Description of problem: After installing FC5 test 2 on machine with only sata disks, default kernel 2.6.15-1.1826.2.10_FC5 works fine, any of the updated kernels since then fail with can't mount /dev/root Remaking initrd with something like: mkinitrd --with=sd_mod.ko /boot/initrd-2.6.15-1.1955_FC5.img 2.6.15-1.1955_FC5 This is an x86_64 bit machine if that makes any difference. Allows machine to boot correctly. Version-Release number of selected component (if applicable): Fedora Core release 4.91 (Pre-FC5), all kernels after 1826 (that I have checked) How reproducible: Very. Steps to Reproduce: 1. Update kernel on machine with sata 2. Reboot, kernel will panic, saying it can't mount /dev/root 3. Actual results: Expected results: Additional info:
hmm. it's present in the initrd's I just checked. Can you attach the output of lsmod and /etc/modprobe.conf ? Could be that whatever sata driver you use doesn't have a dependancy for some reason.
It doesn't look like a dependancy is listed with lsmod, or modprobe.conf, but it "just worked" in 1826. initrd-2.6.15-1.1826.2.10_FC5.img is the initrd that came with the rpm initrd-2.6.15-1.1955_FC5.img-dist is the initrd that came with the rpm initrd-2.6.15-1.1955_FC5.img is the one I made using mkinitrd b3:~% uname -a Linux b3 2.6.15-1.1955_FC5 #1 SMP Wed Feb 15 15:45:28 EST 2006 x86_64 x86_64 x86_64 GNU/Linux b3:~% ls -l /etc/modprobe.conf -rw-r--r-- 1 root root 150 Jan 29 17:10 /etc/modprobe.conf b3:~% cat /etc/modprobe.conf options snd-intel8x0 index=0 remove snd-intel8x0 { /usr/sbin/alsactl store 0 >/dev/null 2>&1 || : ; }; /sbin/modprobe -r --ignore-remove snd-intel8x0 b3:~% lsmod Module Size Used by ppdev 42569 0 autofs4 55369 1 it87 58085 0 hwmon_vid 35777 1 it87 hwmon 36553 1 it87 eeprom 41297 0 i2c_isa 39233 1 it87 rfcomm 105057 0 l2cap 91841 5 rfcomm bluetooth 117189 2 rfcomm,l2cap sunrpc 192265 1 reiserfs 270793 1 dm_mirror 54721 0 dm_mod 90385 1 dm_mirror video 50633 0 button 40673 0 battery 43464 0 ac 38601 0 ipv6 399073 12 usb_storage 104097 0 lp 47761 0 parport_pc 62825 0 parport 74189 3 ppdev,lp,parport_pc floppy 100745 0 nvram 42569 0 ohci1394 68120 0 ieee1394 392025 1 ohci1394 ehci_hcd 65357 0 ohci_hcd 55005 0 sg 69225 0 snd_intel8x0 68329 1 snd_ac97_codec 136217 1 snd_intel8x0 snd_ac97_bus 35905 1 snd_ac97_codec snd_seq_dummy 37189 0 snd_seq_oss 66369 0 snd_seq_midi_event 41281 1 snd_seq_oss snd_seq 89953 5 snd_seq_dummy,snd_seq_oss,snd_seq_midi_event snd_seq_device 42961 3 snd_seq_dummy,snd_seq_oss,snd_seq snd_pcm_oss 85713 0 snd_mixer_oss 51009 1 snd_pcm_oss nvidia 4893104 8 snd_pcm 126537 3 snd_intel8x0,snd_ac97_codec,snd_pcm_oss i2c_nforce2 40897 0 sunhme 56061 0 i2c_core 57537 4 it87,eeprom,i2c_isa,i2c_nforce2 sky2 73797 0 qla1280 151633 0 snd_timer 59593 2 snd_seq,snd_pcm snd 97129 11 snd_intel8x0,snd_ac97_codec,snd_seq_oss,snd_seq,snd_seq_device,snd_pcm_oss ,snd_mixer_oss,snd_pcm,snd_timer soundcore 44257 1 snd snd_page_alloc 44625 2 snd_intel8x0,snd_pcm sata_sil24 45381 0 forcedeth 57925 0 sd_mod 50881 9 sata_nv 43589 7 libata 93913 2 sata_sil24,sata_nv scsi_mod 180753 6 usb_storage,sg,qla1280,sata_sil24,sd_mod,libata