IB/mthca: Query port fix

[linux-2.6/verdex.git] / Documentation / MSI-HOWTO.txt

                The MSI Driver Guide HOWTO
        Tom L Nguyen tom.l.nguyen@intel.com
                        10/03/2003
        Revised Feb 12, 2004 by Martine Silbermann
                email: Martine.Silbermann@hp.com
        Revised Jun 25, 2004 by Tom L Nguyen

1. About this guide

This guide describes the basics of Message Signaled Interrupts (MSI),
the advantages of using MSI over traditional interrupt mechanisms,
and how to enable your driver to use MSI or MSI-X. Also included is
a Frequently Asked Questions (FAQ) section.

1.1 Terminology

PCI devices can be single-function or multi-function.  In either case,
when this text talks about enabling or disabling MSI on a "device
function," it is referring to one specific PCI device and function and
not to all functions on a PCI device (unless the PCI device has only
one function).

2. Copyright 2003 Intel Corporation

3. What is MSI/MSI-X?

Message Signaled Interrupt (MSI), as described in the PCI Local Bus
Specification Revision 2.3 or later, is an optional feature, and a
required feature for PCI Express devices. MSI enables a device function
to request service by sending an Inbound Memory Write on its PCI bus to
the FSB as a Message Signal Interrupt transaction. Because MSI is
generated in the form of a Memory Write, all transaction conditions,
such as a Retry, Master-Abort, Target-Abort or normal completion, are
supported.

A PCI device that supports MSI must also support pin IRQ assertion
interrupt mechanism to provide backward compatibility for systems that
do not support MSI. In systems which support MSI, the bus driver is
responsible for initializing the message address and message data of
the device function's MSI/MSI-X capability structure during device
initial configuration.

An MSI capable device function indicates MSI support by implementing
the MSI/MSI-X capability structure in its PCI capability list. The
device function may implement both the MSI capability structure and
the MSI-X capability structure; however, the bus driver should not
enable both.

The MSI capability structure contains Message Control register,
Message Address register and Message Data register. These registers
provide the bus driver control over MSI. The Message Control register
indicates the MSI capability supported by the device. The Message
Address register specifies the target address and the Message Data
register specifies the characteristics of the message. To request
service, the device function writes the content of the Message Data
register to the target address. The device and its software driver
are prohibited from writing to these registers.

The MSI-X capability structure is an optional extension to MSI. It
uses an independent and separate capability structure. There are
some key advantages to implementing the MSI-X capability structure
over the MSI capability structure as described below.

        - Support a larger maximum number of vectors per function.

        - Provide the ability for system software to configure
        each vector with an independent message address and message
        data, specified by a table that resides in Memory Space.

        - MSI and MSI-X both support per-vector masking. Per-vector
        masking is an optional extension of MSI but a required
        feature for MSI-X. Per-vector masking provides the kernel the
        ability to mask/unmask a single MSI while running its
        interrupt service routine. If per-vector masking is
        not supported, then the device driver should provide the
        hardware/software synchronization to ensure that the device
        generates MSI when the driver wants it to do so.

4. Why use MSI?

As a benefit to the simplification of board design, MSI allows board
designers to remove out-of-band interrupt routing. MSI is another
step towards a legacy-free environment.

Due to increasing pressure on chipset and processor packages to
reduce pin count, the need for interrupt pins is expected to
diminish over time. Devices, due to pin constraints, may implement
messages to increase performance.

PCI Express endpoints uses INTx emulation (in-band messages) instead
of IRQ pin assertion. Using INTx emulation requires interrupt
sharing among devices connected to the same node (PCI bridge) while
MSI is unique (non-shared) and does not require BIOS configuration
support. As a result, the PCI Express technology requires MSI
support for better interrupt performance.

Using MSI enables the device functions to support two or more
vectors, which can be configured to target different CPUs to
increase scalability.

5. Configuring a driver to use MSI/MSI-X

By default, the kernel will not enable MSI/MSI-X on all devices that
support this capability. The CONFIG_PCI_MSI kernel option
must be selected to enable MSI/MSI-X support.

5.1 Including MSI/MSI-X support into the kernel

To allow MSI/MSI-X capable device drivers to selectively enable
MSI/MSI-X (using pci_enable_msi()/pci_enable_msix() as described
below), the VECTOR based scheme needs to be enabled by setting
CONFIG_PCI_MSI during kernel config.

