2011-07-23 08:43:04 +02:00
|
|
|
/******************************************************************************
|
|
|
|
* This file contains error recovery level zero functions used by
|
|
|
|
* the iSCSI Target driver.
|
|
|
|
*
|
2013-09-06 00:29:12 +02:00
|
|
|
* (c) Copyright 2007-2013 Datera, Inc.
|
2011-07-23 08:43:04 +02:00
|
|
|
*
|
|
|
|
* Author: Nicholas A. Bellinger <nab@linux-iscsi.org>
|
|
|
|
*
|
|
|
|
* This program is free software; you can redistribute it and/or modify
|
|
|
|
* it under the terms of the GNU General Public License as published by
|
|
|
|
* the Free Software Foundation; either version 2 of the License, or
|
|
|
|
* (at your option) any later version.
|
|
|
|
*
|
|
|
|
* This program is distributed in the hope that it will be useful,
|
|
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
|
|
* GNU General Public License for more details.
|
|
|
|
******************************************************************************/
|
|
|
|
|
|
|
|
#include <scsi/iscsi_proto.h>
|
|
|
|
#include <target/target_core_base.h>
|
2011-11-16 15:46:48 +01:00
|
|
|
#include <target/target_core_fabric.h>
|
2011-07-23 08:43:04 +02:00
|
|
|
|
|
|
|
#include "iscsi_target_core.h"
|
|
|
|
#include "iscsi_target_seq_pdu_list.h"
|
|
|
|
#include "iscsi_target_tq.h"
|
|
|
|
#include "iscsi_target_erl0.h"
|
|
|
|
#include "iscsi_target_erl1.h"
|
|
|
|
#include "iscsi_target_erl2.h"
|
|
|
|
#include "iscsi_target_util.h"
|
|
|
|
#include "iscsi_target.h"
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Used to set values in struct iscsi_cmd that iscsit_dataout_check_sequence()
|
|
|
|
* checks against to determine a PDU's Offset+Length is within the current
|
|
|
|
* DataOUT Sequence. Used for DataSequenceInOrder=Yes only.
|
|
|
|
*/
|
|
|
|
void iscsit_set_dataout_sequence_values(
|
|
|
|
struct iscsi_cmd *cmd)
|
|
|
|
{
|
|
|
|
struct iscsi_conn *conn = cmd->conn;
|
|
|
|
/*
|
|
|
|
* Still set seq_start_offset and seq_end_offset for Unsolicited
|
|
|
|
* DataOUT, even if DataSequenceInOrder=No.
|
|
|
|
*/
|
|
|
|
if (cmd->unsolicited_data) {
|
|
|
|
cmd->seq_start_offset = cmd->write_data_done;
|
|
|
|
cmd->seq_end_offset = (cmd->write_data_done +
|
2012-10-02 10:22:48 +02:00
|
|
|
((cmd->se_cmd.data_length >
|
|
|
|
conn->sess->sess_ops->FirstBurstLength) ?
|
|
|
|
conn->sess->sess_ops->FirstBurstLength : cmd->se_cmd.data_length));
|
2011-07-23 08:43:04 +02:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!conn->sess->sess_ops->DataSequenceInOrder)
|
|
|
|
return;
|
|
|
|
|
|
|
|
if (!cmd->seq_start_offset && !cmd->seq_end_offset) {
|
|
|
|
cmd->seq_start_offset = cmd->write_data_done;
|
2012-04-04 00:51:24 +02:00
|
|
|
cmd->seq_end_offset = (cmd->se_cmd.data_length >
|
2011-07-23 08:43:04 +02:00
|
|
|
conn->sess->sess_ops->MaxBurstLength) ?
|
|
|
|
(cmd->write_data_done +
|
2012-04-04 00:51:24 +02:00
|
|
|
conn->sess->sess_ops->MaxBurstLength) : cmd->se_cmd.data_length;
|
2011-07-23 08:43:04 +02:00
|
|
|
} else {
|
|
|
|
cmd->seq_start_offset = cmd->seq_end_offset;
|
|
|
|
cmd->seq_end_offset = ((cmd->seq_end_offset +
|
|
|
|
conn->sess->sess_ops->MaxBurstLength) >=
|
2012-04-04 00:51:24 +02:00
|
|
|
cmd->se_cmd.data_length) ? cmd->se_cmd.data_length :
|
2011-07-23 08:43:04 +02:00
|
|
|
(cmd->seq_end_offset +
|
|
|
|
conn->sess->sess_ops->MaxBurstLength);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static int iscsit_dataout_within_command_recovery_check(
|
|
|
|
struct iscsi_cmd *cmd,
|
|
|
|
unsigned char *buf)
|
|
|
|
{
|
|
|
|
struct iscsi_conn *conn = cmd->conn;
|
|
|
|
struct iscsi_data *hdr = (struct iscsi_data *) buf;
|
|
|
|
u32 payload_length = ntoh24(hdr->dlength);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* We do the within-command recovery checks here as it is
|
|
|
|
* the first function called in iscsi_check_pre_dataout().
|
|
|
|
* Basically, if we are in within-command recovery and
|
|
|
|
* the PDU does not contain the offset the sequence needs,
|
|
|
|
* dump the payload.
|
|
|
|
*
|
|
|
|
* This only applies to DataPDUInOrder=Yes, for
|
|
|
|
* DataPDUInOrder=No we only re-request the failed PDU
|
|
|
|
* and check that all PDUs in a sequence are received
|
|
|
|
* upon end of sequence.
|
|
|
|
*/
|
|
|
|
if (conn->sess->sess_ops->DataSequenceInOrder) {
|
|
|
|
if ((cmd->cmd_flags & ICF_WITHIN_COMMAND_RECOVERY) &&
|
2012-09-26 14:00:40 +02:00
|
|
|
cmd->write_data_done != be32_to_cpu(hdr->offset))
|
2011-07-23 08:43:04 +02:00
|
|
|
goto dump;
|
|
|
|
|
|
|
|
cmd->cmd_flags &= ~ICF_WITHIN_COMMAND_RECOVERY;
|
|
|
|
} else {
|
|
|
|
struct iscsi_seq *seq;
|
|
|
|
|
2012-09-26 14:00:40 +02:00
|
|
|
seq = iscsit_get_seq_holder(cmd, be32_to_cpu(hdr->offset),
|
|
|
|
payload_length);
|
2011-07-23 08:43:04 +02:00
|
|
|
if (!seq)
|
|
|
|
return DATAOUT_CANNOT_RECOVER;
|
|
|
|
/*
|
|
|
|
* Set the struct iscsi_seq pointer to reuse later.