ext3 164177 4 jbd 93673 1 ext3 b3:~% cd /boot b3:/boot% zcat initrd-2.6.15-1.1955_FC5.img | cpio -it sys sysroot init sbin lib lib/jbd.ko lib/scsi_mod.ko lib/ext3.ko lib/libata.ko lib/sata_nv.ko lib/sd_mod.ko etc proc bin bin/insmod bin/modprobe bin/nash dev dev/ttyS2 dev/ptmx dev/mapper dev/tty1 dev/tty12 dev/tty dev/rtc dev/tty10 dev/tty11 dev/tty3 dev/ttyS3 dev/ttyS0 dev/tty9 dev/console dev/ttyS1 dev/ram1 dev/tty5 dev/zero dev/tty6 dev/tty4 dev/tty2 dev/null dev/tty8 dev/tty0 dev/tty7 dev/systty dev/ram0 5391 blocks b3:/boot% zcat initrd-2.6.15-1.1955_FC5.img-dist | cpio -it sys sysroot init sbin lib lib/jbd.ko lib/scsi_mod.ko lib/ext3.ko lib/libata.ko lib/sata_nv.ko etc proc bin bin/insmod bin/modprobe bin/nash dev dev/ttyS2 dev/ptmx dev/mapper dev/tty1 dev/tty12 dev/tty dev/rtc dev/tty10 dev/tty11 dev/tty3 dev/ttyS3 dev/ttyS0 dev/tty9 dev/console dev/ttyS1 dev/ram1 dev/tty5 dev/zero dev/tty6 dev/tty4 dev/tty2 dev/null dev/tty8 dev/tty0 dev/tty7 dev/systty dev/ram0 5272 blocks b3:/boot% zcat initrd-2.6.15-1.1826.2.10_FC5.img | cpio -it sys sysroot init sbin lib lib/jbd.ko lib/scsi_mod.ko lib/ext3.ko lib/libata.ko lib/sata_nv.ko lib/sd_mod.ko etc proc bin bin/insmod bin/hotplug bin/modprobe bin/nash dev dev/ttyS2 dev/mapper dev/tty1 dev/tty12 dev/tty dev/rtc dev/tty10 dev/tty11 dev/tty3 dev/ttyS3 dev/ttyS0 dev/tty9 dev/ram0 5272 blocks b3:/boot% zcat initrd-2.6.15-1.1826.2.10_FC5.img | cpio -it sys sysroot init sbin lib lib/jbd.ko lib/scsi_mod.ko lib/ext3.ko lib/libata.ko lib/sata_nv.ko lib/sd_mod.ko etc proc bin bin/insmod bin/hotplug bin/modprobe bin/nash dev dev/ttyS2 dev/mapper dev/tty1 dev/tty12 dev/tty dev/rtc dev/tty10 dev/tty11 dev/tty3 dev/ttyS3 dev/ttyS0 dev/tty9 dev/console dev/ttyS1 dev/tty5 dev/zero dev/ram dev/tty6 dev/tty4 dev/tty2 dev/null dev/tty8 dev/tty0 dev/tty7 dev/systty 4927 blocks
your /etc/modprobe.conf is broken. You should have at least one 'alias scsi_hostadapter...' line in there. Hmm, Bill, it seems kudzu has stopped detecting sata controllers. (15:38:59:davej@nwo:~)$ cat /etc/modprobe.conf alias eth0 e1000 alias scsi_hostadapter ata_piix alias snd-card-0 snd-emu10k1 options snd-card-0 index=0 options snd-emu10k1 index=0 remove snd-emu10k1 { /usr/sbin/alsactl store 0 >/dev/null 2>&1 || : ; }; /sbin/modprobe -r --ignore-remove snd-emu10k1 alias snd-card-1 snd-intel8x0 options snd-card-1 index=1 options snd-intel8x0 index=1 remove snd-intel8x0 { /usr/sbin/alsactl store 1 >/dev/null 2>&1 || : ; }; /sbin/modprobe -r --ignore-remove snd-intel8x0 (15:40:17:davej@nwo:~)$ sudo bash (15:40:59:root@nwo:~)# mv /etc/sysconfig/hwconf /etc/sysconfig/hwconf.broken (15:40:59:root@nwo:~)# mv /etc/modprobe.conf /etc/modprobe.conf.broken (15:40:59:root@nwo:~)# kudzu (15:41:01:root@nwo:~)# diff -u /etc/modprobe.conf.broken /etc/modprobe.conf --- /etc/modprobe.conf.broken 2006-02-14 21:46:34.000000000 -0500 +++ /etc/modprobe.conf 2006-02-19 15:41:01.000000000 -0500 @@ -1,5 +1,4 @@ alias eth0 e1000 -alias scsi_hostadapter ata_piix alias snd-card-0 snd-emu10k1 options snd-card-0 index=0 options snd-emu10k1 index=0 That's not good(TM).
