Helium Sources (kernel patchset)

Message

hirakendu · Post by **hirakendu** » Sat Aug 07, 2010 4:51 am

2.6.35-helium0: patch, broken-out and precompiled.

Contains reiser4 (upstream), aufs2 (git-20100802), tuxonice (git-2010805), fbcondecor (0.9.6-2.6.35-rc4).

Reiser4 is broken for me. It segfaults while unmounting (and doesn't sync before unmounting). So I would not recommend it to anyone using Reiser4. (Please note that this was while mounting and unmounting a reiser4 formatted file as a loop device.) Some dmesg details indicate problems due to changes in kernel/exit.c:

Code: Select all

reiser4: loop2: found disk format 4.0.0.
------------[ cut here ]------------
kernel BUG at fs/reiser4/plugin/item/extent_flush_ops.c:551!
invalid opcode: 0000 [#1] SMP 
last sysfs file: /sys/devices/virtual/block/loop2/range
CPU 0 
Modules linked in: snd_seq_dummy snd_seq_oss snd_seq_midi_event snd_seq snd_seq_device snd_pcm_oss snd_mixer_oss ipv6 sco bnep rfcomm l2cap hfsplus coretemp hwmon nvidia(P) snd_hda_codec_nvhdmi snd_hda_codec_realtek snd_hda_intel snd_hda_codec snd_hwdep uvcvideo snd_pcm arc4 snd_timer ecb snd ttm ath9k ath9k_common ath9k_hw videodev drm_kms_helper drm ohci1394 ieee1394 soundcore btusb asus_laptop i2c_algo_bit r8169 psmouse intel_agp sparse_keymap snd_page_alloc sdhci_pci sdhci mmc_core ath usbhid hid mac80211 cfg80211 bluetooth v4l2_compat_ioctl32 rfkill i2c_core video output led_class

