From: akpm@linux-foundation.org
To: mm-commits@vger.kernel.org
Cc: shannon.nelson@intel.com, davem@davemloft.net
Subject: + i-oat-add-support-for-msi-and-msi-x.patch added to -mm tree
Date: Fri, 31 Aug 2007 12:28:08 -0700 [thread overview]
Message-ID: <200708311928.l7VJS87Q022930@imap1.linux-foundation.org> (raw)
The patch titled
I/OAT: Add support for MSI and MSI-X
has been added to the -mm tree. Its filename is
i-oat-add-support-for-msi-and-msi-x.patch
*** Remember to use Documentation/SubmitChecklist when testing your code ***
See http://www.zip.com.au/~akpm/linux/patches/stuff/added-to-mm.txt to find
out what to do about this
------------------------------------------------------
Subject: I/OAT: Add support for MSI and MSI-X
From: Shannon Nelson <shannon.nelson@intel.com>
Add support for MSI and MSI-X interrupt handling, including the ability
to choose the desired interrupt method.
Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Acked-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
---
drivers/dma/ioat_dma.c | 348 ++++++++++++++++++++++++------
drivers/dma/ioatdma.h | 12 +
drivers/dma/ioatdma_registers.h | 6
include/linux/bitops.h | 6
4 files changed, 305 insertions(+), 67 deletions(-)
diff -puN drivers/dma/ioat_dma.c~i-oat-add-support-for-msi-and-msi-x drivers/dma/ioat_dma.c
--- a/drivers/dma/ioat_dma.c~i-oat-add-support-for-msi-and-msi-x
+++ a/drivers/dma/ioat_dma.c
@@ -47,6 +47,65 @@
static void ioat_dma_start_null_desc(struct ioat_dma_chan *ioat_chan);
static void ioat_dma_memcpy_cleanup(struct ioat_dma_chan *ioat_chan);
+struct ioat_dma_chan *ioat_lookup_chan_by_index(struct ioatdma_device *device,
+ int index)
+{
+ return device->idx[index];
+}
+
+/**
+ * ioat_dma_do_interrupt - handler used for single vector interrupt mode
+ * @irq: interrupt id
+ * @data: interrupt data
+ */
+static irqreturn_t ioat_dma_do_interrupt(int irq, void *data)
+{
+ struct ioatdma_device *instance = data;
+ struct ioat_dma_chan *ioat_chan;
+ unsigned long attnstatus;
+ int bit;
+ u8 intrctrl;
+
+ intrctrl = readb(instance->reg_base + IOAT_INTRCTRL_OFFSET);
+
+ if (!(intrctrl & IOAT_INTRCTRL_MASTER_INT_EN))
+ return IRQ_NONE;
+
+ if (!(intrctrl & IOAT_INTRCTRL_INT_STATUS)) {
+ writeb(intrctrl, instance->reg_base + IOAT_INTRCTRL_OFFSET);
+ return IRQ_NONE;
+ }
+
+ attnstatus = readl(instance->reg_base + IOAT_ATTNSTATUS_OFFSET);
+ for_each_bit(bit, &attnstatus, BITS_PER_LONG) {
+ ioat_chan = ioat_lookup_chan_by_index(instance, bit);
+ tasklet_schedule(&ioat_chan->cleanup_task);
+ }
+
+ writeb(intrctrl, instance->reg_base + IOAT_INTRCTRL_OFFSET);
+ return IRQ_HANDLED;
+}
+
+/**
+ * ioat_dma_do_interrupt_msix - handler used for vector-per-channel interrupt mode
