* [PATCH V4] scsi: hpsa: drop shutdown callback
From: Sinan Kaya @ 2018-06-01 13:46 UTC (permalink / raw)
To: linux-arm-kernel
'Commit cc27b735ad3a ("PCI/portdrv: Turn off PCIe services during
shutdown")' has been added to kernel to shutdown pending PCIe port
service interrupts during reboot so that a newly started kexec kernel
wouldn't observe pending interrupts.
pcie_port_device_remove() is disabling the root port and switches by
calling pci_disable_device() after all PCIe service drivers are shutdown.
This has been found to cause crashes on HP DL360 Gen9 machines during
reboot due to hpsa driver not clearing the bus master bit during the
shutdown procedure by calling pci_disable_device().
Disable device as part of the shutdown sequence.
Signed-off-by: Sinan Kaya <okaya@codeaurora.org>
Link: https://bugzilla.kernel.org/show_bug.cgi?id=199779
Fixes: cc27b735ad3a ("PCI/portdrv: Turn off PCIe services during shutdown")
Cc: stable at vger.kernel.org
Reported-by: Ryan Finnie <ryan@finnie.org>
---
drivers/scsi/hpsa.c | 10 ++++++++--
1 file changed, 8 insertions(+), 2 deletions(-)
diff --git a/drivers/scsi/hpsa.c b/drivers/scsi/hpsa.c
index 3a9eca1..b92f86a 100644
--- a/drivers/scsi/hpsa.c
+++ b/drivers/scsi/hpsa.c
@@ -8869,7 +8869,7 @@ static void hpsa_disable_rld_caching(struct ctlr_info *h)
kfree(options);
}
-static void hpsa_shutdown(struct pci_dev *pdev)
+static void __hpsa_shutdown(struct pci_dev *pdev)
{
struct ctlr_info *h;
@@ -8884,6 +8884,12 @@ static void hpsa_shutdown(struct pci_dev *pdev)
hpsa_disable_interrupt_mode(h); /* pci_init 2 */
}
+static void hpsa_shutdown(struct pci_dev *pdev)
+{
+ __hpsa_shutdown(pdev);
+ pci_disable_device(pdev);
+}
+
static void hpsa_free_device_info(struct ctlr_info *h)
{
int i;
@@ -8927,7 +8933,7 @@ static void hpsa_remove_one(struct pci_dev *pdev)
scsi_remove_host(h->scsi_host); /* init_one 8 */
/* includes hpsa_free_irqs - init_one 4 */
/* includes hpsa_disable_interrupt_mode - pci_init 2 */
- hpsa_shutdown(pdev);
+ __hpsa_shutdown(pdev);
hpsa_free_device_info(h); /* scan */
--
2.7.4
^ permalink raw reply related
* [PATCH v2]irqchip/irq-gic-v3:Avoid a waste of LPI resource
From: Zhang, Lei @ 2018-06-01 13:47 UTC (permalink / raw)
To: linux-arm-kernel
In-Reply-To: <31bede40-dbc5-0a88-d12c-5fc79ea02391@arm.com>
Hi Marc
Thanks. Now I understood the policy of MSI allocation.
So this patch is great for our bus, and looks no problem for PCI.
Best Regards,
Lei Zhang zhang.lei at jp.fujitsu.com
FUJITSU LIMITED
> -----Original Message-----
> From: Marc Zyngier [mailto:marc.zyngier at arm.com]
> Sent: Friday, June 01, 2018 9:57 PM
> To: Zhang, Lei/? ?; 'linux-arm-kernel at lists.infradead.org'
> Subject: Re: [PATCH v2]irqchip/irq-gic-v3:Avoid a waste of LPI resource
>
> Hi Lei,
>
> On 01/06/18 13:44, Zhang, Lei wrote:
> > Hi Marc
> >
> > I have reviewed your patch.
> > I think the approach is same between your patch and mine.
> > Your patch is simpler and more beautiful, and match our bus's
> requirement.
> >
> > I have only one question.
> > According your patch, if there are no enough lpis, amount of lpis
> required
> > will be divide by 2. it means someone want 16 lpis, maybe they can only
> get 8?
> > I don?t' understand why we need it.
>
> That's the way the MSI allocation works in the kernel.
>
> A driver asks a number of MSIs (let's imagine, for example, one MSI per
> CPU in the system), but the underlying HW can only provide a smaller
> number.
>
> Instead of failing and just returning an error, we reduce the allocation
> in order to provide the driver with something. If that's not enough,
> well, the driver itself will have the opportunity to give up.
>
> See pci_alloc_irq_vectors_affinity(), which takes a min and a max number
> of vectors, for example.
>
> Thanks,
>
> M.
> --
> Jazz is not dead. It just smells funny...
>
^ permalink raw reply
* [PATCH V4] scsi: hpsa: drop shutdown callback
From: Sinan Kaya @ 2018-06-01 13:47 UTC (permalink / raw)
To: linux-arm-kernel
In-Reply-To: <1527860768-11367-1-git-send-email-okaya@codeaurora.org>
On 6/1/2018 9:46 AM, Sinan Kaya wrote:
> 'Commit cc27b735ad3a ("PCI/portdrv: Turn off PCIe services during
> shutdown")' has been added to kernel to shutdown pending PCIe port
> service interrupts during reboot so that a newly started kexec kernel
> wouldn't observe pending interrupts.
>
> pcie_port_device_remove() is disabling the root port and switches by
> calling pci_disable_device() after all PCIe service drivers are shutdown.
>
> This has been found to cause crashes on HP DL360 Gen9 machines during
> reboot due to hpsa driver not clearing the bus master bit during the
> shutdown procedure by calling pci_disable_device().
>
> Disable device as part of the shutdown sequence.
Forgot to change the commit title. Apologies for that. The rest of the
patch should work. Waiting for tested-bys.
>
> Signed-off-by: Sinan Kaya <okaya@codeaurora.org>
> Link: https://bugzilla.kernel.org/show_bug.cgi?id=199779
> Fixes: cc27b735ad3a ("PCI/portdrv: Turn off PCIe services during shutdown")
> Cc: stable at vger.kernel.org
> Reported-by: Ryan Finnie <ryan@finnie.org>
> ---
> drivers/scsi/hpsa.c | 10 ++++++++--
> 1 file changed, 8 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/scsi/hpsa.c b/drivers/scsi/hpsa.c
> index 3a9eca1..b92f86a 100644
> --- a/drivers/scsi/hpsa.c
> +++ b/drivers/scsi/hpsa.c
> @@ -8869,7 +8869,7 @@ static void hpsa_disable_rld_caching(struct ctlr_info *h)
> kfree(options);
> }
>
> -static void hpsa_shutdown(struct pci_dev *pdev)
> +static void __hpsa_shutdown(struct pci_dev *pdev)
> {
> struct ctlr_info *h;
>
> @@ -8884,6 +8884,12 @@ static void hpsa_shutdown(struct pci_dev *pdev)
> hpsa_disable_interrupt_mode(h); /* pci_init 2 */
> }
>
> +static void hpsa_shutdown(struct pci_dev *pdev)
> +{
> + __hpsa_shutdown(pdev);
> + pci_disable_device(pdev);
> +}
> +
> static void hpsa_free_device_info(struct ctlr_info *h)
> {
> int i;
> @@ -8927,7 +8933,7 @@ static void hpsa_remove_one(struct pci_dev *pdev)
> scsi_remove_host(h->scsi_host); /* init_one 8 */
> /* includes hpsa_free_irqs - init_one 4 */
> /* includes hpsa_disable_interrupt_mode - pci_init 2 */
> - hpsa_shutdown(pdev);
> + __hpsa_shutdown(pdev);
>
> hpsa_free_device_info(h); /* scan */
>
>
--
Sinan Kaya
Qualcomm Datacenter Technologies, Inc. as an affiliate of Qualcomm Technologies, Inc.
Qualcomm Technologies, Inc. is a member of the Code Aurora Forum, a Linux Foundation Collaborative Project.
^ permalink raw reply
* [PATCHv2 0/4] arm64: SMCCC conduit cleanup
From: Catalin Marinas @ 2018-06-01 14:14 UTC (permalink / raw)
To: linux-arm-kernel
In-Reply-To: <20180531173223.9668-1-mark.rutland@arm.com>
On Thu, May 31, 2018 at 06:32:19PM +0100, Mark Rutland wrote:
> Mark Rutland (4):
> arm/arm64: smccc/psci: add arm_smccc_1_1_get_conduit()
> arm64: errata: use arm_smccc_get_conduit()
> firmware/psci: use common SMCCC_CONDUIT_*
> firmware: arm_sdei: use common SMCCC_CONDUIT_*
Queued for 4.18. Thanks.
--
Catalin
^ permalink raw reply
* [PATCH v2 3/6] clk: ti: dra7: Add clkctrl clock data for the mcan clocks
From: Tony Lindgren @ 2018-06-01 14:26 UTC (permalink / raw)
To: linux-arm-kernel
In-Reply-To: <d7532c7a-d2ae-5cbf-cfd0-c347b975ed6b@ti.com>
* Faiz Abbas <faiz_abbas@ti.com> [180601 06:49]:
> Hi,
>
> On Thursday 31 May 2018 06:59 PM, Tero Kristo wrote:
> > On 31/05/18 13:14, Faiz Abbas wrote:
> >> Hi,
> >>
> >> On Thursday 31 May 2018 09:33 AM, Rob Herring wrote:
> >>> On Wed, May 30, 2018 at 07:41:30PM +0530, Faiz Abbas wrote:
> >>>> Add clkctrl data for the m_can clocks and register it within the
> >> ...
> >>>> ? diff --git a/include/dt-bindings/clock/dra7.h
> >>>> b/include/dt-bindings/clock/dra7.h
> >>>> index 5e1061b15aed..d7549c57cac3 100644
> >>>> --- a/include/dt-bindings/clock/dra7.h
> >>>> +++ b/include/dt-bindings/clock/dra7.h
> >>>> @@ -168,5 +168,6 @@
> >>>> ? #define DRA7_COUNTER_32K_CLKCTRL??? DRA7_CLKCTRL_INDEX(0x50)
> >>>> ? #define DRA7_UART10_CLKCTRL??? DRA7_CLKCTRL_INDEX(0x80)
> >>>> ? #define DRA7_DCAN1_CLKCTRL??? DRA7_CLKCTRL_INDEX(0x88)
> >>>> +#define DRA7_ADC_CLKCTRL??? DRA7_CLKCTRL_INDEX(0xa0)
> >>>
> >>> ADC and mcan are the same thing?
> >>>
> >>
> >> The register to control MCAN clocks is called ADC_CLKCTRL, Yes.
> >
> > Is there any reason for this or is that just a documentation bug?
> >
>
> Looks like they meant to have an ADC in dra74 or dra72 but decided
> against it and then many years later used the same registers for MCAN
> instead. You can see ADC_CLKCTRL exists in dra72 TRM but is explicitly
> disabled.
>
> http://www.ti.com/lit/ug/spruic2b/spruic2b.pdf pg:1524
How about make add also something like to dra7.h:
#define DRA7_MCAN_CLKCTRL DRA7_ADC_CLKCTRL
And you can add a comment to the dts file to avoid people
getting confused with this constantly.
Regards,
Tony
^ permalink raw reply
* [PATCHv2 0/4] arm64: SMCCC conduit cleanup
From: Russell King - ARM Linux @ 2018-06-01 14:26 UTC (permalink / raw)
To: linux-arm-kernel
In-Reply-To: <20180601141410.e7dwvmqiqi2esbnk@armageddon.cambridge.arm.com>
On Fri, Jun 01, 2018 at 03:14:10PM +0100, Catalin Marinas wrote:
> On Thu, May 31, 2018 at 06:32:19PM +0100, Mark Rutland wrote:
> > Mark Rutland (4):
> > arm/arm64: smccc/psci: add arm_smccc_1_1_get_conduit()
> > arm64: errata: use arm_smccc_get_conduit()
> > firmware/psci: use common SMCCC_CONDUIT_*
> > firmware: arm_sdei: use common SMCCC_CONDUIT_*
>
> Queued for 4.18. Thanks.
Have you considered how this affects the 32-bit ARM spectre stuff, which
makes use of the firmware PSCI interfaces for the workaround?
--
RMK's Patch system: http://www.armlinux.org.uk/developer/patches/
FTTC broadband for 0.8mile line in suburbia: sync at 8.8Mbps down 630kbps up
According to speedtest.net: 8.21Mbps down 510kbps up
^ permalink raw reply
* [PATCH v2 3/4] dt-bindings: soc: Add TmFifo binding for Mellanox BlueField SoC
From: Liming Sun @ 2018-06-01 14:31 UTC (permalink / raw)
To: linux-arm-kernel
In-Reply-To: <20180531034352.GA32666@rob-hp-laptop>
Thanks for the comments. Please see response inline.
