Solid state drive

From ArchWiki
(Redirected from TRIM)

This article covers special topics for operating solid state drives (SSDs) and other flash-memory based storage devices.

If you want to partition an SSD for a specific purpose, it may be useful to consider the List of file systems optimized for flash memory.

For general usage, simply choose your preferred filesystem and enable #TRIM.

Usage

TRIM

Compared to hard drives, where deleting a file is only handled at the file system level[1], SSDs benefit from informing the disk controller when blocks of memory are free to be reused. Since the flash cells they are made of are worn out a little with each write operation, the disk controllers use algorithms to share the write operations on all the cells: this process is called wear leveling. Without the NVMe DEALLOCATE, SAS UNMAP or ATA_TRIM command (supported by most SSDs), the disk controller takes more time to do a write operation as soon as there is no empty memory blocks, as it has to shuffle data around to erase a cell before writing to it (see Wikipedia:Write amplification): a TechSpot benchmark shows the performance impact before and after filling an SSD with data.

Note: If you want to use TRIM, use either periodic TRIM or continuous TRIM. Continuous TRIM is not the most preferred way to issue TRIM commands among the Linux community. For example, Ubuntu enables periodic TRIM by default [2], Debian does not recommend using continuous TRIM and Red Hat recommends using periodic TRIM over using continuous TRIM if feasible [3].

As of Linux kernel version 3.8 onwards, support for TRIM was continually added for the different filesystems. See the following table for an indicative overview:

File system Continuous TRIM
(discard option)
Periodic TRIM
(fstrim)
References
and notes
Bcachefs Yes No
Btrfs Yes Yes Asynchronous discard is enabled by default since kernel 6.2.
exFAT Yes Yes fstrim is supported since kernel 5.13, [4]
ext3 Yes Yes
ext4 Yes Yes "discard, nodiscard(*)" in [5]
F2FS Yes Yes
JFS Yes Yes [6]
NILFS2 Yes Yes
NTFS Yes No ntfs3 kernel driver only supports continuous TRIM.
No Yes NTFS-3G driver only supports periodic TRIM.
VFAT Yes Yes fstrim is supported since kernel 4.19, [7]
XFS Yes Yes [8][dead link 2024-10-12 ⓘ]
Swap Yes No Not technically a "filesystem", but TRIM is still relevant. The "once" option provides scheduled trimming at startup. See swapon(8).
Warning: Users need to be certain that their SSD supports TRIM before attempting to use it. Data loss can occur otherwise!

To verify TRIM support, run:

$ lsblk --discard

And check the values of DISC-GRAN (discard granularity) and DISC-MAX (discard max bytes) columns. Non-zero values indicate TRIM support.

For SATA SSDs only, the hdparm package can detect TRIM support via hdparm -I /dev/sda | grep TRIM as the root user. hdparm does however not support NVMe SSDs.

Periodic TRIM

The util-linux package provides fstrim.service and fstrim.timer systemd unit files. Enabling the timer will activate the service weekly. The service executes fstrim(8) on all mounted filesystems on devices that support the discard operation.

The timer relies on the timestamp of /var/lib/systemd/timers/stamp-fstrim.timer (which it will create upon first invocation) to know whether a week has elapsed since it last ran. Therefore there is no need to worry about too frequent invocations, in an anacron-like fashion.

To query the units activity and status, see journalctl. To change the periodicity of the timer or the command run, edit the provided unit files.

Continuous TRIM

Instead of issuing TRIM commands once in a while (by default once a week if using fstrim.timer), it is also possible to issue TRIM commands each time files are deleted instead. The latter is known as the continuous TRIM.

Warning: Before SATA 3.1 all TRIM commands were non-queued, so continuous trimming would produce frequent system freezes. In this case, applying #Periodic TRIM less often is better alternative. Similar issue holds also for a number of devices, see ata_device_blacklist in Linux source code, for which queued TRIM command execution was blacklisted due to serious data corruption. In such case, depending on the device, the system may be forced to send non-queued TRIM commands the SSD instead of queued TRIM. See Wikipedia:Trim_(computing)#Disadvantages for details.