Pid: 3964, comm: umount Tainted: P            2.6.35-helium0 #1 G50VT     /G50VT               
RIP: 0010:[<ffffffff81195b07>]  [<ffffffff81195b07>] assign_real_blocknrs+0x107/0x120
RSP: 0018:ffff8801316ebb08  EFLAGS: 00010246
RAX: 0000000000000000 RBX: 0000000000000000 RCX: 00000000000013f3
RDX: 0000000000000000 RSI: 00000000000113f3 RDI: ffff880132aa5820
RBP: ffff88012c769590 R08: 00000000000004e0 R09: 0000000000000500
R10: dead000000200200 R11: dead000000100100 R12: 0000000000000001
R13: 0000000000000000 R14: 00000000000004e0 R15: ffff880132aa5820
FS:  00007fb3c429a740(0000) GS:ffff880001800000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
CR2: 00007fb3c3975e60 CR3: 000000010f7c3000 CR4: 00000000000406f0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
Process umount (pid: 3964, threadinfo ffff8801316ea000, task ffff88013a644440)
Stack:
 ffff88012c769600 00000000000113f3 00000000000004e0 ffff88011d190000
<0> ffff88012c769598 ffff8801316ebbb8 ffff88012c769590 ffff88012c769598
<0> ffff88012c769600 0000000000000001 00000000000113f3 ffffffff8119656a
Call Trace:
 [<ffffffff8119656a>] ? reiser4_alloc_extent+0x1aa/0x510
 [<ffffffff8115a094>] ? move_lh_internal+0x104/0x1c0
 [<ffffffff81164836>] ? handle_pos_to_twig+0x66/0x130
 [<ffffffff811649f0>] ? handle_pos_on_twig+0xf0/0x2a0
 [<ffffffff811672f3>] ? flush_current_atom+0x713/0x810
 [<ffffffff81162672>] ? reiser4_txn_end+0x562/0x660
 [<ffffffff811622a6>] ? reiser4_txn_end+0x196/0x660
 [<ffffffff8115a952>] ? longterm_lock_znode+0x3d2/0x530
 [<ffffffff81162779>] ? reiser4_txn_restart+0x9/0x20
 [<ffffffff81162fb5>] ? force_commit_atom+0x45/0x60
 [<ffffffff8116317c>] ? txnmgr_force_commit_all+0x11c/0x160
 [<ffffffff81199286>] ? release_format40+0x66/0x110
 [<ffffffff8115d723>] ? reiser4_init_context+0x53/0x90
 [<ffffffff8116daf9>] ? reiser4_put_super+0x49/0xf0
 [<ffffffff810d3e3f>] ? generic_shutdown_super+0x4f/0xd0
 [<ffffffff810d3ee8>] ? kill_block_super+0x28/0x50
 [<ffffffff810d2e63>] ? deactivate_locked_super+0x43/0x70
 [<ffffffff810ea805>] ? sys_umount+0x85/0x3c0
 [<ffffffff8100242b>] ? system_call_fastpath+0x16/0x1b
Code: 4d 39 ec 77 8f 48 8b 54 24 18 8b 02 89 c2 c1 fa 08 38 c2 74 22 48 8b 44 24 18 fe 00 48 83 c4 28 5b 5d 41 5c 41 5d 41 5e 41 5f c3 <0f> 0b eb fe 0f 0b eb fe 0f 0b eb fe 0f 0b eb fe 66 0f 1f 84 00 
RIP  [<ffffffff81195b07>] assign_real_blocknrs+0x107/0x120
 RSP <ffff8801316ebb08>
---[ end trace dc625ea8bb60d99c ]---
------------[ cut here ]------------
WARNING: at kernel/exit.c:896 do_exit+0x753/0x780()
Hardware name: G50VT               
Modules linked in: snd_seq_dummy snd_seq_oss snd_seq_midi_event snd_seq snd_seq_device snd_pcm_oss snd_mixer_oss ipv6 sco bnep rfcomm l2cap hfsplus coretemp hwmon nvidia(P) snd_hda_codec_nvhdmi snd_hda_codec_realtek snd_hda_intel snd_hda_codec snd_hwdep uvcvideo snd_pcm arc4 snd_timer ecb snd ttm ath9k ath9k_common ath9k_hw videodev drm_kms_helper drm ohci1394 ieee1394 soundcore btusb asus_laptop i2c_algo_bit r8169 psmouse intel_agp sparse_keymap snd_page_alloc sdhci_pci sdhci mmc_core ath usbhid hid mac80211 cfg80211 bluetooth v4l2_compat_ioctl32 rfkill i2c_core video output led_class
Pid: 3964, comm: umount Tainted: P      D     2.6.35-helium0 #1
Call Trace:
 [<ffffffff8103be8b>] ? warn_slowpath_common+0x7b/0xc0
 [<ffffffff810400a3>] ? do_exit+0x753/0x780
 [<ffffffff81478536>] ? printk+0x40/0x45
 [<ffffffff8103ce8c>] ? kmsg_dump+0x7c/0x150
 [<ffffffff810066f2>] ? oops_end+0x72/0xa0
 [<ffffffff81003e94>] ? do_invalid_op+0x84/0xa0
 [<ffffffff81195b07>] ? assign_real_blocknrs+0x107/0x120
 [<ffffffff81198cc1>] ? T.650+0x221/0x230
 [<ffffffff81002fd5>] ? invalid_op+0x15/0x20
 [<ffffffff81195b07>] ? assign_real_blocknrs+0x107/0x120
 [<ffffffff81195a8d>] ? assign_real_blocknrs+0x8d/0x120
 [<ffffffff8119656a>] ? reiser4_alloc_extent+0x1aa/0x510
 [<ffffffff8115a094>] ? move_lh_internal+0x104/0x1c0
 [<ffffffff81164836>] ? handle_pos_to_twig+0x66/0x130
 [<ffffffff811649f0>] ? handle_pos_on_twig+0xf0/0x2a0
 [<ffffffff811672f3>] ? flush_current_atom+0x713/0x810
 [<ffffffff81162672>] ? reiser4_txn_end+0x562/0x660
 [<ffffffff811622a6>] ? reiser4_txn_end+0x196/0x660
 [<ffffffff8115a952>] ? longterm_lock_znode+0x3d2/0x530
 [<ffffffff81162779>] ? reiser4_txn_restart+0x9/0x20
 [<ffffffff81162fb5>] ? force_commit_atom+0x45/0x60
 [<ffffffff8116317c>] ? txnmgr_force_commit_all+0x11c/0x160
 [<ffffffff81199286>] ? release_format40+0x66/0x110
 [<ffffffff8115d723>] ? reiser4_init_context+0x53/0x90
 [<ffffffff8116daf9>] ? reiser4_put_super+0x49/0xf0
 [<ffffffff810d3e3f>] ? generic_shutdown_super+0x4f/0xd0
 [<ffffffff810d3ee8>] ? kill_block_super+0x28/0x50
 [<ffffffff810d2e63>] ? deactivate_locked_super+0x43/0x70
 [<ffffffff810ea805>] ? sys_umount+0x85/0x3c0
 [<ffffffff8100242b>] ? system_call_fastpath+0x16/0x1b
---[ end trace dc625ea8bb60d99d ]---

dusanc · Post by **dusanc** » Sun Aug 08, 2010 8:04 pm

As I usually cc these messages to maintainer cause none else will, I got this back:

"loop is over a file on what file system?
Doesn't he want to fsck the last one?"