> -----Original Message-----
> From: Rob Herring [mailto:robh at kernel.org]
> Sent: Wednesday, May 30, 2018 11:44 PM
> To: Liming Sun <lsun@mellanox.com>
> Cc: Olof Johansson <olof@lixom.net>; Arnd Bergmann <arnd@arndb.de>;
> David Woods <dwoods@mellanox.com>; Robin Murphy
> <robin.murphy@arm.com>; devicetree at vger.kernel.org; linux-arm-
> kernel at lists.infradead.org
> Subject: Re: [PATCH v2 3/4] dt-bindings: soc: Add TmFifo binding for
> Mellanox BlueField SoC
>
> On Fri, May 25, 2018 at 04:17:15PM -0400, Liming Sun wrote:
>
> Commit msg?
Updated in patch v3 3/4.
>
> > Reviewed-by: David Woods <dwoods@mellanox.com>
> > Signed-off-by: Liming Sun <lsun@mellanox.com>
> > ---
> > .../devicetree/bindings/soc/mellanox/tmfifo.txt | 20
> ++++++++++++++++++++
> > 1 file changed, 20 insertions(+)
> > create mode 100644
> Documentation/devicetree/bindings/soc/mellanox/tmfifo.txt
> >
> > diff --git a/Documentation/devicetree/bindings/soc/mellanox/tmfifo.txt
> b/Documentation/devicetree/bindings/soc/mellanox/tmfifo.txt
> > new file mode 100644
> > index 0000000..0a362f5
> > --- /dev/null
> > +++ b/Documentation/devicetree/bindings/soc/mellanox/tmfifo.txt
> > @@ -0,0 +1,20 @@
> > +* Mellanox BlueField SoC TmFifo
> > +
> > +BlueField TmFifo provides a shared FIFO between the target and the
> > +external host machine, which can be accessed via USB or PCIe.
>
> A FIFO for what? I'd like to find a better spot than bindings/soc/
This is a generic HW FIFO which can be accessed by the SoC SW and external Host machine to exchange any data. In the current Linux tmfifo driver, this FIFO has been used (demuxed) to implement a virtual console and network interface based on virtio framework.
Updated the tmfifo.txt in patch v3 3/4 with the above explanation. Please advise if there is a better place for this file.
> > +
> > +Required properties:
> > +
> > +- compatible: Should be "mellanox,bf-tmfifo"
> > +- reg: Physical base address and length of Rx/Tx block
> > +- interrupts: The interrupt number of Rx low water mark, Rx high water
> mark
> > + Tx low water mark, Tx high water mark respectively.
> > +
> > +Example:
> > +
> > +tmfifo at 800a20 {
> > + compatible = "mellanox,bf-tmfifo";
> > + reg = <0x00800a20 0x00000018
> > + 0x00800a40 0x00000018>;
> > + interrupts = <41, 42, 43, 44>;
> > +};
> > --
> > 1.8.3.1
> >
> > --
> > To unsubscribe from this list: send the line "unsubscribe devicetree" in
> > the body of a message to majordomo at vger.kernel.org
> > More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply
* [PATCH v3 1/4] soc: Add TmFifo driver for Mellanox BlueField Soc
From: Liming Sun @ 2018-06-01 14:31 UTC (permalink / raw)
To: linux-arm-kernel
In-Reply-To: <b143b40446c1870fb8d422b364ead95d54552be9.1527264077.git.lsun@mellanox.com>
This commit adds the TmFifo driver for Mellanox BlueField Soc.
TmFifo is a shared FIFO which enables external host machine to
exchange data with the SoC via USB or PCIe. The driver is based on
virtio framework and has console and network access enabled.
Reviewed-by: David Woods <dwoods@mellanox.com>
Signed-off-by: Liming Sun <lsun@mellanox.com>
---
drivers/soc/Kconfig | 1 +
drivers/soc/Makefile | 1 +
drivers/soc/mellanox/Kconfig | 18 +
drivers/soc/mellanox/Makefile | 5 +
drivers/soc/mellanox/tmfifo.c | 1239 ++++++++++++++++++++++++++++++++++++
drivers/soc/mellanox/tmfifo_regs.h | 75 +++
6 files changed, 1339 insertions(+)
create mode 100644 drivers/soc/mellanox/Kconfig
create mode 100644 drivers/soc/mellanox/Makefile
create mode 100644 drivers/soc/mellanox/tmfifo.c
create mode 100644 drivers/soc/mellanox/tmfifo_regs.h
diff --git a/drivers/soc/Kconfig b/drivers/soc/Kconfig
index c07b4a8..fa87dc8 100644
--- a/drivers/soc/Kconfig
+++ b/drivers/soc/Kconfig
@@ -7,6 +7,7 @@ source "drivers/soc/bcm/Kconfig"
source "drivers/soc/fsl/Kconfig"
source "drivers/soc/imx/Kconfig"
source "drivers/soc/mediatek/Kconfig"
+source "drivers/soc/mellanox/Kconfig"
source "drivers/soc/qcom/Kconfig"
source "drivers/soc/renesas/Kconfig"
source "drivers/soc/rockchip/Kconfig"
diff --git a/drivers/soc/Makefile b/drivers/soc/Makefile
index 4052357..868163f 100644
--- a/drivers/soc/Makefile
+++ b/drivers/soc/Makefile
@@ -13,6 +13,7 @@ obj-$(CONFIG_ARCH_GEMINI) += gemini/
obj-$(CONFIG_ARCH_MXC) += imx/
obj-$(CONFIG_SOC_XWAY) += lantiq/
obj-y += mediatek/
+obj-$(CONFIG_SOC_MLNX) += mellanox/
obj-$(CONFIG_ARCH_MESON) += amlogic/
obj-$(CONFIG_ARCH_QCOM) += qcom/
obj-y += renesas/
diff --git a/drivers/soc/mellanox/Kconfig b/drivers/soc/mellanox/Kconfig
new file mode 100644
index 0000000..d88efa1
--- /dev/null
+++ b/drivers/soc/mellanox/Kconfig
@@ -0,0 +1,18 @@
+menuconfig SOC_MLNX
+ bool "Mellanox SoC drivers"
+ default y if ARCH_MLNX_BLUEFIELD
+
+if ARCH_MLNX_BLUEFIELD || COMPILE_TEST
+
+config MLNX_BLUEFIELD_TMFIFO
+ tristate "Mellanox BlueField SoC TmFifo driver"
+ depends on ARM64
+ default m
+ select VIRTIO_CONSOLE
+ select VIRTIO_NET
+ help
+ Say y here to enable TmFifo support. The TmFifo driver provides the
+ virtio driver framework for the TMFIFO of Mellanox BlueField SoC and
+ the implementation of a console and network driver.
+
+endif # ARCH_MLNX_BLUEFIELD
diff --git a/drivers/soc/mellanox/Makefile b/drivers/soc/mellanox/Makefile
new file mode 100644
index 0000000..c44c0e2
--- /dev/null
+++ b/drivers/soc/mellanox/Makefile
@@ -0,0 +1,5 @@
+# SPDX-License-Identifier: GPL-2.0
+#
+# Makefile for Mellanox SoC drivers.
+#
+obj-$(CONFIG_MLNX_BLUEFIELD_TMFIFO) += tmfifo.o
diff --git a/drivers/soc/mellanox/tmfifo.c b/drivers/soc/mellanox/tmfifo.c
new file mode 100644
index 0000000..5647cb6
--- /dev/null
+++ b/drivers/soc/mellanox/tmfifo.c
@@ -0,0 +1,1239 @@
+/*
+ * Copyright (c) 2018, Mellanox Technologies. 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 version 2 and
+ * only version 2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ */
+
+#include <linux/acpi.h>
+#include <linux/bitfield.h>
+#include <linux/cache.h>
+#include <linux/device.h>
+#include <linux/dma-mapping.h>
+#include <linux/efi.h>
+#include <linux/io.h>
+#include <linux/interrupt.h>
+#include <linux/irq.h>
+#include <linux/kernel.h>
+#include <linux/math64.h>
+#include <linux/module.h>
+#include <linux/moduleparam.h>
+#include <linux/mutex.h>
+#include <linux/platform_device.h>
+#include <linux/resource.h>
+#include <linux/slab.h>
+#include <linux/types.h>
+#include <linux/version.h>
+#include <linux/virtio.h>
+#include <linux/virtio_config.h>
+#include <linux/virtio_console.h>
+#include <linux/virtio_ids.h>
+#include <linux/virtio_net.h>
+#include <linux/virtio_ring.h>
+#include <asm/byteorder.h>
+
+#include "tmfifo_regs.h"
+
+#define TMFIFO_GET_FIELD(reg, mask) FIELD_GET(mask, reg)
+
+#define TMFIFO_SET_FIELD(reg, mask, value) \
+ ((reg & ~mask) | FIELD_PREP(mask, value))
+
+/* Vring size. */
+#define TMFIFO_VRING_SIZE 1024
+
+/* Console Tx buffer size. */
+#define TMFIFO_CONS_TX_BUF_SIZE (32 * 1024)
+
+/* Use a timer for house-keeping. */
+static int tmfifo_timer_interval = HZ / 10;
+
+/* Global lock. */
+static struct mutex tmfifo_lock;
+
+/* Virtio ring size. */
+static int tmfifo_vring_size = TMFIFO_VRING_SIZE;
+module_param(tmfifo_vring_size, int, 0444);
+MODULE_PARM_DESC(tmfifo_vring_size, "Size of the vring.");
+
+struct tmfifo;
+
+/* A flag to indicate TmFifo ready. */
+static bool tmfifo_ready;
+
+/* Virtual devices sharing the TM FIFO. */
+#define TMFIFO_VDEV_MAX (VIRTIO_ID_CONSOLE + 1)
+
+/* Spin lock. */
+static DEFINE_SPINLOCK(tmfifo_spin_lock);
+
+/* Structure to maintain the ring state. */
+struct tmfifo_vring {
+ void *va; /* virtual address */
+ dma_addr_t dma; /* dma address */
+ struct virtqueue *vq; /* virtqueue pointer */
+ struct vring_desc *desc; /* current desc */
+ struct vring_desc *desc_head; /* current desc head */
+ int cur_len; /* processed len in current desc */
+ int rem_len; /* remaining length to be processed */
+ int size; /* vring size */
+ int align; /* vring alignment */
+ int id; /* vring id */
+ int vdev_id; /* TMFIFO_VDEV_xxx */
+ u32 pkt_len; /* packet total length */
+ __virtio16 next_avail; /* next avail desc id */
+ struct tmfifo *fifo; /* pointer back to the tmfifo */
+};
+
+/* Interrupt types. */
+enum {
+ TM_RX_LWM_IRQ, /* Rx low water mark irq */
+ TM_RX_HWM_IRQ, /* Rx high water mark irq */
+ TM_TX_LWM_IRQ, /* Tx low water mark irq */
+ TM_TX_HWM_IRQ, /* Tx high water mark irq */
+ TM_IRQ_CNT
+};
+
+/* Ring types (Rx & Tx). */
+enum {
+ TMFIFO_VRING_RX, /* Rx ring */
+ TMFIFO_VRING_TX, /* Tx ring */
+ TMFIFO_VRING_NUM
+};
+
+struct tmfifo_vdev {
+ struct virtio_device vdev; /* virtual device */
+ u8 status;
+ u64 features;
+ union { /* virtio config space */
+ struct virtio_console_config cons;
+ struct virtio_net_config net;
+ } config;
+ struct tmfifo_vring vrings[TMFIFO_VRING_NUM];
+ u8 *tx_buf; /* tx buffer */
+ u32 tx_head; /* tx buffer head */
+ u32 tx_tail; /* tx buffer tail */
+};
+
+#define TMFIFO_VDEV_TX_BUF_AVAIL(vdev) \
+ (((vdev)->tx_tail >= (vdev)->tx_head) ? \
+ (TMFIFO_CONS_TX_BUF_SIZE - 8 - ((vdev)->tx_tail - (vdev)->tx_head)) : \
+ ((vdev)->tx_head - (vdev)->tx_tail - 8))
+
+#define TMFIFO_VDEV_TX_BUF_PUSH(vdev, len) do { \
+ (vdev)->tx_tail += (len); \
+ if ((vdev)->tx_tail >= TMFIFO_CONS_TX_BUF_SIZE) \
+ (vdev)->tx_tail -= TMFIFO_CONS_TX_BUF_SIZE; \
+} while (0)
+
+#define TMFIFO_VDEV_TX_BUF_POP(vdev, len) do { \
+ (vdev)->tx_head += (len); \
+ if ((vdev)->tx_head >= TMFIFO_CONS_TX_BUF_SIZE) \
+ (vdev)->tx_head -= TMFIFO_CONS_TX_BUF_SIZE; \
+} while (0)
+
+/* TMFIFO device structure */
+struct tmfifo {
+ struct tmfifo_vdev *vdev[TMFIFO_VDEV_MAX]; /* virtual devices */
+ struct platform_device *pdev; /* platform device */
+ struct mutex lock;
+ void __iomem *rx_base; /* mapped register base */
+ void __iomem *tx_base; /* mapped register base */
+ int tx_fifo_size; /* number of entries of the Tx FIFO */
+ int rx_fifo_size; /* number of entries of the Rx FIFO */
+ unsigned long pend_events; /* pending bits for deferred process */
+ int irq[TM_IRQ_CNT]; /* irq numbers */
+ struct work_struct work; /* work struct for deferred process */
+ struct timer_list timer; /* keepalive timer */
+ struct tmfifo_vring *vring[2]; /* current Tx/Rx ring */
+};
+
+union tmfifo_msg_hdr {
+ struct {
+ u8 type; /* message type */
+ __be16 len; /* payload length */
+ u8 unused[5]; /* reserved, set to 0 */
+ } __packed;
+ u64 data;
+};
+
+/*
+ * Default MAC.
