OpenCloudOS-Kernel/drivers/scsi/aacraid/commsup.c

2032 lines
57 KiB
C
Raw Normal View History

/*
* Adaptec AAC series RAID controller driver
* (c) Copyright 2001 Red Hat Inc.
*
* based on the old aacraid driver that is..
* Adaptec aacraid device driver for Linux.
*
* Copyright (c) 2000-2010 Adaptec, Inc.
* 2010 PMC-Sierra, Inc. (aacraid@pmc-sierra.com)
*
* 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, 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.
*
* You should have received a copy of the GNU General Public License
* along with this program; see the file COPYING. If not, write to
* the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA.
*
* Module Name:
* commsup.c
*
* Abstract: Contain all routines that are required for FSA host/adapter
* communication.
*
*/
#include <linux/kernel.h>
#include <linux/init.h>
#include <linux/types.h>
#include <linux/sched.h>
#include <linux/pci.h>
#include <linux/spinlock.h>
#include <linux/slab.h>
#include <linux/completion.h>
#include <linux/blkdev.h>
#include <linux/delay.h>
#include <linux/kthread.h>
#include <linux/interrupt.h>
#include <linux/semaphore.h>
#include <scsi/scsi.h>
#include <scsi/scsi_host.h>
#include <scsi/scsi_device.h>
#include <scsi/scsi_cmnd.h>
#include "aacraid.h"
/**
* fib_map_alloc - allocate the fib objects
* @dev: Adapter to allocate for
*
* Allocate and map the shared PCI space for the FIB blocks used to
* talk to the Adaptec firmware.
*/
static int fib_map_alloc(struct aac_dev *dev)
{
dprintk((KERN_INFO
"allocate hardware fibs pci_alloc_consistent(%p, %d * (%d + %d), %p)\n",
dev->pdev, dev->max_fib_size, dev->scsi_host_ptr->can_queue,
AAC_NUM_MGT_FIB, &dev->hw_fib_pa));
dev->hw_fib_va = pci_alloc_consistent(dev->pdev,
(dev->max_fib_size + sizeof(struct aac_fib_xporthdr))
* (dev->scsi_host_ptr->can_queue + AAC_NUM_MGT_FIB) + (ALIGN32 - 1),
&dev->hw_fib_pa);
if (dev->hw_fib_va == NULL)
return -ENOMEM;
return 0;
}
/**
* aac_fib_map_free - free the fib objects
* @dev: Adapter to free
*
* Free the PCI mappings and the memory allocated for FIB blocks
* on this adapter.
*/
void aac_fib_map_free(struct aac_dev *dev)
{
pci_free_consistent(dev->pdev,
dev->max_fib_size * (dev->scsi_host_ptr->can_queue + AAC_NUM_MGT_FIB),
dev->hw_fib_va, dev->hw_fib_pa);
dev->hw_fib_va = NULL;
dev->hw_fib_pa = 0;
}
/**
* aac_fib_setup - setup the fibs
* @dev: Adapter to set up
*
* Allocate the PCI space for the fibs, map it and then initialise the
* fib area, the unmapped fib data and also the free list
*/
int aac_fib_setup(struct aac_dev * dev)
{
struct fib *fibptr;
struct hw_fib *hw_fib;
dma_addr_t hw_fib_pa;
int i;
while (((i = fib_map_alloc(dev)) == -ENOMEM)
&& (dev->scsi_host_ptr->can_queue > (64 - AAC_NUM_MGT_FIB))) {
dev->init->MaxIoCommands = cpu_to_le32((dev->scsi_host_ptr->can_queue + AAC_NUM_MGT_FIB) >> 1);
dev->scsi_host_ptr->can_queue = le32_to_cpu(dev->init->MaxIoCommands) - AAC_NUM_MGT_FIB;
}
if (i<0)
return -ENOMEM;
/* 32 byte alignment for PMC */
hw_fib_pa = (dev->hw_fib_pa + (ALIGN32 - 1)) & ~(ALIGN32 - 1);
dev->hw_fib_va = (struct hw_fib *)((unsigned char *)dev->hw_fib_va +
(hw_fib_pa - dev->hw_fib_pa));
dev->hw_fib_pa = hw_fib_pa;
memset(dev->hw_fib_va, 0,
(dev->max_fib_size + sizeof(struct aac_fib_xporthdr)) *
(dev->scsi_host_ptr->can_queue + AAC_NUM_MGT_FIB));
/* add Xport header */
dev->hw_fib_va = (struct hw_fib *)((unsigned char *)dev->hw_fib_va +
sizeof(struct aac_fib_xporthdr));
dev->hw_fib_pa += sizeof(struct aac_fib_xporthdr);
hw_fib = dev->hw_fib_va;
hw_fib_pa = dev->hw_fib_pa;
/*
* Initialise the fibs
*/
for (i = 0, fibptr = &dev->fibs[i];
i < (dev->scsi_host_ptr->can_queue + AAC_NUM_MGT_FIB);
i++, fibptr++)
{
fibptr->flags = 0;
fibptr->dev = dev;
fibptr->hw_fib_va = hw_fib;
fibptr->data = (void *) fibptr->hw_fib_va->data;
fibptr->next = fibptr+1; /* Forward chain the fibs */
sema_init(&fibptr->event_wait, 0);
spin_lock_init(&fibptr->event_lock);
hw_fib->header.XferState = cpu_to_le32(0xffffffff);
hw_fib->header.SenderSize = cpu_to_le16(dev->max_fib_size);
fibptr->hw_fib_pa = hw_fib_pa;
hw_fib = (struct hw_fib *)((unsigned char *)hw_fib +
dev->max_fib_size + sizeof(struct aac_fib_xporthdr));
hw_fib_pa = hw_fib_pa +
dev->max_fib_size + sizeof(struct aac_fib_xporthdr);
}
/*
* Add the fib chain to the free list
*/
dev->fibs[dev->scsi_host_ptr->can_queue + AAC_NUM_MGT_FIB - 1].next = NULL;
/*
* Enable this to debug out of queue space
*/
dev->free_fib = &dev->fibs[0];
return 0;
}
/**
* aac_fib_alloc - allocate a fib
* @dev: Adapter to allocate the fib for
*
* Allocate a fib from the adapter fib pool. If the pool is empty we
* return NULL.
*/
struct fib *aac_fib_alloc(struct aac_dev *dev)
{
struct fib * fibptr;
unsigned long flags;
spin_lock_irqsave(&dev->fib_lock, flags);
fibptr = dev->free_fib;
if(!fibptr){
spin_unlock_irqrestore(&dev->fib_lock, flags);
return fibptr;
}
dev->free_fib = fibptr->next;
spin_unlock_irqrestore(&dev->fib_lock, flags);
/*
* Set the proper node type code and node byte size
*/
fibptr->type = FSAFS_NTC_FIB_CONTEXT;
fibptr->size = sizeof(struct fib);
/*
* Null out fields that depend on being zero at the start of
* each I/O
*/
fibptr->hw_fib_va->header.XferState = 0;
fibptr->flags = 0;
fibptr->callback = NULL;
fibptr->callback_data = NULL;
return fibptr;
}
/**
* aac_fib_free - free a fib
* @fibptr: fib to free up
*
* Frees up a fib and places it on the appropriate queue
*/
void aac_fib_free(struct fib *fibptr)
{
unsigned long flags;
[SCSI] aacraid: fix File System going into read-only mode These particular problems were reported by Cisco and SAP and customers as well. Cisco reported on RHEL4 U6 and SAP reported on SLES9 SP4 and SLES10 SP2. We added these fixes on RHEL4 U6 and gave a private build to IBM and Cisco. Cisco and IBM tested it for more than 15 days and they reported that they did not see the issue so far. Before the fix, Cisco used to see the issue within 5 days. We generated a patch for SLES9 SP4 and SLES10 SP2 and submitted to Novell. Novell applied the patch and gave a test build to SAP. SAP tested and reported that the build is working properly. We also tested in our lab using the tools "dishogsync", which is IO stress tool and the tool was provided by Cisco. Issue1: File System going into read-only mode Root cause: The driver tends to not free the memory (FIB) when the management request exits prematurely. The accumulation of such un-freed memory causes the driver to fail to allocate anymore memory (FIB) and hence return 0x70000 value to the upper layer, which puts the file system into read only mode. Fix details: The fix makes sure to free the memory (FIB) even if the request exits prematurely hence ensuring the driver wouldn't run out of memory (FIBs). Issue2: False Raid Alert occurs When the Physical Drives and Logical drives are reported as deleted or added, even though there is no change done on the system Root cause: Driver IOCTLs is signaled with EINTR while waiting on response from the lower layers. Returning "EINTR" will never initiate internal retry. Fix details: The issue was fixed by replacing "EINTR" with "ERESTARTSYS" for mid-layer retries. Signed-off-by: Penchala Narasimha Reddy <ServeRAIDDriver@hcl.in> Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2009-12-21 21:09:27 +08:00
if (fibptr->done == 2)
[SCSI] aacraid: fix File System going into read-only mode These particular problems were reported by Cisco and SAP and customers as well. Cisco reported on RHEL4 U6 and SAP reported on SLES9 SP4 and SLES10 SP2. We added these fixes on RHEL4 U6 and gave a private build to IBM and Cisco. Cisco and IBM tested it for more than 15 days and they reported that they did not see the issue so far. Before the fix, Cisco used to see the issue within 5 days. We generated a patch for SLES9 SP4 and SLES10 SP2 and submitted to Novell. Novell applied the patch and gave a test build to SAP. SAP tested and reported that the build is working properly. We also tested in our lab using the tools "dishogsync", which is IO stress tool and the tool was provided by Cisco. Issue1: File System going into read-only mode Root cause: The driver tends to not free the memory (FIB) when the management request exits prematurely. The accumulation of such un-freed memory causes the driver to fail to allocate anymore memory (FIB) and hence return 0x70000 value to the upper layer, which puts the file system into read only mode. Fix details: The fix makes sure to free the memory (FIB) even if the request exits prematurely hence ensuring the driver wouldn't run out of memory (FIBs). Issue2: False Raid Alert occurs When the Physical Drives and Logical drives are reported as deleted or added, even though there is no change done on the system Root cause: Driver IOCTLs is signaled with EINTR while waiting on response from the lower layers. Returning "EINTR" will never initiate internal retry. Fix details: The issue was fixed by replacing "EINTR" with "ERESTARTSYS" for mid-layer retries. Signed-off-by: Penchala Narasimha Reddy <ServeRAIDDriver@hcl.in> Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2009-12-21 21:09:27 +08:00
return;
spin_lock_irqsave(&fibptr->dev->fib_lock, flags);
if (unlikely(fibptr->flags & FIB_CONTEXT_FLAG_TIMED_OUT))
aac_config.fib_timeouts++;
if (fibptr->hw_fib_va->header.XferState != 0) {
printk(KERN_WARNING "aac_fib_free, XferState != 0, fibptr = 0x%p, XferState = 0x%x\n",
(void*)fibptr,
le32_to_cpu(fibptr->hw_fib_va->header.XferState));
}
fibptr->next = fibptr->dev->free_fib;
fibptr->dev->free_fib = fibptr;
spin_unlock_irqrestore(&fibptr->dev->fib_lock, flags);
}
/**
* aac_fib_init - initialise a fib
* @fibptr: The fib to initialize
*
* Set up the generic fib fields ready for use
*/
void aac_fib_init(struct fib *fibptr)
{
struct hw_fib *hw_fib = fibptr->hw_fib_va;
memset(&hw_fib->header, 0, sizeof(struct aac_fibhdr));
hw_fib->header.StructType = FIB_MAGIC;
hw_fib->header.Size = cpu_to_le16(fibptr->dev->max_fib_size);
hw_fib->header.XferState = cpu_to_le32(HostOwned | FibInitialized | FibEmpty | FastResponseCapable);
hw_fib->header.u.ReceiverFibAddress = cpu_to_le32(fibptr->hw_fib_pa);
hw_fib->header.SenderSize = cpu_to_le16(fibptr->dev->max_fib_size);
}
/**
* fib_deallocate - deallocate a fib
* @fibptr: fib to deallocate
*
* Will deallocate and return to the free pool the FIB pointed to by the
* caller.
