08:16:22

by Norman Diamond

[permalink] [raw]
Subject: Trying to run 2.6.0-test3

Sorry I cannot keep up with the list. Any advice or questions, personal
mail please. The following problems still exist in 2.6.0-test3. The
first one is still a show stopper.

1. Although both yenta and i82365 are compiled in, my 16-bit NE2000 clone
isn't recognized. If I boot kernel 2.4.19 I can use the network, if I
boot kernel 2.6.0 I can't find any way to use the network. Partial output
of various commands and files are shown below.

2. When I attach a USB hard disk drive, 2.6.0 drivers log the following:
[...] no runnable /etc/hotplug/scsi_device.agent is installed
2.4.19 didn't have this problem. 2.4.19 automatically updated /etc/fstab.

3. For 2.6.0-test1, a few people kindly explained and provided a patch
to make the keyboard work with a plain text console, to get a pipe symbol
when not running under X. How come 2.6.0-test3 still doesn't incorporate
that patch?

4. A hotkey on the keyboard, Fn+F10, is interpreted by the BIOS to do a
suspend-to-RAM. In an OS without APM support, the BIOS does it directly.
In Windows 98, it does what Microsoft calls Standby but is really suspend-
to-RAM. In Windows 2000, a vendor-supplied driver makes it do the same.
In Linux 2.4.19, it did what Intel calls Standby, what the apm command
calls "apm -S", which still drains the battery quickly. I patched 2.4.19
to make it do a suspend-to-RAM, what the apm command calls "apm -s".
But in 2.6.0 the hotkey is ignored. The kapmd driver recognizes things
like the battery level but never gets notified of standby requests.
(No I didn't configure the kernel to ignore them.) "apm -s" still works.

5. Modules seem to work except for module symbols. This seems to be a
result of compiling the new modules packages manually at a time when I
could not persuade the rpm --rebuild command to target the correct cpu.
Later I persuaded rpm --rebuild to work. modprobe and lsmod and rmmod
work, only kernel symbols think that modules are disabled.

6. It is still necessary to omit vga=788 from the kernel command line.
In kernel 2.4.19 this worked fine, but in 2.6.0 it blanks the screen.

7. The UDF driver makes some very strange observations about a CD drive
and CD media at boot time when no CD device even exists. I have used a
USB DVD+RW drive under 2.4.19 but have not even tried connecting it under
2.6.0 yet. (I also have a PCMCIA CD-ROM drive for which 2.6.0 finally
copied necessary code from the external PCMCIA package, but haven't tried
connecting it under 2.6.0 yet.)

8. As always, the OPL3/SA-whatever chip driver doesn't work. In some
previous versions I could make it buzz if I tried hard enough, but never
music, and usually only silence. Now it's silence. I'd better delete
the driver from my .config in order to hear the PCMCIA driver's tones
in case PCMCIA starts working.


(lspci)

00:0a.0 CardBus bridge: Texas Instruments PCI1250 (rev 02)
Flags: bus master, medium devsel, latency 168, IRQ 10
Memory at e0000000 (32-bit, non-prefetchable) [size=4K]
Bus: primary=00, secondary=01, subordinate=04, sec-latency=176
Memory window 0: 10400000-107ff000 (prefetchable)
Memory window 1: 10800000-10bff000
I/O window 0: 00001400-000014ff
I/O window 1: 00001800-000018ff
16-bit legacy interface ports at 0001

00:0a.1 CardBus bridge: Texas Instruments PCI1250 (rev 02)
Flags: bus master, medium devsel, latency 168, IRQ 10
Memory at e8000000 (32-bit, non-prefetchable) [size=4K]
Bus: primary=00, secondary=05, subordinate=08, sec-latency=176
Memory window 0: 10c00000-10fff000 (prefetchable)
Memory window 1: 11000000-113ff000
I/O window 0: 00001c00-00001cff
I/O window 1: 00002000-000020ff
16-bit legacy interface ports at 0001


(cardctl status)

Socket 0:
5V 16-bit PC Card
function 0: [ready], [wp]
Socket 1:
no card


(cardctl ident)

Socket 0:
no product info available
Socket 1:
no product info available


(lsmod)
(This is after doing modprobe pcnet_cs manually;
8390 and crc32 were loaded by that, but
videodev, evdev, and ntfs were already
loaded automatically.)

Module Size Used by
pcnet_cs 16644 0
8390 8960 1 pcnet_cs
crc32 4224 1 8390
videodev 7936 0
evdev 8192 0
ntfs 82708 1


(.config)

CONFIG_HOTPLUG=y

#
# PCMCIA/CardBus support
#
CONFIG_PCMCIA=y
CONFIG_YENTA=y
CONFIG_CARDBUS=y
# CONFIG_I82092 is not set
CONFIG_I82365=y
# CONFIG_TCIC is not set
CONFIG_PCMCIA_PROBE=y

#
# PCI Hotplug Support
#
CONFIG_HOTPLUG_PCI=m
CONFIG_HOTPLUG_PCI_FAKE=m


(/var/log/messages)

Aug 14 22:07:55 diamondpana syslogd 1.4.1: restart.
Aug 14 22:07:56 diamondpana /etc/hotplug/usb.rc[472]: FATAL: Module joydev not found.
Aug 14 22:08:00 diamondpana kernel: klogd 1.4.1, log source = /proc/kmsg started.
Aug 14 22:08:00 diamondpana kernel: Inspecting /boot/System.map-2.6.0-test3
Aug 14 22:08:01 diamondpana kernel: Loaded 26800 symbols from /boot/System.map-2.6.0-test3.
Aug 14 22:08:01 diamondpana kernel: Symbols match kernel version 2.6.0.
Aug 14 22:08:01 diamondpana kernel: No module symbols loaded - kernel modules not enabled.
Aug 14 22:08:04 diamondpana SuSEfirewall2: Firewall rules successfully set in QUICKMODE for device(s) " ppp+"
Aug 14 22:08:06 diamondpana modprobe: FATAL: Module ipv6 not found.
Aug 14 22:08:06 diamondpana sshd[669]: Server listening on 0.0.0.0 port 22.
Aug 14 22:08:17 diamondpana SuSEfirewall2: Firewall rules successfully set in QUICKMODE for device(s) " ppp+"
Aug 14 22:08:19 diamondpana smpppd[835]: smpppd version 0.78 started
Aug 14 22:08:31 diamondpana /usr/sbin/cron[1065]: (CRON) STARTUP (fork ok)
Aug 14 22:08:38 diamondpana modprobe: FATAL: Module sr_mod not found.
Aug 14 22:08:47 diamondpana kernel: Linux video capture interface: v1.00
Aug 14 22:09:07 diamondpana kdm[1154]: pam_unix2: session started for user ndiamond, service xdm
Aug 14 22:10:00 diamondpana /USR/SBIN/CRON[1260]: (root) CMD ( /usr/lib/sa/sa1 )
Aug 14 22:10:52 diamondpana su: (to root) ndiamond on /dev/pts/1
Aug 14 22:10:52 diamondpana su: pam_unix2: session started for user root, service su
Aug 14 22:20:00 diamondpana /USR/SBIN/CRON[1493]: (root) CMD ( /usr/lib/sa/sa1 )
Aug 14 22:21:37 diamondpana kernel: cs: warning: no high memory space available!
Aug 14 22:21:37 diamondpana kernel: cs: unable to map card memory!
Aug 14 22:21:37 diamondpana last message repeated 3 times
Aug 14 22:22:49 diamondpana last message repeated 7 times
Aug 14 22:23:35 diamondpana last message repeated 3 times
Aug 14 22:24:40 diamondpana last message repeated 3 times
Aug 14 22:29:07 diamondpana kernel: cs: unable to map card memory!


(dmesg)

Linux version 2.6.0-test3 (ndiamond@diamondpana) (gcc version 3.2) #1 Thu Aug 14 21:00:20 JST 2003
Video mode to be used for restore is f00
BIOS-provided physical RAM map:
BIOS-e820: 0000000000000000 - 000000000009fc00 (usable)
BIOS-e820: 000000000009fc00 - 00000000000a0000 (reserved)
BIOS-e820: 00000000000f0000 - 0000000000100000 (reserved)
BIOS-e820: 0000000000100000 - 0000000005ff0000 (usable)
BIOS-e820: 0000000005ff0000 - 0000000006000000 (ACPI data)
BIOS-e820: 0000000010000000 - 0000000010100000 (reserved)
BIOS-e820: 00000000fff80000 - 0000000100000000 (reserved)
95MB LOWMEM available.
On node 0 totalpages: 24560
DMA zone: 4096 pages, LIFO batch:1
Normal zone: 20464 pages, LIFO batch:4
HighMem zone: 0 pages, LIFO batch:1
Building zonelist for node : 0
Kernel command line: acpi=off apm=on root=/dev/hda8
No local APIC present or hardware disabled
Initializing CPU#0
PID hash table entries: 512 (order 9: 4096 bytes)
Detected 198.983 MHz processor.
Console: colour VGA+ 80x25
Calibrating delay loop... 392.19 BogoMIPS
Memory: 92704k/98240k available (2549k kernel code, 5036k reserved, 1130k data, 172k init, 0k highmem)
Security Scaffold v1.0.0 initialized
Capability LSM initialized
Dentry cache hash table entries: 16384 (order: 4, 65536 bytes)
Inode-cache hash table entries: 8192 (order: 3, 32768 bytes)
Mount-cache hash table entries: 512 (order: 0, 4096 bytes)
-> /dev
-> /dev/console
-> /root
CPU: After generic identify, caps: 008001bf 00000000 00000000 00000000
CPU: After vendor identify, caps: 008001bf 00000000 00000000 00000000
Intel Pentium with F0 0F bug - workaround enabled.
CPU: After all inits, caps: 008001bf 00000000 00000000 00000000
CPU: Intel Mobile Pentium MMX stepping 01
Checking 'hlt' instruction... OK.
POSIX conformance testing by UNIFIX
Initializing RT netlink socket
PCI: PCI BIOS revision 2.10 entry at 0xfd74c, last bus=0
PCI: Using configuration type 1
mtrr: v2.0 (20020519)
BIO: pool of 256 setup, 15Kb (60 bytes/bio)
biovec pool[0]: 1 bvecs: 180 entries (12 bytes)
biovec pool[1]: 4 bvecs: 180 entries (48 bytes)
biovec pool[2]: 16 bvecs: 180 entries (192 bytes)
biovec pool[3]: 64 bvecs: 90 entries (768 bytes)
biovec pool[4]: 128 bvecs: 45 entries (1536 bytes)
biovec pool[5]: 256 bvecs: 22 entries (3072 bytes)
Linux Plug and Play Support v0.97 (c) Adam Belay
PnPBIOS: Scanning system for PnP BIOS support...
PnPBIOS: Found PnP BIOS installation structure at 0xc00fdf90
PnPBIOS: PnP BIOS version 1.0, entry 0xf0000:0x58ed, dseg 0xf0000
PnPBIOS: 23 nodes reported by PnP BIOS; 23 recorded by driver
SCSI subsystem initialized
Linux Kernel Card Services 3.1.22
options: [pci] [cardbus] [pm]
drivers/usb/core/usb.c: registered new driver usbfs
drivers/usb/core/usb.c: registered new driver hub
PCI: Probing PCI hardware
PCI: Probing PCI hardware (bus 00)
PCI: Using IRQ router PIIX [8086/7110] at 0000:00:01.0
PCI: IRQ 0 for device 0000:00:01.2 doesn't match PIRQ mask - try pci=usepirqmask
PCI: IRQ 0 for device 0000:00:0a.0 doesn't match PIRQ mask - try pci=usepirqmask
PCI: IRQ 0 for device 0000:00:0a.1 doesn't match PIRQ mask - try pci=usepirqmask
PCI: Found IRQ 10 for device 0000:00:02.0
neofb: mapped io at c6800000
Autodetected internal display
Panel is a 800x600 color TFT display
neofb: mapped framebuffer at c6a01000
neofb v0.4.1: 2048kB VRAM, using 800x600, 37.878kHz, 60Hz
fb0: MagicGraph 128XD frame buffer device
vga16fb: initializing
vga16fb: mapped to 0xc00a0000
fb1: VGA16 VGA frame buffer device
pty: 256 Unix98 ptys configured
apm: BIOS version 1.2 Flags 0x03 (Driver version 1.16ac)
Journalled Block Device driver loaded
devfs: v1.22 (20021013) Richard Gooch ([email protected])
devfs: boot_options: 0x0
Installing knfsd (copyright (C) 1996 [email protected]).
udf: registering filesystem
Initializing Cryptographic API
Limiting direct PCI/PCI transfers.
isapnp: Scanning for PnP cards...
isapnp: No Plug & Play device found
request_module: failed /sbin/modprobe -- parport_lowlevel. error = -16
lp: driver loaded but no devices found
Serial: 8250/16550 driver $Revision: 1.90 $ IRQ sharing disabled
ttyS0 at I/O 0x3f8 (irq = 4) is a 16550A
ttyS1 at I/O 0x2f8 (irq = 3) is a NS16550A
parport0: PC-style at 0x378 (0x778) [PCSPP,TRISTATE]
parport0: irq 7 detected
parport0: cpp_daisy: aa5500ff(38)
parport0: assign_addrs: aa5500ff(38)
parport0: cpp_daisy: aa5500ff(38)
parport0: assign_addrs: aa5500ff(38)
lp0: using parport0 (polling).
Using anticipatory scheduling elevator
Floppy drive(s): fd0 is 1.44M
FDC 0 is a National Semiconductor PC87306
Uniform Multi-Platform E-IDE driver Revision: 7.00alpha2
ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
PIIX4: IDE controller at PCI slot 0000:00:01.1
PIIX4: chipset revision 1
PIIX4: not 100% native mode: will probe irqs later
ide0: BM-DMA at 0xffa0-0xffa7, BIOS settings: hda:pio, hdb:pio
hda: TOSHIBA MK4018GAP, ATA DISK drive
ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
hda: max request size: 128KiB
hda: 78140160 sectors (40007 MB), CHS=65535/16/63, UDMA(33)
/dev/ide/host0/bus0/target0/lun0: p1 p2 p3 < p5 p6 p7 p8 p9 p10 >
PCI: IRQ 0 for device 0000:00:0a.0 doesn't match PIRQ mask - try pci=usepirqmask
PCI: Assigned IRQ 10 for device 0000:00:0a.0
Yenta: CardBus bridge found at 0000:00:0a.0 [0000:0000]
Yenta IRQ list 0ab8, PCI irq10
Socket status: 30000010
PCI: IRQ 0 for device 0000:00:0a.1 doesn't match PIRQ mask - try pci=usepirqmask
PCI: Assigned IRQ 10 for device 0000:00:0a.1
Yenta: CardBus bridge found at 0000:00:0a.1 [0000:0000]
Yenta IRQ list 0ab8, PCI irq10
Socket status: 30000006
Intel PCIC probe: not found.
drivers/usb/host/uhci-hcd.c: USB Universal Host Controller Interface driver v2.1
PCI: Enabling device 0000:00:01.2 (0000 -> 0001)
PCI: IRQ 0 for device 0000:00:01.2 doesn't match PIRQ mask - try pci=usepirqmask
PCI: Assigned IRQ 10 for device 0000:00:01.2
uhci-hcd 0000:00:01.2: UHCI Host Controller
uhci-hcd 0000:00:01.2: irq 10, io base 00001000
uhci-hcd 0000:00:01.2: new USB bus registered, assigned bus number 1
hub 1-0:0: USB hub found
hub 1-0:0: 2 ports detected
Initializing USB Mass Storage driver...
drivers/usb/core/usb.c: registered new driver usb-storage
USB Mass Storage support registered.
mice: PS/2 mouse device common for all mice
input: PS/2 Logitech Mouse on isa0060/serio1
serio: i8042 AUX port at 0x60,0x64 irq 12
input: AT Set 2 keyboard on isa0060/serio0
serio: i8042 KBD port at 0x60,0x64 irq 1
NET4: Linux TCP/IP 1.0 for NET4.0
IP: routing cache hash table of 512 buckets, 4Kbytes
TCP: Hash tables configured (established 8192 bind 16384)
NET4: Unix domain sockets 1.0/SMP for Linux NET4.0.
BIOS EDD facility v0.09 2003-Jan-22, 1 devices found
UDF-fs DEBUG fs/udf/lowlevel.c:65:udf_get_last_session: CDROMMULTISESSION not supported: rc=-22
UDF-fs DEBUG fs/udf/super.c:1472:udf_fill_super: Multi-session=0
UDF-fs DEBUG fs/udf/super.c:460:udf_vrs: Starting at sector 16 (2048 byte sectors)
UDF-fs DEBUG fs/udf/super.c:1208:udf_check_valid: Failed to read byte 32768. Assuming open disc. Skipping validity check
UDF-fs DEBUG fs/udf/misc.c:286:udf_read_tagged: location mismatch block 256, tag 1413234186 != 256
UDF-fs DEBUG fs/udf/super.c:1262:udf_load_partition: No Anchor block found
UDF-fs: No partition found (1)
found reiserfs format "3.6" with standard journal
Reiserfs journal params: device hda8, size 8192, journal first block 18, max trans len 1024, max batch 900, max commit age 30, max
trans age 30
reiserfs: checking transaction log (hda8) for (hda8)
Using r5 hash to sort names
VFS: Mounted root (reiserfs filesystem) readonly.
Freeing unused kernel memory: 172k freed
Adding 1092380k swap on /dev/hda7. Priority:42 extents:1
NTFS driver 2.1.4 [Flags: R/O MODULE].
NTFS volume version 3.1.
Linux video capture interface: v1.00


2003-08-15 08:53:52

by Felipe Alfaro Solana

[permalink] [raw]
Subject: Re: Trying to run 2.6.0-test3

On Fri, 2003-08-15 at 10:11, Norman Diamond wrote:

> 1. Although both yenta and i82365 are compiled in, my 16-bit NE2000 clone
> isn't recognized. If I boot kernel 2.4.19 I can use the network, if I
> boot kernel 2.6.0 I can't find any way to use the network. Partial output
> of various commands and files are shown below.

What brand/model? What chipset does the card use? Are your sure it's
PCMCIA (16-bit) and not CardBus (32-bit)?

> 3. For 2.6.0-test1, a few people kindly explained and provided a patch
> to make the keyboard work with a plain text console, to get a pipe symbol
> when not running under X. How come 2.6.0-test3 still doesn't incorporate
> that patch?

Uh? A patch to make the pipe (|) symbol work?

> 5. Modules seem to work except for module symbols. This seems to be a
> result of compiling the new modules packages manually at a time when I
> could not persuade the rpm --rebuild command to target the correct cpu.
> Later I persuaded rpm --rebuild to work. modprobe and lsmod and rmmod
> work, only kernel symbols think that modules are disabled.

Have you upgraded to latest modutils package? Modules are implemented
quite differently in 2.6.

> #
> # PCI Hotplug Support
> #
> CONFIG_HOTPLUG_PCI=m
> CONFIG_HOTPLUG_PCI_FAKE=m

You don't need that.

>
>
> (/var/log/messages)
>
> Aug 14 22:21:37 diamondpana kernel: cs: warning: no high memory space available!
> Aug 14 22:21:37 diamondpana kernel: cs: unable to map card memory!
> Aug 14 22:21:37 diamondpana last message repeated 3 times
> Aug 14 22:22:49 diamondpana last message repeated 7 times
> Aug 14 22:23:35 diamondpana last message repeated 3 times
> Aug 14 22:24:40 diamondpana last message repeated 3 times
> Aug 14 22:29:07 diamondpana kernel: cs: unable to map card memory!

It seems the kernel is having problems assigning resources to your card.
Plase, edit "/etc/pcmcia/config.opts" and uncomment the following line:

#include port 0x1000-0x17ff

>
>
> (dmesg)
>
> Kernel command line: acpi=off apm=on root=/dev/hda8

I suggest you to recompile your kernel with ACPI support to see if it
works correctly. Else, recompile it disabling ACPI support and enabling
APM support.

> No local APIC present or hardware disabled

You can remove APIC support from your config. It's not supported by your
hardware.

> PCI: IRQ 0 for device 0000:00:01.2 doesn't match PIRQ mask - try pci=usepirqmask
> PCI: IRQ 0 for device 0000:00:0a.0 doesn't match PIRQ mask - try pci=usepirqmask
> PCI: IRQ 0 for device 0000:00:0a.1 doesn't match PIRQ mask - try pci=usepirqmask

Please, add "pci=usepirqmask" to your kernel command line in GRUB/Lilo.

2003-08-15 10:02:17

by Norman Diamond

[permalink] [raw]
Subject: Re: Trying to run 2.6.0-test3

"Felipe Alfaro Solana" <[email protected]> replied to me:

> > 1. Although both yenta and i82365 are compiled in, my 16-bit NE2000 clone
> > isn't recognized. If I boot kernel 2.4.19 I can use the network, if I
> > boot kernel 2.6.0 I can't find any way to use the network. Partial output
> > of various commands and files are shown below.
>
> What brand/model? What chipset does the card use? Are your sure it's
> PCMCIA (16-bit) and not CardBus (32-bit)?

I think it's Greenhouse ELP-10T. But I think it doesn't matter. Under
2.6.0-test3, "cardctl ident" couldn't even read the card's ID. If it's
really necessary I can boot 2.4.19 to run "cardctl ident". I don't know the
chipset but I think it doesn't matter. As for knowing that it's 16-bit and
not 32-bit, even though I know you're even more of a volunteer than I am,
it's hard to keep civil about this. Yes I know that it's 16-bit. But even
if it magically changed to 32-bit, then the yenta driver should take over,
right? But neither PCMCIA driver recognized the thing under 2.6.0-test3.
Oh, you might notice that "cardctl status" figured out that the thing takes
5 volts. This is why I posted the outputs of "cardctl ident" and "cardctl
status" and the relevant part of .config.

> > 3. For 2.6.0-test1, a few people kindly explained and provided a patch
> > to make the keyboard work with a plain text console, to get a pipe symbol
> > when not running under X. How come 2.6.0-test3 still doesn't incorporate
> > that patch?
>
> Uh? A patch to make the pipe (|) symbol work?

Yes. Read the archives about Japanese keyboards. A patch was even POSTED,
in the -test1 days. How come the patch didn't get into -test3?

> > 5. Modules seem to work except for module symbols. This seems to be a
> > result of compiling the new modules packages manually at a time when I
> > could not persuade the rpm --rebuild command to target the correct cpu.
> > Later I persuaded rpm --rebuild to work. modprobe and lsmod and rmmod
> > work, only kernel symbols think that modules are disabled.
>
> Have you upgraded to latest modutils package? Modules are implemented
> quite differently in 2.6.

"seems to be a result of compiling the new modules packages manually at a
time when I could not persuade the rpm --rebuild command to target the
correct cpu." Yes, "the new modules packages" meant the new modules
packages. I did have to ask about this at the beginning of -test1 days. (I
never had time to try testing a 2.5 kernel but the appearance of 2.6.0-test1
made it look like testing would become valuable.)

> > #
> > # PCI Hotplug Support
> > #
> > CONFIG_HOTPLUG_PCI=m
> > CONFIG_HOTPLUG_PCI_FAKE=m
>
> You don't need that.

I didn't think I needed it, but I didn't think it would hurt. After all I
was trying to get some PCMCIA cards to work, and someday might try getting a
Cardbus card to work, and PCMCIA wasn't working yet, so I tried it.

> > (/var/log/messages)
> >
> > Aug 14 22:21:37 diamondpana kernel: cs: warning: no high memory space available!
> > Aug 14 22:21:37 diamondpana kernel: cs: unable to map card memory!
> > Aug 14 22:21:37 diamondpana last message repeated 3 times
> > Aug 14 22:22:49 diamondpana last message repeated 7 times
> > Aug 14 22:23:35 diamondpana last message repeated 3 times
> > Aug 14 22:24:40 diamondpana last message repeated 3 times
> > Aug 14 22:29:07 diamondpana kernel: cs: unable to map card memory!
>
> It seems the kernel is having problems assigning resources to your card.
> Plase, edit "/etc/pcmcia/config.opts" and uncomment the following line:
>
> #include port 0x1000-0x17ff

This I will take a look at. Thank you. Though it makes me curious why
2.4.19 didn't need this, I will try it. And I think I will have no problem
remaining civil while discussing it. Thank you.

> > Kernel command line: acpi=off apm=on root=/dev/hda8
>
> I suggest you to recompile your kernel with ACPI support to see if it
> works correctly. Else, recompile it disabling ACPI support and enabling
> APM support.

Guess why I compiled it without ACPI support and with APM support. Guess
why my kernel command line has acpi=off apm=on. (Although the command line
options are redundant with the self-compiled kernel configuration, they are
necessary when booting a generic kernel. Yes I know that the machine has
just enough ACPI hooks to cause problems when anyone other than Windows 98
tries to use ACPI. It's not even recommended to force ACPI on when
installing Windows 98 on this machine. Windows 2000 blue screens if ACPI is
forced on. Linux doesn't panic when its default ACPI takes over, but it
does prevent APM from working.)

> > No local APIC present or hardware disabled
>
> You can remove APIC support from your config. It's not supported by your
> hardware.

Fine, but I didn't think I had to worry about that, since it looked like
Linux already properly detected the fact.

> > PCI: IRQ 0 for device 0000:00:01.2 doesn't match PIRQ mask - try pci=usepirqmask
> > PCI: IRQ 0 for device 0000:00:0a.0 doesn't match PIRQ mask - try pci=usepirqmask
> > PCI: IRQ 0 for device 0000:00:0a.1 doesn't match PIRQ mask - try pci=usepirqmask
>
> Please, add "pci=usepirqmask" to your kernel command line in GRUB/Lilo.

Oh, that could be it, I just noticed that the second and third occurences
relate to the PCMCIA slots. May I ask though, what PIRQ mask is it, that is
failing to be matched? It seems to me that it might be more useful for
2.6.0 to do as good a job as 2.4.19 did in figuring out the IRQs, and I
would like to continue testing this. (Probably I can only spend about one
day a week on it though.)

2003-08-15 10:14:50

by Russell King

[permalink] [raw]
Subject: Re: Trying to run 2.6.0-test3

On Fri, Aug 15, 2003 at 05:11:42PM +0900, Norman Diamond wrote:
> 1. Although both yenta and i82365 are compiled in, my 16-bit NE2000 clone
> isn't recognized. If I boot kernel 2.4.19 I can use the network, if I
> boot kernel 2.6.0 I can't find any way to use the network. Partial output
> of various commands and files are shown below.

As a general rule, you should be using yenta not i82365. There have
been some historical problems when you build both into the kernel,
so you might like to try disabling i82365.

I don't see any sign of cardmgr starting up in your message logs, and
cardmgr is still required to tell the kernel which resources it can
use, and to bind the driver to the device.

Start cardmgr and it should work.

--
Russell King ([email protected]) The developer of ARM Linux
http://www.arm.linux.org.uk/personal/aboutme.html

2003-08-15 11:54:26

by Felipe Alfaro Solana

[permalink] [raw]
Subject: Re: Trying to run 2.6.0-test3

On Fri, 2003-08-15 at 12:00, Norman Diamond wrote:

> > Have you upgraded to latest modutils package? Modules are implemented
> > quite differently in 2.6.
>
> "seems to be a result of compiling the new modules packages manually at a
> time when I could not persuade the rpm --rebuild command to target the
> correct cpu." Yes, "the new modules packages" meant the new modules
> packages. I did have to ask about this at the beginning of -test1 days. (I
> never had time to try testing a 2.5 kernel but the appearance of 2.6.0-test1
> made it look like testing would become valuable.)

I can't understand why manually compiling modules (make modules) could
cause you problems that "rpm --rebuild" couldn't.

> > > Kernel command line: acpi=off apm=on root=/dev/hda8
> >
> > I suggest you to recompile your kernel with ACPI support to see if it
> > works correctly. Else, recompile it disabling ACPI support and enabling
> > APM support.
>
> Guess why I compiled it without ACPI support and with APM support. Guess
> why my kernel command line has acpi=off apm=on. (Although the command line
> options are redundant with the self-compiled kernel configuration, they are
> necessary when booting a generic kernel. Yes I know that the machine has
> just enough ACPI hooks to cause problems when anyone other than Windows 98
> tries to use ACPI. It's not even recommended to force ACPI on when
> installing Windows 98 on this machine. Windows 2000 blue screens if ACPI is
> forced on. Linux doesn't panic when its default ACPI takes over, but it
> does prevent APM from working.)

If you turn ACPI on, you won't need APM support. Well, ACPI has still
some broken functionality, like S3 states, but usually, you can do
pretty well with ACPI. There was a long time ago since I used APM for
the very last time.

> Fine, but I didn't think I had to worry about that, since it looked like
> Linux already properly detected the fact.
>
> > > PCI: IRQ 0 for device 0000:00:01.2 doesn't match PIRQ mask - try pci=usepirqmask
> > > PCI: IRQ 0 for device 0000:00:0a.0 doesn't match PIRQ mask - try pci=usepirqmask
> > > PCI: IRQ 0 for device 0000:00:0a.1 doesn't match PIRQ mask - try pci=usepirqmask
> >
> > Please, add "pci=usepirqmask" to your kernel command line in GRUB/Lilo.
>
> Oh, that could be it, I just noticed that the second and third occurences
> relate to the PCMCIA slots. May I ask though, what PIRQ mask is it, that is
> failing to be matched? It seems to me that it might be more useful for
> 2.6.0 to do as good a job as 2.4.19 did in figuring out the IRQs, and I
> would like to continue testing this. (Probably I can only spend about one
> day a week on it though.)

To be sincere, I don't know exactly why "pci=usepirqmask" needs to be
used. I'm no hardware expert. But I know that I needed it when I wasn't
using ACPI.

Probably, if you disable APM support and enable ACPI, you won't need to
boot using "pci=usepirqmask".

2003-08-15 12:42:24

by Norman Diamond

[permalink] [raw]
Subject: Re: Trying to run 2.6.0-test3

"Russell King" <[email protected]> replied to me:

> > 1. Although both yenta and i82365 are compiled in, my 16-bit NE2000 clone
> > isn't recognized. If I boot kernel 2.4.19 I can use the network, if I
> > boot kernel 2.6.0 I can't find any way to use the network. Partial output
> > of various commands and files are shown below.
>
> As a general rule, you should be using yenta not i82365. There have
> been some historical problems when you build both into the kernel,
> so you might like to try disabling i82365.

I will do that in my next build. For some reason I wasn't sure if yenta
would handle 16-bit cards. But this turns out not to be necessary. Also
the PCMCIA suggestions which Felipe Alfaro Solana suggested (the suggestions
which I intended to try) turned out not to be necessary. The winner is the
next one:

> I don't see any sign of cardmgr starting up in your message logs,

You're right. I started it from the command line. It didn't recognize the
existing card, but "cardctl eject 0" followed by "cardctl insert 0" taught it.

Now the question is why cardmgr doesn't start automatically in 2.6.0-test3.
In 2.4.19, PCMCIA support was a module, for which I guess I never bothered
to change SuSE's default. In 2.6.0-test1 since had to set all options
myself, I compiled in PCMCIA support, and it's still that way in test3.
I'll try to see if I can find the reason, though of course you might be able
to guess it instantaneously.

Thank you.

2003-08-15 13:04:44

by Mikael Pettersson

[permalink] [raw]
Subject: Re: Trying to run 2.6.0-test3

Norman Diamond writes:
> Now the question is why cardmgr doesn't start automatically in 2.6.0-test3.

You probably need fixes to /etc/rc.d/init.d/pcmcia (drop module ".o"
extensions from modprobe commands) and /etc/hotplug/net.agent (case
$ACTION needs "add|register" not just "register").

/Mikael

2003-08-15 13:24:03

by Norman Diamond

[permalink] [raw]
Subject: Re: Trying to run 2.6.0-test3

"Felipe Alfaro Solana" <[email protected]> replied to me:

> > "seems to be a result of compiling the new modules packages manually at a
> > time when I could not persuade the rpm --rebuild command to target the
> > correct cpu." Yes, "the new modules packages" meant the new modules
> > packages. I did have to ask about this at the beginning of -test1 days. (I
> > never had time to try testing a 2.5 kernel but the appearance of 2.6.0-test1
> > made it look like testing would become valuable.)
>
> I can't understand why manually compiling modules (make modules) could
> cause you problems that "rpm --rebuild" couldn't.

Manually compiling the two items that are contained in the rpm (modutils and
module-init-tools) did indeed give different problems from "rpm --rebuild".
I reported them around two weeks ago or so. "make modules" has nothing to
do with this.

> > > > Kernel command line: acpi=off apm=on root=/dev/hda8
> > >
> > > I suggest you to recompile your kernel with ACPI support to see if it
> > > works correctly. Else, recompile it disabling ACPI support and enabling
> > > APM support.
> >
> > Guess why I compiled it without ACPI support and with APM support. Guess
> > why my kernel command line has acpi=off apm=on. (Although the command line
> > options are redundant with the self-compiled kernel configuration, they are
> > necessary when booting a generic kernel. Yes I know that the machine has
> > just enough ACPI hooks to cause problems when anyone other than Windows 98
> > tries to use ACPI. It's not even recommended to force ACPI on when
> > installing Windows 98 on this machine. Windows 2000 blue screens if ACPI is
> > forced on. Linux doesn't panic when its default ACPI takes over, but it
> > does prevent APM from working.)
>
> If you turn ACPI on, you won't need APM support.

WRONG. ACPI DOESN'T WORK ON THE MACHINE I'M DOING THIS ON. DID
YOU TRY READING WHAT YOU QUOTED THERE? Yes I know you volunteer
more effort on Linux than I do, but you're asking me to flame you anyway.
How many times do you need to be told?

> Well, ACPI has still some broken functionality, like S3 states,

S3 has nothing to do with it. Whether and how much ACPI functionality the
vendor put in the BIOS has everything to do with it. If you read what you
quoted from me, you might reach the same opinion I have, that it would have
been better for the vendor to include no ACPI functionality at all instead
of some experiments at partial ACPI support. But even if you don't agree,
ACPI still doesn't work on that machine.

> To be sincere, I don't know exactly why "pci=usepirqmask" needs to be
> used. I'm no hardware expert. But I know that I needed it when I wasn't
> using ACPI.

Hmm. Then some dependency seems to be broken in kernel compilation. When
ACPI is not compiled in, it should know that the effect of "pci=usepirqmask"
should be compiled in (whatever that effect is).

> Probably, if you disable APM support and enable ACPI, you won't need to
> boot using "pci=usepirqmask".

Could be, but then I'll have no power management at all because ACPI doesn't
work but APM will disable itself when APM thinks that ACPI looks like it's
running. How many times do you need to be told, guess why I did "acpi=off
apm=on".

2003-08-15 16:44:46

by Joshua Kwan

[permalink] [raw]
Subject: Re: Trying to run 2.6.0-test3

On Fri, Aug 15, 2003 at 09:39:07PM +0900, Norman Diamond wrote:
> I will do that in my next build. For some reason I wasn't sure if yenta
> would handle 16-bit cards. But this turns out not to be necessary. Also
> the PCMCIA suggestions which Felipe Alfaro Solana suggested (the suggestions
> which I intended to try) turned out not to be necessary. The winner is the
> next one:

Note that if you don't already have CONFIG_ISA enabled, 16-bit CardBus
devices will refuse to work.

--
Joshua Kwan


Attachments:
(No filename) (502.00 B)
(No filename) (827.00 B)
Download all attachments

2003-08-15 17:15:16

by Russell King

[permalink] [raw]
Subject: Re: Trying to run 2.6.0-test3

On Fri, Aug 15, 2003 at 10:06:56AM -0700, Randy.Dunlap wrote:
> 16-bit PCMCIA is ISA-like, 32-bit CardBus is basically PCI.
> I've never heard of 16-bit CardBus. Are there some?

No.

--
Russell King ([email protected]) The developer of ARM Linux
http://www.arm.linux.org.uk/personal/aboutme.html

2003-08-15 17:10:26

by Randy.Dunlap

[permalink] [raw]
Subject: Re: Trying to run 2.6.0-test3

On Fri, 15 Aug 2003 09:43:00 -0700 Joshua Kwan <[email protected]> wrote:

| On Fri, Aug 15, 2003 at 09:39:07PM +0900, Norman Diamond wrote:
| > I will do that in my next build. For some reason I wasn't sure if yenta
| > would handle 16-bit cards. But this turns out not to be necessary. Also
| > the PCMCIA suggestions which Felipe Alfaro Solana suggested (the suggestions
| > which I intended to try) turned out not to be necessary. The winner is the
| > next one:
|
| Note that if you don't already have CONFIG_ISA enabled, 16-bit CardBus
| devices will refuse to work.

16-bit PCMCIA is ISA-like, 32-bit CardBus is basically PCI.
I've never heard of 16-bit CardBus. Are there some?

--
~Randy

2003-08-15 16:55:12

by Felipe Alfaro Solana

[permalink] [raw]
Subject: Re: Trying to run 2.6.0-test3

On Fri, 2003-08-15 at 15:20, Norman Diamond wrote:

> > > Guess why I compiled it without ACPI support and with APM support. Guess
> > > why my kernel command line has acpi=off apm=on. (Although the command line
> > > options are redundant with the self-compiled kernel configuration, they are
> > > necessary when booting a generic kernel. Yes I know that the machine has
> > > just enough ACPI hooks to cause problems when anyone other than Windows 98
> > > tries to use ACPI. It's not even recommended to force ACPI on when
> > > installing Windows 98 on this machine. Windows 2000 blue screens if ACPI is
> > > forced on. Linux doesn't panic when its default ACPI takes over, but it
> > > does prevent APM from working.)
> >
> > If you turn ACPI on, you won't need APM support.
>
> WRONG. ACPI DOESN'T WORK ON THE MACHINE I'M DOING THIS ON. DID
> YOU TRY READING WHAT YOU QUOTED THERE? Yes I know you volunteer
> more effort on Linux than I do, but you're asking me to flame you anyway.
> How many times do you need to be told?

Yes, I tried reading. You said Linux doesn't panic while using ACPI, so
I supposed ACPI just worked but the problem was you wanted APM support.

> > To be sincere, I don't know exactly why "pci=usepirqmask" needs to be
> > used. I'm no hardware expert. But I know that I needed it when I wasn't
> > using ACPI.
>
> Hmm. Then some dependency seems to be broken in kernel compilation. When
> ACPI is not compiled in, it should know that the effect of "pci=usepirqmask"
> should be compiled in (whatever that effect is).

It's not a problem with dependencies. On ACPI-enabled kernels, you using
ACPI routing. If you boot using "acpi=off", then you're using standard
PCI routing and that, in turn, on same machines, it warns you to use
"pci=usepirqmask". Don't know why, on some machines using standard PCI
routing, you don't need to boot with "pci=usepirqmask". I suppose it
will be some kind of hardware incompatibility or a bad implementation.


2003-08-15 17:21:42

by Greg KH

[permalink] [raw]
Subject: Re: Trying to run 2.6.0-test3

On Fri, Aug 15, 2003 at 05:11:42PM +0900, Norman Diamond wrote:
>
> 2. When I attach a USB hard disk drive, 2.6.0 drivers log the following:
> [...] no runnable /etc/hotplug/scsi_device.agent is installed
> 2.4.19 didn't have this problem. 2.4.19 automatically updated /etc/fstab.

Try getting the latest update of the hotplug scripts from
http://linux-hotplug.sf.net/ if you don't like seeing this in your log
files. It's a very harmless message.

thanks,

greg k-h

2003-08-16 13:22:37

by Norman Diamond

[permalink] [raw]
Subject: Re: Trying to run 2.6.0-test3

"Felipe Alfaro Solana" <[email protected]> replied to me:

> > > > Guess why I compiled it without ACPI support and with APM support.
> > > > [...]. Linux doesn't panic when its default ACPI takes over, but it
> > > > does prevent APM from working.)
> > >
> > > If you turn ACPI on, you won't need APM support.
> >
> > WRONG. ACPI DOESN'T WORK ON THE MACHINE I'M DOING THIS ON. DID
> > YOU TRY READING WHAT YOU QUOTED THERE?
[though I deleted it this time]
>
> Yes, I tried reading. You said Linux doesn't panic while using ACPI, so
> I supposed ACPI just worked but the problem was you wanted APM support.

Since Linux doesn't panic, ACPI turns into a no-op. Yes that's better than
Windows 2000 blue screening, but no it's not as good as APM support.

The present status of APM support is that the command "apm -s" still
suspends the laptop but the hotkey Fn+F10 gets ignored. In kernel 2.4.19
the hotkey was interpreted as a more power-hungry variation of standby (same
as "apm -S") so I hacked 2.4.19 to make it do suspend, but in kernel
2.6.0-test3 the hotkey doesn't even reach the apm driver. But this is a
separate issue from the one that caused you to think I should turn on ACPI's
failures.

> > > To be sincere, I don't know exactly why "pci=usepirqmask" needs to be
> > > used. I'm no hardware expert. But I know that I needed it when I
> > > wasn't using ACPI.
> >
> > Hmm. Then some dependency seems to be broken in kernel compilation.
> > When ACPI is not compiled in, it should know that the effect of
> > "pci=usepirqmask" should be compiled in (whatever that effect is).
>
> It's not a problem with dependencies. On ACPI-enabled kernels, you using
> ACPI routing.

Then it *is* a problem with dependencies. In kernel 2.6.0-test1 through
test3, I set all configuration options myself, instead of inheriting
anything from SuSE's 2.4.19 defaults. I compiled 2.6.0 without ACPI. Since
this is not an ACPI-enabled kernel, no one should be expecting me to use
ACPI routing.

> If you boot using "acpi=off", then you're using standard
> PCI routing and that, in turn, on same machines, it warns you to use
> "pci=usepirqmask".

But this combination of facts remains very curious:
In 2.4.19, where the kernel is still ACPI-enabled, where it is absolutely
necessary for me to use "acpi=off apm=on", it doesn't warn to use
"pci=usepirqmask".
In 2.6.0-test3, where the kernel is not ACPI-enabled (because I
config'ed it not to be), where it is redundant for me to use
"acpi=off apm=on", it is warning me to use "pci=usepirqmask".

This combination of facts is exactly the opposite of what you think it
should be. I'd say it looks like a bug in a dependency condition.

2003-08-16 13:22:31

by Norman Diamond

[permalink] [raw]
Subject: Re: Trying to run 2.6.0-test3

"Greg KH" <[email protected]> replied to me:

> > 2. When I attach a USB hard disk drive, 2.6.0 drivers log:
> > [...] no runnable /etc/hotplug/scsi_device.agent is installed
> > 2.4.19 didn't have this problem. 2.4.19 automatically updated
> > /etc/fstab.
>
> Try getting the latest update of the hotplug scripts from
> http://linux-hotplug.sf.net/ if you don't like seeing this in your log
> files. It's a very harmless message.

Once I know that the message is harmless, I don't need to worry about it.
But it would be nice if 2.6.0 would automatically update /etc/fstab the way
2.4.19 did. I assume that the new hotplug scripts will do that.

By the way, since we had some discussion of USB keyboards, I want to mention
that it is separate (so far) from the keyboard problems that I've reported
for 2.6.0-test1 through test3. So far I've only tested 2.6.0 on machines
with PS/2 keyboards (or emulations thereof). Considering the patch for
test1 which still hasn't made it into test3, maybe I ought to try connecting
a USB keyboard to one machine soon. Meanwhile I don't dare to even try
2.6.0 on a machine which depends on its USB keyboard (no PS/2 ports and
cessation of BIOS emulation when the OS boots).

2003-08-17 02:23:48

by Norman Diamond

[permalink] [raw]
Subject: Re: Trying to run 2.6.0-test3

"Mikael Pettersson" <[email protected]> replied to me:

> > [...] cardmgr doesn't start automatically in 2.6.0-test3.
>
> You probably need fixes to /etc/rc.d/init.d/pcmcia (drop module ".o"
> extensions from modprobe commands) and /etc/hotplug/net.agent (case
> $ACTION needs "add|register" not just "register").

Very close. /etc/rc.d/pcmcia is a SuSE script which "knows" that some
things weren't compiled into the kernel, it "knows" that the things were
modules. I didn't want to spend a few hours learning enough bash
programming to make the thing deal with all the possibilities of 2.4 and
2.6. Instead, I changed the .config to compile PCMCIA stuff as modules (and
also some network stuff and serial stuff that are giving problems), and
munged some ".o" to ".ko" extensions. Fortunately it isn't completely
necessary to delete the ".ko" extension.

/etc/hotplug/net.agent was editable exactly as you said.

When I have time I'll hunt for the SuSE scripts hotplug scripts for serial
stuff that is still giving problems.

2003-08-17 08:00:30

by Norman Diamond

[permalink] [raw]
Subject: Re: Trying to run 2.6.0-test3

Additional defect:

Support for the 16-bit IDE card KME KXL-C005 was added to the external
PCMCIA package a couple of years ago, and has finally been copied into the
kernel for 2.6.0. But it doesn't work in 2.6.0-test3.

cardctl ident gets it right:
Socket 1:
product info: "KME", "KXLC005", "00"
manfid: 0x0032, 0x0704
function: 8 (SCSI)
Of course the card is lying, it's really IDE.

2.6.0-test3 gets it wrong:
Aug 17 16:39:01 diamondpana cardmgr[540]: socket 1: KME KXLC005 CD-ROM
Aug 17 16:39:01 diamondpana cardmgr[540]: executing: 'modprobe -v ide_cs'
Aug 17 16:39:02 diamondpana cardmgr[540]: + insmod /lib/modules/2.6.0-test3/kernel/drivers/ide/legacy/ide-cs.ko
Aug 17 16:39:02 diamondpana cardmgr[540]: bind 'ide_cs' to socket 1 failed: Invalid argument

lsmod agrees that the module name is different from the filename:
Module Size Used by
ide_cs 6400 0

So the problem indeed seems to be an invalid argument in binding.

2003-08-17 10:34:51

by Norman Diamond

[permalink] [raw]
Subject: Re: Trying to run 2.6.0-test3

I knew it. I knew it.

With a PS/2 keyboard, on a text console, only the yen and pipe key is
broken. But with a USB keyboard, on a text console, the backslash and
underbar key is broken right along with the yen and pipe key. These famous
two keys produce no input.

A few years ago I tried to persuade Andries Brouwer or Vojtech Pavlik to
accept a USB keyboard but they refused. The patch which I sent to Vojtech
Pavlik was ignored and these two keys continued not to work (except on my
machine). Finally Mike Fabian accepted a gift of a USB keyboard and this
defect in Linux got fixed. But only for somewhere around the last half of
the 2.4 releases, not for 2.6.

What will it take this time?

2003-08-17 13:33:26

by Russell King

[permalink] [raw]
Subject: Re: Trying to run 2.6.0-test3

On Sun, Aug 17, 2003 at 04:59:34PM +0900, Norman Diamond wrote:
> 2.6.0-test3 gets it wrong:
> Aug 17 16:39:01 diamondpana cardmgr[540]: socket 1: KME KXLC005 CD-ROM
> Aug 17 16:39:01 diamondpana cardmgr[540]: executing: 'modprobe -v ide_cs'
> Aug 17 16:39:02 diamondpana cardmgr[540]: + insmod /lib/modules/2.6.0-test3/kernel/drivers/ide/legacy/ide-cs.ko
> Aug 17 16:39:02 diamondpana cardmgr[540]: bind 'ide_cs' to socket 1 failed: Invalid argument

Not quite. This is an old problem dating back several years to early 2.4
times. Back in the dark old days, ide-cs used to use the name "ide_cs" to
bind the driver to the device. It now uses "ide-cs" for binding.

Make sure that "ide_cs" isn't mentioned in /etc/pcmcia/config - if it
is, change it to "ide-cs".

--
Russell King ([email protected]) The developer of ARM Linux
http://www.arm.linux.org.uk/personal/aboutme.html

2003-08-17 17:25:40

by Alan

[permalink] [raw]
Subject: Re: Trying to run 2.6.0-test3

On Sul, 2003-08-17 at 10:57, Norman Diamond wrote:
> accept a USB keyboard but they refused. The patch which I sent to Vojtech
> Pavlik was ignored and these two keys continued not to work (except on my
> machine). Finally Mike Fabian accepted a gift of a USB keyboard and this
> defect in Linux got fixed. But only for somewhere around the last half of
> the 2.4 releases, not for 2.6.
>
> What will it take this time?

Posting the patch with any luck ?

2003-08-18 10:55:08

by Norman Diamond

[permalink] [raw]
Subject: Re: Trying to run 2.6.0-test3

"Russell King" <[email protected]> replied to me:

> > 2.6.0-test3 gets it wrong:
> > Aug 17 16:39:01 diamondpana cardmgr[540]: socket 1: KME KXLC005 CD-ROM
> > Aug 17 16:39:01 diamondpana cardmgr[540]: executing: 'modprobe -v ide_cs'
> > Aug 17 16:39:02 diamondpana cardmgr[540]: + insmod /lib/modules/2.6.0-test3/kernel/drivers/ide/legacy/ide-cs.ko
> > Aug 17 16:39:02 diamondpana cardmgr[540]: bind 'ide_cs' to socket 1 failed: Invalid argument
>
> Not quite. This is an old problem dating back several years to early 2.4
> times. Back in the dark old days, ide-cs used to use the name "ide_cs" to
> bind the driver to the device. It now uses "ide-cs" for binding.

Yes the problem is reminiscent of early 2.4 days, but how is it that 2.4.19
doesn't have the problem and 2.6.0-test3 does have the problem on the same
machine?

> Make sure that "ide_cs" isn't mentioned in /etc/pcmcia/config - if it
> is, change it to "ide-cs".

I will hope to have time to check this next weekend, but I'm sure the 2.4
name ide-cs is used. I was startled to see 2.6.0-test3 load a module named
ide_cs instead of ide-cs, but there it was. It was listed that way by
lsmod, but failing in the bind as shown above. Notice that modprobe used
the same name as lsmod showed.

2003-08-18 11:00:40

by Vojtech Pavlik

[permalink] [raw]
Subject: Re: Trying to run 2.6.0-test3

On Mon, Aug 18, 2003 at 07:35:09PM +0900, Norman Diamond wrote:
> "Alan Cox" <[email protected]> replied to me.
>
> > > accept a USB keyboard but they refused. The patch which I sent to Vojtech
> > > Pavlik was ignored and these two keys continued not to work (except on my
> > > machine). Finally Mike Fabian accepted a gift of a USB keyboard and this
> > > defect in Linux got fixed. But only for somewhere around the last half of
> > > the 2.4 releases, not for 2.6.
> > > What will it take this time?
> >
> > Posting the patch with any luck ?
>
> Hirofumi Ogawa posted a patch for the yen-sign pipe key on 2003.07.23 for
> test1 but his patch still didn't get into test3.

It will get into 2.6 sooner or later.

> On a PS/2 keyboard that
> seems to be the only key with any problem.
>
> Yesterday when I finally tried a USB keyboard and found that the backslash
> underbar has the same problem, maybe I was the first person to even try a
> Japanese USB keyboard in 2.6, and maybe no one at all tried some number of
> 2.5 series kernels.

If you can find out what input event the key generates (using the evtest
program attached), then please tell me, and I'll fix in the same way as
the yen key was fixed.

> As mentioned, usually I can only spend one day a week
> testing 2.6. I'll try to spend one day next weekend trying to figure out
> the new necessary patch. If I succeed, but if it gets ignored again, I'll
> probably rejoin the set of users who never have time to test.
>
> I really do think that if Andries Brouwer or Vojtech Pavlik would accept a
> gift of a USB keyboard then this kind of bug would be avoided a lot earlier.

Well, I have a bunch of USB keyboards, all working perfectly. None of
them is a Japanese keyboard, though.

--
Vojtech Pavlik
SuSE Labs, SuSE CR

2003-08-18 10:55:05

by Norman Diamond

[permalink] [raw]
Subject: Re: Trying to run 2.6.0-test3

"Alan Cox" <[email protected]> replied to me.

> > accept a USB keyboard but they refused. The patch which I sent to Vojtech
> > Pavlik was ignored and these two keys continued not to work (except on my
> > machine). Finally Mike Fabian accepted a gift of a USB keyboard and this
> > defect in Linux got fixed. But only for somewhere around the last half of
> > the 2.4 releases, not for 2.6.
> > What will it take this time?
>
> Posting the patch with any luck ?

Hirofumi Ogawa posted a patch for the yen-sign pipe key on 2003.07.23 for
test1 but his patch still didn't get into test3. On a PS/2 keyboard that
seems to be the only key with any problem.

Yesterday when I finally tried a USB keyboard and found that the backslash
underbar has the same problem, maybe I was the first person to even try a
Japanese USB keyboard in 2.6, and maybe no one at all tried some number of
2.5 series kernels. As mentioned, usually I can only spend one day a week
testing 2.6. I'll try to spend one day next weekend trying to figure out
the new necessary patch. If I succeed, but if it gets ignored again, I'll
probably rejoin the set of users who never have time to test.

I really do think that if Andries Brouwer or Vojtech Pavlik would accept a
gift of a USB keyboard then this kind of bug would be avoided a lot earlier.

2003-08-18 11:21:10

by Russell King

[permalink] [raw]
Subject: Re: Trying to run 2.6.0-test3

On Mon, Aug 18, 2003 at 07:41:34PM +0900, Norman Diamond wrote:
> "Russell King" <[email protected]> replied to me:
> > Not quite. This is an old problem dating back several years to early 2.4
> > times. Back in the dark old days, ide-cs used to use the name "ide_cs" to
> > bind the driver to the device. It now uses "ide-cs" for binding.
>
> Yes the problem is reminiscent of early 2.4 days, but how is it that 2.4.19
> doesn't have the problem and 2.6.0-test3 does have the problem on the same
> machine?

In 2.6.0-test3:

linux/drivers/ide/legacy/ide-cs.c:

static dev_info_t dev_info = "ide-cs";

In 2.4.21:

linux/drivers/ide/legacy/ide-cs.c

static dev_info_t dev_info = "ide-cs";

The pcmcia-cs package may have ide-cs called ide_cs though.

> > Make sure that "ide_cs" isn't mentioned in /etc/pcmcia/config - if it
> > is, change it to "ide-cs".
>
> I will hope to have time to check this next weekend, but I'm sure the 2.4
> name ide-cs is used. I was startled to see 2.6.0-test3 load a module named
> ide_cs instead of ide-cs, but there it was. It was listed that way by
> lsmod, but failing in the bind as shown above. Notice that modprobe used
> the same name as lsmod showed.

The output of lsmod can be confusing. '-' is replaced by '_' in the
module name. This doesn't affect module binding, loading or unloading.
You can still rmmod ide-cs even if it's called ide_cs.

In either case, this is what should be in /etc/pcmcia/config:

device "ide-cs"
class "ide" module "ide-cs"

pcmcia-cs 3.1.24 ships with:

device "ide_cs"
class "ide" module "ide_cs"

which is wrong for the kernels ide-cs implementation.

The first line contains the device name, and is also the name used to
bind the driver to the device. (this is referenced by the "bind"
statements in the "card" sections.) The second line contains the
module name(s) to load.

--
Russell King ([email protected]) The developer of ARM Linux
http://www.arm.linux.org.uk/personal/aboutme.html

2003-08-18 11:25:51

by John Bradford

[permalink] [raw]
Subject: Re: Trying to run 2.6.0-test3

> > I really do think that if Andries Brouwer or Vojtech Pavlik would accept a
> > gift of a USB keyboard then this kind of bug would be avoided a lot earlier.
>
> Well, I have a bunch of USB keyboards, all working perfectly. None of
> them is a Japanese keyboard, though.

I strongly recommend Japanese keyboards to all developers.

My touch typing speed for code is probably about 10% faster using a
Japanese keyboard, compared with a U.K. keyboard, and no slower for
English text.

John.

2003-08-23 12:35:21

by Norman Diamond

[permalink] [raw]
Subject: Re: Trying to run 2.6.0-test3

"Vojtech Pavlik" <[email protected]> replied to me with the program evtest.c.

> > Hirofumi Ogawa posted a patch for the yen-sign pipe key on 2003.07.23
> > for test1 but his patch still didn't get into test3.
>
> It will get into 2.6 sooner or later.
>
> > On a PS/2 keyboard that
> > seems to be the only key with any problem.
> >
> > Yesterday when I finally tried a USB keyboard and found that the
> > backslash underbar has the same problem, maybe I was the first person to
> > even try a Japanese USB keyboard in 2.6, and maybe no one at all tried
> > some number of 2.5 series kernels.
>
> If you can find out what input event the key generates (using the evtest
> program attached), then please tell me, and I'll fix in the same way as
> the yen key was fixed.
>
> > As mentioned, usually I can only spend one day a week
> > testing 2.6.

For this test, I used a laptop with a built-in emulated PS/2 keyboard and
mouse, and plugged in a USB keyboard. I do not dare yet to try 2.6.0-testN
on a small-size desktop which depends entirely on USB. The small-size
desktop provides BIOS emulation of a PS/2 keyboard from boot until the OS
detects USB devices. It does not provide any emulation of an i8042 chip.
(Windows NT4 blue-screens if I forget to disable its i8042 driver.) It
does not have PS/2 ports.

Back to the laptop used in this test. Since the built-in emulated PS/2
keyboard had some "?" events, I ran evtest on all devices. Sorry this
is still 2.6.0-test3. I wanted to finish this test before experimenting
with 2.6.0-test4. I already patched the 2.6.0-test3 keyboard map, so the
yen-sign pipe key produces input in both X11 and text consoles. The USB
problem is with the yen-sign or-bar key, which seems to produce events
properly, and which produces correct input under X11, but produces no
input to a plain text console. I wonder what needs patching for this.

In a plain text console, in the built-in emulated PS/2 keyboard, both the
yen-sign or-bar and backslash underbar keys are working because of the
patch. But in the USB keyboard, the yen-sign or-bar key is working while
the backslash underbar key yields no input.

First, here are results of running evtest under X11.

The emulated PS/2 mouse looks fine.

ndiamond@diamondpana:~/linux-pavlik-evtest> evtest /dev/input/event0
Input driver version is 1.0.0
Input device ID: bus 0x11 vendor 0x2 product 0x1 version 0x29
Input device name: "PS/2 Logitech Mouse"
Supported events:
Event type 0 (Reset)
Event code 0 (Reset)
Event code 1 (Key)
Event code 2 (Relative)
Event type 1 (Key)
Event code 272 (LeftBtn)
Event code 273 (RightBtn)
Event code 275 (SideBtn)
Event type 2 (Relative)
Event code 0 (X)
Event code 1 (Y)
Testing ... (interrupt to exit)
Event: time 1061637735.598896, type 2 (Relative), code 1 (Y), value 1
Event: time 1061637735.598936, type 0 (Reset), code 0 (Reset), value 0
[...]
Event: time 1061637735.634618, type 2 (Relative), code 0 (X), value 1
Event: time 1061637735.634657, type 0 (Reset), code 0 (Reset), value 0
[...]
Event: time 1061637737.663347, type 1 (Key), code 273 (RightBtn), value 1
Event: time 1061637737.663390, type 0 (Reset), code 0 (Reset), value 0
Event: time 1061637737.806503, type 1 (Key), code 273 (RightBtn), value 0
Event: time 1061637737.806543, type 0 (Reset), code 0 (Reset), value 0
Event: time 1061637739.053020, type 1 (Key), code 272 (LeftBtn), value 1
Event: time 1061637739.053068, type 0 (Reset), code 0 (Reset), value 0
Event: time 1061637739.144967, type 1 (Key), code 272 (LeftBtn), value 0
Event: time 1061637739.145016, type 0 (Reset), code 0 (Reset), value 0

Here is the emulated PS/2 keyboard. I did not try all keys. In fact I
cannot try a number of keys which appear in the list but don't exist.
I did try the ones which are frequently mishandled, yen-sign or-bar and
backslash underbar.

ndiamond@diamondpana:~/linux-pavlik-evtest> evtest /dev/input/event1
Input driver version is 1.0.0
Input device ID: bus 0x11 vendor 0x1 product 0x2 version 0xab02
Input device name: "AT Set 2 keyboard"
Supported events:
Event type 0 (Reset)
Event code 0 (Reset)
Event code 1 (Key)
Event code 17 (LED)
Event code 20 (Repeat)
Event type 1 (Key)
Event code 1 (Esc)
Event code 2 (1)
Event code 3 (2)
Event code 4 (3)
Event code 5 (4)
Event code 6 (5)
Event code 7 (6)
Event code 8 (7)
Event code 9 (8)
Event code 10 (9)
Event code 11 (0)
Event code 12 (Minus)
Event code 13 (Equal)
Event code 14 (Backspace)
Event code 15 (Tab)
Event code 16 (Q)
Event code 17 (W)
Event code 18 (E)
Event code 19 (R)
Event code 20 (T)
Event code 21 (Y)
Event code 22 (U)
Event code 23 (I)
Event code 24 (O)
Event code 25 (P)
Event code 26 (LeftBrace)
Event code 27 (RightBrace)
Event code 28 (Enter)
Event code 29 (LeftControl)
Event code 30 (A)
Event code 31 (S)
Event code 32 (D)
Event code 33 (F)
Event code 34 (G)
Event code 35 (H)
Event code 36 (J)
Event code 37 (K)
Event code 38 (L)
Event code 39 (Semicolon)
Event code 40 (Apostrophe)
Event code 41 (Grave)
Event code 42 (LeftShift)
Event code 43 (BackSlash)
Event code 44 (Z)
Event code 45 (X)
Event code 46 (C)
Event code 47 (V)
Event code 48 (B)
Event code 49 (N)
Event code 50 (M)
Event code 51 (Comma)
Event code 52 (Dot)
Event code 53 (Slash)
Event code 54 (RightShift)
Event code 55 (KPAsterisk)
Event code 56 (LeftAlt)
Event code 57 (Space)
Event code 58 (CapsLock)
Event code 59 (F1)
Event code 60 (F2)
Event code 61 (F3)
Event code 62 (F4)
Event code 63 (F5)
Event code 64 (F6)
Event code 65 (F7)
Event code 66 (F8)
Event code 67 (F9)
Event code 68 (F10)
Event code 69 (NumLock)
Event code 70 (ScrollLock)
Event code 71 (KP7)
Event code 72 (KP8)
Event code 73 (KP9)
Event code 74 (KPMinus)
Event code 75 (KP4)
Event code 76 (KP5)
Event code 77 (KP6)
Event code 78 (KPPlus)
Event code 79 (KP1)
Event code 80 (KP2)
Event code 81 (KP3)
Event code 82 (KP0)
Event code 83 (KPDot)
Event code 85 (F13)
Event code 86 (102nd)
Event code 87 (F11)
Event code 88 (F12)
Event code 89 (F14)
Event code 90 (F15)
Event code 91 (F16)
Event code 92 (F17)
Event code 93 (F18)
Event code 94 (F19)
Event code 95 (F20)
Event code 96 (KPEnter)
Event code 97 (RightCtrl)
Event code 98 (KPSlash)
Event code 99 (SysRq)
Event code 100 (RightAlt)
Event code 102 (Home)
Event code 103 (Up)
Event code 104 (PageUp)
Event code 105 (Left)
Event code 106 (Right)
Event code 107 (End)
Event code 108 (Down)
Event code 109 (PageDown)
Event code 110 (Insert)
Event code 111 (Delete)
Event code 112 (Macro)
Event code 113 (Mute)
Event code 114 (VolumeDown)
Event code 115 (VolumeUp)
Event code 116 (Power)
Event code 118 (KPPlusMinus)
Event code 119 (Pause)
Event code 120 (F21)
Event code 121 (F22)
Event code 122 (F23)
Event code 123 (F24)
Event code 125 (LeftMeta)
Event code 126 (RightMeta)
Event code 127 (Compose)
Event code 128 (Stop)
Event code 133 (Copy)
Event code 137 (Cut)
Event code 138 (Help)
Event code 140 (Calc)
Event code 142 (Sleep)
Event code 143 (WakeUp)
Event code 144 (File)
Event code 147 (X-fer)
Event code 150 (WWW)
Event code 151 (MSDOS)
Event code 152 (Coffee)
Event code 153 (Direction)
Event code 155 (Mail)
Event code 156 (Bookmarks)
Event code 157 (Computer)
Event code 158 (Back)
Event code 159 (Forward)
Event code 160 (CloseCD)
Event code 163 (NextSong)
Event code 164 (PlayPause)
Event code 165 (PreviousSong)
Event code 166 (StopCD)
Event code 167 (Record)
Event code 168 (Rewind)
Event code 173 (Refresh)
Event code 174 (Exit)
Event code 182 (International2)
Event code 183 (International3)
Event code 217 (?)
Event code 226 (?)
Event type 17 (LED)
Event code 0 (NumLock)
Event code 1 (CapsLock)
Event code 2 (ScrollLock)
Event type 20 (Repeat)

[Here is yen-sign or-bar:]
Event: time 1061638812.315540, type 1 (Key), code 183 (International3), value 1
Event: time 1061638812.315548, type 0 (Reset), code 0 (Reset), value 0
\Event: time 1061638812.467647, type 1 (Key), code 183 (International3), value 0
Event: time 1061638812.467655, type 0 (Reset), code 0 (Reset), value 0
Event: time 1061638813.909474, type 1 (Key), code 42 (LeftShift), value 1
Event: time 1061638813.909482, type 0 (Reset), code 0 (Reset), value 0
Event: time 1061638814.159276, type 1 (Key), code 42 (LeftShift), value 2
Event: time 1061638814.159286, type 0 (Reset), code 0 (Reset), value 0
Event: time 1061638814.189239, type 1 (Key), code 42 (LeftShift), value 2
Event: time 1061638814.189246, type 0 (Reset), code 0 (Reset), value 0
Event: time 1061638814.209031, type 1 (Key), code 183 (International3), value 1
Event: time 1061638814.209039, type 0 (Reset), code 0 (Reset), value 0
|Event: time 1061638814.361093, type 1 (Key), code 183 (International3), value 0
Event: time 1061638814.361102, type 0 (Reset), code 0 (Reset), value 0
Event: time 1061638814.706204, type 1 (Key), code 42 (LeftShift), value 0
Event: time 1061638814.706211, type 0 (Reset), code 0 (Reset), value 0

[Here is backslash underbar:]
Event: time 1061638897.294886, type 1 (Key), code 89 (F14), value 1
Event: time 1061638897.294894, type 0 (Reset), code 0 (Reset), value 0
\Event: time 1061638897.416553, type 1 (Key), code 89 (F14), value 0
Event: time 1061638897.416560, type 0 (Reset), code 0 (Reset), value 0
Event: time 1061638898.462432, type 1 (Key), code 42 (LeftShift), value 1
Event: time 1061638898.462440, type 0 (Reset), code 0 (Reset), value 0
Event: time 1061638898.650346, type 1 (Key), code 89 (F14), value 1
Event: time 1061638898.650354, type 0 (Reset), code 0 (Reset), value 0
_Event: time 1061638898.761825, type 1 (Key), code 89 (F14), value 0
Event: time 1061638898.761833, type 0 (Reset), code 0 (Reset), value 0
Event: time 1061638899.046039, type 1 (Key), code 42 (LeftShift), value 0
Event: time 1061638899.046051, type 0 (Reset), code 0 (Reset), value 0

Here is the external USB keyboard. I did not try all keys. In fact I
cannot try a number of keys which appear in the list but don't exist.
(Though the keyboard which I sent to Mike Fabian might really have some
of these, among its ton of funny extra buttons.) I did try the ones
which are frequently mishandled, yen-sign or-bar and backslash underbar.

ndiamond@diamondpana:~/linux-pavlik-evtest> evtest /dev/input/event2
Input driver version is 1.0.0
Input device ID: bus 0x3 vendor 0x409 product 0x14 version 0x100
Input device name: "NEC 109 JPN USB KBD/M"
Supported events:
Event type 0 (Reset)
Event code 0 (Reset)
Event code 1 (Key)
Event code 17 (LED)
Event code 20 (Repeat)
Event type 1 (Key)
Event code 1 (Esc)
Event code 2 (1)
Event code 3 (2)
Event code 4 (3)
Event code 5 (4)
Event code 6 (5)
Event code 7 (6)
Event code 8 (7)
Event code 9 (8)
Event code 10 (9)
Event code 11 (0)
Event code 12 (Minus)
Event code 13 (Equal)
Event code 14 (Backspace)
Event code 15 (Tab)
Event code 16 (Q)
Event code 17 (W)
Event code 18 (E)
Event code 19 (R)
Event code 20 (T)
Event code 21 (Y)
Event code 22 (U)
Event code 23 (I)
Event code 24 (O)
Event code 25 (P)
Event code 26 (LeftBrace)
Event code 27 (RightBrace)
Event code 28 (Enter)
Event code 29 (LeftControl)
Event code 30 (A)
Event code 31 (S)
Event code 32 (D)
Event code 33 (F)
Event code 34 (G)
Event code 35 (H)
Event code 36 (J)
Event code 37 (K)
Event code 38 (L)
Event code 39 (Semicolon)
Event code 40 (Apostrophe)
Event code 41 (Grave)
Event code 42 (LeftShift)
Event code 43 (BackSlash)
Event code 44 (Z)
Event code 45 (X)
Event code 46 (C)
Event code 47 (V)
Event code 48 (B)
Event code 49 (N)
Event code 50 (M)
Event code 51 (Comma)
Event code 52 (Dot)
Event code 53 (Slash)
Event code 54 (RightShift)
Event code 55 (KPAsterisk)
Event code 56 (LeftAlt)
Event code 57 (Space)
Event code 58 (CapsLock)
Event code 59 (F1)
Event code 60 (F2)
Event code 61 (F3)
Event code 62 (F4)
Event code 63 (F5)
Event code 64 (F6)
Event code 65 (F7)
Event code 66 (F8)
Event code 67 (F9)
Event code 68 (F10)
Event code 69 (NumLock)
Event code 70 (ScrollLock)
Event code 71 (KP7)
Event code 72 (KP8)
Event code 73 (KP9)
Event code 74 (KPMinus)
Event code 75 (KP4)
Event code 76 (KP5)
Event code 77 (KP6)
Event code 78 (KPPlus)
Event code 79 (KP1)
Event code 80 (KP2)
Event code 81 (KP3)
Event code 82 (KP0)
Event code 83 (KPDot)
Event code 84 (103rd)
Event code 85 (F13)
Event code 86 (102nd)
Event code 87 (F11)
Event code 88 (F12)
Event code 89 (F14)
Event code 90 (F15)
Event code 91 (F16)
Event code 92 (F17)
Event code 93 (F18)
Event code 94 (F19)
Event code 95 (F20)
Event code 96 (KPEnter)
Event code 97 (RightCtrl)
Event code 98 (KPSlash)
Event code 99 (SysRq)
Event code 100 (RightAlt)
Event code 102 (Home)
Event code 103 (Up)
Event code 104 (PageUp)
Event code 105 (Left)
Event code 106 (Right)
Event code 107 (End)
Event code 108 (Down)
Event code 109 (PageDown)
Event code 110 (Insert)
Event code 111 (Delete)
Event code 113 (Mute)
Event code 114 (VolumeDown)
Event code 115 (VolumeUp)
Event code 116 (Power)
Event code 117 (KPEqual)
Event code 119 (Pause)
Event code 120 (F21)
Event code 121 (F22)
Event code 122 (F23)
Event code 123 (F24)
Event code 124 (KPComma)
Event code 125 (LeftMeta)
Event code 126 (RightMeta)
Event code 127 (Compose)
Event code 128 (Stop)
Event code 129 (Again)
Event code 130 (Props)
Event code 131 (Undo)
Event code 132 (Front)
Event code 133 (Copy)
Event code 134 (Open)
Event code 135 (Paste)
Event code 136 (Find)
Event code 137 (Cut)
Event code 138 (Help)
Event code 140 (Calc)
Event code 142 (Sleep)
Event code 150 (WWW)
Event code 152 (Coffee)
Event code 158 (Back)
Event code 159 (Forward)
Event code 161 (EjectCD)
Event code 163 (NextSong)
Event code 164 (PlayPause)
Event code 165 (PreviousSong)
Event code 166 (StopCD)
Event code 173 (Refresh)
Event code 176 (Edit)
Event code 177 (ScrollUp)
Event code 178 (ScrollDown)
Event code 181 (International1)
Event code 182 (International2)
Event code 183 (International3)
Event code 184 (International4)
Event code 185 (International5)
Event code 186 (International6)
Event code 187 (International7)
Event code 188 (International8)
Event code 189 (International9)
Event code 190 (Language1)
Event code 191 (Language2)
Event code 192 (Language3)
Event code 193 (Language4)
Event code 194 (Language5)
Event code 195 (Language6)
Event code 196 (Language7)
Event code 197 (Language8)
Event code 198 (Language9)
Event code 240 (?)
Event type 17 (LED)
Event code 0 (NumLock)
Event code 1 (CapsLock)
Event code 2 (ScrollLock)
Event type 20 (Repeat)
Testing ... (interrupt to exit)

[Here is yen-sign or-bar:]
Event: time 1061639164.238590, type 1 (Key), code 183 (International3), value 1
Event: time 1061639164.238602, type 0 (Reset), code 0 (Reset), value 0
Event: time 1061639164.318573, type 1 (Key), code 183 (International3), value 0
Event: time 1061639164.318584, type 0 (Reset), code 0 (Reset), value 0
Event: time 1061639168.837822, type 1 (Key), code 54 (RightShift), value 1
Event: time 1061639168.837838, type 0 (Reset), code 0 (Reset), value 0
Event: time 1061639169.087280, type 1 (Key), code 54 (RightShift), value 2
Event: time 1061639169.087291, type 0 (Reset), code 0 (Reset), value 0
Event: time 1061639169.117272, type 1 (Key), code 54 (RightShift), value 2
Event: time 1061639169.117279, type 0 (Reset), code 0 (Reset), value 0
Event: time 1061639169.125757, type 1 (Key), code 183 (International3), value 1
Event: time 1061639169.125768, type 0 (Reset), code 0 (Reset), value 0
Event: time 1061639169.237759, type 1 (Key), code 183 (International3), value 0
Event: time 1061639169.237772, type 0 (Reset), code 0 (Reset), value 0
Event: time 1061639169.477720, type 1 (Key), code 54 (RightShift), value 0
Event: time 1061639169.477736, type 0 (Reset), code 0 (Reset), value 0

[Here is backslash underbar:]
Event: time 1061639230.403570, type 1 (Key), code 181 (International1), value 1
Event: time 1061639230.403585, type 0 (Reset), code 0 (Reset), value 0
Event: time 1061639230.499543, type 1 (Key), code 181 (International1), value 0
Event: time 1061639230.499556, type 0 (Reset), code 0 (Reset), value 0
Event: time 1061639231.275459, type 1 (Key), code 54 (RightShift), value 1
Event: time 1061639231.275483, type 0 (Reset), code 0 (Reset), value 0
Event: time 1061639231.419385, type 1 (Key), code 181 (International1), value 1
Event: time 1061639231.419396, type 0 (Reset), code 0 (Reset), value 0
Event: time 1061639231.491366, type 1 (Key), code 181 (International1), value 0
Event: time 1061639231.491379, type 0 (Reset), code 0 (Reset), value 0
Event: time 1061639231.691448, type 1 (Key), code 54 (RightShift), value 0
Event: time 1061639231.691467, type 0 (Reset), code 0 (Reset), value 0

Next I switched to a text-mode console and tested the same two keyboards.
The key codes seem to be the same.

2003-08-25 14:24:23

by Vojtech Pavlik

[permalink] [raw]
Subject: Re: Trying to run 2.6.0-test3

On Sat, Aug 23, 2003 at 09:29:53PM +0900, Norman Diamond wrote:
> "Vojtech Pavlik" <[email protected]> replied to me with the program evtest.c.
>
> > > Hirofumi Ogawa posted a patch for the yen-sign pipe key on 2003.07.23
> > > for test1 but his patch still didn't get into test3.
> >
> > It will get into 2.6 sooner or later.
> >
> > > On a PS/2 keyboard that
> > > seems to be the only key with any problem.
> > >
> > > Yesterday when I finally tried a USB keyboard and found that the
> > > backslash underbar has the same problem, maybe I was the first person to
> > > even try a Japanese USB keyboard in 2.6, and maybe no one at all tried
> > > some number of 2.5 series kernels.
> >
> > If you can find out what input event the key generates (using the evtest
> > program attached), then please tell me, and I'll fix in the same way as
> > the yen key was fixed.
> >
> > > As mentioned, usually I can only spend one day a week
> > > testing 2.6.
>
> For this test, I used a laptop with a built-in emulated PS/2 keyboard and
> mouse, and plugged in a USB keyboard. I do not dare yet to try 2.6.0-testN
> on a small-size desktop which depends entirely on USB. The small-size
> desktop provides BIOS emulation of a PS/2 keyboard from boot until the OS
> detects USB devices. It does not provide any emulation of an i8042 chip.
> (Windows NT4 blue-screens if I forget to disable its i8042 driver.) It
> does not have PS/2 ports.

Don't fear, go ahead, 2.6 should work fine without an i8042.

> Back to the laptop used in this test. Since the built-in emulated PS/2
> keyboard had some "?" events, I ran evtest on all devices. Sorry this
> is still 2.6.0-test3. I wanted to finish this test before experimenting
> with 2.6.0-test4. I already patched the 2.6.0-test3 keyboard map, so the
> yen-sign pipe key produces input in both X11 and text consoles. The USB
> problem is with the yen-sign or-bar key, which seems to produce events
> properly, and which produces correct input under X11, but produces no
> input to a plain text console. I wonder what needs patching for this.

The default keymap. Patch was already created by somebody, I'll
include it soon.

> In a plain text console, in the built-in emulated PS/2 keyboard, both the
> yen-sign or-bar and backslash underbar keys are working because of the
> patch. But in the USB keyboard, the yen-sign or-bar key is working while
> the backslash underbar key yields no input.
>
> First, here are results of running evtest under X11.
>
> The emulated PS/2 mouse looks fine.
>
> ndiamond@diamondpana:~/linux-pavlik-evtest> evtest /dev/input/event0
> Input driver version is 1.0.0
> Input device ID: bus 0x11 vendor 0x2 product 0x1 version 0x29
> Input device name: "PS/2 Logitech Mouse"

> Event: time 1061637735.598896, type 2 (Relative), code 1 (Y), value 1
> Event: time 1061637735.634618, type 2 (Relative), code 0 (X), value 1
> Event: time 1061637737.806503, type 1 (Key), code 273 (RightBtn), value 0
> Event: time 1061637739.053020, type 1 (Key), code 272 (LeftBtn), value 1
>
> Here is the emulated PS/2 keyboard. I did not try all keys. In fact I
> cannot try a number of keys which appear in the list but don't exist.
> I did try the ones which are frequently mishandled, yen-sign or-bar and
> backslash underbar.
>
> ndiamond@diamondpana:~/linux-pavlik-evtest> evtest /dev/input/event1
> Input driver version is 1.0.0
> Input device ID: bus 0x11 vendor 0x1 product 0x2 version 0xab02
> Input device name: "AT Set 2 keyboard"
> Event type 20 (Repeat)
>
> [Here is yen-sign or-bar:]
> Event: time 1061638812.315540, type 1 (Key), code 183 (International3), value 1

> [Here is backslash underbar:]
> Event: time 1061638897.294886, type 1 (Key), code 89 (F14), value 1

This one should probably be Intl1 and not F14. Scancode collission
between japanese keyboards and some old keyboard.

We will never be able to create a scancode->keycode table that matches
all existing keyboards. This is why these tables can be changed from
userspace via ioctls.

> Here is the external USB keyboard. I did not try all keys. In fact I
> cannot try a number of keys which appear in the list but don't exist.
> (Though the keyboard which I sent to Mike Fabian might really have some
> of these, among its ton of funny extra buttons.) I did try the ones
> which are frequently mishandled, yen-sign or-bar and backslash underbar.
>
> ndiamond@diamondpana:~/linux-pavlik-evtest> evtest /dev/input/event2
> Input driver version is 1.0.0
> Input device ID: bus 0x3 vendor 0x409 product 0x14 version 0x100
> Input device name: "NEC 109 JPN USB KBD/M"
>
> [Here is yen-sign or-bar:]
> Event: time 1061639164.238590, type 1 (Key), code 183 (International3), value 1
> [Here is backslash underbar:]
> Event: time 1061639230.403570, type 1 (Key), code 181 (International1), value 1
>
> Next I switched to a text-mode console and tested the same two keyboards.
> The key codes seem to be the same.

--
Vojtech Pavlik
SuSE Labs, SuSE CR