+ * This MAC address will be read from EFI persistent variable if configured.
+ * It can also be reconfigured with standard Linux tools.
+ */
+static u8 tmfifo_net_default_mac[6] = {0x00, 0x1A, 0xCA, 0xFF, 0xFF, 0x01};
+
+/* MTU setting of the virtio-net interface. */
+#define TMFIFO_NET_MTU 1500
+
+/* Supported virtio-net features. */
+#define TMFIFO_NET_FEATURES ((1UL << VIRTIO_NET_F_MTU) | \
+ (1UL << VIRTIO_NET_F_STATUS) | \
+ (1UL << VIRTIO_NET_F_MAC))
+
+/* Forward declaration. */
+static void tmfifo_virtio_rxtx(struct virtqueue *vq, bool is_rx);
+static void tmfifo_release_pkt(struct virtio_device *vdev,
+ struct tmfifo_vring *vring,
+ struct vring_desc **desc);
+
+/* Allocate vrings for the fifo. */
+static int tmfifo_alloc_vrings(struct tmfifo *fifo,
+ struct tmfifo_vdev *tm_vdev, int vdev_id)
+{
+ dma_addr_t dma;
+ void *va;
+ int i, size;
+ struct tmfifo_vring *vring;
+
+ for (i = 0; i < ARRAY_SIZE(tm_vdev->vrings); i++) {
+ vring = &tm_vdev->vrings[i];
+ vring->fifo = fifo;
+ vring->size = tmfifo_vring_size;
+ vring->align = SMP_CACHE_BYTES;
+ vring->id = i;
+ vring->vdev_id = vdev_id;
+
+ size = PAGE_ALIGN(vring_size(vring->size, vring->align));
+ va = dma_alloc_coherent(tm_vdev->vdev.dev.parent, size, &dma,
+ GFP_KERNEL);
+ if (!va) {
+ dev_err(tm_vdev->vdev.dev.parent,
+ "vring allocation failed\n");
+ return -EINVAL;
+ }
+
+ vring->va = va;
+ vring->dma = dma;
+ }
+
+ return 0;
+}
+
+/* Free vrings of the fifo device. */
+static void tmfifo_free_vrings(struct tmfifo *fifo, int vdev_id)
+{
+ int i, size;
+ struct tmfifo_vring *vring;
+ struct tmfifo_vdev *tm_vdev = fifo->vdev[vdev_id];
+
+ for (i = 0; i < ARRAY_SIZE(tm_vdev->vrings); i++) {
+ vring = &tm_vdev->vrings[i];
+
+ size = PAGE_ALIGN(vring_size(vring->size, vring->align));
+ if (vring->va) {
+ dma_free_coherent(tm_vdev->vdev.dev.parent, size,
+ vring->va, vring->dma);
+ vring->va = NULL;
+ if (vring->vq) {
+ vring_del_virtqueue(vring->vq);
+ vring->vq = NULL;
+ }
+ }
+ }
+}
+
+/* Free interrupts of the fifo device. */
+static void tmfifo_free_irqs(struct tmfifo *fifo)
+{
+ int i, irq;
+
+ for (i = 0; i < TM_IRQ_CNT; i++) {
+ irq = fifo->irq[i];
+ if (irq) {
+ fifo->irq[i] = 0;
+ disable_irq(irq);
+ free_irq(irq, (u8 *)fifo + i);
+ }
+ }
+}
+
+/* Work handler for Rx, Tx or activity monitoring. */
+static void tmfifo_work_handler(struct work_struct *work)
+{
+ int i;
+ struct tmfifo_vdev *tm_vdev;
+ struct tmfifo *fifo = container_of(work, struct tmfifo, work);
+
+ if (!tmfifo_ready)
+ return;
+
+ mutex_lock(&fifo->lock);
+
+ /* Tx. */
+ if (test_and_clear_bit(TM_TX_LWM_IRQ, &fifo->pend_events) &&
+ fifo->irq[TM_TX_LWM_IRQ]) {
+ for (i = 0; i < TMFIFO_VDEV_MAX; i++) {
+ tm_vdev = fifo->vdev[i];
+ if (tm_vdev != NULL) {
+ tmfifo_virtio_rxtx(
+ tm_vdev->vrings[TMFIFO_VRING_TX].vq,
+ false);
+ }
+ }
+ }
+
+ /* Rx. */
+ if (test_and_clear_bit(TM_RX_HWM_IRQ, &fifo->pend_events) &&
+ fifo->irq[TM_RX_HWM_IRQ]) {
+ for (i = 0; i < TMFIFO_VDEV_MAX; i++) {
+ tm_vdev = fifo->vdev[i];
+ if (tm_vdev != NULL) {
+ tmfifo_virtio_rxtx(
+ tm_vdev->vrings[TMFIFO_VRING_RX].vq,
+ true);
+ }
+ }
+ }
+
+ mutex_unlock(&fifo->lock);
+}
+
+/* Interrupt handler. */
+static irqreturn_t tmfifo_irq_handler(int irq, void *dev_id)
+{
+ int i = (uintptr_t)dev_id % sizeof(void *);
+ struct tmfifo *fifo = dev_id - i;
+
+ if (i < TM_IRQ_CNT && !test_and_set_bit(i, &fifo->pend_events))
+ schedule_work(&fifo->work);
+
+ return IRQ_HANDLED;
+}
+
+/* Nothing to do for now. */
+static void tmfifo_virtio_dev_release(struct device *dev)
+{
+}
+
+/* Get the next packet descriptor from the vring. */
+static inline struct vring_desc *
+tmfifo_virtio_get_next_desc(struct virtqueue *vq)
+{
+ unsigned int idx, head;
+ struct vring *vr = (struct vring *)virtqueue_get_vring(vq);
+ struct tmfifo_vring *vring = (struct tmfifo_vring *)vq->priv;
+
+ if (!vr || vring->next_avail == vr->avail->idx)
+ return NULL;
+
+ idx = vring->next_avail % vr->num;
+ head = vr->avail->ring[idx];
+ BUG_ON(head >= vr->num);
+ vring->next_avail++;
+ return &vr->desc[head];
+}
+
+static inline void tmfifo_virtio_release_desc(struct virtio_device *vdev,
+ struct vring *vr,
+ struct vring_desc *desc, u32 len)
+{
+ unsigned int idx;
+
+ idx = vr->used->idx % vr->num;
+ vr->used->ring[idx].id = desc - vr->desc;
+ vr->used->ring[idx].len = cpu_to_virtio32(vdev, len);
+
+ /* Virtio could poll and check the 'idx' to decide
+ * whether the desc is done or not. Add a memory
+ * barrier here to make sure the update above completes
+ * before updating the idx.
+ */
+ mb();
+ vr->used->idx++;
+}
+
+/* Get the total length of a descriptor chain. */
+static inline u32 tmfifo_virtio_get_pkt_len(struct virtio_device *vdev,
+ struct vring_desc *desc, struct vring *vr)
+{
+ u32 len = 0, idx;
+
+ while (desc) {
+ len += virtio32_to_cpu(vdev, desc->len);
+ if (!(virtio16_to_cpu(vdev, desc->flags) & VRING_DESC_F_NEXT))
+ break;
+ idx = virtio16_to_cpu(vdev, desc->next);
+ desc = &vr->desc[idx];
+ }
+
+ return len;
+}
+
+static void tmfifo_release_pkt(struct virtio_device *vdev,
+ struct tmfifo_vring *vring,
+ struct vring_desc **desc)
+{
+ struct vring *vr = (struct vring *)virtqueue_get_vring(vring->vq);
+ struct vring_desc *desc_head;
+ uint32_t pkt_len = 0;
+
+ if (!vr)
+ return;
+
+ if (desc != NULL && *desc != NULL && vring->desc_head != NULL) {
+ desc_head = vring->desc_head;
+ pkt_len = vring->pkt_len;
+ } else {
+ desc_head = tmfifo_virtio_get_next_desc(vring->vq);
+ if (desc_head != NULL) {
+ pkt_len = tmfifo_virtio_get_pkt_len(vdev,
+ desc_head, vr);
+ }
+ }
+
+ if (desc_head != NULL)
+ tmfifo_virtio_release_desc(vdev, vr, desc_head, pkt_len);
+
+ if (desc != NULL)
+ *desc = NULL;
+ vring->pkt_len = 0;
+}
+
+/* House-keeping timer. */
+static void tmfifo_timer(struct timer_list *arg)
+{
+ struct tmfifo *fifo = container_of(arg, struct tmfifo, timer);
+
+ /*
+ * Wake up the work handler to poll the Rx FIFO in case interrupt
+ * missing or any leftover bytes stuck in the FIFO.
+ */
+ test_and_set_bit(TM_RX_HWM_IRQ, &fifo->pend_events);
+
+ /*
+ * Wake up Tx handler in case virtio has queued too many packets
+ * and are waiting for buffer return.