hirakendu · Post by **hirakendu** » Sun Aug 08, 2010 8:44 pm

@dusanc: Thank you for cc'ing the kernel messages to upstream (assuming I understand you correctly). The loop is over an ext3 file system, and here are the steps to reproduce:

Code: Select all

# dd if=/dev/zero of=abc.rfs bs=1024 count=500000  # Create a 512 MB file, and assuming current directory is on a ext3 filesystem.
# mkfs.reiser4 -f abc.rfs  # Force format this file with Reiser4.
# mkdir -p /mnt/test && mount abc.rfs /mnt/test/ -o loop  # Mount this file as a loop device.
# cp /boot/vmlinuz /mnt/test  #  Copy some file.
# umount /mnt/test  # Try to unmount quickly, potentially before filesystem syncs, and we should get a Segmentation fault.

And the same kernel messages:

Code: Select all

reiser4: loop2: found disk format 4.0.0.
------------[ cut here ]------------
kernel BUG at fs/reiser4/plugin/item/extent_flush_ops.c:551!
invalid opcode: 0000 [#1] SMP 
last sysfs file: /sys/devices/virtual/block/loop2/range
CPU 0 
Modules linked in: snd_seq_dummy snd_seq_oss snd_seq_midi_event snd_seq snd_seq_device snd_pcm_oss snd_mixer_oss ipv6 sco bnep rfcomm l2cap hfsplus snd_hda_codec_nvhdmi snd_hda_codec_realtek arc4 snd_hda_intel ecb nvidia(P) ath9k snd_hda_codec snd_hwdep asus_laptop btusb bluetooth snd_pcm ath9k_common ath9k_hw ath mac80211 cfg80211 snd_timer sparse_keymap snd rfkill usbhid soundcore sdhci_pci sdhci uvcvideo mmc_core snd_page_alloc video videodev v4l2_compat_ioctl32 ohci1394 hid ieee1394 intel_agp output led_class i2c_core r8169 psmouse