|
|
|
|
*/
|
|
|
|
cmd->seq_ptr = seq;
|
|
|
|
|
|
|
|
if (conn->sess->sess_ops->DataPDUInOrder) {
|
2012-09-26 14:00:40 +02:00
|
|
|
if (seq->status ==
|
|
|
|
DATAOUT_SEQUENCE_WITHIN_COMMAND_RECOVERY &&
|
|
|
|
(seq->offset != be32_to_cpu(hdr->offset) ||
|
|
|
|
seq->data_sn != be32_to_cpu(hdr->datasn)))
|
2011-07-23 08:43:04 +02:00
|
|
|
goto dump;
|
|
|
|
} else {
|
2012-09-26 14:00:40 +02:00
|
|
|
if (seq->status ==
|
|
|
|
DATAOUT_SEQUENCE_WITHIN_COMMAND_RECOVERY &&
|
|
|
|
seq->data_sn != be32_to_cpu(hdr->datasn))
|
2011-07-23 08:43:04 +02:00
|
|
|
goto dump;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (seq->status == DATAOUT_SEQUENCE_COMPLETE)
|
|
|
|
goto dump;
|
|
|
|
|
|
|
|
if (seq->status != DATAOUT_SEQUENCE_COMPLETE)
|
|
|
|
seq->status = 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
return DATAOUT_NORMAL;
|
|
|
|
|
|
|
|
dump:
|
|
|
|
pr_err("Dumping DataOUT PDU Offset: %u Length: %d DataSN:"
|
|
|
|
" 0x%08x\n", hdr->offset, payload_length, hdr->datasn);
|
|
|
|
return iscsit_dump_data_payload(conn, payload_length, 1);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int iscsit_dataout_check_unsolicited_sequence(
|
|
|
|
struct iscsi_cmd *cmd,
|
|
|
|
unsigned char *buf)
|
|
|
|
{
|
|
|
|
u32 first_burst_len;
|
|
|
|
struct iscsi_conn *conn = cmd->conn;
|
|
|
|
struct iscsi_data *hdr = (struct iscsi_data *) buf;
|
|
|
|
u32 payload_length = ntoh24(hdr->dlength);
|
|
|
|
|
|
|
|
|
2012-09-26 14:00:40 +02:00
|
|
|
if ((be32_to_cpu(hdr->offset) < cmd->seq_start_offset) ||
|
|
|
|
((be32_to_cpu(hdr->offset) + payload_length) > cmd->seq_end_offset)) {
|
2011-07-23 08:43:04 +02:00
|
|
|
pr_err("Command ITT: 0x%08x with Offset: %u,"
|
|
|
|
" Length: %u outside of Unsolicited Sequence %u:%u while"
|
|
|
|
" DataSequenceInOrder=Yes.\n", cmd->init_task_tag,
|
2012-09-26 14:00:40 +02:00
|
|
|
be32_to_cpu(hdr->offset), payload_length, cmd->seq_start_offset,
|
2011-07-23 08:43:04 +02:00
|
|
|
cmd->seq_end_offset);
|
|
|
|
return DATAOUT_CANNOT_RECOVER;
|
|
|
|
}
|
|
|
|
|
|
|
|
first_burst_len = (cmd->first_burst_len + payload_length);
|
|
|
|
|
|
|
|
if (first_burst_len > conn->sess->sess_ops->FirstBurstLength) {
|
|
|
|
pr_err("Total %u bytes exceeds FirstBurstLength: %u"
|
|
|
|
" for this Unsolicited DataOut Burst.\n",
|
|
|
|
first_burst_len, conn->sess->sess_ops->FirstBurstLength);
|
|
|
|
transport_send_check_condition_and_sense(&cmd->se_cmd,
|
|
|
|
TCM_INCORRECT_AMOUNT_OF_DATA, 0);
|
|
|
|
return DATAOUT_CANNOT_RECOVER;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Perform various MaxBurstLength and ISCSI_FLAG_CMD_FINAL sanity
|
|
|
|
* checks for the current Unsolicited DataOUT Sequence.
|
|
|
|
*/
|
|
|
|
if (hdr->flags & ISCSI_FLAG_CMD_FINAL) {
|
|
|
|
/*
|
|
|
|
* Ignore ISCSI_FLAG_CMD_FINAL checks while DataPDUInOrder=No, end of
|
|
|
|
* sequence checks are handled in
|
|
|
|
* iscsit_dataout_datapduinorder_no_fbit().
|
|
|
|
*/
|
|
|
|
if (!conn->sess->sess_ops->DataPDUInOrder)
|
|
|
|
goto out;
|
|
|
|
|
2012-04-04 00:51:24 +02:00
|
|
|
if ((first_burst_len != cmd->se_cmd.data_length) &&
|
2011-07-23 08:43:04 +02:00
|
|
|
(first_burst_len != conn->sess->sess_ops->FirstBurstLength)) {
|
|
|
|
pr_err("Unsolicited non-immediate data"
|
|
|
|
" received %u does not equal FirstBurstLength: %u, and"
|
|
|
|
" does not equal ExpXferLen %u.\n", first_burst_len,
|
|
|
|
conn->sess->sess_ops->FirstBurstLength,
|
2012-04-04 00:51:24 +02:00
|
|
|
cmd->se_cmd.data_length);
|
2011-07-23 08:43:04 +02:00
|
|
|
transport_send_check_condition_and_sense(&cmd->se_cmd,
|
|
|
|
TCM_INCORRECT_AMOUNT_OF_DATA, 0);
|
|
|
|
return DATAOUT_CANNOT_RECOVER;
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
if (first_burst_len == conn->sess->sess_ops->FirstBurstLength) {
|
|
|
|
pr_err("Command ITT: 0x%08x reached"
|
|
|
|
" FirstBurstLength: %u, but ISCSI_FLAG_CMD_FINAL is not set. protocol"
|
|
|
|
" error.\n", cmd->init_task_tag,
|
|
|
|
conn->sess->sess_ops->FirstBurstLength);
|
|
|
|
return DATAOUT_CANNOT_RECOVER;
|
|
|
|
}
|
2012-04-04 00:51:24 +02:00
|
|
|
if (first_burst_len == cmd->se_cmd.data_length) {
|
2011-07-23 08:43:04 +02:00
|
|
|
pr_err("Command ITT: 0x%08x reached"
|
|
|
|
" ExpXferLen: %u, but ISCSI_FLAG_CMD_FINAL is not set. protocol"
|
2012-04-04 00:51:24 +02:00
|
|
|
" error.\n", cmd->init_task_tag, cmd->se_cmd.data_length);
|
2011-07-23 08:43:04 +02:00
|
|
|
return DATAOUT_CANNOT_RECOVER;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
out:
|
|
|
|
return DATAOUT_NORMAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int iscsit_dataout_check_sequence(
|
|
|
|
struct iscsi_cmd *cmd,
|
|
|
|
unsigned char *buf)
|
|
|
|
{
|
|
|
|
u32 next_burst_len;
|
|
|
|
struct iscsi_conn *conn = cmd->conn;
|
|
|
|
struct iscsi_seq *seq = NULL;
|
|
|
|
struct iscsi_data *hdr = (struct iscsi_data *) buf;
|
|
|
|
u32 payload_length = ntoh24(hdr->dlength);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* For DataSequenceInOrder=Yes: Check that the offset and offset+length
|
|
|
|
* is within range as defined by iscsi_set_dataout_sequence_values().
|
|
|
|
*
|
|
|
|
* For DataSequenceInOrder=No: Check that an struct iscsi_seq exists for
|
|
|
|
* offset+length tuple.
|
|
|
|
*/
|
|
|
|
if (conn->sess->sess_ops->DataSequenceInOrder) {
|
|
|
|
/*
|
|
|
|
* Due to possibility of recovery DataOUT sent by the initiator
|
|
|
|
* fullfilling an Recovery R2T, it's best to just dump the
|
|
|
|
* payload here, instead of erroring out.
|
|
|
|
*/
|
2012-09-26 14:00:40 +02:00
|
|
|
if ((be32_to_cpu(hdr->offset) < cmd->seq_start_offset) ||
|
|
|
|
((be32_to_cpu(hdr->offset) + payload_length) > cmd->seq_end_offset)) {
|
2011-07-23 08:43:04 +02:00
|
|
|
pr_err("Command ITT: 0x%08x with Offset: %u,"
|
|
|
|
" Length: %u outside of Sequence %u:%u while"
|
|
|
|
" DataSequenceInOrder=Yes.\n", cmd->init_task_tag,
|
2012-09-26 14:00:40 +02:00
|
|
|
be32_to_cpu(hdr->offset), payload_length, cmd->seq_start_offset,
|
2011-07-23 08:43:04 +02:00
|
|
|
cmd->seq_end_offset);
|
|
|
|
|
|
|
|
if (iscsit_dump_data_payload(conn, payload_length, 1) < 0)
|
|
|
|
return DATAOUT_CANNOT_RECOVER;
|
|
|
|
return DATAOUT_WITHIN_COMMAND_RECOVERY;
|
|
|
|
}
|
|
|
|
|
|
|
|
next_burst_len = (cmd->next_burst_len + payload_length);
|
|
|
|
} else {
|
2012-09-26 14:00:40 +02:00
|
|
|
seq = iscsit_get_seq_holder(cmd, be32_to_cpu(hdr->offset),
|
|
|
|
payload_length);
|
2011-07-23 08:43:04 +02:00
|
|
|
if (!seq)
|
|
|
|
return DATAOUT_CANNOT_RECOVER;
|
|
|
|
/*
|
|
|
|
* Set the struct iscsi_seq pointer to reuse later.
