openkylin
/
linux
mirror of https://gitee.com/openkylin/linux.git
9
0
Fork 0
Code Issues Projects Releases Wiki Activity

Merge drm/drm-next into drm-misc-next 

4.19 is out, Lyude asked for a backmerge, and it's been a while. All
very good reasons on their own :-)

Signed-off-by: Sean Paul <seanpaul@chromium.org>
This commit is contained in:
Sean Paul 2018-10-24 14:26:04 -04:00
parent 2b02a05bdc f2bfc71aee
commit 6542e9adc0
629 changed files with 13840 additions and 9844 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

25
Documentation/devicetree/bindings/display/exynos/exynos_dsim.txt
View File

@ -21,6 +21,9 @@ Required properties:
- samsung,pll-clock-frequency: specifies frequency of the oscillator clock
- #address-cells, #size-cells: should be set respectively to <1> and <0>
according to DSI host bindings (see MIPI DSI bindings [1])
- samsung,burst-clock-frequency: specifies DSI frequency in high-speed burst
mode
- samsung,esc-clock-frequency: specifies DSI frequency in escape mode
Optional properties:
- power-domains: a phandle to DSIM power domain node
@ -29,25 +32,9 @@ Child nodes:
Should contain DSI peripheral nodes (see MIPI DSI bindings [1]).
Video interfaces:
Device node can contain video interface port nodes according to [2].
The following are properties specific to those nodes:
port node inbound:
- reg: (required) must be 0.
port node outbound:
- reg: (required) must be 1.
endpoint node connected from mic node (reg = 0):
- remote-endpoint: specifies the endpoint in mic node. This node is required
for Exynos5433 mipi dsi. So mic can access to panel node
throughout this dsi node.
endpoint node connected to panel node (reg = 1):
- remote-endpoint: specifies the endpoint in panel node. This node is
required in all kinds of exynos mipi dsi to represent
the connection between mipi dsi and panel.
- samsung,burst-clock-frequency: specifies DSI frequency in high-speed burst
mode
- samsung,esc-clock-frequency: specifies DSI frequency in escape mode
Device node can contain following video interface port nodes according to [2]:
0: RGB input,
1: DSI output
[1]: Documentation/devicetree/bindings/display/mipi-dsi-bus.txt
[2]: Documentation/devicetree/bindings/media/video-interfaces.txt

2
Documentation/devicetree/bindings/input/gpio-keys.txt
View File

@ -1,4 +1,4 @@
Device-Tree bindings for input/gpio_keys.c keyboard driver
Device-Tree bindings for input/keyboard/gpio_keys.c keyboard driver
Required properties:
- compatible = "gpio-keys";

1
Documentation/media/uapi/dvb/video_function_calls.rst
View File

@ -33,4 +33,3 @@ Video Function Calls
video-clear-buffer
video-set-streamtype
video-set-format
video-set-attributes

24
MAINTAINERS
View File

@ -4761,8 +4761,8 @@ F: include/uapi/drm/
F: include/linux/vga*
DRM DRIVERS AND MISC GPU PATCHES
M: Gustavo Padovan <gustavo@padovan.org>
M: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
M: Maxime Ripard <maxime.ripard@bootlin.com>
M: Sean Paul <sean@poorly.run>
W: https://01.org/linuxgraphics/gfx-docs/maintainer-tools/drm-misc.html
S: Maintained
@ -9727,13 +9727,6 @@ Q: http://patchwork.linuxtv.org/project/linux-media/list/
S: Maintained
F: drivers/media/dvb-frontends/mn88473*
PCI DRIVER FOR MOBIVEIL PCIE IP
M: Subrahmanya Lingappa <l.subrahmanya@mobiveil.co.in>
L: linux-pci@vger.kernel.org
S: Supported
F: Documentation/devicetree/bindings/pci/mobiveil-pcie.txt
F: drivers/pci/controller/pcie-mobiveil.c
MODULE SUPPORT
M: Jessica Yu <jeyu@kernel.org>
T: git git://git.kernel.org/pub/scm/linux/kernel/git/jeyu/linux.git modules-next
@ -10963,7 +10956,7 @@ M: Willy Tarreau <willy@haproxy.com>
M: Ksenija Stanojevic <ksenija.stanojevic@gmail.com>
S: Odd Fixes
F: Documentation/auxdisplay/lcd-panel-cgram.txt
F: drivers/misc/panel.c
F: drivers/auxdisplay/panel.c
PARALLEL PORT SUBSYSTEM
M: Sudip Mukherjee <sudipm.mukherjee@gmail.com>
@ -11151,6 +11144,13 @@ F: include/uapi/linux/switchtec_ioctl.h
F: include/linux/switchtec.h
F: drivers/ntb/hw/mscc/
PCI DRIVER FOR MOBIVEIL PCIE IP
M: Subrahmanya Lingappa <l.subrahmanya@mobiveil.co.in>
L: linux-pci@vger.kernel.org
S: Supported
F: Documentation/devicetree/bindings/pci/mobiveil-pcie.txt
F: drivers/pci/controller/pcie-mobiveil.c
PCI DRIVER FOR MVEBU (Marvell Armada 370 and Armada XP SOC support)
M: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
M: Jason Cooper <jason@lakedaemon.net>
@ -11217,8 +11217,14 @@ F: tools/pci/
PCI ENHANCED ERROR HANDLING (EEH) FOR POWERPC
M: Russell Currey <ruscur@russell.cc>
M: Sam Bobroff <sbobroff@linux.ibm.com>
M: Oliver O'Halloran <oohall@gmail.com>
L: linuxppc-dev@lists.ozlabs.org
S: Supported
F: Documentation/PCI/pci-error-recovery.txt
F: drivers/pci/pcie/aer.c
F: drivers/pci/pcie/dpc.c
F: drivers/pci/pcie/err.c
F: Documentation/powerpc/eeh-pci-error-recovery.txt
F: arch/powerpc/kernel/eeh*.c
F: arch/powerpc/platforms/*/eeh*.c

2
Makefile
View File

@ -2,7 +2,7 @@
VERSION = 4
PATCHLEVEL = 19
SUBLEVEL = 0
EXTRAVERSION = -rc5
EXTRAVERSION = -rc6
NAME = Merciless Moray
# *DOCUMENTATION*

1
arch/powerpc/include/asm/setup.h
View File

@ -9,6 +9,7 @@ extern void ppc_printk_progress(char *s, unsigned short hex);
extern unsigned int rtas_data;
extern unsigned long long memory_limit;
extern bool init_mem_is_free;
extern unsigned long klimit;
extern void *zalloc_maybe_bootmem(size_t size, gfp_t mask);

4
arch/powerpc/kernel/exceptions-64s.S
View File

@ -1314,9 +1314,7 @@ EXC_REAL_BEGIN(denorm_exception_hv, 0x1500, 0x100)
#ifdef CONFIG_PPC_DENORMALISATION
mfspr r10,SPRN_HSRR1
mfspr r11,SPRN_HSRR0 /* save HSRR0 */
andis. r10,r10,(HSRR1_DENORM)@h /* denorm? */
addi r11,r11,-4 /* HSRR0 is next instruction */
bne+ denorm_assist
#endif
@ -1382,6 +1380,8 @@ END_FTR_SECTION_IFCLR(CPU_FTR_ARCH_207S)
*/
XVCPSGNDP32(32)
denorm_done:
mfspr r11,SPRN_HSRR0
subi r11,r11,4
mtspr SPRN_HSRR0,r11
mtcrf 0x80,r9
ld r9,PACA_EXGEN+EX_R9(r13)

20
arch/powerpc/kernel/tm.S
View File

@ -176,13 +176,27 @@ _GLOBAL(tm_reclaim)
std r1, PACATMSCRATCH(r13)
ld r1, PACAR1(r13)
/* Store the PPR in r11 and reset to decent value */
std r11, GPR11(r1) /* Temporary stash */
/*
* Move the saved user r1 to the kernel stack in case PACATMSCRATCH is
* clobbered by an exception once we turn on MSR_RI below.
*/
ld r11, PACATMSCRATCH(r13)
std r11, GPR1(r1)
/*
* Store r13 away so we can free up the scratch SPR for the SLB fault
* handler (needed once we start accessing the thread_struct).
*/
GET_SCRATCH0(r11)
std r11, GPR13(r1)
/* Reset MSR RI so we can take SLB faults again */
li r11, MSR_RI
mtmsrd r11, 1
/* Store the PPR in r11 and reset to decent value */
mfspr r11, SPRN_PPR
HMT_MEDIUM
@ -207,11 +221,11 @@ _GLOBAL(tm_reclaim)
SAVE_GPR(8, r7) /* user r8 */
SAVE_GPR(9, r7) /* user r9 */
SAVE_GPR(10, r7) /* user r10 */
ld r3, PACATMSCRATCH(r13) /* user r1 */
ld r3, GPR1(r1) /* user r1 */
ld r4, GPR7(r1) /* user r7 */
ld r5, GPR11(r1) /* user r11 */
ld r6, GPR12(r1) /* user r12 */
GET_SCRATCH0(8) /* user r13 */
ld r8, GPR13(r1) /* user r13 */
std r3, GPR1(r7)
std r4, GPR7(r7)
std r5, GPR11(r7)

3
arch/powerpc/lib/checksum_64.S
View File

@ -443,6 +443,9 @@ _GLOBAL(csum_ipv6_magic)
addc r0, r8, r9
ld r10, 0(r4)
ld r11, 8(r4)
#ifdef CONFIG_CPU_LITTLE_ENDIAN
rotldi r5, r5, 8
#endif
adde r0, r0, r10
add r5, r5, r7
adde r0, r0, r11

6
arch/powerpc/lib/code-patching.c
View File

@ -28,6 +28,12 @@ static int __patch_instruction(unsigned int *exec_addr, unsigned int instr,
{
int err;
/* Make sure we aren't patching a freed init section */
if (init_mem_is_free && init_section_contains(exec_addr, 4)) {
pr_debug("Skipping init section patching addr: 0x%px\n", exec_addr);
return 0;
}
__put_user_size(instr, patch_addr, 4, err);
if (err)
return err;

2
arch/powerpc/mm/mem.c
View File

@ -63,6 +63,7 @@
#endif
unsigned long long memory_limit;
bool init_mem_is_free;
#ifdef CONFIG_HIGHMEM
pte_t *kmap_pte;
@ -396,6 +397,7 @@ void free_initmem(void)
{
ppc_md.progress = ppc_printk_progress;
mark_initmem_nx();
init_mem_is_free = true;
free_initmem_default(POISON_FREE_INITMEM);
}

7
arch/powerpc/mm/numa.c
View File

@ -1204,7 +1204,9 @@ int find_and_online_cpu_nid(int cpu)
int new_nid;
/* Use associativity from first thread for all siblings */
vphn_get_associativity(cpu, associativity);
if (vphn_get_associativity(cpu, associativity))
return cpu_to_node(cpu);
new_nid = associativity_to_nid(associativity);
if (new_nid < 0 || !node_possible(new_nid))
new_nid = first_online_node;
@ -1452,7 +1454,8 @@ static struct timer_list topology_timer;
static void reset_topology_timer(void)
{
mod_timer(&topology_timer, jiffies + topology_timer_secs * HZ);
if (vphn_enabled)
mod_timer(&topology_timer, jiffies + topology_timer_secs * HZ);
}
#ifdef CONFIG_SMP

2
arch/powerpc/mm/pkeys.c
View File

@ -45,7 +45,7 @@ static void scan_pkey_feature(void)
* Since any pkey can be used for data or execute, we will just treat
* all keys as equal and track them as one entity.
*/
pkeys_total = be32_to_cpu(vals[0]);
pkeys_total = vals[0];
pkeys_devtree_defined = true;
}

2
arch/powerpc/platforms/powernv/pci-ioda-tce.c
View File

@ -276,7 +276,7 @@ long pnv_pci_ioda2_table_alloc_pages(int nid, __u64 bus_offset,
level_shift = entries_shift + 3;
level_shift = max_t(unsigned int, level_shift, PAGE_SHIFT);
if ((level_shift - 3) * levels + page_shift >= 60)
if ((level_shift - 3) * levels + page_shift >= 55)
return -EINVAL;
/* Allocate TCE table */

7
arch/riscv/include/asm/asm-prototypes.h Normal file
View File

@ -0,0 +1,7 @@
/* SPDX-License-Identifier: GPL-2.0 */
#ifndef _ASM_RISCV_PROTOTYPES_H
#include <linux/ftrace.h>
#include <asm-generic/asm-prototypes.h>
#endif /* _ASM_RISCV_PROTOTYPES_H */

19
arch/x86/boot/compressed/mem_encrypt.S
View File

@ -25,20 +25,6 @@ ENTRY(get_sev_encryption_bit)
push %ebx
push %ecx
push %edx
push %edi
/*
* RIP-relative addressing is needed to access the encryption bit
* variable. Since we are running in 32-bit mode we need this call/pop
* sequence to get the proper relative addressing.
*/
call 1f
1: popl %edi
subl $1b, %edi
movl enc_bit(%edi), %eax
cmpl $0, %eax
jge .Lsev_exit
/* Check if running under a hypervisor */
movl $1, %eax
@ -69,15 +55,12 @@ ENTRY(get_sev_encryption_bit)
movl %ebx, %eax
andl $0x3f, %eax /* Return the encryption bit location */
movl %eax, enc_bit(%edi)
jmp .Lsev_exit
.Lno_sev:
xor %eax, %eax
movl %eax, enc_bit(%edi)
.Lsev_exit:
pop %edi
pop %edx
pop %ecx
pop %ebx
@ -113,8 +96,6 @@ ENTRY(set_sev_encryption_mask)
ENDPROC(set_sev_encryption_mask)
.data
enc_bit:
.int 0xffffffff
#ifdef CONFIG_AMD_MEM_ENCRYPT
.balign 8