Pid: 19888, comm: umount Tainted: P            2.6.35-helium0 #1 G50VT     /G50VT               
RIP: 0010:[<ffffffff81195b07>]  [<ffffffff81195b07>] assign_real_blocknrs+0x107/0x120
RSP: 0018:ffff880037b09b08  EFLAGS: 00010246
RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000000
RDX: 0000000000000000 RSI: 0000000000010000 RDI: ffff880037888020
RBP: ffff88011fa0ed90 R08: 000000000000001a R09: ffff880037bb6008
R10: dead000000200200 R11: dead000000100100 R12: 000000000000032b
R13: 0000000000000000 R14: 000000000000001a R15: ffff880037888020
FS:  00007f2d00c39740(0000) GS:ffff880001800000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
CR2: 00007f2d00313e60 CR3: 00000000b08f8000 CR4: 00000000000406f0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
Process umount (pid: 19888, threadinfo ffff880037b08000, task ffff880129884fa0)
Stack:
 ffff88011fa0ee00 0000000000010000 000000000000001a ffff880096ada000
<0> ffff88011fa0ed98 ffff880037b09bb8 ffff88011fa0ed90 ffff88011fa0ed98
<0> ffff88011fa0ee00 0000000000000001 0000000000010000 ffffffff8119656a
Call Trace:
 [<ffffffff8119656a>] ? reiser4_alloc_extent+0x1aa/0x510
 [<ffffffff8115a094>] ? move_lh_internal+0x104/0x1c0
 [<ffffffff81164836>] ? handle_pos_to_twig+0x66/0x130
 [<ffffffff811649f0>] ? handle_pos_on_twig+0xf0/0x2a0
 [<ffffffff811672f3>] ? flush_current_atom+0x713/0x810
 [<ffffffff81162672>] ? reiser4_txn_end+0x562/0x660
 [<ffffffff811622a6>] ? reiser4_txn_end+0x196/0x660
 [<ffffffff8115a952>] ? longterm_lock_znode+0x3d2/0x530
 [<ffffffff81162779>] ? reiser4_txn_restart+0x9/0x20
 [<ffffffff81162fb5>] ? force_commit_atom+0x45/0x60
 [<ffffffff8116317c>] ? txnmgr_force_commit_all+0x11c/0x160
 [<ffffffff81163062>] ? txnmgr_force_commit_all+0x2/0x160
 [<ffffffff81199286>] ? release_format40+0x66/0x110
 [<ffffffff8115d723>] ? reiser4_init_context+0x53/0x90
 [<ffffffff8116daf9>] ? reiser4_put_super+0x49/0xf0
 [<ffffffff810d3e3f>] ? generic_shutdown_super+0x4f/0xd0
 [<ffffffff810d3ee8>] ? kill_block_super+0x28/0x50
 [<ffffffff810d2e63>] ? deactivate_locked_super+0x43/0x70
 [<ffffffff810ea805>] ? sys_umount+0x85/0x3c0
 [<ffffffff8100242b>] ? system_call_fastpath+0x16/0x1b
Code: 4d 39 ec 77 8f 48 8b 54 24 18 8b 02 89 c2 c1 fa 08 38 c2 74 22 48 8b 44 24 18 fe 00 48 83 c4 28 5b 5d 41 5c 41 5d 41 5e 41 5f c3 <0f> 0b eb fe 0f 0b eb fe 0f 0b eb fe 0f 0b eb fe 66 0f 1f 84 00 
RIP  [<ffffffff81195b07>] assign_real_blocknrs+0x107/0x120
 RSP <ffff880037b09b08>
---[ end trace 02a4a3ab1ef37ff4 ]---
------------[ cut here ]------------
WARNING: at kernel/exit.c:896 do_exit+0x753/0x780()
Hardware name: G50VT               
Modules linked in: snd_seq_dummy snd_seq_oss snd_seq_midi_event snd_seq snd_seq_device snd_pcm_oss snd_mixer_oss ipv6 sco bnep rfcomm l2cap hfsplus snd_hda_codec_nvhdmi snd_hda_codec_realtek arc4 snd_hda_intel ecb nvidia(P) ath9k snd_hda_codec snd_hwdep asus_laptop btusb bluetooth snd_pcm ath9k_common ath9k_hw ath mac80211 cfg80211 snd_timer sparse_keymap snd rfkill usbhid soundcore sdhci_pci sdhci uvcvideo mmc_core snd_page_alloc video videodev v4l2_compat_ioctl32 ohci1394 hid ieee1394 intel_agp output led_class i2c_core r8169 psmouse
Pid: 19888, comm: umount Tainted: P      D     2.6.35-helium0 #1
Call Trace:
 [<ffffffff8103be8b>] ? warn_slowpath_common+0x7b/0xc0
 [<ffffffff810400a3>] ? do_exit+0x753/0x780
 [<ffffffff81478536>] ? printk+0x40/0x45
 [<ffffffff8103ce8c>] ? kmsg_dump+0x7c/0x150
 [<ffffffff810066f2>] ? oops_end+0x72/0xa0
 [<ffffffff81003e94>] ? do_invalid_op+0x84/0xa0
 [<ffffffff81195b07>] ? assign_real_blocknrs+0x107/0x120
 [<ffffffff81198cc1>] ? T.650+0x221/0x230
 [<ffffffff81002fd5>] ? invalid_op+0x15/0x20
 [<ffffffff81195b07>] ? assign_real_blocknrs+0x107/0x120
 [<ffffffff81195a8d>] ? assign_real_blocknrs+0x8d/0x120
 [<ffffffff8119656a>] ? reiser4_alloc_extent+0x1aa/0x510
 [<ffffffff8115a094>] ? move_lh_internal+0x104/0x1c0
 [<ffffffff81164836>] ? handle_pos_to_twig+0x66/0x130
 [<ffffffff811649f0>] ? handle_pos_on_twig+0xf0/0x2a0
 [<ffffffff811672f3>] ? flush_current_atom+0x713/0x810
 [<ffffffff81162672>] ? reiser4_txn_end+0x562/0x660
 [<ffffffff811622a6>] ? reiser4_txn_end+0x196/0x660
 [<ffffffff8115a952>] ? longterm_lock_znode+0x3d2/0x530
 [<ffffffff81162779>] ? reiser4_txn_restart+0x9/0x20
 [<ffffffff81162fb5>] ? force_commit_atom+0x45/0x60
 [<ffffffff8116317c>] ? txnmgr_force_commit_all+0x11c/0x160
 [<ffffffff81163062>] ? txnmgr_force_commit_all+0x2/0x160
 [<ffffffff81199286>] ? release_format40+0x66/0x110
 [<ffffffff8115d723>] ? reiser4_init_context+0x53/0x90
 [<ffffffff8116daf9>] ? reiser4_put_super+0x49/0xf0
 [<ffffffff810d3e3f>] ? generic_shutdown_super+0x4f/0xd0
 [<ffffffff810d3ee8>] ? kill_block_super+0x28/0x50
 [<ffffffff810d2e63>] ? deactivate_locked_super+0x43/0x70
 [<ffffffff810ea805>] ? sys_umount+0x85/0x3c0
 [<ffffffff8100242b>] ? system_call_fastpath+0x16/0x1b
---[ end trace 02a4a3ab1ef37ff5 ]---

Edit 1: Would be happy to provide more information if needed. Other (possibly irrelevant) information: using sys-apps/util-linux-2.18-r1. And my kernel has CONFIG_NUMA=y (seems some numa specific code has changed in kernel/exit.c.)

