linux-5.15.105 broke /dev/sd* naming (probing)

linux-5.15.105 broke /dev/sd* naming (probing)

[Ilari Jääskeläinen]

As I attached a USB SSD into CentOS 9 Stream computer, after a short
while it swaps /dev/sdb into /dev/sdc and the I/O gets ruined.
Kind regards, Ilari Jääskeläinen.

Re: linux-5.15.105 broke /dev/sd* naming (probing)

[Greg KH]

On Fri, Mar 31, 2023 at 10:52:55AM +0300, Ilari Jääskeläinen wrote:
> As I attached a USB SSD into CentOS 9 Stream computer, after a short
> while it swaps /dev/sdb into /dev/sdc and the I/O gets ruined.
> Kind regards, Ilari Jääskeläinen.
> 

Is this using the CentOS kernel, or a kernel.org release?

And if a device changes names like that, it implies it was disconnected
and then reconnected, what does the kernel logs say when this happened?

thanks,

greg k-h

Re: linux-5.15.105 broke /dev/sd* naming (probing)

[Ilari Jääskeläinen]

I am afraid I cant do that. I already almost wiped my root partition by
accident because of this flaw.
pe, 2023-03-31 kello 10:05 +0200, Greg KH kirjoitti:
> On Fri, Mar 31, 2023 at 10:52:55AM +0300, Ilari Jääskeläinen wrote:
> > As I attached a USB SSD into CentOS 9 Stream computer, after a
> > short
> > while it swaps /dev/sdb into /dev/sdc and the I/O gets ruined.
> > Kind regards, Ilari Jääskeläinen.
> > 
> 
> Is this using the CentOS kernel, or a kernel.org release?
> 
> And if a device changes names like that, it implies it was
> disconnected
> and then reconnected, what does the kernel logs say when this
> happened?
> 
> thanks,
> 
> greg k-h

Re: linux-5.15.105 broke /dev/sd* naming (probing)

[Ilari Jääskeläinen]

I am using https://www.kernel.org/ 5.15.105.
pe, 2023-03-31 kello 10:05 +0200, Greg KH kirjoitti:
> On Fri, Mar 31, 2023 at 10:52:55AM +0300, Ilari Jääskeläinen wrote:
> > As I attached a USB SSD into CentOS 9 Stream computer, after a
> > short
> > while it swaps /dev/sdb into /dev/sdc and the I/O gets ruined.
> > Kind regards, Ilari Jääskeläinen.
> > 
> 
> Is this using the CentOS kernel, or a kernel.org release?
> 
> And if a device changes names like that, it implies it was
> disconnected
> and then reconnected, what does the kernel logs say when this
> happened?
> 
> thanks,
> 
> greg k-h

Re: linux-5.15.105 broke /dev/sd* naming (probing)

[Ilari Jääskeläinen]

roger that. good advice. thanks.
pe, 2023-03-31 kello 09:35 +0000, David Laight kirjoitti:
> From: Ilari Jääskeläinen
> > Sent: 31 March 2023 10:09
> > 
> > I am afraid I cant do that. I already almost wiped my root
> > partition by
> > accident because of this flaw.
> 
> Always mount filsystems by uuid, not device name.
> 
>         David
> 
> -
> Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes,
> MK1 1PT, UK
> Registration No: 1397386 (Wales)

Re: linux-5.15.105 broke /dev/sd* naming (probing)

[Ilari Jääskeläinen]

I am fairly certain that CentOS 9 Stream GNOME mounted it without UUID,
though.
pe, 2023-03-31 kello 09:35 +0000, David Laight kirjoitti:
> From: Ilari Jääskeläinen
> > Sent: 31 March 2023 10:09
> > 
> > I am afraid I cant do that. I already almost wiped my root
> > partition by
> > accident because of this flaw.
> 
> Always mount filsystems by uuid, not device name.
> 
>         David
> 
> -
> Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes,
> MK1 1PT, UK
> Registration No: 1397386 (Wales)

Re: linux-5.15.105 broke /dev/sd* naming (probing)

[Ilari Jääskeläinen]

Here is the output of dmesg without doing anything with the disk:

[ 7632.544792] usb 2-2: new SuperSpeed USB device number 2 using
xhci_hcd
[ 7632.575851] usb 2-2: New USB device found, idVendor=152d,
idProduct=b567, bcdDevice= 2.23
[ 7632.575856] usb 2-2: New USB device strings: Mfr=1, Product=2,
SerialNumber=3
[ 7632.575858] usb 2-2: Product: USB3.0 to SATA adapter
[ 7632.575860] usb 2-2: Manufacturer: JMicron
[ 7632.575861] usb 2-2: SerialNumber: 0000AB123531
[ 7632.636013] usbcore: registered new interface driver usb-storage
[ 7632.655897] scsi host4: uas
[ 7632.656277] usbcore: registered new interface driver uas
[ 7632.656480] scsi 4:0:0:0: Direct-Access     Samsung  SSD 860 EVO
500G 0223 PQ: 0 ANSI: 6
[ 7632.657492] sd 4:0:0:0: Attached scsi generic sg2 type 0
[ 7632.657753] sd 4:0:0:0: [sdb] Spinning up disk...
[ 7633.694677] ...ready
[ 7635.775548] sd 4:0:0:0: [sdb] 976773168 512-byte logical blocks:
(500 GB/466 GiB)
[ 7635.775551] sd 4:0:0:0: [sdb] 4096-byte physical blocks
[ 7635.775707] sd 4:0:0:0: [sdb] Write Protect is off
[ 7635.775710] sd 4:0:0:0: [sdb] Mode Sense: 53 00 00 08
[ 7635.776003] sd 4:0:0:0: [sdb] Write cache: enabled, read cache:
enabled, doesn't support DPO or FUA
[ 7635.776308] sd 4:0:0:0: [sdb] Optimal transfer size 33553920 bytes
not a multiple of physical block size (4096 bytes)
[ 7635.783827]  sdb: sdb1
[ 7635.786814] sd 4:0:0:0: [sdb] Attached SCSI disk
[ 7636.072316] EXT4-fs (sdb1): mounting ext2 file system using the ext4
subsystem
[ 7636.076261] EXT4-fs (sdb1): warning: mounting unchecked fs, running
e2fsck is recommended
[ 7636.083896] EXT4-fs (sdb1): mounted filesystem without journal.
Opts: errors=remount-ro. Quota mode: none.

