Re: lenny kernel memory leak?
On Thu, Dec 17, 2009 at 08:05:08AM +0200, Mikko Rapeli wrote: Since then I have switched system shell from bash to dash, shut down a few non-essential services and mounted all fs with noatime. These have helped a bit but I still see slab growing about 1 mb in day. $ uptime 08:01:02 up 2 days, 17:35, 3 users, load average: 0.11, 0.11, 0.06 $ grep Slab /proc/meminfo Slab: 6448 kB # uptime 10:42:01 up 5 days, 20:16, 5 users, load average: 2.09, 1.45, 1.17 # grep Slab /proc/meminfo Slab:16600 kB So I'll start testing a few patches to lenny kernel. I guess these two are good candidates: http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=e7087a828f8714e464fff18d93618727530dfd89;hp=5adf6d63c1697ce1835daf2b5393488a71ee0dca http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=0f3e63a55b1a7b695a79bf3eec2ff5ab6b336037;hp=9381be059bf5831d259e8735005cfa35b7488543 -Mikko -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#561776: linux: ath5k no longer works with kernel since 2.6.30
Package: linux-2.6 Version: 2.6.31-2 Severity: important File: linux Since i upgraded to 2.6.31 (and 2.6.32) my wireless no longer succeeds in connecting to my AP. Symptons are diverse: * i can connect but get disconnected immediately after which scanning starts again * wicd can't start scanning * wicd can't find networks Booting back to 2.6.30 'fixes' the issue. I've messed with rfkill too and that seems to help everynow and then (but mostly not and is very inconvenient) rmmod ath5k; rfkill block all; rfkill unblock all; modprobe ath5k; rfkill unblock all; iwconfig wlan0 up -- Package-specific info: ** Kernel log: boot messages should be attached ** Model information not available ** Network interface configuration: # This file describes the network interfaces available on your system # and how to activate them. For more information, see interfaces(5). # The loopback network interface auto lo iface lo inet loopback # The primary network interface allow-hotplug eth0 iface eth0 inet dhcp ** PCI devices: 00:00.0 Host bridge [0600]: ATI Technologies Inc RS200/RS200M AGP Bridge [IGP 340M] [1002:cbb2] (rev 02) Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx- Status: Cap+ 66MHz+ UDF- FastB2B- ParErr- DEVSEL=medium TAbort- TAbort- MAbort- SERR- PERR- INTx- Latency: 64 Region 0: Memory at a400 (32-bit, prefetchable) [size=64M] Region 1: Memory at a050 (32-bit, prefetchable) [size=4K] Capabilities: access denied Kernel driver in use: agpgart-ati 00:01.0 PCI bridge [0604]: ATI Technologies Inc PCI Bridge [IGP 340M] [1002:7010] (prog-if 00 [Normal decode]) Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx- Status: Cap- 66MHz+ UDF- FastB2B- ParErr- DEVSEL=medium TAbort- TAbort- MAbort- SERR- PERR- INTx- Latency: 99 Bus: primary=00, secondary=01, subordinate=01, sec-latency=66 I/O behind bridge: 2000-2fff Memory behind bridge: 9000-900f Prefetchable memory behind bridge: 9800-9fff Secondary status: 66MHz+ FastB2B- ParErr- DEVSEL=medium TAbort- TAbort- MAbort+ SERR- PERR- BridgeCtl: Parity- SERR- NoISA+ VGA+ MAbort- Reset- FastB2B- PriDiscTmr- SecDiscTmr- DiscTmrStat- DiscTmrSERREn- 00:06.0 Multimedia audio controller [0401]: ALi Corporation M5451 PCI AC-Link Controller Audio Device [10b9:5451] (rev 02) Subsystem: Compaq Computer Corporation Device [0e11:005a] Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx- Status: Cap+ 66MHz- UDF- FastB2B+ ParErr- DEVSEL=medium TAbort- TAbort- MAbort- SERR+ PERR+ INTx- Latency: 64 (500ns min, 6000ns max) Interrupt: pin A routed to IRQ 11 Region 0: I/O ports at 3000 [size=256] Region 1: Memory at a010 (32-bit, non-prefetchable) [size=4K] Capabilities: access denied Kernel driver in use: ALI 5451 00:07.0 ISA bridge [0601]: ALi Corporation M1533/M1535/M1543 PCI to ISA Bridge [Aladdin IV/V/V+] [10b9:1533] Subsystem: Compaq Computer Corporation Device [0e11:005a] Control: I/O+ Mem+ BusMaster+ SpecCycle+ MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx- Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=medium TAbort- TAbort- MAbort- SERR- PERR- INTx- Latency: 0 Capabilities: access denied 00:08.0 Modem [0703]: ALi Corporation M5457 AC'97 Modem Controller [10b9:5457] (prog-if 00 [Generic]) Subsystem: Compaq Computer Corporation Device [0e11:005a] Control: I/O+ Mem+ BusMaster- SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx- Status: Cap+ 66MHz- UDF- FastB2B+ ParErr- DEVSEL=medium TAbort- TAbort- MAbort- SERR- PERR- INTx- Interrupt: pin A routed to IRQ 11 Region 0: Memory at a018 (32-bit, non-prefetchable) [size=4K] Region 1: I/O ports at 3400 [size=256] Capabilities: access denied 00:09.0 Ethernet controller [0200]: Atheros Communications Inc. Atheros AR5001X+ Wireless Network Adapter [168c:0013] (rev 01) Subsystem: Compaq Computer Corporation NC6000 laptop [0e11:00e5] Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx- Status: Cap+ 66MHz- UDF- FastB2B+ ParErr- DEVSEL=medium TAbort- TAbort- MAbort- SERR- PERR- INTx- Latency: 168 (2500ns min, 7000ns max), Cache Line Size: 64 bytes Interrupt: pin A routed to IRQ 5 Region 0: Memory at a008 (32-bit, non-prefetchable) [size=64K] Capabilities: access denied Kernel driver in use: ath5k 00:0b.0 CardBus bridge [0607]: O2 Micro, Inc. OZ711M1/MC1 4-in-1 MemoryCardBus Controller [1217:7114] (rev 20) Subsystem: Compaq Computer Corporation Device [0e11:005a]
Site recommendation from a friend
My Name is Cynthia I am average in height and fair in complexion ,am a loving, romantic and caring angel. In my searching for friend today and i came across your email in my Google search after going through it. then i decided to contact you. I really want to have a good relationship with you. Beside i have a special something i want to discuss with you ,so reach me here(cynthiamaa...@yahoo.com) Hope to hear from you soon. I will send my beautiful pictures to you and also tell you more about myself. Remember distance or color does not matter but love matter's alot. I will be waiting for your mail. Your's Cynthia Kisses.