13
block/blk-mq-tag.c
View File

@ -322,16 +322,11 @@ void blk_mq_queue_tag_busy_iter(struct request_queue *q, busy_iter_fn *fn,
/*
* __blk_mq_update_nr_hw_queues will update the nr_hw_queues and
* queue_hw_ctx after freeze the queue. So we could use q_usage_counter
* to avoid race with it. __blk_mq_update_nr_hw_queues will users
* synchronize_rcu to ensure all of the users go out of the critical
* section below and see zeroed q_usage_counter.
* queue_hw_ctx after freeze the queue, so we use q_usage_counter
* to avoid race with it.
*/
rcu_read_lock();
if (percpu_ref_is_zero(&q->q_usage_counter)) {
rcu_read_unlock();
if (!percpu_ref_tryget(&q->q_usage_counter))
return;
}
queue_for_each_hw_ctx(q, hctx, i) {
struct blk_mq_tags *tags = hctx->tags;
@ -347,7 +342,7 @@ void blk_mq_queue_tag_busy_iter(struct request_queue *q, busy_iter_fn *fn,
bt_for_each(hctx, &tags->breserved_tags, fn, priv, true);
bt_for_each(hctx, &tags->bitmap_tags, fn, priv, false);
}
rcu_read_unlock();
blk_queue_exit(q);
}
static int bt_alloc(struct sbitmap_queue *bt, unsigned int depth,

4
block/blk-mq.c
View File

@ -1628,7 +1628,7 @@ void blk_mq_flush_plug_list(struct blk_plug *plug, bool from_schedule)
BUG_ON(!rq->q);
if (rq->mq_ctx != this_ctx) {
if (this_ctx) {
trace_block_unplug(this_q, depth, from_schedule);
trace_block_unplug(this_q, depth, !from_schedule);
blk_mq_sched_insert_requests(this_q, this_ctx,
&ctx_list,
from_schedule);
@ -1648,7 +1648,7 @@ void blk_mq_flush_plug_list(struct blk_plug *plug, bool from_schedule)
* on 'ctx_list'. Do those.
*/
if (this_ctx) {
trace_block_unplug(this_q, depth, from_schedule);
trace_block_unplug(this_q, depth, !from_schedule);
blk_mq_sched_insert_requests(this_q, this_ctx, &ctx_list,
from_schedule);
}

2
block/elevator.c
View File

@ -609,7 +609,7 @@ void elv_drain_elevator(struct request_queue *q)
while (e->type->ops.sq.elevator_dispatch_fn(q, 1))
;
if (q->nr_sorted && printed++ < 10) {
if (q->nr_sorted && !blk_queue_is_zoned(q) && printed++ < 10 ) {
printk(KERN_ERR "%s: forced dispatching is broken "
"(nr_sorted=%u), please report this\n",
q->elevator->type->elevator_name, q->nr_sorted);

4
drivers/block/xen-blkfront.c
View File

@ -2670,8 +2670,8 @@ static void purge_persistent_grants(struct blkfront_info *info)
list_del(&gnt_list_entry->node);
gnttab_end_foreign_access(gnt_list_entry->gref, 0, 0UL);
rinfo->persistent_gnts_c--;
__free_page(gnt_list_entry->page);
kfree(gnt_list_entry);
gnt_list_entry->gref = GRANT_INVALID_REF;
list_add_tail(&gnt_list_entry->node, &rinfo->grants);
}
spin_unlock_irqrestore(&rinfo->ring_lock, flags);

20
drivers/clocksource/timer-atmel-pit.c
View File

@ -180,26 +180,29 @@ static int __init at91sam926x_pit_dt_init(struct device_node *node)
data->base = of_iomap(node, 0);
if (!data->base) {
pr_err("Could not map PIT address\n");
return -ENXIO;
ret = -ENXIO;
goto exit;
}
data->mck = of_clk_get(node, 0);
if (IS_ERR(data->mck)) {
pr_err("Unable to get mck clk\n");
return PTR_ERR(data->mck);
ret = PTR_ERR(data->mck);
goto exit;
}
ret = clk_prepare_enable(data->mck);
if (ret) {
pr_err("Unable to enable mck\n");
return ret;
goto exit;
}
/* Get the interrupts property */
data->irq = irq_of_parse_and_map(node, 0);
if (!data->irq) {
pr_err("Unable to get IRQ from DT\n");
return -EINVAL;
ret = -EINVAL;
goto exit;
}
/*
@ -227,7 +230,7 @@ static int __init at91sam926x_pit_dt_init(struct device_node *node)
ret = clocksource_register_hz(&data->clksrc, pit_rate);
if (ret) {
pr_err("Failed to register clocksource\n");
return ret;
goto exit;
}
/* Set up irq handler */
@ -236,7 +239,8 @@ static int __init at91sam926x_pit_dt_init(struct device_node *node)
"at91_tick", data);
if (ret) {
pr_err("Unable to setup IRQ\n");
return ret;
clocksource_unregister(&data->clksrc);
goto exit;
}
/* Set up and register clockevents */
@ -254,6 +258,10 @@ static int __init at91sam926x_pit_dt_init(struct device_node *node)
clockevents_register_device(&data->clkevt);
return 0;
exit:
kfree(data);
return ret;
}
TIMER_OF_DECLARE(at91sam926x_pit, "atmel,at91sam9260-pit",
at91sam926x_pit_dt_init);

18
drivers/clocksource/timer-fttmr010.c
View File

@ -130,13 +130,17 @@ static int fttmr010_timer_set_next_event(unsigned long cycles,
cr &= ~fttmr010->t1_enable_val;
writel(cr, fttmr010->base + TIMER_CR);
/* Setup the match register forward/backward in time */
cr = readl(fttmr010->base + TIMER1_COUNT);
if (fttmr010->count_down)
cr -= cycles;
else
cr += cycles;
writel(cr, fttmr010->base + TIMER1_MATCH1);
if (fttmr010->count_down) {
/*
* ASPEED Timer Controller will load TIMER1_LOAD register
* into TIMER1_COUNT register when the timer is re-enabled.
*/
writel(cycles, fttmr010->base + TIMER1_LOAD);
} else {
/* Setup the match register forward in time */
cr = readl(fttmr010->base + TIMER1_COUNT);
writel(cr + cycles, fttmr010->base + TIMER1_MATCH1);
}
/* Start */
cr = readl(fttmr010->base + TIMER_CR);

3
drivers/clocksource/timer-ti-32k.c
View File

@ -97,6 +97,9 @@ static int __init ti_32k_timer_init(struct device_node *np)
return -ENXIO;
}
if (!of_machine_is_compatible("ti,am43"))
ti_32k_timer.cs.flags |= CLOCK_SOURCE_SUSPEND_NONSTOP;
ti_32k_timer.counter = ti_32k_timer.base;
/*

4
drivers/cpufreq/qcom-cpufreq-kryo.c
View File

@ -44,7 +44,7 @@ enum _msm8996_version {
struct platform_device *cpufreq_dt_pdev, *kryo_cpufreq_pdev;
static enum _msm8996_version __init qcom_cpufreq_kryo_get_msm_id(void)
static enum _msm8996_version qcom_cpufreq_kryo_get_msm_id(void)
{
size_t len;
u32 *msm_id;
@ -222,7 +222,7 @@ static int __init qcom_cpufreq_kryo_init(void)
}
module_init(qcom_cpufreq_kryo_init);
static void __init qcom_cpufreq_kryo_exit(void)
static void __exit qcom_cpufreq_kryo_exit(void)
{
platform_device_unregister(kryo_cpufreq_pdev);
platform_driver_unregister(&qcom_cpufreq_kryo_driver);

6
drivers/dax/device.c
View File

@ -535,6 +535,11 @@ static unsigned long dax_get_unmapped_area(struct file *filp,
return current->mm->get_unmapped_area(filp, addr, len, pgoff, flags);
}
static const struct address_space_operations dev_dax_aops = {
.set_page_dirty = noop_set_page_dirty,
.invalidatepage = noop_invalidatepage,
};
static int dax_open(struct inode *inode, struct file *filp)
{
struct dax_device *dax_dev = inode_dax(inode);
@ -544,6 +549,7 @@ static int dax_open(struct inode *inode, struct file *filp)
dev_dbg(&dev_dax->dev, "trace\n");
inode->i_mapping = __dax_inode->i_mapping;
inode->i_mapping->host = __dax_inode;
inode->i_mapping->a_ops = &dev_dax_aops;
filp->f_mapping = inode->i_mapping;
filp->f_wb_err = filemap_sample_wb_err(filp->f_mapping);
filp->private_data = dev_dax;

20
drivers/gpu/drm/Kconfig
View File

@ -110,6 +110,26 @@ config DRM_FBDEV_OVERALLOC
is 100. Typical values for double buffering will be 200,
triple buffering 300.
config DRM_FBDEV_LEAK_PHYS_SMEM
bool "Shamelessly allow leaking of fbdev physical address (DANGEROUS)"
depends on DRM_FBDEV_EMULATION && EXPERT
default n
help
In order to keep user-space compatibility, we want in certain
use-cases to keep leaking the fbdev physical address to the
user-space program handling the fbdev buffer.
This affects, not only, Amlogic, Allwinner or Rockchip devices
with ARM Mali GPUs using an userspace Blob.
This option is not supported by upstream developers and should be
removed as soon as possible and be considered as a broken and
legacy behaviour from a modern fbdev device driver.
Please send any bug reports when using this to your proprietary
software vendor that requires this.
If in doubt, say "N" or spread the word to your closed source
library vendor.
config DRM_LOAD_EDID_FIRMWARE
bool "Allow to specify an EDID data set instead of probing for it"
depends on DRM

73
drivers/gpu/drm/amd/amdgpu/amdgpu.h
View File

@ -81,6 +81,23 @@
#include "amdgpu_bo_list.h"
#include "amdgpu_gem.h"
#define MAX_GPU_INSTANCE 16
struct amdgpu_gpu_instance
{
struct amdgpu_device *adev;
int mgpu_fan_enabled;
};
struct amdgpu_mgpu_info
{
struct amdgpu_gpu_instance gpu_ins[MAX_GPU_INSTANCE];
struct mutex mutex;
uint32_t num_gpu;
uint32_t num_dgpu;
uint32_t num_apu;
};
/*
* Modules parameters.
*/
@ -134,6 +151,7 @@ extern int amdgpu_compute_multipipe;
extern int amdgpu_gpu_recovery;
extern int amdgpu_emu_mode;
extern uint amdgpu_smu_memory_pool_size;
extern struct amdgpu_mgpu_info mgpu_info;
#ifdef CONFIG_DRM_AMDGPU_SI
extern int amdgpu_si_support;
@ -146,6 +164,7 @@ extern int amdgpu_cik_support;
#define AMDGPU_DEFAULT_GTT_SIZE_MB 3072ULL /* 3GB by default */
#define AMDGPU_WAIT_IDLE_TIMEOUT_IN_MS 3000
#define AMDGPU_MAX_USEC_TIMEOUT 100000 /* 100 ms */
#define AMDGPU_FENCE_JIFFIES_TIMEOUT (HZ / 2)
/* AMDGPU_IB_POOL_SIZE must be a power of 2 */
#define AMDGPU_IB_POOL_SIZE 16
#define AMDGPU_DEBUGFS_MAX_COMPONENTS 32
@ -408,16 +427,25 @@ typedef enum _AMDGPU_DOORBELL64_ASSIGNMENT
AMDGPU_DOORBELL64_GFX_RING0 = 0x8b,
/*
* Other graphics doorbells can be allocated here: from 0x8c to 0xef
* Other graphics doorbells can be allocated here: from 0x8c to 0xdf
* Graphics voltage island aperture 1
* default non-graphics QWORD index is 0xF0 - 0xFF inclusive
* default non-graphics QWORD index is 0xe0 - 0xFF inclusive
*/
/* sDMA engines */
AMDGPU_DOORBELL64_sDMA_ENGINE0 = 0xF0,
AMDGPU_DOORBELL64_sDMA_HI_PRI_ENGINE0 = 0xF1,
AMDGPU_DOORBELL64_sDMA_ENGINE1 = 0xF2,
AMDGPU_DOORBELL64_sDMA_HI_PRI_ENGINE1 = 0xF3,
/* sDMA engines reserved from 0xe0 -oxef */
AMDGPU_DOORBELL64_sDMA_ENGINE0 = 0xE0,
AMDGPU_DOORBELL64_sDMA_HI_PRI_ENGINE0 = 0xE1,
AMDGPU_DOORBELL64_sDMA_ENGINE1 = 0xE8,
AMDGPU_DOORBELL64_sDMA_HI_PRI_ENGINE1 = 0xE9,
/* For vega10 sriov, the sdma doorbell must be fixed as follow
* to keep the same setting with host driver, or it will
* happen conflicts
*/
AMDGPU_VEGA10_DOORBELL64_sDMA_ENGINE0 = 0xF0,
AMDGPU_VEGA10_DOORBELL64_sDMA_HI_PRI_ENGINE0 = 0xF1,
AMDGPU_VEGA10_DOORBELL64_sDMA_ENGINE1 = 0xF2,
AMDGPU_VEGA10_DOORBELL64_sDMA_HI_PRI_ENGINE1 = 0xF3,
/* Interrupt handler */
AMDGPU_DOORBELL64_IH = 0xF4, /* For legacy interrupt ring buffer */
@ -588,31 +616,6 @@ void amdgpu_benchmark(struct amdgpu_device *adev, int test_number);
*/
void amdgpu_test_moves(struct amdgpu_device *adev);
/*
* amdgpu smumgr functions
*/
struct amdgpu_smumgr_funcs {
int (*check_fw_load_finish)(struct amdgpu_device *adev, uint32_t fwtype);
int (*request_smu_load_fw)(struct amdgpu_device *adev);
int (*request_smu_specific_fw)(struct amdgpu_device *adev, uint32_t fwtype);
};
/*
* amdgpu smumgr
*/
struct amdgpu_smumgr {
struct amdgpu_bo *toc_buf;
struct amdgpu_bo *smu_buf;
/* asic priv smu data */
void *priv;
spinlock_t smu_lock;
/* smumgr functions */
const struct amdgpu_smumgr_funcs *smumgr_funcs;
/* ucode loading complete flag */
uint32_t fw_flags;
};
/*
* ASIC specific register table accessible by UMD
*/
@ -948,9 +951,6 @@ struct amdgpu_device {
u32 cg_flags;
u32 pg_flags;
/* amdgpu smumgr */
struct amdgpu_smumgr smu;
/* gfx */
struct amdgpu_gfx gfx;
@ -1015,6 +1015,9 @@ struct amdgpu_device {
bool has_hw_reset;
u8 reset_magic[AMDGPU_RESET_MAGIC_NUM];
/* s3/s4 mask */
bool in_suspend;
/* record last mm index being written through WREG32*/
unsigned long last_mm_index;
bool in_gpu_reset;