Edit 2: And this doesn't happen when I use another filesystem like hfsplus instead of reiser4.

dusanc · Post by **dusanc** » Mon Aug 09, 2010 2:58 pm

Would be nice if we could cut the man in the middle

Can you report it directly to reiserfs-devel, would be much faster?

hirakendu · Post by **hirakendu** » Mon Aug 09, 2010 4:45 pm

Very true, it feels silly on my part to make someone report it. I'll try to report to reiserfs-devel directly. Just that I hate to join a new mailing list each time.

dusanc · Post by **dusanc** » Mon Aug 09, 2010 5:58 pm

U don't have to join, U'll be cced

Skotlex · Post by **Skotlex** » Wed Aug 11, 2010 7:47 am

I am a daredevil (I got backup power), so I'll just add that Reiser4 works just fine when it's your main filesystem of choice. :P

Unfortunately I can't stick to 2.6.35, it has a strange bug with my wireless card where it gives me microcode errors :S Probably will have to report it as I can't find any references to it online.

hirakendu · Post by **hirakendu** » Tue Aug 17, 2010 9:02 am

@Skotlex: Well I almost expected it to work on real partitions, otherwise upstream wouldn't have released it.

About wireless issues in 2.6.35: (1) What wireless card you have? (2) You may try the new patch (same version 2.6.35-helium0) that includes 2.6.35.2. Seems from the Changelog that there has been quite a bit of patching for Atheros cards.

punloh · Post by **punloh** » Sat Sep 04, 2010 9:16 am

Hi. Every time I try to compile openoffice in 3 threads using reiser 4 partition I get the following kernel error and always in the same place, specifically where some java part of openoffice is being compiled

Code: Select all

[ 3820.957906] BUG: unable to handle kernel NULL pointer dereference at 0000000000000050 [ 3820.957910] IP: [<ffffffff814af285>] _raw_spin_lock+0x5/0x20 [ 3820.957916] PGD 4ae65067 PUD 0 [ 3820.957919] Oops: 0002 [#1] SMP [ 3820.957921] last sysfs
file: /sys/devices/pci0000:00/0000:00:11.0/host0/target0:0:0/0:0:0:0/block/sda/uevent
[ 3820.957923] CPU 0
[ 3820.957924] Modules linked in: it87 hwmon_vid snd_pcm_oss snd_mixer_oss 8250_pnp firewire_ohci 8250 firewire_core serial_core k10temp hwmon [ 3820.957931] [ 3820.957934] Pid: 11617, comm: java Not tainted 2.6.35-helium #1
GA-MA790X-UD4/GA-MA790X-UD4
[ 3820.957936] RIP: 0010:[<ffffffff814af285>]  [<ffffffff814af285>] _raw_spin_lock+0x5/0x20 [ 3820.957939] RSP: 0018:ffff8800ac8edaf0  EFLAGS: 00010292 [ 3820.957941] RAX: 0000000000000100 RBX: ffff8800ac8edbc8 RCX:
0000000000000010
[ 3820.957942] RDX: 000000000000006e RSI: ffff88006ba37508 RDI:
0000000000000050
[ 3820.957944] RBP: ffff88006ba37508 R08: 0000000000000026 R09:
0000000000000026
[ 3820.957945] R10: ffff880048c810d0 R11: 0000000000000026 R12:
0000000000000000
[ 3820.957947] R13: ffff88006ba37508 R14: ffff8800ac8edbc8 R15:
0000000000000000
[ 3820.957949] FS:  00002ab1018b3710(0000) GS:ffff880028200000(0000) knlGS:00000000f6160700 [ 3820.957950] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 3820.957952] CR2: 0000000000000050 CR3: 00000000ba04b000 CR4:
00000000000006f0
[ 3820.957953] DR0: 0000000000000000 DR1: 0000000000000000 DR2:
0000000000000000
[ 3820.957955] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7:
0000000000000400
[ 3820.957957] Process java (pid: 11617, threadinfo ffff8800ac8ec000, task ffff88010af5dbe0) [ 3820.957958] Stack:
[ 3820.957959]  ffffffff81157336 0000000000000000 ffffffff81146070 ffff880048c810c0 [ 3820.957962] <0> ffff8800ac8edbc8 ffff88006ba37508 ffffea0003c19040
ffff88006ba37508
[ 3820.957964] <0> ffff8800ac8edbc8 0000000000000000 ffffffff8115aa34 0000000000000000 [ 3820.957967] Call Trace:
[ 3820.957971]  [<ffffffff81157336>] ? checkin_logical_cluster
+0x36/0x210
[ 3820.957975]  [<ffffffff81146070>] ? reiser4_get_file_fsdata+0x60/0xb0 [ 3820.957977]  [<ffffffff8115aa34>] ? capture_page_cluster+0x84/0x150 [ 3820.957981]  [<ffffffff8112b6c9>] ? reiser4_preempt_point+0x9/0x30 [ 3820.957983]  [<ffffffff8115ac00>] ? write_end_cryptcompress
+0x100/0x2a0
[ 3820.957985]  [<ffffffff81155d42>] ? reiser4_write_end_careful
+0xb2/0x1f0
[ 3820.957989]  [<ffffffff810f6122>] ? pipe_to_file+0x162/0x1a0 [ 3820.957992]  [<ffffffff81042c8e>] ? current_fs_time+0x1e/0x30 [ 3820.957994]  [<ffffffff810f5fc0>] ? pipe_to_file+0x0/0x1a0 [ 3820.957996]  [<ffffffff810f4a90>] ? splice_from_pipe_feed+0x60/0xf0 [ 3820.957998]  [<ffffffff810f5f3a>] ? generic_file_splice_write
+0xfa/0x180
[ 3820.958001]  [<ffffffff810f5d6b>] ? direct_splice_actor+0x1b/0x20 [ 3820.958003]  [<ffffffff810f5578>] ? splice_direct_to_actor+0xc8/0x1b0 [ 3820.958003]  [<ffffffff810f5d50>] ? direct_splice_actor+0x0/0x20 [ 3820.958003]  [<ffffffff810f56d5>] ? do_splice_direct+0x75/0x90 [ 3820.958003]  [<ffffffff810d3f9e>] ? do_sendfile+0x19e/0x200 [ 3820.958003]  [<ffffffff810d4089>] ? sys_sendfile64+0x89/0xb0 [ 3820.958003]  [<ffffffff810023ab>] ? system_call_fastpath+0x16/0x1b [ 3820.958003] Code: 00 75 05 f0 66 0f b1 17 0f 94 c2 0f b6 c2 85 c0 0f
95 c0 0f b6 c0 c3 66 66 66 66 66 66 2e 0f 1f 84 00 00 00 00 00 b8 00 01 00 00 <f0> 66 0f c1 07 38 e0 74 06 f3 90 8a 07 eb f6 c3 66 66 2e 0f 1f [ 3820.958003] RIP  [<ffffffff814af285>] _raw_spin_lock+0x5/0x20 [ 3820.958003]  RSP <ffff8800ac8edaf0> [ 3820.958003] CR2: 0000000000000050 [ 3820.958048] ---[ end trace 854f280adc0057e0 ]---