Bug#477377: rsync: page allocation failure. order:1, mode:0x20
On Sun, Dec 21, 2008 at 01:01:50AM +0100, Moritz Muehlenhoff wrote: On Tue, Apr 22, 2008 at 08:40:45PM +0200, Kurt Roeckx wrote: Package: linux-2.6 Version: 2.6.22-6 Hi, I found this message today: rsync: page allocation failure. order:1, mode:0x20 Call Trace: IRQ [8026d94d] __alloc_pages+0x2a3/0x2bc [80287a1a] kmem_getpages+0x69/0x111 [80287a1a] kmem_getpages+0x69/0x111 [80287f92] fallback_alloc+0x115/0x180 [80287bf2] kmem_cache_alloc_node+0x93/0xed [80396ce5] __alloc_skb+0x64/0x137 [803ca52c] tcp_collapse+0x169/0x37a [803ca8bc] tcp_prune_queue+0x17f/0x27b [803cab85] tcp_data_queue+0x1cd/0xafd [803cce04] tcp_rcv_established+0x802/0x8da [803d2591] tcp_v4_do_rcv+0x30/0x36c [803d4e30] tcp_v4_rcv+0x98f/0xa23 [803b9c27] ip_local_deliver+0x1c9/0x28e [803b9a15] ip_rcv+0x551/0x59a [8039acfd] netif_receive_skb+0x7d/0x392 [880abb21] :tg3:tg3_poll+0x6ff/0x94d [8039d20b] net_rx_action+0xa8/0x1b6 [80237552] __do_softirq+0x55/0xc3 [8021b7ff] ack_apic_level+0x3a/0x4e [8020af2c] call_softirq+0x1c/0x28 [8020c855] do_softirq+0x2c/0x7d [8020ca8b] do_IRQ+0xb6/0xd3 [8020a2b1] ret_from_intr+0x0/0xa EOI [8814cbcd] :jbd:journal_dirty_metadata+0x8f/0x1be [8816ccae] :ext3:__ext3_journal_dirty_metadata+0x1e/0x46 [88160cdd] :ext3:ext3_mark_iloc_dirty+0x2b0/0x339 [881610f4] :ext3:ext3_mark_inode_dirty+0x28/0x31 [88163d7d] :ext3:ext3_dirty_inode+0x63/0x7b [802a9f5f] __mark_inode_dirty+0x29/0x17d [8815e7d1] :ext3:ext3_new_blocks+0x5ec/0x7d3 [802ae8ec] __bread+0x6/0x76 [8816182e] :ext3:ext3_get_blocks_handle+0x405/0x8e6 [80287aa8] kmem_getpages+0xf7/0x111 [88161ffb] :ext3:ext3_get_block+0xc2/0xe4 [802addca] __block_prepare_write+0x18a/0x441 [88161f39] :ext3:ext3_get_block+0x0/0xe4 [802ae09b] block_prepare_write+0x1a/0x25 [8816338a] :ext3:ext3_prepare_write+0xb2/0x17b [8026a432] generic_file_buffered_write+0x288/0x61a [8026ab07] __generic_file_aio_write_nolock+0x343/0x3ad [803901c8] sock_aio_read+0x112/0x19a [8026abd2] generic_file_aio_write+0x61/0xc1 [8815f416] :ext3:ext3_file_write+0x16/0x94 [8028d7f8] do_sync_write+0xc9/0x10c [80220d53] do_page_fault+0x41f/0x78c [80244a5a] autoremove_wake_function+0x0/0x2e [8028dfc9] vfs_write+0xce/0x157 [8028e527] sys_write+0x45/0x6e [80209d9e] system_call+0x7e/0x83 Does this error still occur with more recent kernel versions? It still occurs, but I'm now using a work around. I use this setting in /etc/sysctl.conf: vm.min_free_kbytes = 65536 See http://www.acc.umu.se/~maswan/linux-netperf.txt Kurt -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Bits from the kernel team
Hi, Now that we have a 2.6.32 kernel in unstable, can you updates us on the various things mentioned in this mail? For instance, as I understand it, most other distro's recently had a release with a 2.6.31 kernel? Do you know if there are plans to have a kernel with backported drivers, one used by multiple distributions? Kurt On Mon, Oct 19, 2009 at 10:54:47PM +0100, Vincent Sanders wrote: The Debian Kernel team recently had a series of face to face meetings during the Linux Plumbers Conference [1]. The DPL managed to arrange for the whole team to be present in Oregon at the same time, a representative of the release team was also present. The LPC conference venue allowed the kernel team to interact with the upstream developers and other distributions kernel teams in an positive and productive way. The Debian kernel team meetings ran over four days and covered a large number of topics, the abridged minutes are presented here, the full meeting minutes are also available [2]. Co-operation and version synchronisation with other distributions - This discussion involved timing of the Debian freeze and what implications this might have on the kernel version selected for squeeze. The version selected by other distributions was also discussed. In conclusion the 2.6.32 release will probably be the initial kernel version shipped with squeeze. Separate firmware, what is left to do? -- A constructive discussion was held about the outstanding firmware issues, how the team addresses them and how we might work with upstream to address our DSFG issues with kernel sources. Kernel Mode Setting transition -- It was resolved that KMS will be enabled at build time but disabled at run time by default. The X packages will be able to use modprobe config files to enable KMS at run time as required. Feature patches --- These are patches the Debian kernels have for major features which are not upstream. openvz ++ Debian will continue to support this system with assistance from the openvz developers. rt patchset +++ This is apparently not ready for production use and will not be present in Debian kernels. vserver +++ This feature will be present in squeeze but will be marked as deprecated and a migration path to Linux containers investigated. xen dom 0 + This feature will be included in the squeeze kernel release subject to ongoing stabilisation work. The feature will be marked as deprecated and will not appear in future releases. IDE to libata decision -- Debian will perform this transition using the udev packages in a similar way to Ubuntu. The Ubuntu developers have offered their assistance with this transition. preemption -- This feature will be enabled for the squeeze release. OSS --- This has been a deprecated kernel interface for some time and will be disabled for squeeze with mechanisms put in place to deal with legacy users. bug triage and tagging -- The kernel team has a large number of bugs, many of which contain inadequate information. The team decided that a policy for bugs and patches will be produced and enforced. We will also be improving the bug reporting by improving the reportbug usage. Moving the Debian Kernel packaging to Git - A robust discussion happened with several views and ideas expressed. The final outcome was that the team as a whole favoured the move to git and that further investigation and implementation would occur. Coordination with release team and D-I -- Several issues were covered the main item from this session was an investigation as to if udeb generation should be merged with the main kernel source package. Out of tree modules --- After some discussion it was resolved to remove linux-modules-extra and -nonfree as they are an impossible to support properly. A few modules the project really must have will be placed directly into the linux-2.6 source The kernel team will endorse the use of dkms as a way for out-of-tree module maintainers to get their modules auto-built. Leveraging upstream .deb building - This became a discussion about the general kernel packaging and how we might use the upstream provided facilities better. There was some discussion we have way too many ways to build a kernel. We will be rationalising this to two methods, an upstream merged make deb-pkg target and the linux-2.6 Debian source. We will also be rationalising the kernel postinst and co-ordinating our efforts with the Ubuntu developers. New lists to co-ordinate There is
Re: lenny kernel memory leak?
