Dotaz: USB 2.0 radic a pametova zarizeni

[Eregon]

Zdravim vsechny, mam malou potiz s USB 2.0 radicem EU-CV4N (PCI karta se ctyrmi USB 2.0 porty, detaily zde.)

CDMA modem pripojeny pres tento radic funguje bez potizi, problem nastava, kdyz pripojim USB Flash disk nebo ctecku pametovych karet. Po provedeni mount se normalne na flash/kartu dostanu, ale po nejake dobe se zarizeni resetuje a ja mam po ptakach (viz vypisy nize). Obe zarizeni (flash disk i ctecka pametovych karet) ale funguji spravne, kdyz je pripojim do USB 1.1 portu (vyvedeny primo z motherboardu). Obe zarizeni ale USB 2.0 podporuji.

Pri prohledavani diskusi a googlovani jsem nasel pomerne dost informaci, ale mam dojem ze nejsou relevantni pro muj pripad. Jedine, co myslim, ze by mi mohlo pomoci, je nastaveni CONFIG_BLK_DEV_UB parametru jadra, ovsem neprisel jsem na to, jestli se to da provest bez kompilace jadra noveho (zapisem hodnoty do souboru nekde v /proc ???)

Chci Vas tedy poprosit o nejake rady, alespon nejaky tip, abych se mel od ceho odrazit pri dalsim zkoumani.

Jeste pridavam vypisy (je jich pozehnane: lsmod, lspci, odpovidajici casti fstab a syslog). Na vypisech syslogu je zajimava jeste jedna vec: cast tykajici se pripojeni flashdisku na USB 2.0 - musel jsem najit vcerejsi pokus, protoze kdyz ted prave flashdisk pripojim na USB 2.0 radic, v syslogu se to vubec nijak neprojevuje. Uz jsem ho dnes (tzn. od posledniho startu) jednou pripojeny mel, ale nezkousel jsem ho ani mountnout.

[root@masterlord etc]# lsmod
Module                  Size  Used by
isofs                  31352  1 
nls_iso8859-2           4224  1 
udf                    86340  0 
nls_iso8859-1           3680  0 
nls_cp437               5312  0 
vfat                   11168  0 
fat                    39776  1 vfat
sd_mod                 19232  0 
usb-storage            65504  0 
scsi_mod              104044  2 sd_mod,usb-storage
bsd_comp                5152  0 
ppp_async               8896  1 
crc-ccitt               1664  1 ppp_async
ppp_generic            24468  6 bsd_comp,ppp_async
slhc                    6112  1 ppp_generic
vmnet                  24912  0 
parport_pc             30976  0 
parport                33896  1 parport_pc
vmmon                  44792  0 
nfsd                  189408  8 
exportfs                4736  1 nfsd
lockd                  58344  2 nfsd
sunrpc                127652  2 nfsd,lockd
md5                     3584  1 
ipv6                  230916  18 
snd-seq-oss            31232  0 
snd-seq-midi-event      6080  1 snd-seq-oss
snd-seq                47440  4 snd-seq-oss,snd-seq-midi-event
snd-pcm-oss            49480  0 
snd-mixer-oss          17376  2 snd-pcm-oss
snd-via82xx            22372  2 
snd-ac97-codec         69392  1 snd-via82xx
snd-pcm                81800  2 snd-pcm-oss,snd-via82xx
snd-timer              20356  2 snd-seq,snd-pcm
snd-page-alloc          7400  2 snd-via82xx,snd-pcm
gameport                3328  1 snd-via82xx
snd-mpu401-uart         5856  1 snd-via82xx
snd-rawmidi            19300  1 snd-mpu401-uart
snd-seq-device          6344  3 snd-seq-oss,snd-seq,snd-rawmidi
snd                    45988  13 snd-seq-oss,snd-seq,snd-pcm-oss,snd-mixer-oss,snd-via82xx,snd-ac97-codec,snd-pcm,snd-timer,snd-mpu401-uart,snd-rawmidi,snd-seq-device
soundcore               7008  2 snd
af_packet              16072  0 
floppy                 55088  0 
ipt_MASQUERADE          2720  1 
iptable_nat            21164  2 ipt_MASQUERADE
ip_conntrack           28616  2 ipt_MASQUERADE,iptable_nat
ipt_LOG                 5568  2 
iptable_filter          2080  1 
ip_tables              14624  4 ipt_MASQUERADE,iptable_nat,ipt_LOG,iptable_filter
cdc-acm                 9696  2 
r8169                  14180  0 
ide-cd                 37280  1 
cdrom                  37724  1 ide-cd
loop                   12520  0 
supermount             34804  2 
via-agp                 7360  1 
agpgart                27752  1 via-agp
ehci-hcd               26244  0 
uhci-hcd               28752  0 
usbcore               103172  6 usb-storage,cdc-acm,ehci-hcd,uhci-hcd
ext3                  120680  4 
jbd                    49080  1 ext3

