Files
scst/iscsi-scst/ChangeLog
Vladislav Bolkhovitin 73e27412fb - Update for 2.6.26
- Zero-copy feature, introduced by put_page_callback patch, made kernel compile time configurable via CONFIG_TCP_ZERO_COPY_TRANSFER_COMPLETION_NOTIFICATION option. If you upgrade from earlier versions, see the corresponding "[CAUTION]" e-mail in scst-devel mailing list about possible compatibility problems.
 - SRP target temporary removed from "all" and "install" the main Makefile targets, since it isn't compiled on 2.6.26
 - Minor fixes and cleanups



git-svn-id: http://svn.code.sf.net/p/scst/svn/trunk@475 d57e44dd-8a1f-0410-8b47-8ef2f437770f
2008-07-24 09:22:55 +00:00

60 lines
2.3 KiB
Plaintext

Summary of changes in iSCSI-SCST between versions 1.0.0 and 1.0.1
-----------------------------------------------------------------
- Update for 2.6.26
- Zero-copy feature, introduced by put_page_callback patch made
kernel compile time configurable via
CONFIG_TCP_ZERO_COPY_TRANSFER_COMPLETION_NOTIFICATION option. If you
upgrade from earlier versions, see the corresponding "[CAUTION]"
e-mail in scst-devel mailing list about possible compatibility
problems.
Summary of changes in iSCSI-SCST since it was IET
-------------------------------------------------
- Commands processing flow changed to work with SCST. Significant
cleanup was done. Modules and config files were renamed to allow to
run with IET on the same host.
- In IET the iSCSI negotiation isn't fully iSCSI RFC confirmed: it doesn't
support ranges of values and, more important, violates RFC in case
when in the IET config file some value for some parameter is set and
a remote initiator doesn't initiate the negotiation for this
parameter or declare its value. According to RFC, in this case IET
shall use the RFC-specified default value, but it uses config file
specified one instead. Looks like the implementation confuses IET
config file default values and iSCSI RFC ones. The default values
handling was fixed. But support for ranges remains unfixed (see ToDo
file).
- All shutdown/restart problems (resource leaks, hangups, etc.), especially
under load, were fixed.
- Full duplex network IO implemented.
- Threading model reimplemented.
- Digests calculation made multi-threaded.
- TX data digest errors handling iSCSI RFC violation was fixed. Another one
(SNACK handling) remains unfixed.
- Ability to send data with sense in a single response added.
- Ability to handle initiators misbehavior a bit improved: several BUG()'s
were relaced by the proper handling. Hovewer, there is a plenty of work in
this area left, IET is known to have a lot of weaknesses here.
- Task management made more robust. Particularly, possible data
corruption scenarios were fixed.
- Implemented SN based iSCSI IO flow control
- Response data send timeout implemented: now if initiator doesn't
accept data for too long (30 sec), target closes connection
- A lot of other bugfixes and code cleanups.