+ */
+ test_and_set_bit(TM_TX_LWM_IRQ, &fifo->pend_events);
+
+ schedule_work(&fifo->work);
+
+ mod_timer(&fifo->timer, jiffies + tmfifo_timer_interval);
+}
+
+/* Buffer the console output. */
+static void tmfifo_console_output(struct tmfifo_vdev *cons,
+ struct virtqueue *vq)
+{
+ u32 len, pkt_len, idx;
+ struct vring_desc *head_desc, *desc = NULL;
+ struct vring *vr = (struct vring *)virtqueue_get_vring(vq);
+ struct virtio_device *vdev = &cons->vdev;
+ void *addr;
+ union tmfifo_msg_hdr *hdr;
+
+ for (;;) {
+ head_desc = tmfifo_virtio_get_next_desc(vq);
+ if (head_desc == NULL)
+ break;
+
+ /* Release the packet if no more space. */
+ pkt_len = tmfifo_virtio_get_pkt_len(vdev, head_desc, vr);
+ if (pkt_len + sizeof(*hdr) > TMFIFO_VDEV_TX_BUF_AVAIL(cons)) {
+ tmfifo_virtio_release_desc(vdev, vr, head_desc,
+ pkt_len);
+ break;
+ }
+
+ hdr = (union tmfifo_msg_hdr *)&cons->tx_buf[cons->tx_tail];
+ hdr->data = 0;
+ hdr->type = VIRTIO_ID_CONSOLE;
+ hdr->len = htons(pkt_len);
+
+ TMFIFO_VDEV_TX_BUF_PUSH(cons, sizeof(*hdr));
+ desc = head_desc;
+
+ while (desc != NULL) {
+ addr = phys_to_virt(virtio64_to_cpu(vdev, desc->addr));
+ len = virtio32_to_cpu(vdev, desc->len);
+
+ if (len <= TMFIFO_CONS_TX_BUF_SIZE - cons->tx_tail) {
+ memcpy(cons->tx_buf + cons->tx_tail, addr, len);
+ } else {
+ u32 seg;
+
+ seg = TMFIFO_CONS_TX_BUF_SIZE - cons->tx_tail;
+ memcpy(cons->tx_buf + cons->tx_tail, addr, seg);
+ addr += seg;
+ memcpy(cons->tx_buf, addr, len - seg);
+ }
+ TMFIFO_VDEV_TX_BUF_PUSH(cons, len);
+
+ if (!(virtio16_to_cpu(vdev, desc->flags) &
+ VRING_DESC_F_NEXT))
+ break;
+ idx = virtio16_to_cpu(vdev, desc->next);
+ desc = &vr->desc[idx];
+ }
+
+ /* Make each packet 8-byte aligned. */
+ TMFIFO_VDEV_TX_BUF_PUSH(cons, ((pkt_len + 7) & -8) - pkt_len);
+
+ tmfifo_virtio_release_desc(vdev, vr, head_desc, pkt_len);
+ }
+}
+
+/* Rx & Tx processing of a virtual queue. */
+static void tmfifo_virtio_rxtx(struct virtqueue *vq, bool is_rx)
+{
+ struct tmfifo_vring *vring;
+ struct tmfifo *fifo;
+ struct vring *vr;
+ struct virtio_device *vdev;
+ u64 sts, data;
+ int num_avail = 0, hdr_len, tx_reserve;
+ void *addr;
+ u32 len, idx;
+ struct vring_desc *desc;
+ unsigned long flags;
+ struct tmfifo_vdev *cons;
+
+ if (!vq)
+ return;
+
+ vring = (struct tmfifo_vring *)vq->priv;
+ fifo = vring->fifo;
+ vr = (struct vring *)virtqueue_get_vring(vq);
+
+ if (!fifo->vdev[vring->vdev_id])
+ return;
+ vdev = &fifo->vdev[vring->vdev_id]->vdev;
+ cons = fifo->vdev[VIRTIO_ID_CONSOLE];
+
+ /* Don't continue if another vring is running. */
+ if (fifo->vring[is_rx] != NULL && fifo->vring[is_rx] != vring)
+ return;
+
+ /* tx_reserve is used to reserved some room in FIFO for console. */
+ if (vring->vdev_id == VIRTIO_ID_NET) {
+ hdr_len = sizeof(struct virtio_net_hdr);
+ tx_reserve = fifo->tx_fifo_size / 16;
+ } else {
+ BUG_ON(vring->vdev_id != VIRTIO_ID_CONSOLE);
+ hdr_len = 0;
+ tx_reserve = 1;
+ }
+
+ desc = vring->desc;
+
+again:
+ while (1) {
+ /* Get available FIFO space. */
+ if (num_avail == 0) {
+ if (is_rx) {
+ /* Get the number of available words in FIFO. */
+ sts = readq(fifo->rx_base + TMFIFO_RX_STS);
+ num_avail = TMFIFO_GET_FIELD(sts,
+ TMFIFO_RX_STS__COUNT_MASK);
+
+ /* Don't continue if nothing in FIFO. */
+ if (num_avail <= 0)
+ break;
+ } else {
+ /* Get available space in FIFO. */
+ sts = readq(fifo->tx_base + TMFIFO_TX_STS);
+ num_avail = fifo->tx_fifo_size - tx_reserve -
+ TMFIFO_GET_FIELD(sts,
+ TMFIFO_TX_STS__COUNT_MASK);
+
+ if (num_avail <= 0)
+ break;
+ }
+ }
+
+ /* Console output always comes from the Tx buffer. */
+ if (!is_rx && vring->vdev_id == VIRTIO_ID_CONSOLE &&
+ cons != NULL && cons->tx_buf != NULL) {
+ for (;;) {
+ spin_lock_irqsave(&tmfifo_spin_lock, flags);
+ if (cons->tx_head == cons->tx_tail) {
+ spin_unlock_irqrestore(
+ &tmfifo_spin_lock, flags);
+ return;
+ }
+ addr = cons->tx_buf + cons->tx_head;
+ writeq(cpu_to_le64(*(u64 *)addr),
+ fifo->tx_base + TMFIFO_TX_DATA);
+ TMFIFO_VDEV_TX_BUF_POP(cons, sizeof(u64));
+ spin_unlock_irqrestore(&tmfifo_spin_lock,
+ flags);
+ if (--num_avail <= 0)
+ goto again;
+ }
+ }
+
+ /* Get the desc of next packet. */
+ if (!desc) {
+ /* Save the head desc of the chain. */
+ vring->desc_head = tmfifo_virtio_get_next_desc(vq);
+ if (!vring->desc_head) {
+ vring->desc = NULL;
+ return;
+ }
+ desc = vring->desc_head;
+ vring->desc = desc;
+
+ if (is_rx && vring->vdev_id == VIRTIO_ID_NET) {
+ struct virtio_net_hdr *net_hdr;
+
+ /* Initialize the packet header. */
+ net_hdr = (struct virtio_net_hdr *)
+ phys_to_virt(virtio64_to_cpu(
+ vdev, desc->addr));
+ memset(net_hdr, 0, sizeof(*net_hdr));
+ }
+ }
+
+ /* Beginning of each packet. */
+ if (vring->pkt_len == 0) {
+ int vdev_id, vring_change = 0;
+ union tmfifo_msg_hdr hdr;
+
+ num_avail--;
+
+ /* Read/Write packet length. */
+ if (is_rx) {
+ hdr.data = readq(fifo->rx_base +
+ TMFIFO_RX_DATA);
+ hdr.data = le64_to_cpu(hdr.data);
+
+ /* Skip the length 0 packet (keepalive). */
+ if (hdr.len == 0)
+ continue;
+
+ /* Check packet type. */
+ if (hdr.type == VIRTIO_ID_NET) {
+ vdev_id = VIRTIO_ID_NET;
+ hdr_len = sizeof(struct virtio_net_hdr);
+ } else if (hdr.type == VIRTIO_ID_CONSOLE) {
+ vdev_id = VIRTIO_ID_CONSOLE;
+ hdr_len = 0;
+ } else {
+ continue;
+ }
+
+ /*
+ * Check whether the new packet still belongs
+ * to this vring or not. If not, update the
+ * pkt_len of the new vring and return.
+ */
+ if (vdev_id != vring->vdev_id) {
+ struct tmfifo_vdev *dev2 =
+ fifo->vdev[vdev_id];
+
+ if (!dev2)
+ break;
+ vring->desc = desc;
+ vring = &dev2->vrings[TMFIFO_VRING_RX];
+ vring_change = 1;
+ }
+ vring->pkt_len = ntohs(hdr.len) + hdr_len;
+ } else {
+ vring->pkt_len = tmfifo_virtio_get_pkt_len(
+ vdev, desc, vr);
+
+ hdr.data = 0;
+ hdr.type = (vring->vdev_id == VIRTIO_ID_NET) ?
+ VIRTIO_ID_NET :
+ VIRTIO_ID_CONSOLE;
+ hdr.len = htons(vring->pkt_len - hdr_len);
+ writeq(cpu_to_le64(hdr.data),
+ fifo->tx_base + TMFIFO_TX_DATA);
+ }
+
+ vring->cur_len = hdr_len;
+ vring->rem_len = vring->pkt_len;
+ fifo->vring[is_rx] = vring;
+
+ if (vring_change)
+ return;
+ continue;
+ }
+
+ /* Check available space in this desc. */
+ len = virtio32_to_cpu(vdev, desc->len);
+ if (len > vring->rem_len)
+ len = vring->rem_len;
+
+ /* Check if the current desc is already done. */
+ if (vring->cur_len == len)
+ goto check_done;
+
+ addr = phys_to_virt(virtio64_to_cpu(vdev, desc->addr));
+
+ /* Read a word from FIFO for Rx. */
+ if (is_rx) {
+ data = readq(fifo->rx_base + TMFIFO_RX_DATA);
+ data = le64_to_cpu(data);
+ }
+
+ if (vring->cur_len + sizeof(u64) <= len) {
+ /* The whole word. */
+ if (is_rx) {
+ memcpy(addr + vring->cur_len, &data,
+ sizeof(u64));
+ } else {
+ memcpy(&data, addr + vring->cur_len,
+ sizeof(u64));
+ }
+ vring->cur_len += sizeof(u64);
+ } else {
+ /* Leftover bytes. */
+ BUG_ON(vring->cur_len > len);
+ if (is_rx) {
+ memcpy(addr + vring->cur_len, &data,
+ len - vring->cur_len);
+ } else {
+ memcpy(&data, addr + vring->cur_len,
+ len - vring->cur_len);
+ }
+ vring->cur_len = len;
+ }
+
+ /* Write the word into FIFO for Tx. */
+ if (!is_rx) {
+ writeq(cpu_to_le64(data),
+ fifo->tx_base + TMFIFO_TX_DATA);
+ }
+
+ num_avail--;
+
+check_done:
+ /* Check whether this desc is full or completed. */
+ if (vring->cur_len == len) {
+ vring->cur_len = 0;
+ vring->rem_len -= len;
+
+ /* Get the next desc on the chain. */
+ if (vring->rem_len > 0 &&
+ (virtio16_to_cpu(vdev, desc->flags) &
+ VRING_DESC_F_NEXT)) {
+ idx = virtio16_to_cpu(vdev, desc->next);
+ desc = &vr->desc[idx];
+ continue;
+ }
+
+ /* Done and release the desc. */
+ tmfifo_release_pkt(vdev, vring, &desc);
+ fifo->vring[is_rx] = NULL;
+
+ /* Notify upper layer that packet is done. */
+ spin_lock_irqsave(&tmfifo_spin_lock, flags);
+ vring_interrupt(0, vq);
+ spin_unlock_irqrestore(&tmfifo_spin_lock, flags);
+ continue;
+ }
+ }
+
+ /* Save the current desc. */
+ vring->desc = desc;
+}
+
+/* The notify function is called when new buffers are posted. */
+static bool tmfifo_virtio_notify(struct virtqueue *vq)
+{
+ struct tmfifo_vring *vring = (struct tmfifo_vring *)vq->priv;
+ struct tmfifo *fifo = vring->fifo;
+ unsigned long flags;
+
+ /*
+ * Virtio maintains vrings in pairs, even number ring for Rx
+ * and odd number ring for Tx.
+ */
+ if (!(vring->id & 1)) {
+ /* Set the RX HWM bit to start Rx. */
+ if (!test_and_set_bit(TM_RX_HWM_IRQ, &fifo->pend_events))
+ schedule_work(&fifo->work);
+ } else {
+ /*
+ * Console could make blocking call with interrupts disabled.
+ * In such case, the vring needs to be served right away. For
+ * other cases, just set the TX LWM bit to start Tx in the
+ * worker handler.