|
|
|
|
*/
|
|
|
|
cmd->seq_ptr = seq;
|
|
|
|
|
|
|
|
if (seq->status == DATAOUT_SEQUENCE_COMPLETE) {
|
|
|
|
if (iscsit_dump_data_payload(conn, payload_length, 1) < 0)
|
|
|
|
return DATAOUT_CANNOT_RECOVER;
|
|
|
|
return DATAOUT_WITHIN_COMMAND_RECOVERY;
|
|
|
|
}
|
|
|
|
|
|
|
|
next_burst_len = (seq->next_burst_len + payload_length);
|
|
|
|
}
|
|
|
|
|
|
|
|
if (next_burst_len > conn->sess->sess_ops->MaxBurstLength) {
|
|
|
|
pr_err("Command ITT: 0x%08x, NextBurstLength: %u and"
|
|
|
|
" Length: %u exceeds MaxBurstLength: %u. protocol"
|
|
|
|
" error.\n", cmd->init_task_tag,
|
|
|
|
(next_burst_len - payload_length),
|
|
|
|
payload_length, conn->sess->sess_ops->MaxBurstLength);
|
|
|
|
return DATAOUT_CANNOT_RECOVER;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Perform various MaxBurstLength and ISCSI_FLAG_CMD_FINAL sanity
|
|
|
|
* checks for the current DataOUT Sequence.
|
|
|
|
*/
|
|
|
|
if (hdr->flags & ISCSI_FLAG_CMD_FINAL) {
|
|
|
|
/*
|
|
|
|
* Ignore ISCSI_FLAG_CMD_FINAL checks while DataPDUInOrder=No, end of
|
|
|
|
* sequence checks are handled in
|
|
|
|
* iscsit_dataout_datapduinorder_no_fbit().
|
|
|
|
*/
|
|
|
|
if (!conn->sess->sess_ops->DataPDUInOrder)
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
if (conn->sess->sess_ops->DataSequenceInOrder) {
|
|
|
|
if ((next_burst_len <
|
|
|
|
conn->sess->sess_ops->MaxBurstLength) &&
|
|
|
|
((cmd->write_data_done + payload_length) <
|
2012-04-04 00:51:24 +02:00
|
|
|
cmd->se_cmd.data_length)) {
|
2011-07-23 08:43:04 +02:00
|
|
|
pr_err("Command ITT: 0x%08x set ISCSI_FLAG_CMD_FINAL"
|
|
|
|
" before end of DataOUT sequence, protocol"
|
|
|
|
" error.\n", cmd->init_task_tag);
|
|
|
|
return DATAOUT_CANNOT_RECOVER;
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
if (next_burst_len < seq->xfer_len) {
|
|
|
|
pr_err("Command ITT: 0x%08x set ISCSI_FLAG_CMD_FINAL"
|
|
|
|
" before end of DataOUT sequence, protocol"
|
|
|
|
" error.\n", cmd->init_task_tag);
|
|
|
|
return DATAOUT_CANNOT_RECOVER;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
if (conn->sess->sess_ops->DataSequenceInOrder) {
|
|
|
|
if (next_burst_len ==
|
|
|
|
conn->sess->sess_ops->MaxBurstLength) {
|
|
|
|
pr_err("Command ITT: 0x%08x reached"
|
|
|
|
" MaxBurstLength: %u, but ISCSI_FLAG_CMD_FINAL is"
|
|
|
|
" not set, protocol error.", cmd->init_task_tag,
|
|
|
|
conn->sess->sess_ops->MaxBurstLength);
|
|
|
|
return DATAOUT_CANNOT_RECOVER;
|
|
|
|
}
|
|
|
|
if ((cmd->write_data_done + payload_length) ==
|
2012-04-04 00:51:24 +02:00
|
|
|
cmd->se_cmd.data_length) {
|
2011-07-23 08:43:04 +02:00
|
|
|
pr_err("Command ITT: 0x%08x reached"
|
|
|
|
" last DataOUT PDU in sequence but ISCSI_FLAG_"
|
|
|
|
"CMD_FINAL is not set, protocol error.\n",
|
|
|
|
cmd->init_task_tag);
|
|
|
|
return DATAOUT_CANNOT_RECOVER;
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
if (next_burst_len == seq->xfer_len) {
|
|
|
|
pr_err("Command ITT: 0x%08x reached"
|
|
|
|
" last DataOUT PDU in sequence but ISCSI_FLAG_"
|
|
|
|
"CMD_FINAL is not set, protocol error.\n",
|
|
|
|
cmd->init_task_tag);
|
|
|
|
return DATAOUT_CANNOT_RECOVER;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
out:
|
|
|
|
return DATAOUT_NORMAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int iscsit_dataout_check_datasn(
|
|
|
|
struct iscsi_cmd *cmd,
|
|
|
|
unsigned char *buf)
|
|
|
|
{
|
|
|
|
int dump = 0, recovery = 0;
|
|
|
|
u32 data_sn = 0;
|
|
|
|
struct iscsi_conn *conn = cmd->conn;
|
|
|
|
struct iscsi_data *hdr = (struct iscsi_data *) buf;
|
|
|
|
u32 payload_length = ntoh24(hdr->dlength);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Considering the target has no method of re-requesting DataOUT
|
|
|
|
* by DataSN, if we receieve a greater DataSN than expected we
|
|
|
|
* assume the functions for DataPDUInOrder=[Yes,No] below will
|
|
|
|
* handle it.
|
|
|
|
*
|
|
|
|
* If the DataSN is less than expected, dump the payload.
|
|
|
|
*/
|
|
|
|
if (conn->sess->sess_ops->DataSequenceInOrder)
|
|
|
|
data_sn = cmd->data_sn;
|
|
|
|
else {
|
|
|
|
struct iscsi_seq *seq = cmd->seq_ptr;
|
|
|
|
data_sn = seq->data_sn;
|
|
|
|
}
|
|
|
|
|
2012-09-26 14:00:40 +02:00
|
|
|
if (be32_to_cpu(hdr->datasn) > data_sn) {
|
2011-07-23 08:43:04 +02:00
|
|
|
pr_err("Command ITT: 0x%08x, received DataSN: 0x%08x"
|
|
|
|
" higher than expected 0x%08x.\n", cmd->init_task_tag,
|
2012-09-26 14:00:40 +02:00
|
|
|
be32_to_cpu(hdr->datasn), data_sn);
|
2011-07-23 08:43:04 +02:00
|
|
|
recovery = 1;
|
|
|
|
goto recover;
|
2012-09-26 14:00:40 +02:00
|
|
|
} else if (be32_to_cpu(hdr->datasn) < data_sn) {
|
2011-07-23 08:43:04 +02:00
|
|
|
pr_err("Command ITT: 0x%08x, received DataSN: 0x%08x"
|
|
|
|
" lower than expected 0x%08x, discarding payload.\n",
|
2012-09-26 14:00:40 +02:00
|
|
|
cmd->init_task_tag, be32_to_cpu(hdr->datasn), data_sn);
|
2011-07-23 08:43:04 +02:00
|
|
|
dump = 1;
|
|
|
|
goto dump;
|
|
|
|
}
|
|
|
|
|
|
|
|
return DATAOUT_NORMAL;
|
|
|
|
|
|
|
|
recover:
|
|
|
|
if (!conn->sess->sess_ops->ErrorRecoveryLevel) {
|
|
|
|
pr_err("Unable to perform within-command recovery"
|
|
|
|
" while ERL=0.\n");
|
|
|
|
return DATAOUT_CANNOT_RECOVER;
|
|
|
|
}
|
|
|
|
dump:
|
|
|
|
if (iscsit_dump_data_payload(conn, payload_length, 1) < 0)
|
|
|
|
return DATAOUT_CANNOT_RECOVER;
|
|
|
|
|
|
|
|
return (recovery || dump) ? DATAOUT_WITHIN_COMMAND_RECOVERY :
|
|
|
|
DATAOUT_NORMAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int iscsit_dataout_pre_datapduinorder_yes(
|
|
|
|
struct iscsi_cmd *cmd,
|
|
|
|
unsigned char *buf)
|
|
|
|
{
|
|
|
|
int dump = 0, recovery = 0;
|
|
|
|
struct iscsi_conn *conn = cmd->conn;
|
|
|
|
struct iscsi_data *hdr = (struct iscsi_data *) buf;
|
|
|
|
u32 payload_length = ntoh24(hdr->dlength);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* For DataSequenceInOrder=Yes: If the offset is greater than the global
|
|
|
|
* DataPDUInOrder=Yes offset counter in struct iscsi_cmd a protcol error has
|
2012-08-27 15:46:00 +02:00
|
|
|
* occurred and fail the connection.
|
2011-07-23 08:43:04 +02:00
|
|
|
*
|
|
|
|
* For DataSequenceInOrder=No: If the offset is greater than the per
|
|
|
|
* sequence DataPDUInOrder=Yes offset counter in struct iscsi_seq a protocol
|
2012-08-27 15:46:00 +02:00
|
|
|
* error has occurred and fail the connection.
|
2011-07-23 08:43:04 +02:00
|
|
|
*/
|
|
|
|
if (conn->sess->sess_ops->DataSequenceInOrder) {
|
2012-09-26 14:00:40 +02:00
|
|
|
if (be32_to_cpu(hdr->offset) != cmd->write_data_done) {
|
2011-07-23 08:43:04 +02:00
|
|
|
pr_err("Command ITT: 0x%08x, received offset"
|
|
|
|
" %u different than expected %u.\n", cmd->init_task_tag,
|
2012-09-26 14:00:40 +02:00
|
|
|
be32_to_cpu(hdr->offset), cmd->write_data_done);
|
2011-07-23 08:43:04 +02:00
|
|
|
recovery = 1;
|
|
|
|
goto recover;
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
struct iscsi_seq *seq = cmd->seq_ptr;
|
|
|
|
|
2012-09-26 14:00:40 +02:00
|
|
|
if (be32_to_cpu(hdr->offset) > seq->offset) {
|
2011-07-23 08:43:04 +02:00
|
|
|
pr_err("Command ITT: 0x%08x, received offset"
|
|
|
|
" %u greater than expected %u.\n", cmd->init_task_tag,
|
2012-09-26 14:00:40 +02:00
|
|
|
be32_to_cpu(hdr->offset), seq->offset);
|
2011-07-23 08:43:04 +02:00
|
|
|
recovery = 1;
|
|
|
|
goto recover;
|
2012-09-26 14:00:40 +02:00
|
|
|
} else if (be32_to_cpu(hdr->offset) < seq->offset) {
|
2011-07-23 08:43:04 +02:00
|
|
|
pr_err("Command ITT: 0x%08x, received offset"
|
|
|
|
" %u less than expected %u, discarding payload.\n",
|
2012-09-26 14:00:40 +02:00
|
|
|
cmd->init_task_tag, be32_to_cpu(hdr->offset),
|
|
|
|
seq->offset);
|
2011-07-23 08:43:04 +02:00
|
|
|
dump = 1;
|
|
|
|
goto dump;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return DATAOUT_NORMAL;
|
|
|
|
|
|
|
|
recover:
|
|
|
|
if (!conn->sess->sess_ops->ErrorRecoveryLevel) {
|
|
|
|
pr_err("Unable to perform within-command recovery"
|
|
|
|
" while ERL=0.\n");
|
|
|
|
return DATAOUT_CANNOT_RECOVER;
|
|
|
|
}
|
|
|
|
dump:
|
|
|
|
if (iscsit_dump_data_payload(conn, payload_length, 1) < 0)
|
|
|
|
return DATAOUT_CANNOT_RECOVER;
|
|
|
|
|
|
|
|
return (recovery) ? iscsit_recover_dataout_sequence(cmd,
|
2012-09-26 14:00:40 +02:00
|
|
|
be32_to_cpu(hdr->offset), payload_length) :
|
2011-07-23 08:43:04 +02:00
|
|
|
(dump) ? DATAOUT_WITHIN_COMMAND_RECOVERY : DATAOUT_NORMAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int iscsit_dataout_pre_datapduinorder_no(
|
|
|
|
struct iscsi_cmd *cmd,
|
|
|
|
unsigned char *buf)
|
|
|
|
{
|
|
|
|
struct iscsi_pdu *pdu;
|
|
|
|
struct iscsi_data *hdr = (struct iscsi_data *) buf;
|
|
|
|
u32 payload_length = ntoh24(hdr->dlength);
|
|
|
|
|
2012-09-26 14:00:40 +02:00
|
|
|
pdu = iscsit_get_pdu_holder(cmd, be32_to_cpu(hdr->offset),
|
|
|
|
payload_length);
|
2011-07-23 08:43:04 +02:00
|
|
|
if (!pdu)
|
|
|
|
return DATAOUT_CANNOT_RECOVER;
|
|
|
|
|
|
|
|
cmd->pdu_ptr = pdu;
|
|
|
|
|
|
|
|
switch (pdu->status) {
|
|
|
|
case ISCSI_PDU_NOT_RECEIVED:
|
|
|
|
case ISCSI_PDU_CRC_FAILED:
|
|
|
|
case ISCSI_PDU_TIMED_OUT:
|
|
|
|
break;
|
|
|
|
case ISCSI_PDU_RECEIVED_OK:
|
|
|
|
pr_err("Command ITT: 0x%08x received already gotten"
|
|
|
|
" Offset: %u, Length: %u\n", cmd->init_task_tag,
|
2012-09-26 14:00:40 +02:00
|
|
|
be32_to_cpu(hdr->offset), payload_length);
|
2011-07-23 08:43:04 +02:00
|
|
|
return iscsit_dump_data_payload(cmd->conn, payload_length, 1);
|
|
|
|
default:
|
|
|
|
return DATAOUT_CANNOT_RECOVER;
|
|
|
|
}
|
|
|
|
|
|
|
|
return DATAOUT_NORMAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int iscsit_dataout_update_r2t(struct iscsi_cmd *cmd, u32 offset, u32 length)
|
|
|
|
{
|
|
|
|
struct iscsi_r2t *r2t;
|
|
|
|
|
|
|
|
if (cmd->unsolicited_data)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
r2t = iscsit_get_r2t_for_eos(cmd, offset, length);
|
|
|
|
if (!r2t)
|
|
|
|
return -1;
|
|
|
|
|
|
|
|
spin_lock_bh(&cmd->r2t_lock);
|
|
|
|
r2t->seq_complete = 1;
|
|
|
|
cmd->outstanding_r2ts--;
|
|
|
|
spin_unlock_bh(&cmd->r2t_lock);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int iscsit_dataout_update_datapduinorder_no(
|
|
|
|
struct iscsi_cmd *cmd,
|
|
|
|
u32 data_sn,
|
|
|
|
int f_bit)
|
|
|
|
{
|
|
|
|
int ret = 0;
|
|
|
|
struct iscsi_pdu *pdu = cmd->pdu_ptr;
|
|
|
|
|
|
|
|
pdu->data_sn = data_sn;
|
|
|
|
|
|
|
|
switch (pdu->status) {
|
|
|
|
case ISCSI_PDU_NOT_RECEIVED:
|
|
|
|
pdu->status = ISCSI_PDU_RECEIVED_OK;
|
|
|
|
break;
|
|
|
|
case ISCSI_PDU_CRC_FAILED:
|
|
|
|
pdu->status = ISCSI_PDU_RECEIVED_OK;
|
|
|
|
break;
|
|
|
|
case ISCSI_PDU_TIMED_OUT:
|
|
|
|
pdu->status = ISCSI_PDU_RECEIVED_OK;
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
return DATAOUT_CANNOT_RECOVER;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (f_bit) {
|
|
|
|
ret = iscsit_dataout_datapduinorder_no_fbit(cmd, pdu);
|
|
|
|
if (ret == DATAOUT_CANNOT_RECOVER)
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
return DATAOUT_NORMAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int iscsit_dataout_post_crc_passed(
|
|
|
|
struct iscsi_cmd *cmd,
|
|
|
|
unsigned char *buf)
|
|
|
|
