3 files changed, 75 insertions, 0 deletions
diff --git a/open_issues/performance/io_system.mdwn b/open_issues/performance/io_system.mdwn
new file mode 100644
index 00000000..bf59b5df
--- /dev/null
+++ b/open_issues/performance/io_system.mdwn
@@ -0,0 +1,42 @@
+[[!meta copyright="Copyright © 2008, 2009, 2010 Free Software Foundation,
+Inc."]]
+
+[[!meta license="""[[!toggle id="license" text="GFDL 1.2+"]][[!toggleable
+id="license" text="Permission is granted to copy, distribute and/or modify this
+document under the terms of the GNU Free Documentation License, Version 1.2 or
+any later version published by the Free Software Foundation; with no Invariant
+Sections, no Front-Cover Texts, and no Back-Cover Texts.  A copy of the license
+is included in the section entitled
+[[GNU Free Documentation License|/fdl]]."]]"""]]
+
+[[!meta title="I/O System"]]
+
+[[!tag open_issue_hurd]]
+
+The most obvious reason for the Hurd feeling slow compared to mainstream
+systems like GNU/Linux, is a low I/O system performance, in particular very
+slow hard disk access.
+
+The reason for this slowness is lack and/or bad implementation of common
+optimization techniques, like scheduling reads and writes to minimize head
+movement; effective block caching; effective reads/writes to partial blocks;
+reading/writing multiple blocks at once; and read-ahead.  The
+[[ext2_filesystem_server|hurd/translator/ext2fs]] might also need some
+optimizations at a higher logical level.
+
+The goal of this project is to analyze the current situation, and implement/fix
+various optimizations, to achieve significantly better disk performance. It
+requires understanding the data flow through the various layers involved in
+disk access on the Hurd ([[filesystem|hurd/virtual_file_system]],
+[[pager|hurd/libpager]], driver), and general experience with
+optimizing complex systems.  That said, the killing feature we are definitely
+missing is the read-ahead, and even a very simple implementation would bring
+very big performance speedups.
+
+Here's a real use-case: [[binutils_ld_64ksec]].
+
+Possible mentors: Samuel Thibault (youpi)
+
+Exercise: Look through all the code involved in disk I/O, and try something
+easy to improve. It's quite likely though that you will find nothing obvious --
+in this case, please contact us about a different exercise task.
diff --git a/open_issues/performance/io_system/binutils_ld_64ksec.mdwn b/open_issues/performance/io_system/binutils_ld_64ksec.mdwn
new file mode 100644
index 00000000..60dca510
--- /dev/null
+++ b/open_issues/performance/io_system/binutils_ld_64ksec.mdwn
@@ -0,0 +1,33 @@
+[[!meta copyright="Copyright © 2010 Free Software Foundation, Inc."]]
+
+[[!meta license="""[[!toggle id="license" text="GFDL 1.2+"]][[!toggleable
+id="license" text="Permission is granted to copy, distribute and/or modify this
+document under the terms of the GNU Free Documentation License, Version 1.2 or
+any later version published by the Free Software Foundation; with no Invariant
+Sections, no Front-Cover Texts, and no Back-Cover Texts.  A copy of the license
+is included in the section entitled [[GNU Free Documentation
+License|/fdl]]."]]"""]]
+
+[[!tag open_issue_hurd]]
+
+This one may be considered as a testcase for I/O system optimization.
+
+It is taken from the [[binutils_testsuite]], `ld/ld-elf/sec64k.exp`, where this
+test may occasionally [[trigger a timeout|binutils_testsuite#64ksec]].  It is
+extracted from cdf7c161ebd4a934c9e705d33f5247fd52975612 sources, 2010-10-24.
+
+    $ wget -O - http://www.gnu.org/software/hurd/open_issues/performance/io_system/binutils_ld_64ksec/test.tar.xz | xz -d | tar -x
+    $ cd test/
+    $ \time ./ld-new.stripped -o dump dump?.o dump??.o
+    0.00user 0.00system 2:46.11elapsed 0%CPU (0avgtext+0avgdata 0maxresident)k
+    0inputs+0outputs (0major+0minor)pagefaults 0swaps
+
+On the idle grubber, this one repeatedly takes a few minutes wall time to
+complete successfully, contrary to a few seconds on a GNU/Linux system.
+
+While processing the object files, there is heavy interaction with the relevant
+[[hurd/translator/ext2fs]] process .  Running [[hurd/debugging/rpctrace]] on
+the testee shows that (primarily) an ever-repeating series of `io_seek` and
+`io_read` is being processed.  Running the testee on GNU/Linux with strace
+shows the equivalent thing (`_llseek`, `read`) -- but Linux' I/O system isn't
+as slow as the Hurd's.
diff --git a/open_issues/performance/io_system/binutils_ld_64ksec/test.tar.xz b/open_issues/performance/io_system/binutils_ld_64ksec/test.tar.xz
new file mode 100644
index 00000000..6d7c606c
--- /dev/null
+++ b/open_issues/performance/io_system/binutils_ld_64ksec/test.tar.xz