17
drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c
View File

@ -359,7 +359,9 @@ static int amdgpu_atif_get_sbios_requests(struct amdgpu_atif *atif,
*
* Checks the acpi event and if it matches an atif event,
* handles it.
* Returns NOTIFY code
*
* Returns:
* NOTIFY_BAD or NOTIFY_DONE, depending on the event.
*/
static int amdgpu_atif_handler(struct amdgpu_device *adev,
struct acpi_bus_event *event)
@ -373,11 +375,16 @@ static int amdgpu_atif_handler(struct amdgpu_device *adev,
if (strcmp(event->device_class, ACPI_VIDEO_CLASS) != 0)
return NOTIFY_DONE;
/* Is this actually our event? */
if (!atif ||
!atif->notification_cfg.enabled ||
event->type != atif->notification_cfg.command_code)
/* Not our event */
return NOTIFY_DONE;
event->type != atif->notification_cfg.command_code) {
/* These events will generate keypresses otherwise */
if (event->type == ACPI_VIDEO_NOTIFY_PROBE)
return NOTIFY_BAD;
else
return NOTIFY_DONE;
}
if (atif->functions.sbios_requests) {
struct atif_sbios_requests req;
@ -386,7 +393,7 @@ static int amdgpu_atif_handler(struct amdgpu_device *adev,
count = amdgpu_atif_get_sbios_requests(atif, &req);
if (count <= 0)
return NOTIFY_DONE;
return NOTIFY_BAD;
DRM_DEBUG_DRIVER("ATIF: %d pending SBIOS requests\n", count);

50
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
View File

@ -76,6 +76,7 @@ void amdgpu_amdkfd_device_probe(struct amdgpu_device *adev)
kfd2kgd = amdgpu_amdkfd_gfx_8_0_get_functions();
break;
case CHIP_VEGA10:
case CHIP_VEGA20:
case CHIP_RAVEN:
kfd2kgd = amdgpu_amdkfd_gfx_9_0_get_functions();
break;
@ -123,7 +124,7 @@ static void amdgpu_doorbell_get_kfd_info(struct amdgpu_device *adev,
void amdgpu_amdkfd_device_init(struct amdgpu_device *adev)
{
int i;
int i, n;
int last_valid_bit;
if (adev->kfd) {
struct kgd2kfd_shared_resources gpu_resources = {
@ -162,7 +163,15 @@ void amdgpu_amdkfd_device_init(struct amdgpu_device *adev)
&gpu_resources.doorbell_physical_address,
&gpu_resources.doorbell_aperture_size,
&gpu_resources.doorbell_start_offset);
if (adev->asic_type >= CHIP_VEGA10) {
if (adev->asic_type < CHIP_VEGA10) {
kgd2kfd->device_init(adev->kfd, &gpu_resources);
return;
}
n = (adev->asic_type < CHIP_VEGA20) ? 2 : 8;
for (i = 0; i < n; i += 2) {
/* On SOC15 the BIF is involved in routing
* doorbells using the low 12 bits of the
* address. Communicate the assignments to
@ -170,20 +179,31 @@ void amdgpu_amdkfd_device_init(struct amdgpu_device *adev)
* process in case of 64-bit doorbells so we
* can use each doorbell assignment twice.
*/
gpu_resources.sdma_doorbell[0][0] =
AMDGPU_DOORBELL64_sDMA_ENGINE0;
gpu_resources.sdma_doorbell[0][1] =
AMDGPU_DOORBELL64_sDMA_ENGINE0 + 0x200;
gpu_resources.sdma_doorbell[1][0] =
AMDGPU_DOORBELL64_sDMA_ENGINE1;
gpu_resources.sdma_doorbell[1][1] =
AMDGPU_DOORBELL64_sDMA_ENGINE1 + 0x200;
/* Doorbells 0x0f0-0ff and 0x2f0-2ff are reserved for
* SDMA, IH and VCN. So don't use them for the CP.
*/
gpu_resources.reserved_doorbell_mask = 0x1f0;
gpu_resources.reserved_doorbell_val = 0x0f0;
if (adev->asic_type == CHIP_VEGA10) {
gpu_resources.sdma_doorbell[0][i] =
AMDGPU_VEGA10_DOORBELL64_sDMA_ENGINE0 + (i >> 1);
gpu_resources.sdma_doorbell[0][i+1] =
AMDGPU_VEGA10_DOORBELL64_sDMA_ENGINE0 + 0x200 + (i >> 1);
gpu_resources.sdma_doorbell[1][i] =
AMDGPU_VEGA10_DOORBELL64_sDMA_ENGINE1 + (i >> 1);
gpu_resources.sdma_doorbell[1][i+1] =
AMDGPU_VEGA10_DOORBELL64_sDMA_ENGINE1 + 0x200 + (i >> 1);
} else {
gpu_resources.sdma_doorbell[0][i] =
AMDGPU_DOORBELL64_sDMA_ENGINE0 + (i >> 1);
gpu_resources.sdma_doorbell[0][i+1] =
AMDGPU_DOORBELL64_sDMA_ENGINE0 + 0x200 + (i >> 1);
gpu_resources.sdma_doorbell[1][i] =
AMDGPU_DOORBELL64_sDMA_ENGINE1 + (i >> 1);
gpu_resources.sdma_doorbell[1][i+1] =
AMDGPU_DOORBELL64_sDMA_ENGINE1 + 0x200 + (i >> 1);
}
}
/* Doorbells 0x0e0-0ff and 0x2e0-2ff are reserved for
* SDMA, IH and VCN. So don't use them for the CP.
*/
gpu_resources.reserved_doorbell_mask = 0x1e0;
gpu_resources.reserved_doorbell_val = 0x0e0;
kgd2kfd->device_init(adev->kfd, &gpu_resources);
}

2
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
View File

@ -174,7 +174,7 @@ void amdgpu_amdkfd_gpuvm_destroy_cb(struct amdgpu_device *adev,
struct amdgpu_vm *vm);
void amdgpu_amdkfd_gpuvm_destroy_process_vm(struct kgd_dev *kgd, void *vm);
void amdgpu_amdkfd_gpuvm_release_process_vm(struct kgd_dev *kgd, void *vm);
uint32_t amdgpu_amdkfd_gpuvm_get_process_page_dir(void *vm);
uint64_t amdgpu_amdkfd_gpuvm_get_process_page_dir(void *vm);
int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu(
struct kgd_dev *kgd, uint64_t va, uint64_t size,
void *vm, struct kgd_mem **mem,

7
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c
View File

@ -142,7 +142,7 @@ static uint16_t get_fw_version(struct kgd_dev *kgd, enum kgd_engine_type type);
static void set_scratch_backing_va(struct kgd_dev *kgd,
uint64_t va, uint32_t vmid);
static void set_vm_context_page_table_base(struct kgd_dev *kgd, uint32_t vmid,
uint32_t page_table_base);
uint64_t page_table_base);
static int invalidate_tlbs(struct kgd_dev *kgd, uint16_t pasid);
static int invalidate_tlbs_vmid(struct kgd_dev *kgd, uint16_t vmid);
static uint32_t read_vmid_from_vmfault_reg(struct kgd_dev *kgd);
@ -874,7 +874,7 @@ static uint16_t get_fw_version(struct kgd_dev *kgd, enum kgd_engine_type type)
}
static void set_vm_context_page_table_base(struct kgd_dev *kgd, uint32_t vmid,
uint32_t page_table_base)
uint64_t page_table_base)
{
struct amdgpu_device *adev = get_amdgpu_device(kgd);
@ -882,7 +882,8 @@ static void set_vm_context_page_table_base(struct kgd_dev *kgd, uint32_t vmid,
pr_err("trying to set page table base for wrong VMID\n");
return;
}
WREG32(mmVM_CONTEXT8_PAGE_TABLE_BASE_ADDR + vmid - 8, page_table_base);
WREG32(mmVM_CONTEXT8_PAGE_TABLE_BASE_ADDR + vmid - 8,
lower_32_bits(page_table_base));
}
static int invalidate_tlbs(struct kgd_dev *kgd, uint16_t pasid)

12
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c
View File

@ -45,8 +45,6 @@ enum hqd_dequeue_request_type {
RESET_WAVES
};
struct vi_sdma_mqd;
/*
* Register access functions
*/
@ -100,7 +98,7 @@ static uint16_t get_fw_version(struct kgd_dev *kgd, enum kgd_engine_type type);
static void set_scratch_backing_va(struct kgd_dev *kgd,
uint64_t va, uint32_t vmid);
static void set_vm_context_page_table_base(struct kgd_dev *kgd, uint32_t vmid,
uint32_t page_table_base);
uint64_t page_table_base);
static int invalidate_tlbs(struct kgd_dev *kgd, uint16_t pasid);
static int invalidate_tlbs_vmid(struct kgd_dev *kgd, uint16_t vmid);
@ -282,7 +280,8 @@ static int kgd_init_interrupts(struct kgd_dev *kgd, uint32_t pipe_id)
lock_srbm(kgd, mec, pipe, 0, 0);
WREG32(mmCPC_INT_CNTL, CP_INT_CNTL_RING0__TIME_STAMP_INT_ENABLE_MASK);
WREG32(mmCPC_INT_CNTL, CP_INT_CNTL_RING0__TIME_STAMP_INT_ENABLE_MASK |
CP_INT_CNTL_RING0__OPCODE_ERROR_INT_ENABLE_MASK);
unlock_srbm(kgd);
@ -834,7 +833,7 @@ static uint16_t get_fw_version(struct kgd_dev *kgd, enum kgd_engine_type type)
}
static void set_vm_context_page_table_base(struct kgd_dev *kgd, uint32_t vmid,
uint32_t page_table_base)
uint64_t page_table_base)
{
struct amdgpu_device *adev = get_amdgpu_device(kgd);
@ -842,7 +841,8 @@ static void set_vm_context_page_table_base(struct kgd_dev *kgd, uint32_t vmid,
pr_err("trying to set page table base for wrong VMID\n");
return;
}
WREG32(mmVM_CONTEXT8_PAGE_TABLE_BASE_ADDR + vmid - 8, page_table_base);
WREG32(mmVM_CONTEXT8_PAGE_TABLE_BASE_ADDR + vmid - 8,
lower_32_bits(page_table_base));
}
static int invalidate_tlbs(struct kgd_dev *kgd, uint16_t pasid)

7
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v9.c
View File

@ -138,7 +138,7 @@ static bool get_atc_vmid_pasid_mapping_valid(struct kgd_dev *kgd,
static uint16_t get_atc_vmid_pasid_mapping_pasid(struct kgd_dev *kgd,
uint8_t vmid);
static void set_vm_context_page_table_base(struct kgd_dev *kgd, uint32_t vmid,
uint32_t page_table_base);
uint64_t page_table_base);
static uint16_t get_fw_version(struct kgd_dev *kgd, enum kgd_engine_type type);
static void set_scratch_backing_va(struct kgd_dev *kgd,
uint64_t va, uint32_t vmid);
@ -1013,11 +1013,10 @@ static uint16_t get_fw_version(struct kgd_dev *kgd, enum kgd_engine_type type)
}
static void set_vm_context_page_table_base(struct kgd_dev *kgd, uint32_t vmid,
uint32_t page_table_base)
uint64_t page_table_base)
{
struct amdgpu_device *adev = get_amdgpu_device(kgd);
uint64_t base = (uint64_t)page_table_base << PAGE_SHIFT |
AMDGPU_PTE_VALID;
uint64_t base = page_table_base | AMDGPU_PTE_VALID;
if (!amdgpu_amdkfd_is_kfd_vmid(adev, vmid)) {
pr_err("trying to set page table base for wrong VMID %u\n",

8
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
View File

@ -1131,11 +1131,15 @@ void amdgpu_amdkfd_gpuvm_release_process_vm(struct kgd_dev *kgd, void *vm)
amdgpu_vm_release_compute(adev, avm);
}
uint32_t amdgpu_amdkfd_gpuvm_get_process_page_dir(void *vm)
uint64_t amdgpu_amdkfd_gpuvm_get_process_page_dir(void *vm)
{
struct amdgpu_vm *avm = (struct amdgpu_vm *)vm;
struct amdgpu_bo *pd = avm->root.base.bo;
struct amdgpu_device *adev = amdgpu_ttm_adev(pd->tbo.bdev);
return avm->pd_phys_addr >> AMDGPU_GPU_PAGE_SHIFT;
if (adev->asic_type < CHIP_VEGA10)
return avm->pd_phys_addr >> AMDGPU_GPU_PAGE_SHIFT;
return avm->pd_phys_addr;
}
int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu(

12
drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c
View File

@ -826,21 +826,13 @@ int amdgpu_debugfs_regs_init(struct amdgpu_device *adev)
{
struct drm_minor *minor = adev->ddev->primary;
struct dentry *ent, *root = minor->debugfs_root;
unsigned i, j;
unsigned int i;
for (i = 0; i < ARRAY_SIZE(debugfs_regs); i++) {
ent = debugfs_create_file(debugfs_regs_names[i],
S_IFREG | S_IRUGO, root,
adev, debugfs_regs[i]);
if (IS_ERR(ent)) {
for (j = 0; j < i; j++) {
debugfs_remove(adev->debugfs_regs[i]);
adev->debugfs_regs[i] = NULL;
}
return PTR_ERR(ent);
}
if (!i)
if (!i && !IS_ERR_OR_NULL(ent))
i_size_write(ent->d_inode, adev->rmmio_size);
adev->debugfs_regs[i] = ent;
}

