The iscsi eh could be tearing down the session/connection while
the scsi eh is still sending task management functions. If when
we drop the session lock to grab the recv lock, the iscsi eh
tears down the connection we will oops.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
We do not want to send data if we are aborting a task. There is
a check in iscsi_xmit_ctask, but right before calling this we overwrite
the state so we always go right past the test. Sending data causes problems
because when we clean up from a successful abort the LLD assumes that
the task is not running.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
We should not be checking the cmd windown for just handling r2t responses.
And if the window closes in on us, always have scsi-ml requeue the command
from our queuecommand function.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
When we logout we block the session since we are not taking any more
commands, but when we call remove host we want to make sure any
IO that got queued up and blocked gets failed upwards quickly, so
we unblock the session and fail it.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
- remove the unnecessary map_single path.
- convert to use the new accessors for the sg lists and the
parameters.
TODO: use scsi_for_each_sg().
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Acked-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
iSCSI must support software iscsi (iscsi_tcp, iser), hardware iscsi (qla4xxx),
and partial offload (broadcom). To be able to allow each stack or driver
or port (virtual or physical) to be able to log into the same target portal
we use the initiator tuple [[HWADDRESS | NETDEVNAME], INITIATOR_NAME] and
the target tuple [TARGETNAME, CONN_ADDRESS, CONN_PORT] to id a session.
This patch adds the netdev name, which is used by software iscsi when
it binds a session to a netdevice using the SO_BINDTODEVICE sock opt.
It cannot use HWADDRESS because if someone did vlans then the same netdevice
will have the same mac and the initiator,target id will not be unique.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Cc: Roland Dreier <rdreier@cisco.com>
Cc: David C Somayajulu <david.somayajulu@qlogic.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
This patch allows us to set can_queue and cmds_per_lun from userspace
when we create the session/host. From there we can set it on a per
target basis. The patch fully converts iscsi_tcp, but only hooks
up ib_iser for cmd_per_lun since it currently has a lots of preallocations
based on can_queue.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Cc: Roland Dreier <rdreier@cisco.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
The cmdsn allocation and pdu transmit code can race, and we can end
up sending a pdu with cmdsn 10 before a pdu with 5. The target will
then fail the connection/session. This patch fixes the problem by
delaying the cmdsn allocation until we are about to send the pdu.
This also removes the xmitmutex. We were using the connection xmitmutex
during error handling to handle races with mtask and ctask cleanup and
completion. For ctasks we now have nice refcounting and for the mtask,
if we hit the case where the mtask timesout and it is floating
around somewhere in the driver, we end up dropping the session.
And to handle session level cleanup, we use the xmit suspend bit
along with scsi_flush_queue and the session lock to make sure
that the xmit thread is not possibly transmitting a task while
we are trying to kill it.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Cc: Roland Dreier <rdreier@cisco.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
If iscsi_tcp partially sends a header, it would recalculate the
header size and readd the size of the digest (if header digests
are used).This would cause us to send sizeof(digest) extra bytes
when we sent the rest of the header.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
The attached patches add sysfs files for the chap settings
to the iscsi transport class, iscsi_tcp and ib_iser. This is
needed for software iscsi because there are times when iscsid
can die and it will need to reread the values it was using.
And it is needed by qla4xxx for basic management opertaions.
This patch does not hook in qla4xxx yet, because I am not sure
the mbx command to use.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Cc: Roland Dreier <rdreier@cisco.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
- Remove shadow of request length from struct iscsi_cmd_task.
- change all users to use scsi_cmnd->request_bufflen directly
(With bidi we will use scsi-ml API to retrieve in/out length)
Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Cc: Roland Dreier <rdreier@cisco.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
This patch fixes handling of expected datasn/r2tsn as received from
target. It is done according to: T10 rfc3720 section 3.2.2.3. Data Sequencing.
. unify expected datasn/r2tsn into one counter
. calculate than check expected datasn/r2tsn. On error print a message
and fail the request. (TODO use iscsi retransmits)
. remove the FIXME ;)
. avoid zero length memset
Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
For iscsi root boot, software iscsi needs to know what the BIOS/OF
initiator used for the initiator name so this puts it in sysfs
for userspace to be able to pick up.
For hw iscsi, it is nice to see what the card is using.
This patch adds the new param, and hooks in qla4xxx, iscsi_tcp, and ib_iser.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Cc: Roland Dreier <rdreier@cisco.com>
Cc: David C Somayajulu <david.somayajulu@qlogic.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
iscsid and udev need to key off the hw address being
used so add some helpers for iser and iscsi tcp.
