CVS log for src/sys/dev/iscsi/iscsi_main.c

Up to [cvs.NetBSD.org] / src / sys / dev / iscsi

Request diff between arbitrary revisions

Keyword substitution: kv
Default branch: MAIN

Pull up following revision(s) (requested by rin in ticket #997):

	sys/dev/iscsi/iscsi_ioctl.c: revision 1.35
	sys/dev/iscsi/iscsi_main.c: revision 1.42
	sys/dev/iscsi/iscsi_send.c: revision 1.41

Use correct status value SCSI_BUSY (0x08) instead of XS_BUSY (7) when
running out of sessions. The bug had no impact as scsipi was only comparing
against SCSI_CHECK (0x02) and SCSI_QUEUE_FULL (0x28).

Avoid race in timeout handling.

Don't try to wake up CCB without connection (which led to a NULL pointer
deref).

Use correct status value SCSI_BUSY (0x08) instead of XS_BUSY (7) when running
out of sessions. The bug had no impact as scsipi was only comparing
against SCSI_CHECK (0x02) and SCSI_QUEUE_FULL (0x28).

kill_session now uses the session id to avoid stale session pointers.
protect network socket with rwlock to handle recconnects.
always take over socket from iscsid to prevent leaks.
keep a good connection alive.
don't forget child device when config_detach fails.
fix locking when reassigning CCBs.
pducount is protected by lock, no need for atomic.
some code rework, refined debug messages.

Split some common stuff into scsi_subr module.  This enables loading
of the iscsi module whether or not there are any scsi things built
into the kernel.

Addresses the iscsi portion of kern/56772

For device modules that provide both auto-config and /dev/xxx
interfaces, make sure that initialization and destruction
follow the proper sequence.  This is triggered by the recent
changes to the devsw stuff; per riastradh@ the required call
sequence is:

	devsw_attach()
	config_init_component() or config_cf*_attach()
	...
	config_fini_component() or config_cf*_detach()
	devsw_detach()

While here, add a few missing calls to some of the detach
routines.

Testing of these changes has been limited to:
	1. compile without build break
	2. no related test failures from atf
	3. modload/modunload work as well as
	   before.

No functional device testing done, since I don't have any
of these devices.  Let me know of any damage I might cause
here!

XXX Some of the modules affected by this commit are already
XXX broken;  see kern/56772.  This commit does not break
any additional modules (as far as I know).

iscsi(4): Fix config_found usage -- needs kernel lock.

Candidate fix for PR kern/56709.

Merge thorpej-cfargs2.

Adapt to CFARGS().

Sync w/ HEAD.

sync with head

Provide missing file ops. Before a simple read() would panic the kernel.
No longer fail close().

Merge thorpej-cfargs branch:

Simplify and make extensible the config_search() / config_found() /
config_attach() interfaces: rather than having different variants for
which arguments you want pass along, just have a single call that
takes a variadic list of tag-value arguments.

Adjust all call sites:
- Simplify wherever possible; don't pass along arguments that aren't
  actually needed.
- Don't be explicit about what interface attribute is attaching if
  the device only has one.  (More simplification.)
- Add a config_probe() function to be used in indirect configuiration
  situations, making is visibly easier to see when indirect config is
  in play, and allowing for future change in semantics.  (As of now,
  this is just a wrapper around config_match(), but that is an
  implementation detail.)

Remove unnecessary or redundant interface attributes where they're not
needed.

There are currently 5 "cfargs" defined:
- CFARG_SUBMATCH (submatch function for direct config)
- CFARG_SEARCH (search function for indirect config)
- CFARG_IATTR (interface attribte)
- CFARG_LOCATORS (locators array)
- CFARG_DEVHANDLE (devhandle_t - wraps OFW, ACPI, etc. handles)

...and a sentinel value CFARG_EOL.

Add some extra sanity checking to ensure that interface attributes
aren't ambiguous.

Use CFARG_DEVHANDLE in MI FDT, OFW, and ACPI code, and macppc and shark
ports to associate those device handles with device_t instance.  This
will trickle trough to more places over time (need back-end for pre-OFW
Sun OBP; any others?).

Sync with HEAD.

Give config_found() the same variadic arguments treatment as
config_search().  This commit only adds the CFARG_EOL sentinel
to the existing config_found() calls.  Conversion of config_found_sm_loc()
and config_found_ia() call sites will be in subsequent commits.

Stop advertising more openings after limit is reached.

make the iscsi attached message aprint_verbose(), no need to show normally

Mostly merge changes from HEAD upto 20200411

in many device attach paths, allocate memory with KM_SLEEP instead of KM_NOSLEEP
and remove code to handle failures that can no longer happen.

