Re: [PATCH] libata: implement ata_wait_after_reset()

["Indan Zupancic" <indan@nul.nu>]

Hello,

Disabling sd_resume() gives me a "quick" resume again (few seconds),
though it doesn't get rid of the COMRESET errors:

[    2.476417] ata1: device not ready (errno=-19), forcing hardreset
[    2.476440] ata2: SATA link down (SStatus 0 SControl 310)
[    2.824896] usb_endpoint usbdev2.3_ep00: PM: resume from 0, parent 2-1 still 2
[    2.824902] usb 2-1:1.0: PM: resume from 2, parent 2-1 still 2
[    2.824906] usb 2-1:1.1: PM: resume from 2, parent 2-1 still 2
[    2.824910] usb 2-1:1.2: PM: resume from 2, parent 2-1 still 2
[    2.824914] usb_endpoint usbdev2.3_ep05: PM: resume from 0, parent 2-1:1.2 still 2
[    2.824919] usb_endpoint usbdev2.3_ep85: PM: resume from 0, parent 2-1:1.2 still 2
[    2.824923] usb_endpoint usbdev2.3_ep81: PM: resume from 0, parent 2-1:1.0 still 2
[    2.825392] Restarting tasks ... done.
[    2.931812] ata1: COMRESET failed (errno=-19)
[    2.931880] ata1: reset failed (errno=-19), retrying in 10 secs
[    3.136543] usb 2-1: USB disconnect, address 3
[    3.249382] usb 2-1: new full speed USB device using ohci_hcd and address 4
[    3.396768] usb 2-1: configuration #1 chosen from 1 choice
[    4.225560] usb 2-1: reset full speed USB device using ohci_hcd and address 4
[    4.372298] usbcore: registered new interface driver speedtch
[    4.456118] speedtch 2-1:1.0: found stage 1 firmware speedtch-1.bin.0.00
[    4.621056] speedtch 2-1:1.0: found stage 2 firmware speedtch-2.bin.0.00
[    9.153941] ATM dev 0: ADSL line is synchronising
[   12.918440] ata1: SATA link up 1.5 Gbps (SStatus 113 SControl 310)
[   12.921971] ata1.00: ata_hpa_resize 1: sectors = 234441648, hpa_sectors = 234441648
[   12.924924] ata1.00: ata_hpa_resize 1: sectors = 234441648, hpa_sectors = 234441648

Doesn't the controller generate an interrupt when it detects a harddisk?
Is it really needed to do polling?

Idea: What about implementing a sata_sil specific .check_status function,
instead of using the generic ata_check_status()? That one could check other
registers to find out what's really going on in the case ENODEV is returned
by ata_check_status(), or something like that.

Greetings,

Indan