183
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
View File

@ -1525,6 +1525,92 @@ static int amdgpu_device_ip_early_init(struct amdgpu_device *adev)
return 0;
}
static int amdgpu_device_ip_hw_init_phase1(struct amdgpu_device *adev)
{
int i, r;
for (i = 0; i < adev->num_ip_blocks; i++) {
if (!adev->ip_blocks[i].status.sw)
continue;
if (adev->ip_blocks[i].status.hw)
continue;
if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_COMMON ||
adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_IH) {
r = adev->ip_blocks[i].version->funcs->hw_init(adev);
if (r) {
DRM_ERROR("hw_init of IP block <%s> failed %d\n",
adev->ip_blocks[i].version->funcs->name, r);
return r;
}
adev->ip_blocks[i].status.hw = true;
}
}
return 0;
}
static int amdgpu_device_ip_hw_init_phase2(struct amdgpu_device *adev)
{
int i, r;
for (i = 0; i < adev->num_ip_blocks; i++) {
if (!adev->ip_blocks[i].status.sw)
continue;
if (adev->ip_blocks[i].status.hw)
continue;
r = adev->ip_blocks[i].version->funcs->hw_init(adev);
if (r) {
DRM_ERROR("hw_init of IP block <%s> failed %d\n",
adev->ip_blocks[i].version->funcs->name, r);
return r;
}
adev->ip_blocks[i].status.hw = true;
}
return 0;
}
static int amdgpu_device_fw_loading(struct amdgpu_device *adev)
{
int r = 0;
int i;
if (adev->asic_type >= CHIP_VEGA10) {
for (i = 0; i < adev->num_ip_blocks; i++) {
if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_PSP) {
if (adev->in_gpu_reset || adev->in_suspend) {
if (amdgpu_sriov_vf(adev) && adev->in_gpu_reset)
break; /* sriov gpu reset, psp need to do hw_init before IH because of hw limit */
r = adev->ip_blocks[i].version->funcs->resume(adev);
if (r) {
DRM_ERROR("resume of IP block <%s> failed %d\n",
adev->ip_blocks[i].version->funcs->name, r);
return r;
}
} else {
r = adev->ip_blocks[i].version->funcs->hw_init(adev);
if (r) {
DRM_ERROR("hw_init of IP block <%s> failed %d\n",
adev->ip_blocks[i].version->funcs->name, r);
return r;
}
}
adev->ip_blocks[i].status.hw = true;
}
}
}
if (adev->powerplay.pp_funcs->load_firmware) {
r = adev->powerplay.pp_funcs->load_firmware(adev->powerplay.pp_handle);
if (r) {
pr_err("firmware loading failed\n");
return r;
}
}
return 0;
}
/**
* amdgpu_device_ip_init - run init for hardware IPs
*
@ -1581,19 +1667,21 @@ static int amdgpu_device_ip_init(struct amdgpu_device *adev)
}
}
for (i = 0; i < adev->num_ip_blocks; i++) {
if (!adev->ip_blocks[i].status.sw)
continue;
if (adev->ip_blocks[i].status.hw)
continue;
r = adev->ip_blocks[i].version->funcs->hw_init((void *)adev);
if (r) {
DRM_ERROR("hw_init of IP block <%s> failed %d\n",
adev->ip_blocks[i].version->funcs->name, r);
return r;
}
adev->ip_blocks[i].status.hw = true;
}
r = amdgpu_ucode_create_bo(adev); /* create ucode bo when sw_init complete*/
if (r)
return r;
r = amdgpu_device_ip_hw_init_phase1(adev);
if (r)
return r;
r = amdgpu_device_fw_loading(adev);
if (r)
return r;
r = amdgpu_device_ip_hw_init_phase2(adev);
if (r)
return r;
amdgpu_xgmi_add_device(adev);
amdgpu_amdkfd_device_init(adev);
@ -1656,7 +1744,7 @@ static int amdgpu_device_set_cg_state(struct amdgpu_device *adev,
for (j = 0; j < adev->num_ip_blocks; j++) {
i = state == AMD_CG_STATE_GATE ? j : adev->num_ip_blocks - j - 1;
if (!adev->ip_blocks[i].status.valid)
if (!adev->ip_blocks[i].status.late_initialized)
continue;
/* skip CG for VCE/UVD, it's handled specially */
if (adev->ip_blocks[i].version->type != AMD_IP_BLOCK_TYPE_UVD &&
@ -1686,7 +1774,7 @@ static int amdgpu_device_set_pg_state(struct amdgpu_device *adev, enum amd_power
for (j = 0; j < adev->num_ip_blocks; j++) {
i = state == AMD_PG_STATE_GATE ? j : adev->num_ip_blocks - j - 1;
if (!adev->ip_blocks[i].status.valid)
if (!adev->ip_blocks[i].status.late_initialized)
continue;
/* skip CG for VCE/UVD, it's handled specially */
if (adev->ip_blocks[i].version->type != AMD_IP_BLOCK_TYPE_UVD &&
@ -1723,7 +1811,7 @@ static int amdgpu_device_ip_late_init(struct amdgpu_device *adev)
int i = 0, r;
for (i = 0; i < adev->num_ip_blocks; i++) {
if (!adev->ip_blocks[i].status.valid)
if (!adev->ip_blocks[i].status.hw)
continue;
if (adev->ip_blocks[i].version->funcs->late_init) {
r = adev->ip_blocks[i].version->funcs->late_init((void *)adev);
@ -1732,8 +1820,8 @@ static int amdgpu_device_ip_late_init(struct amdgpu_device *adev)
adev->ip_blocks[i].version->funcs->name, r);
return r;
}
adev->ip_blocks[i].status.late_initialized = true;
}
adev->ip_blocks[i].status.late_initialized = true;
}
amdgpu_device_set_cg_state(adev, AMD_CG_STATE_GATE);
@ -1803,6 +1891,7 @@ static int amdgpu_device_ip_fini(struct amdgpu_device *adev)
continue;
if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_GMC) {
amdgpu_ucode_free_bo(adev);
amdgpu_free_static_csa(adev);
amdgpu_device_wb_fini(adev);
amdgpu_device_vram_scratch_fini(adev);
@ -1833,6 +1922,43 @@ static int amdgpu_device_ip_fini(struct amdgpu_device *adev)
return 0;
}
static int amdgpu_device_enable_mgpu_fan_boost(void)
{
struct amdgpu_gpu_instance *gpu_ins;
struct amdgpu_device *adev;
int i, ret = 0;
mutex_lock(&mgpu_info.mutex);
/*
* MGPU fan boost feature should be enabled
* only when there are two or more dGPUs in
* the system
*/
if (mgpu_info.num_dgpu < 2)
goto out;
for (i = 0; i < mgpu_info.num_dgpu; i++) {
gpu_ins = &(mgpu_info.gpu_ins[i]);
adev = gpu_ins->adev;
if (!(adev->flags & AMD_IS_APU) &&
!gpu_ins->mgpu_fan_enabled &&
adev->powerplay.pp_funcs &&
adev->powerplay.pp_funcs->enable_mgpu_fan_boost) {
ret = amdgpu_dpm_enable_mgpu_fan_boost(adev);
if (ret)
break;
gpu_ins->mgpu_fan_enabled = 1;
}
}
out:
mutex_unlock(&mgpu_info.mutex);
return ret;
}
/**
* amdgpu_device_ip_late_init_func_handler - work handler for ib test
*
@ -1847,6 +1973,10 @@ static void amdgpu_device_ip_late_init_func_handler(struct work_struct *work)
r = amdgpu_ib_ring_tests(adev);
if (r)
DRM_ERROR("ib ring test failed (%d).\n", r);
r = amdgpu_device_enable_mgpu_fan_boost();
if (r)
DRM_ERROR("enable mgpu fan boost failed (%d).\n", r);
}
static void amdgpu_device_delay_enable_gfx_off(struct work_struct *work)
@ -2082,7 +2212,8 @@ static int amdgpu_device_ip_resume_phase2(struct amdgpu_device *adev)
continue;
if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_COMMON ||
adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_GMC ||
adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_IH)
adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_IH ||
adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_PSP)
continue;
r = adev->ip_blocks[i].version->funcs->resume(adev);
if (r) {
@ -2114,6 +2245,11 @@ static int amdgpu_device_ip_resume(struct amdgpu_device *adev)
r = amdgpu_device_ip_resume_phase1(adev);
if (r)
return r;
r = amdgpu_device_fw_loading(adev);
if (r)
return r;
r = amdgpu_device_ip_resume_phase2(adev);
return r;
@ -2608,6 +2744,7 @@ int amdgpu_device_suspend(struct drm_device *dev, bool suspend, bool fbcon)
if (dev->switch_power_state == DRM_SWITCH_POWER_OFF)
return 0;
adev->in_suspend = true;
drm_kms_helper_poll_disable(dev);
if (fbcon)
@ -2793,6 +2930,8 @@ int amdgpu_device_resume(struct drm_device *dev, bool resume, bool fbcon)
#ifdef CONFIG_PM
dev->dev->power.disable_depth--;
#endif
adev->in_suspend = false;
return 0;
}
@ -3061,6 +3200,10 @@ static int amdgpu_device_reset(struct amdgpu_device *adev)
if (r)
goto out;
r = amdgpu_device_fw_loading(adev);
if (r)
return r;
r = amdgpu_device_ip_resume_phase2(adev);
if (r)
goto out;
@ -3117,6 +3260,10 @@ static int amdgpu_device_reset_sriov(struct amdgpu_device *adev,
/* we need recover gart prior to run SMC/CP/SDMA resume */
amdgpu_gtt_mgr_recover(&adev->mman.bdev.man[TTM_PL_TT]);
r = amdgpu_device_fw_loading(adev);
if (r)
return r;
/* now we are okay to resume SMC/CP/SDMA */
r = amdgpu_device_ip_reinit_late_sriov(adev);
if (r)

7
drivers/gpu/drm/amd/amdgpu/amdgpu_dpm.h
View File

@ -278,6 +278,9 @@ enum amdgpu_pcie_gen {
#define amdgpu_dpm_get_fan_speed_rpm(adev, s) \
((adev)->powerplay.pp_funcs->get_fan_speed_rpm)((adev)->powerplay.pp_handle, (s))
#define amdgpu_dpm_set_fan_speed_rpm(adev, s) \
((adev)->powerplay.pp_funcs->set_fan_speed_rpm)((adev)->powerplay.pp_handle, (s))
#define amdgpu_dpm_get_sclk(adev, l) \
((adev)->powerplay.pp_funcs->get_sclk((adev)->powerplay.pp_handle, (l)))
@ -357,6 +360,10 @@ enum amdgpu_pcie_gen {
((adev)->powerplay.pp_funcs->odn_edit_dpm_table(\
(adev)->powerplay.pp_handle, type, parameter, size))
#define amdgpu_dpm_enable_mgpu_fan_boost(adev) \
((adev)->powerplay.pp_funcs->enable_mgpu_fan_boost(\
(adev)->powerplay.pp_handle))
struct amdgpu_dpm {
struct amdgpu_ps *ps;
/* number of valid power states */

3
drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
View File

@ -127,6 +127,9 @@ int amdgpu_compute_multipipe = -1;
int amdgpu_gpu_recovery = -1; /* auto */
int amdgpu_emu_mode = 0;
uint amdgpu_smu_memory_pool_size = 0;
struct amdgpu_mgpu_info mgpu_info = {
.mutex = __MUTEX_INITIALIZER(mgpu_info.mutex),
};
/**
* DOC: vramlimit (int)

66
drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c
View File

@ -195,6 +195,19 @@ int amdgpu_fence_emit_polling(struct amdgpu_ring *ring, uint32_t *s)
return 0;
}
/**
* amdgpu_fence_schedule_fallback - schedule fallback check
*
* @ring: pointer to struct amdgpu_ring
*
* Start a timer as fallback to our interrupts.
*/
static void amdgpu_fence_schedule_fallback(struct amdgpu_ring *ring)
{
mod_timer(&ring->fence_drv.fallback_timer,
jiffies + AMDGPU_FENCE_JIFFIES_TIMEOUT);
}
/**
* amdgpu_fence_process - check for fence activity
*
@ -203,8 +216,10 @@ int amdgpu_fence_emit_polling(struct amdgpu_ring *ring, uint32_t *s)
* Checks the current fence value and calculates the last
* signalled fence value. Wakes the fence queue if the
* sequence number has increased.
*
* Returns true if fence was processed
*/
void amdgpu_fence_process(struct amdgpu_ring *ring)
bool amdgpu_fence_process(struct amdgpu_ring *ring)
{
struct amdgpu_fence_driver *drv = &ring->fence_drv;
uint32_t seq, last_seq;
@ -216,8 +231,12 @@ void amdgpu_fence_process(struct amdgpu_ring *ring)
} while (atomic_cmpxchg(&drv->last_seq, last_seq, seq) != last_seq);
if (del_timer(&ring->fence_drv.fallback_timer) &&
seq != ring->fence_drv.sync_seq)
amdgpu_fence_schedule_fallback(ring);
if (unlikely(seq == last_seq))
return;
return false;
last_seq &= drv->num_fences_mask;
seq &= drv->num_fences_mask;
@ -244,6 +263,24 @@ void amdgpu_fence_process(struct amdgpu_ring *ring)
dma_fence_put(fence);
} while (last_seq != seq);
return true;
}
/**
* amdgpu_fence_fallback - fallback for hardware interrupts
*
* @work: delayed work item
*
* Checks for fence activity.
*/
static void amdgpu_fence_fallback(struct timer_list *t)
{
struct amdgpu_ring *ring = from_timer(ring, t,
fence_drv.fallback_timer);
if (amdgpu_fence_process(ring))
DRM_WARN("Fence fallback timer expired on ring %s\n", ring->name);
}
/**
@ -393,6 +430,8 @@ int amdgpu_fence_driver_init_ring(struct amdgpu_ring *ring,
atomic_set(&ring->fence_drv.last_seq, 0);
ring->fence_drv.initialized = false;
timer_setup(&ring->fence_drv.fallback_timer, amdgpu_fence_fallback, 0);
ring->fence_drv.num_fences_mask = num_hw_submission * 2 - 1;
spin_lock_init(&ring->fence_drv.lock);
ring->fence_drv.fences = kcalloc(num_hw_submission * 2, sizeof(void *),
@ -468,6 +507,7 @@ void amdgpu_fence_driver_fini(struct amdgpu_device *adev)
amdgpu_irq_put(adev, ring->fence_drv.irq_src,
ring->fence_drv.irq_type);
drm_sched_fini(&ring->sched);
del_timer_sync(&ring->fence_drv.fallback_timer);
for (j = 0; j <= ring->fence_drv.num_fences_mask; ++j)
dma_fence_put(ring->fence_drv.fences[j]);
kfree(ring->fence_drv.fences);
@ -560,6 +600,27 @@ static const char *amdgpu_fence_get_timeline_name(struct dma_fence *f)
return (const char *)fence->ring->name;
}
/**
* amdgpu_fence_enable_signaling - enable signalling on fence
* @fence: fence
*
* This function is called with fence_queue lock held, and adds a callback
* to fence_queue that checks if this fence is signaled, and if so it
* signals the fence and removes itself.
*/
static bool amdgpu_fence_enable_signaling(struct dma_fence *f)
{
struct amdgpu_fence *fence = to_amdgpu_fence(f);
struct amdgpu_ring *ring = fence->ring;
if (!timer_pending(&ring->fence_drv.fallback_timer))
amdgpu_fence_schedule_fallback(ring);
DMA_FENCE_TRACE(&fence->base, "armed on ring %i!\n", ring->idx);
return true;
}
/**
* amdgpu_fence_free - free up the fence memory
*
@ -590,6 +651,7 @@ static void amdgpu_fence_release(struct dma_fence *f)
static const struct dma_fence_ops amdgpu_fence_ops = {
.get_driver_name = amdgpu_fence_get_driver_name,
.get_timeline_name = amdgpu_fence_get_timeline_name,
.enable_signaling = amdgpu_fence_enable_signaling,
.release = amdgpu_fence_release,
};

