I expect this is a duplicate of bug 4997.

If you can confirm that adding rdloaddriver=pata_amd to the kernel options
fixes the problem, please close this bug report as a duplicate of 4997.

CC: (none) => davidwhodgins

I'll wait for this to appear:
https://www.mageia.org/pipermail/mageia-dev/2012-March/013410.html

Then I'll test again.

CC: (none) => filip.komar

(In reply to comment #1)
> If you can confirm that adding rdloaddriver=pata_amd to the kernel options
> fixes the problem, please close this bug report as a duplicate of 4997.

Forgot to add that this didn't work for me.

Currently those modules are loaded:
# lspcidrake -v
tg3             : Broadcom Corporation|NetXtreme BCM5751 Gigabit Ethernet PCI Express [NETWORK_ETHERNET] (vendor:14e4 device:1677 subv:1028 subd:0179) (rev: 01)
i2c_i801        : Intel Corporation|82801FB/FBM/FR/FW/FRW (ICH6 Family) SMBus Controller [SERIAL_SMBUS] (vendor:8086 device:266a subv:1028 subd:0179) (rev: 03)
ata_piix        : Intel Corporation|82801FB/FW (ICH6/ICH6W) SATA Controller [STORAGE_IDE] (vendor:8086 device:2651 subv:1028 subd:0179) (rev: 03)
ata_piix        : Intel Corporation|82801FB/FBM/FR/FW/FRW (ICH6 Family) IDE Controller [STORAGE_IDE] (vendor:8086 device:266f subv:1028 subd:0179) (rev: 03)
iTCO_wdt        : Intel Corporation|82801FB/FR (ICH6/ICH6R) LPC Interface Bridge [BRIDGE_ISA] (vendor:8086 device:2640) (rev: 03)
snd_intel8x0    : Intel Corporation|82801FB/FBM/FR/FW/FRW (ICH6 Family) AC'97 Audio Controller [MULTIMEDIA_AUDIO] (vendor:8086 device:266e subv:1028 subd:0179) (rev: 03)
unknown         : Intel Corporation|82801 PCI Bridge [BRIDGE_PCI] (vendor:8086 device:244e) (rev: d3)
ehci_hcd        : Intel Corporation|82801FB/FBM/FR/FW/FRW (ICH6 Family) USB2 EHCI Controller [SERIAL_USB] (vendor:8086 device:265c subv:1028 subd:0179) (rev: 03)
uhci_hcd        : Intel Corporation|82801FB/FBM/FR/FW/FRW (ICH6 Family) USB UHCI #4 [SERIAL_USB] (vendor:8086 device:265b subv:1028 subd:0179) (rev: 03)
uhci_hcd        : Intel Corporation|82801FB/FBM/FR/FW/FRW (ICH6 Family) USB UHCI #3 [SERIAL_USB] (vendor:8086 device:265a subv:1028 subd:0179) (rev: 03)
uhci_hcd        : Intel Corporation|82801FB/FBM/FR/FW/FRW (ICH6 Family) USB UHCI #2 [SERIAL_USB] (vendor:8086 device:2659 subv:1028 subd:0179) (rev: 03) 
uhci_hcd        : Intel Corporation|82801FB/FBM/FR/FW/FRW (ICH6 Family) USB UHCI #1 [SERIAL_USB] (vendor:8086 device:2658 subv:1028 subd:0179) (rev: 03) 
shpchp          : Intel Corporation|82801FB/FBM/FR/FW/FRW (ICH6 Family) PCI Express Port 2 [BRIDGE_PCI] (vendor:8086 device:2662) (rev: 03)              
shpchp          : Intel Corporation|82801FB/FBM/FR/FW/FRW (ICH6 Family) PCI Express Port 1 [BRIDGE_PCI] (vendor:8086 device:2660) (rev: 03)              
unknown         : Intel Corporation|82915G Integrated Graphics Controller [DISPLAY_OTHER] (vendor:8086 device:2782 subv:1028 subd:0179) (rev: 04)        
Card:Intel 810 and later: Intel Corporation|82915G/GV/910GL Integrated Graphics Controller [DISPLAY_VGA] (vendor:8086 device:2582 subv:1028 subd:0179) (rev: 04)                                                                                                                                                  
shpchp          : Intel Corporation|82915G/P/GV/GL/PL/910GL PCI Express Root Port [BRIDGE_PCI] (vendor:8086 device:2581) (rev: 04)                       
unknown         : Intel Corporation|82915G/P/GV/GL/PL/910GL Memory Controller Hub [BRIDGE_HOST] (vendor:8086 device:2580 subv:1028 subd:0179) (rev: 04)


But bug 4997 and this one have the same roots in my opinion and that is the reason for me to wait that new kernel from comment #2.

