Re: [PATCH PCI PCIe portdrv: Fix allocation of interrupts (rev. 5)

From: Hidetoshi Seto
Date: Sun Jan 18 2009 - 22:40:37 EST

Next message: Stephen Rothwell: "Re: [GIT PULL -tip] fix 41 'make headers_check' warnings"
Previous message: David Miller: "Re: [PATCH] tcp: splice as many packets as possible at once"
In reply to: Rafael J. Wysocki: "[PATCH PCI PCIe portdrv: Fix allocation of interrupts (rev. 5) (was: Re: [PATCH] PCI PCIe portdrv: Fix allocation ...)"
Next in thread: Kenji Kaneshige: "Re: [PATCH PCI PCIe portdrv: Fix allocation of interrupts (rev. 5)"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

Rafael J. Wysocki wrote:
> On Saturday 17 January 2009, Rafael J. Wysocki wrote:
(snip)
>> If MSI-X are supported, it allocates as many vectors as there are entries
>> in the port's MSI-X table, but no more than 32, and figures out which of them
>> will be used for the port services.
>
> The patch didn't check which services are available during the MSI-X set up
> which was wrong.
>
> Also, in the meantime, i thought it might be a good idea to free the interrupt
> routing table entries that aren't going to be used after all.
>
> The patch below adds this to the previous version and checks for the
> availability of port services in the MSI-X setup resume. I hope it will
> be acceptable to everyone.
>
> Thanks,
> Rafael
>
> ---
> Subject: PCI PCIe portdrv: Fix allocation of interrupts (rev. 5)
> From: Rafael J. Wysocki <rjw@xxxxxxx>
>
> If MSI-X interrupt mode is used by the PCI Express port driver, too
> many vectors are allocated and it is not ensured that the right
> vectors will be used for the right services. Namely, the PCI Express
> specification states that both PCI Express native PME and PCI Express
> hotplug will always use the same MSI or MSI-X message for signalling
> interrupts, which implies that the same vector will be used by both
> of them. Also, the VC service does not use interrupts at all.
> Moreover, is not clear which of the vectors allocated by
> pci_enable_msix() in the current code will be used for PME and
> hotplug and which of them will be used for AER if all of these
> services are configured.
>
> For these reasons, rework the allocation of interrupts for PCI
> Express ports so that if MSI-X are enabled, the right vectors will be
> used for the right purposes.
>
> Signed-off-by: Rafael J. Wysocki <rjw@xxxxxxx>
> ---
> drivers/pci/msi.c | 24 +++-
> drivers/pci/pcie/portdrv.h | 6 +
> drivers/pci/pcie/portdrv_core.c | 195 ++++++++++++++++++++++++++++++++--------
> include/linux/pci.h | 5 +
> include/linux/pcieport_if.h | 12 +-
> 5 files changed, 194 insertions(+), 48 deletions(-)
>
> Index: linux-2.6/drivers/pci/pcie/portdrv_core.c
> ===================================================================
> --- linux-2.6.orig/drivers/pci/pcie/portdrv_core.c
> +++ linux-2.6/drivers/pci/pcie/portdrv_core.c
> @@ -31,6 +31,141 @@ static void release_pcie_device(struct d
> }
>
> /**
> + * pcie_port_msix_add_entry - add entry to given array of MSI-X entries
> + * @entries: Array of MSI-X entries
> + * @new_entry: Index of the entry to add to the array
> + * @nr_entries: Number of entries aleady in the array
> + *
> + * Return value: Position of the added entry in the array
> + */
> +static int pcie_port_msix_add_entry(
> + struct msix_entry *entries, int new_entry, int nr_entries)
> +{
> + int j;
> +
> + for (j = 0; j < nr_entries; j++)
> + if (entries[j].entry == new_entry)
> + return j;
> +
> + entries[j].entry = new_entry;
> + return j;
> +}
> +
> +/**
> + * pcie_port_enable_msix - try to set up MSI-X as interrupt mode for given port
> + * @dev: PCI Express port to handle
> + * @vectors: Array of interrupt vectors to populate
> + * @mask: Bitmask of port capabilities returned by get_port_device_capability()
> + *
> + * Return value: 0 on success, error code on failure
> + */
> +static int pcie_port_enable_msix(struct pci_dev *dev, int *vectors, int mask)
> +{
> + struct msix_entry *msix_entries;
> + int idx[PCIE_PORT_DEVICE_MAXSERVICES];
> + int nr_entries, status, pos, i, nvec;
> + u16 reg16;
> + u32 reg32;
> +
> + nr_entries = pci_msix_table_size(dev);
> + if (!nr_entries)
> + return -EINVAL;
> + if (nr_entries > PCIE_PORT_MAX_MSIX_ENTRIES)
> + nr_entries = PCIE_PORT_MAX_MSIX_ENTRIES;
> +
> + msix_entries = kzalloc(sizeof(*msix_entries) * nr_entries, GFP_KERNEL);
> + if (!msix_entries)
> + return -ENOMEM;
> +
> + /*
> + * Allocate as many entries as the device wants temporarily, so that we
> + * can check which of them will be useful.
> + */
> + for (i = 0; i < nr_entries; i++)
> + msix_entries[i].entry = i;