3
drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h
View File

@ -297,8 +297,7 @@ struct amdgpu_gfx {
/* reset mask */
uint32_t grbm_soft_reset;
uint32_t srbm_soft_reset;
/* s3/s4 mask */
bool in_suspend;
/* NGG */
struct amdgpu_ngg ngg;

13
drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
View File

@ -146,6 +146,8 @@ void amdgpu_gmc_gart_location(struct amdgpu_device *adev, struct amdgpu_gmc *mc)
{
const uint64_t four_gb = 0x100000000ULL;
u64 size_af, size_bf;
/*To avoid the hole, limit the max mc address to AMDGPU_GMC_HOLE_START*/
u64 max_mc_address = min(adev->gmc.mc_mask, AMDGPU_GMC_HOLE_START - 1);
mc->gart_size += adev->pm.smu_prv_buffer_size;
@ -153,7 +155,7 @@ void amdgpu_gmc_gart_location(struct amdgpu_device *adev, struct amdgpu_gmc *mc)
* the GART base on a 4GB boundary as well.
*/
size_bf = mc->fb_start;
size_af = adev->gmc.mc_mask + 1 - ALIGN(mc->fb_end + 1, four_gb);
size_af = max_mc_address + 1 - ALIGN(mc->fb_end + 1, four_gb);
if (mc->gart_size > max(size_bf, size_af)) {
dev_warn(adev->dev, "limiting GART\n");
@ -164,7 +166,7 @@ void amdgpu_gmc_gart_location(struct amdgpu_device *adev, struct amdgpu_gmc *mc)
(size_af < mc->gart_size))
mc->gart_start = 0;
else
mc->gart_start = mc->mc_mask - mc->gart_size + 1;
mc->gart_start = max_mc_address - mc->gart_size + 1;
mc->gart_start &= ~(four_gb - 1);
mc->gart_end = mc->gart_start + mc->gart_size - 1;
@ -200,16 +202,13 @@ void amdgpu_gmc_agp_location(struct amdgpu_device *adev, struct amdgpu_gmc *mc)
}
if (size_bf > size_af) {
mc->agp_start = mc->fb_start > mc->gart_start ?
mc->gart_end + 1 : 0;
mc->agp_start = (mc->fb_start - size_bf) & sixteen_gb_mask;
mc->agp_size = size_bf;
} else {
mc->agp_start = (mc->fb_start > mc->gart_start ?
mc->fb_end : mc->gart_end) + 1,
mc->agp_start = ALIGN(mc->fb_end + 1, sixteen_gb);
mc->agp_size = size_af;
}
mc->agp_start = ALIGN(mc->agp_start, sixteen_gb);
mc->agp_end = mc->agp_start + mc->agp_size - 1;
dev_info(adev->dev, "AGP: %lluM 0x%016llX - 0x%016llX\n",
mc->agp_size >> 20, mc->agp_start, mc->agp_end);

8
drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
View File

@ -354,6 +354,14 @@ int amdgpu_ib_ring_tests(struct amdgpu_device *adev)
if (!ring || !ring->ready)
continue;
/* skip IB tests for KIQ in general for the below reasons:
* 1. We never submit IBs to the KIQ
* 2. KIQ doesn't use the EOP interrupts,
* we use some other CP interrupt.
*/
if (ring->funcs->type == AMDGPU_RING_TYPE_KIQ)
continue;
/* MM engine need more time */
if (ring->funcs->type == AMDGPU_RING_TYPE_UVD ||
ring->funcs->type == AMDGPU_RING_TYPE_VCE ||

2
drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c
View File

@ -574,7 +574,7 @@ void amdgpu_vmid_mgr_init(struct amdgpu_device *adev)
/* skip over VMID 0, since it is the system VM */
for (j = 1; j < id_mgr->num_ids; ++j) {
amdgpu_vmid_reset(adev, i, j);
amdgpu_sync_create(&id_mgr->ids[i].active);
amdgpu_sync_create(&id_mgr->ids[j].active);
list_add_tail(&id_mgr->ids[j].list, &id_mgr->ids_lru);
}
}

164
drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c
View File

@ -24,46 +24,21 @@
#include <drm/drmP.h>
#include "amdgpu.h"
#include "amdgpu_ih.h"
#include "amdgpu_amdkfd.h"
/**
* amdgpu_ih_ring_alloc - allocate memory for the IH ring
*
* @adev: amdgpu_device pointer
*
* Allocate a ring buffer for the interrupt controller.
* Returns 0 for success, errors for failure.
*/
static int amdgpu_ih_ring_alloc(struct amdgpu_device *adev)
{
int r;
/* Allocate ring buffer */
if (adev->irq.ih.ring_obj == NULL) {
r = amdgpu_bo_create_kernel(adev, adev->irq.ih.ring_size,
PAGE_SIZE, AMDGPU_GEM_DOMAIN_GTT,
&adev->irq.ih.ring_obj,
&adev->irq.ih.gpu_addr,
(void **)&adev->irq.ih.ring);
if (r) {
DRM_ERROR("amdgpu: failed to create ih ring buffer (%d).\n", r);
return r;
}
}
return 0;
}
/**
* amdgpu_ih_ring_init - initialize the IH state
*
* @adev: amdgpu_device pointer
* @ih: ih ring to initialize
* @ring_size: ring size to allocate
* @use_bus_addr: true when we can use dma_alloc_coherent
*
* Initializes the IH state and allocates a buffer
* for the IH ring buffer.
* Returns 0 for success, errors for failure.
*/
int amdgpu_ih_ring_init(struct amdgpu_device *adev, unsigned ring_size,
bool use_bus_addr)
int amdgpu_ih_ring_init(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih,
unsigned ring_size, bool use_bus_addr)
{
u32 rb_bufsz;
int r;
@ -71,70 +46,76 @@ int amdgpu_ih_ring_init(struct amdgpu_device *adev, unsigned ring_size,
/* Align ring size */
rb_bufsz = order_base_2(ring_size / 4);
ring_size = (1 << rb_bufsz) * 4;
adev->irq.ih.ring_size = ring_size;
adev->irq.ih.ptr_mask = adev->irq.ih.ring_size - 1;
adev->irq.ih.rptr = 0;
adev->irq.ih.use_bus_addr = use_bus_addr;
ih->ring_size = ring_size;
ih->ptr_mask = ih->ring_size - 1;
ih->rptr = 0;
ih->use_bus_addr = use_bus_addr;
if (adev->irq.ih.use_bus_addr) {
if (!adev->irq.ih.ring) {
/* add 8 bytes for the rptr/wptr shadows and
* add them to the end of the ring allocation.
*/
adev->irq.ih.ring = pci_alloc_consistent(adev->pdev,
adev->irq.ih.ring_size + 8,
&adev->irq.ih.rb_dma_addr);
if (adev->irq.ih.ring == NULL)
return -ENOMEM;
memset((void *)adev->irq.ih.ring, 0, adev->irq.ih.ring_size + 8);
adev->irq.ih.wptr_offs = (adev->irq.ih.ring_size / 4) + 0;
adev->irq.ih.rptr_offs = (adev->irq.ih.ring_size / 4) + 1;
}
return 0;
if (use_bus_addr) {
if (ih->ring)
return 0;
/* add 8 bytes for the rptr/wptr shadows and
* add them to the end of the ring allocation.
*/
ih->ring = dma_alloc_coherent(adev->dev, ih->ring_size + 8,
&ih->rb_dma_addr, GFP_KERNEL);
if (ih->ring == NULL)
return -ENOMEM;
memset((void *)ih->ring, 0, ih->ring_size + 8);
ih->wptr_offs = (ih->ring_size / 4) + 0;
ih->rptr_offs = (ih->ring_size / 4) + 1;
} else {
r = amdgpu_device_wb_get(adev, &adev->irq.ih.wptr_offs);
r = amdgpu_device_wb_get(adev, &ih->wptr_offs);
if (r)
return r;
r = amdgpu_device_wb_get(adev, &ih->rptr_offs);
if (r) {
dev_err(adev->dev, "(%d) ih wptr_offs wb alloc failed\n", r);
amdgpu_device_wb_free(adev, ih->wptr_offs);
return r;
}
r = amdgpu_device_wb_get(adev, &adev->irq.ih.rptr_offs);
r = amdgpu_bo_create_kernel(adev, ih->ring_size, PAGE_SIZE,
AMDGPU_GEM_DOMAIN_GTT,
&ih->ring_obj, &ih->gpu_addr,
(void **)&ih->ring);
if (r) {
amdgpu_device_wb_free(adev, adev->irq.ih.wptr_offs);
dev_err(adev->dev, "(%d) ih rptr_offs wb alloc failed\n", r);
amdgpu_device_wb_free(adev, ih->rptr_offs);
amdgpu_device_wb_free(adev, ih->wptr_offs);
return r;
}
return amdgpu_ih_ring_alloc(adev);
}
return 0;
}
/**
* amdgpu_ih_ring_fini - tear down the IH state
*
* @adev: amdgpu_device pointer
* @ih: ih ring to tear down
*
* Tears down the IH state and frees buffer
* used for the IH ring buffer.
*/
void amdgpu_ih_ring_fini(struct amdgpu_device *adev)
void amdgpu_ih_ring_fini(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih)
{
if (adev->irq.ih.use_bus_addr) {
if (adev->irq.ih.ring) {
/* add 8 bytes for the rptr/wptr shadows and
* add them to the end of the ring allocation.
*/
pci_free_consistent(adev->pdev, adev->irq.ih.ring_size + 8,
(void *)adev->irq.ih.ring,
adev->irq.ih.rb_dma_addr);
adev->irq.ih.ring = NULL;
}
if (ih->use_bus_addr) {
if (!ih->ring)
return;
/* add 8 bytes for the rptr/wptr shadows and
* add them to the end of the ring allocation.
*/
dma_free_coherent(adev->dev, ih->ring_size + 8,
(void *)ih->ring, ih->rb_dma_addr);
ih->ring = NULL;
} else {
amdgpu_bo_free_kernel(&adev->irq.ih.ring_obj,
&adev->irq.ih.gpu_addr,
(void **)&adev->irq.ih.ring);
amdgpu_device_wb_free(adev, adev->irq.ih.wptr_offs);
amdgpu_device_wb_free(adev, adev->irq.ih.rptr_offs);
amdgpu_bo_free_kernel(&ih->ring_obj, &ih->gpu_addr,
(void **)&ih->ring);
amdgpu_device_wb_free(adev, ih->wptr_offs);
amdgpu_device_wb_free(adev, ih->rptr_offs);
}
}
@ -142,56 +123,43 @@ void amdgpu_ih_ring_fini(struct amdgpu_device *adev)
* amdgpu_ih_process - interrupt handler
*
* @adev: amdgpu_device pointer
* @ih: ih ring to process
*
* Interrupt hander (VI), walk the IH ring.
* Returns irq process return code.
*/
int amdgpu_ih_process(struct amdgpu_device *adev)
int amdgpu_ih_process(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih,
void (*callback)(struct amdgpu_device *adev,
struct amdgpu_ih_ring *ih))
{
struct amdgpu_iv_entry entry;
u32 wptr;
if (!adev->irq.ih.enabled || adev->shutdown)
if (!ih->enabled || adev->shutdown)
return IRQ_NONE;
wptr = amdgpu_ih_get_wptr(adev);
restart_ih:
/* is somebody else already processing irqs? */
if (atomic_xchg(&adev->irq.ih.lock, 1))
if (atomic_xchg(&ih->lock, 1))
return IRQ_NONE;
DRM_DEBUG("%s: rptr %d, wptr %d\n", __func__, adev->irq.ih.rptr, wptr);
DRM_DEBUG("%s: rptr %d, wptr %d\n", __func__, ih->rptr, wptr);
/* Order reading of wptr vs. reading of IH ring data */
rmb();
while (adev->irq.ih.rptr != wptr) {
u32 ring_index = adev->irq.ih.rptr >> 2;
/* Prescreening of high-frequency interrupts */
if (!amdgpu_ih_prescreen_iv(adev)) {
adev->irq.ih.rptr &= adev->irq.ih.ptr_mask;
continue;
}
/* Before dispatching irq to IP blocks, send it to amdkfd */
amdgpu_amdkfd_interrupt(adev,
(const void *) &adev->irq.ih.ring[ring_index]);
entry.iv_entry = (const uint32_t *)
&adev->irq.ih.ring[ring_index];
amdgpu_ih_decode_iv(adev, &entry);
adev->irq.ih.rptr &= adev->irq.ih.ptr_mask;
amdgpu_irq_dispatch(adev, &entry);
while (ih->rptr != wptr) {
callback(adev, ih);
ih->rptr &= ih->ptr_mask;
}
amdgpu_ih_set_rptr(adev);
atomic_set(&adev->irq.ih.lock, 0);
atomic_set(&ih->lock, 0);
/* make sure wptr hasn't changed while processing */
wptr = amdgpu_ih_get_wptr(adev);
if (wptr != adev->irq.ih.rptr)
if (wptr != ih->rptr)
goto restart_ih;
return IRQ_HANDLED;

32
drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h
View File

@ -24,12 +24,8 @@
#ifndef __AMDGPU_IH_H__
#define __AMDGPU_IH_H__
#include "soc15_ih_clientid.h"
struct amdgpu_device;
#define AMDGPU_IH_CLIENTID_LEGACY 0
#define AMDGPU_IH_CLIENTID_MAX SOC15_IH_CLIENTID_MAX
struct amdgpu_iv_entry;
/*
* R6xx+ IH ring
@ -51,22 +47,6 @@ struct amdgpu_ih_ring {
dma_addr_t rb_dma_addr; /* only used when use_bus_addr = true */
};
#define AMDGPU_IH_SRC_DATA_MAX_SIZE_DW 4
struct amdgpu_iv_entry {
unsigned client_id;
unsigned src_id;
unsigned ring_id;
unsigned vmid;
unsigned vmid_src;
uint64_t timestamp;
unsigned timestamp_src;
unsigned pasid;
unsigned pasid_src;
unsigned src_data[AMDGPU_IH_SRC_DATA_MAX_SIZE_DW];
const uint32_t *iv_entry;
};
/* provided by the ih block */
struct amdgpu_ih_funcs {
/* ring read/write ptr handling, called from interrupt context */
@ -82,9 +62,11 @@ struct amdgpu_ih_funcs {
#define amdgpu_ih_decode_iv(adev, iv) (adev)->irq.ih_funcs->decode_iv((adev), (iv))
#define amdgpu_ih_set_rptr(adev) (adev)->irq.ih_funcs->set_rptr((adev))
int amdgpu_ih_ring_init(struct amdgpu_device *adev, unsigned ring_size,
bool use_bus_addr);
void amdgpu_ih_ring_fini(struct amdgpu_device *adev);
int amdgpu_ih_process(struct amdgpu_device *adev);
int amdgpu_ih_ring_init(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih,
unsigned ring_size, bool use_bus_addr);
void amdgpu_ih_ring_fini(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih);
int amdgpu_ih_process(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih,
void (*callback)(struct amdgpu_device *adev,
struct amdgpu_ih_ring *ih));
#endif