Unfortunately it seems that kernel-desktop-3.3.0-2.mga2 still use hda for PATA drives.

This bug is very serious. Some systems will be unbootable after upgrade. Luckily they are not very common these days because in that case this would be release critical bug. This should go to Errata in my opinion.

I just tried with kernel option "rdloaddriver=ata_piix". See bug 4997.

Although sfdisk now sees PATA HD as sda grub still claims same old hda thing:

# echo quit | grub --batch --device-map=/tmp/device.map
# cat tmp/device.map                                                                                                               
(fd0)   /dev/fd0
(hd0)   /dev/hda
(hd1)   /dev/sda


So this issue is also a grub bug?

My guess is that you need to reinstall grub, while booted with
the rdloaddriver option, or edit /boot/grub/device.map

(In reply to comment #6)
> My guess is that you need to reinstall grub, while booted with
> the rdloaddriver option, or edit /boot/grub/device.map
I tried to recreate device.map as you see in comment #5. It's different than before:
(hd0) /dev/sdb
(hd1) /dev/sda
Since originally "(hd0) /dev/sdb" is my boot device which is now "(hd1)   /dev/sda" grub fails to boot.

I solve that with hand editing menu.lst. This is easy for me. But for others it can be showstopper.

I've got the same symptoms on laptop HP NC8230.

After upgrade from mga1 to mga2b2 behavior was normal. All drives are checked in schema /dev/s[d|r]*. After upgrading kernel from 3.3.0 rc7 to 3.3.0-2 hd drive mapping changed from sd* to hd*. I recognised this because of old style modified fstab without UUIDs. After complete change to UUID in fstab and grub/menu.lst system was bootable again. 
Application k3b isn't recognising dvd burner. But I'm not sure if there is a junction. In the end of kde4 start I get an dbus error together with bluetooth error and missing bluetooth device. Possibly this dbus erroro prevents dvd burning (k3b wasn't long ago switched from hal to dbus).

My other older testing system with module "pata_via" hasn't changed from /dev/sd* during update to "3.3.0-2". This points to changes connected with kernel module "pata_piix".

This behaviour change is showstopper for most users and should be reverted with 3.3.1.

Thanks for your interest,

with regards Juergen

CC: (none) => preugen

Jürgen, what is the result of:
lspcidrake -v | grep STORAGE_IDE

Did you try rdloaddriver option with that module?

Do you use grub2? Grub legacy doesn't have support for UUID.

(In reply to comment #5)
Now it works for me even without kernel option with the new kernel 3.3.1-desktop-1.mga2 and udev.


I think that this issue is also a grub bug because of a difference in hda and sda notation of kernel and grub. Is it a regression of bug #570?

# echo quit | grub --batch --device-map=/tmp/device.map
# cat /tmp/device.map 
(fd0)   /dev/fd0
(hd0)   /dev/hda
(hd1)   /dev/sda


# sfdisk -l

Disk /dev/sda: 19929 stez, 255 glav, 63 sektorjev/stezo
Enote = steze po 8225280 bajtov, bloki po 1024 bajtov, Å¡tejem od 0

 Naprava Zagon ZaÄetek Konec  #steze    #bloki    Id  Sistem
/dev/sda1   *      0+  19929-  19930- 160085984+  83  Linux
/dev/sda2          0       -       0          0    0  Prazno
/dev/sda3          0       -       0          0    0  Prazno
/dev/sda4          0       -       0          0    0  Prazno

Disk /dev/sdb: 77825 stez, 255 glav, 63 sektorjev/stezo
Enote = steze po 8225280 bajtov, bloki po 1024 bajtov, Å¡tejem od 0