/*
* So, if msix_entries is correctly equal to the number of entries this
* port actually uses, we'll happily go through without using trick.
*/
> +
> + status = pci_enable_msix(dev, msix_entries, nr_entries);
> + if (status)
> + goto Exit;
> +
> + for (i = 0; i < PCIE_PORT_DEVICE_MAXSERVICES; i++)
> + idx[i] = -1;
> + status = -EIO;
> + nvec = 0;
> +
> + if (mask & (PCIE_PORT_SERVICE_PME | PCIE_PORT_SERVICE_HP)) {
> + int entry;
> +
> + /*
> + * The code below follows the PCI Express Base Specification 2.0
> + * stating in Section 6.1.6 that "PME and Hot-Plug Event
> + * interrupts (when both are implemented) always share the same
> + * MSI or MSI-X vector, as indicated by the Interrupt Message
> + * Number field in the PCI Express Capabilities register", where
> + * according to Section 7.8.2 of the specification "For MSI-X,
> + * the value in this field indicates which MSI-X Table entry is
> + * used to generate the interrupt message."
> + */
> + pos = pci_find_capability(dev, PCI_CAP_ID_EXP);
> + pci_read_config_word(dev, pos + PCIE_CAPABILITIES_REG, &reg16);
> + entry = (reg16 >> 9) & PCIE_PORT_MSI_VECTOR_MASK;
> + if (entry >= nr_entries)
> + goto Error;
> +
> + i = pcie_port_msix_add_entry(msix_entries, entry, nvec);
> + if (i == nvec)
> + nvec++;
> +
> + idx[PCIE_PORT_SERVICE_PME_SHIFT] = i;
> + idx[PCIE_PORT_SERVICE_HP_SHIFT] = i;
> + }
> +
> + if (mask & PCIE_PORT_SERVICE_AER) {
> + int entry;
> +
> + /*
> + * The code below follows Section 7.10.10 of the PCI Express
> + * Base Specification 2.0 stating that bits 31-27 of the Root
> + * Error Status Register contain a value indicating which of the
> + * MSI/MSI-X vectors assigned to the port is going to be used
> + * for AER, where "For MSI-X, the value in this register
> + * indicates which MSI-X Table entry is used to generate the
> + * interrupt message."
> + */
> + pos = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_ERR);
> + pci_read_config_dword(dev, pos + PCI_ERR_ROOT_STATUS, &reg32);
> + entry = reg32 >> 27;
> + if (entry >= nr_entries)
> + goto Error;
> +
> + i = pcie_port_msix_add_entry(msix_entries, entry, nvec);
> + if (i == nvec)
> + nvec++;
> +
> + idx[PCIE_PORT_SERVICE_AER_SHIFT] = i;
> + }
> +

/* Are there any unused entries? */
if (nr_allocated > nvec) {
/* this port have extra entries not for services we know... */

> + /* Drop the temporary MSI-X setup */
> + pci_disable_msix(dev);
> +
> + /* Now allocate the MSI-X vectors for real */
> + status = pci_enable_msix(dev, msix_entries, nvec);
> + if (status)
> + goto Error;

/*
* World have broken hardwares, so even spec says numbers are constant,
* it would be better to re-check registers after 2nd pci_enable_msix.
* Or we just skip this. (However this was what your concern, Rafael?)
*/
if (func_foo_do_paranoia_check(dev, msix_entries, nvec))
goto Error;
}