Unfortunately my mails to reiserfs-devel don't seem to get delivered to the list so I'm posting it in this thread.

Skotlex · Post by **Skotlex** » Mon Sep 06, 2010 10:44 am

hirakendu wrote:@Skotlex: Well I almost expected it to work on real partitions, otherwise upstream wouldn't have released it.

About wireless issues in 2.6.35: (1) What wireless card you have? (2) You may try the new patch (same version 2.6.35-helium0) that includes 2.6.35.2. Seems from the Changelog that there has been quite a bit of patching for Atheros cards.

The wireless card mysteriously resolved itself after many reinstalls of different parts of the wireless stack :[

Anyway, I reverted back to the previous version because Reiser4 seems to have a few problems! Last time I suspended to RAM after compiling a few packages, one of the files included by the compile (in /usr/include/) process got "one character changed" (I think it was an underscore that became an opening bracket), which then caused compilation to fail until I fixed the file. How did a system include file owned by root got modified while a user-process was executing (and said file was open read-only)? :/

Because the errors caused by this kind of bug could be undetectable (and things just randomly start failing) I had to revert back to the previous kernel version. Hopefully the other reported bugs are somehow tied, so that fixing them will resolve all my issues. I'll try again for 2.6.36.

hirakendu · Post by **hirakendu** » Fri Oct 29, 2010 4:55 am

Here is a initial version of 2.6.36-helium0: patch, broken-out and precompiled.

It does not contain reiser4 for now and will be added if and when released upstream. At the moment, it only contains aufs2 (from here and here), tuxonice (3.2-rc2), fbcondecor (0.9.6-2.6.36-rc8). All tested OK.

Skotlex · Post by **Skotlex** » Sun Nov 21, 2010 1:41 am

Hi. I am only posting here because....

kernelOfTruth wrote:Reiser4 for 2.6.36 is out !
Edward Shishkin wrote:Changes since 2.6.35:

. Fixed ->write_end() method: the old one didn't update i_size, and this
resulted in empty files and "orphan" items. The problem took place only
for splice users (loop devices, pipe2, etc).

