RE: [PATCH] usb: uas: fix usb subsystem hang after power off hub port

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi,

>no I am sorry, that is an assumption you just cannot make.
>Anything can trigger a reset. That being SCSI is the common case certainly, but not the only case. And in those cases we cannot depend on upper layers doing the right thing, if we just ignore an error.
While we investigate this issue, we debugged and found usb_reset_and_verify_device will return -NODEV before enter post_reset operation.
And the return value(-ENODEV) will be returned to error handler.
uas_eh_device_reset_handler->usb_reset_device -> usb_reset_and_verify_device (return -ENODEV) Then I wrote that commit message that we think even if we ignore "ENODEV" in post reset to avoid hang issue but the error will also be reported to error handler.
#If ignore an error and the error will not be reported then it is not good.

Additional information about usb-storage driver(usb/storage/usb.c) in usb_stor_post_reset() function, it always return 0 that means rebind will not be execute and this issue doesn't happen.

Regards,
Kento Kobayashi

-----Original Message-----
From: Oliver Neukum <oneukum@xxxxxxxx> 
Sent: Saturday, March 9, 2019 1:52 AM
To: Kobayashi, Kento (Sony) <Kento.A.Kobayashi@xxxxxxxx>; gregkh@xxxxxxxxxxxxxxxxxxx; stern@xxxxxxxxxxxxxxxxxxx
Cc: usb-storage@xxxxxxxxxxxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx; linux-scsi@xxxxxxxxxxxxxxx; linux-usb@xxxxxxxxxxxxxxx
Subject: Re: [PATCH] usb: uas: fix usb subsystem hang after power off hub port

On Fr, 2019-03-08 at 09:13 +0000, Kento.A.Kobayashi@xxxxxxxx wrote:
> The usb_reset_and_verify_device included in usb_reset_device fails 
> with -ENODEV after power off hub port, and the -ENODEV error will be 
> reported to uas_eh_bus_reset_handler and upper layer, so it doesn't 
> need to do rebind if -ENODEV happens.

Hi,

no I am sorry, that is an assumption you just cannot make.
Anything can trigger a reset. That being SCSI is the common case certainly, but not the only case. And in those cases we cannot depend on upper layers doing the right thing, if we just ignore an error.

NACK

	Sorry
		Oliver





[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Index of Archives]     [SCSI Target Devel]     [Linux SCSI Target Infrastructure]     [Kernel Newbies]     [IDE]     [Security]     [Git]     [Netfilter]     [Bugtraq]     [Yosemite News]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Linux ATA RAID]     [Linux IIO]     [Samba]     [Device Mapper]

  Powered by Linux