[root@masterlord etc]# lspci
00:00.0 Host bridge: VIA Technologies, Inc. VT8366/A/7 [Apollo KT266/A/333]
00:01.0 PCI bridge: VIA Technologies, Inc. VT8366/A/7 [Apollo KT266/A/333 AGP]
00:09.0 USB Controller: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller (rev 61)
00:09.1 USB Controller: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller (rev 61)
00:09.2 USB Controller: VIA Technologies, Inc. USB 2.0 (rev 63)
00:0a.0 Communication controller: Conexant HCF 56k Data/Fax/Voice Modem (rev 08)
00:0c.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL-8169 Gigabit Ethernet (rev 10)
00:11.0 ISA bridge: VIA Technologies, Inc. VT8233A ISA Bridge
00:11.1 IDE interface: VIA Technologies, Inc. VT82C586A/B/VT82C686/A/B/VT823x/A/C PIPC Bus Master IDE (rev 06)
00:11.2 USB Controller: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller (rev 23)
00:11.3 USB Controller: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller (rev 23)
00:11.5 Multimedia audio controller: VIA Technologies, Inc. VT8233/A/8235/8237 AC97 Audio Controller (rev 40)
01:00.0 VGA compatible controller: nVidia Corporation NV17 [GeForce4 MX 440] (rev a3)

[root@masterlord etc]# cat fstab
...
/dev/sda /mnt/flash vfat user,noauto,noatime  
...

Ted relevantni casti syslogu. Nejprve tesne po pripojeni flashdisku k USB 1.1:

May 25 17:30:23 masterlord kernel: usb 3-1: new full speed USB device using address 4
May 25 17:30:23 masterlord kernel: usb 3-1: not running at top speed; connect to a high speed hub
May 25 17:30:23 masterlord kernel: hub 3-1:1.0: USB hub found
May 25 17:30:23 masterlord kernel: hub 3-1:1.0: 1 port detected
May 25 17:30:24 masterlord kernel: usb 3-1.1: new full speed USB device using address 5
May 25 17:30:24 masterlord kernel: usb 3-1.1: not running at top speed; connect to a high speed hub
May 25 17:30:24 masterlord kernel: scsi2 : SCSI emulation for USB Mass Storage devices
May 25 17:30:24 masterlord kernel:   Vendor: USB 2.0   Model: Flash Disk        Rev: PROL
May 25 17:30:24 masterlord kernel:   Type:   Direct-Access                      ANSI SCSI revision: 02
May 25 17:30:24 masterlord kernel: SCSI device sda: 256000 512-byte hdwr sectors (131 MB)
May 25 17:30:25 masterlord kernel: sda: Write Protect is off
May 25 17:30:25 masterlord kernel: sda: Mode Sense: 00 06 00 00
May 25 17:30:25 masterlord kernel: sda: assuming drive cache: write through
May 25 17:30:25 masterlord kernel:  /dev/scsi/host2/bus0/target0/lun0: unknown partition table
May 25 17:30:25 masterlord kernel: Attached scsi removable disk sda at scsi2, channel 0, id 0, lun 0
May 25 17:30:25 masterlord kernel: USB Mass Storage device found at 5
May 25 17:30:25 masterlord scsi.agent[8795]: disk at /devices/pci0000:00/0000:00:11.2/usb3/3-1/3-1.1/3-1.1:1.0/host2/2:0:0:0