+ * @irq: interrupt id
+ * @data: interrupt data
+ */
+static irqreturn_t ioat_dma_do_interrupt_msix(int irq, void *data)
+{
+ struct ioat_dma_chan *ioat_chan = data;
+
+ tasklet_schedule(&ioat_chan->cleanup_task);
+
+ return IRQ_HANDLED;
+}
+
+static void ioat_dma_cleanup_tasklet(unsigned long data);
+
+/**
+ * ioat_dma_enumerate_channels - find and initialize the device's channels
+ * @device: the device to be enumerated
+ */
static int ioat_dma_enumerate_channels(struct ioatdma_device *device)
{
u8 xfercap_scale;
@@ -76,6 +135,11 @@ static int ioat_dma_enumerate_channels(s
ioat_chan->common.device = &device->common;
list_add_tail(&ioat_chan->common.device_node,
&device->common.channels);
+ device->idx[i] = ioat_chan;
+ tasklet_init(&ioat_chan->cleanup_task,
+ ioat_dma_cleanup_tasklet,
+ (unsigned long) ioat_chan);
+ tasklet_disable(&ioat_chan->cleanup_task);
}
return device->common.chancnt;
}
@@ -234,6 +298,7 @@ static int ioat_dma_alloc_chan_resources
writel(((u64) ioat_chan->completion_addr) >> 32,
ioat_chan->reg_base + IOAT_CHANCMP_OFFSET_HIGH);
+ tasklet_enable(&ioat_chan->cleanup_task);
ioat_dma_start_null_desc(ioat_chan);
return i;
}
@@ -245,9 +310,14 @@ static void ioat_dma_free_chan_resources
struct ioat_desc_sw *desc, *_desc;
int in_use_descs = 0;
+ tasklet_disable(&ioat_chan->cleanup_task);
ioat_dma_memcpy_cleanup(ioat_chan);
+ /* Delay 100ms after reset to allow internal DMA logic to quiesce
+ * before removing DMA descriptor resources.
+ */
writeb(IOAT_CHANCMD_RESET, ioat_chan->reg_base + IOAT_CHANCMD_OFFSET);
+ mdelay(100);
spin_lock_bh(&ioat_chan->desc_lock);
list_for_each_entry_safe(desc, _desc, &ioat_chan->used_desc, node) {
@@ -276,6 +346,34 @@ static void ioat_dma_free_chan_resources
in_use_descs - 1);
ioat_chan->last_completion = ioat_chan->completion_addr = 0;
+ ioat_chan->pending = 0;
+}
+/**
+ * ioat_dma_get_next_descriptor - return the next available descriptor
+ * @ioat_chan: IOAT DMA channel handle
+ *
+ * Gets the next descriptor from the chain, and must be called with the
+ * channel's desc_lock held. Allocates more descriptors if the channel
+ * has run out.
+ */
+static struct ioat_desc_sw *ioat_dma_get_next_descriptor(
+ struct ioat_dma_chan *ioat_chan)
+{
+ struct ioat_desc_sw *new = NULL;
+
+ if (!list_empty(&ioat_chan->free_desc)) {
+ new = to_ioat_desc(ioat_chan->free_desc.next);
+ list_del(&new->node);
+ } else {
+ /* try to get another desc */
+ new = ioat_dma_alloc_descriptor(ioat_chan, GFP_ATOMIC);
+ /* will this ever happen? */
+ /* TODO add upper limit on these */
+ BUG_ON(!new);
+ }
+
+ prefetch(new->hw);
+ return new;
}
static struct dma_async_tx_descriptor *ioat_dma_prep_memcpy(