Since the target of the inbound message is the local APIC, providing
CONFIG_X86_LOCAL_APIC must be enabled as well as CONFIG_PCI_MSI.

5.2 Configuring for MSI support

Due to the non-contiguous fashion in vector assignment of the
existing Linux kernel, this version does not support multiple
messages regardless of a device function is capable of supporting
more than one vector. To enable MSI on a device function's MSI
capability structure requires a device driver to call the function
pci_enable_msi() explicitly.

5.2.1 API pci_enable_msi

int pci_enable_msi(struct pci_dev *dev)

With this new API, a device driver that wants to have MSI
enabled on its device function must call this API to enable MSI.
A successful call will initialize the MSI capability structure
with ONE vector, regardless of whether a device function is
capable of supporting multiple messages. This vector replaces the
pre-assigned dev->irq with a new MSI vector. To avoid a conflict
of the new assigned vector with existing pre-assigned vector requires
a device driver to call this API before calling request_irq().

5.2.2 API pci_disable_msi

void pci_disable_msi(struct pci_dev *dev)

This API should always be used to undo the effect of pci_enable_msi()
when a device driver is unloading. This API restores dev->irq with
the pre-assigned IOAPIC vector and switches a device's interrupt
mode to PCI pin-irq assertion/INTx emulation mode.

Note that a device driver should always call free_irq() on the MSI vector
that it has done request_irq() on before calling this API. Failure to do
so results in a BUG_ON() and a device will be left with MSI enabled and
leaks its vector.

5.2.3 MSI mode vs. legacy mode diagram

The below diagram shows the events which switch the interrupt
mode on the MSI-capable device function between MSI mode and
PIN-IRQ assertion mode.

         ------------   pci_enable_msi   ------------------------
        |            | <=============== |                        |
        | MSI MODE   |                  | PIN-IRQ ASSERTION MODE |
        |            | ===============> |                        |
         ------------   pci_disable_msi  ------------------------


Figure 1. MSI Mode vs. Legacy Mode

In Figure 1, a device operates by default in legacy mode. Legacy
in this context means PCI pin-irq assertion or PCI-Express INTx
emulation. A successful MSI request (using pci_enable_msi()) switches
a device's interrupt mode to MSI mode. A pre-assigned IOAPIC vector
stored in dev->irq will be saved by the PCI subsystem and a new
assigned MSI vector will replace dev->irq.

To return back to its default mode, a device driver should always call
pci_disable_msi() to undo the effect of pci_enable_msi(). Note that a
device driver should always call free_irq() on the MSI vector it has
done request_irq() on before calling pci_disable_msi(). Failure to do
so results in a BUG_ON() and a device will be left with MSI enabled and
leaks its vector. Otherwise, the PCI subsystem restores a device's
dev->irq with a pre-assigned IOAPIC vector and marks the released
MSI vector as unused.

Once being marked as unused, there is no guarantee that the PCI
subsystem will reserve this MSI vector for a device. Depending on
the availability of current PCI vector resources and the number of
MSI/MSI-X requests from other drivers, this MSI may be re-assigned.

For the case where the PCI subsystem re-assigns this MSI vector to
another driver, a request to switch back to MSI mode may result
in being assigned a different MSI vector or a failure if no more
vectors are available.

5.3 Configuring for MSI-X support

Due to the ability of the system software to configure each vector of
the MSI-X capability structure with an independent message address
and message data, the non-contiguous fashion in vector assignment of
the existing Linux kernel has no impact on supporting multiple
messages on an MSI-X capable device functions. To enable MSI-X on
a device function's MSI-X capability structure requires its device
driver to call the function pci_enable_msix() explicitly.

The function pci_enable_msix(), once invoked, enables either
all or nothing, depending on the current availability of PCI vector
resources. If the PCI vector resources are available for the number
of vectors requested by a device driver, this function will configure
the MSI-X table of the MSI-X capability structure of a device with
requested messages. To emphasize this reason, for example, a device
may be capable for supporting the maximum of 32 vectors while its
software driver usually may request 4 vectors. It is recommended
that the device driver should call this function once during the
initialization phase of the device driver.