41
drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c
View File

@ -51,6 +51,7 @@
#include "atom.h"
#include "amdgpu_connectors.h"
#include "amdgpu_trace.h"
#include "amdgpu_amdkfd.h"
#include <linux/pm_runtime.h>
@ -123,7 +124,7 @@ void amdgpu_irq_disable_all(struct amdgpu_device *adev)
int r;
spin_lock_irqsave(&adev->irq.lock, irqflags);
for (i = 0; i < AMDGPU_IH_CLIENTID_MAX; ++i) {
for (i = 0; i < AMDGPU_IRQ_CLIENTID_MAX; ++i) {
if (!adev->irq.client[i].sources)
continue;
@ -146,6 +147,34 @@ void amdgpu_irq_disable_all(struct amdgpu_device *adev)
spin_unlock_irqrestore(&adev->irq.lock, irqflags);
}
/**
* amdgpu_irq_callback - callback from the IH ring
*
* @adev: amdgpu device pointer
* @ih: amdgpu ih ring
*
* Callback from IH ring processing to handle the entry at the current position
* and advance the read pointer.
*/
static void amdgpu_irq_callback(struct amdgpu_device *adev,
struct amdgpu_ih_ring *ih)
{
u32 ring_index = ih->rptr >> 2;
struct amdgpu_iv_entry entry;
/* Prescreening of high-frequency interrupts */
if (!amdgpu_ih_prescreen_iv(adev))
return;
/* Before dispatching irq to IP blocks, send it to amdkfd */
amdgpu_amdkfd_interrupt(adev, (const void *) &ih->ring[ring_index]);
entry.iv_entry = (const uint32_t *)&ih->ring[ring_index];
amdgpu_ih_decode_iv(adev, &entry);
amdgpu_irq_dispatch(adev, &entry);
}
/**
* amdgpu_irq_handler - IRQ handler
*
@ -163,7 +192,7 @@ irqreturn_t amdgpu_irq_handler(int irq, void *arg)
struct amdgpu_device *adev = dev->dev_private;
irqreturn_t ret;
ret = amdgpu_ih_process(adev);
ret = amdgpu_ih_process(adev, &adev->irq.ih, amdgpu_irq_callback);
if (ret == IRQ_HANDLED)
pm_runtime_mark_last_busy(dev->dev);
return ret;
@ -273,7 +302,7 @@ void amdgpu_irq_fini(struct amdgpu_device *adev)
cancel_work_sync(&adev->reset_work);
}
for (i = 0; i < AMDGPU_IH_CLIENTID_MAX; ++i) {
for (i = 0; i < AMDGPU_IRQ_CLIENTID_MAX; ++i) {
if (!adev->irq.client[i].sources)
continue;
@ -313,7 +342,7 @@ int amdgpu_irq_add_id(struct amdgpu_device *adev,
unsigned client_id, unsigned src_id,
struct amdgpu_irq_src *source)
{
if (client_id >= AMDGPU_IH_CLIENTID_MAX)
if (client_id >= AMDGPU_IRQ_CLIENTID_MAX)
return -EINVAL;
if (src_id >= AMDGPU_MAX_IRQ_SRC_ID)
@ -367,7 +396,7 @@ void amdgpu_irq_dispatch(struct amdgpu_device *adev,
trace_amdgpu_iv(entry);
if (client_id >= AMDGPU_IH_CLIENTID_MAX) {
if (client_id >= AMDGPU_IRQ_CLIENTID_MAX) {
DRM_DEBUG("Invalid client_id in IV: %d\n", client_id);
return;
}
@ -440,7 +469,7 @@ void amdgpu_irq_gpu_reset_resume_helper(struct amdgpu_device *adev)
{
int i, j, k;
for (i = 0; i < AMDGPU_IH_CLIENTID_MAX; ++i) {
for (i = 0; i < AMDGPU_IRQ_CLIENTID_MAX; ++i) {
if (!adev->irq.client[i].sources)
continue;

25
drivers/gpu/drm/amd/amdgpu/amdgpu_irq.h
View File

@ -25,19 +25,38 @@
#define __AMDGPU_IRQ_H__
#include <linux/irqdomain.h>
#include "soc15_ih_clientid.h"
#include "amdgpu_ih.h"
#define AMDGPU_MAX_IRQ_SRC_ID 0x100
#define AMDGPU_MAX_IRQ_SRC_ID 0x100
#define AMDGPU_MAX_IRQ_CLIENT_ID 0x100
#define AMDGPU_IRQ_CLIENTID_LEGACY 0
#define AMDGPU_IRQ_CLIENTID_MAX SOC15_IH_CLIENTID_MAX
#define AMDGPU_IRQ_SRC_DATA_MAX_SIZE_DW 4
struct amdgpu_device;
struct amdgpu_iv_entry;
enum amdgpu_interrupt_state {
AMDGPU_IRQ_STATE_DISABLE,
AMDGPU_IRQ_STATE_ENABLE,
};
struct amdgpu_iv_entry {
unsigned client_id;
unsigned src_id;
unsigned ring_id;
unsigned vmid;
unsigned vmid_src;
uint64_t timestamp;
unsigned timestamp_src;
unsigned pasid;
unsigned pasid_src;
unsigned src_data[AMDGPU_IRQ_SRC_DATA_MAX_SIZE_DW];
const uint32_t *iv_entry;
};
struct amdgpu_irq_src {
unsigned num_types;
atomic_t *enabled_types;
@ -63,7 +82,7 @@ struct amdgpu_irq {
bool installed;
spinlock_t lock;
/* interrupt sources */
struct amdgpu_irq_client client[AMDGPU_IH_CLIENTID_MAX];
struct amdgpu_irq_client client[AMDGPU_IRQ_CLIENTID_MAX];
/* status, etc. */
bool msi_enabled; /* msi enabled */

52
drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
View File

@ -40,6 +40,30 @@
#include "amdgpu_gem.h"
#include "amdgpu_display.h"
static void amdgpu_unregister_gpu_instance(struct amdgpu_device *adev)
{
struct amdgpu_gpu_instance *gpu_instance;
int i;
mutex_lock(&mgpu_info.mutex);
for (i = 0; i < mgpu_info.num_gpu; i++) {
gpu_instance = &(mgpu_info.gpu_ins[i]);
if (gpu_instance->adev == adev) {
mgpu_info.gpu_ins[i] =
mgpu_info.gpu_ins[mgpu_info.num_gpu - 1];
mgpu_info.num_gpu--;
if (adev->flags & AMD_IS_APU)
mgpu_info.num_apu--;
else
mgpu_info.num_dgpu--;
break;
}
}
mutex_unlock(&mgpu_info.mutex);
}
/**
* amdgpu_driver_unload_kms - Main unload function for KMS.
*
@ -55,6 +79,8 @@ void amdgpu_driver_unload_kms(struct drm_device *dev)
if (adev == NULL)
return;
amdgpu_unregister_gpu_instance(adev);
if (adev->rmmio == NULL)
goto done_free;
@ -75,6 +101,31 @@ void amdgpu_driver_unload_kms(struct drm_device *dev)
dev->dev_private = NULL;
}
static void amdgpu_register_gpu_instance(struct amdgpu_device *adev)
{
struct amdgpu_gpu_instance *gpu_instance;
mutex_lock(&mgpu_info.mutex);
if (mgpu_info.num_gpu >= MAX_GPU_INSTANCE) {
DRM_ERROR("Cannot register more gpu instance\n");
mutex_unlock(&mgpu_info.mutex);
return;
}
gpu_instance = &(mgpu_info.gpu_ins[mgpu_info.num_gpu]);
gpu_instance->adev = adev;
gpu_instance->mgpu_fan_enabled = 0;
mgpu_info.num_gpu++;
if (adev->flags & AMD_IS_APU)
mgpu_info.num_apu++;
else
mgpu_info.num_dgpu++;
mutex_unlock(&mgpu_info.mutex);
}
/**
* amdgpu_driver_load_kms - Main load function for KMS.
*
@ -169,6 +220,7 @@ int amdgpu_driver_load_kms(struct drm_device *dev, unsigned long flags)
pm_runtime_put_autosuspend(dev->dev);
}
amdgpu_register_gpu_instance(adev);
out:
if (r) {
/* balance pm_runtime_get_sync in amdgpu_driver_unload_kms */