*/
static void fib_dealloc(struct fib * fibptr)
{
struct hw_fib *hw_fib = fibptr->hw_fib_va;
hw_fib->header.XferState = 0;
}
/*
* Commuication primitives define and support the queuing method we use to
* support host to adapter commuication. All queue accesses happen through
* these routines and are the only routines which have a knowledge of the
* how these queues are implemented.
*/
/**
* aac_get_entry - get a queue entry
* @dev: Adapter
* @qid: Queue Number
* @entry: Entry return
* @index: Index return
* @nonotify: notification control
*
* With a priority the routine returns a queue entry if the queue has free entries. If the queue
* is full(no free entries) than no entry is returned and the function returns 0 otherwise 1 is
* returned.
*/
static int aac_get_entry (struct aac_dev * dev, u32 qid, struct aac_entry **entry, u32 * index, unsigned long *nonotify)
{
struct aac_queue * q;
unsigned long idx;
/*
* All of the queues wrap when they reach the end, so we check
* to see if they have reached the end and if they have we just
* set the index back to zero. This is a wrap. You could or off
* the high bits in all updates but this is a bit faster I think.
*/
q = &dev->queues->queue[qid];
idx = *index = le32_to_cpu(*(q->headers.producer));
/* Interrupt Moderation, only interrupt for first two entries */
if (idx != le32_to_cpu(*(q->headers.consumer))) {
if (--idx == 0) {
if (qid == AdapNormCmdQueue)
idx = ADAP_NORM_CMD_ENTRIES;
else
idx = ADAP_NORM_RESP_ENTRIES;
}
if (idx != le32_to_cpu(*(q->headers.consumer)))
*nonotify = 1;
}
if (qid == AdapNormCmdQueue) {
if (*index >= ADAP_NORM_CMD_ENTRIES)
*index = 0; /* Wrap to front of the Producer Queue. */
} else {
if (*index >= ADAP_NORM_RESP_ENTRIES)
*index = 0; /* Wrap to front of the Producer Queue. */
}
/* Queue is full */
if ((*index + 1) == le32_to_cpu(*(q->headers.consumer))) {
printk(KERN_WARNING "Queue %d full, %u outstanding.\n",
qid, atomic_read(&q->numpending));
return 0;
} else {
*entry = q->base + *index;
return 1;
}
}
/**
* aac_queue_get - get the next free QE
* @dev: Adapter
* @index: Returned index
* @priority: Priority of fib
* @fib: Fib to associate with the queue entry
* @wait: Wait if queue full
* @fibptr: Driver fib object to go with fib
* @nonotify: Don't notify the adapter
*
* Gets the next free QE off the requested priorty adapter command
* queue and associates the Fib with the QE. The QE represented by
* index is ready to insert on the queue when this routine returns
* success.
*/
int aac_queue_get(struct aac_dev * dev, u32 * index, u32 qid, struct hw_fib * hw_fib, int wait, struct fib * fibptr, unsigned long *nonotify)
{
struct aac_entry * entry = NULL;
int map = 0;
if (qid == AdapNormCmdQueue) {
/* if no entries wait for some if caller wants to */
while (!aac_get_entry(dev, qid, &entry, index, nonotify)) {
printk(KERN_ERR "GetEntries failed\n");
}
/*
* Setup queue entry with a command, status and fib mapped
*/
entry->size = cpu_to_le32(le16_to_cpu(hw_fib->header.Size));
map = 1;
} else {
while (!aac_get_entry(dev, qid, &entry, index, nonotify)) {
/* if no entries wait for some if caller wants to */
}
/*
* Setup queue entry with command, status and fib mapped
*/
entry->size = cpu_to_le32(le16_to_cpu(hw_fib->header.Size));
entry->addr = hw_fib->header.SenderFibAddress;
/* Restore adapters pointer to the FIB */
hw_fib->header.u.ReceiverFibAddress = hw_fib->header.SenderFibAddress; /* Let the adapter now where to find its data */
map = 0;
}
/*
* If MapFib is true than we need to map the Fib and put pointers
* in the queue entry.
*/
if (map)
entry->addr = cpu_to_le32(fibptr->hw_fib_pa);
return 0;
}
/*
* Define the highest level of host to adapter communication routines.
* These routines will support host to adapter FS commuication. These
* routines have no knowledge of the commuication method used. This level
* sends and receives FIBs. This level has no knowledge of how these FIBs
* get passed back and forth.
*/
/**
* aac_fib_send - send a fib to the adapter
* @command: Command to send
* @fibptr: The fib
* @size: Size of fib data area
* @priority: Priority of Fib
* @wait: Async/sync select
* @reply: True if a reply is wanted
* @callback: Called with reply
* @callback_data: Passed to callback
*
* Sends the requested FIB to the adapter and optionally will wait for a
* response FIB. If the caller does not wish to wait for a response than
* an event to wait on must be supplied. This event will be set when a
* response FIB is received from the adapter.
*/
int aac_fib_send(u16 command, struct fib *fibptr, unsigned long size,
int priority, int wait, int reply, fib_callback callback,
void *callback_data)
{
struct aac_dev * dev = fibptr->dev;
struct hw_fib * hw_fib = fibptr->hw_fib_va;
unsigned long flags = 0;
[SCSI] aacraid: fix File System going into read-only mode These particular problems were reported by Cisco and SAP and customers as well. Cisco reported on RHEL4 U6 and SAP reported on SLES9 SP4 and SLES10 SP2. We added these fixes on RHEL4 U6 and gave a private build to IBM and Cisco. Cisco and IBM tested it for more than 15 days and they reported that they did not see the issue so far. Before the fix, Cisco used to see the issue within 5 days. We generated a patch for SLES9 SP4 and SLES10 SP2 and submitted to Novell. Novell applied the patch and gave a test build to SAP. SAP tested and reported that the build is working properly. We also tested in our lab using the tools "dishogsync", which is IO stress tool and the tool was provided by Cisco. Issue1: File System going into read-only mode Root cause: The driver tends to not free the memory (FIB) when the management request exits prematurely. The accumulation of such un-freed memory causes the driver to fail to allocate anymore memory (FIB) and hence return 0x70000 value to the upper layer, which puts the file system into read only mode. Fix details: The fix makes sure to free the memory (FIB) even if the request exits prematurely hence ensuring the driver wouldn't run out of memory (FIBs). Issue2: False Raid Alert occurs When the Physical Drives and Logical drives are reported as deleted or added, even though there is no change done on the system Root cause: Driver IOCTLs is signaled with EINTR while waiting on response from the lower layers. Returning "EINTR" will never initiate internal retry. Fix details: The issue was fixed by replacing "EINTR" with "ERESTARTSYS" for mid-layer retries. Signed-off-by: Penchala Narasimha Reddy <ServeRAIDDriver@hcl.in> Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2009-12-21 21:09:27 +08:00
unsigned long mflags = 0;
unsigned long sflags = 0;
[SCSI] aacraid: fix File System going into read-only mode These particular problems were reported by Cisco and SAP and customers as well. Cisco reported on RHEL4 U6 and SAP reported on SLES9 SP4 and SLES10 SP2. We added these fixes on RHEL4 U6 and gave a private build to IBM and Cisco. Cisco and IBM tested it for more than 15 days and they reported that they did not see the issue so far. Before the fix, Cisco used to see the issue within 5 days. We generated a patch for SLES9 SP4 and SLES10 SP2 and submitted to Novell. Novell applied the patch and gave a test build to SAP. SAP tested and reported that the build is working properly. We also tested in our lab using the tools "dishogsync", which is IO stress tool and the tool was provided by Cisco. Issue1: File System going into read-only mode Root cause: The driver tends to not free the memory (FIB) when the management request exits prematurely. The accumulation of such un-freed memory causes the driver to fail to allocate anymore memory (FIB) and hence return 0x70000 value to the upper layer, which puts the file system into read only mode. Fix details: The fix makes sure to free the memory (FIB) even if the request exits prematurely hence ensuring the driver wouldn't run out of memory (FIBs). Issue2: False Raid Alert occurs When the Physical Drives and Logical drives are reported as deleted or added, even though there is no change done on the system Root cause: Driver IOCTLs is signaled with EINTR while waiting on response from the lower layers. Returning "EINTR" will never initiate internal retry. Fix details: The issue was fixed by replacing "EINTR" with "ERESTARTSYS" for mid-layer retries. Signed-off-by: Penchala Narasimha Reddy <ServeRAIDDriver@hcl.in> Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2009-12-21 21:09:27 +08:00
if (!(hw_fib->header.XferState & cpu_to_le32(HostOwned)))
return -EBUSY;
/*
* There are 5 cases with the wait and response requested flags.
* The only invalid cases are if the caller requests to wait and
* does not request a response and if the caller does not want a
* response and the Fib is not allocated from pool. If a response
* is not requesed the Fib will just be deallocaed by the DPC
* routine when the response comes back from the adapter. No
* further processing will be done besides deleting the Fib. We
* will have a debug mode where the adapter can notify the host
* it had a problem and the host can log that fact.
*/
fibptr->flags = 0;
if (wait && !reply) {
return -EINVAL;
} else if (!wait && reply) {
hw_fib->header.XferState |= cpu_to_le32(Async | ResponseExpected);
FIB_COUNTER_INCREMENT(aac_config.AsyncSent);
} else if (!wait && !reply) {
hw_fib->header.XferState |= cpu_to_le32(NoResponseExpected);
FIB_COUNTER_INCREMENT(aac_config.NoResponseSent);
} else if (wait && reply) {
hw_fib->header.XferState |= cpu_to_le32(ResponseExpected);
FIB_COUNTER_INCREMENT(aac_config.NormalSent);
}
/*
* Map the fib into 32bits by using the fib number
*/
hw_fib->header.SenderFibAddress = cpu_to_le32(((u32)(fibptr - dev->fibs)) << 2);
hw_fib->header.Handle = (u32)(fibptr - dev->fibs) + 1;
/*
* Set FIB state to indicate where it came from and if we want a
* response from the adapter. Also load the command from the
* caller.
*
* Map the hw fib pointer as a 32bit value
*/
hw_fib->header.Command = cpu_to_le16(command);
hw_fib->header.XferState |= cpu_to_le32(SentFromHost);
/*
* Set the size of the Fib we want to send to the adapter
*/
hw_fib->header.Size = cpu_to_le16(sizeof(struct aac_fibhdr) + size);
if (le16_to_cpu(hw_fib->header.Size) > le16_to_cpu(hw_fib->header.SenderSize)) {
return -EMSGSIZE;
}
/*
* Get a queue entry connect the FIB to it and send an notify
* the adapter a command is ready.
*/
hw_fib->header.XferState |= cpu_to_le32(NormalPriority);
/*
* Fill in the Callback and CallbackContext if we are not
* going to wait.