Unlike the function pci_enable_msi(), the function pci_enable_msix()
does not replace the pre-assigned IOAPIC dev->irq with a new MSI
vector because the PCI subsystem writes the 1:1 vector-to-entry mapping
into the field vector of each element contained in a second argument.
Note that the pre-assigned IOAPIC dev->irq is valid only if the device
operates in PIN-IRQ assertion mode. In MSI-X mode, any attempt at
using dev->irq by the device driver to request for interrupt service
may result unpredictabe behavior.

For each MSI-X vector granted, a device driver is responsible for calling
other functions like request_irq(), enable_irq(), etc. to enable
this vector with its corresponding interrupt service handler. It is
a device driver's choice to assign all vectors with the same
interrupt service handler or each vector with a unique interrupt
service handler.

5.3.1 Handling MMIO address space of MSI-X Table

The PCI 3.0 specification has implementation notes that MMIO address
space for a device's MSI-X structure should be isolated so that the
software system can set different pages for controlling accesses to the
MSI-X structure. The implementation of MSI support requires the PCI
subsystem, not a device driver, to maintain full control of the MSI-X
table/MSI-X PBA (Pending Bit Array) and MMIO address space of the MSI-X
table/MSI-X PBA.  A device driver is prohibited from requesting the MMIO
address space of the MSI-X table/MSI-X PBA. Otherwise, the PCI subsystem
will fail enabling MSI-X on its hardware device when it calls the function
pci_enable_msix().

5.3.2 Handling MSI-X allocation

Determining the number of MSI-X vectors allocated to a function is
dependent on the number of MSI capable devices and MSI-X capable
devices populated in the system. The policy of allocating MSI-X
vectors to a function is defined as the following:

#of MSI-X vectors allocated to a function = (x - y)/z where

x =     The number of available PCI vector resources by the time
        the device driver calls pci_enable_msix(). The PCI vector
        resources is the sum of the number of unassigned vectors
        (new) and the number of released vectors when any MSI/MSI-X
        device driver switches its hardware device back to a legacy
        mode or is hot-removed. The number of unassigned vectors
        may exclude some vectors reserved, as defined in parameter
        NR_HP_RESERVED_VECTORS, for the case where the system is
        capable of supporting hot-add/hot-remove operations. Users
        may change the value defined in NR_HR_RESERVED_VECTORS to
        meet their specific needs.

y =     The number of MSI capable devices populated in the system.
        This policy ensures that each MSI capable device has its
        vector reserved to avoid the case where some MSI-X capable
        drivers may attempt to claim all available vector resources.

z =     The number of MSI-X capable devices populated in the system.
        This policy ensures that maximum (x - y) is distributed
        evenly among MSI-X capable devices.

Note that the PCI subsystem scans y and z during a bus enumeration.
When the PCI subsystem completes configuring MSI/MSI-X capability
structure of a device as requested by its device driver, y/z is
decremented accordingly.

5.3.3 Handling MSI-X shortages

For the case where fewer MSI-X vectors are allocated to a function
than requested, the function pci_enable_msix() will return the
maximum number of MSI-X vectors available to the caller. A device
driver may re-send its request with fewer or equal vectors indicated
in the return. For example, if a device driver requests 5 vectors, but
the number of available vectors is 3 vectors, a value of 3 will be
returned as a result of pci_enable_msix() call. A function could be
designed for its driver to use only 3 MSI-X table entries as
different combinations as ABC--, A-B-C, A--CB, etc. Note that this
patch does not support multiple entries with the same vector. Such
attempt by a device driver to use 5 MSI-X table entries with 3 vectors
as ABBCC, AABCC, BCCBA, etc will result as a failure by the function
pci_enable_msix(). Below are the reasons why supporting multiple
entries with the same vector is an undesirable solution.

        - The PCI subsystem cannot determine the entry that
          generated the message to mask/unmask MSI while handling
          software driver ISR. Attempting to walk through all MSI-X
          table entries (2048 max) to mask/unmask any match vector
          is an undesirable solution.

        - Walking through all MSI-X table entries (2048 max) to handle
          SMP affinity of any match vector is an undesirable solution.

5.3.4 API pci_enable_msix

int pci_enable_msix(struct pci_dev *dev, struct msix_entry *entries, int nvec)

This API enables a device driver to request the PCI subsystem
to enable MSI-X messages on its hardware device. Depending on
the availability of PCI vectors resources, the PCI subsystem enables
either all or none of the requested vectors.

