- MPSAFE flag no longer present in syscalls.master, all system calls are
MPSAFE.
- printf(9) synchronization.
- Mutex profiling has become lock profiling.
and resort some current content into them. Add a note that this page
is largely historic (and especially that the task list may be out of
date), since we mostly run without Giant now. Move credits to sponsors
higher up, and expand. We now run on 24/32-thread sun4v hardware, bump
the process/thread count.
level bullets to recognize completed work:
- The network stack infrastructure has been MPSAFE for several years.
Mark as done.
- The NFS client has been MPSAFE since May. Mark as done.
- Move includes.nav*.sgml to share/sgml/navibar.ent and
<lang>/share/sgml/nabibar.l10n.ent.
- Move includes.sgml and includes.xsl to
share/sgml/common.ent, share/sgml/header.ent, <lang>/share/sgml/l10n.ent,
and <lang>?share/sgml/header.l10n.ent.
- Move most of XSLT libraries to share/sgml/*.xsl and
<lang>/share/sgml/*.xsl.
- Move news.xml and other *.xml files for the similar purpose
to share/sgml/*.xml and <lang>/share/sgml/*.xml.
- Switch to use a custom DTD for HTML document. Now we use
"-//FreeBSD//DTD HTML 4.01 Transitional-Based Extension", which is
HTML 4.01 + some entities previously pulled via
"<!ENTITY % includes SYSTEM "includes.sgml"> %includes;" line.
The location of entity file will be resolved by using catalog file.
- Add DOCTYPE declearation to XML documents. This makes the followings
possible:
* Use of &foo; entities for SGML in an XML file instead of defining
{$foo} as the same content.
* &symbolic; entities for Latin characters.
- Duplicated information between SGML and XML, or English and
translated doc, has been removed as much as possible.
- Update the tasks to push Giant down into the various compat ABIs (really
sync up the MPSAFE flags) as done.
- Update the task to add witness checks on thread exit as done.
- Update the atomic_fetchadd() and refcount API tasks as done.
- Add a new WIP task describing the intr_event stuff.
- Discard some poorly thought not started tasks (or ones OBE).
- iedowse added ktr tracing for mutex contention.
- jeff allowed ktr to trace to a file using alq.
- I added ktr tracing for UMA, geom, busdma, critical sections.
- jeff added ktr tracing for scheduler run queues, buffer cache events,
additional VFS events.
critical sections for synchronizing access to UMA per-cpu caches (rwatson),
and converting malloc(9) to per-CPU statistics protected by critical
sections (rwatson).
done. That was in 2002.
Mark cdevsw flag for MPSAFEty as done, phk did it in 2002.
Mark linker as locked down in 2002.
Mark a number of VFS-related activities as done on 01 January 2005, as
that's when Jeff moved ahead with the MPSAFE VFS default.
Add completed task for making UFS MPSAFE.
will no doubt follow, but marking it as "wip" is misleading.
Separate IPv4, IPv6 tasks. Mark IPv4 as done, as it works quite well,
and leave IPv6 as WIP, since there are some important loose ends that
need to be covered.
Mark IPX/SPX as done.
Mark netatalk as done.
Mark NFS client as a work in progress--mohans has the network pieces
running MPSAFE, but the VFS MPSAFEty needs to happen still.
- Mention KSE 1:1 and M:N threading as occuring in parallel, allowing
applications to take advantage of parallelism.
- Mention WITNESS.
- Put in a specific recognition of BSD/OS-BSDi contributions early in the
project.
- Since a lot of the development work is done, focus less on the project
management approach (especially since things have diverged a lot from
what was discussed).
- Point at the Netperf cluster.
with a new introduction that reflects the work that has gone into the 5.x
and 6.x branches. In particular, update the status to reflect MPSAFE
networking, MPSAFE VFS, MPSAFE VM, on-going performance work, and some of
the important steps to get there. Highlight the iterative nature of the
work, and the focus on infrastructure.
Quoting trhodes:
``If we cannot produce the article,
it should be removed.''
PR: www/76817
Noticed by: Radko Keves <rado@daemon.sk>
Submitted by: Dan Gerzo <danger@rulez.sk>
Approved by: trhodes (mentor)
jail is in fact not an issue due to integer reads being atomic and
the fact that any races there would exist regardless of the presence
of synchronization. Remove from the todo list.
* correct rwatson's SGML author entity so his name is
displayed correctly on the page.
* mention Rick Macklem, not only his e-mail address;
update the last-modified date to reflect this change.
* fix the task line for the TTY subsystem so it is displayed
correctly on the page and also add a `Last updated' date
to the task.