*/
if (!wait) {
fibptr->callback = callback;
fibptr->callback_data = callback_data;
fibptr->flags = FIB_CONTEXT_FLAG;
}
fibptr->done = 0;
FIB_COUNTER_INCREMENT(aac_config.FibsSent);
dprintk((KERN_DEBUG "Fib contents:.\n"));
dprintk((KERN_DEBUG " Command = %d.\n", le32_to_cpu(hw_fib->header.Command)));
dprintk((KERN_DEBUG " SubCommand = %d.\n", le32_to_cpu(((struct aac_query_mount *)fib_data(fibptr))->command)));
dprintk((KERN_DEBUG " XferState = %x.\n", le32_to_cpu(hw_fib->header.XferState)));
dprintk((KERN_DEBUG " hw_fib va being sent=%p\n",fibptr->hw_fib_va));
dprintk((KERN_DEBUG " hw_fib pa being sent=%lx\n",(ulong)fibptr->hw_fib_pa));
dprintk((KERN_DEBUG " fib being sent=%p\n",fibptr));
if (!dev->queues)
return -EBUSY;
[SCSI] aacraid: fix File System going into read-only mode These particular problems were reported by Cisco and SAP and customers as well. Cisco reported on RHEL4 U6 and SAP reported on SLES9 SP4 and SLES10 SP2. We added these fixes on RHEL4 U6 and gave a private build to IBM and Cisco. Cisco and IBM tested it for more than 15 days and they reported that they did not see the issue so far. Before the fix, Cisco used to see the issue within 5 days. We generated a patch for SLES9 SP4 and SLES10 SP2 and submitted to Novell. Novell applied the patch and gave a test build to SAP. SAP tested and reported that the build is working properly. We also tested in our lab using the tools "dishogsync", which is IO stress tool and the tool was provided by Cisco. Issue1: File System going into read-only mode Root cause: The driver tends to not free the memory (FIB) when the management request exits prematurely. The accumulation of such un-freed memory causes the driver to fail to allocate anymore memory (FIB) and hence return 0x70000 value to the upper layer, which puts the file system into read only mode. Fix details: The fix makes sure to free the memory (FIB) even if the request exits prematurely hence ensuring the driver wouldn't run out of memory (FIBs). Issue2: False Raid Alert occurs When the Physical Drives and Logical drives are reported as deleted or added, even though there is no change done on the system Root cause: Driver IOCTLs is signaled with EINTR while waiting on response from the lower layers. Returning "EINTR" will never initiate internal retry. Fix details: The issue was fixed by replacing "EINTR" with "ERESTARTSYS" for mid-layer retries. Signed-off-by: Penchala Narasimha Reddy <ServeRAIDDriver@hcl.in> Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2009-12-21 21:09:27 +08:00
if (wait) {
spin_lock_irqsave(&dev->manage_lock, mflags);
if (dev->management_fib_count >= AAC_NUM_MGT_FIB) {
printk(KERN_INFO "No management Fibs Available:%d\n",
dev->management_fib_count);
spin_unlock_irqrestore(&dev->manage_lock, mflags);
return -EBUSY;
}
dev->management_fib_count++;
spin_unlock_irqrestore(&dev->manage_lock, mflags);
spin_lock_irqsave(&fibptr->event_lock, flags);
[SCSI] aacraid: fix File System going into read-only mode These particular problems were reported by Cisco and SAP and customers as well. Cisco reported on RHEL4 U6 and SAP reported on SLES9 SP4 and SLES10 SP2. We added these fixes on RHEL4 U6 and gave a private build to IBM and Cisco. Cisco and IBM tested it for more than 15 days and they reported that they did not see the issue so far. Before the fix, Cisco used to see the issue within 5 days. We generated a patch for SLES9 SP4 and SLES10 SP2 and submitted to Novell. Novell applied the patch and gave a test build to SAP. SAP tested and reported that the build is working properly. We also tested in our lab using the tools "dishogsync", which is IO stress tool and the tool was provided by Cisco. Issue1: File System going into read-only mode Root cause: The driver tends to not free the memory (FIB) when the management request exits prematurely. The accumulation of such un-freed memory causes the driver to fail to allocate anymore memory (FIB) and hence return 0x70000 value to the upper layer, which puts the file system into read only mode. Fix details: The fix makes sure to free the memory (FIB) even if the request exits prematurely hence ensuring the driver wouldn't run out of memory (FIBs). Issue2: False Raid Alert occurs When the Physical Drives and Logical drives are reported as deleted or added, even though there is no change done on the system Root cause: Driver IOCTLs is signaled with EINTR while waiting on response from the lower layers. Returning "EINTR" will never initiate internal retry. Fix details: The issue was fixed by replacing "EINTR" with "ERESTARTSYS" for mid-layer retries. Signed-off-by: Penchala Narasimha Reddy <ServeRAIDDriver@hcl.in> Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2009-12-21 21:09:27 +08:00
}
if (dev->sync_mode) {
if (wait)
spin_unlock_irqrestore(&fibptr->event_lock, flags);
spin_lock_irqsave(&dev->sync_lock, sflags);
if (dev->sync_fib) {
list_add_tail(&fibptr->fiblink, &dev->sync_fib_list);
spin_unlock_irqrestore(&dev->sync_lock, sflags);
} else {
dev->sync_fib = fibptr;
spin_unlock_irqrestore(&dev->sync_lock, sflags);
aac_adapter_sync_cmd(dev, SEND_SYNCHRONOUS_FIB,
(u32)fibptr->hw_fib_pa, 0, 0, 0, 0, 0,
NULL, NULL, NULL, NULL, NULL);
}
if (wait) {
fibptr->flags |= FIB_CONTEXT_FLAG_WAIT;
if (down_interruptible(&fibptr->event_wait)) {
fibptr->flags &= ~FIB_CONTEXT_FLAG_WAIT;
return -EFAULT;
}
return 0;
}
return -EINPROGRESS;
}
[SCSI] aacraid: fix File System going into read-only mode These particular problems were reported by Cisco and SAP and customers as well. Cisco reported on RHEL4 U6 and SAP reported on SLES9 SP4 and SLES10 SP2. We added these fixes on RHEL4 U6 and gave a private build to IBM and Cisco. Cisco and IBM tested it for more than 15 days and they reported that they did not see the issue so far. Before the fix, Cisco used to see the issue within 5 days. We generated a patch for SLES9 SP4 and SLES10 SP2 and submitted to Novell. Novell applied the patch and gave a test build to SAP. SAP tested and reported that the build is working properly. We also tested in our lab using the tools "dishogsync", which is IO stress tool and the tool was provided by Cisco. Issue1: File System going into read-only mode Root cause: The driver tends to not free the memory (FIB) when the management request exits prematurely. The accumulation of such un-freed memory causes the driver to fail to allocate anymore memory (FIB) and hence return 0x70000 value to the upper layer, which puts the file system into read only mode. Fix details: The fix makes sure to free the memory (FIB) even if the request exits prematurely hence ensuring the driver wouldn't run out of memory (FIBs). Issue2: False Raid Alert occurs When the Physical Drives and Logical drives are reported as deleted or added, even though there is no change done on the system Root cause: Driver IOCTLs is signaled with EINTR while waiting on response from the lower layers. Returning "EINTR" will never initiate internal retry. Fix details: The issue was fixed by replacing "EINTR" with "ERESTARTSYS" for mid-layer retries. Signed-off-by: Penchala Narasimha Reddy <ServeRAIDDriver@hcl.in> Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2009-12-21 21:09:27 +08:00
if (aac_adapter_deliver(fibptr) != 0) {
printk(KERN_ERR "aac_fib_send: returned -EBUSY\n");
if (wait) {
spin_unlock_irqrestore(&fibptr->event_lock, flags);
spin_lock_irqsave(&dev->manage_lock, mflags);
dev->management_fib_count--;
spin_unlock_irqrestore(&dev->manage_lock, mflags);
}
return -EBUSY;
}
/*
* If the caller wanted us to wait for response wait now.
*/
if (wait) {
spin_unlock_irqrestore(&fibptr->event_lock, flags);
/* Only set for first known interruptable command */
if (wait < 0) {
/*
* *VERY* Dangerous to time out a command, the
* assumption is made that we have no hope of
* functioning because an interrupt routing or other
* hardware failure has occurred.
*/
unsigned long timeout = jiffies + (180 * HZ); /* 3 minutes */
while (down_trylock(&fibptr->event_wait)) {
int blink;
if (time_is_before_eq_jiffies(timeout)) {
struct aac_queue * q = &dev->queues->queue[AdapNormCmdQueue];
atomic_dec(&q->numpending);
if (wait == -1) {
printk(KERN_ERR "aacraid: aac_fib_send: first asynchronous command timed out.\n"
"Usually a result of a PCI interrupt routing problem;\n"
"update mother board BIOS or consider utilizing one of\n"
"the SAFE mode kernel options (acpi, apic etc)\n");
}
return -ETIMEDOUT;
}
if ((blink = aac_adapter_check_health(dev)) > 0) {
if (wait == -1) {
printk(KERN_ERR "aacraid: aac_fib_send: adapter blinkLED 0x%x.\n"
"Usually a result of a serious unrecoverable hardware problem\n",
blink);
}
return -EFAULT;
}
/* We used to udelay() here but that absorbed
* a CPU when a timeout occured. Not very
* useful. */
cpu_relax();
}
} else if (down_interruptible(&fibptr->event_wait)) {
[SCSI] aacraid: fix File System going into read-only mode These particular problems were reported by Cisco and SAP and customers as well. Cisco reported on RHEL4 U6 and SAP reported on SLES9 SP4 and SLES10 SP2. We added these fixes on RHEL4 U6 and gave a private build to IBM and Cisco. Cisco and IBM tested it for more than 15 days and they reported that they did not see the issue so far. Before the fix, Cisco used to see the issue within 5 days. We generated a patch for SLES9 SP4 and SLES10 SP2 and submitted to Novell. Novell applied the patch and gave a test build to SAP. SAP tested and reported that the build is working properly. We also tested in our lab using the tools "dishogsync", which is IO stress tool and the tool was provided by Cisco. Issue1: File System going into read-only mode Root cause: The driver tends to not free the memory (FIB) when the management request exits prematurely. The accumulation of such un-freed memory causes the driver to fail to allocate anymore memory (FIB) and hence return 0x70000 value to the upper layer, which puts the file system into read only mode. Fix details: The fix makes sure to free the memory (FIB) even if the request exits prematurely hence ensuring the driver wouldn't run out of memory (FIBs). Issue2: False Raid Alert occurs When the Physical Drives and Logical drives are reported as deleted or added, even though there is no change done on the system Root cause: Driver IOCTLs is signaled with EINTR while waiting on response from the lower layers. Returning "EINTR" will never initiate internal retry. Fix details: The issue was fixed by replacing "EINTR" with "ERESTARTSYS" for mid-layer retries. Signed-off-by: Penchala Narasimha Reddy <ServeRAIDDriver@hcl.in> Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2009-12-21 21:09:27 +08:00
/* Do nothing ... satisfy
* down_interruptible must_check */
}
[SCSI] aacraid: fix File System going into read-only mode These particular problems were reported by Cisco and SAP and customers as well. Cisco reported on RHEL4 U6 and SAP reported on SLES9 SP4 and SLES10 SP2. We added these fixes on RHEL4 U6 and gave a private build to IBM and Cisco. Cisco and IBM tested it for more than 15 days and they reported that they did not see the issue so far. Before the fix, Cisco used to see the issue within 5 days. We generated a patch for SLES9 SP4 and SLES10 SP2 and submitted to Novell. Novell applied the patch and gave a test build to SAP. SAP tested and reported that the build is working properly. We also tested in our lab using the tools "dishogsync", which is IO stress tool and the tool was provided by Cisco. Issue1: File System going into read-only mode Root cause: The driver tends to not free the memory (FIB) when the management request exits prematurely. The accumulation of such un-freed memory causes the driver to fail to allocate anymore memory (FIB) and hence return 0x70000 value to the upper layer, which puts the file system into read only mode. Fix details: The fix makes sure to free the memory (FIB) even if the request exits prematurely hence ensuring the driver wouldn't run out of memory (FIBs). Issue2: False Raid Alert occurs When the Physical Drives and Logical drives are reported as deleted or added, even though there is no change done on the system Root cause: Driver IOCTLs is signaled with EINTR while waiting on response from the lower layers. Returning "EINTR" will never initiate internal retry. Fix details: The issue was fixed by replacing "EINTR" with "ERESTARTSYS" for mid-layer retries. Signed-off-by: Penchala Narasimha Reddy <ServeRAIDDriver@hcl.in> Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2009-12-21 21:09:27 +08:00
