summaryrefslogtreecommitdiffstats
path: root/libnetdata/procfile/README.md
diff options
context:
space:
mode:
authorDaniel Baumann <daniel.baumann@progress-linux.org>2018-11-07 12:19:29 +0000
committerDaniel Baumann <daniel.baumann@progress-linux.org>2018-11-07 12:20:17 +0000
commita64a253794ac64cb40befee54db53bde17dd0d49 (patch)
treec1024acc5f6e508814b944d99f112259bb28b1be /libnetdata/procfile/README.md
parentNew upstream version 1.10.0+dfsg (diff)
downloadnetdata-a64a253794ac64cb40befee54db53bde17dd0d49.tar.xz
netdata-a64a253794ac64cb40befee54db53bde17dd0d49.zip
New upstream version 1.11.0+dfsgupstream/1.11.0+dfsg
Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>
Diffstat (limited to '')
-rw-r--r--libnetdata/procfile/README.md61
1 files changed, 61 insertions, 0 deletions
diff --git a/libnetdata/procfile/README.md b/libnetdata/procfile/README.md
new file mode 100644
index 000000000..279885f93
--- /dev/null
+++ b/libnetdata/procfile/README.md
@@ -0,0 +1,61 @@
+
+# PROCFILE
+
+procfile is a library for reading text data files (i.e `/proc` files) in the fastest possible way.
+
+## How it works
+
+The library automatically adapts (through the iterations) its memory so that each file
+is read with single `read()` call.
+
+Then the library splits the file into words, using the supplied separators.
+The library also supported quoted words (i.e. strings within of which the separators are ignored).
+
+### Initialization
+
+Initially the caller:
+
+- calls `procfile_open()` to open the file and allocate the structures needed.
+
+### Iterations
+
+For each iteration, the caller:
+
+- calls `procfile_readall()` to read updated contents.
+ This call also rewinds (`lseek()` to 0) before reading it.
+
+ For every file, a [BUFFER](../buffer/) is used that is automatically adjusted to fit
+ the entire file contents of the file. So the file is read with a single `read()` call
+ (providing atomicity / consistency when the data are read from the kernel).
+
+ Once the data are read, 2 arrays of pointers are updated:
+
+ - a `words` array, pointing to each word in the data read
+ - a `lines` array, pointing to the first word for each line
+
+ This is highly optimized. Both arrays are automatically adjusted to
+ fit all contents and are updated in a single pass on the data.
+
+ The library provides a number of macros:
+
+ - `procfile_lines()` returns the # of lines read
+ - `procfile_linewords()` returns the # of words in the given line
+ - `procfile_word()` returns a pointer the given word #
+ - `procfile_line()` returns a pointer to the first word of the given line #
+ - `procfile_lineword()` returns a pointer to the given word # of the given line #
+
+### Cleanup
+
+When the caller exits:
+
+- calls `procfile_free()` to close the file and free all memory used.
+
+### Performance
+
+- a **raspberry Pi 1** (the oldest single core one) can process 5.000+ `/proc` files per second.
+- a **J1900 Celeron** processor can process 23.000+ `/proc` files per second per core.
+
+To achieve this kind of performance, the library tries to work in batches so that the code
+and the data are inside the processor's caches.
+
+This library is extensively used in netdata and its plugins.