 Naprava Zagon ZaÄetek Konec  #steze    #bloki    Id  Sistem
/dev/sdb1   *      0+   8679-   8680-  69719008+   7  HPFS/NTFS/exFAT
/dev/sdb2       8679+  12595-   3917-  31457280   83  Linux
/dev/sdb3      12595+  16511-   3917-  31456256   83  Linux
/dev/sdb4      16512   77825-  61314- 492498880    5  Razširjen
/dev/sdb5      16512+  64490-  47979- 385390293+  83  Linux
/dev/sdb6      64491+  77187-  12697- 101986507    7  HPFS/NTFS/exFAT
/dev/sdb7      77187+  77825-    638-   5120000   82  Linux izmenjalni / Solaris

(In reply to comment #8)
> After upgrade from mga1 to mga2b2 behavior was normal. All drives are checked
> in schema /dev/s[d|r]*. After upgrading kernel from 3.3.0 rc7 to 3.3.0-2 hd
> drive mapping changed from sd* to hd*.

3.3.0-2 added http://svnweb.mageia.org/packages/cauldron/kernel/current/PATCHES/patches/ata-prefer-ata-drivers-over-ide-drivers-when-both-are-built.patch?revision=225786&view=markup&sortby=date which should do the opposite (switch you back to sd*).
Thinking about it, the problem may be that you were running the previous one when installing 3.3.0-2, it should then have been fixed by the next update (installed while running 3.3.0-2).

CC: (none) => pterjan

(In reply to comment #9)
> Jürgen, what is the result of:
> lspcidrake -v | grep STORAGE_IDE
> 
> Did you try rdloaddriver option with that module?
> 
> Do you use grub2? Grub legacy doesn't have support for UUID.

Hallo Filip,

I haven't sorting problems on this laptop, because there is no SATA HW build in. I haven't tried cardbus SATA.

to your question:

   [root@xara ~]# lspcidrake -v | grep STORAGE_IDE
ata_piix        : Intel Corporation|82801FB/FBM/FR/FW/FRW (ICH6 Family) IDE Controller [STORAGE_IDE] (vendor:8086 device:266f subv:103c subd:0934) (rev: 03)

Yesterday I updated to new kernel 3.3.1, but no self healing features.

   cat /var/log/dmesg

gives
...
[    0.000000] Linux version 3.3.1-desktop-1.mga2 (iurt@jonund.mageia.org) (gcc version 4.6.3 (GCC) ) #1 SMP Mon Apr 2 23:56:28 UTC 
2012
...
[    0.000000] DMI: Hewlett-Packard HP Compaq nc8230 (EK066ES#ABD)/0934, BIOS 68DTV Ver. F.18 07/24/2009
...
[    0.000000] Kernel command line: BOOT_IMAGE=linux root=UUID=c847c22f-5a4a-4f21-aec6-39a713435a97 splash=silent resume=UUID=d1b9a973-8f6c-40e9-851e-f86b8b42015a vga=866
...
[    4.249377] piix 0000:00:1f.1: IDE controller (0x8086:0x266f rev 0x03)
[    4.249405] piix 0000:00:1f.1: IDE port disabled
[    4.249417] piix 0000:00:1f.1: not 100% native mode: will probe irqs later
[    4.249424]     ide0: BM-DMA at 0x3240-0x3247
[    4.249432] Probing IDE interface ide0...
...
[    4.617304] hda: SAMSUNG HM160HC, ATA DISK drive
[    5.076009] hdb: MATSHITADVD-RAM UJ-832S, ATAPI CD/DVD-ROM drive
[    5.076073] hda: host max PIO4 wanted PIO255(auto-tune) selected PIO4
[    5.076170] hda: UDMA/100 mode selected
[    5.076257] hdb: host max PIO4 wanted PIO255(auto-tune) selected PIO4
[    5.076543] hdb: MWDMA2 mode selected
[    5.076949] ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
...
[    5.705341] SCSI subsystem initialized
[    5.725673] libata version 3.00 loaded.
[    5.741128] ide-gd driver 1.18
[    5.741176] hda: max request size: 512KiB
[    5.741257] ide-cd driver 5.00
[    5.743056] hda: 312581808 sectors (160041 MB) w/8192KiB Cache, CHS=19457/255/63
[    5.744965] hda: cache flushes supported
[    5.793771]  hda: hda1 hda2 hda4 < hda5 hda6 hda7 >
[    5.796213] ide-cd: hdb: ATAPI 24X DVD-ROM DVD-R/RAM CD-R/RW drive, 2048kB Cache
[    5.796222] cdrom: Uniform CD-ROM driver Revision: 3.20

Yesterday I've also tried old kernel vmlinuz-2.6.38.8-desktop-8.mga, which definitly was using /dev/sd* schema. But it gave same results, so it won't depend of intrinsics of already bundled initrd compared to newly created. So there might be some magic of udev/systemd for device naming.