spin_lock_irqsave(&fibptr->event_lock, flags);
[SCSI] aacraid: fix File System going into read-only mode These particular problems were reported by Cisco and SAP and customers as well. Cisco reported on RHEL4 U6 and SAP reported on SLES9 SP4 and SLES10 SP2. We added these fixes on RHEL4 U6 and gave a private build to IBM and Cisco. Cisco and IBM tested it for more than 15 days and they reported that they did not see the issue so far. Before the fix, Cisco used to see the issue within 5 days. We generated a patch for SLES9 SP4 and SLES10 SP2 and submitted to Novell. Novell applied the patch and gave a test build to SAP. SAP tested and reported that the build is working properly. We also tested in our lab using the tools "dishogsync", which is IO stress tool and the tool was provided by Cisco. Issue1: File System going into read-only mode Root cause: The driver tends to not free the memory (FIB) when the management request exits prematurely. The accumulation of such un-freed memory causes the driver to fail to allocate anymore memory (FIB) and hence return 0x70000 value to the upper layer, which puts the file system into read only mode. Fix details: The fix makes sure to free the memory (FIB) even if the request exits prematurely hence ensuring the driver wouldn't run out of memory (FIBs). Issue2: False Raid Alert occurs When the Physical Drives and Logical drives are reported as deleted or added, even though there is no change done on the system Root cause: Driver IOCTLs is signaled with EINTR while waiting on response from the lower layers. Returning "EINTR" will never initiate internal retry. Fix details: The issue was fixed by replacing "EINTR" with "ERESTARTSYS" for mid-layer retries. Signed-off-by: Penchala Narasimha Reddy <ServeRAIDDriver@hcl.in> Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2009-12-21 21:09:27 +08:00
if (fibptr->done == 0) {
fibptr->done = 2; /* Tell interrupt we aborted */
spin_unlock_irqrestore(&fibptr->event_lock, flags);
[SCSI] aacraid: fix File System going into read-only mode These particular problems were reported by Cisco and SAP and customers as well. Cisco reported on RHEL4 U6 and SAP reported on SLES9 SP4 and SLES10 SP2. We added these fixes on RHEL4 U6 and gave a private build to IBM and Cisco. Cisco and IBM tested it for more than 15 days and they reported that they did not see the issue so far. Before the fix, Cisco used to see the issue within 5 days. We generated a patch for SLES9 SP4 and SLES10 SP2 and submitted to Novell. Novell applied the patch and gave a test build to SAP. SAP tested and reported that the build is working properly. We also tested in our lab using the tools "dishogsync", which is IO stress tool and the tool was provided by Cisco. Issue1: File System going into read-only mode Root cause: The driver tends to not free the memory (FIB) when the management request exits prematurely. The accumulation of such un-freed memory causes the driver to fail to allocate anymore memory (FIB) and hence return 0x70000 value to the upper layer, which puts the file system into read only mode. Fix details: The fix makes sure to free the memory (FIB) even if the request exits prematurely hence ensuring the driver wouldn't run out of memory (FIBs). Issue2: False Raid Alert occurs When the Physical Drives and Logical drives are reported as deleted or added, even though there is no change done on the system Root cause: Driver IOCTLs is signaled with EINTR while waiting on response from the lower layers. Returning "EINTR" will never initiate internal retry. Fix details: The issue was fixed by replacing "EINTR" with "ERESTARTSYS" for mid-layer retries. Signed-off-by: Penchala Narasimha Reddy <ServeRAIDDriver@hcl.in> Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2009-12-21 21:09:27 +08:00
return -ERESTARTSYS;
}
spin_unlock_irqrestore(&fibptr->event_lock, flags);
BUG_ON(fibptr->done == 0);
if(unlikely(fibptr->flags & FIB_CONTEXT_FLAG_TIMED_OUT))
return -ETIMEDOUT;
return 0;
}
/*
* If the user does not want a response than return success otherwise
* return pending
*/
if (reply)
return -EINPROGRESS;
else
return 0;
}
/**
* aac_consumer_get - get the top of the queue
* @dev: Adapter
* @q: Queue
* @entry: Return entry
*
* Will return a pointer to the entry on the top of the queue requested that
* we are a consumer of, and return the address of the queue entry. It does
* not change the state of the queue.
*/
int aac_consumer_get(struct aac_dev * dev, struct aac_queue * q, struct aac_entry **entry)
{
u32 index;
int status;
if (le32_to_cpu(*q->headers.producer) == le32_to_cpu(*q->headers.consumer)) {
status = 0;
} else {
/*
* The consumer index must be wrapped if we have reached
* the end of the queue, else we just use the entry
* pointed to by the header index
*/
if (le32_to_cpu(*q->headers.consumer) >= q->entries)
index = 0;
else
index = le32_to_cpu(*q->headers.consumer);
*entry = q->base + index;
status = 1;
}
return(status);
}
/**
* aac_consumer_free - free consumer entry
* @dev: Adapter
* @q: Queue
* @qid: Queue ident
*
* Frees up the current top of the queue we are a consumer of. If the
* queue was full notify the producer that the queue is no longer full.
*/
void aac_consumer_free(struct aac_dev * dev, struct aac_queue *q, u32 qid)
{
int wasfull = 0;
u32 notify;
if ((le32_to_cpu(*q->headers.producer)+1) == le32_to_cpu(*q->headers.consumer))
wasfull = 1;
if (le32_to_cpu(*q->headers.consumer) >= q->entries)
*q->headers.consumer = cpu_to_le32(1);
else
le32_add_cpu(q->headers.consumer, 1);
if (wasfull) {
switch (qid) {
case HostNormCmdQueue:
notify = HostNormCmdNotFull;
break;
case HostNormRespQueue:
notify = HostNormRespNotFull;
break;
default:
BUG();
return;
}
aac_adapter_notify(dev, notify);
}
}
/**
* aac_fib_adapter_complete - complete adapter issued fib
* @fibptr: fib to complete
* @size: size of fib
*
* Will do all necessary work to complete a FIB that was sent from
* the adapter.
*/
int aac_fib_adapter_complete(struct fib *fibptr, unsigned short size)
{
struct hw_fib * hw_fib = fibptr->hw_fib_va;
struct aac_dev * dev = fibptr->dev;
struct aac_queue * q;
unsigned long nointr = 0;
unsigned long qflags;
if (dev->comm_interface == AAC_COMM_MESSAGE_TYPE1 ||
dev->comm_interface == AAC_COMM_MESSAGE_TYPE2) {
kfree(hw_fib);
return 0;
}
if (hw_fib->header.XferState == 0) {
if (dev->comm_interface == AAC_COMM_MESSAGE)
kfree(hw_fib);
return 0;
}
/*
* If we plan to do anything check the structure type first.
*/
if (hw_fib->header.StructType != FIB_MAGIC &&
hw_fib->header.StructType != FIB_MAGIC2 &&
hw_fib->header.StructType != FIB_MAGIC2_64) {
if (dev->comm_interface == AAC_COMM_MESSAGE)
kfree(hw_fib);
return -EINVAL;
}
/*
* This block handles the case where the adapter had sent us a
* command and we have finished processing the command. We
* call completeFib when we are done processing the command
* and want to send a response back to the adapter. This will
* send the completed cdb to the adapter.
*/
if (hw_fib->header.XferState & cpu_to_le32(SentFromAdapter)) {
if (dev->comm_interface == AAC_COMM_MESSAGE) {
kfree (hw_fib);
} else {
u32 index;
hw_fib->header.XferState |= cpu_to_le32(HostProcessed);
if (size) {
size += sizeof(struct aac_fibhdr);
if (size > le16_to_cpu(hw_fib->header.SenderSize))
return -EMSGSIZE;
hw_fib->header.Size = cpu_to_le16(size);
}
q = &dev->queues->queue[AdapNormRespQueue];
spin_lock_irqsave(q->lock, qflags);
aac_queue_get(dev, &index, AdapNormRespQueue, hw_fib, 1, NULL, &nointr);
*(q->headers.producer) = cpu_to_le32(index + 1);
spin_unlock_irqrestore(q->lock, qflags);
if (!(nointr & (int)aac_config.irq_mod))
aac_adapter_notify(dev, AdapNormRespQueue);
}
} else {
printk(KERN_WARNING "aac_fib_adapter_complete: "
"Unknown xferstate detected.\n");
BUG();
}
return 0;
}
/**
* aac_fib_complete - fib completion handler
* @fib: FIB to complete
*
* Will do all necessary work to complete a FIB.
*/
int aac_fib_complete(struct fib *fibptr)
{
struct hw_fib * hw_fib = fibptr->hw_fib_va;
/*
* Check for a fib which has already been completed
*/
if (hw_fib->header.XferState == 0)
return 0;
/*
* If we plan to do anything check the structure type first.
*/
if (hw_fib->header.StructType != FIB_MAGIC &&
hw_fib->header.StructType != FIB_MAGIC2 &&
hw_fib->header.StructType != FIB_MAGIC2_64)
return -EINVAL;
/*
* This block completes a cdb which orginated on the host and we
* just need to deallocate the cdb or reinit it. At this point the
* command is complete that we had sent to the adapter and this
* cdb could be reused.
*/
[SCSI] aacraid: fix File System going into read-only mode These particular problems were reported by Cisco and SAP and customers as well. Cisco reported on RHEL4 U6 and SAP reported on SLES9 SP4 and SLES10 SP2. We added these fixes on RHEL4 U6 and gave a private build to IBM and Cisco. Cisco and IBM tested it for more than 15 days and they reported that they did not see the issue so far. Before the fix, Cisco used to see the issue within 5 days. We generated a patch for SLES9 SP4 and SLES10 SP2 and submitted to Novell. Novell applied the patch and gave a test build to SAP. SAP tested and reported that the build is working properly. We also tested in our lab using the tools "dishogsync", which is IO stress tool and the tool was provided by Cisco. Issue1: File System going into read-only mode Root cause: The driver tends to not free the memory (FIB) when the management request exits prematurely. The accumulation of such un-freed memory causes the driver to fail to allocate anymore memory (FIB) and hence return 0x70000 value to the upper layer, which puts the file system into read only mode. Fix details: The fix makes sure to free the memory (FIB) even if the request exits prematurely hence ensuring the driver wouldn't run out of memory (FIBs). Issue2: False Raid Alert occurs When the Physical Drives and Logical drives are reported as deleted or added, even though there is no change done on the system Root cause: Driver IOCTLs is signaled with EINTR while waiting on response from the lower layers. Returning "EINTR" will never initiate internal retry. Fix details: The issue was fixed by replacing "EINTR" with "ERESTARTSYS" for mid-layer retries. Signed-off-by: Penchala Narasimha Reddy <ServeRAIDDriver@hcl.in> Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2009-12-21 21:09:27 +08:00
if((hw_fib->header.XferState & cpu_to_le32(SentFromHost)) &&
(hw_fib->header.XferState & cpu_to_le32(AdapterProcessed)))
{
fib_dealloc(fibptr);
}
else if(hw_fib->header.XferState & cpu_to_le32(SentFromHost))
{
/*
* This handles the case when the host has aborted the I/O
* to the adapter because the adapter is not responding
*/
fib_dealloc(fibptr);
} else if(hw_fib->header.XferState & cpu_to_le32(HostOwned)) {
fib_dealloc(fibptr);
} else {
BUG();
}
return 0;
}
/**
* aac_printf - handle printf from firmware
* @dev: Adapter
* @val: Message info
*
* Print a message passed to us by the controller firmware on the
* Adaptec board
*/
void aac_printf(struct aac_dev *dev, u32 val)
{
char *cp = dev->printfbuf;
if (dev->printf_enabled)
{
int length = val & 0xffff;
int level = (val >> 16) & 0xffff;
/*
* The size of the printfbuf is set in port.c
* There is no variable or define for it
*/
if (length > 255)
length = 255;
if (cp[length] != 0)
cp[length] = 0;
if (level == LOG_AAC_HIGH_ERROR)
printk(KERN_WARNING "%s:%s", dev->name, cp);
else
printk(KERN_INFO "%s:%s", dev->name, cp);
}
memset(cp, 0, 256);
}
/**
* aac_handle_aif - Handle a message from the firmware
* @dev: Which adapter this fib is from
* @fibptr: Pointer to fibptr from adapter
*
* This routine handles a driver notify fib from the adapter and
* dispatches it to the appropriate routine for handling.