Many years ago someone created a new __link_set_sysctl_funcs to hold
the list of routines that need to be called for setting up sysctl
variables.  This worked great for all code included in the kernel
itself, but didn't deal with modules that want to create their own
sysctl data.  So, we ended up with a lot of #ifdef _MODULE blocks
so modules could explicitly call their setup functions when loaded
as non-built-in modules.

So today, we complete the task that was started so many years ago.

When modules are loaded, after we've called xxx_modcmd(INIT...) we
check if the module contains its own __link_set_sysctl_funcs, and
if so we call the functions listed.  We add a struct sysctllog member
to the struct module so we can call sysctl_teardown() when the module
gets unloaded.  (The sequence of events ensures that the sysctl stuff
doesn't get created until the rest of the module's init code does any
required memory allocation.)

So, no more need to explicitly call the sysctl setup routines when
built as a loadable module.

register with pmf to allow suspend.

Sync with HEAD

Replace build option to enable hex encoded bignum parameters with a sysctl.

Fix CVS Id

NFCI

Pull up following revision(s) (requested by mlelstv in ticket #437):
	sys/dev/iscsi/iscsi_ioctl.c: 1.26-1.27
	sys/dev/iscsi/iscsi_main.c: 1.26
	sys/dev/iscsi/iscsi_send.c: 1.35
Fix session cleanup.
--
add debug messages
-
use same lock for ref/deref.
--
unreference session only for responses to SCSI commands.

- add a prefix to struct members (but not the ioctl ones because they affect
  the API)
- use consistent variable names.

update from HEAD

unreference session only for responses to SCSI commands.

add fo_name so we can identify the fileops in a simple way.

At suggestion of chuq@, modify config_attach_pseudo() to return with a
reference held on the device.

Adapt callers to expect the reference to exist, and to ensure that the
reference is released.

Sync with HEAD

Sync with HEAD.  (Note that most of these changes are simply $NetBSD$
tag issues.)

Lock correctly around CV calls.
Fix handling of session termination.
Enable MPSAFE processing for scsipi.

Sync with HEAD

Remove throttling code, instead signal scsipi layer to reduce the openings
and retry the command. Start with a small openings number and let scsipi
request to grow it up to the current send window.

Adjust ccb and pdu counts to avoid ressource shortages. These are still
very ad-hoc numbers, but seem to be sufficient for a Gigabit link.

Use separate condvar for PDU pool and add counter to help debugging.

Revert setting PDU disposition to UNUSED before freeing. free_pdu
uses this as a flag to detect already returned PDUs.

Add reference counter for open commands to defer unmapping a session
that would lead to crashes in scsipi.

Move session cleanup to cleanup thread.

Use get_sernum to retrieve current serial number where possible and
make it check for immediate commands itself.

Adjust debug output.

Keep 1 CCB free for private use. This avoids frequent recovery in scsipi when
resources appear to be exhausted.

Tear down sysctls only after detaching children. This lets you enable
debugging while detaching hangs.

Whitespace

More debugging output, changed debug levels for some.

Several improvements to the ISCSI driver.

- Enable debug messages but set log level to be quiet. Provide a
  system (hw.iscsi.debug) to set the log level at run time.
- Replace old tsleep/wakeup synchronization with mutexes and condvars.
- Defer actions from callouts (basically timeouts) to the cleanup thread.
- Protect lists and unique ids with mutexes. protect connection usecount
  by using atomic operations.
- Assert kernel lock when calling into scsipi and network code.
- Use this to make send/receive/cleanup threads MPSAFE.

- Fix handling of out-of-CCB/out-of-PDU conditions against the scsipi layer.
- Bump number of PDUs to 128 to avoid virtually all out-of-PDU conditions

- Make use of softc structure for attach/detach operations.
- Track open file handles to prevent detach when busy.

- Move some global variables to make them static.

- Fix 'Overlapping Commands Attempted' error by marking commands as
  simply ordered (ATTR_SIMPLE) like FreeBSD.

Sync with HEAD

include "ioconf.h" to get the 'void <driver>attach(int count);' prototype.

Sync with HEAD

Create a per-device state for iscsi(4), effectively making it a cloning
device.

Remove ISCSI_PERFTEST.

Don't pre-compute string sizes in a relatively cold function.

Remove conditionals for NetBSD before 4.0.

Sync with HEAD

Revert and add comment that this dance is for the sake of builtin module
registration.