Using the discard option for a mount in /etc/fstab enables continuous TRIM in device operations:

/dev/sda1  /           ext4  defaults,discard   0  1
Note: Specifying the discard mount option in /etc/fstab does not work with an XFS / partition. According to this thread, it has to be set using the rootflags=discard kernel parameter.

On the ext4 filesystem, the discard flag can also be set as a default mount option using tune2fs:

# tune2fs -o discard /dev/sdXY

Using the default mount options instead of an entry in /etc/fstab is particularly useful for external drives, because such partition will be mounted with the default options also on other machines. This way, there is no need to edit /etc/fstab on every machine.

Note: The default mount options are not listed in /proc/mounts.

Trim an entire device

If you want to trim your entire SSD at once, e.g. for a new install or if you want to sell the drive, you can use the blkdiscard command.

LVM

TRIM requests that get passed from the file system to the logical volume are automatically passed to the physical volume(s). No additional configuration is necessary.

No LVM operations (lvremove, lvreduce and all others) issue TRIM requests to physical volume(s) by default. This is done to allow restoring previous volume group configuration with vgcfgrestore(8). The setting issue_discards in /etc/lvm/lvm.conf controls whether discards are sent to a logical volume's underlying physical volumes when the logical volume is no longer using the physical volumes' space.

Note: Carefully read the comment in /etc/lvm/lvm.conf before changing the issue_discards setting. It does not in any way affect TRIM requests that get passed from the file system to the disk (e.g. file deletions inside a file system) nor does it affect space management within a thin pool.
Warning: Enabling issue_discards will prevent volume group metadata restoration with vgcfgrestore. There will be no recovery options in case of a mistakenly issued LVM command.

dm-crypt

Warning: The discard option allows discard requests to be passed through the encrypted block device. This may or may not improve performance on SSD storage [9] but has security implications. See the following for more information:

Follow the instructions in dm-crypt/Specialties#Discard/TRIM support for solid state drives (SSD) to enable discard support.

swap

To activate discard for the swap space, use:

# swapon --discard

See swapon(8) for the discussion on discard=once or discard=pages.

This article or section needs expansion.

Reason: Is discard automatically enabled when using Systemd#GPT partition automounting ? (Discuss in Talk:Solid state drive)

To enable discard when booting, add discard to the options in fstab for the swap device.

Warning: Enabling discard on RAID setups using mdadm will cause system lockup on boot and during runtime, if using swapon.

Maximizing performance

Follow the tips in Improving performance#Storage devices to maximize the performance of your drives.

Sector size

See Advanced Format#NVMe solid state drives.

SSD memory cell clearing

On occasion, users may wish to completely reset an SSD's cells to the same virgin state they were at the time the device was installed, thus restoring it to its factory default write performance. Write performance is known to degrade over time even on SSDs with native TRIM support: TRIM only safeguards against file deletes, not replacements such as an incremental save.

The reset can be accomplished by following the appropriate procedure denoted in Solid state drive/Memory cell clearing, either for SATA or NVMe SSDs.

Note: If the reason for the reset is to wipe data, you may not want to rely on the SSD controller to perform it securely, e.g. if you do not trust the manufacturer or are wary of potential bugs. In this case, see Securely wipe disk#Flash memory for further information and examples to perform a manual wipe.

Security

Frozen mode

Some motherboard firmware issue a ATA SECURITY FREEZE LOCK command to SATA devices on initialization, setting the drive to frozen mode which transitions it to SEC2 state (security disabled, not locked, frozen). Likewise some SSD (and HDD) are set to this state in the factory already. This can be seen in hdparm and smartctl output:

# hdparm -I /dev/sda
Security:
 	Master password revision code = 65534
 		supported
 	not	enabled
 	not	locked
 		frozen
 	not	expired: security count
 		supported: enhanced erase
 	4min for SECURITY ERASE UNIT. 2min for ENHANCED SECURITY ERASE UNIT.
# smartctl -g security /dev/sda
ATA Security is:  Disabled, frozen [SEC2]

Operations like formatting the device or installing operating systems are not affected by the frozen mode.

The above hdparm output shows the device is not locked by a HDD-password on boot and the frozen state safeguards the device against malwares which may try to lock it by setting a password to it at runtime.