I'll do another comment for an old HP Compaq Armada M700 with Intel BX Chipset, which is not affected. No clue.

With regards, Juergen

(In reply to comment #11)
> (In reply to comment #8)
> > After upgrade from mga1 to mga2b2 behavior was normal. All drives are checked
> > in schema /dev/s[d|r]*. After upgrading kernel from 3.3.0 rc7 to 3.3.0-2 hd
> > drive mapping changed from sd* to hd*.
> 
> 3.3.0-2 added
> http://svnweb.mageia.org/packages/cauldron/kernel/current/PATCHES/patches/ata-prefer-ata-drivers-over-ide-drivers-when-both-are-built.patch?revision=225786&view=markup&sortby=date
> which should do the opposite (switch you back to sd*).
> Thinking about it, the problem may be that you were running the previous one
> when installing 3.3.0-2, it should then have been fixed by the next update
> (installed while running 3.3.0-2).

Thanks for your interest.
Self healing isn't working in this part. Updated yesterday to 3.3.1 which has the proposed driver order. Also tried old mga-Kernel 2.6.38.8-8 and linus-kernel 3.3.1.

Now I assume, numbering information is external to kernel (comment #12) at boot time. 
What has changed in udev/systemd in time frame of kernel change 3.3.0rc7 -> 3.3.0-2?
Also to note, the dvd drive (hdb) is actually not acessible.

with regards, Juergen

(In reply to comment #13)
> Also to note, the dvd drive (hdb) is actually not acessible.
See bug 4997 for that.

Hi,

I've discovered something, which might bring us a solution.

I researched with lsmod command the stack of kernel modules. I discovered, that there are two modules for for Intel pata chips are loaded:

piix and pata_piix

The first is connected with ide*-modules, the latter with pata* modules.

Now I've tried to delete piix. Wasn't possible, because it was in use. But it's pata* sibling was removeable. So the ide* modules are loading earlier then the pata* ones.

What to do to prove?
I renamed /lib/modules/3.3.1-desktop*/kernel/drivers/ide to somewhat. Then I rebuild initrd and rebooted. Voila! Because the lack of ide* stack the pata* modules occupied the hardware and I'm happy. Also dvd is working again and is autodetecting media.

But where is root cause? I searched in /etc/modprobe.*. Most significant in my eyes is /etc/modprobe.d/00_modprobe.conf.

[root@xara ~]# ll /etc/modprobe.d/00_modprobe.conf
-rw-r--r-- 1 root root 5035 Mär 23 13:27 /etc/modprobe.d/00_modprobe.conf
[root@xara ~]# rpm -qf /etc/modprobe.d/00_modprobe.conf
module-init-tools-3.16-16.mga2

The content of this file dates to 2004. The chapter for block devices is  pointing to old style ide drivers. Possibly was this the game changer.

Can someone check my findings and refine it for solution? Are the old style ide drivers really useful for general audience of MGA2?

Good evening in Europe,

with regards Juergen

I tried upgrade on adjusted copy of working M1 also. I used method "Via mgaapplet" first from https://wiki.mageia.org/en/QA_process_for_testing_upgrades but it stalled during third hdlist download. Probably mirror issue but it works OK on M1. Tried twice with the same results.

So I decided to try also "Via a Mageia mirror" method. Upgrade worked but I still had to edit menu.lst to boot.


There is no more hda notation in grub:
# echo quit | grub --batch --device-map=/tmp/device.map
# cat /tmp/device.map 
(fd0)   /dev/fd0
(hd0)   /dev/sda
(hd1)   /dev/sdb


Note that before upgrade lspcidrake uses "ata_piix" and after "ata_generic" module.



Summary: apart from GRUB bug (hd0 and hd1 mixup) it works on this system.

Hi,

This bug was filed against cauldron, but we do not have cauldron at the moment.

Please report whether this bug is still valid for Mageia 2.