@@ -300,17 +398,7 @@ static struct dma_async_tx_descriptor *i
spin_lock_bh(&ioat_chan->desc_lock);
while (len) {
- if (!list_empty(&ioat_chan->free_desc)) {
- new = to_ioat_desc(ioat_chan->free_desc.next);
- list_del(&new->node);
- } else {
- /* try to get another desc */
- new = ioat_dma_alloc_descriptor(ioat_chan, GFP_ATOMIC);
- /* will this ever happen? */
- /* TODO add upper limit on these */
- BUG_ON(!new);
- }
-
+ new = ioat_dma_get_next_descriptor(ioat_chan);
copy = min((u32) len, ioat_chan->xfercap);
new->hw->size = copy;
@@ -360,6 +448,14 @@ static void ioat_dma_memcpy_issue_pendin
}
}
+static void ioat_dma_cleanup_tasklet(unsigned long data)
+{
+ struct ioat_dma_chan *chan = (void *)data;
+ ioat_dma_memcpy_cleanup(chan);
+ writew(IOAT_CHANCTRL_INT_DISABLE,
+ chan->reg_base + IOAT_CHANCTRL_OFFSET);
+}
+
static void ioat_dma_memcpy_cleanup(struct ioat_dma_chan *ioat_chan)
{
unsigned long phys_complete;
@@ -397,6 +493,7 @@ static void ioat_dma_memcpy_cleanup(stru
return;
}
+ cookie = 0;
spin_lock_bh(&ioat_chan->desc_lock);
list_for_each_entry_safe(desc, _desc, &ioat_chan->used_desc, node) {
@@ -509,48 +606,13 @@ static enum dma_status ioat_dma_is_compl
/* PCI API */
-static irqreturn_t ioat_do_interrupt(int irq, void *data)
-{
- struct ioatdma_device *instance = data;
- unsigned long attnstatus;
- u8 intrctrl;
-
- intrctrl = readb(instance->reg_base + IOAT_INTRCTRL_OFFSET);
-
- if (!(intrctrl & IOAT_INTRCTRL_MASTER_INT_EN))
- return IRQ_NONE;
-
- if (!(intrctrl & IOAT_INTRCTRL_INT_STATUS)) {
- writeb(intrctrl, instance->reg_base + IOAT_INTRCTRL_OFFSET);
- return IRQ_NONE;
- }
-
- attnstatus = readl(instance->reg_base + IOAT_ATTNSTATUS_OFFSET);
-
- printk(KERN_ERR "ioatdma: interrupt! status %lx\n", attnstatus);
-
- writeb(intrctrl, instance->reg_base + IOAT_INTRCTRL_OFFSET);
- return IRQ_HANDLED;
-}
-
static void ioat_dma_start_null_desc(struct ioat_dma_chan *ioat_chan)
{
struct ioat_desc_sw *desc;
spin_lock_bh(&ioat_chan->desc_lock);
- if (!list_empty(&ioat_chan->free_desc)) {
- desc = to_ioat_desc(ioat_chan->free_desc.next);
- list_del(&desc->node);
- } else {
- /* try to get another desc */
- spin_unlock_bh(&ioat_chan->desc_lock);
- desc = ioat_dma_alloc_descriptor(ioat_chan, GFP_KERNEL);
- spin_lock_bh(&ioat_chan->desc_lock);
- /* will this ever happen? */
- BUG_ON(!desc);
- }
-
+ desc = ioat_dma_get_next_descriptor(ioat_chan);
desc->hw->ctl = IOAT_DMA_DESCRIPTOR_NUL;
desc->hw->next = 0;
desc->async_tx.ack = 1;
@@ -571,7 +633,11 @@ static void ioat_dma_start_null_desc(str
*/
#define IOAT_TEST_SIZE 2000
-static int ioat_self_test(struct ioatdma_device *device)
+/**
+ * ioat_dma_self_test - Perform a IOAT transaction to verify the HW works.