{
|
|
|
|
int ret, send_r2t = 0;
|
|
|
|
struct iscsi_conn *conn = cmd->conn;
|
|
|
|
struct iscsi_seq *seq = NULL;
|
|
|
|
struct iscsi_data *hdr = (struct iscsi_data *) buf;
|
|
|
|
u32 payload_length = ntoh24(hdr->dlength);
|
|
|
|
|
|
|
|
if (cmd->unsolicited_data) {
|
|
|
|
if ((cmd->first_burst_len + payload_length) ==
|
|
|
|
conn->sess->sess_ops->FirstBurstLength) {
|
2012-09-26 14:00:40 +02:00
|
|
|
if (iscsit_dataout_update_r2t(cmd, be32_to_cpu(hdr->offset),
|
2011-07-23 08:43:04 +02:00
|
|
|
payload_length) < 0)
|
|
|
|
return DATAOUT_CANNOT_RECOVER;
|
|
|
|
send_r2t = 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!conn->sess->sess_ops->DataPDUInOrder) {
|
|
|
|
ret = iscsit_dataout_update_datapduinorder_no(cmd,
|
2012-09-26 14:00:40 +02:00
|
|
|
be32_to_cpu(hdr->datasn),
|
|
|
|
(hdr->flags & ISCSI_FLAG_CMD_FINAL));
|
2011-07-23 08:43:04 +02:00
|
|
|
if (ret == DATAOUT_CANNOT_RECOVER)
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
cmd->first_burst_len += payload_length;
|
|
|
|
|
|
|
|
if (conn->sess->sess_ops->DataSequenceInOrder)
|
|
|
|
cmd->data_sn++;
|
|
|
|
else {
|
|
|
|
seq = cmd->seq_ptr;
|
|
|
|
seq->data_sn++;
|
|
|
|
seq->offset += payload_length;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (send_r2t) {
|
|
|
|
if (seq)
|
|
|
|
seq->status = DATAOUT_SEQUENCE_COMPLETE;
|
|
|
|
cmd->first_burst_len = 0;
|
|
|
|
cmd->unsolicited_data = 0;
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
if (conn->sess->sess_ops->DataSequenceInOrder) {
|
|
|
|
if ((cmd->next_burst_len + payload_length) ==
|
|
|
|
conn->sess->sess_ops->MaxBurstLength) {
|
2012-09-26 14:00:40 +02:00
|
|
|
if (iscsit_dataout_update_r2t(cmd,
|
|
|
|
be32_to_cpu(hdr->offset),
|
2011-07-23 08:43:04 +02:00
|
|
|
payload_length) < 0)
|
|
|
|
return DATAOUT_CANNOT_RECOVER;
|
|
|
|
send_r2t = 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!conn->sess->sess_ops->DataPDUInOrder) {
|
|
|
|
ret = iscsit_dataout_update_datapduinorder_no(
|
2012-09-26 14:00:40 +02:00
|
|
|
cmd, be32_to_cpu(hdr->datasn),
|
2011-07-23 08:43:04 +02:00
|
|
|
(hdr->flags & ISCSI_FLAG_CMD_FINAL));
|
|
|
|
if (ret == DATAOUT_CANNOT_RECOVER)
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
cmd->next_burst_len += payload_length;
|
|
|
|
cmd->data_sn++;
|
|
|
|
|
|
|
|
if (send_r2t)
|
|
|
|
cmd->next_burst_len = 0;
|
|
|
|
} else {
|
|
|
|
seq = cmd->seq_ptr;
|
|
|
|
|
|
|
|
if ((seq->next_burst_len + payload_length) ==
|
|
|
|
seq->xfer_len) {
|
2012-09-26 14:00:40 +02:00
|
|
|
if (iscsit_dataout_update_r2t(cmd,
|
|
|
|
be32_to_cpu(hdr->offset),
|
2011-07-23 08:43:04 +02:00
|
|
|
payload_length) < 0)
|
|
|
|
return DATAOUT_CANNOT_RECOVER;
|
|
|
|
send_r2t = 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!conn->sess->sess_ops->DataPDUInOrder) {
|
|
|
|
ret = iscsit_dataout_update_datapduinorder_no(
|
2012-09-26 14:00:40 +02:00
|
|
|
cmd, be32_to_cpu(hdr->datasn),
|
2011-07-23 08:43:04 +02:00
|
|
|
(hdr->flags & ISCSI_FLAG_CMD_FINAL));
|
|
|
|
if (ret == DATAOUT_CANNOT_RECOVER)
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
seq->data_sn++;
|
|
|
|
seq->offset += payload_length;
|
|
|
|
seq->next_burst_len += payload_length;
|
|
|
|
|
|
|
|
if (send_r2t) {
|
|
|
|
seq->next_burst_len = 0;
|
|
|
|
seq->status = DATAOUT_SEQUENCE_COMPLETE;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (send_r2t && conn->sess->sess_ops->DataSequenceInOrder)
|
|
|
|
cmd->data_sn = 0;
|
|
|
|
|
|
|
|
cmd->write_data_done += payload_length;
|
|
|
|
|
2012-04-04 00:51:24 +02:00
|
|
|
if (cmd->write_data_done == cmd->se_cmd.data_length)
|
2012-04-04 00:51:02 +02:00
|
|
|
return DATAOUT_SEND_TO_TRANSPORT;
|
|
|
|
else if (send_r2t)
|
|
|
|
return DATAOUT_SEND_R2T;
|
|
|
|
else
|
|
|
|
return DATAOUT_NORMAL;
|
2011-07-23 08:43:04 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
static int iscsit_dataout_post_crc_failed(
|
|
|
|
struct iscsi_cmd *cmd,
|
|
|
|
unsigned char *buf)
|
|
|
|
{
|
|
|
|
struct iscsi_conn *conn = cmd->conn;
|
|
|
|
struct iscsi_pdu *pdu;
|
|
|
|
struct iscsi_data *hdr = (struct iscsi_data *) buf;
|
|
|
|
u32 payload_length = ntoh24(hdr->dlength);
|
|
|
|
|
|
|
|
if (conn->sess->sess_ops->DataPDUInOrder)
|
|
|
|
goto recover;
|
|
|
|
/*
|
|
|
|
* The rest of this function is only called when DataPDUInOrder=No.
|
|
|
|
*/
|
|
|
|
pdu = cmd->pdu_ptr;
|
|
|
|
|
|
|
|
switch (pdu->status) {
|
|
|
|
case ISCSI_PDU_NOT_RECEIVED:
|
|
|
|
pdu->status = ISCSI_PDU_CRC_FAILED;
|
|
|
|
break;
|
|
|
|
case ISCSI_PDU_CRC_FAILED:
|
|
|
|
break;
|
|
|
|
case ISCSI_PDU_TIMED_OUT:
|
|
|
|
pdu->status = ISCSI_PDU_CRC_FAILED;
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
return DATAOUT_CANNOT_RECOVER;
|
|
|
|
}
|
|
|
|
|
|
|
|
recover:
|
2012-09-26 14:00:40 +02:00
|
|
|
return iscsit_recover_dataout_sequence(cmd, be32_to_cpu(hdr->offset),
|
|
|
|
payload_length);
|
2011-07-23 08:43:04 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Called from iscsit_handle_data_out() before DataOUT Payload is received
|
|
|
|
* and CRC computed.
