Using an NVMe drive causes huge power drain

Related to :
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1579917
(marked as invalid)

This change was made by a bot.

Solved upstream with kernel 4.11 RC1 from Ubuntu kernel archive.
I'm back to 4.23 W instead of 8.26 W in same power conditions.

I reach PC8 power state (powertop) instead of PC3.

Hope the APST Andy's patch (I guess it's the key ?) will be backported soon to Zesty's kernel.

Francois - I doubt if that is going to happen for 17.04. It appears that single patch requires a substantial amount of infrastructure and prerequisite patches.

I know that my bugs seem to me more important than other people's bugs because that's MY bugs :-) .

But *IMHO* this lack of NVMe power management (and consequently Skylake CPU low power states usage) is critical since :
1) it voids a lot of power management efforts these last years in Ubuntu/Linux ;
2) low power states are (AFAIK) required for Skylake CPU (and NVMe drive ?) long term durability.

I was hoping that for Zesty as this bug is not wontfixed and seems to be useful only for APST :
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1666401

I can confirm Francois experience with my 950 pro. Although I saved one or two watts. It's a bit hard to tell. However my PC states were already reaching PC8 before this patch.

I found that APST patches can be cleanly cherry-picked into zesty tree, please try it:
http://people.canonical.com/~khfeng/lp1664602/

I've tested it now and it seems to work flawlessly.
Power usage is very similar to 4.11 kernel.

It would be great to see your patches in Zesty.

Kai-Heng Feng: Please post a git branch from which I can pull your patches (or perhaps just attach them to this bug). Make sure the upstream commit SHA1 is included by using the '-x' option of 'git cherry-pick'.

I can confirm Kai-Heng's patched kernel (linux-image-4.10.0-14-generic_4.10.0-14.16_amd64.deb) triggers pc8 state with a Samsung 960 EVO, where the stock kernel only attains pc3 at best.

@Tim, When I checked the backport it was not a clean cherry-pick. As you mentioned earlier there was some prerequisite patches that git pulled in. I did not continue investigating, but I guess there should only be a few. Either way the patches seem nvme

@Kai-Heng Feng, please make sure you appropriately document and separate out patches that cleanly cherry-pick. Also please be careful for context patches that git pulls in to make the patch apply.

@Dave

I was wrong, you are right, it's not a clean cherry-pick.

I picked 7ac8b7abdb8d903c83fa15cfd952316e824fe6f3 from nvme tree (which can be clean applied) instead of bd4da3abaabffdd2472fb7085fcadd5d1d8c2153 from master.

The later one's diff contextually depends 8a9ae523282f324989850fcf41312b42a2fb9296, which is not a functional dependendency to APST, so the conflict can be easily resolved.

@Kai-Heng have you been able to make any progress on this? I think Tim is simply waiting on a git branch or git send-email submission to the kernel-team mailing list.

@Dave,

For Zesty, it's already in the tree.
For Yakkety, I already sent patches to kernel-team ML.
For Xenial, I'd like to see if there's any regression after some real world usages in Yakkety and Zesty - after all the patch author made a blacklist table for a reason. If everything goes smoothly, I'll backport it to Xenial.

This bug was fixed in the package linux - 4.10.0-14.16

---------------
linux (4.10.0-14.16) zesty; urgency=low

  [ Tim Gardner ]

  * Release Tracking Bug
    - LP: #1673805

  * msleep() bug causes Nuvoton I2C TPM device driver delays (LP: #1667567)
    - tpm: msleep() delays - replace with usleep_range() in i2c nuvoton driver
    - SAUCE: tpm: add sleep only for retry in i2c_nuvoton_write_status()

  * C++ demangling support missing from perf (LP: #1396654)
    - [Config] added binutils-dev to Build-deps

  * dm-queue-length module is not included in installer/initramfs (LP: #1673350)
    - [Config] d-i: Also add dm-queue-length to multipath modules

  * move aufs.ko from -extra to linux-image package (LP: #1673498)
    - [config] aufs.ko moved to linux-image package

  * Using an NVMe drive causes huge power drain (LP: #1664602)
    - nvme: Add a quirk mechanism that uses identify_ctrl
    - nvme: Enable autonomous power state transitions