On Sun, 2009-12-20 at 10:46 +0200, Mikko Rapeli wrote: On Thu, Dec 17, 2009 at 08:05:08AM +0200, Mikko Rapeli wrote: Since then I have switched system shell from bash to dash, shut down a few non-essential services and mounted all fs with noatime. These have helped a bit but I still see slab growing about 1 mb in day. $ uptime 08:01:02 up 2 days, 17:35, 3 users, load average: 0.11, 0.11, 0.06 $ grep Slab /proc/meminfo Slab: 6448 kB # uptime 10:42:01 up 5 days, 20:16, 5 users, load average: 2.09, 1.45, 1.17 # grep Slab /proc/meminfo Slab:16600 kB So I'll start testing a few patches to lenny kernel. I guess these two are good candidates: Try c95edf5432f097c926dd3f59239ecde80da3b214 rt2x00: Properly clean up beacon skbs. http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=e7087a828f8714e464fff18d93618727530dfd89;hp=5adf6d63c1697ce1835daf2b5393488a71ee0dca Could be relevant, but you reported you didn't see the error message above this. http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=0f3e63a55b1a7b695a79bf3eec2ff5ab6b336037;hp=9381be059bf5831d259e8735005cfa35b7488543 This should only have an effect if you unplug the device. Ben. -- Ben Hutchings Once a job is fouled up, anything done to improve it makes it worse. signature.asc Description: This is a digitally signed message part
Bug#561843: ethtool/mii-tool report incorrect/missing values on Marvell 88E1011
Source: linux-2.6 Version: 2.6.32-2 Hi, I get this: # ethtool eth0 |grep -i pause Advertised pause frame use: No Link partner advertised pause frame use: No # ethtool -a eth0 Pause parameters for eth0: Autonegotiate: on RX: on TX: on Those seem to conflict with each other. I assume this is a kernel bug. It also reports: Link partner advertised link modes: Not reported It would be nice that it could report it and that atleast seems to be supported with other cards. As I understand this, this should be supported since a 2.6.31 kernel. It also ways: Link partner advertised auto-negotiation: No Auto-negotiation: on Those also seem to conflict and it really should advertise it. But it might be related to not reporting the link modes and ethtool should show that instead? On the other hand, mii-tool reports this: # mii-tool -v eth0: negotiated 100baseTx-FD flow-control, link ok product info: Yukon 88E1011 rev 5 basic mode: autonegotiation enabled basic status: autonegotiation complete, link ok capabilities: 1000baseT-HD 1000baseT-FD 100baseTx-FD 100baseTx-HD 10baseT-FD 10baseT-HD advertising: 100baseTx-FD 100baseTx-HD 10baseT-FD 10baseT-HD flow-control link partner: 1000baseT-HD 1000baseT-FD 100baseTx-FD 100baseTx-HD 10baseT-FD 10baseT-HD flow-control The link partner is only a 100 mbit switch, so it shouldn't advertise 1000. It looks like the 2 last lines are switched. lspci -v says: 00:0a.0 Ethernet controller: Marvell Technology Group Ltd. 88E8001 Gigabit Ethernet Controller (rev 13) Subsystem: ASUSTeK Computer Inc. Marvell 88E8001 Gigabit Ethernet Controller (Asus) Flags: bus master, 66MHz, medium devsel, latency 64, IRQ 17 Memory at fdd0 (32-bit, non-prefetchable) [size=16K] I/O ports at b000 [size=256] Expansion ROM at fdc0 [disabled] [size=128K] Capabilities: [48] Power Management version 2 Capabilities: [50] Vital Product Data Kernel driver in use: skge Kurt -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Bits from the kernel team
On Sun, 2009-12-20 at 17:22 +0100, Kurt Roeckx wrote: Hi, Now that we have a 2.6.32 kernel in unstable, can you updates us on the various things mentioned in this mail? For instance, as I understand it, most other distro's recently had a release with a 2.6.31 kernel? Fedora 12, Ubuntu 9.10 and openSUSE 11.2 have this kernel version. Do you know if there are plans to have a kernel with backported drivers, one used by multiple distributions? I've not heard of a general plan for that. However there is a compat-wireless project (covering Wifi and Bluetooth drivers) that I think Fedora and Ubuntu are pulling from. That might be something we should do too. I'm also keen to pull in nouveau if the X maintainers are happy with that. Kernel Mode Setting transition -- It was resolved that KMS will be enabled at build time but disabled at run time by default. Done. The X packages will be able to use modprobe config files to enable KMS at run time as required. This is not for the kernel team to do. Feature patches --- These are patches the Debian kernels have for major features which are not upstream. openvz ++ Debian will continue to support this system with assistance from the openvz developers. In progress, I believe. rt patchset +++ This is apparently not ready for production use and will not be present in Debian kernels. vserver +++ This feature will be present in squeeze but will be marked as deprecated and a migration path to Linux containers investigated. Don't know. xen dom 0 + This feature will be included in the squeeze kernel release subject to ongoing stabilisation work. The feature will be marked as deprecated and will not appear in future releases. In progress. IDE to libata decision -- Debian will perform this transition using the udev packages in a similar way to Ubuntu. The Ubuntu developers have offered their assistance with this transition. Since we also have a wide variety of bootloaders to update, we now intend to do this separately from udev. I have written an update script (written in Perl) which I will be sending out for review shortly. preemption -- This feature will be enabled for the squeeze release. OSS --- This has been a deprecated kernel interface for some time and will be disabled for squeeze Done. with mechanisms put in place to deal with legacy users. Er, not sure. bug triage and tagging -- The kernel team has a large number of bugs, many of which contain inadequate information. The team decided that a policy for bugs and patches will be produced and enforced. We will also be improving the bug reporting by improving the reportbug usage. See http://lists.debian.org/debian-kernel/2009/11/msg00114.html. This still needs to be incorporated into the kernel handbook. Moving the Debian Kernel packaging to Git - A robust discussion happened with several views and ideas expressed. The final outcome was that the team as a whole favoured the move to git and that further investigation and implementation would occur. Still to do; I think this may have to be post-squeeze. Coordination with release team and D-I -- Several issues were covered the main item from this session was an investigation as to if udeb generation should be merged with the main kernel source package. No change as yet. Out of tree modules --- After some discussion it was resolved to remove linux-modules-extra and -nonfree as they are an impossible to support properly. A few modules the project really must have will be placed directly into the linux-2.6 source The kernel team will endorse the use of dkms as a way for out-of-tree module maintainers to get their modules auto-built. Done. Leveraging upstream .deb building - This became a discussion about the general kernel packaging and how we might use the upstream provided facilities better. There was some discussion we have way too many ways to build a kernel. We will be rationalising this to two methods, an upstream merged make deb-pkg target and the linux-2.6 Debian source. Done. We will also be rationalising the kernel postinst and co-ordinating our efforts with the Ubuntu developers. In progress. Debug Packages -- This refers to debugging information from current packages, not a separate configuration, useful for crash tools. This will be investigated further. Not done. Automated build and test This might be a useful tool in the future and work is ongoing. Don't know. Experimental Some upload experimental
Bug#561843: ethtool/mii-tool report incorrect/missing values on Marvell 88E1011
On Sun, Dec 20, 2009 at 05:27:33PM +0100, Kurt Roeckx wrote: On the other hand, mii-tool reports this: # mii-tool -v eth0: negotiated 100baseTx-FD flow-control, link ok product info: Yukon 88E1011 rev 5 basic mode: autonegotiation enabled basic status: autonegotiation complete, link ok capabilities: 1000baseT-HD 1000baseT-FD 100baseTx-FD 100baseTx-HD 10baseT-FD 10baseT-HD advertising: 100baseTx-FD 100baseTx-HD 10baseT-FD 10baseT-HD flow-control link partner: 1000baseT-HD 1000baseT-FD 100baseTx-FD 100baseTx-HD 10baseT-FD 10baseT-HD flow-control The link partner is only a 100 mbit switch, so it shouldn't advertise 1000. It looks like the 2 last lines are switched. This part is part is probably #511395 Kurt -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#561843: ethtool/mii-tool report incorrect/missing values on Marvell 88E1011
On Sun, 2009-12-20 at 17:27 +0100, Kurt Roeckx wrote: Source: linux-2.6 Version: 2.6.32-2 Hi, I get this: # ethtool eth0 |grep -i pause Advertised pause frame use: No Link partner advertised pause frame use: No # ethtool -a eth0 Pause parameters for eth0: Autonegotiate: on RX: on TX: on Those seem to conflict with each other. I assume this is a kernel bug. Driver bug. A lot of drivers do not report specific pause frame auto-negotiation flags. It also reports: Link partner advertised link modes: Not reported It would be nice that it could report it and that atleast seems to be supported with other cards. As I understand this, this should be supported since a 2.6.31 kernel. It depends on the driver. I provided patches for the generic MII code but many drivers do not use that and do not yet provide any information about link partner advertising. It also ways: Link partner advertised auto-negotiation: No Auto-negotiation: on Those also seem to conflict and it really should advertise it. [...] No, Auto-negotiation: on does not mean auto-negotiation was successful. Ben. -- Ben Hutchings Once a job is fouled up, anything done to improve it makes it worse. signature.asc Description: This is a digitally signed message part
Nouveau kernel driver
I'd like to pull nouveau from 2.6.33 into Debian's 2.6.32. I'm hoping this would allow for replacement of nv and its dodgy source with nouveau (I realise nouveau has its own issues with non-free bits, but they're more easily separable). Does this sound like a good idea? Ben. -- Ben Hutchings Once a job is fouled up, anything done to improve it makes it worse. signature.asc Description: This is a digitally signed message part
Bug#561764: some warnings fly off screen and are not saved in any file
OK, via
$ zcat /boot/initrd.img-2.6.32-trunk-686|strings|less
I think I found one of the messages I saw:
SYSFS{}= will be removed in a future udev version
which is in /sbin/udevadm ... bug #561229 perhaps.