+ */
+ if (vring->vdev_id == VIRTIO_ID_CONSOLE) {
+ spin_lock_irqsave(&tmfifo_spin_lock, flags);
+ tmfifo_console_output(fifo->vdev[VIRTIO_ID_CONSOLE],
+ vq);
+ spin_unlock_irqrestore(&tmfifo_spin_lock, flags);
+ schedule_work(&fifo->work);
+ } else if (!test_and_set_bit(TM_TX_LWM_IRQ, &fifo->pend_events))
+ schedule_work(&fifo->work);
+ }
+
+ return true;
+}
+
+/* Get the array of feature bits for this device. */
+static u64 tmfifo_virtio_get_features(struct virtio_device *vdev)
+{
+ struct tmfifo_vdev *tm_vdev = container_of(vdev, struct tmfifo_vdev,
+ vdev);
+
+ return tm_vdev->features;
+}
+
+/* Confirm device features to use. */
+static int tmfifo_virtio_finalize_features(struct virtio_device *vdev)
+{
+ struct tmfifo_vdev *tm_vdev = container_of(vdev, struct tmfifo_vdev,
+ vdev);
+
+ tm_vdev->features = vdev->features;
+ return 0;
+}
+
+/* Free virtqueues found by find_vqs(). */
+static void tmfifo_virtio_del_vqs(struct virtio_device *vdev)
+{
+ int i;
+ struct tmfifo_vring *vring;
+ struct virtqueue *vq;
+ struct tmfifo_vdev *tm_vdev = container_of(vdev, struct tmfifo_vdev,
+ vdev);
+
+ for (i = 0; i < ARRAY_SIZE(tm_vdev->vrings); i++) {
+ vring = &tm_vdev->vrings[i];
+
+ /* Release the pending packet. */
+ if (vring->desc != NULL)
+ tmfifo_release_pkt(&tm_vdev->vdev, vring, &vring->desc);
+
+ vq = vring->vq;
+ if (vq) {
+ vring->vq = NULL;
+ vring_del_virtqueue(vq);
+ }
+ }
+}
+
+/* Create and initialize the virtual queues. */
+static int tmfifo_virtio_find_vqs(struct virtio_device *vdev,
+ unsigned int nvqs,
+ struct virtqueue *vqs[],
+ vq_callback_t *callbacks[],
+ const char * const names[],
+ const bool *ctx,
+ struct irq_affinity *desc)
+{
+ int i, ret = -EINVAL, size;
+ struct tmfifo_vring *vring;
+ struct virtqueue *vq;
+ struct tmfifo_vdev *tm_vdev = container_of(vdev, struct tmfifo_vdev,
+ vdev);
+
+ if (nvqs > ARRAY_SIZE(tm_vdev->vrings))
+ return -EINVAL;
+
+ for (i = 0; i < nvqs; ++i) {
+ if (!names[i])
+ goto error;
+ vring = &tm_vdev->vrings[i];
+
+ /* zero vring */
+ size = vring_size(vring->size, vring->align);
+ memset(vring->va, 0, size);
+ vq = vring_new_virtqueue(i, vring->size, vring->align, vdev,
+ false, false, vring->va,
+ tmfifo_virtio_notify,
+ callbacks[i], names[i]);
+ if (!vq) {
+ dev_err(&vdev->dev, "vring_new_virtqueue failed\n");
+ ret = -ENOMEM;
+ goto error;
+ }
+
+ vqs[i] = vq;
+ vring->vq = vq;
+ vq->priv = vring;
+ }
+
+ return 0;
+
+error:
+ tmfifo_virtio_del_vqs(vdev);
+ return ret;
+}
+
+/* Read the status byte. */
+static u8 tmfifo_virtio_get_status(struct virtio_device *vdev)
+{
+ struct tmfifo_vdev *tm_vdev = container_of(vdev, struct tmfifo_vdev,
+ vdev);
+
+ return tm_vdev->status;
+}
+
+/* Write the status byte. */
+static void tmfifo_virtio_set_status(struct virtio_device *vdev, u8 status)
+{
+ struct tmfifo_vdev *tm_vdev = container_of(vdev, struct tmfifo_vdev,
+ vdev);
+
+ tm_vdev->status = status;
+}
+
+/* Reset the device. Not much here for now. */
+static void tmfifo_virtio_reset(struct virtio_device *vdev)
+{
+ struct tmfifo_vdev *tm_vdev = container_of(vdev, struct tmfifo_vdev,
+ vdev);
+
+ tm_vdev->status = 0;
+}
+
+/* Read the value of a configuration field. */
+static void tmfifo_virtio_get(struct virtio_device *vdev,
+ unsigned int offset,
+ void *buf,
+ unsigned int len)
+{
+ struct tmfifo_vdev *tm_vdev = container_of(vdev, struct tmfifo_vdev,
+ vdev);
+
+ if (offset + len > sizeof(tm_vdev->config) || offset + len < len) {
+ dev_err(vdev->dev.parent, "virtio_get access out of bounds\n");
+ return;
+ }
+
+ memcpy(buf, (u8 *)&tm_vdev->config + offset, len);
+}
+
+/* Write the value of a configuration field. */
+static void tmfifo_virtio_set(struct virtio_device *vdev,
+ unsigned int offset,
+ const void *buf,
+ unsigned int len)
+{
+ struct tmfifo_vdev *tm_vdev = container_of(vdev, struct tmfifo_vdev,
+ vdev);
+
+ if (offset + len > sizeof(tm_vdev->config) || offset + len < len) {
+ dev_err(vdev->dev.parent, "virtio_get access out of bounds\n");
+ return;
+ }
+
+ memcpy((u8 *)&tm_vdev->config + offset, buf, len);
+}
+
+/* Virtio config operations. */
+static const struct virtio_config_ops tmfifo_virtio_config_ops = {
+ .get_features = tmfifo_virtio_get_features,
+ .finalize_features = tmfifo_virtio_finalize_features,
+ .find_vqs = tmfifo_virtio_find_vqs,
+ .del_vqs = tmfifo_virtio_del_vqs,
+ .reset = tmfifo_virtio_reset,
+ .set_status = tmfifo_virtio_set_status,
+ .get_status = tmfifo_virtio_get_status,
+ .get = tmfifo_virtio_get,
+ .set = tmfifo_virtio_set,
+};
+
+/* Create vdev type in a tmfifo. */
+int tmfifo_create_vdev(struct tmfifo *fifo, int vdev_id, u64 features,
+ void *config, u32 size)
+{
+ struct tmfifo_vdev *tm_vdev;
+ int ret = 0;
+
+ mutex_lock(&fifo->lock);
+
+ tm_vdev = fifo->vdev[vdev_id];
+ if (tm_vdev != NULL) {
+ pr_err("vdev %d already exists\n", vdev_id);
+ ret = -EEXIST;
+ goto already_exist;
+ }
+
+ tm_vdev = kzalloc(sizeof(*tm_vdev), GFP_KERNEL);
+ if (!tm_vdev) {
+ ret = -ENOMEM;
+ goto already_exist;
+ }
+
+ tm_vdev->vdev.id.device = vdev_id;
+ tm_vdev->vdev.config = &tmfifo_virtio_config_ops;
+ tm_vdev->vdev.dev.parent = &fifo->pdev->dev;
+ tm_vdev->vdev.dev.release = tmfifo_virtio_dev_release;
+ tm_vdev->features = features;
+ if (config)
+ memcpy(&tm_vdev->config, config, size);
+ if (tmfifo_alloc_vrings(fifo, tm_vdev, vdev_id)) {
+ pr_err("Unable to allocate vring\n");
+ ret = -ENOMEM;
+ goto alloc_vring_fail;
+ }
+ if (vdev_id == VIRTIO_ID_CONSOLE) {
+ tm_vdev->tx_buf = kmalloc(TMFIFO_CONS_TX_BUF_SIZE,
+ GFP_KERNEL);
+ }
+ fifo->vdev[vdev_id] = tm_vdev;
+
+ /* Register the virtio device. */
+ ret = register_virtio_device(&tm_vdev->vdev);
+ if (ret) {
+ dev_err(&fifo->pdev->dev, "register_virtio_device() failed\n");
+ goto register_fail;
+ }
+
+ mutex_unlock(&fifo->lock);
+ return 0;
+
+register_fail:
+ tmfifo_free_vrings(fifo, vdev_id);
+ fifo->vdev[vdev_id] = NULL;
+alloc_vring_fail:
+ kfree(tm_vdev);
+already_exist:
+ mutex_unlock(&fifo->lock);
+ return ret;
+}
+
+/* Delete vdev type from a tmfifo. */
+int tmfifo_delete_vdev(struct tmfifo *fifo, int vdev_id)
+{
+ struct tmfifo_vdev *tm_vdev;
+
+ mutex_lock(&fifo->lock);
+
+ /* Unregister vdev. */
+ tm_vdev = fifo->vdev[vdev_id];
+ if (tm_vdev) {
+ unregister_virtio_device(&tm_vdev->vdev);
+ tmfifo_free_vrings(fifo, vdev_id);
+ kfree(tm_vdev->tx_buf);
+ kfree(tm_vdev);
+ fifo->vdev[vdev_id] = NULL;
+ }
+
+ mutex_unlock(&fifo->lock);
+
+ return 0;
+}
+
+/* Device remove function. */
+static int tmfifo_remove(struct platform_device *pdev)
+{
+ int i;
+ struct tmfifo *fifo = platform_get_drvdata(pdev);
+ struct resource *rx_res, *tx_res;
+
+ tmfifo_ready = false;
+
+ if (fifo) {
+ mutex_lock(&tmfifo_lock);
+
+ /* Stop the timer. */
+ del_timer_sync(&fifo->timer);
+
+ /* Release interrupts. */
+ tmfifo_free_irqs(fifo);
+
+ /* Cancel the pending work. */
+ cancel_work_sync(&fifo->work);
+
+ for (i = 0; i < TMFIFO_VDEV_MAX; i++)
+ tmfifo_delete_vdev(fifo, i);
+
+ /* Release IO resources. */
+ if (fifo->rx_base)
+ iounmap(fifo->rx_base);
+ if (fifo->tx_base)
+ iounmap(fifo->tx_base);
+
+ platform_set_drvdata(pdev, NULL);
+ kfree(fifo);
+
+ mutex_unlock(&tmfifo_lock);
+ }
+
+ rx_res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+ if (rx_res)
+ release_mem_region(rx_res->start, resource_size(rx_res));
+ tx_res = platform_get_resource(pdev, IORESOURCE_MEM, 1);
+ if (tx_res)
+ release_mem_region(tx_res->start, resource_size(tx_res));
+
+ return 0;
+}
+
+/* Read the configured network MAC address from efi variable. */
+static void tmfifo_get_cfg_mac(u8 *mac)
+{
+ u8 buf[6];
+ efi_status_t status;
+ unsigned long size = sizeof(buf);
+ efi_char16_t name[] = { 'R', 's', 'h', 'i', 'm', 'M', 'a', 'c',
+ 'A', 'd', 'd', 'r', 0 };
+ efi_guid_t guid = EFI_GLOBAL_VARIABLE_GUID;
+
+ status = efi.get_variable(name, &guid, NULL, &size, buf);
+ if (status == EFI_SUCCESS && size == sizeof(buf))
+ memcpy(mac, buf, sizeof(buf));
+}
+
+/* Probe the TMFIFO. */
+static int tmfifo_probe(struct platform_device *pdev)
+{
+ u64 ctl;
+ struct tmfifo *fifo;
+ struct resource *rx_res, *tx_res;
+ struct virtio_net_config net_config;
+ int i, ret;
+
+ /* Get the resource of the Rx & Tx FIFO. */
+ rx_res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+ tx_res = platform_get_resource(pdev, IORESOURCE_MEM, 1);
+ if (!rx_res || !tx_res) {
+ ret = -EINVAL;
+ goto err;
+ }
+
+ if (request_mem_region(rx_res->start,
+ resource_size(rx_res), "bf-tmfifo") == NULL) {
+ ret = -EBUSY;
+ goto early_err;
+ }
+
+ if (request_mem_region(tx_res->start,
+ resource_size(tx_res), "bf-tmfifo") == NULL) {
+ release_mem_region(rx_res->start, resource_size(rx_res));
+ ret = -EBUSY;
+ goto early_err;
+ }
+
+ ret = -ENOMEM;
+ fifo = kzalloc(sizeof(struct tmfifo), GFP_KERNEL);
+ if (!fifo)
+ goto err;
+
+ fifo->pdev = pdev;
+ platform_set_drvdata(pdev, fifo);
+
+ INIT_WORK(&fifo->work, tmfifo_work_handler);
+
+ timer_setup(&fifo->timer, tmfifo_timer, 0);
+ fifo->timer.function = tmfifo_timer;
+
+ for (i = 0; i < TM_IRQ_CNT; i++) {
+ fifo->irq[i] = platform_get_irq(pdev, i);
+ ret = request_irq(fifo->irq[i], tmfifo_irq_handler, 0,
+ "tmfifo", (u8 *)fifo + i);
+ if (ret) {
+ pr_err("Unable to request irq\n");
+ fifo->irq[i] = 0;
+ goto err;
+ }
+ }
+
+ fifo->rx_base = ioremap(rx_res->start, resource_size(rx_res));
+ if (!fifo->rx_base)
+ goto err;
+
+ fifo->tx_base = ioremap(tx_res->start, resource_size(tx_res));
+ if (!fifo->tx_base)
+ goto err;
+
+ /* Get Tx FIFO size and set the low/high watermark. */
+ ctl = readq(fifo->tx_base + TMFIFO_TX_CTL);
+ fifo->tx_fifo_size =
+ TMFIFO_GET_FIELD(ctl, TMFIFO_TX_CTL__MAX_ENTRIES_MASK);
+ ctl = TMFIFO_SET_FIELD(ctl, TMFIFO_TX_CTL__LWM_MASK,
+ fifo->tx_fifo_size / 2);
+ ctl = TMFIFO_SET_FIELD(ctl, TMFIFO_TX_CTL__HWM_MASK,
+ fifo->tx_fifo_size - 1);
+ writeq(ctl, fifo->tx_base + TMFIFO_TX_CTL);
+
+ /* Get Rx FIFO size and set the low/high watermark. */
+ ctl = readq(fifo->rx_base + TMFIFO_RX_CTL);
+ fifo->rx_fifo_size =
+ TMFIFO_GET_FIELD(ctl, TMFIFO_RX_CTL__MAX_ENTRIES_MASK);
+ ctl = TMFIFO_SET_FIELD(ctl, TMFIFO_RX_CTL__LWM_MASK, 0);
+ ctl = TMFIFO_SET_FIELD(ctl, TMFIFO_RX_CTL__HWM_MASK, 1);
+ writeq(ctl, fifo->rx_base + TMFIFO_RX_CTL);
+
+ mutex_init(&fifo->lock);
+
+ /* Create the console vdev. */
+ ret = tmfifo_create_vdev(fifo, VIRTIO_ID_CONSOLE, 0, NULL, 0);
+ if (ret)
+ goto err;
+
+ /* Create the network vdev. */
+ memset(&net_config, 0, sizeof(net_config));
+ net_config.mtu = TMFIFO_NET_MTU;
+ net_config.status = VIRTIO_NET_S_LINK_UP;
+ memcpy(net_config.mac, tmfifo_net_default_mac, 6);
+ tmfifo_get_cfg_mac(net_config.mac);
+ ret = tmfifo_create_vdev(fifo, VIRTIO_ID_NET, TMFIFO_NET_FEATURES,
+ &net_config, sizeof(net_config));
+ if (ret)
+ goto err;
+
+ mod_timer(&fifo->timer, jiffies + tmfifo_timer_interval);
+
+ tmfifo_ready = true;
+
+ return 0;
+
+err:
+ tmfifo_remove(pdev);
+early_err:
+ dev_err(&pdev->dev, "Probe Failed\n");
+ return ret;
+}
+
+static const struct of_device_id tmfifo_match[] = {
+ { .compatible = "mellanox,bf-tmfifo" },
+ {},
+};
+MODULE_DEVICE_TABLE(of, tmfifo_match);
+
+static const struct acpi_device_id bf_tmfifo_acpi_match[] = {
+ { "MLNXBF01", 0 },
+ {},
+};
+MODULE_DEVICE_TABLE(acpi, bf_tmfifo_acpi_match);
+
+static struct platform_driver tmfifo_driver = {
+ .probe = tmfifo_probe,
+ .remove = tmfifo_remove,
+ .driver = {
+ .name = "bf-tmfifo",
+ .of_match_table = tmfifo_match,
+ .acpi_match_table = ACPI_PTR(bf_tmfifo_acpi_match),
+ },
+};
+
+static int __init tmfifo_init(void)
+{
+ int ret;
+
+ mutex_init(&tmfifo_lock);
+
+ ret = platform_driver_register(&tmfifo_driver);
+ if (ret)
+ pr_err("Failed to register tmfifo driver.\n");
+
+ return ret;
+}
+
+static void __exit tmfifo_exit(void)
+{
+ platform_driver_unregister(&tmfifo_driver);
+}
+
+module_init(tmfifo_init);
+module_exit(tmfifo_exit);
+
+MODULE_DESCRIPTION("Mellanox BlueField SoC TMFIFO Driver");
+MODULE_LICENSE("GPL v2");
+MODULE_AUTHOR("Mellanox Technologies");
diff --git a/drivers/soc/mellanox/tmfifo_regs.h b/drivers/soc/mellanox/tmfifo_regs.h
new file mode 100644
index 0000000..f42c9d6
--- /dev/null
+++ b/drivers/soc/mellanox/tmfifo_regs.h
@@ -0,0 +1,75 @@
+/*
+ * Copyright (c) 2018, Mellanox Technologies. 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 version 2 and
+ * only version 2 as published by the Free Software Foundation.