  * Broadcom bluetooth modules sometimes fail to initialize (LP: #1483101)
    - Bluetooth: btbcm: Add a delay for module reset

  * Need support of Broadcom bluetooth device [413c:8143] (LP: #1166113)
    - Bluetooth: btusb: Add support for 413c:8143

  * Zesty update to v4.10.3 stable release (LP: #1673118)
    - serial: 8250_pci: Add MKS Tenta SCOM-0800 and SCOM-0801 cards
    - KVM: s390: Disable dirty log retrieval for UCONTROL guests
    - KVM: VMX: use correct vmcs_read/write for guest segment selector/base
    - Bluetooth: Add another AR3012 04ca:3018 device
    - phy: qcom-ufs: Don't kfree devres resource
    - phy: qcom-ufs: Fix misplaced jump label
    - s390/qdio: clear DSCI prior to scanning multiple input queues
    - s390/dcssblk: fix device size calculation in dcssblk_direct_access()
    - s390/kdump: Use "LINUX" ELF note name instead of "CORE"
    - s390/chsc: Add exception handler for CHSC instruction
    - s390: TASK_SIZE for kernel threads
    - s390/topology: correct allocation of topology information
    - s390: make setup_randomness work
    - s390: use correct input data address for setup_randomness
    - net: mvpp2: fix DMA address calculation in mvpp2_txq_inc_put()
    - cxl: Prevent read/write to AFU config space while AFU not configured
    - cxl: fix nested locking hang during EEH hotplug
    - brcmfmac: fix incorrect event channel deduction
    - mnt: Tuck mounts under others instead of creating shadow/side mounts.
    - IB/ipoib: Fix deadlock between rmmod and set_mode
    - IB/IPoIB: Add destination address when re-queue packet
    - IB/mlx5: Fix out-of-bound access
    - IB/SRP: Avoid using IB_MR_TYPE_SG_GAPS
    - IB/srp: Avoid that duplicate responses trigger a kernel bug
    - IB/srp: Fix race conditions related to task management
    - Btrfs: fix data loss after truncate when using the no-holes feature
    - orangefs: Use RCU for destroy_inode
    - memory/atmel-ebi: Fix ns <-> cycles conversions
    - tracing: Fix return value check in trace_benchmark_reg()
    - ktest: Fix child exit code processing
    - ceph: remove req from unsafe list when unregistering it
    - target: Fix NULL dereference during LUN lookup + active I/O shutdown
    - drivers/pci/hotplug: Han...

After some reasonable big time of use, I get (as far I noticed in user interface or system journals) no errors and not any hang. I did not notice any slowdown either.

Thanks a lot, my battery is smiling again.

Status changed to 'Confirmed' because the bug affects multiple users.

This bug is awaiting verification that the kernel in -proposed solves the problem. Please test the kernel and update this bug with the results. If the problem is solved, change the tag 'verification-needed-yakkety' to 'verification-done-yakkety'. If the problem still exists, change the tag 'verification-needed-yakkety' to 'verification-failed-yakkety'.

If verification is not done by 5 working days from today, this fix will be dropped from the source code, and this bug will be closed.

See https://wiki.ubuntu.com/Testing/EnableProposed for documentation how to enable and use -proposed. Thank you!

This bug was fixed in the package linux - 4.8.0-49.52

---------------
linux (4.8.0-49.52) yakkety; urgency=low

  * linux: 4.8.0-49.52 -proposed tracker (LP: #1684427)

  * [Hyper-V] hv: util: move waiting for release to hv_utils_transport itself
    (LP: #1682561)
    - Drivers: hv: util: move waiting for release to hv_utils_transport itself

linux (4.8.0-48.51) yakkety; urgency=low

  * linux: 4.8.0-48.51 -proposed tracker (LP: #1682034)

  * [Hyper-V] hv: vmbus: Raise retry/wait limits in vmbus_post_msg()
    (LP: #1681893)
    - Drivers: hv: vmbus: Raise retry/wait limits in vmbus_post_msg()

linux (4.8.0-47.50) yakkety; urgency=low

  * linux: 4.8.0-47.50 -proposed tracker (LP: #1679678)