However, what I'm saying is how bad it is that there be any messages
that fly off the screen at boot and are not also 'remembered into some
log file' one can read later?.
Either remember all messages into those /var/log/ files, or remember
none. Remembering 95% makes one think that they see all of them.
There is some step so early in the boot process that recording has not
been turned on yet. That step is making error messages that fly off the
screen... with only one small laptop one is not going to be able to see these
messages. One should not be required to attach other equipment to see them.
--
To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
DRBD in Squeeze kernel?
Hi, what's the opinion on merging DRBD into the Squeeze kernel? DRBD is proven, popular technology and now that is has been accepted into 2.6.33 I'd like to merge it into the Squeeze kernel so that it's available out of the box. Cheers, Moritz -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#561764: some warnings fly off screen and are not saved in any file
On Mon, 2009-12-21 at 01:39 +0800, jida...@jidanni.org wrote:
OK, via
$ zcat /boot/initrd.img-2.6.32-trunk-686|strings|less
I think I found one of the messages I saw:
SYSFS{}= will be removed in a future udev version
which is in /sbin/udevadm ... bug #561229 perhaps.
However, what I'm saying is how bad it is that there be any messages
that fly off the screen at boot and are not also 'remembered into some
log file' one can read later?.
[...]
Early kernel messages should be copied into /var/log/messages by the log
daemon. Maybe you need to turn up the log level. If you think this is
not a configuration error, please reassign to the log daemon's package
(rsyslog is the default).
Ben.
--
Ben Hutchings
Once a job is fouled up, anything done to improve it makes it worse.
signature.asc
Description: This is a digitally signed message part
Bug#561843: ethtool/mii-tool report incorrect/missing values on Marvell 88E1011
On Sun, Dec 20, 2009 at 04:54:42PM +, Ben Hutchings wrote: It also ways: Link partner advertised auto-negotiation: No Auto-negotiation: on Those also seem to conflict and it really should advertise it. [...] No, Auto-negotiation: on does not mean auto-negotiation was successful. But the link parter does support auto-negotiation, while it says no. Kurt -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#561843: ethtool/mii-tool report incorrect/missing values on Marvell 88E1011
On Sun, 2009-12-20 at 19:36 +0100, Kurt Roeckx wrote: On Sun, Dec 20, 2009 at 04:54:42PM +, Ben Hutchings wrote: It also ways: Link partner advertised auto-negotiation: No Auto-negotiation: on Those also seem to conflict and it really should advertise it. [...] No, Auto-negotiation: on does not mean auto-negotiation was successful. But the link parter does support auto-negotiation, while it says no. That's really a bug in ethtool; it should say not reported. I'll submit a patch. Ben. -- Ben Hutchings Once a job is fouled up, anything done to improve it makes it worse. signature.asc Description: This is a digitally signed message part
Processed: your mail
Processing commands for cont...@bugs.debian.org: notfound 557720 2.6.32-2 Bug #557720 [linux-2.6] FTBFS on powerpc: undefined reference to `.unmask_msi_irq' There is no source info for the package 'linux-2.6' at version '2.6.32-2' with architecture '' Unable to make a source version for version '2.6.32-2' Ignoring request to alter found versions of bug #557720 to the same values previously set thanks Stopping processing here. Please contact me if you need assistance. Debian bug tracking system administrator (administrator, Debian Bugs database) -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#561843: [PATCH] ethtool: Do not report link partner advertising flags if set to 0
Only some drivers (and none before kernel version 2.6.31) currently set these flags. When the flags are equal to 0 and so we don't know what the link partner advertised, don't report anything. Signed-off-by: Ben Hutchings b...@decadent.org.uk --- ethtool.c |4 +++- 1 files changed, 3 insertions(+), 1 deletions(-) diff --git a/ethtool.c b/ethtool.c index 10dfc80..298b690 100644 --- a/ethtool.c +++ b/ethtool.c @@ -948,7 +948,9 @@ static int dump_ecmd(struct ethtool_cmd *ep) dump_supported(ep); dump_advertised(ep, Advertised, ep-advertising); - dump_advertised(ep, Link partner advertised, ep-lp_advertising); + if (ep-lp_advertising) + dump_advertised(ep, Link partner advertised, + ep-lp_advertising); fprintf(stdout,Speed: ); speed = ethtool_cmd_speed(ep); -- 1.6.5.7 -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
2.6.30-2-amd64 kernel panic ... libm.so.6
things were smooth prior to major sid package update last week, have been down since, seems very similar to the bug #550320 filed in October. Unfortunately don't have a second Debian machine to fix (or file bugs from) [] Freeing unused kernel memory: 560k freed [] Write protecting the kernel read-only data: 3808k /bin/sh: error while loading shared libraries: libm.so.6: cannot open shared object file: No such file or directory [] Kernel panic - not synchingL Attempted to kill init! Call Trace... think the distribution screwed something, since the same kernel was stable before. Anyone suggestions welcome. I can boot from USB stick, mount clean, chroot and run aptitude. Thanks. -- Pat.
Bug#523735: Make that visible
Hello, whatever you do to resolve this bug or non-bug, just make it visible to the user. I just wasted some time to debug this problem (or maybe not a real problem, haven't figured that out yet). Anyhow, please print a proper message and write a sane rationale into /etc/kernel/postinst.d/initramfs-tools , the current comment is anything but helpful. Or what do you expect a user to think when seeing: # kernel-package passes an extra arg; hack to not run under kernel-package [ -z $2 ] || exit 0 Maybe WTF, some internal war so they are fighting each other? Regards, Eduard. -- Uneigennützige Freundschaft gibt es nur unter Leuten gleicher Einkommensklasse. -- Jean-Paul Getty -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#561764: some warnings fly off screen and are not saved in any file
On 12/20/2009 11:39 AM, jida...@jidanni.org wrote:
OK, via
$ zcat /boot/initrd.img-2.6.32-trunk-686|strings|less
I think I found one of the messages I saw:
SYSFS{}= will be removed in a future udev version
which is in /sbin/udevadm ... bug #561229 perhaps.
However, what I'm saying is how bad it is that there be any messages
that fly off the screen at boot and are not also 'remembered into some
log file' one can read later?.
Either remember all messages into those /var/log/ files, or remember
none. Remembering 95% makes one think that they see all of them.
There is some step so early in the boot process that recording has not
been turned on yet. That step is making error messages that fly off the
screen... with only one small laptop one is not going to be able to see these
messages. One should not be required to attach other equipment to see them.
Not sure why this is CCed to LKML, the kernel isn't responsible for
logging anything to files..
--
To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#561764: some warnings fly off screen and are not saved in any file
BH == Ben Hutchings b...@decadent.org.uk writes: BH Early kernel messages should be copied into /var/log/messages by the log BH daemon. Maybe you need to turn up the log level. If you think this is BH not a configuration error, please reassign to the log daemon's package BH (rsyslog is the default). All this is happening during the phase when /boot/initrd.img-2.6.32-trunk-686 is the only file operational. I.e., it can't be me to blame as I only use the stock kernel. -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#561843: [PATCH] ethtool: Do not report link partner advertising flags if set to 0
On 12/20/2009 02:12 PM, Ben Hutchings wrote: Only some drivers (and none before kernel version 2.6.31) currently set these flags. When the flags are equal to 0 and so we don't know what the link partner advertised, don't report anything. Signed-off-by: Ben Hutchingsb...@decadent.org.uk --- ethtool.c |4 +++- 1 files changed, 3 insertions(+), 1 deletions(-) applied -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#561606: Kernel fails to boot
On Sun, 2009-12-20 at 21:59 +0100, Nicolas DEGAND wrote: I have been able to get that (crappy) picture. Hope this helps. Sorry, no, we need to know the task names. Try adding the parameter vga=6 to the kernel command line; this should fit more lines on the screen. Ben. -- Ben Hutchings Once a job is fouled up, anything done to improve it makes it worse. signature.asc Description: This is a digitally signed message part
Re: lenny kernel memory leak?
On Sun, Dec 20, 2009 at 04:30:54PM +, Ben Hutchings wrote: Try c95edf5432f097c926dd3f59239ecde80da3b214 rt2x00: Properly clean up beacon skbs. With this I got kernel panics at boot when loading the modules. Maybe it depends on some other patches. http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=e7087a828f8714e464fff18d93618727530dfd89;hp=5adf6d63c1697ce1835daf2b5393488a71ee0dca This is now in. Will see if it makes any difference in a few days. -Mikko -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#550665: linux-image-2.6.30-2-amd64: ALSA: When i use a kernel most newer than 2.6.26 alsamixer section Capture is empty with the exception of the presence of the Digital channel
Hi, The next release of Debian (6.0, code name Squeeze) will be based on 2.6.32. Please test the current 2.6.32 from unstable and tell us whether the problem persists. If so, we should report it upstream to the kernel.org developers. The 2.6.32 kernel is available from packages.debian.org and can be installed in both Debian stable, testing and unstable installations. Thanks, Moritz Hi, I try with .32 kernel and with that the situation sounds better. No new channel appear but i can record with the digital one. There is another issue anyway the volume is too low. I can just hear my voice, as if it were far, only by using the microphone of my headset (with the one integrated in the laptop no because it has a too low volume) and setting almost the maximum volume of PCM, mic and digital channels. I try many different combination of volumes with no result. I hope you understand my English is not very good thanks
Bug#561880: linux-image-2.6-parisc64: inotify_add_watch does not give a unique watch descriptor
Package: linux-image-2.6-parisc64
Severity: important
Blocks: 558981
The attatched minimal test case shows these results:
on my i386 machine, it correctly gives two different numbers:
$ gcc test.c; ./a.out
First: 1
Second: 2
however on paer.debian.org (hppa):
$ gcc test.c; ./a.out
First: 1
Second: 1
Based on the manpage of inotify_add_watch, it sounds like the correct
behavior is to have different numbers. It seems that the author of
inotify-tools also interpreted it that way.
This bug is causing my package's tests to fail, which causes it to FTBFS:
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=558981
Cheers,
Ryan
--
_
Ryan Niebur
ryanrya...@gmail.com
#include sys/inotify.h
#include stdio.h
int main()
{
int fd, wd;
fd = inotify_init ();
wd = inotify_add_watch (fd,
test.c,
IN_OPEN | IN_CLOSE);
printf(First: %d\n, wd);
inotify_rm_watch(fd, wd);
wd = inotify_add_watch (fd,
a.out,
IN_OPEN | IN_CLOSE);
printf(Second: %d\n, wd);
inotify_rm_watch(fd, wd);
}
signature.asc
Description: Digital signature
Bug#550665: linux-image-2.6.30-2-amd64: ALSA: When i use a kernel most newer than 2.6.26 alsamixer section Capture is empty with the exception of the presence of the Digital channel
On Mon, 2009-12-21 at 00:08 +0100, kiroken wrote: [...] I try with .32 kernel and with that the situation sounds better. No new channel appear but i can record with the digital one. There is another issue anyway the volume is too low. I can just hear my voice, as if it were far, only by using the microphone of my headset (with the one integrated in the laptop no because it has a too low volume) and setting almost the maximum volume of PCM, mic and digital channels. I try many different combination of volumes with no result. I suggest that you run alsamixer -V capture so that you only see the recording controls. You may need to use a switch, rather than a sliding control, to boost the recording volume. For example, there may be a switch to change the input from line-in to microphone mode. I'm going to close this bug because the original problem is fixed. Ben. -- Ben Hutchings Once a job is fouled up, anything done to improve it makes it worse. signature.asc Description: This is a digitally signed message part
Processed: closing 550665
Processing commands for cont...@bugs.debian.org: close 550665 2.6.32-1 Bug#550665: linux-image-2.6.30-2-amd64: ALSA: When i use a kernel most newer than 2.6.26 alsamixer section Capture is empty with the exception of the presence of the Digital channel 'close' is deprecated; see http://www.debian.org/Bugs/Developer#closing. Bug marked as fixed in version 2.6.32-1, send any further explanations to kiroken kiro...@gmail.com End of message, stopping processing here. Please contact me if you need assistance. Debian bug tracking system administrator (administrator, Debian Bugs database) -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#561880: linux-image-2.6-parisc64: inotify_add_watch does not give a unique watch descriptor
On Sun, 2009-12-20 at 15:46 -0800, Ryan Niebur wrote: Package: linux-image-2.6-parisc64 Yes but which version? Severity: important Blocks: 558981 The attatched minimal test case shows these results: on my i386 machine, it correctly gives two different numbers: $ gcc test.c; ./a.out First: 1 Second: 2 however on paer.debian.org (hppa): $ gcc test.c; ./a.out First: 1 Second: 1 Based on the manpage of inotify_add_watch, it sounds like the correct behavior is to have different numbers. It seems that the author of inotify-tools also interpreted it that way. This bug is causing my package's tests to fail, which causes it to FTBFS: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=558981 I think the test case is broken; why should two watches that don't exist at the same time have unique ids? You might as well test: fd1 = open(foo, O_RDONLY); close(fd1); fd2 = open(bar, O_RDONLY); assert(fd2 != fd1); Ben. -- Ben Hutchings Once a job is fouled up, anything done to improve it makes it worse. signature.asc Description: This is a digitally signed message part
Bug#561880: linux-image-2.6-parisc64: inotify_add_watch does not give a unique watch descriptor
Hi Ben, thanks for the response. On Mon, Dec 21, 2009 at 12:12:44AM +, Ben Hutchings wrote: On Sun, 2009-12-20 at 15:46 -0800, Ryan Niebur wrote: Package: linux-image-2.6-parisc64 Yes but which version? the version that paer is running (apparently linux-image-2.6.32-trunk-parisc64 2.6.32-1) and whatever version peri is running. Severity: important Blocks: 558981 The attatched minimal test case shows these results: on my i386 machine, it correctly gives two different numbers: $ gcc test.c; ./a.out First: 1 Second: 2 however on paer.debian.org (hppa): $ gcc test.c; ./a.out First: 1 Second: 1 Based on the manpage of inotify_add_watch, it sounds like the correct behavior is to have different numbers. It seems that the author of inotify-tools also interpreted it that way. This bug is causing my package's tests to fail, which causes it to FTBFS: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=558981 I think the test case is broken; why should two watches that don't exist at the same time have unique ids? You might as well test: the manpage makes it sound like it should be unique for each filename. and if the test case is broken, then I don't understand why this only fails on hppa. fd1 = open(foo, O_RDONLY); close(fd1); fd2 = open(bar, O_RDONLY); assert(fd2 != fd1); and you would expect this assertion to fail on all architectures, yes? however my test only fails on hppa. Cheers, Ryan -- _ Ryan Niebur ryanrya...@gmail.com signature.asc Description: Digital signature
Processed: retitle 561880 to inotify/hppa: Not assigning sequential ids as expected
Processing commands for cont...@bugs.debian.org: retitle 561880 inotify/hppa: Not assigning sequential ids as expected Bug #561880 [linux-image-2.6-parisc64] linux-image-2.6-parisc64: inotify_add_watch does not give a unique watch descriptor Changed Bug title to 'inotify/hppa: Not assigning sequential ids as expected' from 'linux-image-2.6-parisc64: inotify_add_watch does not give a unique watch descriptor' End of message, stopping processing here. Please contact me if you need assistance. Debian bug tracking system administrator (administrator, Debian Bugs database) -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Processed: reassign 561880 to linux-2.6