+ *
+ * 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.
+ */
+
+#ifndef __TMFIFO_REGS_H__
+#define __TMFIFO_REGS_H__
+
+#include <linux/types.h>
+
+#define TMFIFO_TX_DATA 0x0
+
+#define TMFIFO_TX_STS 0x8
+#define TMFIFO_TX_STS__LENGTH 0x0001
+#define TMFIFO_TX_STS__COUNT_SHIFT 0
+#define TMFIFO_TX_STS__COUNT_WIDTH 9
+#define TMFIFO_TX_STS__COUNT_RESET_VAL 0
+#define TMFIFO_TX_STS__COUNT_RMASK 0x1ff
+#define TMFIFO_TX_STS__COUNT_MASK 0x1ff
+
+#define TMFIFO_TX_CTL 0x10
+#define TMFIFO_TX_CTL__LENGTH 0x0001
+#define TMFIFO_TX_CTL__LWM_SHIFT 0
+#define TMFIFO_TX_CTL__LWM_WIDTH 8
+#define TMFIFO_TX_CTL__LWM_RESET_VAL 128
+#define TMFIFO_TX_CTL__LWM_RMASK 0xff
+#define TMFIFO_TX_CTL__LWM_MASK 0xff
+#define TMFIFO_TX_CTL__HWM_SHIFT 8
+#define TMFIFO_TX_CTL__HWM_WIDTH 8
+#define TMFIFO_TX_CTL__HWM_RESET_VAL 128
+#define TMFIFO_TX_CTL__HWM_RMASK 0xff
+#define TMFIFO_TX_CTL__HWM_MASK 0xff00
+#define TMFIFO_TX_CTL__MAX_ENTRIES_SHIFT 32
+#define TMFIFO_TX_CTL__MAX_ENTRIES_WIDTH 9
+#define TMFIFO_TX_CTL__MAX_ENTRIES_RESET_VAL 256
+#define TMFIFO_TX_CTL__MAX_ENTRIES_RMASK 0x1ff
+#define TMFIFO_TX_CTL__MAX_ENTRIES_MASK 0x1ff00000000ULL
+
+#define TMFIFO_RX_DATA 0x0
+
+#define TMFIFO_RX_STS 0x8
+#define TMFIFO_RX_STS__LENGTH 0x0001
+#define TMFIFO_RX_STS__COUNT_SHIFT 0
+#define TMFIFO_RX_STS__COUNT_WIDTH 9
+#define TMFIFO_RX_STS__COUNT_RESET_VAL 0
+#define TMFIFO_RX_STS__COUNT_RMASK 0x1ff
+#define TMFIFO_RX_STS__COUNT_MASK 0x1ff
+
+#define TMFIFO_RX_CTL 0x10
+#define TMFIFO_RX_CTL__LENGTH 0x0001
+#define TMFIFO_RX_CTL__LWM_SHIFT 0
+#define TMFIFO_RX_CTL__LWM_WIDTH 8
+#define TMFIFO_RX_CTL__LWM_RESET_VAL 128
+#define TMFIFO_RX_CTL__LWM_RMASK 0xff
+#define TMFIFO_RX_CTL__LWM_MASK 0xff
+#define TMFIFO_RX_CTL__HWM_SHIFT 8
+#define TMFIFO_RX_CTL__HWM_WIDTH 8
+#define TMFIFO_RX_CTL__HWM_RESET_VAL 128
+#define TMFIFO_RX_CTL__HWM_RMASK 0xff
+#define TMFIFO_RX_CTL__HWM_MASK 0xff00
+#define TMFIFO_RX_CTL__MAX_ENTRIES_SHIFT 32
+#define TMFIFO_RX_CTL__MAX_ENTRIES_WIDTH 9
+#define TMFIFO_RX_CTL__MAX_ENTRIES_RESET_VAL 256
+#define TMFIFO_RX_CTL__MAX_ENTRIES_RMASK 0x1ff
+#define TMFIFO_RX_CTL__MAX_ENTRIES_MASK 0x1ff00000000ULL
+
+#endif /* !defined(__TMFIFO_REGS_H__) */
--
1.8.3.1
^ permalink raw reply related
* [PATCH v3 2/4] arm64: Add Mellanox BlueField SoC config option
From: Liming Sun @ 2018-06-01 14:31 UTC (permalink / raw)
To: linux-arm-kernel
In-Reply-To: <1527863467-45472-1-git-send-email-lsun@mellanox.com>
This commit introduces config option for Mellanox BlueField SoC,
which can be used to build the SoC specific drivers, and enables
it by default in configs/defconfig.
Reviewed-by: David Woods <dwoods@mellanox.com>
Signed-off-by: Liming Sun <lsun@mellanox.com>
---
arch/arm64/Kconfig.platforms | 6 ++++++
arch/arm64/configs/defconfig | 1 +
2 files changed, 7 insertions(+)
diff --git a/arch/arm64/Kconfig.platforms b/arch/arm64/Kconfig.platforms
index 2b1535c..74ad03f 100644
--- a/arch/arm64/Kconfig.platforms
+++ b/arch/arm64/Kconfig.platforms
@@ -110,6 +110,12 @@ config ARCH_MESON
help
This enables support for the Amlogic S905 SoCs.
+config ARCH_MLNX_BLUEFIELD
+ bool "Mellanox BlueField SoC Family"
+ select SOC_MLNX
+ help
+ This enables support for the Mellanox BlueField SoC.
+
config ARCH_MVEBU
bool "Marvell EBU SoC Family"
select ARMADA_AP806_SYSCON
diff --git a/arch/arm64/configs/defconfig b/arch/arm64/configs/defconfig
index 1c98939..842f607 100644
--- a/arch/arm64/configs/defconfig
+++ b/arch/arm64/configs/defconfig
@@ -43,6 +43,7 @@ CONFIG_ARCH_LG1K=y
CONFIG_ARCH_HISI=y
CONFIG_ARCH_MEDIATEK=y
CONFIG_ARCH_MESON=y
+CONFIG_ARCH_MLNX_BLUEFIELD=y
CONFIG_ARCH_MVEBU=y
CONFIG_ARCH_QCOM=y
CONFIG_ARCH_ROCKCHIP=y
--
1.8.3.1
^ permalink raw reply related
* [PATCH v3 3/4] dt-bindings: soc: Add TmFifo binding for Mellanox BlueField SoC
From: Liming Sun @ 2018-06-01 14:31 UTC (permalink / raw)
To: linux-arm-kernel
In-Reply-To: <1527863467-45472-1-git-send-email-lsun@mellanox.com>
Add devicetree bindings for the TmFifo which is found on Mellanox
BlueField SoCs.
Reviewed-by: David Woods <dwoods@mellanox.com>
Signed-off-by: Liming Sun <lsun@mellanox.com>
---
.../devicetree/bindings/soc/mellanox/tmfifo.txt | 23 ++++++++++++++++++++++
1 file changed, 23 insertions(+)
create mode 100644 Documentation/devicetree/bindings/soc/mellanox/tmfifo.txt
diff --git a/Documentation/devicetree/bindings/soc/mellanox/tmfifo.txt b/Documentation/devicetree/bindings/soc/mellanox/tmfifo.txt
new file mode 100644
index 0000000..8a13fa6
--- /dev/null
+++ b/Documentation/devicetree/bindings/soc/mellanox/tmfifo.txt
@@ -0,0 +1,23 @@
+* Mellanox BlueField SoC TmFifo
+
+BlueField TmFifo provides a shared FIFO between the target and the
+external host machine, which can be accessed by external host via
+USB or PCIe. In the current tmfifo driver, this FIFO has been demuxed
+to implement virtual console and network interface based on the virtio
+framework.
+
+Required properties:
+
+- compatible: Should be "mellanox,bf-tmfifo"
+- reg: Physical base address and length of Rx/Tx block
+- interrupts: The interrupt number of Rx low water mark, Rx high water mark
+ Tx low water mark, Tx high water mark respectively.
+
+Example:
+
+tmfifo at 800a20 {
+ compatible = "mellanox,bf-tmfifo";
+ reg = <0x00800a20 0x00000018
+ 0x00800a40 0x00000018>;
+ interrupts = <41, 42, 43, 44>;
+};
--
1.8.3.1
^ permalink raw reply related
* [PATCH v3 4/4] MAINTAINERS: Add entry for Mellanox Bluefield Soc
From: Liming Sun @ 2018-06-01 14:31 UTC (permalink / raw)
To: linux-arm-kernel
In-Reply-To: <1527863467-45472-1-git-send-email-lsun@mellanox.com>
Add maintainer information for Mellanox BlueField SoC.
Reviewed-by: David Woods <dwoods@mellanox.com>
Signed-off-by: Liming Sun <lsun@mellanox.com>
---
MAINTAINERS | 8 ++++++++
1 file changed, 8 insertions(+)
diff --git a/MAINTAINERS b/MAINTAINERS
index 58b9861..85d5639 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -1636,6 +1636,14 @@ L: linux-mediatek at lists.infradead.org (moderated for non-subscribers)
S: Maintained
F: drivers/phy/mediatek/phy-mtk-tphy.c
+ARM/Mellanox BlueField SoC support
+M: David Woods <dwoods@mellanox.com>
+M: Liming Sun <lsun@mellanox.com>
+L: linux-arm-kernel at lists.infradead.org (moderated for non-subscribers)
+S: Maintained
+F: drivers/soc/mellanox/*
+F: Documentation/devicetree/bindings/soc/mellanox/tmfifo.txt
+
ARM/MICREL KS8695 ARCHITECTURE
M: Greg Ungerer <gerg@uclinux.org>
L: linux-arm-kernel at lists.infradead.org (moderated for non-subscribers)
--
1.8.3.1
^ permalink raw reply related
* [PATCHv2 0/4] arm64: SMCCC conduit cleanup
From: Catalin Marinas @ 2018-06-01 14:37 UTC (permalink / raw)
To: linux-arm-kernel
In-Reply-To: <20180601142615.GU17671@n2100.armlinux.org.uk>
On Fri, Jun 01, 2018 at 03:26:16PM +0100, Russell King wrote:
> On Fri, Jun 01, 2018 at 03:14:10PM +0100, Catalin Marinas wrote:
> > On Thu, May 31, 2018 at 06:32:19PM +0100, Mark Rutland wrote:
> > > Mark Rutland (4):
> > > arm/arm64: smccc/psci: add arm_smccc_1_1_get_conduit()
> > > arm64: errata: use arm_smccc_get_conduit()
> > > firmware/psci: use common SMCCC_CONDUIT_*
> > > firmware: arm_sdei: use common SMCCC_CONDUIT_*
> >
> > Queued for 4.18. Thanks.