Simplify.

iscsi_attrs is not used when iscsi(4) is compiled into the kernel.

Rebase to HEAD as of a few days ago.

Rebase.

Add d_discard to all struct cdevsw instances I could find.

All have been set to "nodiscard"; some should get a real implementation.

sync with head.

for a reference, the tree before this commit was tagged
as yamt-pagecache-tag8.

this commit was splitted into small chunks to avoid
a limitation of cvs.  ("Protocol error: too many arguments")

sync with head

Change (mostly mechanically) every cdevsw/bdevsw I can find to use
designated initializers.

I have not built every extant kernel so I have probably broken at
least one build; however I've also found and fixed some wrong
cdevsw/bdevsw entries so even if so I think we come out ahead.

resync with head

sync with head

- defer session cleanup to not force detachments
- use more and explicit locking
- improve connection recovery
- use larger timeouts
- handle ccb buffer underflow correctly
- simplify throttling code

Sessions can now temporarily exist without a valid
connection, you also need to update iscsid(8).

sync with head

Don't use common variables, also prefix global variables to avoid
namespace pollution.

Pull up revisions:
  src/sys/dev/iscsi/iscsi_globals.h revisions 1.3,1.4
  src/sys/dev/iscsi/iscsi_main.c revisions 1.4,1.5
  src/sys/dev/iscsi/iscsi_text.c revisions 1.4,1.5
  src/sys/dev/iscsi/iscsi_ioctl.c revisions 1.3,1.4
  src/sys/dev/iscsi/iscsi_send.c revisions 1.3,1.4,1.5
  src/sys/dev/iscsi/iscsi_utils.c revisions 1.2,1.3,1.4
  src/sbin/iscsid/iscsid_discover.c revision 1.4
  src/sys/dev/iscsi/iscsi_rcv.c revision 1.3
  src/sbin/iscsictl/iscsic_driverif.c revisions 1.5,1.6
(requested by martin in ticket #357, with updates from mhitch and mlelstv).

avoid namespace pollution when debugging
don't panic when printing data for a dead (NULL) connection

Close file descriptor passed into the kernel on success.
Fix locking of file handle. More cleanup on error paths.
Keep track of CCBs, so they cannot be used after a session ends.
Handle CCB timeouts even when the connection is terminated.
Compute firstdata, firstimmed correctly.

assemble_login_parameters() has a strange error reporting convention:
errors are positive numbers, so <= 0 is good, not >= ...
This makes CHAP authenticated iscsi logins work.

use first successful connection to a ISNS server

Add more debugging, fix filehandle usage, login negotiation and session
shutdown.
Add #ifdef'd code to send negotiation parameters in hex instead of base64,
so it works against older Linux targets.

report luns returns a 32bit list length.

Make digests work also on big endian machines.

it's really 32bit

Close file descriptor passed into the kernel on success.
Fix locking of file handle. More cleanup on error paths.
Keep track of CCBs, so they cannot be used after a session ends.
Handle CCB timeouts even when the connection is terminated.
Compute firstdata, firstimmed correctly.

avoid namespace pollution when debugging
don't panic when printing data for a dead (NULL) connection

sync to latest -current.

Pull up the following revisions(s) (requested by riz in ticket #291):
	etc/MAKEDEV.tmpl:		revision 1.155
	sbin/iscsictl/iscsictl.8:	revision 1.2-1.4
	sbin/iscsid/iscsid_driverif.c:	revision 1.4-1.5
	sbin/iscsid/iscsid_lists.c:	revision 1.4-1.7
	sbin/iscsid/iscsid_targets.c:	revision 1.4
	sbin/iscsid/iscsid_globals.h:	revision 1.5-1.7
	sbin/iscsid/iscsid_main.c:	revision 1.4-1.7
	sbin/iscsid/Makefile:		revision 1.2-1.4
	sbin/iscsid/iscsid.8:		revision 1.3-1.8
	sbin/iscsid/iscsid.h:		revision 1.3
	sys/dev/iscsi/iscsi_main.c:	revision 1.2-1.3

Fix bugs in iscsid target list handling, and improve documentation
somewhat for the in-kernel iSCSI initiator.

Expose some of the module framework when _MODULE is not defined, so
iscsi becomes a "built in" module in that case.

Don't allow iscsi to autounload.  The default unload timeout is short,
and too much state is lost.

Start to move the in-kernel iSCSI initiator, kindly contributed to the
NetBSD Foundation by Wasabi Systems, from

        othersrc/external/bsd/iscsi/sys/dev/iscsi

to

        src/sys/dev/iscsi