Also convert them
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Cc: Roland Dreier <rdreier@cisco.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
Dave Miller meantioned that the data buffer in a past
sense fixup patch was not gauranteed to be aligned
properly for ia64. This patch has libiscsi use get_unalinged
to make sure. There are a couple more places in the
digest handling we may need to do this, but we are in the middle
of fixing that code for big endien systems so just the sense
access is fixed here.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
This patch renames DEFAULT_MAX_RECV_DATA_SEGMENT_LENGTH to avoid
confusion with the drivers default values (DEFAULT_MAX_RECV_DATA_SEGMENT_LENGTH
is the iscsi RFC specific default).
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
It's possible that we call iscsi_xmitworker after iscsi_conn_release
which causes a oops. This patch flushes the workqueue.
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
Yanling Qi, noted that when the sense data length of
a check-condition is greater than 0x7f (127), senselen = (data[0] << 8)
| data[1] will become negative. It causes different kinds of panics from
GPF, spin_lock deadlock to spin_lock recursion.
We were also swapping this value on big endien machines.
This patch fixes both issues by using be16_to_cpu().
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
Some messages from debug_scsi do not have trailing newlines,
making console messages difficult to read. Fix that.
Signed-off-by: Pete Wyckoff <pw@osc.edu>
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
According to the iscsi RFC, we cannot send other requests if
we have sent a logout pdu. This patch enforces this requirement
by blocking the session and suspending the send thread. Userspace
decides if we restart the connection or if we just free everything.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
We have been dropping the pdu. We should just send it to userspace
and let it handle it.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
from bhalevy@gmail.com:
It looks like change 652 to libiscsi.c added some dead code around line
670
if (rc) {
spin_unlock_bh(&conn->session->lock);
goto again;
}
since 5 lines above we goto again if (rc).
It looks like the previous if (rc) should go away if we want to put the
ctask before
breaking out of the while loop with "goto again" (see following patch).
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
If connection creation fails we end up calling list_del
on a invalid struct. This then causes an oops. We are not
acutally using the lists (old MCS code we thought might
be useful elsewhere) so this patch just removes that
code.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
In the normal IO path we should not be calling back
into the LLD since the LLD will have cleaned up the
task before or after calling complete pdu.
For the fail_command path we still need to do this
to force the cleanup.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
If the scsi eh sends a TUR and the session is down we could
return SCSI_ML_HOST_BUSY. scsi eh will ignore this and send
ask us to abort the command and we blindly accesst the
command ptr.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
The first burst length is only relevant if immedate data = Yes
or if Initial R2T is No
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
iscsi_tcp calculates padding by using the expected transfer length. This
has the problem where if we have immediate data = no and initial R2T =
yes, and the transfer length ended up needing padding then we send:
1. header
2. padding which should have gone after data
3. data
Besides this bug, we also assume the target will always ask for nice
transfer lengths and the first burst length will always be a nice value.
As far as I can tell form the RFC this is not a requirement. It would be
silly to do this, but if someone did it we will end doing bad things.
Finally the last bug in that bit of code is in our handling of the
recalculation of data digests when we do not send a whole iscsi_buf in
one try. The bug here is that we call crypto_digest_final on a
iscsi_sendpage error, then when we send the rest of the iscsi_buf, we
doiscsi_data_digest_init and this causes the previous data digest to be
lost.
And to make matters worse, some of these bugs are replicated over and
over and over again for immediate data, solicited data and unsolicited
data. So the attached patch made over the iscsi git tree (see
kernel.org/git for details) which I updated today to include the patches
I said I merged, consolidates the sending of data, padding and digests
and calculation of data digests and fixes the above bugs.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
It is possible that a ctask could be completing and getting
cleaned up at the same time, we are finishing up the last
data transfer. This could then result in the data transfer
code using stale or invalid values. This patch adds a refcount
to the ctask. When the count goes to zero then we know the
transmit thread and recv thread or softirq are not touching
it and we can safely release it.
The eh should not need to grab a reference because it only cleans
up a task if it has both the xmit mutex and recv lock (or recv
side suspended).
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
iSCSI RFC states that the first burst length must be smaller than the
max burst length. We currently assume targets will be good, but that may
not be the case, so this patch adds a check.