>
> Have you considered how this affects the 32-bit ARM spectre stuff, which
> makes use of the firmware PSCI interfaces for the workaround?
No, I haven't. Thanks for the heads up. Talking to Mark, I think it's
best if I drop this series for now.
--
Catalin
^ permalink raw reply
* [PATCHv2 11/19] arm64: don't reload GPRs after apply_ssbd
From: Marc Zyngier @ 2018-06-01 14:41 UTC (permalink / raw)
To: linux-arm-kernel
In-Reply-To: <20180601112441.37810-12-mark.rutland@arm.com>
On 01/06/18 12:24, Mark Rutland wrote:
> Now that all of the syscall logic works on the saved pt_regs, apply_ssbd
> can safely corrupt x0-x3 in the entry paths, and we no longer need to
> restore them. So let's remotve the logic doing so.
remove*
>
> With that logic gone, we can fold the branch target into the macro, so
> that callers need not deal with this. GAS provides \@, which provides a
> unique value per macro invocation, which we can use to create a unique
> label.
>
> Signed-off-by: Mark Rutland <mark.rutland@arm.com>
> Cc: Catalin Marinas <catalin.marinas@arm.com>
> Cc: Marc Zyngier <marc.zyngier@arm.com>
> Cc: Will Deacon <will.deacon@arm.com>
Acked-by: Marc Zyngier <marc.zyngier@arm.com>
M.
--
Jazz is not dead. It just smells funny...
^ permalink raw reply
* [PATCH] PCI: move early dump functionality from x86 arch into the common code
From: Sinan Kaya @ 2018-06-01 14:45 UTC (permalink / raw)
To: linux-arm-kernel
In-Reply-To: <1527654876-26716-1-git-send-email-okaya@codeaurora.org>
On 5/30/2018 12:34 AM, Sinan Kaya wrote:
> Move early dump functionality into common code so that it is available for
> all archtiectures. No need to carry arch specific reads around as the read
> hooks are already initialized by the time pci_setup_device() is getting
> called during scan.
>
> Signed-off-by: Sinan Kaya <okaya@codeaurora.org>
> ---
> Documentation/admin-guide/kernel-parameters.txt | 2 +-
> arch/x86/include/asm/pci-direct.h | 5 ---
> arch/x86/kernel/setup.c | 5 ---
> arch/x86/pci/common.c | 4 --
> arch/x86/pci/early.c | 50 -------------------------
> drivers/pci/pci.c | 4 ++
> drivers/pci/pci.h | 2 +-
> drivers/pci/probe.c | 19 ++++++++++
> 8 files changed, 25 insertions(+), 66 deletions(-)
Any feedback on the direction?
--
Sinan Kaya
Qualcomm Datacenter Technologies, Inc. as an affiliate of Qualcomm Technologies, Inc.
Qualcomm Technologies, Inc. is a member of the Code Aurora Forum, a Linux Foundation Collaborative Project.
^ permalink raw reply
* [PATCH v3 2/8] dt-bindings: media: Document data-enable-active property
From: jacopo mondi @ 2018-06-01 14:58 UTC (permalink / raw)
To: linux-arm-kernel
In-Reply-To: <20180601102910.3qhe6bhg3w263chq@paasikivi.fi.intel.com>
Hi Sakari,
On Fri, Jun 01, 2018 at 01:29:10PM +0300, Sakari Ailus wrote:
> Hi Jacopo,
>
> Thanks for the patch.
>
> On Tue, May 29, 2018 at 05:05:53PM +0200, Jacopo Mondi wrote:
> > Add 'data-enable-active' property to endpoint node properties list.
> >
> > The property allows to specify the polarity of the data-enable signal, which
> > when in active state determinates when data lanes have to sampled for valid
> > pixel data.
>
> Lanes or lines? I understand this is forthe parallel interface.
>
Now I'm confused. Are the parallel data 'lines' and the CSI-2 one
'lanes'? I thought 'lanes' applies to both.. am I wrong?
Thanks
j
> >
> > Signed-off-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
> > ---
> > v3:
> > - new patch
> > ---
> > Documentation/devicetree/bindings/media/video-interfaces.txt | 2 ++
> > 1 file changed, 2 insertions(+)
> >
> > diff --git a/Documentation/devicetree/bindings/media/video-interfaces.txt b/Documentation/devicetree/bindings/media/video-interfaces.txt
> > index 258b8df..9839d26 100644
> > --- a/Documentation/devicetree/bindings/media/video-interfaces.txt
> > +++ b/Documentation/devicetree/bindings/media/video-interfaces.txt
> > @@ -109,6 +109,8 @@ Optional endpoint properties
> > Note, that if HSYNC and VSYNC polarities are not specified, embedded
> > synchronization may be required, where supported.
> > - data-active: similar to HSYNC and VSYNC, specifies data line polarity.
> > +- data-enable-active: similar to HSYNC and VSYNC, specifies the data enable
> > + signal polarity.
> > - field-even-active: field signal level during the even field data transmission.
> > - pclk-sample: sample data on rising (1) or falling (0) edge of the pixel clock
> > signal.
>
> --
> Sakari Ailus
> sakari.ailus at linux.intel.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20180601/da16dd9b/attachment.sig>
^ permalink raw reply
* [PATCH] PCI: move early dump functionality from x86 arch into the common code
From: Andy Shevchenko @ 2018-06-01 15:02 UTC (permalink / raw)
To: linux-arm-kernel
In-Reply-To: <1527654876-26716-1-git-send-email-okaya@codeaurora.org>
On Wed, May 30, 2018 at 7:34 AM, Sinan Kaya <okaya@codeaurora.org> wrote:
> Move early dump functionality into common code so that it is available for
> all archtiectures. No need to carry arch specific reads around as the read
> hooks are already initialized by the time pci_setup_device() is getting
> called during scan.
> /* If set, the PCIe ATS capability will not be used. */
> static bool pcie_ats_disabled;
>
> +bool pci_early_dump;
> +
I didn't check above these, but maybe a good idea to add one line
comment what is this about?
> extern const unsigned char pcie_link_speed[];
> -
> +extern bool pci_early_dump;
> bool pcie_cap_has_lnkctl(const struct pci_dev *dev);
>
Hmm... I would rather not attach this line to some function declarations.
> +static void early_dump_pci_device(struct pci_dev *pdev)
> +{
> + u32 value[256 / 4];
> + int i;
> +
> + dev_info(&pdev->dev, "pci 0000:%02x:%02x.%d config space:\n",
> + pdev->bus->number, PCI_SLOT(pdev->devfn),
> + PCI_FUNC(pdev->devfn));
Shouldn't be this changed to pci_info() ?
> +
> + for (i = 0; i < 256; i += 4)
> + pci_read_config_dword(pdev, i, &value[i / 4]);
> +
> + print_hex_dump(KERN_INFO, "", DUMP_PREFIX_OFFSET, 16, 1, value,
> + 256, false);
> +}
--
With Best Regards,
Andy Shevchenko
^ permalink raw reply
* [PATCH v2] arm64: allwinner: a64-amarula-relic: Enable AP6330 WiFi support
From: Maxime Ripard @ 2018-06-01 15:06 UTC (permalink / raw)
To: linux-arm-kernel
In-Reply-To: <20180531165109.8476-1-jagan@amarulasolutions.com>
Hi,
On Thu, May 31, 2018 at 10:21:09PM +0530, Jagan Teki wrote:
> Enable AP6330 WiFi/BT combo chip on Amarula A64-Relic board:
> - WiFi SDIO interface is connected to MMC1
> - WiFi WL-PMU-EN pin connected to gpio PL2: attach to mmc-pwrseq
> - WiFi WL-WAKE-AP pin connected to gpio PL3
> - 32kHz external oscillator gate clock from RTC
>
> Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
> ---
> Changes for v2:
> - Move external rtc clock nodes into main rtc node definition
> of sun50i-a64.dtsi
Please make that a separate patch.
Maxime
--
Maxime Ripard, Bootlin (formerly Free Electrons)
Embedded Linux and Kernel engineering
https://bootlin.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20180601/d1f386eb/attachment-0001.sig>
^ permalink raw reply
* [PATCH] KVM: arm/arm64: drop resource size check for GICV window
From: Ard Biesheuvel @ 2018-06-01 15:06 UTC (permalink / raw)
To: linux-arm-kernel
When booting a 64 KB pages kernel on a ACPI GICv3 system that
implements support for v2 emulation, the following warning is
produced
GICV size 0x2000 not a multiple of page size 0x10000
and support for v2 emulation is disabled, preventing GICv2 VMs
from being able to run on such hosts.
The reason is that vgic_v3_probe() performs a sanity check on the
size of the window (it should be a multiple of the page size),
while the ACPI MADT parsing code hardcodes the size of the window
to 8 KB. This makes sense, considering that ACPI does not bother
to describe the size in the first place, under the assumption that
platforms implementing ACPI will follow the architecture and not
put anything else in the same 64 KB window.
So let's just drop the sanity check altogether, and assume that
the window is at least 64 KB in size.
Fixes: 909777324588 ("KVM: arm/arm64: vgic-new: vgic_init: implement kvm_vgic_hyp_init")
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
---
virt/kvm/arm/vgic/vgic-v3.c | 5 -----
1 file changed, 5 deletions(-)
diff --git a/virt/kvm/arm/vgic/vgic-v3.c b/virt/kvm/arm/vgic/vgic-v3.c
index bdcf8e7a6161..72fc688c3e9d 100644
--- a/virt/kvm/arm/vgic/vgic-v3.c
+++ b/virt/kvm/arm/vgic/vgic-v3.c
@@ -552,11 +552,6 @@ int vgic_v3_probe(const struct gic_kvm_info *info)
pr_warn("GICV physical address 0x%llx not page aligned\n",
(unsigned long long)info->vcpu.start);
kvm_vgic_global_state.vcpu_base = 0;
- } else if (!PAGE_ALIGNED(resource_size(&info->vcpu))) {
- pr_warn("GICV size 0x%llx not a multiple of page size 0x%lx\n",
- (unsigned long long)resource_size(&info->vcpu),
- PAGE_SIZE);
- kvm_vgic_global_state.vcpu_base = 0;
} else {
kvm_vgic_global_state.vcpu_base = info->vcpu.start;
kvm_vgic_global_state.can_emulate_gicv2 = true;
--
2.17.0
^ permalink raw reply related
* [PATCH] PCI: move early dump functionality from x86 arch into the common code
From: Sinan Kaya @ 2018-06-01 15:06 UTC (permalink / raw)
To: linux-arm-kernel
In-Reply-To: <CAHp75Vc3YKAR1P_F0fwD+ueGub4gVUjP0=wv4Egf_fTSVkSUjw@mail.gmail.com>
On 6/1/2018 11:02 AM, Andy Shevchenko wrote:
> On Wed, May 30, 2018 at 7:34 AM, Sinan Kaya <okaya@codeaurora.org> wrote:
>> Move early dump functionality into common code so that it is available for
>> all archtiectures. No need to carry arch specific reads around as the read
>> hooks are already initialized by the time pci_setup_device() is getting
>> called during scan.
>
>> /* If set, the PCIe ATS capability will not be used. */
>> static bool pcie_ats_disabled;
>>
>> +bool pci_early_dump;
>> +
>
> I didn't check above these, but maybe a good idea to add one line
> comment what is this about?
This was discussed here:
https://www.spinics.net/lists/linux-pci/msg72859.html
>
>
>> extern const unsigned char pcie_link_speed[];
>> -
>> +extern bool pci_early_dump;
>> bool pcie_cap_has_lnkctl(const struct pci_dev *dev);
>>
>
> Hmm... I would rather not attach this line to some function declarations.
Sure
>
>> +static void early_dump_pci_device(struct pci_dev *pdev)
>> +{
>> + u32 value[256 / 4];
>> + int i;
>> +
>> + dev_info(&pdev->dev, "pci 0000:%02x:%02x.%d config space:\n",
>> + pdev->bus->number, PCI_SLOT(pdev->devfn),
>> + PCI_FUNC(pdev->devfn));
>
> Shouldn't be this changed to pci_info() ?
Yeah, I need to get used to pci_info().
>
>> +
>> + for (i = 0; i < 256; i += 4)
>> + pci_read_config_dword(pdev, i, &value[i / 4]);
>> +
>> + print_hex_dump(KERN_INFO, "", DUMP_PREFIX_OFFSET, 16, 1, value,
>> + 256, false);
>> +}
>
--
Sinan Kaya
Qualcomm Datacenter Technologies, Inc. as an affiliate of Qualcomm Technologies, Inc.
Qualcomm Technologies, Inc. is a member of the Code Aurora Forum, a Linux Foundation Collaborative Project.