Processing commands for cont...@bugs.debian.org: reassign 561880 linux-2.6 2.6.32-1 Bug #561880 [linux-image-2.6-parisc64] inotify/hppa: Not assigning sequential ids as expected Bug reassigned from package 'linux-image-2.6-parisc64' to 'linux-2.6'. Bug #561880 [linux-2.6] inotify/hppa: Not assigning sequential ids as expected There is no source info for the package 'linux-2.6' at version '2.6.32-1' with architecture '' Unable to make a source version for version '2.6.32-1' Bug Marked as found in versions 2.6.32-1. End of message, stopping processing here. Please contact me if you need assistance. Debian bug tracking system administrator (administrator, Debian Bugs database) -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#561880: linux-image-2.6-parisc64: inotify_add_watch does not give a unique watch descriptor
On Sun, 2009-12-20 at 16:35 -0800, Ryan Niebur wrote: Hi Ben, thanks for the response. On Mon, Dec 21, 2009 at 12:12:44AM +, Ben Hutchings wrote: On Sun, 2009-12-20 at 15:46 -0800, Ryan Niebur wrote: Package: linux-image-2.6-parisc64 Yes but which version? the version that paer is running (apparently linux-image-2.6.32-trunk-parisc64 2.6.32-1) Wow, already? OK. and whatever version peri is running. Severity: important Blocks: 558981 The attatched minimal test case shows these results: on my i386 machine, it correctly gives two different numbers: $ gcc test.c; ./a.out First: 1 Second: 2 however on paer.debian.org (hppa): $ gcc test.c; ./a.out First: 1 Second: 1 Based on the manpage of inotify_add_watch, it sounds like the correct behavior is to have different numbers. It seems that the author of inotify-tools also interpreted it that way. This bug is causing my package's tests to fail, which causes it to FTBFS: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=558981 I think the test case is broken; why should two watches that don't exist at the same time have unique ids? You might as well test: the manpage makes it sound like it should be unique for each filename. and if the test case is broken, then I don't understand why this only fails on hppa. I'm saying there is no guarantee in the manual page about what order watch ids will be assigned in, and that it is unreasonable to expect unique to mean unique among all watches that ever existed for this handle. The assignments could vary depending on the architecture or on the phase of the moon. The test cases should not make any assertions about them, other than that any two watches *that exist at the same time* have different ids. However, it does look like the implementation is intended to assign sequential ids, so this may indicate a bug in the implementation. But it shouldn't block you. fd1 = open(foo, O_RDONLY); close(fd1); fd2 = open(bar, O_RDONLY); assert(fd2 != fd1); and you would expect this assertion to fail on all architectures, yes? however my test only fails on hppa. In a single-threaded program, yes. Otherwise, who knows? But open() is specified to use the lowest currently unused entry in the file descriptor table, whereas inotify_add_watch() can use any currently unused id - including the one just removed. Ben. -- Ben Hutchings Once a job is fouled up, anything done to improve it makes it worse. signature.asc Description: This is a digitally signed message part
Processed: reassign 561309 to linux-2.6
Processing commands for cont...@bugs.debian.org:
# Should not be a hard dependency in the driver
reassign 561309 linux-2.6 2.6.32~rc8-1~experimental.1
Bug #561309 [firmware-linux-nonfree] firmware-linux-nonfree: needs firmware for
module r8169 (/rtl8168d-{1, 2}.fw)
Bug reassigned from package 'firmware-linux-nonfree' to 'linux-2.6'.
Bug No longer marked as found in versions firmware-nonfree/0.21 and
firmware-nonfree/0.22.
Bug #561309 [linux-2.6] firmware-linux-nonfree: needs firmware for module r8169
(/rtl8168d-{1, 2}.fw)
There is no source info for the package 'linux-2.6' at version
'2.6.32~rc8-1~experimental.1' with architecture ''
Unable to make a source version for version '2.6.32~rc8-1~experimental.1'
Bug Marked as found in versions 2.6.32~rc8-1~experimental.1.
End of message, stopping processing here.
Please contact me if you need assistance.
Debian bug tracking system administrator
(administrator, Debian Bugs database)
--
To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Bits from the kernel team
On Mon, Dec 21, 2009 at 12:50 AM, Ben Hutchings b...@decadent.org.uk wrote: OSS --- This has been a deprecated kernel interface for some time and will be disabled for squeeze Done. with mechanisms put in place to deal with legacy users. Er, not sure. I guess oss4-dkms will be enough to take care of these users, hopefully it will reach squeeze in time. -- bye, pabs http://wiki.debian.org/PaulWise -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Processed: tagging 561309
Processing commands for cont...@bugs.debian.org:
# Automatically generated email from bts, devscripts version 2.10.35lenny7
tags 561309 + pending
Bug #561309 [linux-2.6] firmware-linux-nonfree: needs firmware for module r8169
(/rtl8168d-{1, 2}.fw)
Added tag(s) pending.
End of message, stopping processing here.
Please contact me if you need assistance.
Debian bug tracking system administrator
(administrator, Debian Bugs database)
--
To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#561890: lenny kernel memory leak?
Package: linux-2.6 Version: 2.6.26-19lenny2 Severity: important Forwarded Message From: Mikko Rapeli mikko.rap...@iki.fi To: debian-kernel@lists.debian.org Subject: lenny kernel memory leak? Date: Tue, 15 Dec 2009 22:36:56 +0200 Newsgroups: gmane.linux.debian.devel.kernel (maybe debian-kernel is a better place to ask than debian-user) I have an old home gateway with 32 MB of ram running email, apache 1.3 with some static files and an rt2500 based ad-hoc mode WLAN AP with dnsmasq. Since upgrade from etch to lenny, I've seen a number of OOM situations and slow downs after a few days of uptime. Is there some perhaps known memory leak in lenny 2.6.26 kernels, perhaps in rt2500 driver? In ad-hoc mode the rt2500 driver very chatty about so I disabled klogd for now, but AFAIK kernel uses a ring buffer which should not slow anything down. Here's a diff of /proc/meminfo after boot and when seeing a dramatic slow down after a few days of uptime. I think the anonpages difference shows a memory leak in kernel, am I correct? -Mikko ps. please cc me in replies, thanks. --- meminfo_after_boot.txt 2009-12-14 14:28:25.0 +0200 +++ meminfo_slow.txt2009-12-14 14:24:01.0 +0200 @@ -1,28 +1,28 @@ MemTotal:29228 kB -MemFree: 1172 kB -Buffers: 552 kB -Cached: 7488 kB -SwapCached:404 kB -Active: 17292 kB -Inactive: 2632 kB +MemFree: 1392 kB +Buffers: 168 kB +Cached: 1744 kB +SwapCached:736 kB +Active: 1360 kB +Inactive: 1348 kB SwapTotal: 204792 kB -SwapFree: 195728 kB +SwapFree: 164480 kB Dirty: 0 kB -Writeback: 0 kB -AnonPages: 11652 kB -Mapped: 5132 kB -Slab: 4276 kB -SReclaimable: 1108 kB -SUnreclaim: 3168 kB -PageTables:556 kB +Writeback: 260 kB +AnonPages: 384 kB +Mapped:928 kB +Slab:21048 kB +SReclaimable: 952 kB +SUnreclaim: 20096 kB +PageTables:688 kB NFS_Unstable:0 kB Bounce: 0 kB WritebackTmp:0 kB CommitLimit:219404 kB -Committed_AS:33172 kB +Committed_AS:53648 kB VmallocTotal: 1007300 kB VmallocUsed: 2500 kB -VmallocChunk: 1004556 kB +VmallocChunk: 1004624 kB HugePages_Total: 0 HugePages_Free: 0 HugePages_Rsvd: 0 -- Ben Hutchings Once a job is fouled up, anything done to improve it makes it worse. -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Bits from the kernel team
On Mon, 2009-12-21 at 09:24 +0800, Paul Wise wrote: On Mon, Dec 21, 2009 at 12:50 AM, Ben Hutchings b...@decadent.org.uk wrote: OSS --- This has been a deprecated kernel interface for some time and will be disabled for squeeze Done. with mechanisms put in place to deal with legacy users. Er, not sure. I guess oss4-dkms will be enough to take care of these users, hopefully it will reach squeeze in time. Hopefully not. OSS4 on Linux is part of the problem, not part of the solution. Linux applications should not use /dev/dsp any more. Those that do may be handled by some kind of bridge to ALSA, which was what this item refers to. (The existing snd-pcm-oss and snd-mixer-oss don't seem to be good enough.) Ben. -- Ben Hutchings Once a job is fouled up, anything done to improve it makes it worse. signature.asc Description: This is a digitally signed message part
Bug#561890: lenny kernel memory leak?