. Implemented ->sync_fs() method of super operations. Since introducing a
new per-bdi flushing stuff reiser4 wasn't able to perform sync on umount.
Now such sync is performed by reiser4_sync_fs().

. Fixed an old bug which led to non-deletable directories.

. Fixed an old bug in tail conversion, which led to data corruption when
running in disk space pressure situations.

And I am excited about it.

hirakendu · Post by **hirakendu** » Sun Nov 21, 2010 4:56 am

Thanks for the information. I updated the patch (without name change).

Skotlex · Post by **Skotlex** » Mon Nov 22, 2010 12:21 pm

I recently decided to uninstall gentoo-sources (since I am not using any other kernel than Helium), but now Portage bugs me that it needs some default kernel installed in order to update my nvidia-drivers :/

So I went ahead and wrote a small ebuild for this. Can we not put it as a "quote" or "code" block in the first post? I'll gladly take care of any inquiries or problems that arises from the ebuild. It should work for all kernel versions (as long as it's helium0, otherwise one char needs to be changed within).

Code: Select all

# Copyright 1999-2010 Gentoo Foundation
# Distributed under the terms of the GNU General Public License v2
# $Header: $

ETYPE="sources"

inherit kernel-2

DESCRIPTION="Hirakendu's light patchset. Contains: reiser4, aufs, tuxonice, fbcondecor (gensplash)"
HOMEPAGE="http://hirakendu.mooo.com/helium-sources/"
K_SECURITY_UNSUPPORTED="1"
HE_VERSION="${PV}-helium0"
UNIPATCH_LIST="${DISTDIR}/${HE_VERSION}.patch.bz2"

SRC_URI="${KERNEL_URI} http://hirakendu.mooo.com/${PN}/${HE_VERSION}/${HE_VERSION}.patch.bz2"
IUSE=""
KEYWORDS="alpha amd64 arm hppa ia64 ppc ppc64 sh sparc x86"

Saved as helium-sources-2.6.36.ebuild works just fine for me.

PS: I was thinking of using the revision as the version of the patchset (eg: helium-sources-2.6.36-r2 would be helium2), but I don't think this convention falls into the Gentoo guidelines. :B

hirakendu · Post by **hirakendu** » Tue Nov 23, 2010 7:48 am

@ Skotlex: Thanks for the ebuild and suggestions. Hopefully I will find some time today and clean things up.

hirakendu · Post by **hirakendu** » Tue Nov 23, 2010 2:25 pm

Here is my take on an ebuild, which you can also get from here. Since I have written it more as a bash script, let me know if there have been any potential abuses on my part.

1. Copy and digest this ebuild as "helium-sources-2.6.36.ebuild" or "helium-sources-2.6.36-rX.ebuild" (where X=1,2,3, etc) in sys-kernel/helium-sources/ of your local overlay. Emerging it will fetch the latest 2.6.36.X vanilla kernel patches (along with the reiser4, aufs patches etc of course). So next time a minor update to stable kernel comes out, just bump the revision of this ebuild. The corresponding kernel source is installed with version 2.6.36-helium or 2.6.36-helium-rX.

2. True to Gentoo's configurability, depending on the features you need, the ebuild supports the USE flags "resier4 aufs tuxonice fbcondecor config". (It is more for fun than for any serious purpose.)

Code: Select all

# Copyright 1999-2010 Gentoo Foundation
# Distributed under the terms of the GNU General Public License v2
# $Header: portage-delta/sys-kernel/helium-sources/helium-sources-2.6.36-r1.ebuild,v0.1 2010/11/23 Hirakendu Das$

EAPI="2"

K_SECURITY_UNSUPPORTED="1"
ETYPE="sources"
inherit kernel-2
detect_version
K_NOSETEXTRAVERSION="don't_set_it"

KV3="$(get_version_component_range 1-3)"
HE_BROKEN_OUT="broken-out-${PN}-${KV3}.tar.gz"
HE_URI="http://hirakendu.mooo.com/helium-sources/${KV3}/${HE_BROKEN_OUT}"
VANILLA_EXTRA_PATCH="http://www.kernel.org/pub/linux/kernel/v2.6/patch-${KV3}.${PR/r/}.bz2"

DESCRIPTION="A minimal patchset with USE'able support for reiser4, aufs, tuxonice, fbcondecor"
HOMEPAGE="http://hirakendu.mooo.com/helium-sources/"
SRC_URI="${KERNEL_URI} ${HE_URI}"
if [ ${PR} != "r0" ]; then
  SRC_URI="${SRC_URI} ${VANILLA_EXTRA_PATCH}"
fi