If you intend to set a password to a "frozen" device yourself, a motherboard BIOS with support for it is required. A lot of notebooks have support, because it is required for hardware encryption, but support may not be trivial for a desktop/server board. For the Intel DH67CL/BL motherboard, for example, the motherboard has to be set to "maintenance mode" by a physical jumper to access the settings.[10]

Warning: Do not try to change the above lock security settings with hdparm unless you know exactly what you are doing.

If you intend to erase the SSD, see Securely wipe disk#hdparm and /Memory cell clearing.

Setting the SATA SSD state to frozen mode after waking up from sleep

When waking up from S3 sleep, the SATA SSD will most likely have reverted to SEC1 state (security disabled, not locked, not frozen), leaving it vulnerable to ATA SECURITY ERASE UNIT commands like those described in /Memory cell clearing.

In order to prevent this issue, a script can be run after waking up from sleep:

/usr/lib/systemd/system-sleep/ssd-freeze.sh
#!/bin/sh
if [ "$1" = 'post' ]; then
	sleep 1
	if hdparm --security-freeze /dev/disk/by-id/ata-name-of-disk; then
		logger "$0: SSD freeze command executed successfully"
	else
		logger "$0: SSD freeze command failed"
	fi	
fi

If the system has multiple storage devices and/or portable USB-drives, another option is to adapt Hdparm#Persistent configuration using udev rule to issue a --security-freeze for all drives (incl. HDD).

Hardware encryption

As noted in #Frozen mode, setting a password for a storage device (SSD/HDD) in the BIOS may also initialize the hardware encryption of devices supporting it. If the device also conforms to the OPAL standard, this may also be achieved without a respective BIOS feature to set the passphrase. See Self-encrypting drives.

Troubleshooting

It is possible that the issue you are encountering is a firmware bug which is not Linux specific, so before trying to troubleshoot an issue affecting the SSD device, you should first check if updates are available for:

Even if it is a firmware bug it might be possible to avoid it, so if there are no updates to the firmware or you hesitant on updating firmware then the following might help.

Resolving NCQ errors

Some SSDs and SATA chipsets do not work properly with Linux Native Command Queueing (NCQ). The tell-tale errors in the journal look like:

ata9: exception Emask 0x0 SAct 0xf SErr 0x0 action 0x10 frozen
ata9.00: failed command: READ FPDMA QUEUED
ata9.00: cmd 60/04:00:d4:82:85/00:00:1f:00:00/40 tag 0 ncq 2048 in
res 40/00:18:d3:82:85/00:00:1f:00:00/40 Emask 0x4 (timeout)

To disable NCQ on boot, add libata.force=noncq to the kernel command line in the boot loader configuration. To disable NCQ only for disk 0 on port 9 use: libata.force=9.00:noncq

Alternatively, you may disable NCQ for a specific drive without rebooting via sysfs:

# echo 1 > /sys/block/sdX/device/queue_depth

If this (and also updating the firmware) does not resolve the problem or causes other issues, then file a bug report.

Resolving SATA power management related errors

Some SSDs (e.g. Transcend MTS400 or Crucial M550 SSDs) are failing with certain SATA controllers when SATA Active Link Power Management (ALPM), is enabled.

ALPM is enabled by default since linux-4.16, or may be enabled at runtime by a power saving daemon (e.g. TLP, Laptop Mode Tools). See Power management#SATA Active Link Power Management for more on this.

External SSD with TRIM support

This article or section needs language, wiki syntax or style improvements. See Help:Style for reference.

Reason: Several style and formatting problems, especially too many explicit commands - see Help:Style#File editing requests. (Discuss in Talk:Solid state drive)

Several USB-to-SATA bridge chips (like VL715, VL716 etc.) and also USB-to-PCIe bridge chips (like the JMicron JMS583 used in external NVMe enclosures like IB-1817M-C31) support TRIM-like commands that can be sent through the USB Attached SCSI driver (named "uas" under Linux).