Argument 'dev' points to the device (pci_dev) structure.

Argument 'entries' is a pointer to an array of msix_entry structs.
The number of entries is indicated in argument 'nvec'.
struct msix_entry is defined in /driver/pci/msi.h:

struct msix_entry {
        u16     vector; /* kernel uses to write alloc vector */
        u16     entry; /* driver uses to specify entry */
};

A device driver is responsible for initializing the field 'entry' of
each element with a unique entry supported by MSI-X table. Otherwise,
-EINVAL will be returned as a result. A successful return of zero
indicates the PCI subsystem completed initializing each of the requested
entries of the MSI-X table with message address and message data.
Last but not least, the PCI subsystem will write the 1:1
vector-to-entry mapping into the field 'vector' of each element. A
device driver is responsible for keeping track of allocated MSI-X
vectors in its internal data structure.

A return of zero indicates that the number of MSI-X vectors was
successfully allocated. A return of greater than zero indicates
MSI-X vector shortage. Or a return of less than zero indicates
a failure. This failure may be a result of duplicate entries
specified in second argument, or a result of no available vector,
or a result of failing to initialize MSI-X table entries.

5.3.5 API pci_disable_msix

void pci_disable_msix(struct pci_dev *dev)

This API should always be used to undo the effect of pci_enable_msix()
when a device driver is unloading. Note that a device driver should
always call free_irq() on all MSI-X vectors it has done request_irq()
on before calling this API. Failure to do so results in a BUG_ON() and
a device will be left with MSI-X enabled and leaks its vectors.

5.3.6 MSI-X mode vs. legacy mode diagram

The below diagram shows the events which switch the interrupt
mode on the MSI-X capable device function between MSI-X mode and
PIN-IRQ assertion mode (legacy).

         ------------   pci_enable_msix(,,n) ------------------------
        |            | <===============     |                        |
        | MSI-X MODE |                      | PIN-IRQ ASSERTION MODE |
        |            | ===============>     |                        |
         ------------   pci_disable_msix     ------------------------

Figure 2. MSI-X Mode vs. Legacy Mode

In Figure 2, a device operates by default in legacy mode. A
successful MSI-X request (using pci_enable_msix()) switches a
device's interrupt mode to MSI-X mode. A pre-assigned IOAPIC vector
stored in dev->irq will be saved by the PCI subsystem; however,
unlike MSI mode, the PCI subsystem will not replace dev->irq with
assigned MSI-X vector because the PCI subsystem already writes the 1:1
vector-to-entry mapping into the field 'vector' of each element
specified in second argument.

To return back to its default mode, a device driver should always call
pci_disable_msix() to undo the effect of pci_enable_msix(). Note that
a device driver should always call free_irq() on all MSI-X vectors it
has done request_irq() on before calling pci_disable_msix(). Failure
to do so results in a BUG_ON() and a device will be left with MSI-X
enabled and leaks its vectors. Otherwise, the PCI subsystem switches a
device function's interrupt mode from MSI-X mode to legacy mode and
marks all allocated MSI-X vectors as unused.

Once being marked as unused, there is no guarantee that the PCI
subsystem will reserve these MSI-X vectors for a device. Depending on
the availability of current PCI vector resources and the number of
MSI/MSI-X requests from other drivers, these MSI-X vectors may be
re-assigned.

For the case where the PCI subsystem re-assigned these MSI-X vectors
to other drivers, a request to switch back to MSI-X mode may result
being assigned with another set of MSI-X vectors or a failure if no
more vectors are available.

5.4 Handling function implementing both MSI and MSI-X capabilities

For the case where a function implements both MSI and MSI-X
capabilities, the PCI subsystem enables a device to run either in MSI
mode or MSI-X mode but not both. A device driver determines whether it
wants MSI or MSI-X enabled on its hardware device. Once a device
driver requests for MSI, for example, it is prohibited from requesting
MSI-X; in other words, a device driver is not permitted to ping-pong
between MSI mod MSI-X mode during a run-time.

5.5 Hardware requirements for MSI/MSI-X support

MSI/MSI-X support requires support from both system hardware and
individual hardware device functions.