|
|
|
|
*/
|
2012-09-26 14:00:36 +02:00
|
|
|
int iscsit_check_pre_dataout(
|
2011-07-23 08:43:04 +02:00
|
|
|
struct iscsi_cmd *cmd,
|
|
|
|
unsigned char *buf)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
struct iscsi_conn *conn = cmd->conn;
|
|
|
|
|
|
|
|
ret = iscsit_dataout_within_command_recovery_check(cmd, buf);
|
|
|
|
if ((ret == DATAOUT_WITHIN_COMMAND_RECOVERY) ||
|
|
|
|
(ret == DATAOUT_CANNOT_RECOVER))
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
ret = iscsit_dataout_check_datasn(cmd, buf);
|
|
|
|
if ((ret == DATAOUT_WITHIN_COMMAND_RECOVERY) ||
|
|
|
|
(ret == DATAOUT_CANNOT_RECOVER))
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
if (cmd->unsolicited_data) {
|
|
|
|
ret = iscsit_dataout_check_unsolicited_sequence(cmd, buf);
|
|
|
|
if ((ret == DATAOUT_WITHIN_COMMAND_RECOVERY) ||
|
|
|
|
(ret == DATAOUT_CANNOT_RECOVER))
|
|
|
|
return ret;
|
|
|
|
} else {
|
|
|
|
ret = iscsit_dataout_check_sequence(cmd, buf);
|
|
|
|
if ((ret == DATAOUT_WITHIN_COMMAND_RECOVERY) ||
|
|
|
|
(ret == DATAOUT_CANNOT_RECOVER))
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
return (conn->sess->sess_ops->DataPDUInOrder) ?
|
|
|
|
iscsit_dataout_pre_datapduinorder_yes(cmd, buf) :
|
|
|
|
iscsit_dataout_pre_datapduinorder_no(cmd, buf);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Called from iscsit_handle_data_out() after DataOUT Payload is received
|
|
|
|
* and CRC computed.
|
|
|
|
*/
|
|
|
|
int iscsit_check_post_dataout(
|
|
|
|
struct iscsi_cmd *cmd,
|
|
|
|
unsigned char *buf,
|
|
|
|
u8 data_crc_failed)
|
|
|
|
{
|
|
|
|
struct iscsi_conn *conn = cmd->conn;
|
|
|
|
|
|
|
|
cmd->dataout_timeout_retries = 0;
|
|
|
|
|
|
|
|
if (!data_crc_failed)
|
|
|
|
return iscsit_dataout_post_crc_passed(cmd, buf);
|
|
|
|
else {
|
|
|
|
if (!conn->sess->sess_ops->ErrorRecoveryLevel) {
|
|
|
|
pr_err("Unable to recover from DataOUT CRC"
|
|
|
|
" failure while ERL=0, closing session.\n");
|
2013-07-03 12:48:24 +02:00
|
|
|
iscsit_reject_cmd(cmd, ISCSI_REASON_DATA_DIGEST_ERROR,
|
|
|
|
buf);
|
2011-07-23 08:43:04 +02:00
|
|
|
return DATAOUT_CANNOT_RECOVER;
|
|
|
|
}
|
|
|
|
|
2013-07-03 12:48:24 +02:00
|
|
|
iscsit_reject_cmd(cmd, ISCSI_REASON_DATA_DIGEST_ERROR, buf);
|
2011-07-23 08:43:04 +02:00
|
|
|
return iscsit_dataout_post_crc_failed(cmd, buf);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void iscsit_handle_time2retain_timeout(unsigned long data)
|
|
|
|
{
|
|
|
|
struct iscsi_session *sess = (struct iscsi_session *) data;
|
|
|
|
struct iscsi_portal_group *tpg = ISCSI_TPG_S(sess);
|
|
|
|
struct se_portal_group *se_tpg = &tpg->tpg_se_tpg;
|
|
|
|
|
|
|
|
spin_lock_bh(&se_tpg->session_lock);
|
|
|
|
if (sess->time2retain_timer_flags & ISCSI_TF_STOP) {
|
|
|
|
spin_unlock_bh(&se_tpg->session_lock);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
if (atomic_read(&sess->session_reinstatement)) {
|
|
|
|
pr_err("Exiting Time2Retain handler because"
|
|
|
|
" session_reinstatement=1\n");
|
|
|
|
spin_unlock_bh(&se_tpg->session_lock);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
sess->time2retain_timer_flags |= ISCSI_TF_EXPIRED;
|
|
|
|
|
|
|
|
pr_err("Time2Retain timer expired for SID: %u, cleaning up"
|
|
|
|
" iSCSI session.\n", sess->sid);
|
|
|
|
{
|
|
|
|
struct iscsi_tiqn *tiqn = tpg->tpg_tiqn;
|
|
|
|
|
|
|
|
if (tiqn) {
|
|
|
|
spin_lock(&tiqn->sess_err_stats.lock);
|
|
|
|
strcpy(tiqn->sess_err_stats.last_sess_fail_rem_name,
|
|
|
|
(void *)sess->sess_ops->InitiatorName);
|
|
|
|
tiqn->sess_err_stats.last_sess_failure_type =
|
|
|
|
ISCSI_SESS_ERR_CXN_TIMEOUT;
|
|
|
|
tiqn->sess_err_stats.cxn_timeout_errors++;
|
|
|
|
sess->conn_timeout_errors++;
|
|
|
|
spin_unlock(&tiqn->sess_err_stats.lock);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
spin_unlock_bh(&se_tpg->session_lock);
|
2012-02-27 10:43:32 +01:00
|
|
|
target_put_session(sess->se_sess);
|
2011-07-23 08:43:04 +02:00
|
|
|
}
|
|
|
|
|
2012-09-26 14:00:36 +02:00
|
|
|
void iscsit_start_time2retain_handler(struct iscsi_session *sess)
|
2011-07-23 08:43:04 +02:00
|
|
|
{
|
|
|
|
int tpg_active;
|
|
|
|
/*
|
2012-08-27 15:46:00 +02:00
|
|
|
* Only start Time2Retain timer when the associated TPG is still in
|
2011-07-23 08:43:04 +02:00
|
|
|
* an ACTIVE (eg: not disabled or shutdown) state.
|
|
|
|
*/
|
|
|
|
spin_lock(&ISCSI_TPG_S(sess)->tpg_state_lock);
|
|
|
|
tpg_active = (ISCSI_TPG_S(sess)->tpg_state == TPG_STATE_ACTIVE);
|
|
|
|
spin_unlock(&ISCSI_TPG_S(sess)->tpg_state_lock);
|
|
|
|
|
|
|
|
if (!tpg_active)
|
|
|
|
return;
|
|
|
|
|
|
|
|
if (sess->time2retain_timer_flags & ISCSI_TF_RUNNING)
|
|
|
|
return;
|
|
|
|
|
|
|
|
pr_debug("Starting Time2Retain timer for %u seconds on"
|
|
|
|
" SID: %u\n", sess->sess_ops->DefaultTime2Retain, sess->sid);
|
|
|
|
|
|
|
|
init_timer(&sess->time2retain_timer);
|
|
|
|
sess->time2retain_timer.expires =
|
|
|
|
(get_jiffies_64() + sess->sess_ops->DefaultTime2Retain * HZ);
|
|
|
|
sess->time2retain_timer.data = (unsigned long)sess;
|
|
|
|
sess->time2retain_timer.function = iscsit_handle_time2retain_timeout;
|
|
|
|
sess->time2retain_timer_flags &= ~ISCSI_TF_STOP;
|
|
|
|
sess->time2retain_timer_flags |= ISCSI_TF_RUNNING;
|
|
|
|
add_timer(&sess->time2retain_timer);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Called with spin_lock_bh(&struct se_portal_group->session_lock) held
|
|
|
|
*/
|
2012-09-26 14:00:36 +02:00
|
|
|
int iscsit_stop_time2retain_timer(struct iscsi_session *sess)
|
2011-07-23 08:43:04 +02:00
|
|
|
{
|
|
|
|
struct iscsi_portal_group *tpg = ISCSI_TPG_S(sess);
|
|
|
|
struct se_portal_group *se_tpg = &tpg->tpg_se_tpg;
|
|
|
|
|
|
|
|
if (sess->time2retain_timer_flags & ISCSI_TF_EXPIRED)
|
|
|
|
return -1;
|
|
|
|
|
|
|
|
if (!(sess->time2retain_timer_flags & ISCSI_TF_RUNNING))
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
sess->time2retain_timer_flags |= ISCSI_TF_STOP;
|
target/iscsi: don't corrupt bh_count in iscsit_stop_time2retain_timer()
Here is a fun one. Bug seems to have been introduced by commit 140854cb,
almost two years ago. I have no idea why we only started seeing it now,
but we did.