*/
#define AIF_SNIFF_TIMEOUT (500*HZ)
static void aac_handle_aif(struct aac_dev * dev, struct fib * fibptr)
{
struct hw_fib * hw_fib = fibptr->hw_fib_va;
struct aac_aifcmd * aifcmd = (struct aac_aifcmd *)hw_fib->data;
u32 channel, id, lun, container;
struct scsi_device *device;
enum {
NOTHING,
DELETE,
ADD,
CHANGE
} device_config_needed = NOTHING;
/* Sniff for container changes */
if (!dev || !dev->fsa_dev)
return;
container = channel = id = lun = (u32)-1;
/*
* We have set this up to try and minimize the number of
* re-configures that take place. As a result of this when
* certain AIF's come in we will set a flag waiting for another
* type of AIF before setting the re-config flag.
*/
switch (le32_to_cpu(aifcmd->command)) {
case AifCmdDriverNotify:
switch (le32_to_cpu(((__le32 *)aifcmd->data)[0])) {
case AifRawDeviceRemove:
container = le32_to_cpu(((__le32 *)aifcmd->data)[1]);
if ((container >> 28)) {
container = (u32)-1;
break;
}
channel = (container >> 24) & 0xF;
if (channel >= dev->maximum_num_channels) {
container = (u32)-1;
break;
}
id = container & 0xFFFF;
if (id >= dev->maximum_num_physicals) {
container = (u32)-1;
break;
}
lun = (container >> 16) & 0xFF;
container = (u32)-1;
channel = aac_phys_to_logical(channel);
device_config_needed =
(((__le32 *)aifcmd->data)[0] ==
cpu_to_le32(AifRawDeviceRemove)) ? DELETE : ADD;
if (device_config_needed == ADD) {
device = scsi_device_lookup(
dev->scsi_host_ptr,
channel, id, lun);
if (device) {
scsi_remove_device(device);
scsi_device_put(device);
}
}
break;
/*
* Morph or Expand complete
*/
case AifDenMorphComplete:
case AifDenVolumeExtendComplete:
container = le32_to_cpu(((__le32 *)aifcmd->data)[1]);
if (container >= dev->maximum_num_containers)
break;
/*
* Find the scsi_device associated with the SCSI
* address. Make sure we have the right array, and if
* so set the flag to initiate a new re-config once we
* see an AifEnConfigChange AIF come through.
*/
if ((dev != NULL) && (dev->scsi_host_ptr != NULL)) {
device = scsi_device_lookup(dev->scsi_host_ptr,
CONTAINER_TO_CHANNEL(container),
CONTAINER_TO_ID(container),
CONTAINER_TO_LUN(container));
if (device) {
dev->fsa_dev[container].config_needed = CHANGE;
dev->fsa_dev[container].config_waiting_on = AifEnConfigChange;
dev->fsa_dev[container].config_waiting_stamp = jiffies;
scsi_device_put(device);
}
}
}
/*
* If we are waiting on something and this happens to be
* that thing then set the re-configure flag.
*/
if (container != (u32)-1) {
if (container >= dev->maximum_num_containers)
break;
if ((dev->fsa_dev[container].config_waiting_on ==
le32_to_cpu(*(__le32 *)aifcmd->data)) &&
time_before(jiffies, dev->fsa_dev[container].config_waiting_stamp + AIF_SNIFF_TIMEOUT))
dev->fsa_dev[container].config_waiting_on = 0;
} else for (container = 0;
container < dev->maximum_num_containers; ++container) {
if ((dev->fsa_dev[container].config_waiting_on ==
le32_to_cpu(*(__le32 *)aifcmd->data)) &&
time_before(jiffies, dev->fsa_dev[container].config_waiting_stamp + AIF_SNIFF_TIMEOUT))
dev->fsa_dev[container].config_waiting_on = 0;
}
break;
case AifCmdEventNotify:
switch (le32_to_cpu(((__le32 *)aifcmd->data)[0])) {
case AifEnBatteryEvent:
dev->cache_protected =
(((__le32 *)aifcmd->data)[1] == cpu_to_le32(3));
break;
/*
* Add an Array.
*/
case AifEnAddContainer:
container = le32_to_cpu(((__le32 *)aifcmd->data)[1]);
if (container >= dev->maximum_num_containers)
break;
dev->fsa_dev[container].config_needed = ADD;
dev->fsa_dev[container].config_waiting_on =
AifEnConfigChange;
dev->fsa_dev[container].config_waiting_stamp = jiffies;
break;
/*
* Delete an Array.
*/
case AifEnDeleteContainer:
container = le32_to_cpu(((__le32 *)aifcmd->data)[1]);
if (container >= dev->maximum_num_containers)
break;
dev->fsa_dev[container].config_needed = DELETE;
dev->fsa_dev[container].config_waiting_on =
AifEnConfigChange;
dev->fsa_dev[container].config_waiting_stamp = jiffies;
break;
/*
* Container change detected. If we currently are not
* waiting on something else, setup to wait on a Config Change.
*/
case AifEnContainerChange:
container = le32_to_cpu(((__le32 *)aifcmd->data)[1]);
if (container >= dev->maximum_num_containers)
break;
if (dev->fsa_dev[container].config_waiting_on &&
time_before(jiffies, dev->fsa_dev[container].config_waiting_stamp + AIF_SNIFF_TIMEOUT))
break;
dev->fsa_dev[container].config_needed = CHANGE;
dev->fsa_dev[container].config_waiting_on =
AifEnConfigChange;
dev->fsa_dev[container].config_waiting_stamp = jiffies;
break;
case AifEnConfigChange:
break;
case AifEnAddJBOD:
case AifEnDeleteJBOD:
container = le32_to_cpu(((__le32 *)aifcmd->data)[1]);
if ((container >> 28)) {
container = (u32)-1;
break;
}
channel = (container >> 24) & 0xF;
if (channel >= dev->maximum_num_channels) {
container = (u32)-1;
break;
}
id = container & 0xFFFF;
if (id >= dev->maximum_num_physicals) {
container = (u32)-1;
break;
}
lun = (container >> 16) & 0xFF;
container = (u32)-1;
channel = aac_phys_to_logical(channel);
device_config_needed =
(((__le32 *)aifcmd->data)[0] ==
cpu_to_le32(AifEnAddJBOD)) ? ADD : DELETE;
if (device_config_needed == ADD) {
device = scsi_device_lookup(dev->scsi_host_ptr,
channel,
id,
lun);
if (device) {
scsi_remove_device(device);
scsi_device_put(device);
}
}
break;
case AifEnEnclosureManagement:
/*
* If in JBOD mode, automatic exposure of new
* physical target to be suppressed until configured.
*/
if (dev->jbod)
break;
switch (le32_to_cpu(((__le32 *)aifcmd->data)[3])) {
case EM_DRIVE_INSERTION:
case EM_DRIVE_REMOVAL:
case EM_SES_DRIVE_INSERTION:
case EM_SES_DRIVE_REMOVAL:
container = le32_to_cpu(
((__le32 *)aifcmd->data)[2]);
if ((container >> 28)) {
container = (u32)-1;
break;
}
channel = (container >> 24) & 0xF;
if (channel >= dev->maximum_num_channels) {
container = (u32)-1;
break;
}
id = container & 0xFFFF;
lun = (container >> 16) & 0xFF;
container = (u32)-1;
if (id >= dev->maximum_num_physicals) {
/* legacy dev_t ? */
if ((0x2000 <= id) || lun || channel ||
((channel = (id >> 7) & 0x3F) >=
dev->maximum_num_channels))
break;
lun = (id >> 4) & 7;
id &= 0xF;
}
channel = aac_phys_to_logical(channel);
device_config_needed =
((((__le32 *)aifcmd->data)[3]
== cpu_to_le32(EM_DRIVE_INSERTION)) ||
(((__le32 *)aifcmd->data)[3]
== cpu_to_le32(EM_SES_DRIVE_INSERTION))) ?
ADD : DELETE;
break;
}
break;
}
/*
* If we are waiting on something and this happens to be
* that thing then set the re-configure flag.
*/
if (container != (u32)-1) {
if (container >= dev->maximum_num_containers)
break;
if ((dev->fsa_dev[container].config_waiting_on ==
le32_to_cpu(*(__le32 *)aifcmd->data)) &&
time_before(jiffies, dev->fsa_dev[container].config_waiting_stamp + AIF_SNIFF_TIMEOUT))
dev->fsa_dev[container].config_waiting_on = 0;
} else for (container = 0;
container < dev->maximum_num_containers; ++container) {
if ((dev->fsa_dev[container].config_waiting_on ==
le32_to_cpu(*(__le32 *)aifcmd->data)) &&
time_before(jiffies, dev->fsa_dev[container].config_waiting_stamp + AIF_SNIFF_TIMEOUT))
dev->fsa_dev[container].config_waiting_on = 0;
}
break;
case AifCmdJobProgress:
/*
* These are job progress AIF's. When a Clear is being
* done on a container it is initially created then hidden from
* the OS. When the clear completes we don't get a config
* change so we monitor the job status complete on a clear then
* wait for a container change.
*/
if (((__le32 *)aifcmd->data)[1] == cpu_to_le32(AifJobCtrZero) &&
(((__le32 *)aifcmd->data)[6] == ((__le32 *)aifcmd->data)[5] ||
((__le32 *)aifcmd->data)[4] == cpu_to_le32(AifJobStsSuccess))) {
for (container = 0;
container < dev->maximum_num_containers;
++container) {
/*
* Stomp on all config sequencing for all
* containers?
*/
dev->fsa_dev[container].config_waiting_on =
AifEnContainerChange;
dev->fsa_dev[container].config_needed = ADD;
dev->fsa_dev[container].config_waiting_stamp =
jiffies;
}
}
if (((__le32 *)aifcmd->data)[1] == cpu_to_le32(AifJobCtrZero) &&
((__le32 *)aifcmd->data)[6] == 0 &&
((__le32 *)aifcmd->data)[4] == cpu_to_le32(AifJobStsRunning)) {
for (container = 0;
container < dev->maximum_num_containers;
++container) {
/*
* Stomp on all config sequencing for all
* containers?
*/
dev->fsa_dev[container].config_waiting_on =
AifEnContainerChange;
dev->fsa_dev[container].config_needed = DELETE;
dev->fsa_dev[container].config_waiting_stamp =
jiffies;
}
}
break;
}
container = 0;
retry_next:
if (device_config_needed == NOTHING)
for (; container < dev->maximum_num_containers; ++container) {
if ((dev->fsa_dev[container].config_waiting_on == 0) &&
(dev->fsa_dev[container].config_needed != NOTHING) &&
time_before(jiffies, dev->fsa_dev[container].config_waiting_stamp + AIF_SNIFF_TIMEOUT)) {
device_config_needed =
dev->fsa_dev[container].config_needed;
dev->fsa_dev[container].config_needed = NOTHING;
channel = CONTAINER_TO_CHANNEL(container);
id = CONTAINER_TO_ID(container);
lun = CONTAINER_TO_LUN(container);
break;
}
}
if (device_config_needed == NOTHING)
return;
/*
* If we decided that a re-configuration needs to be done,
* schedule it here on the way out the door, please close the door
* behind you.
*/
/*
* Find the scsi_device associated with the SCSI address,
* and mark it as changed, invalidating the cache. This deals
* with changes to existing device IDs.