> +
> + for (i = 0; i < PCIE_PORT_DEVICE_MAXSERVICES; i++)
> + vectors[i] = idx[i] >= 0 ? msix_entries[idx[i]].vector : -1;
> +
> + Exit:
> + kfree(msix_entries);
> + return status;
> +
> + Error:
> + pci_disable_msix(dev);
> + goto Exit;
> +}
> +
> +/**
> * assign_interrupt_mode - choose interrupt mode for PCI Express port services
> * (INTx, MSI-X, MSI) and set up vectors
> * @dev: PCI Express port to handle
> @@ -42,49 +177,31 @@ static void release_pcie_device(struct d
> static int assign_interrupt_mode(struct pci_dev *dev, int *vectors, int mask)
> {
> struct pcie_port_data *port_data = pci_get_drvdata(dev);
> - int i, pos, nvec, status = -EINVAL;
> - int interrupt_mode = PCIE_PORT_NO_IRQ;
> -
> - /* Set INTx as default */
> - for (i = 0, nvec = 0; i < PCIE_PORT_DEVICE_MAXSERVICES; i++) {
> - if (mask & (1 << i))
> - nvec++;
> - vectors[i] = dev->irq;
> - }
> - if (dev->pin)
> - interrupt_mode = PCIE_PORT_INTx_MODE;
> + int irq, interrupt_mode = PCIE_PORT_NO_IRQ;
> + int i;
>
> /* Check MSI quirk */
> if (port_data->port_type == PCIE_RC_PORT && pcie_mch_quirk)
> - return interrupt_mode;
> + goto Fallback;
> +
> + /* Try to use MSI-X if supported */
> + if (!pcie_port_enable_msix(dev, vectors, mask))
> + return PCIE_PORT_MSIX_MODE;
> +
> + /* We're not going to use MSI-X, so try MSI and fall back to INTx */
> + if (!pci_enable_msi(dev))
> + interrupt_mode = PCIE_PORT_MSI_MODE;
> +
> + Fallback:
> + if (interrupt_mode == PCIE_PORT_NO_IRQ && dev->pin)
> + interrupt_mode = PCIE_PORT_INTx_MODE;
> +
> + irq = interrupt_mode != PCIE_PORT_NO_IRQ ? dev->irq : -1;
> + for (i = 0; i < PCIE_PORT_DEVICE_MAXSERVICES; i++)
> + vectors[i] = irq;
> +
> + vectors[PCIE_PORT_SERVICE_VC_SHIFT] = -1;
>
> - /* Select MSI-X over MSI if supported */
> - pos = pci_find_capability(dev, PCI_CAP_ID_MSIX);
> - if (pos) {
> - struct msix_entry msix_entries[PCIE_PORT_DEVICE_MAXSERVICES] =
> - {{0, 0}, {0, 1}, {0, 2}, {0, 3}};
> - status = pci_enable_msix(dev, msix_entries, nvec);
> - if (!status) {
> - int j = 0;
> -
> - interrupt_mode = PCIE_PORT_MSIX_MODE;
> - for (i = 0; i < PCIE_PORT_DEVICE_MAXSERVICES; i++) {
> - if (mask & (1 << i))
> - vectors[i] = msix_entries[j++].vector;
> - }
> - }
> - }
> - if (status) {
> - pos = pci_find_capability(dev, PCI_CAP_ID_MSI);
> - if (pos) {
> - status = pci_enable_msi(dev);
> - if (!status) {
> - interrupt_mode = PCIE_PORT_MSI_MODE;
> - for (i = 0;i < PCIE_PORT_DEVICE_MAXSERVICES;i++)
> - vectors[i] = dev->irq;
> - }
> - }
> - }
> return interrupt_mode;
> }
>
> Index: linux-2.6/include/linux/pcieport_if.h
> ===================================================================
> --- linux-2.6.orig/include/linux/pcieport_if.h
> +++ linux-2.6/include/linux/pcieport_if.h
> @@ -16,10 +16,14 @@
> #define PCIE_ANY_PORT 7
>
> /* Service Type */
> -#define PCIE_PORT_SERVICE_PME 1 /* Power Management Event */
> -#define PCIE_PORT_SERVICE_AER 2 /* Advanced Error Reporting */
> -#define PCIE_PORT_SERVICE_HP 4 /* Native Hotplug */
> -#define PCIE_PORT_SERVICE_VC 8 /* Virtual Channel */
> +#define PCIE_PORT_SERVICE_PME_SHIFT 0 /* Power Management Event */
> +#define PCIE_PORT_SERVICE_PME (1 << PCIE_PORT_SERVICE_PME_SHIFT)
> +#define PCIE_PORT_SERVICE_AER_SHIFT 1 /* Advanced Error Reporting */
> +#define PCIE_PORT_SERVICE_AER (1 << PCIE_PORT_SERVICE_AER_SHIFT)
> +#define PCIE_PORT_SERVICE_HP_SHIFT 2 /* Native Hotplug */
> +#define PCIE_PORT_SERVICE_HP (1 << PCIE_PORT_SERVICE_HP_SHIFT)
> +#define PCIE_PORT_SERVICE_VC_SHIFT 3 /* Virtual Channel */
> +#define PCIE_PORT_SERVICE_VC (1 << PCIE_PORT_SERVICE_VC_SHIFT)
>
> /* Root/Upstream/Downstream Port's Interrupt Mode */
> #define PCIE_PORT_NO_IRQ (-1)
> Index: linux-2.6/drivers/pci/pcie/portdrv.h
> ===================================================================
> --- linux-2.6.orig/drivers/pci/pcie/portdrv.h
> +++ linux-2.6/drivers/pci/pcie/portdrv.h
> @@ -25,6 +25,12 @@
> #define PCIE_CAPABILITIES_REG 0x2
> #define PCIE_SLOT_CAPABILITIES_REG 0x14
> #define PCIE_PORT_DEVICE_MAXSERVICES 4
> +#define PCIE_PORT_MSI_VECTOR_MASK 0x1f
> +/*
> + * According to the PCI Express Base Specification 2.0, the indices of the MSI-X
> + * table entires used by port services must not exceed 31
> + */
> +#define PCIE_PORT_MAX_MSIX_ENTRIES 32
>
> #define get_descriptor_id(type, service) (((type - 4) << 4) | service)
>
> Index: linux-2.6/drivers/pci/msi.c
> ===================================================================
> --- linux-2.6.orig/drivers/pci/msi.c
> +++ linux-2.6/drivers/pci/msi.c
> @@ -670,6 +670,23 @@ static int msi_free_irqs(struct pci_dev*
> }
>
> /**
> + * pci_msix_table_size - return the number of device's MSI-X table entries
> + * @dev: pointer to the pci_dev data structure of MSI-X device function
> + */
> +int pci_msix_table_size(struct pci_dev *dev)
> +{
> + int pos;
> + u16 control;
> +
> + pos = pci_find_capability(dev, PCI_CAP_ID_MSIX);
> + if (!pos)
> + return 0;
> +
> + pci_read_config_word(dev, msi_control_reg(pos), &control);
> + return multi_msix_capable(control);
> +}
> +
> +/**