  * CVE-2017-6353
    - sctp: deny peeloff operation on asocs with threads sleeping on it

  * CVE-2017-5986
    - sctp: avoid BUG_ON on sctp_wait_for_sndbuf

  * vfat: missing iso8859-1 charset (LP: #1677230)
    - [Config] NLS_ISO8859_1=y

  * [Hyper-V] pci-hyperv: Use device serial number as PCI domain (LP: #1667527)
    - net/mlx4_core: Use cq quota in SRIOV when creating completion EQs

  * Regression: KVM modules should be on main kernel package (LP: #1678099)
    - [Config] powerpc: Add kvm-hv and kvm-pr to the generic inclusion list

  * linux-lts-xenial 4.4.0-63.84~14.04.2 ADT test failure with linux-lts-xenial
    4.4.0-63.84~14.04.2 (LP: #1664912)
    - SAUCE: apparmor: fix link auditing failure due to, uninitialized var

  * regession tests failing after stackprofile test is run (LP: #1661030)
    - SAUCE: fix regression with domain change in complain mode

  * Permission denied and inconsistent behavior in complain mode with 'ip netns
    list' command (LP: #1648903)
    - SAUCE: fix regression with domain change in complain mode

  * unexpected errno=13 and disconnected path when trying to open /proc/1/ns/mnt
    from a unshared mount namespace (LP: #1656121)
    - SAUCE: apparmor: null profiles should inherit parent control flags

  * apparmor refcount leak of profile namespace when removing profiles
    (LP: #1660849)
    - SAUCE: apparmor: fix ns ref count link when removing profiles from policy

  * tor in lxd: apparmor="DENIED" operation="change_onexec"
    namespace="root//CONTAINERNAME_<var-lib-lxd>" profile="unconfined"
    name="system_tor" (LP: #1648143)
    - SAUCE: apparmor: Fix no_new_privs blocking change_onexec when using stacked
      namespaces

  * apparmor oops in bind_mnt when dev_path lookup fails (LP: #1660840)
    - SAUCE: apparmor: fix oops in bind_mnt when dev_path lookup fails

  * apparmor auditing denied access of special apparmor .null fi\ le
    (LP: #1660836)
    - SAUCE: apparmor: Don't audit denied access of special apparmor .null file

  * apparmor label leak when new label is unused (LP: #1660834)
    - SAUCE: apparmor: fix label leak when new label is unused

  * apparmor reference count bug in label_merge_insert() (LP: #1660833)
    - SAUCE: apparmor: fix reference count bug in label_merge_insert()

  * apparmor's raw_data file in securityfs is sometimes truncated (LP: #1638996)
    - SAUCE: apparmor: fix replacement race in reading rawdata

  * unix domain socket cross permission check failing with n...

Unfortunately, this fix breaks Dell XPS 9550 with Samsung PM951 drives. I now get a hung system due to the Samsung controller getting reset. Please see https://bugzilla.kernel.org/show_bug.cgi?id=195039 for a discussion of this issue.

The quirk detection is not complete, I fear

Please try https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1678184/comments/44

I'm pretty sure this is not the first time a power-saving feature breaks some subset of hardware. Why even add it in a bugfix update?

After applying the workaround from LP#1678184, I'm happy about the somewhat improved battery life, but I'll never get back the several hours it took to figure out what's going on, and considering whether I should contact Dell support about my failing storage device (which wouldn't have helped, but would probably have wasted several weeks on top of that).

And I'm still on 16.10. Why was this even released in Yakkety, with LP#1678184 having been reported on March 31?

This bug is awaiting verification that the kernel in -proposed solves the problem. Please test the kernel and update this bug with the results. If the problem is solved, change the tag 'verification-needed-xenial' to 'verification-done-xenial'. If the problem still exists, change the tag 'verification-needed-xenial' to 'verification-failed-xenial'.

If verification is not done by 5 working days from today, this fix will be dropped from the source code, and this bug will be closed.

See https://wiki.ubuntu.com/Testing/EnableProposed for documentation how to enable and use -proposed. Thank you!

Verified on 4.4.0-117-generic, APST was enabled.