Provedeni mount ani umount se v syslogu nijak neprojevi. Oba ukony provadim pomoci Kwikdisk z KDE jako user a vse slape spravne tak dlouho, jak potrebuji. Po odpojeni od USB portu se v syslogu objevi:

May 25 17:36:28 masterlord kernel: usb 3-1.1: USB disconnect, address 5

A takhle to vypada, kdyz flash disk pripojim k USB 2.0 na radici:

May 24 14:28:40 masterlord kernel: usb 5-2: new high speed USB device using address 3
May 24 14:28:40 masterlord kernel: hub 5-2:1.0: USB hub found
May 24 14:28:40 masterlord kernel: hub 5-2:1.0: 1 port detected
May 24 14:28:41 masterlord kernel: usb 5-2.1: new high speed USB device using address 4
May 24 14:28:41 masterlord kernel: Initializing USB Mass Storage driver...
May 24 14:28:41 masterlord kernel: scsi0 : SCSI emulation for USB Mass Storage devices
May 24 14:28:42 masterlord kernel:   Vendor: USB 2.0   Model: Flash Disk        Rev: PROL
May 24 14:28:42 masterlord kernel:   Type:   Direct-Access                      ANSI SCSI revision: 02
May 24 14:28:42 masterlord kernel: Attached scsi generic sg0 at scsi0, channel 0, id 0, lun 0,  type 0
May 24 14:28:42 masterlord kernel: USB Mass Storage device found at 4
May 24 14:28:42 masterlord kernel: usbcore: registered new driver usb-storage
May 24 14:28:42 masterlord kernel: USB Mass Storage support registered.
May 24 14:28:43 masterlord scsi.agent[4911]: disk at /devices/pci0000:00/0000:00:09.2/usb5/5-2/5-2.1/5-2.1:1.0/host0/0:0:0:0
May 24 14:28:43 masterlord kernel: SCSI device sda: 256000 512-byte hdwr sectors (131 MB)
May 24 14:28:43 masterlord kernel: sda: Write Protect is off
May 24 14:28:43 masterlord kernel: sda: Mode Sense: 00 06 00 00
May 24 14:28:43 masterlord kernel: sda: assuming drive cache: write through
May 24 14:28:43 masterlord kernel:  /dev/scsi/host0/bus0/target0/lun0: unknown partition table
May 24 14:28:43 masterlord kernel: Attached scsi removable disk sda at scsi0, channel 0, id 0, lun 0

    >>> V tomhle miste jsem provedl mount a snazil jsem se s diskem pracovat, chvili se celkem darilo <<<

May 24 14:32:59 masterlord kernel: usb 5-2.1: reset high speed USB device using address 4
May 24 14:32:59 masterlord kernel: usb 5-2.1: device not accepting address 4, error -71
May 24 14:32:59 masterlord kernel: scsi: Device offlined - not ready after error recovery: host 0 channel 0 id 0 lun 0
May 24 14:32:59 masterlord kernel: SCSI error : <0 0 0 0> return code = 0x50000
May 24 14:32:59 masterlord kernel: end_request: I/O error, dev sda, sector 501
May 24 14:32:59 masterlord kernel: usb 5-2.1: USB disconnect, address 4
May 24 14:32:59 masterlord kernel: FAT: Directory bread(block 501) failed
May 24 14:32:59 masterlord kernel: scsi0 (0:0): rejecting I/O to device being removed
May 24 14:32:59 masterlord kernel: FAT: Directory bread(block 502) failed
May 24 14:32:59 masterlord kernel: scsi0 (0:0): rejecting I/O to device being removed
    ... ...
    >>> Celkem 32 hlaseni lisicich se jen cislem bloku - rada jde po jedne od 501 az po 532 <<<
    ... ...
May 24 14:33:00 masterlord kernel: FAT: Directory bread(block 531) failed
May 24 14:33:00 masterlord kernel: scsi0 (0:0): rejecting I/O to device being removed
May 24 14:33:00 masterlord kernel: FAT: Directory bread(block 532) failed
May 24 14:33:00 masterlord kernel: usb 5-2.1: new high speed USB device using address 5
May 24 14:33:00 masterlord kernel: scsi0 (0:0): rejecting I/O to dead device
May 24 14:33:00 masterlord kernel: FAT: bread(block 1) in fat_access failed
May 24 14:33:00 masterlord kernel: usb 5-2.1: device not accepting address 5, error -71
May 24 14:33:00 masterlord kernel: usb 5-2.1: new high speed USB device using address 6
May 24 14:33:00 masterlord kernel: usb 5-2.1: device not accepting address 6, error -71
May 24 14:33:07 masterlord kernel: scsi0 (0:0): rejecting I/O to dead device
May 24 14:33:07 masterlord kernel: FAT: bread(block 1) in fat_access failed
May 24 14:44:09 masterlord kernel: scsi0 (0:0): rejecting I/O to dead device
May 24 14:44:09 masterlord kernel: FAT: bread(block 1) in fat_access failed