*/
if (!dev || !dev->scsi_host_ptr)
return;
/*
* force reload of disk info via aac_probe_container
*/
if ((channel == CONTAINER_CHANNEL) &&
(device_config_needed != NOTHING)) {
if (dev->fsa_dev[container].valid == 1)
dev->fsa_dev[container].valid = 2;
aac_probe_container(dev, container);
}
device = scsi_device_lookup(dev->scsi_host_ptr, channel, id, lun);
if (device) {
switch (device_config_needed) {
case DELETE:
#if (defined(AAC_DEBUG_INSTRUMENT_AIF_DELETE))
scsi_remove_device(device);
#else
if (scsi_device_online(device)) {
scsi_device_set_state(device, SDEV_OFFLINE);
sdev_printk(KERN_INFO, device,
"Device offlined - %s\n",
(channel == CONTAINER_CHANNEL) ?
"array deleted" :
"enclosure services event");
}
#endif
break;
case ADD:
if (!scsi_device_online(device)) {
sdev_printk(KERN_INFO, device,
"Device online - %s\n",
(channel == CONTAINER_CHANNEL) ?
"array created" :
"enclosure services event");
scsi_device_set_state(device, SDEV_RUNNING);
}
/* FALLTHRU */
case CHANGE:
if ((channel == CONTAINER_CHANNEL)
&& (!dev->fsa_dev[container].valid)) {
#if (defined(AAC_DEBUG_INSTRUMENT_AIF_DELETE))
scsi_remove_device(device);
#else
if (!scsi_device_online(device))
break;
scsi_device_set_state(device, SDEV_OFFLINE);
sdev_printk(KERN_INFO, device,
"Device offlined - %s\n",
"array failed");
#endif
break;
}
scsi_rescan_device(&device->sdev_gendev);
default:
break;
}
scsi_device_put(device);
device_config_needed = NOTHING;
}
if (device_config_needed == ADD)
scsi_add_device(dev->scsi_host_ptr, channel, id, lun);
if (channel == CONTAINER_CHANNEL) {
container++;
device_config_needed = NOTHING;
goto retry_next;
}
}
static int _aac_reset_adapter(struct aac_dev *aac, int forced)
{
int index, quirks;
int retval;
struct Scsi_Host *host;
struct scsi_device *dev;
struct scsi_cmnd *command;
struct scsi_cmnd *command_list;
int jafo = 0;
/*
* Assumptions:
* - host is locked, unless called by the aacraid thread.
* (a matter of convenience, due to legacy issues surrounding
* eh_host_adapter_reset).
* - in_reset is asserted, so no new i/o is getting to the
* card.
* - The card is dead, or will be very shortly ;-/ so no new
* commands are completing in the interrupt service.
*/
host = aac->scsi_host_ptr;
scsi_block_requests(host);
aac_adapter_disable_int(aac);
if (aac->thread->pid != current->pid) {
spin_unlock_irq(host->host_lock);
kthread_stop(aac->thread);
jafo = 1;
}
/*
* If a positive health, means in a known DEAD PANIC
* state and the adapter could be reset to `try again'.
*/
retval = aac_adapter_restart(aac, forced ? 0 : aac_adapter_check_health(aac));
if (retval)
goto out;
/*
* Loop through the fibs, close the synchronous FIBS
*/
for (retval = 1, index = 0; index < (aac->scsi_host_ptr->can_queue + AAC_NUM_MGT_FIB); index++) {
struct fib *fib = &aac->fibs[index];
if (!(fib->hw_fib_va->header.XferState & cpu_to_le32(NoResponseExpected | Async)) &&
(fib->hw_fib_va->header.XferState & cpu_to_le32(ResponseExpected))) {
unsigned long flagv;
spin_lock_irqsave(&fib->event_lock, flagv);
up(&fib->event_wait);
spin_unlock_irqrestore(&fib->event_lock, flagv);
schedule();
retval = 0;
}
}
/* Give some extra time for ioctls to complete. */
if (retval == 0)
ssleep(2);
index = aac->cardtype;
/*
* Re-initialize the adapter, first free resources, then carefully
* apply the initialization sequence to come back again. Only risk
* is a change in Firmware dropping cache, it is assumed the caller
* will ensure that i/o is queisced and the card is flushed in that
* case.
*/
aac_fib_map_free(aac);
pci_free_consistent(aac->pdev, aac->comm_size, aac->comm_addr, aac->comm_phys);
aac->comm_addr = NULL;
aac->comm_phys = 0;
kfree(aac->queues);
aac->queues = NULL;
aac_free_irq(aac);
kfree(aac->fsa_dev);
aac->fsa_dev = NULL;
[SCSI] aacraid: fix driver failure with Dell PowerEdge Expandable RAID Controller 3/Di As reported in http://bugzilla.kernel.org/show_bug.cgi?id=3D9133 it was discovered that the PERC line of controllers lacked a key 64 bit ScatterGather capable SCSI pass-through function. The adapters are still capable of 64 bit ScatterGather I/O commands, but these two can not be mixed. This problem was exacerbated by the introduction of the SCSI Generic access to the DASD physical devices. The fix for users before this patch is applied is aacraid.dacmode=3D0 on the kernel command line to disable 64 bit I/O. The enclosed patch introduces a new adapter quirk and tries to limp along by enabling pass-through in situations where memory is 32 bit addressable on 64 bit machines, or disable the pass-through functions altogether. I expect that the check for 32 bit addressable memory to be controversial in that it can be incorrect in non-Dell non-Intel systems that PERC would never be installed under, the alternative is to disable pass-through in all cases which could be reported as another regression. Pass-through is used for SCSI Generic access to the physical devices, or for the management applications to properly function. In systems where this patch has disabled pass-through because it is unsupportable in combination with I/O performance, the user can choose to enable pass-through by turning off dacmode (aacraid.dacmode=3D0) or limiting the discovered kernel memory (mem=3D4G) with an associated loss in runtime performance. If we chose instead to turn off 64 bit dacmode for the adapters with this quirk, then this would be reported as another regression. Signed-off-by: Mark Salyzyn <aacraid@adaptec.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
2007-12-14 08:14:18 +08:00
quirks = aac_get_driver_ident(index)->quirks;
if (quirks & AAC_QUIRK_31BIT) {
if (((retval = pci_set_dma_mask(aac->pdev, DMA_BIT_MASK(31)))) ||
((retval = pci_set_consistent_dma_mask(aac->pdev, DMA_BIT_MASK(31)))))
goto out;
} else {
if (((retval = pci_set_dma_mask(aac->pdev, DMA_BIT_MASK(32)))) ||
((retval = pci_set_consistent_dma_mask(aac->pdev, DMA_BIT_MASK(32)))))
goto out;
}
if ((retval = (*(aac_get_driver_ident(index)->init))(aac)))
goto out;
[SCSI] aacraid: fix driver failure with Dell PowerEdge Expandable RAID Controller 3/Di As reported in http://bugzilla.kernel.org/show_bug.cgi?id=3D9133 it was discovered that the PERC line of controllers lacked a key 64 bit ScatterGather capable SCSI pass-through function. The adapters are still capable of 64 bit ScatterGather I/O commands, but these two can not be mixed. This problem was exacerbated by the introduction of the SCSI Generic access to the DASD physical devices. The fix for users before this patch is applied is aacraid.dacmode=3D0 on the kernel command line to disable 64 bit I/O. The enclosed patch introduces a new adapter quirk and tries to limp along by enabling pass-through in situations where memory is 32 bit addressable on 64 bit machines, or disable the pass-through functions altogether. I expect that the check for 32 bit addressable memory to be controversial in that it can be incorrect in non-Dell non-Intel systems that PERC would never be installed under, the alternative is to disable pass-through in all cases which could be reported as another regression. Pass-through is used for SCSI Generic access to the physical devices, or for the management applications to properly function. In systems where this patch has disabled pass-through because it is unsupportable in combination with I/O performance, the user can choose to enable pass-through by turning off dacmode (aacraid.dacmode=3D0) or limiting the discovered kernel memory (mem=3D4G) with an associated loss in runtime performance. If we chose instead to turn off 64 bit dacmode for the adapters with this quirk, then this would be reported as another regression. Signed-off-by: Mark Salyzyn <aacraid@adaptec.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
2007-12-14 08:14:18 +08:00
if (quirks & AAC_QUIRK_31BIT)
if ((retval = pci_set_dma_mask(aac->pdev, DMA_BIT_MASK(32))))
goto out;
if (jafo) {
aac->thread = kthread_run(aac_command_thread, aac, "%s",
aac->name);
if (IS_ERR(aac->thread)) {
retval = PTR_ERR(aac->thread);
goto out;
}
}
(void)aac_get_adapter_info(aac);
if ((quirks & AAC_QUIRK_34SG) && (host->sg_tablesize > 34)) {
host->sg_tablesize = 34;
host->max_sectors = (host->sg_tablesize * 8) + 112;
}
if ((quirks & AAC_QUIRK_17SG) && (host->sg_tablesize > 17)) {
host->sg_tablesize = 17;
host->max_sectors = (host->sg_tablesize * 8) + 112;
}
aac_get_config_status(aac, 1);
aac_get_containers(aac);
/*
* This is where the assumption that the Adapter is quiesced
* is important.
*/
command_list = NULL;
__shost_for_each_device(dev, host) {
unsigned long flags;
spin_lock_irqsave(&dev->list_lock, flags);
list_for_each_entry(command, &dev->cmd_list, list)
if (command->SCp.phase == AAC_OWNER_FIRMWARE) {
command->SCp.buffer = (struct scatterlist *)command_list;
command_list = command;
}
spin_unlock_irqrestore(&dev->list_lock, flags);
}
while ((command = command_list)) {
command_list = (struct scsi_cmnd *)command->SCp.buffer;
command->SCp.buffer = NULL;
command->result = DID_OK << 16
| COMMAND_COMPLETE << 8
| SAM_STAT_TASK_SET_FULL;
command->SCp.phase = AAC_OWNER_ERROR_HANDLER;
command->scsi_done(command);
}
retval = 0;
out:
aac->in_reset = 0;
scsi_unblock_requests(host);
if (jafo) {
spin_lock_irq(host->host_lock);
}
return retval;
}
int aac_reset_adapter(struct aac_dev * aac, int forced)
{
unsigned long flagv = 0;
int retval;
struct Scsi_Host * host;
if (spin_trylock_irqsave(&aac->fib_lock, flagv) == 0)
return -EBUSY;
if (aac->in_reset) {
spin_unlock_irqrestore(&aac->fib_lock, flagv);
return -EBUSY;
}
aac->in_reset = 1;
spin_unlock_irqrestore(&aac->fib_lock, flagv);
/*
* Wait for all commands to complete to this specific
* target (block maximum 60 seconds). Although not necessary,
* it does make us a good storage citizen.