#sudo nvme get-feature -f 0x0c -H /dev/nvme0n1

get-feature:0x0c (Autonomous Power State Transition), Current value: 0x000001
 Autonomous Power State Transition Enable (APSTE): Enabled
 Auto PST Entries .................
 Entry[ 0]
 .................
 Idle Time Prior to Transition (ITPT): 1500 ms
 Idle Transition Power State (ITPS): 3
 .................
 Entry[ 1]

This bug was fixed in the package linux - 4.4.0-119.143

---------------
linux (4.4.0-119.143) xenial; urgency=medium

  * linux: 4.4.0-119.143 -proposed tracker (LP: #1760327)

  * Dell XPS 13 9360 bluetooth scan can not detect any device (LP: #1759821)
    - Revert "Bluetooth: btusb: fix QCA Rome suspend/resume"

linux (4.4.0-118.142) xenial; urgency=medium

  * linux: 4.4.0-118.142 -proposed tracker (LP: #1759607)

  * Kernel panic with AWS 4.4.0-1053 / 4.4.0-1015 (Trusty) (LP: #1758869)
    - x86/microcode/AMD: Do not load when running on a hypervisor

  * CVE-2018-8043
    - net: phy: mdio-bcm-unimac: fix potential NULL dereference in
      unimac_mdio_probe()

linux (4.4.0-117.141) xenial; urgency=medium

  * linux: 4.4.0-117.141 -proposed tracker (LP: #1755208)

  * Xenial update to 4.4.114 stable release (LP: #1754592)
    - x86/asm/32: Make sync_core() handle missing CPUID on all 32-bit kernels
    - usbip: prevent vhci_hcd driver from leaking a socket pointer address
    - usbip: Fix implicit fallthrough warning
    - usbip: Fix potential format overflow in userspace tools
    - x86/microcode/intel: Fix BDW late-loading revision check
    - x86/retpoline: Fill RSB on context switch for affected CPUs
    - sched/deadline: Use the revised wakeup rule for suspending constrained dl
      tasks
    - can: af_can: can_rcv(): replace WARN_ONCE by pr_warn_once
    - can: af_can: canfd_rcv(): replace WARN_ONCE by pr_warn_once
    - PM / sleep: declare __tracedata symbols as char[] rather than char
    - time: Avoid undefined behaviour in ktime_add_safe()
    - timers: Plug locking race vs. timer migration
    - Prevent timer value 0 for MWAITX
    - drivers: base: cacheinfo: fix x86 with CONFIG_OF enabled
    - drivers: base: cacheinfo: fix boot error message when acpi is enabled
    - PCI: layerscape: Add "fsl,ls2085a-pcie" compatible ID
    - PCI: layerscape: Fix MSG TLP drop setting
    - mmc: sdhci-of-esdhc: add/remove some quirks according to vendor version
    - fs/select: add vmalloc fallback for select(2)
    - hwpoison, memcg: forcibly uncharge LRU pages
    - cma: fix calculation of aligned offset
    - mm, page_alloc: fix potential false positive in __zone_watermark_ok
    - ipc: msg, make msgrcv work with LONG_MIN
    - x86/ioapic: Fix incorrect pointers in ioapic_setup_resources()
    - ACPI / processor: Avoid reserving IO regions too early
    - ACPI / scan: Prefer devices without _HID/_CID for _ADR matching
    - ACPICA: Namespace: fix operand cache leak
    - netfilter: x_tables: speed up jump target validation
    - netfilter: arp_tables: fix invoking 32bit "iptable -P INPUT ACCEPT" failed
      in 64bit kernel
    - netfilter: nf_dup_ipv6: set again FLOWI_FLAG_KNOWN_NH at flowi6_flags
    - netfilter: nf_ct_expect: remove the redundant slash when policy name is
      empty
    - netfilter: nfnetlink_queue: reject verdict request from different portid
    - netfilter: restart search if moved to other chain
    - netfilter: nf_conntrack_sip: extend request line validation
    - netfilter: use fwmark_reflect in nf_send_reset
    - ext2: Don't clear SGID when inheriting ACLs
    - reiserfs: fix race in prealloc discard
    - re...