Forwarded Message From: Mikko Rapeli mikko.rap...@iki.fi To: Debian kernel team debian-kernel@lists.debian.org Subject: Re: lenny kernel memory leak? Date: Mon, 21 Dec 2009 00:53:35 +0200 On Sun, Dec 20, 2009 at 04:30:54PM +, Ben Hutchings wrote: Try c95edf5432f097c926dd3f59239ecde80da3b214 rt2x00: Properly clean up beacon skbs. With this I got kernel panics at boot when loading the modules. Maybe it depends on some other patches. http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=e7087a828f8714e464fff18d93618727530dfd89;hp=5adf6d63c1697ce1835daf2b5393488a71ee0dca This is now in. Will see if it makes any difference in a few days. -Mikko -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Processed: submitter 561890
Processing commands for cont...@bugs.debian.org: submitter 561890 Mikko Rapeli mikko.rap...@iki.fi Bug #561890 [linux-2.6] lenny kernel memory leak? Changed Bug submitter to 'Mikko Rapeli mikko.rap...@iki.fi' from 'Ben Hutchings b...@decadent.org.uk' End of message, stopping processing here. Please contact me if you need assistance. Debian bug tracking system administrator (administrator, Debian Bugs database) -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#561776: linux: ath5k no longer works with kernel since 2.6.30
On Sun, 2009-12-20 at 09:54 +0100, samuel wrote: Package: linux-2.6 Version: 2.6.31-2 Severity: important File: linux Since i upgraded to 2.6.31 (and 2.6.32) my wireless no longer succeeds in connecting to my AP. Symptons are diverse: * i can connect but get disconnected immediately after which scanning starts again * wicd can't start scanning * wicd can't find networks Booting back to 2.6.30 'fixes' the issue. I've messed with rfkill too and that seems to help everynow and then (but mostly not and is very inconvenient) rmmod ath5k; rfkill block all; rfkill unblock all; modprobe ath5k; rfkill unblock all; iwconfig wlan0 up Please send the kernel messages (output from dmesg) from 2.6.31/2.6.32, as requested by the bug script. Ben. -- Ben Hutchings Once a job is fouled up, anything done to improve it makes it worse. signature.asc Description: This is a digitally signed message part
Bug#561309: needs firmware for module r8169 (/rtl8168d-{1, 2}.fw)
On Sat, 2009-12-19 at 00:52 -0500, Dave Witbrodt wrote: Package: firmware-linux-nonfree Version: 0.22 Severity: normal Ben, your patch works for me. Will this be made part of the Debian 2.6.32 kernel, or were you merely experimenting? (It would really help if the KT used this, unless/until the r8169 firmware can be used in Debian.) [...] This patch will be in the next version (2.6.32-3). Ben. -- Ben Hutchings Once a job is fouled up, anything done to improve it makes it worse. signature.asc Description: This is a digitally signed message part
Bug#561309: needs firmware for module r8169 (/rtl8168d-{1, 2}.fw)
On Thu, 2009-12-17 at 09:32 -0500, Dave Witbrodt wrote: [...] Found some time before work this morning. Find the last series of messages attached, including the kernel oops. I had to manually type this, so I apologize in advance for typos -- I did try to make corrections before rebooting, but no doubt I missed some stuff. [...] This bug will also be fixed by the change to make firmware optional. Ben. -- Ben Hutchings Once a job is fouled up, anything done to improve it makes it worse. signature.asc Description: This is a digitally signed message part
Bug#561764: some warnings fly off screen and are not saved in any file
BH These messages remain in the kernel log buffer and should be copied
BH by the log daemon later.
Ah ha, I can reproduce the messages via
/etc/init.d/udev restart
and indeed they also end up in the /var/log/ files.
(They are complaining about some things in /etc/udev/rules.d/*, but that
is besides the point (that they don't get logged at boot.))
So then I rebooted (regular boot,
/proc/cmdline:
BOOT_IMAGE=/boot/vmlinuz-2.6.32-trunk-686 root=/dev/sda1 ro quiet panic=15)
There those udev warnings go zipping by.
And now that everything is up and running in multi-user mode, I check
the logs, and ah ha, the only warnings that were logged are the ones
from when I did /etc/init.d/udev restart. There are none corresponding
to the time right after reboot. Nor say several hours off due to
timezone settings or anything.
(I don't believe quiet above means less will go to the logs than goes
to the screen.)
10:49 log# find -mtime -1 -type f|xargs egrep -h SYSFS\|reboot|colrm 77|sort -u
Dec 21 10:43:10 jidanni3 udevd[15039]: SYSFS{}= will be removed in a future
Dec 21 10:45:14 jidanni3 shutdown[15063]: shutting down for system reboot
Dec 21 10:45:58 jidanni3 /usr/sbin/cron[1178]: (CRON) INFO (Running @reboot
Maybe some logging is being done, but it is at that time ending up on a
tmpfs that one sees in /etc/init.d/udev, that gets mounted over with
something else, but that is over my head.
--
To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#561894: linux-image-2.6.32-trunk-686: ata reset during moderate load
Package: linux-2.6
Version: 2.6.32-2
Severity: normal
The machine had an uptime of 5 days. Had been suspended (STR) and
resumed many times. Today morning, I tried to preview a half-downloaded
torrent file, which led in too much of disk I/O taking place.
This led to the following ata reset followed by the RO remount of the
ext4 file system.
[476862.000125] ata1.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action
0x6 frozen
[476862.000137] ata1.00: failed command: FLUSH CACHE EXT
[476862.000154] ata1.00: cmd ea/00:00:00:00:00/00:00:00:00:00/a0 tag 0
[476862.000158] res 40/00:00:00:4f:c2/00:00:00:00:00/00 Emask
0x4 (timeout)
[476862.000166] ata1.00: status: { DRDY }
[476867.040082] ata1: link is slow to respond, please be patient
(ready=0)
[476872.024036] ata1: device not ready (errno=-16), forcing hardreset
[476872.024050] ata1: soft resetting link
[476877.220085] ata1: link is slow to respond, please be patient
(ready=0)
[476880.396601] ata1.00: configured for UDMA/100
[476880.396612] ata1.00: device reported invalid CHS sector 0
[476880.396625] end_request: I/O error, dev sda, sector 0
[476880.396650] ata1: EH complete
[476880.445232] Aborting journal on device dm-0-8.