I think this pci_msix_table_size() is useful alone.
It would be nice if we can have separated patches.

i.e.:
[PATCH] PCI/MSI: introduce pci_msix_table_size()
[PATCH] PCI PCIe portdrv: Fix allocation of interrupts (rev. 6)

Thanks,
H.Seto

> * pci_enable_msix - configure device's MSI-X capability structure
> * @dev: pointer to the pci_dev data structure of MSI-X device function
> * @entries: pointer to an array of MSI-X entries
> @@ -686,9 +703,8 @@ static int msi_free_irqs(struct pci_dev*
> **/
> int pci_enable_msix(struct pci_dev* dev, struct msix_entry *entries, int nvec)
> {
> - int status, pos, nr_entries;
> + int status, nr_entries;
> int i, j;
> - u16 control;
>
> if (!entries)
> return -EINVAL;
> @@ -697,9 +713,7 @@ int pci_enable_msix(struct pci_dev* dev,
> if (status)
> return status;
>
> - pos = pci_find_capability(dev, PCI_CAP_ID_MSIX);
> - pci_read_config_word(dev, msi_control_reg(pos), &control);
> - nr_entries = multi_msix_capable(control);
> + nr_entries = pci_msix_table_size(dev);
> if (nvec > nr_entries)
> return -EINVAL;
>
> Index: linux-2.6/include/linux/pci.h
> ===================================================================
> --- linux-2.6.orig/include/linux/pci.h
> +++ linux-2.6/include/linux/pci.h
> @@ -799,6 +799,10 @@ static inline void pci_msi_shutdown(stru
> static inline void pci_disable_msi(struct pci_dev *dev)
> { }
>
> +static inline int pci_msix_table_size(struct pci_dev *dev)
> +{
> + return 0;
> +}
> static inline int pci_enable_msix(struct pci_dev *dev,
> struct msix_entry *entries, int nvec)
> {
> @@ -823,6 +827,7 @@ static inline int pci_msi_enabled(void)
> extern int pci_enable_msi(struct pci_dev *dev);
> extern void pci_msi_shutdown(struct pci_dev *dev);
> extern void pci_disable_msi(struct pci_dev *dev);
> +extern int pci_msix_table_size(struct pci_dev *dev);
> extern int pci_enable_msix(struct pci_dev *dev,
> struct msix_entry *entries, int nvec);
> extern void pci_msix_shutdown(struct pci_dev *dev);
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-pci" in
> the body of a message to majordomo@xxxxxxxxxxxxxxx
> More majordomo info at http://vger.kernel.org/majordomo-info.html
>
>

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Next message: Stephen Rothwell: "Re: [GIT PULL -tip] fix 41 'make headers_check' warnings"
Previous message: David Miller: "Re: [PATCH] tcp: splice as many packets as possible at once"
In reply to: Rafael J. Wysocki: "[PATCH PCI PCIe portdrv: Fix allocation of interrupts (rev. 5) (was: Re: [PATCH] PCI PCIe portdrv: Fix allocation ...)"
Next in thread: Kenji Kaneshige: "Re: [PATCH PCI PCIe portdrv: Fix allocation of interrupts (rev. 5)"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]