192
drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c
View File

@ -1120,12 +1120,19 @@ static ssize_t amdgpu_hwmon_set_pwm1(struct device *dev,
struct amdgpu_device *adev = dev_get_drvdata(dev);
int err;
u32 value;
u32 pwm_mode;
/* Can't adjust fan when the card is off */
if ((adev->flags & AMD_IS_PX) &&
(adev->ddev->switch_power_state != DRM_SWITCH_POWER_ON))
return -EINVAL;
pwm_mode = amdgpu_dpm_get_fan_control_mode(adev);
if (pwm_mode != AMD_FAN_CTRL_MANUAL) {
pr_info("manual fan speed control should be enabled first\n");
return -EINVAL;
}
err = kstrtou32(buf, 10, &value);
if (err)
return err;
@ -1187,6 +1194,148 @@ static ssize_t amdgpu_hwmon_get_fan1_input(struct device *dev,
return sprintf(buf, "%i\n", speed);
}
static ssize_t amdgpu_hwmon_get_fan1_min(struct device *dev,
struct device_attribute *attr,
char *buf)
{
struct amdgpu_device *adev = dev_get_drvdata(dev);
u32 min_rpm = 0;
u32 size = sizeof(min_rpm);
int r;
if (!adev->powerplay.pp_funcs->read_sensor)
return -EINVAL;
r = amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_MIN_FAN_RPM,
(void *)&min_rpm, &size);
if (r)
return r;
return snprintf(buf, PAGE_SIZE, "%d\n", min_rpm);
}
static ssize_t amdgpu_hwmon_get_fan1_max(struct device *dev,
struct device_attribute *attr,
char *buf)
{
struct amdgpu_device *adev = dev_get_drvdata(dev);
u32 max_rpm = 0;
u32 size = sizeof(max_rpm);
int r;
if (!adev->powerplay.pp_funcs->read_sensor)
return -EINVAL;
r = amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_MAX_FAN_RPM,
(void *)&max_rpm, &size);
if (r)
return r;
return snprintf(buf, PAGE_SIZE, "%d\n", max_rpm);
}
static ssize_t amdgpu_hwmon_get_fan1_target(struct device *dev,
struct device_attribute *attr,
char *buf)
{
struct amdgpu_device *adev = dev_get_drvdata(dev);
int err;
u32 rpm = 0;
/* Can't adjust fan when the card is off */
if ((adev->flags & AMD_IS_PX) &&
(adev->ddev->switch_power_state != DRM_SWITCH_POWER_ON))
return -EINVAL;
if (adev->powerplay.pp_funcs->get_fan_speed_rpm) {
err = amdgpu_dpm_get_fan_speed_rpm(adev, &rpm);
if (err)
return err;
}
return sprintf(buf, "%i\n", rpm);
}
static ssize_t amdgpu_hwmon_set_fan1_target(struct device *dev,
struct device_attribute *attr,
const char *buf, size_t count)
{
struct amdgpu_device *adev = dev_get_drvdata(dev);
int err;
u32 value;
u32 pwm_mode;
pwm_mode = amdgpu_dpm_get_fan_control_mode(adev);
if (pwm_mode != AMD_FAN_CTRL_MANUAL)
return -ENODATA;
/* Can't adjust fan when the card is off */
if ((adev->flags & AMD_IS_PX) &&
(adev->ddev->switch_power_state != DRM_SWITCH_POWER_ON))
return -EINVAL;
err = kstrtou32(buf, 10, &value);
if (err)
return err;
if (adev->powerplay.pp_funcs->set_fan_speed_rpm) {
err = amdgpu_dpm_set_fan_speed_rpm(adev, value);
if (err)
return err;
}
return count;
}
static ssize_t amdgpu_hwmon_get_fan1_enable(struct device *dev,
struct device_attribute *attr,
char *buf)
{
struct amdgpu_device *adev = dev_get_drvdata(dev);
u32 pwm_mode = 0;
if (!adev->powerplay.pp_funcs->get_fan_control_mode)
return -EINVAL;
pwm_mode = amdgpu_dpm_get_fan_control_mode(adev);
return sprintf(buf, "%i\n", pwm_mode == AMD_FAN_CTRL_AUTO ? 0 : 1);
}
static ssize_t amdgpu_hwmon_set_fan1_enable(struct device *dev,
struct device_attribute *attr,
const char *buf,
size_t count)
{
struct amdgpu_device *adev = dev_get_drvdata(dev);
int err;
int value;
u32 pwm_mode;
/* Can't adjust fan when the card is off */
if ((adev->flags & AMD_IS_PX) &&
(adev->ddev->switch_power_state != DRM_SWITCH_POWER_ON))
return -EINVAL;
if (!adev->powerplay.pp_funcs->set_fan_control_mode)
return -EINVAL;
err = kstrtoint(buf, 10, &value);
if (err)
return err;
if (value == 0)
pwm_mode = AMD_FAN_CTRL_AUTO;
else if (value == 1)
pwm_mode = AMD_FAN_CTRL_MANUAL;
else
return -EINVAL;
amdgpu_dpm_set_fan_control_mode(adev, pwm_mode);
return count;
}
static ssize_t amdgpu_hwmon_show_vddgfx(struct device *dev,
struct device_attribute *attr,
char *buf)
@ -1406,8 +1555,16 @@ static ssize_t amdgpu_hwmon_set_power_cap(struct device *dev,
*
* - pwm1_max: pulse width modulation fan control maximum level (255)
*
* - fan1_min: an minimum value Unit: revolution/min (RPM)
*
* - fan1_max: an maxmum value Unit: revolution/max (RPM)
*
* - fan1_input: fan speed in RPM
*
* - fan[1-*]_target: Desired fan speed Unit: revolution/min (RPM)
*
* - fan[1-*]_enable: Enable or disable the sensors.1: Enable 0: Disable
*
* You can use hwmon tools like sensors to view this information on your system.
*
*/
@ -1420,6 +1577,10 @@ static SENSOR_DEVICE_ATTR(pwm1_enable, S_IRUGO | S_IWUSR, amdgpu_hwmon_get_pwm1_
static SENSOR_DEVICE_ATTR(pwm1_min, S_IRUGO, amdgpu_hwmon_get_pwm1_min, NULL, 0);
static SENSOR_DEVICE_ATTR(pwm1_max, S_IRUGO, amdgpu_hwmon_get_pwm1_max, NULL, 0);
static SENSOR_DEVICE_ATTR(fan1_input, S_IRUGO, amdgpu_hwmon_get_fan1_input, NULL, 0);
static SENSOR_DEVICE_ATTR(fan1_min, S_IRUGO, amdgpu_hwmon_get_fan1_min, NULL, 0);
static SENSOR_DEVICE_ATTR(fan1_max, S_IRUGO, amdgpu_hwmon_get_fan1_max, NULL, 0);
static SENSOR_DEVICE_ATTR(fan1_target, S_IRUGO | S_IWUSR, amdgpu_hwmon_get_fan1_target, amdgpu_hwmon_set_fan1_target, 0);
static SENSOR_DEVICE_ATTR(fan1_enable, S_IRUGO | S_IWUSR, amdgpu_hwmon_get_fan1_enable, amdgpu_hwmon_set_fan1_enable, 0);
static SENSOR_DEVICE_ATTR(in0_input, S_IRUGO, amdgpu_hwmon_show_vddgfx, NULL, 0);
static SENSOR_DEVICE_ATTR(in0_label, S_IRUGO, amdgpu_hwmon_show_vddgfx_label, NULL, 0);
static SENSOR_DEVICE_ATTR(in1_input, S_IRUGO, amdgpu_hwmon_show_vddnb, NULL, 0);
@ -1438,6 +1599,10 @@ static struct attribute *hwmon_attributes[] = {
&sensor_dev_attr_pwm1_min.dev_attr.attr,
&sensor_dev_attr_pwm1_max.dev_attr.attr,
&sensor_dev_attr_fan1_input.dev_attr.attr,
&sensor_dev_attr_fan1_min.dev_attr.attr,
&sensor_dev_attr_fan1_max.dev_attr.attr,
&sensor_dev_attr_fan1_target.dev_attr.attr,
&sensor_dev_attr_fan1_enable.dev_attr.attr,
&sensor_dev_attr_in0_input.dev_attr.attr,
&sensor_dev_attr_in0_label.dev_attr.attr,
&sensor_dev_attr_in1_input.dev_attr.attr,
@ -1456,13 +1621,16 @@ static umode_t hwmon_attributes_visible(struct kobject *kobj,
struct amdgpu_device *adev = dev_get_drvdata(dev);
umode_t effective_mode = attr->mode;
/* Skip fan attributes if fan is not present */
if (adev->pm.no_fan && (attr == &sensor_dev_attr_pwm1.dev_attr.attr ||
attr == &sensor_dev_attr_pwm1_enable.dev_attr.attr ||
attr == &sensor_dev_attr_pwm1_max.dev_attr.attr ||
attr == &sensor_dev_attr_pwm1_min.dev_attr.attr ||
attr == &sensor_dev_attr_fan1_input.dev_attr.attr))
attr == &sensor_dev_attr_fan1_input.dev_attr.attr ||
attr == &sensor_dev_attr_fan1_min.dev_attr.attr ||
attr == &sensor_dev_attr_fan1_max.dev_attr.attr ||
attr == &sensor_dev_attr_fan1_target.dev_attr.attr ||
attr == &sensor_dev_attr_fan1_enable.dev_attr.attr))
return 0;
/* Skip limit attributes if DPM is not enabled */
@ -1472,7 +1640,12 @@ static umode_t hwmon_attributes_visible(struct kobject *kobj,
attr == &sensor_dev_attr_pwm1.dev_attr.attr ||
attr == &sensor_dev_attr_pwm1_enable.dev_attr.attr ||
attr == &sensor_dev_attr_pwm1_max.dev_attr.attr ||
attr == &sensor_dev_attr_pwm1_min.dev_attr.attr))
attr == &sensor_dev_attr_pwm1_min.dev_attr.attr ||
attr == &sensor_dev_attr_fan1_input.dev_attr.attr ||
attr == &sensor_dev_attr_fan1_min.dev_attr.attr ||
attr == &sensor_dev_attr_fan1_max.dev_attr.attr ||
attr == &sensor_dev_attr_fan1_target.dev_attr.attr ||
attr == &sensor_dev_attr_fan1_enable.dev_attr.attr))
return 0;
/* mask fan attributes if we have no bindings for this asic to expose */
@ -1497,10 +1670,18 @@ static umode_t hwmon_attributes_visible(struct kobject *kobj,
/* hide max/min values if we can't both query and manage the fan */
if ((!adev->powerplay.pp_funcs->set_fan_speed_percent &&
!adev->powerplay.pp_funcs->get_fan_speed_percent) &&
(!adev->powerplay.pp_funcs->set_fan_speed_rpm &&
!adev->powerplay.pp_funcs->get_fan_speed_rpm) &&
(attr == &sensor_dev_attr_pwm1_max.dev_attr.attr ||
attr == &sensor_dev_attr_pwm1_min.dev_attr.attr))
return 0;
if ((!adev->powerplay.pp_funcs->set_fan_speed_rpm &&
!adev->powerplay.pp_funcs->get_fan_speed_rpm) &&
(attr == &sensor_dev_attr_fan1_max.dev_attr.attr ||
attr == &sensor_dev_attr_fan1_min.dev_attr.attr))
return 0;
/* only APUs have vddnb */
if (!(adev->flags & AMD_IS_APU) &&
(attr == &sensor_dev_attr_in1_input.dev_attr.attr ||
@ -1976,6 +2157,7 @@ void amdgpu_pm_compute_clocks(struct amdgpu_device *adev)
static int amdgpu_debugfs_pm_info_pp(struct seq_file *m, struct amdgpu_device *adev)
{
uint32_t value;
uint64_t value64;
uint32_t query = 0;
int size;
@ -2014,6 +2196,10 @@ static int amdgpu_debugfs_pm_info_pp(struct seq_file *m, struct amdgpu_device *a
seq_printf(m, "GPU Load: %u %%\n", value);
seq_printf(m, "\n");
/* SMC feature mask */
if (!amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_ENABLED_SMC_FEATURES_MASK, (void *)&value64, &size))
seq_printf(m, "SMC Feature Mask: 0x%016llx\n", value64);
/* UVD clocks */
if (!amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_UVD_POWER, (void *)&value, &size)) {
if (!value) {

2
drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
View File

@ -452,8 +452,6 @@ static int psp_hw_fini(void *handle)
if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP)
return 0;
amdgpu_ucode_fini_bo(adev);
psp_ring_destroy(psp, PSP_RING_TYPE__KM);
amdgpu_bo_free_kernel(&psp->tmr_bo, &psp->tmr_mc_addr, &psp->tmr_buf);

3
drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h
View File

@ -77,6 +77,7 @@ struct amdgpu_fence_driver {
bool initialized;
struct amdgpu_irq_src *irq_src;
unsigned irq_type;
struct timer_list fallback_timer;
unsigned num_fences_mask;
spinlock_t lock;
struct dma_fence **fences;
@ -96,7 +97,7 @@ void amdgpu_fence_driver_resume(struct amdgpu_device *adev);
int amdgpu_fence_emit(struct amdgpu_ring *ring, struct dma_fence **fence,
unsigned flags);
int amdgpu_fence_emit_polling(struct amdgpu_ring *ring, uint32_t *s);
void amdgpu_fence_process(struct amdgpu_ring *ring);
bool amdgpu_fence_process(struct amdgpu_ring *ring);
int amdgpu_fence_wait_empty(struct amdgpu_ring *ring);
signed long amdgpu_fence_wait_polling(struct amdgpu_ring *ring,
uint32_t wait_seq,

4
drivers/gpu/drm/amd/amdgpu/amdgpu_sdma.h
View File

@ -46,10 +46,6 @@ struct amdgpu_sdma_instance {
struct amdgpu_sdma {
struct amdgpu_sdma_instance instance[AMDGPU_MAX_SDMA_INSTANCES];
#ifdef CONFIG_DRM_AMDGPU_SI
//SI DMA has a difference trap irq number for the second engine
struct amdgpu_irq_src trap_irq_1;
#endif
struct amdgpu_irq_src trap_irq;
struct amdgpu_irq_src illegal_inst_irq;
int num_instances;

2
drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h
View File

@ -103,7 +103,7 @@ TRACE_EVENT(amdgpu_iv,
__entry->src_data[2] = iv->src_data[2];
__entry->src_data[3] = iv->src_data[3];
),
TP_printk("client_id:%u src_id:%u ring:%u vmid:%u timestamp: %llu pasid:%u src_data: %08x %08x %08x %08x\n",
TP_printk("client_id:%u src_id:%u ring:%u vmid:%u timestamp: %llu pasid:%u src_data: %08x %08x %08x %08x",
__entry->client_id, __entry->src_id,
__entry->ring_id, __entry->vmid,
__entry->timestamp, __entry->pasid,

2
drivers/gpu/drm/amd/amdgpu/amdgpu_trace_points.c
View File

@ -1,4 +1,4 @@
// SPDX-License-Identifier: GPL-2.0
// SPDX-License-Identifier: MIT
/* Copyright Red Hat Inc 2010.
*
* Permission is hereby granted, free of charge, to any person obtaining a

84
drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c
View File

@ -297,10 +297,7 @@ amdgpu_ucode_get_load_type(struct amdgpu_device *adev, int load_type)
case CHIP_POLARIS11:
case CHIP_POLARIS12:
case CHIP_VEGAM:
if (!load_type)
return AMDGPU_FW_LOAD_DIRECT;
else
return AMDGPU_FW_LOAD_SMU;
return AMDGPU_FW_LOAD_SMU;
case CHIP_VEGA10:
case CHIP_RAVEN:
case CHIP_VEGA12:
@ -423,32 +420,41 @@ static int amdgpu_ucode_patch_jt(struct amdgpu_firmware_info *ucode,
return 0;
}
int amdgpu_ucode_create_bo(struct amdgpu_device *adev)
{
if (adev->firmware.load_type != AMDGPU_FW_LOAD_DIRECT) {
amdgpu_bo_create_kernel(adev, adev->firmware.fw_size, PAGE_SIZE,
amdgpu_sriov_vf(adev) ? AMDGPU_GEM_DOMAIN_VRAM : AMDGPU_GEM_DOMAIN_GTT,
&adev->firmware.fw_buf,
&adev->firmware.fw_buf_mc,
&adev->firmware.fw_buf_ptr);
if (!adev->firmware.fw_buf) {
dev_err(adev->dev, "failed to create kernel buffer for firmware.fw_buf\n");
return -ENOMEM;
} else if (amdgpu_sriov_vf(adev)) {
memset(adev->firmware.fw_buf_ptr, 0, adev->firmware.fw_size);
}
}
return 0;
}
void amdgpu_ucode_free_bo(struct amdgpu_device *adev)
{
if (adev->firmware.load_type != AMDGPU_FW_LOAD_DIRECT)
amdgpu_bo_free_kernel(&adev->firmware.fw_buf,
&adev->firmware.fw_buf_mc,
&adev->firmware.fw_buf_ptr);
}
int amdgpu_ucode_init_bo(struct amdgpu_device *adev)
{
uint64_t fw_offset = 0;
int i, err;
int i;
struct amdgpu_firmware_info *ucode = NULL;
const struct common_firmware_header *header = NULL;
if (!adev->firmware.fw_size) {
dev_warn(adev->dev, "No ip firmware need to load\n");
/* for baremetal, the ucode is allocated in gtt, so don't need to fill the bo when reset/suspend */
if (!amdgpu_sriov_vf(adev) && (adev->in_gpu_reset || adev->in_suspend))
return 0;
}
if (!adev->in_gpu_reset) {
err = amdgpu_bo_create_kernel(adev, adev->firmware.fw_size, PAGE_SIZE,
amdgpu_sriov_vf(adev) ? AMDGPU_GEM_DOMAIN_VRAM : AMDGPU_GEM_DOMAIN_GTT,
&adev->firmware.fw_buf,
&adev->firmware.fw_buf_mc,
&adev->firmware.fw_buf_ptr);
if (err) {
dev_err(adev->dev, "failed to create kernel buffer for firmware.fw_buf\n");
goto failed;
}
}
memset(adev->firmware.fw_buf_ptr, 0, adev->firmware.fw_size);
/*
* if SMU loaded firmware, it needn't add SMC, UVD, and VCE
* ucode info here
@ -465,7 +471,6 @@ int amdgpu_ucode_init_bo(struct amdgpu_device *adev)
for (i = 0; i < adev->firmware.max_ucodes; i++) {
ucode = &adev->firmware.ucode[i];
if (ucode->fw) {
header = (const struct common_firmware_header *)ucode->fw->data;
amdgpu_ucode_init_single_fw(adev, ucode, adev->firmware.fw_buf_mc + fw_offset,
adev->firmware.fw_buf_ptr + fw_offset);
if (i == AMDGPU_UCODE_ID_CP_MEC1 &&
@ -480,33 +485,4 @@ int amdgpu_ucode_init_bo(struct amdgpu_device *adev)
}
}
return 0;
failed:
if (err)
adev->firmware.load_type = AMDGPU_FW_LOAD_DIRECT;
return err;
}
int amdgpu_ucode_fini_bo(struct amdgpu_device *adev)
{
int i;
struct amdgpu_firmware_info *ucode = NULL;
if (!adev->firmware.fw_size)
return 0;
for (i = 0; i < adev->firmware.max_ucodes; i++) {
ucode = &adev->firmware.ucode[i];
if (ucode->fw) {
ucode->mc_addr = 0;
ucode->kaddr = NULL;
}
}
amdgpu_bo_free_kernel(&adev->firmware.fw_buf,
&adev->firmware.fw_buf_mc,
&adev->firmware.fw_buf_ptr);
return 0;
}

