4 This text file will detail the queue files that are located in the sysfs tree
5 for each block device. Note that stacked devices typically do not export
6 any settings, since their queue merely functions are a remapping target.
7 These files are the ones found in the /sys/block/xxx/queue/ directory.
9 Files denoted with a RO postfix are readonly and the RW postfix means
14 This file allows to turn off the disk entropy contribution. Default
15 value of this file is '1'(on).
19 This file indicates whether the device supports Direct Access (DAX),
20 used by CPU-addressable storage to bypass the pagecache. It shows '1'
23 discard_granularity (RO)
24 -----------------------
25 This shows the size of internal allocation of the device in bytes, if
26 reported by the device. A value of '0' means device does not support
27 the discard functionality.
29 discard_max_hw_bytes (RO)
30 ----------------------
31 Devices that support discard functionality may have internal limits on
32 the number of bytes that can be trimmed or unmapped in a single operation.
33 The discard_max_bytes parameter is set by the device driver to the maximum
34 number of bytes that can be discarded in a single operation. Discard
35 requests issued to the device must not exceed this limit. A discard_max_bytes
36 value of 0 means that the device does not support discard functionality.
38 discard_max_bytes (RW)
39 ----------------------
40 While discard_max_hw_bytes is the hardware limit for the device, this
41 setting is the software limit. Some devices exhibit large latencies when
42 large discards are issued, setting this value lower will make Linux issue
43 smaller discards and potentially help reduce latencies induced by large
48 This is the hardware sector size of the device, in bytes.
52 When read, this file shows whether polling is enabled (1) or disabled
53 (0). Writing '0' to this file will disable polling for this device.
54 Writing any non-zero value will enable this feature.
58 If polling is enabled, this controls what kind of polling will be
59 performed. It defaults to -1, which is classic polling. In this mode,
60 the CPU will repeatedly ask for completions without giving up any time.
61 If set to 0, a hybrid polling mode is used, where the kernel will attempt
62 to make an educated guess at when the IO will complete. Based on this
63 guess, the kernel will put the process issuing IO to sleep for an amount
64 of time, before entering a classic poll loop. This mode might be a
65 little slower than pure classic polling, but it will be more efficient.
66 If set to a value larger than 0, the kernel will put the process issuing
67 IO to sleep for this amount of microseconds before entering classic
72 io_timeout is the request timeout in milliseconds. If a request does not
73 complete in this time then the block driver timeout handler is invoked.
74 That timeout handler can decide to retry the request, to fail it or to start
75 a device recovery strategy.
79 This file is used to control (on/off) the iostats accounting of the
82 logical_block_size (RO)
83 -----------------------
84 This is the logical block size of the device, in bytes.
86 max_hw_sectors_kb (RO)
87 ----------------------
88 This is the maximum number of kilobytes supported in a single data transfer.
90 max_integrity_segments (RO)
91 ---------------------------
92 When read, this file shows the max limit of integrity segments as
93 set by block layer which a hardware controller can handle.
97 This is the maximum number of kilobytes that the block layer will allow
98 for a filesystem request. Must be smaller than or equal to the maximum
99 size allowed by the hardware.
103 Maximum number of segments of the device.
105 max_segment_size (RO)
106 ---------------------
107 Maximum segment size of the device.
111 This is the smallest preferred IO size reported by the device.
115 This enables the user to disable the lookup logic involved with IO
116 merging requests in the block layer. By default (0) all merges are
117 enabled. When set to 1 only simple one-hit merges will be tried. When
118 set to 2 no merge algorithms will be tried (including one-hit or more
119 complex tree/hash lookups).
123 This controls how many requests may be allocated in the block layer for
124 read or write requests. Note that the total allocated number may be twice
125 this amount, since it applies only to reads or writes (not the accumulated
128 To avoid priority inversion through request starvation, a request
129 queue maintains a separate request pool per each cgroup when
130 CONFIG_BLK_CGROUP is enabled, and this parameter applies to each such
131 per-block-cgroup request pool. IOW, if there are N block cgroups,
132 each request queue may have up to N request pools, each independently
133 regulated by nr_requests.
137 This is the optimal IO size reported by the device.
139 physical_block_size (RO)
140 ------------------------
141 This is the physical block size of device, in bytes.
145 Maximum number of kilobytes to read-ahead for filesystems on this block
150 This file is used to stat if the device is of rotational type or
155 If this option is '1', the block layer will migrate request completions to the
156 cpu "group" that originally submitted the request. For some workloads this
157 provides a significant reduction in CPU cycles due to caching effects.
159 For storage configurations that need to maximize distribution of completion
160 processing setting this option to '2' forces the completion to run on the
161 requesting cpu (bypassing the "group" aggregation logic).
165 When read, this file will display the current and available IO schedulers
166 for this block device. The currently active IO scheduler will be enclosed
167 in [] brackets. Writing an IO scheduler name to this file will switch
168 control of this block device to that new IO scheduler. Note that writing
169 an IO scheduler name to this file will attempt to load that IO scheduler
170 module, if it isn't already present in the system.
174 When read, this file will display whether the device has write back
175 caching enabled or not. It will return "write back" for the former
176 case, and "write through" for the latter. Writing to this file can
177 change the kernels view of the device, but it doesn't alter the
178 device state. This means that it might not be safe to toggle the
179 setting from "write back" to "write through", since that will also
180 eliminate cache flushes issued by the kernel.
182 write_same_max_bytes (RO)
183 -------------------------
184 This is the number of bytes the device can write in a single write-same
185 command. A value of '0' means write-same is not supported by this
190 If the device is registered for writeback throttling, then this file shows
191 the target minimum read latency. If this latency is exceeded in a given
192 window of time (see wb_window_usec), then the writeback throttling will start
193 scaling back writes. Writing a value of '0' to this file disables the
194 feature. Writing a value of '-1' to this file resets the value to the
197 throttle_sample_time (RW)
198 -------------------------
199 This is the time window that blk-throttle samples data, in millisecond.
200 blk-throttle makes decision based on the samplings. Lower time means cgroups
201 have more smooth throughput, but higher CPU overhead. This exists only when
202 CONFIG_BLK_DEV_THROTTLING_LOW is enabled.
206 This indicates if the device is a zoned block device and the zone model of the
207 device if it is indeed zoned. The possible values indicated by zoned are
208 "none" for regular block devices and "host-aware" or "host-managed" for zoned
209 block devices. The characteristics of host-aware and host-managed zoned block
210 devices are described in the ZBC (Zoned Block Commands) and ZAC
211 (Zoned Device ATA Command Set) standards. These standards also define the
212 "drive-managed" zone model. However, since drive-managed zoned block devices
213 do not support zone commands, they will be treated as regular block devices
214 and zoned will report "none".
218 For zoned block devices (zoned attribute indicating "host-managed" or
219 "host-aware"), this indicates the total number of zones of the device.
220 This is always 0 for regular block devices.
224 This has different meaning depending on the type of the block device.
225 For a RAID device (dm-raid), chunk_sectors indicates the size in 512B sectors
226 of the RAID volume stripe segment. For a zoned block device, either host-aware
227 or host-managed, chunk_sectors indicates the size in 512B sectors of the zones
228 of the device, with the eventual exception of the last zone of the device which
231 Jens Axboe <jens.axboe@oracle.com>, February 2009