+ * @device: device to be tested
+ */
+static int ioat_dma_self_test(struct ioatdma_device *device)
{
int i;
u8 *src;
@@ -639,6 +705,161 @@ out:
return err;
}
+static char ioat_interrupt_style[32] = "msix";
+module_param_string(ioat_interrupt_style, ioat_interrupt_style,
+ sizeof(ioat_interrupt_style), 0644);
+MODULE_PARM_DESC(ioat_interrupt_style,
+ "set ioat interrupt style: msix (default), "
+ "msix-single-vector, msi, intx)");
+
+/**
+ * ioat_dma_setup_interrupts - setup interrupt handler
+ * @device: ioat device
+ */
+int ioat_dma_setup_interrupts(struct ioatdma_device *device)
+{
+ struct ioat_dma_chan *ioat_chan;
+ int err, i, j, msixcnt;
+ u8 intrctrl = 0;
+
+ if (!strcmp(ioat_interrupt_style, "msix"))
+ goto msix;
+ if (!strcmp(ioat_interrupt_style, "msix-single-vector"))
+ goto msix_single_vector;
+ if (!strcmp(ioat_interrupt_style, "msi"))
+ goto msi;
+ if (!strcmp(ioat_interrupt_style, "intx"))
+ goto intx;
+
+msix:
+ /* The number of MSI-X vectors should equal the number of channels */
+ msixcnt = device->common.chancnt;
+ for (i = 0; i < msixcnt; i++)
+ device->msix_entries[i].entry = i;
+
+ err = pci_enable_msix(device->pdev, device->msix_entries, msixcnt);
+ if (err < 0)
+ goto msi;
+ if (err > 0)
+ goto msix_single_vector;
+
+ for (i = 0; i < msixcnt; i++) {
+ ioat_chan = ioat_lookup_chan_by_index(device, i);
+ err = request_irq(device->msix_entries[i].vector,
+ ioat_dma_do_interrupt_msix,
+ 0, "ioat-msix", ioat_chan);
+ if (err) {
+ for (j = 0; j < i; j++) {
+ ioat_chan =
+ ioat_lookup_chan_by_index(device, j);
+ free_irq(device->msix_entries[j].vector,
+ ioat_chan);
+ }
+ goto msix_single_vector;
+ }
+ }
+ intrctrl |= IOAT_INTRCTRL_MSIX_VECTOR_CONTROL;
+ device->irq_mode = msix_multi_vector;
+ goto done;
+
+msix_single_vector:
+ device->msix_entries[0].entry = 0;
+ err = pci_enable_msix(device->pdev, device->msix_entries, 1);
+ if (err)
+ goto msi;
+
+ err = request_irq(device->msix_entries[0].vector, ioat_dma_do_interrupt,
+ 0, "ioat-msix", device);
+ if (err) {
+ pci_disable_msix(device->pdev);
+ goto msi;
+ }
+ device->irq_mode = msix_single_vector;
+ goto done;
+
+msi:
+ err = pci_enable_msi(device->pdev);
+ if (err)
+ goto intx;
+
+ err = request_irq(device->pdev->irq, ioat_dma_do_interrupt,
+ 0, "ioat-msi", device);
+ if (err) {
+ pci_disable_msi(device->pdev);
+ goto intx;
+ }
+ /*
+ * CB 1.2 devices need a bit set in configuration space to enable MSI
+ */
+ if (device->version == IOAT_VER_1_2) {
+ u32 dmactrl;
+ pci_read_config_dword(device->pdev,
+ IOAT_PCI_DMACTRL_OFFSET, &dmactrl);
+ dmactrl |= IOAT_PCI_DMACTRL_MSI_EN;
+ pci_write_config_dword(device->pdev,
+ IOAT_PCI_DMACTRL_OFFSET, dmactrl);
+ }
+ device->irq_mode = msi;
+ goto done;
+
+intx:
+ err = request_irq(device->pdev->irq, ioat_dma_do_interrupt,
+ IRQF_SHARED, "ioat-intx", device);
+ if (err)
+ goto err_no_irq;
+ device->irq_mode = intx;
+
+done:
+ intrctrl |= IOAT_INTRCTRL_MASTER_INT_EN;
+ writeb(intrctrl, device->reg_base + IOAT_INTRCTRL_OFFSET);
+ return 0;
+
+err_no_irq:
+ /* Disable all interrupt generation */
+ writeb(0, device->reg_base + IOAT_INTRCTRL_OFFSET);