*/
host = aac->scsi_host_ptr;
scsi_block_requests(host);
if (forced < 2) for (retval = 60; retval; --retval) {
struct scsi_device * dev;
struct scsi_cmnd * command;
int active = 0;
__shost_for_each_device(dev, host) {
spin_lock_irqsave(&dev->list_lock, flagv);
list_for_each_entry(command, &dev->cmd_list, list) {
if (command->SCp.phase == AAC_OWNER_FIRMWARE) {
active++;
break;
}
}
spin_unlock_irqrestore(&dev->list_lock, flagv);
if (active)
break;
}
/*
* We can exit If all the commands are complete
*/
if (active == 0)
break;
ssleep(1);
}
/* Quiesce build, flush cache, write through mode */
if (forced < 2)
aac_send_shutdown(aac);
spin_lock_irqsave(host->host_lock, flagv);
retval = _aac_reset_adapter(aac, forced ? forced : ((aac_check_reset != 0) && (aac_check_reset != 1)));
spin_unlock_irqrestore(host->host_lock, flagv);
if ((forced < 2) && (retval == -ENODEV)) {
/* Unwind aac_send_shutdown() IOP_RESET unsupported/disabled */
struct fib * fibctx = aac_fib_alloc(aac);
if (fibctx) {
struct aac_pause *cmd;
int status;
aac_fib_init(fibctx);
cmd = (struct aac_pause *) fib_data(fibctx);
cmd->command = cpu_to_le32(VM_ContainerConfig);
cmd->type = cpu_to_le32(CT_PAUSE_IO);
cmd->timeout = cpu_to_le32(1);
cmd->min = cpu_to_le32(1);
cmd->noRescan = cpu_to_le32(1);
cmd->count = cpu_to_le32(0);
status = aac_fib_send(ContainerCommand,
fibctx,
sizeof(struct aac_pause),
FsaNormal,
-2 /* Timeout silently */, 1,
NULL, NULL);
if (status >= 0)
aac_fib_complete(fibctx);
[SCSI] aacraid: fix File System going into read-only mode These particular problems were reported by Cisco and SAP and customers as well. Cisco reported on RHEL4 U6 and SAP reported on SLES9 SP4 and SLES10 SP2. We added these fixes on RHEL4 U6 and gave a private build to IBM and Cisco. Cisco and IBM tested it for more than 15 days and they reported that they did not see the issue so far. Before the fix, Cisco used to see the issue within 5 days. We generated a patch for SLES9 SP4 and SLES10 SP2 and submitted to Novell. Novell applied the patch and gave a test build to SAP. SAP tested and reported that the build is working properly. We also tested in our lab using the tools "dishogsync", which is IO stress tool and the tool was provided by Cisco. Issue1: File System going into read-only mode Root cause: The driver tends to not free the memory (FIB) when the management request exits prematurely. The accumulation of such un-freed memory causes the driver to fail to allocate anymore memory (FIB) and hence return 0x70000 value to the upper layer, which puts the file system into read only mode. Fix details: The fix makes sure to free the memory (FIB) even if the request exits prematurely hence ensuring the driver wouldn't run out of memory (FIBs). Issue2: False Raid Alert occurs When the Physical Drives and Logical drives are reported as deleted or added, even though there is no change done on the system Root cause: Driver IOCTLs is signaled with EINTR while waiting on response from the lower layers. Returning "EINTR" will never initiate internal retry. Fix details: The issue was fixed by replacing "EINTR" with "ERESTARTSYS" for mid-layer retries. Signed-off-by: Penchala Narasimha Reddy <ServeRAIDDriver@hcl.in> Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2009-12-21 21:09:27 +08:00
/* FIB should be freed only after getting
* the response from the F/W */
if (status != -ERESTARTSYS)
aac_fib_free(fibctx);
}
}
return retval;
}
int aac_check_health(struct aac_dev * aac)
{
int BlinkLED;
unsigned long time_now, flagv = 0;
struct list_head * entry;
struct Scsi_Host * host;
/* Extending the scope of fib_lock slightly to protect aac->in_reset */
if (spin_trylock_irqsave(&aac->fib_lock, flagv) == 0)
return 0;
if (aac->in_reset || !(BlinkLED = aac_adapter_check_health(aac))) {
spin_unlock_irqrestore(&aac->fib_lock, flagv);
return 0; /* OK */
}
aac->in_reset = 1;
/* Fake up an AIF:
* aac_aifcmd.command = AifCmdEventNotify = 1
* aac_aifcmd.seqnum = 0xFFFFFFFF
* aac_aifcmd.data[0] = AifEnExpEvent = 23
* aac_aifcmd.data[1] = AifExeFirmwarePanic = 3
* aac.aifcmd.data[2] = AifHighPriority = 3
* aac.aifcmd.data[3] = BlinkLED
*/
time_now = jiffies/HZ;
entry = aac->fib_list.next;
/*
* For each Context that is on the
* fibctxList, make a copy of the
* fib, and then set the event to wake up the
* thread that is waiting for it.
*/
while (entry != &aac->fib_list) {
/*
* Extract the fibctx
*/
struct aac_fib_context *fibctx = list_entry(entry, struct aac_fib_context, next);
struct hw_fib * hw_fib;
struct fib * fib;
/*
* Check if the queue is getting
* backlogged
*/
if (fibctx->count > 20) {
/*
* It's *not* jiffies folks,
* but jiffies / HZ, so do not
* panic ...
*/
u32 time_last = fibctx->jiffies;
/*
* Has it been > 2 minutes
* since the last read off
* the queue?
*/
if ((time_now - time_last) > aif_timeout) {
entry = entry->next;
aac_close_fib_context(aac, fibctx);
continue;
}
}
/*
* Warning: no sleep allowed while
* holding spinlock
*/
hw_fib = kzalloc(sizeof(struct hw_fib), GFP_ATOMIC);
fib = kzalloc(sizeof(struct fib), GFP_ATOMIC);
if (fib && hw_fib) {
struct aac_aifcmd * aif;
fib->hw_fib_va = hw_fib;
fib->dev = aac;
aac_fib_init(fib);
fib->type = FSAFS_NTC_FIB_CONTEXT;
fib->size = sizeof (struct fib);
fib->data = hw_fib->data;
aif = (struct aac_aifcmd *)hw_fib->data;
aif->command = cpu_to_le32(AifCmdEventNotify);
aif->seqnum = cpu_to_le32(0xFFFFFFFF);
((__le32 *)aif->data)[0] = cpu_to_le32(AifEnExpEvent);
((__le32 *)aif->data)[1] = cpu_to_le32(AifExeFirmwarePanic);
((__le32 *)aif->data)[2] = cpu_to_le32(AifHighPriority);
((__le32 *)aif->data)[3] = cpu_to_le32(BlinkLED);
/*
* Put the FIB onto the
* fibctx's fibs
*/
list_add_tail(&fib->fiblink, &fibctx->fib_list);
fibctx->count++;
/*
* Set the event to wake up the
* thread that will waiting.
*/
up(&fibctx->wait_sem);
} else {
printk(KERN_WARNING "aifd: didn't allocate NewFib.\n");
kfree(fib);
kfree(hw_fib);
}
entry = entry->next;
}
spin_unlock_irqrestore(&aac->fib_lock, flagv);
if (BlinkLED < 0) {
printk(KERN_ERR "%s: Host adapter dead %d\n", aac->name, BlinkLED);
goto out;
}
printk(KERN_ERR "%s: Host adapter BLINK LED 0x%x\n", aac->name, BlinkLED);
if (!aac_check_reset || ((aac_check_reset == 1) &&
(aac->supplement_adapter_info.SupportedOptions2 &
AAC_OPTION_IGNORE_RESET)))
goto out;
host = aac->scsi_host_ptr;
if (aac->thread->pid != current->pid)
spin_lock_irqsave(host->host_lock, flagv);
BlinkLED = _aac_reset_adapter(aac, aac_check_reset != 1);
if (aac->thread->pid != current->pid)
spin_unlock_irqrestore(host->host_lock, flagv);
return BlinkLED;
out:
aac->in_reset = 0;
return BlinkLED;
}
/**
* aac_command_thread - command processing thread
* @dev: Adapter to monitor
*
* Waits on the commandready event in it's queue. When the event gets set
* it will pull FIBs off it's queue. It will continue to pull FIBs off
* until the queue is empty. When the queue is empty it will wait for
* more FIBs.
*/
int aac_command_thread(void *data)
{
struct aac_dev *dev = data;
struct hw_fib *hw_fib, *hw_newfib;
struct fib *fib, *newfib;
struct aac_fib_context *fibctx;
unsigned long flags;
DECLARE_WAITQUEUE(wait, current);
unsigned long next_jiffies = jiffies + HZ;
unsigned long next_check_jiffies = next_jiffies;
long difference = HZ;
/*
* We can only have one thread per adapter for AIF's.
*/
if (dev->aif_thread)
return -EINVAL;
/*
* Let the DPC know it has a place to send the AIF's to.
*/
dev->aif_thread = 1;
add_wait_queue(&dev->queues->queue[HostNormCmdQueue].cmdready, &wait);
set_current_state(TASK_INTERRUPTIBLE);
dprintk ((KERN_INFO "aac_command_thread start\n"));
while (1) {
spin_lock_irqsave(dev->queues->queue[HostNormCmdQueue].lock, flags);
while(!list_empty(&(dev->queues->queue[HostNormCmdQueue].cmdq))) {
struct list_head *entry;
struct aac_aifcmd * aifcmd;
set_current_state(TASK_RUNNING);
entry = dev->queues->queue[HostNormCmdQueue].cmdq.next;
list_del(entry);
spin_unlock_irqrestore(dev->queues->queue[HostNormCmdQueue].lock, flags);
fib = list_entry(entry, struct fib, fiblink);
/*
* We will process the FIB here or pass it to a
* worker thread that is TBD. We Really can't
* do anything at this point since we don't have
* anything defined for this thread to do.
*/
hw_fib = fib->hw_fib_va;
memset(fib, 0, sizeof(struct fib));
fib->type = FSAFS_NTC_FIB_CONTEXT;
fib->size = sizeof(struct fib);
fib->hw_fib_va = hw_fib;
fib->data = hw_fib->data;
fib->dev = dev;
/*
* We only handle AifRequest fibs from the adapter.
*/
aifcmd = (struct aac_aifcmd *) hw_fib->data;
if (aifcmd->command == cpu_to_le32(AifCmdDriverNotify)) {
/* Handle Driver Notify Events */
aac_handle_aif(dev, fib);
*(__le32 *)hw_fib->data = cpu_to_le32(ST_OK);
aac_fib_adapter_complete(fib, (u16)sizeof(u32));
} else {
/* The u32 here is important and intended. We are using
32bit wrapping time to fit the adapter field */
u32 time_now, time_last;
unsigned long flagv;
unsigned num;
struct hw_fib ** hw_fib_pool, ** hw_fib_p;
struct fib ** fib_pool, ** fib_p;
/* Sniff events */
if ((aifcmd->command ==
cpu_to_le32(AifCmdEventNotify)) ||
(aifcmd->command ==
cpu_to_le32(AifCmdJobProgress))) {
aac_handle_aif(dev, fib);
}
time_now = jiffies/HZ;
/*
* Warning: no sleep allowed while
* holding spinlock. We take the estimate
* and pre-allocate a set of fibs outside the
* lock.
*/
num = le32_to_cpu(dev->init->AdapterFibsSize)
/ sizeof(struct hw_fib); /* some extra */
spin_lock_irqsave(&dev->fib_lock, flagv);
entry = dev->fib_list.next;
while (entry != &dev->fib_list) {
entry = entry->next;
++num;
}
spin_unlock_irqrestore(&dev->fib_lock, flagv);
hw_fib_pool = NULL;
fib_pool = NULL;
if (num
&& ((hw_fib_pool = kmalloc(sizeof(struct hw_fib *) * num, GFP_KERNEL)))
&& ((fib_pool = kmalloc(sizeof(struct fib *) * num, GFP_KERNEL)))) {
hw_fib_p = hw_fib_pool;
fib_p = fib_pool;
while (hw_fib_p < &hw_fib_pool[num]) {
if (!(*(hw_fib_p++) = kmalloc(sizeof(struct hw_fib), GFP_KERNEL))) {
--hw_fib_p;
break;
}
if (!(*(fib_p++) = kmalloc(sizeof(struct fib), GFP_KERNEL))) {
kfree(*(--hw_fib_p));
break;
}
}
if ((num = hw_fib_p - hw_fib_pool) == 0) {
kfree(fib_pool);
fib_pool = NULL;
kfree(hw_fib_pool);
hw_fib_pool = NULL;
}
} else {
kfree(hw_fib_pool);
hw_fib_pool = NULL;
}
spin_lock_irqsave(&dev->fib_lock, flagv);
entry = dev->fib_list.next;
/*
* For each Context that is on the
* fibctxList, make a copy of the
* fib, and then set the event to wake up the
* thread that is waiting for it.
*/
hw_fib_p = hw_fib_pool;
fib_p = fib_pool;
while (entry != &dev->fib_list) {
/*
* Extract the fibctx
*/
fibctx = list_entry(entry, struct aac_fib_context, next);
/*
* Check if the queue is getting
* backlogged
*/
if (fibctx->count > 20)
{
/*
* It's *not* jiffies folks,
* but jiffies / HZ so do not
* panic ...
*/
time_last = fibctx->jiffies;
/*
* Has it been > 2 minutes
* since the last read off
* the queue?