But the kernel may not automatically detect this capability, and therefore might not use it. Assuming your block device in question is /dev/sdX, you can find out whether that is the case by using the command from sg3_utils:

# sg_readcap -l /dev/sdX

If in its output you find a line stating "Logical block provisioning: lbpme=0" then you know that the kernel assumes the device does not support "Logical Block Provisioning Management" because the (LBPME) bit is not set.

If this is the case, then you should next find out whether the "Vital Product Data" (VPD) page on "Logical Block Provisioning" of your device tells of supported mechanisms for unmapping data. You can do this using the command:

# sg_vpd -a /dev/sdX

Look for lines in the output that look like this:

Unmap command supported (LBPU): 1
Write same (16) with unmap bit supported (LBPWS): 0
Write same (10) with unmap bit supported (LBPWS10): 0

This example would tell you the device supports the "UNMAP" command.

Have a look at the output of

$ cat /sys/block/sdX/device/scsi_disk/*/provisioning_mode

If the kernel did not detect the capability of your device to unmap data, then this will likely return "full". Apart from "full", the kernel SCSI storage driver currently knows the following values for provisioning_mode:

unmap
writesame_16
writesame_10
writesame_zero
disabled

For the example above, you could now write "unmap" to "provisioning_mode" to ask the kernel to use that:

# echo "unmap" >/sys/block/sdX/device/scsi_disk/*/provisioning_mode

This should immediately enable you to use tools like "blkdiscard" on /dev/sdX or "fstrim" on filesystems mounted on /dev/sdX.

If you want to enable a "provisioning_mode" automatically when an external device of a certain vendor/product is attached, this can be automated via the "udev" mechanism. First find the USB Vendor and Product IDs:

$ cat /sys/block/sdX/../../../../../../idVendor
$ cat /sys/block/sdX/../../../../../../idProduct

Then create or append to a udev rule file (example here using idVendor 152d and idProduct 0583):

# echo 'ACTION=="add|change", ATTRS{idVendor}=="152d", ATTRS{idProduct}=="0583", SUBSYSTEM=="scsi_disk", ATTR{provisioning_mode}="unmap"' >>/etc/udev/rules.d/10-uas-discard.rules

(You can also use the lsusb command to look for the relevant idVendor/idProduct.)

Firmware

If supported by the device vendor, it is recommended to update firmware using the fwupd utility.

To check your current firmware version:

# smartctl -i /dev/ssd_device

ADATA

Updating SSD firmware under Linux is not supported by ADATA. A Windows-only utility called SSD ToolBox is provided by ADATA through their support page and through their ADATA XPG support page to monitor, TRIM, benchmark and update ADATA SSD firmware.

Warning: It is discouraged to attempt to update firmware through Wine as it is not designed to handle hardware interfaces and an incomplete firmware update can potentially brick your device.

Crucial

Crucial provides an option for updating the firmware with an ISO image. These images can be found after selecting the product on their SSD support page and downloading the "Manual Boot File."

Note: ISO images provided by Crucial do not seem to be hybrid. If you will use just the dd command to copy the image to some device, the MBR will not be present, making such device unbootable. Install syslinux and run isohybrid path/to/image.iso.

Owners of an M4 Crucial model, may check if a firmware upgrade is needed with smartctl.

$ smartctl --all /dev/sdX
==> WARNING: This drive may hang after 5184 hours of power-on time:
https://www.tomshardware.com/news/Crucial-m4-Firmware-BSOD,14544.html
See the following web page for firmware updates:
https://www.crucial.com/usa/en/support-ssd

Users seeing this warning are advised to backup all sensible data and consider upgrading immediately. Check this instructions to update Crucial MX100 firmware by using the ISO image and Grub.

Intel

Intel has a Linux live system based Firmware Update Tool for operating systems that are not compatible with its Windows Intel® Memory and Storage Tool (GUI) software.

There is also a newer Linux command-line utility that can reflash firmware called the Intel Memory and Storage (MAS) Tool available as intel-mas-cli-toolAUR. There is a PDF user guide[dead link 2024-07-30 ⓘ] available.

An example for checking the firmware status is:

# intelmas show -intelssd 0
DevicePath : /dev/nvme0n1
DeviceStatus : Healthy
Firmware : 002C
FirmwareUpdateAvailable : The selected Intel SSD contains current firmware as of this tool release.