Rough callgraph:
core_tpg_set_initiator_node_queue_depth()
`-> spin_lock_irqsave(&tpg->session_lock, flags);
`-> lio_tpg_shutdown_session()
`-> iscsit_stop_time2retain_timer()
`-> spin_unlock_bh(&se_tpg->session_lock);
`-> spin_lock_bh(&se_tpg->session_lock);
`-> spin_unlock_irqrestore(&tpg->session_lock, flags);
core_tpg_set_initiator_node_queue_depth() used to call spin_lock_bh(),
but 140854cb changed that to spin_lock_irqsave(). However,
lio_tpg_shutdown_session() still claims to be called with spin_lock_bh()
held, as does iscsit_stop_time2retain_timer():
* Called with spin_lock_bh(&struct se_portal_group->session_lock) held
Stale documentation is mostly annoying, but in this case the dropping
the lock with the _bh variant is plain wrong. It is also wrong to drop
locks two functions below the lock-holder, but I will ignore that bit
for now.
After some more locking and unlocking we eventually hit this backtrace:
------------[ cut here ]------------
WARNING: at kernel/softirq.c:159 local_bh_enable_ip+0xe8/0x100()
Pid: 24645, comm: lio_helper.py Tainted: G O 3.6.11+
Call Trace:
[<ffffffff8103e5ff>] warn_slowpath_common+0x7f/0xc0
[<ffffffffa040ae37>] ? iscsit_inc_conn_usage_count+0x37/0x50 [iscsi_target_mod]
[<ffffffff8103e65a>] warn_slowpath_null+0x1a/0x20
[<ffffffff810472f8>] local_bh_enable_ip+0xe8/0x100
[<ffffffff815b8365>] _raw_spin_unlock_bh+0x15/0x20
[<ffffffffa040ae37>] iscsit_inc_conn_usage_count+0x37/0x50 [iscsi_target_mod]
[<ffffffffa041149a>] iscsit_stop_session+0xfa/0x1c0 [iscsi_target_mod]
[<ffffffffa0417fab>] lio_tpg_shutdown_session+0x7b/0x90 [iscsi_target_mod]
[<ffffffffa033ede4>] core_tpg_set_initiator_node_queue_depth+0xe4/0x290 [target_core_mod]
[<ffffffffa0409032>] iscsit_tpg_set_initiator_node_queue_depth+0x12/0x20 [iscsi_target_mod]
[<ffffffffa0415c29>] lio_target_nacl_store_cmdsn_depth+0xa9/0x180 [iscsi_target_mod]
[<ffffffffa0331b49>] target_fabric_nacl_base_attr_store+0x39/0x40 [target_core_mod]
[<ffffffff811b857d>] configfs_write_file+0xbd/0x120
[<ffffffff81148f36>] vfs_write+0xc6/0x180
[<ffffffff81149251>] sys_write+0x51/0x90
[<ffffffff815c0969>] system_call_fastpath+0x16/0x1b
---[ end trace 3747632b9b164652 ]---
As a pure band-aid, this patch drops the _bh.
Signed-off-by: Joern Engel <joern@logfs.org>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
2013-05-30 22:36:51 +02:00
|
|
|
spin_unlock(&se_tpg->session_lock);
|
2011-07-23 08:43:04 +02:00
|
|
|
|
|
|
|
del_timer_sync(&sess->time2retain_timer);
|
|
|
|
|
target/iscsi: don't corrupt bh_count in iscsit_stop_time2retain_timer()
Here is a fun one. Bug seems to have been introduced by commit 140854cb,
almost two years ago. I have no idea why we only started seeing it now,
but we did.
Rough callgraph:
core_tpg_set_initiator_node_queue_depth()
`-> spin_lock_irqsave(&tpg->session_lock, flags);
`-> lio_tpg_shutdown_session()
`-> iscsit_stop_time2retain_timer()
`-> spin_unlock_bh(&se_tpg->session_lock);
`-> spin_lock_bh(&se_tpg->session_lock);
`-> spin_unlock_irqrestore(&tpg->session_lock, flags);
core_tpg_set_initiator_node_queue_depth() used to call spin_lock_bh(),
but 140854cb changed that to spin_lock_irqsave(). However,
lio_tpg_shutdown_session() still claims to be called with spin_lock_bh()
held, as does iscsit_stop_time2retain_timer():
* Called with spin_lock_bh(&struct se_portal_group->session_lock) held
Stale documentation is mostly annoying, but in this case the dropping
the lock with the _bh variant is plain wrong. It is also wrong to drop
locks two functions below the lock-holder, but I will ignore that bit
for now.
After some more locking and unlocking we eventually hit this backtrace:
------------[ cut here ]------------
WARNING: at kernel/softirq.c:159 local_bh_enable_ip+0xe8/0x100()
Pid: 24645, comm: lio_helper.py Tainted: G O 3.6.11+
Call Trace:
[<ffffffff8103e5ff>] warn_slowpath_common+0x7f/0xc0
[<ffffffffa040ae37>] ? iscsit_inc_conn_usage_count+0x37/0x50 [iscsi_target_mod]
[<ffffffff8103e65a>] warn_slowpath_null+0x1a/0x20
[<ffffffff810472f8>] local_bh_enable_ip+0xe8/0x100
[<ffffffff815b8365>] _raw_spin_unlock_bh+0x15/0x20
[<ffffffffa040ae37>] iscsit_inc_conn_usage_count+0x37/0x50 [iscsi_target_mod]
[<ffffffffa041149a>] iscsit_stop_session+0xfa/0x1c0 [iscsi_target_mod]
[<ffffffffa0417fab>] lio_tpg_shutdown_session+0x7b/0x90 [iscsi_target_mod]
[<ffffffffa033ede4>] core_tpg_set_initiator_node_queue_depth+0xe4/0x290 [target_core_mod]
[<ffffffffa0409032>] iscsit_tpg_set_initiator_node_queue_depth+0x12/0x20 [iscsi_target_mod]
[<ffffffffa0415c29>] lio_target_nacl_store_cmdsn_depth+0xa9/0x180 [iscsi_target_mod]
[<ffffffffa0331b49>] target_fabric_nacl_base_attr_store+0x39/0x40 [target_core_mod]
[<ffffffff811b857d>] configfs_write_file+0xbd/0x120
[<ffffffff81148f36>] vfs_write+0xc6/0x180
[<ffffffff81149251>] sys_write+0x51/0x90
[<ffffffff815c0969>] system_call_fastpath+0x16/0x1b
---[ end trace 3747632b9b164652 ]---
As a pure band-aid, this patch drops the _bh.
