[Qemu-devel] [Bug 1652459] [NEW] kvm rbd driver (and maybe others, i.e. qcow2, qed and so on) does not report DISCARD-ZERO flag

[Коренберг Марк]

Public bug reported:

# lsblk -D
NAME   DISC-ALN DISC-GRAN DISC-MAX DISC-ZERO
sda           0        4K       1G         0
├─sda1        0        4K       1G         0
├─sda2     1024        4K       1G         0
└─sda5        0        4K       1G         0

Last column should be `1` at least for "RBD+discard=unmap" since reading
from discarded regions in RBD MUST return zeroes. The same with QCOW2,
QED and sparse raw images. KVM should copy value of this flag when real
raw device (i.e. real SSD) with discard capability is used as virtual
disk.

** Affects: qemu
     Importance: Undecided
         Status: New

** Description changed:

+ ```
  # lsblk -D
  NAME   DISC-ALN DISC-GRAN DISC-MAX DISC-ZERO
  sda           0        4K       1G         0
  ├─sda1        0        4K       1G         0
  ├─sda2     1024        4K       1G         0
  └─sda5        0        4K       1G         0
+ ```
  
- 
- Last column should be `1` at least for "RBD+discard=unmap" since reading from 
discarded regions in RBD MUST return zeroes. The same with QCOW2, QED and 
sparse raw images. KVM should copy value of this flag when real raw device 
(i.e. real SSD) with discard capability is used as virtual disk.
+ Last column should be `1` at least for "RBD+discard=unmap" since reading
+ from discarded regions in RBD MUST return zeroes. The same with QCOW2,
+ QED and sparse raw images. KVM should copy value of this flag when real
+ raw device (i.e. real SSD) with discard capability is used as virtual
+ disk.

** Description changed:

- ```
- # lsblk -D
- NAME   DISC-ALN DISC-GRAN DISC-MAX DISC-ZERO
- sda           0        4K       1G         0
- ├─sda1        0        4K       1G         0
- ├─sda2     1024        4K       1G         0
- └─sda5        0        4K       1G         0
- ```
+    # lsblk -D
+    NAME   DISC-ALN DISC-GRAN DISC-MAX DISC-ZERO
+    sda           0        4K       1G         0
+    ├─sda1        0        4K       1G         0
+    ├─sda2     1024        4K       1G         0
+    └─sda5        0        4K       1G         0
  
- Last column should be `1` at least for "RBD+discard=unmap" since reading
- from discarded regions in RBD MUST return zeroes. The same with QCOW2,
- QED and sparse raw images. KVM should copy value of this flag when real
- raw device (i.e. real SSD) with discard capability is used as virtual
- disk.
+ 
+ Last column should be `1` at least for "RBD+discard=unmap" since reading from 
discarded regions in RBD MUST return zeroes. The same with QCOW2, QED and 
sparse raw images. KVM should copy value of this flag when real raw device 
(i.e. real SSD) with discard capability is used as virtual disk.

** Description changed:

-    # lsblk -D
-    NAME   DISC-ALN DISC-GRAN DISC-MAX DISC-ZERO
-    sda           0        4K       1G         0
-    ├─sda1        0        4K       1G         0
-    ├─sda2     1024        4K       1G         0
-    └─sda5        0        4K       1G         0
+ <code>
+ # lsblk -D
+ NAME   DISC-ALN DISC-GRAN DISC-MAX DISC-ZERO
+ sda           0        4K       1G         0
+ ├─sda1        0        4K       1G         0
+ ├─sda2     1024        4K       1G         0
+ └─sda5        0        4K       1G         0
+ <code>
  
- 
- Last column should be `1` at least for "RBD+discard=unmap" since reading from 
discarded regions in RBD MUST return zeroes. The same with QCOW2, QED and 
sparse raw images. KVM should copy value of this flag when real raw device 
(i.e. real SSD) with discard capability is used as virtual disk.
+ Last column should be `1` at least for "RBD+discard=unmap" since reading
+ from discarded regions in RBD MUST return zeroes. The same with QCOW2,
+ QED and sparse raw images. KVM should copy value of this flag when real
+ raw device (i.e. real SSD) with discard capability is used as virtual
+ disk.

** Description changed:

- <code>
+ <pre>
  # lsblk -D
  NAME   DISC-ALN DISC-GRAN DISC-MAX DISC-ZERO
  sda           0        4K       1G         0
  ├─sda1        0        4K       1G         0
  ├─sda2     1024        4K       1G         0
  └─sda5        0        4K       1G         0
- <code>
+ </pre>
  
  Last column should be `1` at least for "RBD+discard=unmap" since reading
  from discarded regions in RBD MUST return zeroes. The same with QCOW2,
  QED and sparse raw images. KVM should copy value of this flag when real
  raw device (i.e. real SSD) with discard capability is used as virtual
  disk.

** Description changed:

- <pre>
  # lsblk -D
  NAME   DISC-ALN DISC-GRAN DISC-MAX DISC-ZERO
  sda           0        4K       1G         0
  ├─sda1        0        4K       1G         0
  ├─sda2     1024        4K       1G         0
  └─sda5        0        4K       1G         0
- </pre>
  
  Last column should be `1` at least for "RBD+discard=unmap" since reading
  from discarded regions in RBD MUST return zeroes. The same with QCOW2,
  QED and sparse raw images. KVM should copy value of this flag when real
  raw device (i.e. real SSD) with discard capability is used as virtual
  disk.

-- 
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/1652459

Title:
  kvm rbd driver (and maybe others, i.e. qcow2, qed and so on)  does not
  report DISCARD-ZERO flag

Status in QEMU:
  New

Bug description:
  # lsblk -D
  NAME   DISC-ALN DISC-GRAN DISC-MAX DISC-ZERO
  sda           0        4K       1G         0
  ├─sda1        0        4K       1G         0
  ├─sda2     1024        4K       1G         0
  └─sda5        0        4K       1G         0

  Last column should be `1` at least for "RBD+discard=unmap" since
  reading from discarded regions in RBD MUST return zeroes. The same
  with QCOW2, QED and sparse raw images. KVM should copy value of this
  flag when real raw device (i.e. real SSD) with discard capability is
  used as virtual disk.

To manage notifications about this bug go to:
https://bugs.launchpad.net/qemu/+bug/1652459/+subscriptions