KEYWORDS="~amd64 ~x86"
IUSE="reiser4 aufs tuxonice fbcondecor config"


src_prepare() {
  if [ ${PR} != "r0" ]; then
    epatch ${DISTDIR}/patch-${KV3}.${PR/r/}.bz2
  fi
  unpack ${HE_BROKEN_OUT}
  mv broken-out-${PN}-${KV3} ../
  # if [ -e ../broken-out-${PN}-${KV3}/00* ]; then
  #   EPATCH_FORCE="yes" epatch ../broken-out-${PN}-${KV3}/00*
  # fi
  use reiser4 && epatch ../broken-out-${PN}-${KV3}/01_reiser4.patch
  if use aufs ; then
    if use reiser4 ; then
      epatch ../broken-out-${PN}-${KV3}/02_aufs_reiser4.patch
    else
      epatch ../broken-out-${PN}-${KV3}/02_aufs.patch
    fi
  fi
  use tuxonice && epatch ../broken-out-${PN}-${KV3}/03_tuxonice.patch
  use fbcondecor && epatch ../broken-out-${PN}-${KV3}/04_fbcondecor.patch
  if use config ; then
    cp ../broken-out-${PN}-${KV3}/config* .
    cp -a config .config
  fi
  rm -rf ../broken-out-${PN}-${KV3}/
  if [ ${PR} != "r0" ]; then
    cat Makefile | sed 's/EXTRAVERSION =.*/EXTRAVERSION = -helium-'${PR}'/' > Makefile.new
  else 
    cat Makefile | sed 's/EXTRAVERSION =.*/EXTRAVERSION = -helium/' > Makefile.new
  fi
  mv Makefile.new Makefile
}

Skotlex · Post by **Skotlex** » Thu Dec 02, 2010 9:36 pm

It shows you had plenty of free time in your hands to dedicate yourself into writing such a feature-full ebuild :P

hirakendu · Post by **hirakendu** » Thu Dec 02, 2010 10:03 pm

Thank you and hope it works for you.

hirakendu · Post by **hirakendu** » Sun Jan 23, 2011 1:56 am

As is routine, here is a initial version of 2.6.37-helium0: patch, broken-out and precompiled.

And as can be seen from previous posts, now there is also an ebuild here, that supports different patches as USE flags:

Code: Select all

[ebuild  N    ] sys-kernel/helium-sources-2.6.37  USE="aufs config fbcondecor reiser4 tuxonice -build -symlink" 413 kB [1]

. For point updates 2.6.37.X, just rename the ebuild to helium-sources-2.6.37-rX.ebuild.

Also as usual, it does not contain reiser4 at the moment and will be added if and when released upstream. For now, it contains aufs2 (from aufs 2.1-standalone.tree-37-rcN-20110110), tuxonice (3.2-rc2), fbcondecor (0.9.6-2.6.36-rc8). All tested OK.

Skotlex · Post by **Skotlex** » Thu Jan 27, 2011 10:03 pm

And as usual, I am reporting that the stable patch for R4 is already out (actually, it was just released yesterday, the 26th). :P

Can't wait to update, once more.

hirakendu · Post by **hirakendu** » Thu Jan 27, 2011 10:20 pm

Thanks, I will (try to) update it by tonight.

hirakendu · Post by **hirakendu** » Fri Jan 28, 2011 7:29 am

I have updated to include the reiser4 patch. If using the ebuild, please remove the old Manifest and distfiles and do a fresh ebuild digest. Also, as mentioned earlier, for future point releases of vanilla 2.6.37.X, just rename the ebuild to helium-sources-2.6.37-rX.ebuild and perform the rituals.

hirakendu · Post by **hirakendu** » Mon Apr 04, 2011 3:29 am

2.6.38-helium: patch, ebuild, broken-out and precompiled.

Contains reiser4 (upstream), aufs2.1 (2.1-standalone.tree-38-rcN-20110307), tuxonice (git-20110321), fbcondecor (0.9.6-2.6.37 and gentoo-sources). All tested OK.

Skotlex · Post by **Skotlex** » Mon Apr 04, 2011 11:57 am

Funny, I've been tracking the R4 release for a while, and the same morning it finally comes out for 2.6.38, you also released this update. I am starting to suspect I may be the sole user of this patch-set? :P

In any case, your ebuild worked wonders to update! Such a breeze.

hirakendu · Post by **hirakendu** » Tue Apr 05, 2011 4:25 pm

@ Skotlex: It was a coincidence that it came out on a weekend, when I usually look around. I am happy that at least someone uses this "patchset". BTW, there are just 4 patches, all maintained very well upstream and it is trivial to put them together on my part, so calling it a "patchset" is a bit of a stretch. Still, glad that I have done a working ebuild (by copying from other ones).