Signed-off-by: Joern Engel <joern@logfs.org>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
2013-05-30 22:36:51 +02:00
|
|
|
spin_lock(&se_tpg->session_lock);
|
2011-07-23 08:43:04 +02:00
|
|
|
sess->time2retain_timer_flags &= ~ISCSI_TF_RUNNING;
|
|
|
|
pr_debug("Stopped Time2Retain Timer for SID: %u\n",
|
|
|
|
sess->sid);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
void iscsit_connection_reinstatement_rcfr(struct iscsi_conn *conn)
|
|
|
|
{
|
|
|
|
spin_lock_bh(&conn->state_lock);
|
|
|
|
if (atomic_read(&conn->connection_exit)) {
|
|
|
|
spin_unlock_bh(&conn->state_lock);
|
|
|
|
goto sleep;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (atomic_read(&conn->transport_failed)) {
|
|
|
|
spin_unlock_bh(&conn->state_lock);
|
|
|
|
goto sleep;
|
|
|
|
}
|
|
|
|
spin_unlock_bh(&conn->state_lock);
|
|
|
|
|
|
|
|
iscsi_thread_set_force_reinstatement(conn);
|
|
|
|
|
|
|
|
sleep:
|
|
|
|
wait_for_completion(&conn->conn_wait_rcfr_comp);
|
|
|
|
complete(&conn->conn_post_wait_comp);
|
|
|
|
}
|
|
|
|
|
|
|
|
void iscsit_cause_connection_reinstatement(struct iscsi_conn *conn, int sleep)
|
|
|
|
{
|
|
|
|
spin_lock_bh(&conn->state_lock);
|
|
|
|
if (atomic_read(&conn->connection_exit)) {
|
|
|
|
spin_unlock_bh(&conn->state_lock);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (atomic_read(&conn->transport_failed)) {
|
|
|
|
spin_unlock_bh(&conn->state_lock);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (atomic_read(&conn->connection_reinstatement)) {
|
|
|
|
spin_unlock_bh(&conn->state_lock);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (iscsi_thread_set_force_reinstatement(conn) < 0) {
|
|
|
|
spin_unlock_bh(&conn->state_lock);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
atomic_set(&conn->connection_reinstatement, 1);
|
|
|
|
if (!sleep) {
|
|
|
|
spin_unlock_bh(&conn->state_lock);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
atomic_set(&conn->sleep_on_conn_wait_comp, 1);
|
|
|
|
spin_unlock_bh(&conn->state_lock);
|
|
|
|
|
|
|
|
wait_for_completion(&conn->conn_wait_comp);
|
|
|
|
complete(&conn->conn_post_wait_comp);
|
|
|
|
}
|
2013-07-03 12:05:37 +02:00
|
|
|
EXPORT_SYMBOL(iscsit_cause_connection_reinstatement);
|
2011-07-23 08:43:04 +02:00
|
|
|
|
|
|
|
void iscsit_fall_back_to_erl0(struct iscsi_session *sess)
|
|
|
|
{
|
|
|
|
pr_debug("Falling back to ErrorRecoveryLevel=0 for SID:"
|
|
|
|
" %u\n", sess->sid);
|
|
|
|
|
|
|
|
atomic_set(&sess->session_fall_back_to_erl0, 1);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void iscsit_handle_connection_cleanup(struct iscsi_conn *conn)
|
|
|
|
{
|
|
|
|
struct iscsi_session *sess = conn->sess;
|
|
|
|
|
|
|
|
if ((sess->sess_ops->ErrorRecoveryLevel == 2) &&
|
|
|
|
!atomic_read(&sess->session_reinstatement) &&
|
|
|
|
!atomic_read(&sess->session_fall_back_to_erl0))
|
|
|
|
iscsit_connection_recovery_transport_reset(conn);
|
|
|
|
else {
|
|
|
|
pr_debug("Performing cleanup for failed iSCSI"
|
|
|
|
" Connection ID: %hu from %s\n", conn->cid,
|
|
|
|
sess->sess_ops->InitiatorName);
|
|
|
|
iscsit_close_connection(conn);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2012-09-26 14:00:36 +02:00
|
|
|
void iscsit_take_action_for_connection_exit(struct iscsi_conn *conn)
|
2011-07-23 08:43:04 +02:00
|
|
|
{
|
|
|
|
spin_lock_bh(&conn->state_lock);
|
|
|
|
if (atomic_read(&conn->connection_exit)) {
|
|
|
|
spin_unlock_bh(&conn->state_lock);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
atomic_set(&conn->connection_exit, 1);
|
|
|
|
|
|
|
|
if (conn->conn_state == TARG_CONN_STATE_IN_LOGOUT) {
|
|
|
|
spin_unlock_bh(&conn->state_lock);
|
|
|
|
iscsit_close_connection(conn);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (conn->conn_state == TARG_CONN_STATE_CLEANUP_WAIT) {
|
|
|
|
spin_unlock_bh(&conn->state_lock);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
pr_debug("Moving to TARG_CONN_STATE_CLEANUP_WAIT.\n");
|
|
|
|
conn->conn_state = TARG_CONN_STATE_CLEANUP_WAIT;
|
|
|
|
spin_unlock_bh(&conn->state_lock);
|
|
|
|
|
|
|
|
iscsit_handle_connection_cleanup(conn);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* This is the simple function that makes the magic of
|
|
|
|
* sync and steering happen in the follow paradoxical order:
|
|
|
|
*
|
|
|
|
* 0) Receive conn->of_marker (bytes left until next OFMarker)
|
|
|
|
* bytes into an offload buffer. When we pass the exact number
|
|
|
|
* of bytes in conn->of_marker, iscsit_dump_data_payload() and hence
|
|
|
|
* rx_data() will automatically receive the identical u32 marker
|
|
|
|
* values and store it in conn->of_marker_offset;
|
|
|
|
* 1) Now conn->of_marker_offset will contain the offset to the start
|
|
|
|
* of the next iSCSI PDU. Dump these remaining bytes into another
|
|
|
|
* offload buffer.
|
|
|
|
* 2) We are done!
|
|
|
|
* Next byte in the TCP stream will contain the next iSCSI PDU!
|
|
|
|
* Cool Huh?!
|
|
|
|
*/
|
|
|
|
int iscsit_recover_from_unknown_opcode(struct iscsi_conn *conn)
|
|
|
|
{
|
|
|
|
/*
|
|
|
|
* Make sure the remaining bytes to next maker is a sane value.
|
|
|
|
*/
|
|
|
|
if (conn->of_marker > (conn->conn_ops->OFMarkInt * 4)) {
|
|
|
|
pr_err("Remaining bytes to OFMarker: %u exceeds"
|
|
|
|
" OFMarkInt bytes: %u.\n", conn->of_marker,
|
|
|
|
conn->conn_ops->OFMarkInt * 4);
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
|
|
|
pr_debug("Advancing %u bytes in TCP stream to get to the"
|
|
|
|
" next OFMarker.\n", conn->of_marker);
|
|
|
|
|
|
|
|
if (iscsit_dump_data_payload(conn, conn->of_marker, 0) < 0)
|
|
|
|
return -1;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Make sure the offset marker we retrived is a valid value.
|
|
|
|
*/
|
|
|
|
if (conn->of_marker_offset > (ISCSI_HDR_LEN + (ISCSI_CRC_LEN * 2) +
|
|
|
|
conn->conn_ops->MaxRecvDataSegmentLength)) {
|
|
|
|
pr_err("OfMarker offset value: %u exceeds limit.\n",
|
|
|
|
conn->of_marker_offset);
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
|
|
|
pr_debug("Discarding %u bytes of TCP stream to get to the"
|
|
|
|
" next iSCSI Opcode.\n", conn->of_marker_offset);
|
|
|
|
|
|
|
|
if (iscsit_dump_data_payload(conn, conn->of_marker_offset, 0) < 0)
|
|
|
|
return -1;
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|