This patch also moves the unsol data out offset to the lib so the LLDs
do not have to track it.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
We were leaking some strings. This patch just frees them.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
Must pass ISCSI_ERR values from the recv path and propogate them
upwards.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
We currently try to allocate a max_recv_data_segment_length
which can be very large (default is 64K), and common uses
are up to 1MB. It is very very difficult to allocte this
much contiguous memory and it turns out we never even use it.
We really only need a couple of pages, so this patch has us
allocates just what we know what we need today.
Later if vendors start adding vendor specific data and
we need to handle large buffers we can do this, but for
the last 4 years we have not seen anyone do this or request
it.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
We are touching the cls_session after we have freed
it. This causes a oops.
Signed-off-by: Or Gerlitz <ogerlitz@voltaire.com>
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
When we enter recovery and flush the running commands
we cannot freee the connection before flushing the commands.
Some commands may have a reference to the connection
that needs to be released before. iscsi_stop was forcing
the term and suspend too early and was causing a oops
in iser, so this patch removes those callbacks all together
and allows the LLD to handle that detail.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
Abort handler fixes.
If a connection is dropped and reconnected while an abort is
running then we should assume the recovery code will clean up
the abort. Not doing so causes a oops.
And if a command completes then we get the status for the abort, we do not
need to call into the LLD to cleanup the resources. Doing this causes
and oops in iser because it ends up freeing some resources twice.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
The iscsi tcp code can pluck multiple rt2s from the tasks's r2tqueue
in the xmit code. This can result in the task being queued on the xmit queue
but gettting completed at the same time.
This patch fixes the above bug by making the fifo a list so
we always remove the entry on the list del.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
So the drivers do not use the channel numbers, but some do
use the target numbers. We were just adding some goofy
variable that just increases for the target nr. This is useless
for software iscsi because it is always zero. And for qla4xxx
the target nr is actually the index of the target/session
in its FW or FLASH tables. We needed to expose this to userspace
so apps could access those numbers so this patch just adds the
target nr to the iscsi session creation functions. This way
when qla4xxx's Hw thinks a session is at target nr 4
in its hw, it is exposed as that number in sysfs.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
I do not remember what I was thinking when we added the channel
as a argument to the session create function. It was probably
due to too much cut and paste work from the FC transport class.
The channel is meaningless for iscsi drivers so this patch drops
its usage everywhere in the iscsi related code.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
iscsi_tcp and iser cannot be rmmod from the kernel when sessions
are running because session removal is driven from userspace. For
those modules we get a module reference when a session is
created then drop it when the session is removed.
For qla4xxx, they can jsut remove the sessions from the pci remove
function like normal HW drivers, so this patch moves the module
reference from the transport class functions shared by all
drivers to the libiscsi functions only used be software iscsi
modules.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
Reduce duplication in the software iscsi_transport modules by
adding a libiscsi function to handle the common grunt work.
This also has the drivers return specifc -EXXX values for different
errors so userspace can finally handle them in a sane way.
Also just pass the sysfs buffers to the drivers so HW iscsi can
get/set its string values, like targetname, and initiatorname.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
We can race and misset the suspend bit if iscsi_write_space is
called then iscsi_send returns with a failure indicating
there is no space.
To handle this this patch returns a error upwards allowing xmitworker
to decide if we need to try and transmit again. For the no
write space case xmitworker will not retry, and instead
let iscsi_write_space queue it back up if needed (this relies
on the work queue code to properly requeue us if needed).
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
from davidw@netapp.com:
remove task type should return a task on success.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
from davidw@netapp.com:
We must grab the session lock when modifying the running lists.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
If recovery failed or we are in recovery only overwrite the state
if we are going to terminate the session or if we logged back in.
STOP_CONN_SUSPEND and conn_cnt are not used. We only support
a single connection session ATM, so cleanup that code while
we are working around it.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
Do not flush queues then block session. This will cause commands
to needlessly swing around on us and remove goofy
recovery_failed field and replace with state value.
And do not start recovery from within the host reset function.
This causeis too many problems becuase open-iscsi was desinged to
call out to userspace then have userpscae decide if we should
go into recovery or kill the session.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
We only use the mtask data buffer for login tasks so we do not
need to preallocate a buffer for every mtask. This saves
8 * 31 KB.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
debugged by Ming and Rohan:
The problem Ming and Rohan debugged was that during a normal session
login, open-iscsi is not incrementing the exp_statsn counter. It was
stuck at zero. From the RFC, it looks like if the login response PDU has
a successful status then we should be incrementing that value. Also from
the RFC, it looks like if when we drop a connection then reconnect, we
should be using the exp_statsn from the old connection in the next
relogin attempt.
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
Mike Christie