Aha, that links in the the other problem I'm having (but didn't bother me too much since I hadn't installed any new hardware, oh how I've learnt!), kudzu is seg faulting, both at boot : # strace -fv /sbin/kudzu .... 7059 open("/proc/ide/amd74xx/media", O_RDONLY) = -1 ENOTDIR (Not a directory) 7059 getdents(7, {}, 8192) = 0 7059 close(7) = 0 7059 readlink("/proc/self/fd/0", "/dev/pts/3", 4096) = 10 7059 readlink("/proc/self/fd/1", "/dev/pts/3", 4096) = 10 7059 readlink("/proc/self/fd/2", "/dev/pts/3", 4096) = 10 7059 open("/dev/console", O_RDWR) = 7 7059 ioctl(7, TIOCLINUX, 0x7fffff95d11f) = 6 7059 open("/dev/zero", O_RDWR) = 14 7059 mmap(0x10000, 262144, PROT_READ|PROT_WRITE|PROT_EXEC, MAP_SHARED|MAP_FIXED, 14, 0) = 0x10000 7059 close(14) = 0 7059 open("/dev/mem", O_RDWR) = 14 7059 mmap(NULL, 1282, PROT_READ|PROT_WRITE|PROT_EXEC, MAP_SHARED|MAP_FIXED, 14, 0) = 0 7059 mmap(0xa0000, 393216, PROT_READ|PROT_WRITE, MAP_SHARED|MAP_FIXED, 14, 0xa0000) = 0xa0000 7059 close(14) = 0 7059 iopl(0x3) = 0 7059 ioperm(0, 0x400, 0x1) = 0 7059 iopl(0x3) = 0 7059 ioperm(0, 0x400, 0x1) = 0 7059 --- SIGSEGV (Segmentation fault) @ 0 (0) --- 7059 +++ killed by SIGSEGV +++ I can't debug it since it's been stripped (? why, for a test version), and I can't compile the srpm, it gives a bunch of undefined PCI structures etc. /sbin/kudzu -s works fine but doesn't modify /etc/modules.conf or /etc/sysconfig/hwconf Although I get a bit further if I do this: b3:/etc/sysconfig# rm hwconf ../modprobe.conf b3:/etc/sysconfig# kudzu -s b3:/etc/sysconfig# cat ../modprobe.conf alias dev30686 sunhme alias eth0 forcedeth alias eth2 sky2 alias snd-card-0 snd-intel8x0 options snd-card-0 index=0 options snd-intel8x0 index=0 remove snd-intel8x0 { /usr/sbin/alsactl store 0 >/dev/null 2>&1 || : ; }; /sbin/modprobe -r --ignore-remove snd-intel8x0 syslog logs this: Feb 19 22:24:57 b3 kernel: kudzu[1773]: segfault at 000000000000ffcb rip 0000000000422522 rsp 00007fffffc340e0 error 4
kudzu doesn't write scsi_hostadapter aliases any more - they're not needed. mkinitrd should pick it up just fine without it. Kamal - the segfault is entirely unrelated - see bug 181467.
then I'm at a loss to explain why sd_mod isn't ending up in the initrd. Peter ?
No other modules depend on sd_mod, so it doesn't get pulled in unless we see that we're using scsi -- of course we check for scsi_hostadapter to find that. I'll switch it to looking for scsi_mod and/or libata ...
Have you managed to fix this, I'm still seeing the problem in all kernels up to 2.6.15-1.2009.4.2_FC5.
with the latest mkinitrd package installed ?