+ dev_err(&device->pdev->dev, "no usable interrupts\n");
+ device->irq_mode = none;
+ return -1;
+}
+
+/**
+ * ioat_dma_remove_interrupts - remove whatever interrupts were set
+ * @device: ioat device
+ */
+void ioat_dma_remove_interrupts(struct ioatdma_device *device)
+{
+ struct ioat_dma_chan *ioat_chan;
+ int i;
+
+ /* Disable all interrupt generation */
+ writeb(0, device->reg_base + IOAT_INTRCTRL_OFFSET);
+
+ switch (device->irq_mode) {
+ case msix_multi_vector:
+ for (i = 0; i < device->common.chancnt; i++) {
+ ioat_chan = ioat_lookup_chan_by_index(device, i);
+ free_irq(device->msix_entries[i].vector, ioat_chan);
+ }
+ pci_disable_msix(device->pdev);
+ break;
+ case msix_single_vector:
+ free_irq(device->msix_entries[0].vector, device);
+ pci_disable_msix(device->pdev);
+ break;
+ case msi:
+ free_irq(device->pdev->irq, device);
+ pci_disable_msi(device->pdev);
+ break;
+ case intx:
+ free_irq(device->pdev->irq, device);
+ break;
+ case none:
+ dev_warn(&device->pdev->dev,
+ "call to %s without interrupts setup\n", __func__);
+ }
+ device->irq_mode = none;
+}
+
struct ioatdma_device *ioat_dma_probe(struct pci_dev *pdev,
void __iomem *iobase)
{
@@ -684,21 +905,16 @@ struct ioatdma_device *ioat_dma_probe(st
device->common.device_issue_pending = ioat_dma_memcpy_issue_pending;
device->common.device_dependency_added = ioat_dma_dependency_added;
device->common.dev = &pdev->dev;
- printk(KERN_INFO "ioatdma: Intel(R) I/OAT DMA Engine found,"
- " %d channels, device version 0x%02x\n",
- device->common.chancnt, device->version);
-
- pci_set_drvdata(pdev, device);
- err = request_irq(pdev->irq, &ioat_do_interrupt, IRQF_SHARED, "ioat",
- device);
- if (err)
- goto err_irq;
+ dev_err(&device->pdev->dev,
+ "ioatdma: Intel(R) I/OAT DMA Engine found,"
+ " %d channels, device version 0x%02x\n",
+ device->common.chancnt, device->version);
- writeb(IOAT_INTRCTRL_MASTER_INT_EN,
- device->reg_base + IOAT_INTRCTRL_OFFSET);
- pci_set_master(pdev);
+ err = ioat_dma_setup_interrupts(device);
+ if (err)
+ goto err_setup_interrupts;
- err = ioat_self_test(device);
+ err = ioat_dma_self_test(device);
if (err)
goto err_self_test;
@@ -707,8 +923,8 @@ struct ioatdma_device *ioat_dma_probe(st
return device;
err_self_test:
- free_irq(device->pdev->irq, device);
-err_irq:
+ ioat_dma_remove_interrupts(device);
+err_setup_interrupts:
pci_pool_destroy(device->completion_pool);
err_completion_pool:
pci_pool_destroy(device->dma_pool);
@@ -716,8 +932,8 @@ err_dma_pool:
kfree(device);
err_kzalloc:
iounmap(iobase);
- printk(KERN_ERR
- "ioatdma: Intel(R) I/OAT DMA Engine initialization failed\n");
+ dev_err(&device->pdev->dev,
+ "ioatdma: Intel(R) I/OAT DMA Engine initialization failed\n");
return NULL;
}
@@ -728,7 +944,7 @@ void ioat_dma_remove(struct ioatdma_devi
dma_async_device_unregister(&device->common);
- free_irq(device->pdev->irq, device);
+ ioat_dma_remove_interrupts(device);
pci_pool_destroy(device->dma_pool);
pci_pool_destroy(device->completion_pool);
diff -puN drivers/dma/ioatdma.h~i-oat-add-support-for-msi-and-msi-x drivers/dma/ioatdma.h