^ permalink raw reply
* [PATCH v6 0/2] arm64: signal: Report signal frame size to userspace via auxv
From: Catalin Marinas @ 2018-06-01 15:11 UTC (permalink / raw)
To: linux-arm-kernel
In-Reply-To: <1527847814-22829-1-git-send-email-Dave.Martin@arm.com>
On Fri, Jun 01, 2018 at 11:10:12AM +0100, Dave P Martin wrote:
> This series adds support for telling userspace the size of the signal
> frame via a new AT_MINSIGSTKSZ entry in the aux vector.
>
> This is an update to [1]. An omitted AT_MINSIGSTKSZ is now replaced
> with an AT_IGNORE entry instead of being collapsed out of the auxv
> entirely.
>
> See individual patches for the delta.
>
> [1] [PATCH v5 0/2] arm64: signal: Report signal frame size to userspace via auxv
> http://lists.infradead.org/pipermail/linux-arm-kernel/2018-May/580974.html
>
>
> Dave Martin (2):
> arm64/sve: Thin out initialisation sanity-checks for sve_max_vl
> arm64: signal: Report signal frame size to userspace via auxv
Queued for 4.18. Thanks.
--
Catalin
^ permalink raw reply
* [PATCH 06/15] arm: dts: sun: Add missing cooling device properties for CPUs
From: Maxime Ripard @ 2018-06-01 15:17 UTC (permalink / raw)
To: linux-arm-kernel
In-Reply-To: <20180528105734.bkplm7tfaqsurdrn@vireshk-i7>
Hi,
On Mon, May 28, 2018 at 04:27:34PM +0530, Viresh Kumar wrote:
> On 28-05-18, 10:44, Maxime Ripard wrote:
> > Hi,
> >
> > On Fri, May 25, 2018 at 04:01:52PM +0530, Viresh Kumar wrote:
> > > The cooling device properties, like "#cooling-cells" and
> > > "dynamic-power-coefficient", should either be present for all the CPUs
> > > of a cluster or none. If these are present only for a subset of CPUs of
> > > a cluster then things will start falling apart as soon as the CPUs are
> > > brought online in a different order. For example, this will happen
> > > because the operating system looks for such properties in the CPU node
> > > it is trying to bring up, so that it can register a cooling device.
> > >
> > > Add such missing properties.
> > >
> > > Fix other missing properties (clocks, OPP, clock latency) as well to
> > > make it all work.
> > >
> > > Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
> >
> > The prefix should be sunxi and not sun, but it looks good to me
> > otherwise.
> >
> > Let me know what your preferred merge method is.
>
> Please pick it up directly and send it as part of your pull request.
> Do you want me to resend or can you fix the $subject ?
I just tried to apply it, and it failed. We have usually two different
PR for the H3 SoCs and the others. Could you split the H3 in a
separate patch (and fix the subject in the process?)
Thanks!
Maxime
--
Maxime Ripard, Bootlin (formerly Free Electrons)
Embedded Linux and Kernel engineering
https://bootlin.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20180601/68b1a4e3/attachment.sig>
^ permalink raw reply
* [PATCH 0/4] lib/vsprintf: Remove atomic-unsafe support for printk format %pCr
From: Petr Mladek @ 2018-06-01 15:19 UTC (permalink / raw)
To: linux-arm-kernel
In-Reply-To: <20180601114738.kdoggkha2yosjgbv@pathway.suse.cz>
On Fri 2018-06-01 13:47:38, Petr Mladek wrote:
> On Fri 2018-06-01 06:00:47, Linus Torvalds wrote:
> > On Fri, Jun 1, 2018 at 4:29 AM Geert Uytterhoeven
> > <geert+renesas@glider.be> wrote:
> > >
> > > This patch series:
> > > - Changes all existing users of "%pCr" to print the result of
> > > clk_get_rate() directly, which is safe as they all do this in task
> > > context only,
> > > - Removes support for the "%pCr" printk format.
> >
> > Looks good to me.
> >
> > What tree will this go through? The normal printk one? Just checking
> > that this doesn't end up falling through the cracks because nobody
> > knows who would take it...
>
> I will take it via printk.git. There already is bunch of vsprintf
> changes for-4.18.
It is in printk.git, branch for-4.18-vsprintf-pcr-removal now.
Also I have added Cc: stable at vger.kernel.org into the commit messages.
Best Regards,
Petr
^ permalink raw reply
* [PATCH] PCI: move early dump functionality from x86 arch into the common code
From: Andy Shevchenko @ 2018-06-01 15:24 UTC (permalink / raw)
To: linux-arm-kernel
In-Reply-To: <53e4d6c3-ee19-fd04-4ba3-862d04558689@codeaurora.org>
On Fri, Jun 1, 2018 at 6:06 PM, Sinan Kaya <okaya@codeaurora.org> wrote:
> On 6/1/2018 11:02 AM, Andy Shevchenko wrote:
>> On Wed, May 30, 2018 at 7:34 AM, Sinan Kaya <okaya@codeaurora.org> wrote:
>>> Move early dump functionality into common code so that it is available for
>>> all archtiectures. No need to carry arch specific reads around as the read
>>> hooks are already initialized by the time pci_setup_device() is getting
>>> called during scan.
>>
>>> /* If set, the PCIe ATS capability will not be used. */
>>> static bool pcie_ats_disabled;
>>>
>>> +bool pci_early_dump;
>>> +
>>
>> I didn't check above these, but maybe a good idea to add one line
>> comment what is this about?
>
> This was discussed here:
>
> https://www.spinics.net/lists/linux-pci/msg72859.html
I understand what is it. What I'm pointing is the variable in the
source without any comment. In the context you have in diff previous
one has a comment.
--
With Best Regards,
Andy Shevchenko
^ permalink raw reply
* [PATCH] PCI: move early dump functionality from x86 arch into the common code
From: Sinan Kaya @ 2018-06-01 15:24 UTC (permalink / raw)
To: linux-arm-kernel
In-Reply-To: <CAHp75VdTyvQOqnSm67Uo5heP1otE98BbquM0TjSAorEzTgUwHQ@mail.gmail.com>
On 6/1/2018 11:24 AM, Andy Shevchenko wrote:
>> This was discussed here:
>>
>> https://www.spinics.net/lists/linux-pci/msg72859.html
> I understand what is it. What I'm pointing is the variable in the
> source without any comment. In the context you have in diff previous
> one has a comment.
oh, ok. I can take care of that too.
--
Sinan Kaya
Qualcomm Datacenter Technologies, Inc. as an affiliate of Qualcomm Technologies, Inc.
Qualcomm Technologies, Inc. is a member of the Code Aurora Forum, a Linux Foundation Collaborative Project.
^ permalink raw reply
* [GIT PULL] KVM/ARM updates for 4.18
From: Marc Zyngier @ 2018-06-01 15:27 UTC (permalink / raw)
To: linux-arm-kernel
Paolo, Radim,
This is the set of updates for KVM/ARM for v4.18. This time, lazy
context switching of the floating point registers for arm64, and the
ability to deal with multiple redistributor ranges (which allows for
larger numbers of vcpu).
Note that some additional KVM/ARM updates will come via the arm64 tree
with the SSBD patches. This will generate three conflicts that are
pretty simple to resolve (-next has a good example of the resolution).
Please pull,
M.
The following changes since commit 75bc37fefc4471e718ba8e651aa74673d4e0a9eb:
Linux 4.17-rc4 (2018-05-06 16:57:38 -1000)
are available in the Git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/kvmarm/kvmarm.git tags/kvmarm-for-v4.18
for you to fetch changes up to e25028c8ded011d19f9a11164807507c94febc01:
KVM: arm/arm64: Bump VGIC_V3_MAX_CPUS to 512 (2018-05-25 12:29:27 +0100)
----------------------------------------------------------------
KVM/ARM updates for 4.18
- Lazy context-switching of FPSIMD registers on arm64
- Allow virtual redistributors to be part of two or more MMIO ranges
----------------------------------------------------------------
Christoffer Dall (1):
KVM: arm/arm64: Introduce kvm_arch_vcpu_run_pid_change
Dave Martin (18):
arm64: fpsimd: Fix TIF_FOREIGN_FPSTATE after invalidating cpu regs
thread_info: Add update_thread_flag() helpers
arm64: Use update{,_tsk}_thread_flag()
KVM: arm64: Convert lazy FPSIMD context switch trap to C
arm64: fpsimd: Generalise context saving for non-task contexts
arm64: fpsimd: Avoid FPSIMD context leakage for the init task
arm64: fpsimd: Eliminate task->mm checks
arm64/sve: Refactor user SVE trap maintenance for external use
KVM: arm64: Repurpose vcpu_arch.debug_flags for general-purpose flags
KVM: arm64: Optimise FPSIMD handling to reduce guest/host thrashing
arm64/sve: Move read_zcr_features() out of cpufeature.h
arm64/sve: Switch sve_pffr() argument from task to thread
arm64/sve: Move sve_pffr() to fpsimd.h and make inline
KVM: arm64: Save host SVE context as appropriate
KVM: arm64: Remove eager host SVE state saving
KVM: arm64: Remove redundant *exit_code changes in fpsimd_guest_exit()
KVM: arm64: Fold redundant exit code checks out of fixup_guest_exit()
KVM: arm64: Invoke FPSIMD context switch trap from C
Eric Auger (13):
KVM: arm/arm64: Set dist->spis to NULL after kfree
KVM: arm/arm64: Document KVM_VGIC_V3_ADDR_TYPE_REDIST_REGION
KVM: arm/arm64: Replace the single rdist region by a list
KVM: arm/arm64: Helper to locate free rdist index
KVM: arm/arm64: Revisit Redistributor TYPER last bit computation
KVM: arm/arm64: Adapt vgic_v3_check_base to multiple rdist regions
KVM: arm/arm64: Helper to register a new redistributor region
KVM: arm/arm64: Remove kvm_vgic_vcpu_early_init
KVM: arm/arm64: Check vcpu redist base before registering an iodev
KVM: arm/arm64: Check all vcpu redistributors are set on map_resources
KVM: arm/arm64: Add KVM_VGIC_V3_ADDR_TYPE_REDIST_REGION
KVM: arm/arm64: Implement KVM_VGIC_V3_ADDR_TYPE_REDIST_REGION
KVM: arm/arm64: Bump VGIC_V3_MAX_CPUS to 512
Mark Rutland (1):
arm64: KVM: Use lm_alias() for kvm_ksym_ref()
Documentation/virtual/kvm/devices/arm-vgic-v3.txt | 30 +++-
arch/arm/include/asm/kvm_host.h | 10 +-
arch/arm/include/uapi/asm/kvm.h | 1 +
arch/arm64/Kconfig | 7 +
arch/arm64/include/asm/cpufeature.h | 29 ----
arch/arm64/include/asm/fpsimd.h | 21 +++
arch/arm64/include/asm/kvm_asm.h | 8 +-
arch/arm64/include/asm/kvm_host.h | 45 ++++--
arch/arm64/include/asm/processor.h | 15 +-
arch/arm64/include/asm/thread_info.h | 13 +-
arch/arm64/include/uapi/asm/kvm.h | 1 +
arch/arm64/kernel/fpsimd.c | 177 +++++++++++-----------
arch/arm64/kernel/ptrace.c | 1 +
arch/arm64/kvm/Kconfig | 1 +
arch/arm64/kvm/Makefile | 2 +-
arch/arm64/kvm/debug.c | 8 +-
arch/arm64/kvm/fpsimd.c | 110 ++++++++++++++
arch/arm64/kvm/hyp/debug-sr.c | 6 +-
arch/arm64/kvm/hyp/entry.S | 43 ------
arch/arm64/kvm/hyp/hyp-entry.S | 19 ---
arch/arm64/kvm/hyp/switch.c | 124 ++++++++++-----
arch/arm64/kvm/hyp/sysreg-sr.c | 4 +-
arch/arm64/kvm/sys_regs.c | 9 +-
include/kvm/arm_vgic.h | 17 ++-
include/linux/kvm_host.h | 9 ++
include/linux/sched.h | 6 +
include/linux/thread_info.h | 11 ++
virt/kvm/Kconfig | 3 +
virt/kvm/arm/arm.c | 15 +-
virt/kvm/arm/vgic/vgic-init.c | 100 ++++++------
virt/kvm/arm/vgic/vgic-kvm-device.c | 53 ++++++-
virt/kvm/arm/vgic/vgic-mmio-v3.c | 112 ++++++++++++--
virt/kvm/arm/vgic/vgic-v3.c | 99 +++++++++---
virt/kvm/arm/vgic/vgic.h | 42 ++++-
virt/kvm/kvm_main.c | 7 +-
35 files changed, 809 insertions(+), 349 deletions(-)
create mode 100644 arch/arm64/kvm/fpsimd.c
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox