2009年10月 1日

[vine-users:079720] AMD64 環境での起動時のメモリリークについて

東浜と申します。
いつも有用な情報をありがとうございます。

色々調べてみたのですが、お手上げとなり、アドバイスを頂けましたらと思いました…。


現在の環境は
Vine 5.0 x86_64

AMD Athlon64 X2 4450e
AMD780G+SB700チップセット
DDR2 800 2G×2
Intel 1000MT Desktop Adapter

状況は、インストール後正常に稼動していましたが、データベースのチューンのためにメモリを4Gから8Gへ増設したところ、ネットワークインターフェースがリンクアップするも非常に不安定なのに気づき、様子を見ているうちにリンクアップ・リンクダウンを繰り返すようになりました。
そこでログを見てみると…

Sep 28 07:57:53 localhost kernel: No AGP bridge found
Sep 28 07:57:53 localhost kernel: Node 0: aperture @ 20000000 size 32 MB
Sep 28 07:57:53 localhost kernel: Aperture pointing to e820 RAM. Ignoring.
Sep 28 07:57:53 localhost kernel: Your BIOS doesn't leave a aperture memory
hole
Sep 28 07:57:53 localhost kernel: Please enable the IOMMU option in the BIOS
setup
Sep 28 07:57:53 localhost kernel: This costs you 64 MB of RAM
Sep 28 07:57:53 localhost kernel: Mapping aperture over 65536 KB of RAM @
20000000
Sep 28 07:57:53 localhost kernel: PM: Registered nosave memory:
0000000020000000 - 0000000024000000
Sep 28 07:57:53 localhost kernel: Memory: 7631336k/9175040k available (2889k
kernel code, 229432k reserved, 1450k data, 400k init)
Sep 28 07:57:53 localhost kernel: SLUB: Genslabs=12, HWalign=64, Order=0-3,
MinObjects=0, CPUs=4, Nodes=1
Sep 28 07:57:53 localhost kernel: Calibrating delay loop (skipped), value
calculated using timer frequency.. 2009.28 BogoMIPS (lpj=4018568)
Sep 28 07:57:53 localhost kernel: Mount-cache hash table entries: 256

と出ており、その後にe1000がハングしているようでした。

Sep 28 07:58:50 localhost kernel: ------------[ cut here ]------------
Sep 28 07:58:50 localhost kernel: WARNING: at net/sched/sch_generic.c:219
dev_watchdog+0x136/0x1d8()
Sep 28 07:58:50 localhost kernel: NETDEV WATCHDOG: eth0 (e1000): transmit
timed out
Sep 28 07:58:50 localhost kernel: Modules linked in: autofs4 ipv6 xt_length
ipt_REJECT xt_limit ipt_LOG ipt_recent xt_tcpudp nf_conntrack_ipv4 xt_state
nf_conntrack xt_multiport iptable_filter ip_tables x_tables dm_mod thermal
k8temp processor shpchp hwmon e1000 i2c_piix4 i2c_core button pcspkr wmi sg
usb_storage ahci pata_atiixp libata dock ide_cd_mod cdrom sd_mod scsi_mod
crc_t10dif uhci_hcd ohci_hcd ehci_hcd [last unloaded: freq_table]
Sep 28 07:58:50 localhost kernel: Pid: 0, comm: swapper Not tainted
2.6.27-43vl5 #1
Sep 28 07:58:50 localhost kernel:
Sep 28 07:58:50 localhost kernel: Call Trace:
Sep 28 07:58:50 localhost kernel: <IRQ> [<ffffffff80239eb1>]
warn_slowpath+0xb4/0xe0
Sep 28 07:58:50 localhost kernel: [<ffffffff802303b8>]
enqueue_task_fair+0x22/0x44
Sep 28 07:58:50 localhost kernel: [<ffffffff8022eb50>]
enqueue_task+0x13/0x1e
Sep 28 07:58:50 localhost kernel: [<ffffffff8022ed69>]
source_load+0x2a/0x4f
Sep 28 07:58:50 localhost kernel: [<ffffffff8036b16d>] __next_cpu+0x19/0x26
Sep 28 07:58:50 localhost kernel: [<ffffffff80230f73>]
find_busiest_group+0x256/0x708
Sep 28 07:58:50 localhost kernel: [<ffffffff8022f79b>]
place_entity+0x6c/0x9a
Sep 28 07:58:50 localhost kernel: [<ffffffffa016854c>]
fib6_clean_node+0x2e/0x92 [ipv6]
Sep 28 07:58:50 localhost kernel: [<ffffffff8022edb8>]
target_load+0x2a/0x4f
Sep 28 07:58:50 localhost kernel: [<ffffffffa0167e6b>] fib6_walk+0x87/0x8d
[ipv6]
Sep 28 07:58:50 localhost kernel: [<ffffffffa0167e9e>]
fib6_clean_tree+0x2d/0x32 [ipv6]
Sep 28 07:58:50 localhost kernel: [<ffffffff804cf017>]
_spin_lock_irqsave+0x20/0x27
Sep 28 07:58:50 localhost kernel: [<ffffffff80469bba>]
dev_watchdog+0x136/0x1d8
Sep 28 07:58:50 localhost kernel: [<ffffffff802428c1>]
__mod_timer+0xc5/0xd4
Sep 28 07:58:50 localhost kernel: [<ffffffffa0167acf>] fib6_age+0x0/0x67
[ipv6]
Sep 28 07:58:50 localhost kernel: [<ffffffff80469a84>]
dev_watchdog+0x0/0x1d8
Sep 28 07:58:50 localhost kernel: [<ffffffff802421e3>]
run_timer_softirq+0x16f/0x1ec
Sep 28 07:58:50 localhost kernel: [<ffffffff8023e5e3>]
__do_softirq+0x65/0xdb
Sep 28 07:58:50 localhost kernel: [<ffffffff8021189c>]
call_softirq+0x1c/0x28
Sep 28 07:58:50 localhost kernel: [<ffffffff802139d3>] do_softirq+0x3c/0x81
Sep 28 07:58:50 localhost kernel: [<ffffffff8023e538>] irq_exit+0x3f/0x85
Sep 28 07:58:50 localhost kernel: [<ffffffff8021fc0b>]
smp_apic_timer_interrupt+0x8f/0xa8
Sep 28 07:58:50 localhost kernel: [<ffffffff802110a3>]
apic_timer_interrupt+0x83/0x90
Sep 28 07:58:50 localhost kernel: <EOI> [<ffffffff8021f9b4>]
lapic_next_event+0x0/0x13
Sep 28 07:58:50 localhost kernel: [<ffffffff80223c3c>]
native_safe_halt+0x2/0x3
Sep 28 07:58:50 localhost kernel: [<ffffffff8024f194>]
notifier_call_chain+0x29/0x4c
Sep 28 07:58:50 localhost kernel: [<ffffffff80217454>]
default_idle+0x2a/0x46
Sep 28 07:58:50 localhost kernel: [<ffffffff80217682>] c1e_idle+0x10a/0x10f
Sep 28 07:58:50 localhost kernel: [<ffffffff8020eca5>] cpu_idle+0x9e/0xc8
Sep 28 07:58:50 localhost kernel:
Sep 28 07:58:50 localhost kernel: ---[ end trace 6702b2ba99529d3c ]---
Sep 28 07:58:53 localhost kernel: e1000: eth0: e1000_watchdog: NIC Link is
Up 1000 Mbps Full Duplex, Flow Control: RX/TX
Sep 28 07:58:57 localhost kernel: e1000: eth0: e1000_clean_tx_irq: Detected
Tx Unit Hang
Sep 28 07:58:57 localhost kernel: Tx Queue <0>
Sep 28 07:58:57 localhost kernel: TDH <d7>
Sep 28 07:58:57 localhost kernel: TDT <f1>
Sep 28 07:58:57 localhost kernel: next_to_use <f1>
Sep 28 07:58:57 localhost kernel: next_to_clean <d5>
Sep 28 07:58:57 localhost kernel: buffer_info[next_to_clean]
Sep 28 07:58:57 localhost kernel: time_stamp <1000353e9>
Sep 28 07:58:57 localhost kernel: next_to_watch <d9>
Sep 28 07:58:57 localhost kernel: jiffies <100035556>
Sep 28 07:58:57 localhost kernel: next_to_watch.status <0>
Sep 28 07:58:59 localhost kernel: e1000: eth0: e1000_clean_tx_irq: Detected
Tx Unit Hang
Sep 28 07:58:59 localhost kernel: Tx Queue <0>
Sep 28 07:58:59 localhost kernel: TDH <d7>
Sep 28 07:58:59 localhost kernel: TDT <f1>
Sep 28 07:58:59 localhost kernel: next_to_use <f1>
Sep 28 07:58:59 localhost kernel: next_to_clean <d5>
Sep 28 07:58:59 localhost kernel: buffer_info[next_to_clean]
Sep 28 07:58:59 localhost kernel: time_stamp <1000353e9>
Sep 28 07:58:59 localhost kernel: next_to_watch <d9>
Sep 28 07:58:59 localhost kernel: jiffies <10003574a>
Sep 28 07:58:59 localhost kernel: next_to_watch.status <0>
Sep 28 07:59:01 localhost kernel: e1000: eth0: e1000_clean_tx_irq: Detected
Tx Unit Hang
Sep 28 07:59:01 localhost kernel: Tx Queue <0>
Sep 28 07:59:01 localhost kernel: TDH <d7>
Sep 28 07:59:01 localhost kernel: TDT <f1>
Sep 28 07:59:01 localhost kernel: next_to_use <f1>
Sep 28 07:59:01 localhost kernel: next_to_clean <d5>
Sep 28 07:59:01 localhost kernel: buffer_info[next_to_clean]
Sep 28 07:59:01 localhost kernel: time_stamp <1000353e9>
Sep 28 07:59:01 localhost kernel: next_to_watch <d9>
Sep 28 07:59:01 localhost kernel: jiffies <10003593e>
Sep 28 07:59:01 localhost kernel: next_to_watch.status <0>
Sep 28 07:59:06 localhost kernel: e1000: eth0: e1000_watchdog: NIC Link is
Up 1000 Mbps Full Duplex, Flow Control: RX/TX
Sep 28 07:59:14 localhost kernel: e1000: eth0: e1000_clean_tx_irq: Detected
Tx Unit Hang
Sep 28 07:59:14 localhost kernel: Tx Queue <0>
Sep 28 07:59:14 localhost kernel: TDH <46>
Sep 28 07:59:14 localhost kernel: TDT <7b>
Sep 28 07:59:14 localhost kernel: next_to_use <7b>
Sep 28 07:59:14 localhost kernel: next_to_clean <43>
Sep 28 07:59:14 localhost kernel: buffer_info[next_to_clean]
Sep 28 07:59:14 localhost kernel: time_stamp <100036329>
Sep 28 07:59:14 localhost kernel: next_to_watch <46>
Sep 28 07:59:14 localhost kernel: jiffies <1000365f0>
Sep 28 07:59:14 localhost kernel: next_to_watch.status <0>
Sep 28 07:59:16 localhost kernel: e1000: eth0: e1000_clean_tx_irq: Detected
Tx Unit Hang
Sep 28 07:59:16 localhost kernel: Tx Queue <0>
Sep 28 07:59:16 localhost kernel: TDH <46>
Sep 28 07:59:16 localhost kernel: TDT <7b>
Sep 28 07:59:16 localhost kernel: next_to_use <7b>
Sep 28 07:59:16 localhost kernel: next_to_clean <43>
Sep 28 07:59:16 localhost kernel: buffer_info[next_to_clean]
Sep 28 07:59:16 localhost kernel: time_stamp <100036329>
Sep 28 07:59:16 localhost kernel: next_to_watch <46>
Sep 28 07:59:16 localhost kernel: jiffies <1000367e4>
Sep 28 07:59:16 localhost kernel: next_to_watch.status <0>
Sep 28 07:59:18 localhost kernel: e1000: eth0: e1000_clean_tx_irq: Detected
Tx Unit Hang
Sep 28 07:59:18 localhost kernel: Tx Queue <0>
Sep 28 07:59:18 localhost kernel: TDH <46>
Sep 28 07:59:18 localhost kernel: TDT <7b>
Sep 28 07:59:18 localhost kernel: next_to_use <7b>
Sep 28 07:59:18 localhost kernel: next_to_clean <43>
Sep 28 07:59:18 localhost kernel: buffer_info[next_to_clean]
Sep 28 07:59:18 localhost kernel: time_stamp <100036329>
Sep 28 07:59:18 localhost kernel: next_to_watch <46>
Sep 28 07:59:18 localhost kernel: jiffies <1000369d8>
Sep 28 07:59:18 localhost kernel: next_to_watch.status <0>
Sep 28 07:59:20 localhost kernel: e1000: eth0: e1000_clean_tx_irq: Detected
Tx Unit Hang
Sep 28 07:59:20 localhost kernel: Tx Queue <0>
Sep 28 07:59:20 localhost kernel: TDH <46>
Sep 28 07:59:20 localhost kernel: TDT <7b>
Sep 28 07:59:20 localhost kernel: next_to_use <7b>
Sep 28 07:59:20 localhost kernel: next_to_clean <43>
Sep 28 07:59:20 localhost kernel: buffer_info[next_to_clean]
Sep 28 07:59:20 localhost kernel: time_stamp <100036329>
Sep 28 07:59:20 localhost kernel: next_to_watch <46>
Sep 28 07:59:20 localhost kernel: jiffies <100036bcc>
Sep 28 07:59:20 localhost kernel: next_to_watch.status <0>
Sep 28 07:59:24 localhost kernel: e1000: eth0: e1000_watchdog: NIC Link is
Up 1000 Mbps Full Duplex, Flow Control: RX/TX
Sep 28 07:59:44 localhost kernel: e1000: eth0: e1000_clean_tx_irq: Detected
Tx Unit Hang
Sep 28 07:59:44 localhost kernel: Tx Queue <0>
Sep 28 07:59:44 localhost kernel: TDH <c9>
Sep 28 07:59:44 localhost kernel: TDT <17>
Sep 28 07:59:44 localhost kernel: next_to_use <17>
Sep 28 07:59:44 localhost kernel: next_to_clean <c6>
Sep 28 07:59:44 localhost kernel: buffer_info[next_to_clean]
Sep 28 07:59:44 localhost kernel: time_stamp <1000381a9>
Sep 28 07:59:44 localhost kernel: next_to_watch <c9>
Sep 28 07:59:44 localhost kernel: jiffies <10003833c>
Sep 28 07:59:44 localhost kernel: next_to_watch.status <0>
Sep 28 07:59:46 localhost kernel: e1000: eth0: e1000_clean_tx_irq: Detected
Tx Unit Hang
Sep 28 07:59:46 localhost kernel: Tx Queue <0>
Sep 28 07:59:46 localhost kernel: TDH <c9>
Sep 28 07:59:46 localhost kernel: TDT <17>
Sep 28 07:59:46 localhost kernel: next_to_use <17>
Sep 28 07:59:46 localhost kernel: next_to_clean <c6>
Sep 28 07:59:46 localhost kernel: buffer_info[next_to_clean]
Sep 28 07:59:46 localhost kernel: time_stamp <1000381a9>
Sep 28 07:59:46 localhost kernel: next_to_watch <c9>
Sep 28 07:59:46 localhost kernel: jiffies <100038530>
Sep 28 07:59:46 localhost kernel: next_to_watch.status <0>
Sep 28 07:59:48 localhost kernel: e1000: eth0: e1000_clean_tx_irq: Detected
Tx Unit Hang
Sep 28 07:59:48 localhost kernel: Tx Queue <0>
Sep 28 07:59:48 localhost kernel: TDH <c9>
Sep 28 07:59:48 localhost kernel: TDT <17>
Sep 28 07:59:48 localhost kernel: next_to_use <17>
Sep 28 07:59:48 localhost kernel: next_to_clean <c6>
Sep 28 07:59:48 localhost kernel: buffer_info[next_to_clean]
Sep 28 07:59:48 localhost kernel: time_stamp <1000381a9>
Sep 28 07:59:48 localhost kernel: next_to_watch <c9>
Sep 28 07:59:48 localhost kernel: jiffies <100038724>
Sep 28 07:59:48 localhost kernel: next_to_watch.status <0>
Sep 28 07:59:52 localhost kernel: e1000: eth0: e1000_watchdog: NIC Link is
Up 1000 Mbps Full Duplex, Flow Control: RX/TX


Please enable the IOMMU option in the BIOS setup
で調べて見ますと、カーネルのメモリリークが問題ではないか、との事で、カーネルオプションに
iommu=soft であるとか iommu=off だとか渡して回避するしかないようでしたので、その通り試してみましたが改善されませんでした…。

BIOSでオンボードAperture Sizeを変更してみましたが、いずれも駄目でした。
(このマザーボードではAuto 125M 256M 512Mモードがあります)
しかし、メモリを8Gから4Gへ戻すと何事もなかったように安定してリンクアップします…。

他、BIOSを最新のバージョンへ変更しましたが、これも変わりはありませんでした。
また、K10.5のCPUではメモリコントローラ周りが変わったとの事なので、物は試しとAthlon2
240eへ変えてみましたが、そこは拙い考えで解決はしませんでした。

長文で申し訳ありませんが、どなたかアドバイスを頂けましたら助かります…宜しくお願い致します。


投稿者 xml-rpc : 2009年10月 1日 01:23
役に立ちました?:
過去のフィードバック 平均:(0) 総合:(0) 投票回数:(0)
本記事へのTrackback: http://hoop.euqset.org/blog/mt-tb2006.cgi/89070
トラックバック
コメント
コメントする




画像の中に見える文字を入力してください。