[476880.690442] EXT4-fs error (device dm-0): ext4_journal_start_sb:
Detected aborted journal
[476880.690451] EXT4-fs (dm-0): Remounting filesystem read-only
[476881.064105] EXT4-fs error (device dm-0) in ext4_reserve_inode_write:
Journal has aborted
[476881.065448] journal commit I/O error
[477058.373151] EXT4-fs error (device dm-0): ext4_remount: Abort forced
by user
Regards,
Ritesh
-- Package-specific info:
** Version:
Linux version 2.6.32-trunk-686 (Debian 2.6.32-2) (b...@decadent.org.uk) (gcc
version 4.3.4 (Debian 4.3.4-6) ) #1 SMP Thu Dec 17 00:26:04 UTC 2009
** Command line:
BOOT_IMAGE=//vmlinuz-2.6.32-trunk-686 root=/dev/mapper/VgSd-ROOT ro selinux=1
audit=1 enforcing=0 quiet vga=788
** Not tainted
** Kernel log:
[ 10.55] thinkpad_acpi: WARNING: This firmware may be missing critical
bug fixes and/or important features
[ 10.595560] Registered led device: tpacpi::thinklight
[ 10.595689] Registered led device: tpacpi::power
[ 10.595792] Registered led device: tpacpi::standby
[ 10.600495] thinkpad_acpi: fan_init: initial fan status is unknown, assuming
it is in auto mode
[ 10.601270] input: ThinkPad Extra Buttons as /devices/virtual/input/input5
[ 10.729138] i801_smbus :00:1f.3: PCI INT A - GSI 23 (level, low) - IRQ
23
[ 10.954399] input: PC Speaker as /devices/platform/pcspkr/input/input6
[ 10.968980] yenta_cardbus :0b:00.0: CardBus bridge found [1014:056c]
[ 11.097270] yenta_cardbus :0b:00.0: ISA IRQ mask 0x0cb0, PCI irq 16
[ 11.097279] yenta_cardbus :0b:00.0: Socket status: 3006
[ 11.097294] yenta_cardbus :0b:00.0: pcmcia: parent PCI bridge I/O
window: 0x5000 - 0x8fff
[ 11.097302] pcmcia_socket pcmcia_socket0: cs: IO port probe 0x5000-0x8fff:
clean.
[ 11.099551] yenta_cardbus :0b:00.0: pcmcia: parent PCI bridge Memory
window: 0xb400 - 0xbfff
[ 11.099559] yenta_cardbus :0b:00.0: pcmcia: parent PCI bridge Memory
window: 0xd000 - 0xd7ff
[ 11.123298] lib80211: common routines for IEEE802.11 drivers
[ 11.123306] lib80211_crypt: registered algorithm 'NULL'
[ 11.140520] parport_pc 00:0a: reported by Plug and Play ACPI
[ 11.140585] parport0: PC-style at 0x3bc, irq 7 [PCSPP,TRISTATE]
[ 11.177053] ieee80211: 802.11 data/management/control stack, git-1.1.13
[ 11.177060] ieee80211: Copyright (C) 2004-2005 Intel Corporation
jketr...@linux.intel.com
[ 11.247342] ipw2200: Intel(R) PRO/Wireless 2200/2915 Network Driver,
1.2.2kmprq
[ 11.247349] ipw2200: Copyright(c) 2003-2006 Intel Corporation
[ 11.247456] ipw2200 :0b:02.0: PCI INT A - GSI 21 (level, low) - IRQ 21
[ 11.251933] ipw2200: Detected Intel PRO/Wireless 2200BG Network Connection
[ 11.252060] ipw2200 :0b:02.0: firmware: requesting ipw2200-bss.fw
[ 11.789847] pcmcia_socket pcmcia_socket0: cs: IO port probe 0x100-0x3af:
clean.
[ 11.802279] pcmcia_socket pcmcia_socket0: cs: IO port probe 0x3e0-0x4ff:
excluding 0x4d0-0x4d7
[ 11.808985] pcmcia_socket pcmcia_socket0: cs: IO port probe 0x820-0x8ff:
clean.
[ 11.816772] pcmcia_socket pcmcia_socket0: cs: IO port probe 0xc00-0xcf7:
clean.
[ 11.828285] pcmcia_socket pcmcia_socket0: cs: IO port probe 0xa00-0xaff:
clean.
[ 11.913780] ipw2200: Detected geography ZZR (14 802.11bg channels, 0 802.11a
channels)
[ 11.961458] Synaptics Touchpad, model: 1, fw: 5.9, id: 0x2c6ab1, caps:
0x884793/0x0
[ 11.961469] serio: Synaptics pass-through port at isa0060/serio1/input0
[ 11.983834] Intel ICH :00:1e.2: PCI INT A - GSI 22 (level, low) - IRQ
22
[ 11.983917] Intel ICH :00:1e.2: setting latency timer to 64
[ 12.008811] input: SynPS/2 Synaptics TouchPad as
/devices/platform/i8042/serio1/input/input7
[ 12.908042] intel8x0_measure_ac97_clock: measured 55204 usecs (2660 samples)
[ 12.908049]
Bug#561764: marked as done (some warnings fly off screen and are not saved in any file)
Your message dated Mon, 21 Dec 2009 04:50:22 +0100 with message-id 20091221035022.ga17...@bongo.bofh.it and subject line Re: Bug#561764: some warnings fly off screen and are not saved in any file has caused the Debian Bug report #561764, regarding some warnings fly off screen and are not saved in any file to be marked as done. This means that you claim that the problem has been dealt with. If this is not the case it is now your responsibility to reopen the Bug report if necessary, and/or fix the problem forthwith. (NB: If you are a system administrator and have no idea what this message is talking about, this may indicate a serious mail system misconfiguration somewhere. Please contact ow...@bugs.debian.org immediately.) -- 561764: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=561764 Debian Bug Tracking System Contact ow...@bugs.debian.org with problems ---BeginMessage--- Package: linux-2.6 Version: 2.6.32-2 Severity: wishlist X-debbugs-cc: linux-ker...@vger.kernel.org There is some warning about SYS... being DEPRECATED or something like that. I can't tell as it flies off the screen at boot and is not kept in any /var/log file. I don't even know what to google for, as I can't read the message fast enough. How is one supposed to know what it says if it files off the screen, and even SHIFT PAGE UP can't reach it. How many other warnings fly away and aren't in dmesg etc.? ---End Message--- ---BeginMessage--- On Dec 21, jida...@jidanni.org wrote: Maybe some logging is being done, but it is at that time ending up on a No useful logging can really be done until /var is mounted rw and so on. Not a bug. -- ciao, Marco signature.asc Description: Digital signature ---End Message---
Processed: Re: Bug#557262: 2.6.31+2.6.31.4: XFS - All I/O locks up to D-state after 24-48 hours (sysrq-t+w available) - root cause found = asterisk
Processing commands for cont...@bugs.debian.org: reassign 557262 linux-2.6 Bug #557262 [asterisk] 2.6.31+2.6.31.4: XFS - All I/O locks up to D-state after 24-48 hours (sysrq-t+w available) - root cause found = asterisk Bug reassigned from package 'asterisk' to 'linux-2.6'. Bug No longer marked as found in versions 1.6.2.0~dfsg~rc1-1. thanks Stopping processing here. Please contact me if you need assistance. Debian bug tracking system administrator (administrator, Debian Bugs database) -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#561309: firmware-linux-nonfree: needs firmware for module r8169
Ben (et al): Your patch (with Stefan's addition, replacing dev-name with r8169) also worked for me on an Intel i7 920 system (amd64) running unstable (w/o the firmware). Dec 20 18:39:08 octane kernel: [1.100454] r8169 Gigabit Ethernet driver 2.3LK-NAPI loaded Dec 20 18:39:08 octane kernel: [1.100471] r8169 :06:00.0: PCI INT A - GSI 18 (level, low) - IRQ 18 Dec 20 18:39:08 octane kernel: [1.100960] eth0: RTL8168d/8111d at 0xc9c7a000, 00:1f:bc:08:c4:16, XID 081000c0 IRQ 30 Dec 20 18:39:08 octane kernel: [1.105168] r8169 :06:00.0: firmware: requesting rtl8168d-1.fw Dec 20 18:39:08 octane kernel: [1.106433] Failed to load rtl8168d-1.fw. ... Dec 20 18:39:08 octane kernel: [ 12.736957] r8169: eth0: link up This NIC is onboard the mobo: EVGA X58 SLI LE http://www.evga.com/products/moreInfo.asp?pn=141-BL-E757-TRfamily=Motherboard%20Family Thanks! --Tom -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