5.5.1 System hardware support

Since the target of MSI address is the local APIC CPU, enabling
MSI/MSI-X support in the Linux kernel is dependent on whether existing
system hardware supports local APIC. Users should verify that their
system supports local APIC operation by testing that it runs when
CONFIG_X86_LOCAL_APIC=y.

In SMP environment, CONFIG_X86_LOCAL_APIC is automatically set;
however, in UP environment, users must manually set
CONFIG_X86_LOCAL_APIC. Once CONFIG_X86_LOCAL_APIC=y, setting
CONFIG_PCI_MSI enables the VECTOR based scheme and the option for
MSI-capable device drivers to selectively enable MSI/MSI-X.

Note that CONFIG_X86_IO_APIC setting is irrelevant because MSI/MSI-X
vector is allocated new during runtime and MSI/MSI-X support does not
depend on BIOS support. This key independency enables MSI/MSI-X
support on future IOxAPIC free platforms.

5.5.2 Device hardware support

The hardware device function supports MSI by indicating the
MSI/MSI-X capability structure on its PCI capability list. By
default, this capability structure will not be initialized by
the kernel to enable MSI during the system boot. In other words,
the device function is running on its default pin assertion mode.
Note that in many cases the hardware supporting MSI have bugs,
which may result in system hangs. The software driver of specific
MSI-capable hardware is responsible for deciding whether to call
pci_enable_msi or not. A return of zero indicates the kernel
successfully initialized the MSI/MSI-X capability structure of the
device function. The device function is now running on MSI/MSI-X mode.

5.6 How to tell whether MSI/MSI-X is enabled on device function

At the driver level, a return of zero from the function call of
pci_enable_msi()/pci_enable_msix() indicates to a device driver that
its device function is initialized successfully and ready to run in
MSI/MSI-X mode.

At the user level, users can use the command 'cat /proc/interrupts'
to display the vectors allocated for devices and their interrupt
MSI/MSI-X modes ("PCI-MSI"/"PCI-MSI-X"). Below shows MSI mode is
enabled on a SCSI Adaptec 39320D Ultra320 controller.

           CPU0       CPU1
  0:     324639          0    IO-APIC-edge  timer
  1:       1186          0    IO-APIC-edge  i8042
  2:          0          0          XT-PIC  cascade
 12:       2797          0    IO-APIC-edge  i8042
 14:       6543          0    IO-APIC-edge  ide0
 15:          1          0    IO-APIC-edge  ide1
169:          0          0   IO-APIC-level  uhci-hcd
185:          0          0   IO-APIC-level  uhci-hcd
193:        138         10         PCI-MSI  aic79xx
201:         30          0         PCI-MSI  aic79xx
225:         30          0   IO-APIC-level  aic7xxx
233:         30          0   IO-APIC-level  aic7xxx
NMI:          0          0
LOC:     324553     325068
ERR:          0
MIS:          0

6. FAQ

Q1. Are there any limitations on using the MSI?

A1. If the PCI device supports MSI and conforms to the
specification and the platform supports the APIC local bus,
then using MSI should work.

Q2. Will it work on all the Pentium processors (P3, P4, Xeon,
AMD processors)? In P3 IPI's are transmitted on the APIC local
bus and in P4 and Xeon they are transmitted on the system
bus. Are there any implications with this?

A2. MSI support enables a PCI device sending an inbound
memory write (0xfeexxxxx as target address) on its PCI bus
directly to the FSB. Since the message address has a
redirection hint bit cleared, it should work.

Q3. The target address 0xfeexxxxx will be translated by the
Host Bridge into an interrupt message. Are there any
limitations on the chipsets such as Intel 8xx, Intel e7xxx,
or VIA?

A3. If these chipsets support an inbound memory write with
target address set as 0xfeexxxxx, as conformed to PCI
specification 2.3 or latest, then it should work.

Q4. From the driver point of view, if the MSI is lost because
of errors occurring during inbound memory write, then it may
wait forever. Is there a mechanism for it to recover?

A4. Since the target of the transaction is an inbound memory
write, all transaction termination conditions (Retry,
Master-Abort, Target-Abort, or normal completion) are
supported. A device sending an MSI must abide by all the PCI
rules and conditions regarding that inbound memory write. So,
if a retry is signaled it must retry, etc... We believe that
the recommendation for Abort is also a retry (refer to PCI
specification 2.3 or latest).