--- a/drivers/dma/ioatdma.h~i-oat-add-support-for-msi-and-msi-x
+++ a/drivers/dma/ioatdma.h
@@ -28,6 +28,14 @@
#include <linux/cache.h>
#include <linux/pci_ids.h>
+enum ioat_interrupt {
+ none = 0,
+ msix_multi_vector = 1,
+ msix_single_vector = 2,
+ msi = 3,
+ intx = 4,
+};
+
#define IOAT_LOW_COMPLETION_MASK 0xffffffc0
/**
@@ -46,6 +54,9 @@ struct ioatdma_device {
struct pci_pool *completion_pool;
struct dma_device common;
u8 version;
+ enum ioat_interrupt irq_mode;
+ struct msix_entry msix_entries[4];
+ struct ioat_dma_chan *idx[4];
};
/**
@@ -94,6 +105,7 @@ struct ioat_dma_chan {
u32 high;
};
} *completion_virt;
+ struct tasklet_struct cleanup_task;
};
/* wrapper around hardware descriptor format + additional software fields */
diff -puN drivers/dma/ioatdma_registers.h~i-oat-add-support-for-msi-and-msi-x drivers/dma/ioatdma_registers.h
--- a/drivers/dma/ioatdma_registers.h~i-oat-add-support-for-msi-and-msi-x
+++ a/drivers/dma/ioatdma_registers.h
@@ -1,5 +1,5 @@
/*
- * Copyright(c) 2004 - 2006 Intel Corporation. All rights reserved.
+ * Copyright(c) 2004 - 2007 Intel Corporation. All rights reserved.
*
* 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
@@ -21,6 +21,9 @@
#ifndef _IOAT_REGISTERS_H_
#define _IOAT_REGISTERS_H_
+#define IOAT_PCI_DMACTRL_OFFSET 0x48
+#define IOAT_PCI_DMACTRL_DMA_EN 0x00000001
+#define IOAT_PCI_DMACTRL_MSI_EN 0x00000002
/* MMIO Device Registers */
#define IOAT_CHANCNT_OFFSET 0x00 /* 8-bit */
@@ -39,6 +42,7 @@
#define IOAT_INTRCTRL_MASTER_INT_EN 0x01 /* Master Interrupt Enable */
#define IOAT_INTRCTRL_INT_STATUS 0x02 /* ATTNSTATUS -or- Channel Int */
#define IOAT_INTRCTRL_INT 0x04 /* INT_STATUS -and- MASTER_INT_EN */
+#define IOAT_INTRCTRL_MSIX_VECTOR_CONTROL 0x08 /* Enable all MSI-X vectors */
#define IOAT_ATTNSTATUS_OFFSET 0x04 /* Each bit is a channel */
diff -puN include/linux/bitops.h~i-oat-add-support-for-msi-and-msi-x include/linux/bitops.h
--- a/include/linux/bitops.h~i-oat-add-support-for-msi-and-msi-x
+++ a/include/linux/bitops.h
@@ -8,6 +8,12 @@
*/
#include <asm/bitops.h>
+#define for_each_bit(bit, addr, size) \
+ for ((bit) = find_first_bit((addr), (size)); \
+ (bit) < (size); \
+ (bit) = find_next_bit((addr), (size), (bit) + 1))
+
+
static __inline__ int get_bitmask_order(unsigned int count)
{
int order;
_
Patches currently in -mm which might be from shannon.nelson@intel.com are
i-oat-new-device-ids.patch
i-oat-rename-the-source-file.patch
i-oat-code-cleanup-from-checkpatch-output.patch
i-oat-split-pci-startup-from-dma-handling-code.patch
i-oat-add-support-for-msi-and-msi-x.patch
dca-add-direct-cache-access-driver.patch
i-oat-add-dca-services.patch
remove-unsafe-from-module-struct.patch
reply other threads:[~2007-08-31 19:28 UTC|newest]
Thread overview: [no followups] expand[flat|nested] mbox.gz Atom feed
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=200708311928.l7VJS87Q022930@imap1.linux-foundation.org \
--to=akpm@linux-foundation.org \
--cc=davem@davemloft.net \
--cc=linux-kernel@vger.kernel.org \
--cc=mm-commits@vger.kernel.org \
--cc=shannon.nelson@intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.