*/
if ((time_now - time_last) > aif_timeout) {
entry = entry->next;
aac_close_fib_context(dev, fibctx);
continue;
}
}
/*
* Warning: no sleep allowed while
* holding spinlock
*/
if (hw_fib_p < &hw_fib_pool[num]) {
hw_newfib = *hw_fib_p;
*(hw_fib_p++) = NULL;
newfib = *fib_p;
*(fib_p++) = NULL;
/*
* Make the copy of the FIB
*/
memcpy(hw_newfib, hw_fib, sizeof(struct hw_fib));
memcpy(newfib, fib, sizeof(struct fib));
newfib->hw_fib_va = hw_newfib;
/*
* Put the FIB onto the
* fibctx's fibs
*/
list_add_tail(&newfib->fiblink, &fibctx->fib_list);
fibctx->count++;
/*
* Set the event to wake up the
* thread that is waiting.
*/
up(&fibctx->wait_sem);
} else {
printk(KERN_WARNING "aifd: didn't allocate NewFib.\n");
}
entry = entry->next;
}
/*
* Set the status of this FIB
*/
*(__le32 *)hw_fib->data = cpu_to_le32(ST_OK);
aac_fib_adapter_complete(fib, sizeof(u32));
spin_unlock_irqrestore(&dev->fib_lock, flagv);
/* Free up the remaining resources */
hw_fib_p = hw_fib_pool;
fib_p = fib_pool;
while (hw_fib_p < &hw_fib_pool[num]) {
kfree(*hw_fib_p);
kfree(*fib_p);
++fib_p;
++hw_fib_p;
}
kfree(hw_fib_pool);
kfree(fib_pool);
}
kfree(fib);
spin_lock_irqsave(dev->queues->queue[HostNormCmdQueue].lock, flags);
}
/*
* There are no more AIF's
*/
spin_unlock_irqrestore(dev->queues->queue[HostNormCmdQueue].lock, flags);
/*
* Background activity
*/
if ((time_before(next_check_jiffies,next_jiffies))
&& ((difference = next_check_jiffies - jiffies) <= 0)) {
next_check_jiffies = next_jiffies;
if (aac_check_health(dev) == 0) {
difference = ((long)(unsigned)check_interval)
* HZ;
next_check_jiffies = jiffies + difference;
} else if (!dev->queues)
break;
}
if (!time_before(next_check_jiffies,next_jiffies)
&& ((difference = next_jiffies - jiffies) <= 0)) {
struct timeval now;
int ret;
/* Don't even try to talk to adapter if its sick */
ret = aac_check_health(dev);
if (!ret && !dev->queues)
break;
next_check_jiffies = jiffies
+ ((long)(unsigned)check_interval)
* HZ;
do_gettimeofday(&now);
/* Synchronize our watches */
if (((1000000 - (1000000 / HZ)) > now.tv_usec)
&& (now.tv_usec > (1000000 / HZ)))
difference = (((1000000 - now.tv_usec) * HZ)
+ 500000) / 1000000;
else if (ret == 0) {
struct fib *fibptr;
if ((fibptr = aac_fib_alloc(dev))) {
[SCSI] aacraid: fix File System going into read-only mode These particular problems were reported by Cisco and SAP and customers as well. Cisco reported on RHEL4 U6 and SAP reported on SLES9 SP4 and SLES10 SP2. We added these fixes on RHEL4 U6 and gave a private build to IBM and Cisco. Cisco and IBM tested it for more than 15 days and they reported that they did not see the issue so far. Before the fix, Cisco used to see the issue within 5 days. We generated a patch for SLES9 SP4 and SLES10 SP2 and submitted to Novell. Novell applied the patch and gave a test build to SAP. SAP tested and reported that the build is working properly. We also tested in our lab using the tools "dishogsync", which is IO stress tool and the tool was provided by Cisco. Issue1: File System going into read-only mode Root cause: The driver tends to not free the memory (FIB) when the management request exits prematurely. The accumulation of such un-freed memory causes the driver to fail to allocate anymore memory (FIB) and hence return 0x70000 value to the upper layer, which puts the file system into read only mode. Fix details: The fix makes sure to free the memory (FIB) even if the request exits prematurely hence ensuring the driver wouldn't run out of memory (FIBs). Issue2: False Raid Alert occurs When the Physical Drives and Logical drives are reported as deleted or added, even though there is no change done on the system Root cause: Driver IOCTLs is signaled with EINTR while waiting on response from the lower layers. Returning "EINTR" will never initiate internal retry. Fix details: The issue was fixed by replacing "EINTR" with "ERESTARTSYS" for mid-layer retries. Signed-off-by: Penchala Narasimha Reddy <ServeRAIDDriver@hcl.in> Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2009-12-21 21:09:27 +08:00
int status;
__le32 *info;
aac_fib_init(fibptr);
info = (__le32 *) fib_data(fibptr);
if (now.tv_usec > 500000)
++now.tv_sec;
*info = cpu_to_le32(now.tv_sec);
[SCSI] aacraid: fix File System going into read-only mode These particular problems were reported by Cisco and SAP and customers as well. Cisco reported on RHEL4 U6 and SAP reported on SLES9 SP4 and SLES10 SP2. We added these fixes on RHEL4 U6 and gave a private build to IBM and Cisco. Cisco and IBM tested it for more than 15 days and they reported that they did not see the issue so far. Before the fix, Cisco used to see the issue within 5 days. We generated a patch for SLES9 SP4 and SLES10 SP2 and submitted to Novell. Novell applied the patch and gave a test build to SAP. SAP tested and reported that the build is working properly. We also tested in our lab using the tools "dishogsync", which is IO stress tool and the tool was provided by Cisco. Issue1: File System going into read-only mode Root cause: The driver tends to not free the memory (FIB) when the management request exits prematurely. The accumulation of such un-freed memory causes the driver to fail to allocate anymore memory (FIB) and hence return 0x70000 value to the upper layer, which puts the file system into read only mode. Fix details: The fix makes sure to free the memory (FIB) even if the request exits prematurely hence ensuring the driver wouldn't run out of memory (FIBs). Issue2: False Raid Alert occurs When the Physical Drives and Logical drives are reported as deleted or added, even though there is no change done on the system Root cause: Driver IOCTLs is signaled with EINTR while waiting on response from the lower layers. Returning "EINTR" will never initiate internal retry. Fix details: The issue was fixed by replacing "EINTR" with "ERESTARTSYS" for mid-layer retries. Signed-off-by: Penchala Narasimha Reddy <ServeRAIDDriver@hcl.in> Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2009-12-21 21:09:27 +08:00
status = aac_fib_send(SendHostTime,
fibptr,
sizeof(*info),
FsaNormal,
1, 1,
NULL,
NULL);
[SCSI] aacraid: fix File System going into read-only mode These particular problems were reported by Cisco and SAP and customers as well. Cisco reported on RHEL4 U6 and SAP reported on SLES9 SP4 and SLES10 SP2. We added these fixes on RHEL4 U6 and gave a private build to IBM and Cisco. Cisco and IBM tested it for more than 15 days and they reported that they did not see the issue so far. Before the fix, Cisco used to see the issue within 5 days. We generated a patch for SLES9 SP4 and SLES10 SP2 and submitted to Novell. Novell applied the patch and gave a test build to SAP. SAP tested and reported that the build is working properly. We also tested in our lab using the tools "dishogsync", which is IO stress tool and the tool was provided by Cisco. Issue1: File System going into read-only mode Root cause: The driver tends to not free the memory (FIB) when the management request exits prematurely. The accumulation of such un-freed memory causes the driver to fail to allocate anymore memory (FIB) and hence return 0x70000 value to the upper layer, which puts the file system into read only mode. Fix details: The fix makes sure to free the memory (FIB) even if the request exits prematurely hence ensuring the driver wouldn't run out of memory (FIBs). Issue2: False Raid Alert occurs When the Physical Drives and Logical drives are reported as deleted or added, even though there is no change done on the system Root cause: Driver IOCTLs is signaled with EINTR while waiting on response from the lower layers. Returning "EINTR" will never initiate internal retry. Fix details: The issue was fixed by replacing "EINTR" with "ERESTARTSYS" for mid-layer retries. Signed-off-by: Penchala Narasimha Reddy <ServeRAIDDriver@hcl.in> Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2009-12-21 21:09:27 +08:00
/* Do not set XferState to zero unless
* receives a response from F/W */
if (status >= 0)
aac_fib_complete(fibptr);
/* FIB should be freed only after
* getting the response from the F/W */
if (status != -ERESTARTSYS)
aac_fib_free(fibptr);
}
difference = (long)(unsigned)update_interval*HZ;
} else {
/* retry shortly */
difference = 10 * HZ;
}
next_jiffies = jiffies + difference;
if (time_before(next_check_jiffies,next_jiffies))
difference = next_check_jiffies - jiffies;
}
if (difference <= 0)
difference = 1;
set_current_state(TASK_INTERRUPTIBLE);
schedule_timeout(difference);
if (kthread_should_stop())
break;
}
if (dev->queues)
remove_wait_queue(&dev->queues->queue[HostNormCmdQueue].cmdready, &wait);
dev->aif_thread = 0;
return 0;
}
int aac_acquire_irq(struct aac_dev *dev)
{
int i;
int j;
int ret = 0;
int cpu;
cpu = cpumask_first(cpu_online_mask);
if (!dev->sync_mode && dev->msi_enabled && dev->max_msix > 1) {
for (i = 0; i < dev->max_msix; i++) {
dev->aac_msix[i].vector_no = i;
dev->aac_msix[i].dev = dev;
if (request_irq(dev->msixentry[i].vector,
dev->a_ops.adapter_intr,
0, "aacraid", &(dev->aac_msix[i]))) {
printk(KERN_ERR "%s%d: Failed to register IRQ for vector %d.\n",
dev->name, dev->id, i);
for (j = 0 ; j < i ; j++)
free_irq(dev->msixentry[j].vector,
&(dev->aac_msix[j]));
pci_disable_msix(dev->pdev);
ret = -1;
}
if (irq_set_affinity_hint(dev->msixentry[i].vector,
get_cpu_mask(cpu))) {
printk(KERN_ERR "%s%d: Failed to set IRQ affinity for cpu %d\n",
dev->name, dev->id, cpu);
}
cpu = cpumask_next(cpu, cpu_online_mask);
}
} else {
dev->aac_msix[0].vector_no = 0;
dev->aac_msix[0].dev = dev;
if (request_irq(dev->pdev->irq, dev->a_ops.adapter_intr,
IRQF_SHARED, "aacraid",
&(dev->aac_msix[0])) < 0) {
if (dev->msi)
pci_disable_msi(dev->pdev);
printk(KERN_ERR "%s%d: Interrupt unavailable.\n",
dev->name, dev->id);
ret = -1;
}
}
return ret;
}
void aac_free_irq(struct aac_dev *dev)
{
int i;
int cpu;
cpu = cpumask_first(cpu_online_mask);
if (dev->pdev->device == PMC_DEVICE_S6 ||
dev->pdev->device == PMC_DEVICE_S7 ||
dev->pdev->device == PMC_DEVICE_S8 ||
dev->pdev->device == PMC_DEVICE_S9) {
if (dev->max_msix > 1) {
for (i = 0; i < dev->max_msix; i++) {
if (irq_set_affinity_hint(
dev->msixentry[i].vector, NULL)) {
printk(KERN_ERR "%s%d: Failed to reset IRQ affinity for cpu %d\n",
dev->name, dev->id, cpu);
}
cpu = cpumask_next(cpu, cpu_online_mask);
free_irq(dev->msixentry[i].vector,
&(dev->aac_msix[i]));
}
} else {
free_irq(dev->pdev->irq, &(dev->aac_msix[0]));
}
} else {
free_irq(dev->pdev->irq, dev);
}
if (dev->msi)
pci_disable_msi(dev->pdev);
else if (dev->max_msix > 1)
pci_disable_msix(dev->pdev);
}