    >>> Tady jsem se jeste pokusil, jestli nahodou nepomuze zavedeni modulu ohci-hcd <<<

May 24 14:44:42 masterlord kernel: ohci_hcd: 2004 Feb 02 USB 1.1 'Open' Host Controller (OHCI) Driver (PCI)
May 24 14:44:42 masterlord kernel: ohci_hcd: block sizes: ed 64 td 64
May 24 14:44:56 masterlord kernel: scsi0 (0:0): rejecting I/O to dead device
May 24 14:44:56 masterlord kernel: FAT: bread(block 1) in fat_access failed

    >>> Tady me to prestalo bavit a flashdisk jsem zase vypojil <<<

May 24 14:45:03 masterlord kernel: usb 5-2: USB disconnect, address 3

Odpovědi

[Wildwire]

Pokud se dobre divam, tak se mi moc nelibi tenhle radek:

May 24 14:28:43 masterlord kernel: /dev/scsi/host0/bus0/target0/lun0: unknown partition table

To by znamenalo ze disk/flashka nebo pametova karta ma spatnou tabulku rozdeleni disku. Zalezi na tom, co to je za zarizeni, jinak neni problem na takovem zarizeni novou tabulku vytvorit pomoci fdisku/cfdisku.

Ovsem nevim jak mate nastaveny system, mozna mate na tom flashdisku FATku rovnou od zacatku a mountujete cele zarizeni (/dev/sda) misto partisny (/dev/sda1).

Kazdopadne mam podobny problem, ze mi nejde pripojit externi disk do PCMCIA radice USB 2.0, stejny chipset jako je Vas. Ale tak me napadlo, ze mozna odpojeni USB radice na motherboardu by snad fungovalo. Pridavne USB radice maji jak UHCI (1.5Mb/s a 12Mb/s) tak EHCI (480Mb/s), takze je mozne, ze si s radicem na zakladni desce nerozumi. Az se dostanu k notebooku tak to zkusim.

[Eregon]

Ten radek je divny, to mate pravdu, ale nemyslim, ze to je ten problem - na USB 2.0 radici se problem objevuje, i kdyz pouziju postup presne stejny, jako na integrovanem USB 1.1. Na USB 1.1 muzu bez problemu pracovat, ale na USB 2.0 si pocitac a zarizeni prestanou rozumet (viz komenovany vypis syslogu).

Problem mam s USB Flashdiskem Centon DataStick III (128 MB) a cteckou pametovych karet, kterou pouzivam pro karty SD a RS-MMC, ale jde obecne o USB pametova zarizeni, ktera se ke mne dostanou. Cfdisk by me sice asi zbavil divneho radku v syslogu, ale kdyz mi nekdo na flashce prinese data, asi tezko mu ji muzu zformatovat a poslat ho pro data znova, zvlast, kdyz je prinesl zdaleka Nastesti v tom (doufam) ten problem neni.

Ten konflikt mezi integrovanym a pridavnym USB radicem me vazne nenapadl, mel jsem za to, ze USB se vyvinulo celkem zdrave... ale jak prave ctu ruzne po netu, ani tady si vyrobci neodpustili bordel ve standardech - kdy ja prestanu byt takovy naivni optimista?

Konfliktu radicu by nasvedcovalo i to, ze pametove zarizeni na integrovanem radici chodi, jak ma. Kdyz ho pripojim na 2.0 radic, zarizeni za chvili "umre" a ja se s nim nechytam uz ani na integrovanem radici. Zajimave na tom ale je to, ze CDMA modem pripojeny na 2.0 radic se vubec nevzrusuje a s pripojenim nemam problem ani po vsech tech zverstvech, ktere se mi v syslogu objevi.

Vyzkousim integrovany radic zakazat a dam vedet, jak jsem dopadl, Vy se pak ozvete taky.

[Eregon]

Tak jsem integrovany radic zakazal a stejne to nepomohlo, pametova zarizeni pripojena na USB 2.0 radic jsou porad nepouzitelna

Po celkem kratke chvilce jsem do syslogu dostal takovou hromadu balastu, ze to snad ani mozny neni, lsusb mi zatuhlo (neda se zabit ani pres kill -9). Malicky vzorek prikladam, vypada to jako nejaky developersky vypis, tzn. by teoreticky vse mohlo slapat s novejsim jadrem (ted jedu na uz docela zastarale 2.6.8).

kernel: [pg0+409407861/1069613056] hub_thread+0x35/0x100 [usbcore]
kernel: [<d8a5f175>] hub_thread+0x35/0x100 [usbcore]
kernel: [kernel_thread_helper+5/16] kernel_thread_helper+0x5/0x10
kernel: [<c01052c5>] kernel_thread_helper+0x5/0x10
kernel: Code: c0 74 04 89 cb eb c5 c7 04 24 5c 39 2e c0 8b 45 08 89 44 24 04 e8 92 cc fd ff b9 78 62 3c c0 ff 05 78 62 3c c0 0f 8e eb 18 00 00 <0f> 0b 66 05 e8 31 2e c0 8b 0b 89 cb eb 93 8d 76 00 8b 47 34 c7


[Eregon]

Update: vyhradil jsem si na tenhle problem odpoledne a vyzkousel jsem vsechno mozne - nainstalovat novejsi jadro (konkretne 2.6.11-6 z Mandrake 10.2), odebrat ostatni PCI karty, presouvat USB radic do ruznych PCI slotu, zapinat a vypinat integrovane USB porty, zavadet a odebirat ruzne moduly a tohle vsechno ruzne kombinovat.

Dosel jsem k tomu, ze potize mi zpusobuje modul ehci-hcd (pro USB 2.0). Pokud je zaveden, po nejake dobe prace s pametovym zarizenim se objevi vyse popsana chyba (reset high speed USB device) a vse jde do haje, bez ohledu na to, zde je zaveden modul uhci-hcd (pro USB 1.1). Staci ale ehci-hcd odebrat a chyba se neobjevuje, jenomze vse jede jen rychlosti USB 1.1, tzn. 12 Mbps, tzn. pomalu (vzhledem k tomu, ze jsem si USB 2.0 radic poridil hlavne kvuli te rychlosti)

Budu vdecny za kazdy tip, co ted zkusit.

[CIJOML]

Zkus prekompilovat kernel 2.6.12-vanilla z kernel.org a kdyz to nepomuze, tak presne tenhle report cos tu predved posli v anglickem zneni na linux-kernel@vger.kernel.org

[Eregon]

Cteme si myslenky, presne toto jsem totiz udelal a prave sem jdu vlozit vysledek.

V changelogu jadra 2.6.12 jsem nasel, ze se provadela nejaka oprava kolem ehci (konkretne: [PATCH] spin longer for ehci port reset completion) a pise se k tomu, ze: "This makes the EHCI driver spin a bit longer before concluding that the port reset failed."

Vysledkem je to, ze presun dat z USB zarizeni se sice zastavi, jak jsem popsal vyse, ale neskonci hned chybou - po nejakem case se zase chvilku pokracuje. Trva to ale nesnesitelne dlouho (pri prvnim pokusu 120 MB dat z flashky na disk nekolik minut, pri druhem pokusu doslo k chybe na 82% prenosu - zvlastni je, ze kopirovani jako takove se neprerusilo a porad stoji na 82%, syslog je ale plny hlaseni - uplne novych, takova jeste neznam )

Takze jdu na sestaveni angl. reportu, za vsechny rady dik.