The original fs of the SATA-III SSD used to be xfs, which I changed to
ext2 in hope it would fix the problem about "I/O-errors", but it didnt.

pe, 2023-03-31 kello 10:05 +0200, Greg KH kirjoitti:
> On Fri, Mar 31, 2023 at 10:52:55AM +0300, Ilari Jääskeläinen wrote:
> > As I attached a USB SSD into CentOS 9 Stream computer, after a short
> > while it swaps /dev/sdb into /dev/sdc and the I/O gets ruined.
> > Kind regards, Ilari Jääskeläinen.
> > 
> 
> Is this using the CentOS kernel, or a kernel.org release?
> 
> And if a device changes names like that, it implies it was disconnected
> and then reconnected, what does the kernel logs say when this happened?
> 
> thanks,
> 
> greg k-h

Re: linux-5.15.105 broke /dev/sd* naming (probing)

[Bagas Sanjaya]

[-- Attachment #1: Type: text/plain, Size: 662 bytes --]

On Fri, Mar 31, 2023 at 10:52:55AM +0300, Ilari Jääskeläinen wrote:
> As I attached a USB SSD into CentOS 9 Stream computer, after a short
> while it swaps /dev/sdb into /dev/sdc and the I/O gets ruined.

Is this a kernel regression?

Anyway, for future reports please at least see
Documentation/admin-guide/reporting-issues.rst in kernel sources. Also,
if you have a problem with distro kernel (CentOS in this case), please
report to support channels provided by the distro. Last but not least,
when replying, don't top-post; reply inline with appropriate context
instead.

Thanks.

-- 
An old man doll... just what I always wanted! - Clara

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]

Re: linux-5.15.105 broke /dev/sd* naming (probing)

[Bagas Sanjaya]

[-- Attachment #1: Type: text/plain, Size: 1981 bytes --]

On Fri, Mar 31, 2023 at 12:52:37PM +0300, Ilari Jääskeläinen wrote:
> Here is the output of dmesg without doing anything with the disk:
> 
> [ 7632.544792] usb 2-2: new SuperSpeed USB device number 2 using
> xhci_hcd
> [ 7632.575851] usb 2-2: New USB device found, idVendor=152d,
> idProduct=b567, bcdDevice= 2.23
> [ 7632.575856] usb 2-2: New USB device strings: Mfr=1, Product=2,
> SerialNumber=3
> [ 7632.575858] usb 2-2: Product: USB3.0 to SATA adapter
> [ 7632.575860] usb 2-2: Manufacturer: JMicron
> [ 7632.575861] usb 2-2: SerialNumber: 0000AB123531
> [ 7632.636013] usbcore: registered new interface driver usb-storage
> [ 7632.655897] scsi host4: uas
> [ 7632.656277] usbcore: registered new interface driver uas
> [ 7632.656480] scsi 4:0:0:0: Direct-Access     Samsung  SSD 860 EVO
> 500G 0223 PQ: 0 ANSI: 6
> [ 7632.657492] sd 4:0:0:0: Attached scsi generic sg2 type 0
> [ 7632.657753] sd 4:0:0:0: [sdb] Spinning up disk...
> [ 7633.694677] ...ready
> [ 7635.775548] sd 4:0:0:0: [sdb] 976773168 512-byte logical blocks:
> (500 GB/466 GiB)
> [ 7635.775551] sd 4:0:0:0: [sdb] 4096-byte physical blocks
> [ 7635.775707] sd 4:0:0:0: [sdb] Write Protect is off
> [ 7635.775710] sd 4:0:0:0: [sdb] Mode Sense: 53 00 00 08
> [ 7635.776003] sd 4:0:0:0: [sdb] Write cache: enabled, read cache:
> enabled, doesn't support DPO or FUA
> [ 7635.776308] sd 4:0:0:0: [sdb] Optimal transfer size 33553920 bytes
> not a multiple of physical block size (4096 bytes)
> [ 7635.783827]  sdb: sdb1
> [ 7635.786814] sd 4:0:0:0: [sdb] Attached SCSI disk
> [ 7636.072316] EXT4-fs (sdb1): mounting ext2 file system using the ext4
> subsystem
> [ 7636.076261] EXT4-fs (sdb1): warning: mounting unchecked fs, running
> e2fsck is recommended
> [ 7636.083896] EXT4-fs (sdb1): mounted filesystem without journal.
> Opts: errors=remount-ro. Quota mode: none.

Smells like nothing odd above, right?

-- 
An old man doll... just what I always wanted! - Clara

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]