-intelssd 0 can be omitted if there is only one Intel SSD in the system, or 1 passed for the second SSD, and so on.

If an update is available, it is performed by running intelmas load -intelssd 0. The PDF user guide suggests that this procedure needs to be performed twice in Linux, with a power cycle in between. The latest firmware for all devices is distributed as part of the MAS Tool itself, so does not need to be downloaded separately.

Kingston

KFU tool is available for the Sandforce based drives, kingston_fw_updaterAUR.

Mushkin

The lesser known Mushkin brand solid state drives also use Sandforce controllers, and have a Linux utility (nearly identical to Kingston's) to update the firmware.

OCZ

OCZ has a Command Line Online Update Tool (CLOUT) available for Linux. The existing packages are ocz-ssd-utilityAUR, ocztoolboxAUR and oczcloutAUR.

Samsung

Although Samsung deems firmware update methods outside of their Magician software as "unsupported", they still can work. The Magician software can create a bootable USB drive containing the firmware update, however Samsung no longer provides the software for consumer SSDs. Samsung also provides pre-made bootable ISO images that can be used to update the firmware. Another option is to use Samsung's magician utility provided by samsung_magician-consumer-ssdAUR. Magician only supports Samsung-branded SSDs; those manufactured by Samsung for OEMs (e.g., Lenovo) are not supported.

Note: Samsung does not make it obvious at all that they actually provide these. They seem to have 4 different firmware update pages and each references different ways of doing things.

Users preferring to run the firmware update from a live USB created under Linux (without using Samsung's Magician software under Microsoft Windows) can refer to [11] for more details. Note that this blog post details creating a bootable USB drive with Master Boot Record (MBR) that some newer motherboards, e.g. Intel NUC no longer support.

Update under Linux

The SSD firmware can be updated natively (without making a bootable USB stick) as shown below. First visit the Samsung downloads page, go to the "Samsung SSD Firmware" section, and download the latest firmware for your SSD—it should be an ISO image.

Note: Some ISO images will lack the initrd Linux image mentioned below. See #Older SSDs instead.

Extract the initrd Linux image from the ISO image:

$ bsdtar xf samsung_ssd_firmware.iso initrd

Extract root/fumagician/. This directory contains the firmware update files:

$ bsdtar xf initrd root/fumagician

Finally, run root/fumagician/fumagician with root privileges and reboot your system (if the firmware was successfully updated).

If after reboot the firmware version does not change, run root/fumagician/fumagician 2> log and search for errors in the log file. For example, if the log shows 'unzip is not available', install unzip or extract it from the initrd.

Older SSDs

Some of the SSD firmware ISO images contain a FreeDOS image instead of an initrd Linux image, so the steps needed to update the SSD firmware differ from above. The following table lists these SSDs (and relevant paths):

SSD model FreeDOS image path Firmware package path
470, 830 BTDSK.IMG SSR/
840 isolinux/btdsk.img samsung/DSRD/
840 EVO (mSATA), Pro ISOLINUX/BTDSK.IMG

First, extract the FreeDOS image from the ISO image:

$ bsdtar xf samsung_ssd_firmware.iso freedos_image_path

Mount the FreeDOS image to /mnt/:

# mount freedos_image_path /mnt

Get the disk number of the SSD under Disk Number from the Magician SSD management utility:

# magician --list

Update the SSD firmware for the specified disk by providing the firmware package path:

# magician --disk disk_num --firmware-update --fwpackage-path /mnt/firmware_package_path

Finally, verify whether the firmware was successfully updated by checking the version under Firmware from the output of magician --list (with root privileges). Reboot your system if so.

SanDisk

SanDisk makes ISO firmware images to allow SSD firmware update on operating systems that are unsupported by their SanDisk SSD Toolkit.

One must choose the firmware for the correct SSD model, and the correct capacity that it has (e.g. 60GB, or 256GB). After burning the ISO firmware image, simply restart the PC to boot with the newly created CD/DVD boot disk (may work from a USB stick).

The iso images just contain a linux kernel and an initrd. Extract them to /boot partition and boot them with GRUB or Syslinux to update the firmware.

See also:

See also