Thanks :)

Cheers,
marja

Keywords: (none) => NEEDINFO

Hi all,

I haven't seen the problem anymore. This system is running MGA2 clean and fast. Bug is resolved.

Thanks for your question. Have a nice week,
with regards,

Jürgen

Status: NEW => RESOLVED
Resolution: (none) => FIXED

Unfortunately I have to reopen this bug as grub legacy in mga4 b1 again mess up device.map on this desktop with every kernel upgrade. As a consequence install.sh and menu.lst are of course also wrong.

It's a fresh install of Mageia-4-beta1-i586-DVD. But I think that this doesn't really matter.

As you can see in details bellow I still have one IDE hd which wrongly jumps above SATA in grub's hdx order.




More details:

Here's my manually fixed properly working device.map:
# cat /boot/grub/device.map
#170G SATA SSD, hd1 in grub shell emulation
(hd0) /dev/sdb

# hd1 read error

#600G SATA, hd2 in grub shell emulation
(hd2) /dev/sdc

#160G IDE, hd0 in grub shell emulation
(hd3) /dev/sda



If I run probing devices in grub from terminal I get this non working device.map:
# echo quit | grub --batch --device-map=/tmp/device.map
# cat /tmp/device.map 
(fd0)   /dev/fd0
(hd0)   /dev/sda
(hd1)   /dev/sdb
(hd2)   /dev/sdc



# LC_ALL=C sfdisk -l                                                                                                                                            
                                                                                                                                                                               
Disk /dev/sda: 19929 cylinders, 255 heads, 63 sectors/track                                                                                                                    
sfdisk: Warning: extended partition does not start at a cylinder boundary.                                                                                                     
DOS and Linux will interpret the contents differently.                                                                                                                         
                                                                                                                                                                               
Units: cylinders of 8225280 bytes, blocks of 1024 bytes, counting from 0                                                                                                       
                                                                                                                                                                               
   Device Boot Start     End   #cyls    #blocks   Id  System                                                                                                                   
/dev/sda1          0       -       0          0    0  Empty                                                                                                                    
/dev/sda2          0       -       0          0    0  Empty                                                                                                                    
/dev/sda3          0       -       0          0    0  Empty                                                                                                                    
/dev/sda4          0+  19929-  19930- 160084992    5  Extended
/dev/sda5          0+  19929-  19930- 160083968   83  Linux

Disk /dev/sdb: 21889 cylinders, 255 heads, 63 sectors/track
sfdisk: Warning: extended partition does not start at a cylinder boundary.
DOS and Linux will interpret the contents differently.

Units: cylinders of 8225280 bytes, blocks of 1024 bytes, counting from 0

   Device Boot Start     End   #cyls    #blocks   Id  System
/dev/sdb1          0+  21889-  21890- 175824896    5  Extended
/dev/sdb2          0       -       0          0    0  Empty
/dev/sdb3          0       -       0          0    0  Empty
/dev/sdb4          0       -       0          0    0  Empty
/dev/sdb5          0+   4177-   4178-  33554432   83  Linux
/dev/sdb6       4177+   8354-   4178-  33553408   83  Linux
/dev/sdb7       8355+  21367-  13012- 104518656   83  Linux
/dev/sdb8      21367+  21889-    523-   4194304   82  Linux swap / Solaris

Disk /dev/sdc: 77825 cylinders, 255 heads, 63 sectors/track
sfdisk: Warning: extended partition does not start at a cylinder boundary.
DOS and Linux will interpret the contents differently.

Units: cylinders of 8225280 bytes, blocks of 1024 bytes, counting from 0

   Device Boot Start     End   #cyls    #blocks   Id  System
/dev/sdc1   *      0+   7832-   7833-  62916576+   7  HPFS/NTFS/exFAT
/dev/sdc2          0       -       0          0    0  Empty
/dev/sdc3          0       -       0          0    0  Empty
/dev/sdc4       7832+  77825-  69993- 562214912    5  Extended
/dev/sdc5       7832+  73125-  65293- 524463104   83  Linux
/dev/sdc6      73125+  77302-   4178-  33554432    7  HPFS/NTFS/exFAT

Keywords: NEEDINFO => (none)
Status: RESOLVED => REOPENED
Resolution: FIXED => (none)

Unreproducible due to IDE hard drive removed.

Status: REOPENED => RESOLVED
Resolution: (none) => OLD