4
drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h
View File

@ -276,8 +276,10 @@ void amdgpu_ucode_print_gpu_info_hdr(const struct common_firmware_header *hdr);
int amdgpu_ucode_validate(const struct firmware *fw);
bool amdgpu_ucode_hdr_version(union amdgpu_firmware_header *hdr,
uint16_t hdr_major, uint16_t hdr_minor);
int amdgpu_ucode_init_bo(struct amdgpu_device *adev);
int amdgpu_ucode_fini_bo(struct amdgpu_device *adev);
int amdgpu_ucode_create_bo(struct amdgpu_device *adev);
void amdgpu_ucode_free_bo(struct amdgpu_device *adev);
enum amdgpu_firmware_load_type
amdgpu_ucode_get_load_type(struct amdgpu_device *adev, int load_type);

3
drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c
View File

@ -258,6 +258,8 @@ int amdgpu_vce_suspend(struct amdgpu_device *adev)
{
int i;
cancel_delayed_work_sync(&adev->vce.idle_work);
if (adev->vce.vcpu_bo == NULL)
return 0;
@ -268,7 +270,6 @@ int amdgpu_vce_suspend(struct amdgpu_device *adev)
if (i == AMDGPU_MAX_VCE_HANDLES)
return 0;
cancel_delayed_work_sync(&adev->vce.idle_work);
/* TODO: suspending running encoding sessions isn't supported */
return -EINVAL;
}

187
drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
View File

@ -36,6 +36,7 @@
#include "soc15_common.h"
#include "vcn/vcn_1_0_offset.h"
#include "vcn/vcn_1_0_sh_mask.h"
/* 1 second timeout */
#define VCN_IDLE_TIMEOUT msecs_to_jiffies(1000)
@ -120,8 +121,7 @@ int amdgpu_vcn_sw_init(struct amdgpu_device *adev)
version_major, version_minor, family_id);
}
bo_size = AMDGPU_VCN_STACK_SIZE + AMDGPU_VCN_HEAP_SIZE
+ AMDGPU_VCN_SESSION_SIZE * 40;
bo_size = AMDGPU_VCN_STACK_SIZE + AMDGPU_VCN_CONTEXT_SIZE;
if (adev->firmware.load_type != AMDGPU_FW_LOAD_PSP)
bo_size += AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8);
r = amdgpu_bo_create_kernel(adev, bo_size, PAGE_SIZE,
@ -162,11 +162,11 @@ int amdgpu_vcn_suspend(struct amdgpu_device *adev)
unsigned size;
void *ptr;
cancel_delayed_work_sync(&adev->vcn.idle_work);
if (adev->vcn.vcpu_bo == NULL)
return 0;
cancel_delayed_work_sync(&adev->vcn.idle_work);
size = amdgpu_bo_size(adev->vcn.vcpu_bo);
ptr = adev->vcn.cpu_addr;
@ -212,18 +212,161 @@ int amdgpu_vcn_resume(struct amdgpu_device *adev)
return 0;
}
static int amdgpu_vcn_pause_dpg_mode(struct amdgpu_device *adev,
struct dpg_pause_state *new_state)
{
int ret_code;
uint32_t reg_data = 0;
uint32_t reg_data2 = 0;
struct amdgpu_ring *ring;
/* pause/unpause if state is changed */
if (adev->vcn.pause_state.fw_based != new_state->fw_based) {
DRM_DEBUG("dpg pause state changed %d:%d -> %d:%d",
adev->vcn.pause_state.fw_based, adev->vcn.pause_state.jpeg,
new_state->fw_based, new_state->jpeg);
reg_data = RREG32_SOC15(UVD, 0, mmUVD_DPG_PAUSE) &
(~UVD_DPG_PAUSE__NJ_PAUSE_DPG_ACK_MASK);
if (new_state->fw_based == VCN_DPG_STATE__PAUSE) {
ret_code = 0;
if (!(reg_data & UVD_DPG_PAUSE__JPEG_PAUSE_DPG_ACK_MASK))
SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_POWER_STATUS,
UVD_POWER_STATUS__UVD_POWER_STATUS_TILES_OFF,
UVD_POWER_STATUS__UVD_POWER_STATUS_MASK, ret_code);
if (!ret_code) {
/* pause DPG non-jpeg */
reg_data |= UVD_DPG_PAUSE__NJ_PAUSE_DPG_REQ_MASK;
WREG32_SOC15(UVD, 0, mmUVD_DPG_PAUSE, reg_data);
SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_DPG_PAUSE,
UVD_DPG_PAUSE__NJ_PAUSE_DPG_ACK_MASK,
UVD_DPG_PAUSE__NJ_PAUSE_DPG_ACK_MASK, ret_code);
/* Restore */
ring = &adev->vcn.ring_enc[0];
WREG32_SOC15(UVD, 0, mmUVD_RB_BASE_LO, ring->gpu_addr);
WREG32_SOC15(UVD, 0, mmUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr));
WREG32_SOC15(UVD, 0, mmUVD_RB_SIZE, ring->ring_size / 4);
WREG32_SOC15(UVD, 0, mmUVD_RB_RPTR, lower_32_bits(ring->wptr));
WREG32_SOC15(UVD, 0, mmUVD_RB_WPTR, lower_32_bits(ring->wptr));
ring = &adev->vcn.ring_enc[1];
WREG32_SOC15(UVD, 0, mmUVD_RB_BASE_LO2, ring->gpu_addr);
WREG32_SOC15(UVD, 0, mmUVD_RB_BASE_HI2, upper_32_bits(ring->gpu_addr));
WREG32_SOC15(UVD, 0, mmUVD_RB_SIZE2, ring->ring_size / 4);
WREG32_SOC15(UVD, 0, mmUVD_RB_RPTR2, lower_32_bits(ring->wptr));
WREG32_SOC15(UVD, 0, mmUVD_RB_WPTR2, lower_32_bits(ring->wptr));
ring = &adev->vcn.ring_dec;
WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR,
RREG32_SOC15(UVD, 0, mmUVD_SCRATCH2));
SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_POWER_STATUS,
UVD_PGFSM_CONFIG__UVDM_UVDU_PWR_ON,
UVD_POWER_STATUS__UVD_POWER_STATUS_MASK, ret_code);
}
} else {
/* unpause dpg non-jpeg, no need to wait */
reg_data &= ~UVD_DPG_PAUSE__NJ_PAUSE_DPG_REQ_MASK;
WREG32_SOC15(UVD, 0, mmUVD_DPG_PAUSE, reg_data);
}
adev->vcn.pause_state.fw_based = new_state->fw_based;
}
/* pause/unpause if state is changed */
if (adev->vcn.pause_state.jpeg != new_state->jpeg) {
DRM_DEBUG("dpg pause state changed %d:%d -> %d:%d",
adev->vcn.pause_state.fw_based, adev->vcn.pause_state.jpeg,
new_state->fw_based, new_state->jpeg);
reg_data = RREG32_SOC15(UVD, 0, mmUVD_DPG_PAUSE) &
(~UVD_DPG_PAUSE__JPEG_PAUSE_DPG_ACK_MASK);
if (new_state->jpeg == VCN_DPG_STATE__PAUSE) {
ret_code = 0;
if (!(reg_data & UVD_DPG_PAUSE__NJ_PAUSE_DPG_ACK_MASK))
SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_POWER_STATUS,
UVD_POWER_STATUS__UVD_POWER_STATUS_TILES_OFF,
UVD_POWER_STATUS__UVD_POWER_STATUS_MASK, ret_code);
if (!ret_code) {
/* Make sure JPRG Snoop is disabled before sending the pause */
reg_data2 = RREG32_SOC15(UVD, 0, mmUVD_POWER_STATUS);
reg_data2 |= UVD_POWER_STATUS__JRBC_SNOOP_DIS_MASK;
WREG32_SOC15(UVD, 0, mmUVD_POWER_STATUS, reg_data2);
/* pause DPG jpeg */
reg_data |= UVD_DPG_PAUSE__JPEG_PAUSE_DPG_REQ_MASK;
WREG32_SOC15(UVD, 0, mmUVD_DPG_PAUSE, reg_data);
SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_DPG_PAUSE,
UVD_DPG_PAUSE__JPEG_PAUSE_DPG_ACK_MASK,
UVD_DPG_PAUSE__JPEG_PAUSE_DPG_ACK_MASK, ret_code);
/* Restore */
ring = &adev->vcn.ring_jpeg;
WREG32_SOC15(UVD, 0, mmUVD_LMI_JRBC_RB_VMID, 0);
WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_CNTL,
UVD_JRBC_RB_CNTL__RB_NO_FETCH_MASK |
UVD_JRBC_RB_CNTL__RB_RPTR_WR_EN_MASK);
WREG32_SOC15(UVD, 0, mmUVD_LMI_JRBC_RB_64BIT_BAR_LOW,
lower_32_bits(ring->gpu_addr));
WREG32_SOC15(UVD, 0, mmUVD_LMI_JRBC_RB_64BIT_BAR_HIGH,
upper_32_bits(ring->gpu_addr));
WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_RPTR, ring->wptr);
WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_WPTR, ring->wptr);
WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_CNTL,
UVD_JRBC_RB_CNTL__RB_RPTR_WR_EN_MASK);
ring = &adev->vcn.ring_dec;
WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR,
RREG32_SOC15(UVD, 0, mmUVD_SCRATCH2));
SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_POWER_STATUS,
UVD_PGFSM_CONFIG__UVDM_UVDU_PWR_ON,
UVD_POWER_STATUS__UVD_POWER_STATUS_MASK, ret_code);
}
} else {
/* unpause dpg jpeg, no need to wait */
reg_data &= ~UVD_DPG_PAUSE__JPEG_PAUSE_DPG_REQ_MASK;
WREG32_SOC15(UVD, 0, mmUVD_DPG_PAUSE, reg_data);
}
adev->vcn.pause_state.jpeg = new_state->jpeg;
}
return 0;
}
static void amdgpu_vcn_idle_work_handler(struct work_struct *work)
{
struct amdgpu_device *adev =
container_of(work, struct amdgpu_device, vcn.idle_work.work);
unsigned fences = amdgpu_fence_count_emitted(&adev->vcn.ring_dec);
unsigned i;
unsigned int fences = 0;
unsigned int i;
for (i = 0; i < adev->vcn.num_enc_rings; ++i) {
fences += amdgpu_fence_count_emitted(&adev->vcn.ring_enc[i]);
}
if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) {
struct dpg_pause_state new_state;
if (fences)
new_state.fw_based = VCN_DPG_STATE__PAUSE;
else
new_state.fw_based = VCN_DPG_STATE__UNPAUSE;
if (amdgpu_fence_count_emitted(&adev->vcn.ring_jpeg))
new_state.jpeg = VCN_DPG_STATE__PAUSE;
else
new_state.jpeg = VCN_DPG_STATE__UNPAUSE;
amdgpu_vcn_pause_dpg_mode(adev, &new_state);
}
fences += amdgpu_fence_count_emitted(&adev->vcn.ring_jpeg);
fences += amdgpu_fence_count_emitted(&adev->vcn.ring_dec);
if (fences == 0) {
amdgpu_gfx_off_ctrl(adev, true);
@ -250,6 +393,22 @@ void amdgpu_vcn_ring_begin_use(struct amdgpu_ring *ring)
amdgpu_device_ip_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_VCN,
AMD_PG_STATE_UNGATE);
}
if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG) {
struct dpg_pause_state new_state;
if (ring->funcs->type == AMDGPU_RING_TYPE_VCN_ENC)
new_state.fw_based = VCN_DPG_STATE__PAUSE;
else
new_state.fw_based = adev->vcn.pause_state.fw_based;
if (ring->funcs->type == AMDGPU_RING_TYPE_VCN_JPEG)
new_state.jpeg = VCN_DPG_STATE__PAUSE;
else
new_state.jpeg = adev->vcn.pause_state.jpeg;
amdgpu_vcn_pause_dpg_mode(adev, &new_state);
}
}
void amdgpu_vcn_ring_end_use(struct amdgpu_ring *ring)
@ -264,7 +423,7 @@ int amdgpu_vcn_dec_ring_test_ring(struct amdgpu_ring *ring)
unsigned i;
int r;
WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_CONTEXT_ID), 0xCAFEDEAD);
WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_SCRATCH9), 0xCAFEDEAD);
r = amdgpu_ring_alloc(ring, 3);
if (r) {
DRM_ERROR("amdgpu: cp failed to lock ring %d (%d).\n",
@ -272,11 +431,11 @@ int amdgpu_vcn_dec_ring_test_ring(struct amdgpu_ring *ring)
return r;
}
amdgpu_ring_write(ring,
PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_CONTEXT_ID), 0));
PACKET0(SOC15_REG_OFFSET(UVD, 0, mmUVD_SCRATCH9), 0));
amdgpu_ring_write(ring, 0xDEADBEEF);
amdgpu_ring_commit(ring);
for (i = 0; i < adev->usec_timeout; i++) {
tmp = RREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_CONTEXT_ID));
tmp = RREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_SCRATCH9));
if (tmp == 0xDEADBEEF)
break;
DRM_UDELAY(1);
@ -616,7 +775,7 @@ int amdgpu_vcn_jpeg_ring_test_ring(struct amdgpu_ring *ring)
unsigned i;
int r;
WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_CONTEXT_ID), 0xCAFEDEAD);
WREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_SCRATCH9), 0xCAFEDEAD);
r = amdgpu_ring_alloc(ring, 3);
if (r) {
@ -626,12 +785,12 @@ int amdgpu_vcn_jpeg_ring_test_ring(struct amdgpu_ring *ring)
}
amdgpu_ring_write(ring,
PACKETJ(SOC15_REG_OFFSET(UVD, 0, mmUVD_CONTEXT_ID), 0, 0, 0));
PACKETJ(SOC15_REG_OFFSET(UVD, 0, mmUVD_SCRATCH9), 0, 0, 0));
amdgpu_ring_write(ring, 0xDEADBEEF);
amdgpu_ring_commit(ring);
for (i = 0; i < adev->usec_timeout; i++) {
tmp = RREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_CONTEXT_ID));
tmp = RREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_SCRATCH9));
if (tmp == 0xDEADBEEF)
break;
DRM_UDELAY(1);
@ -665,7 +824,7 @@ static int amdgpu_vcn_jpeg_set_reg(struct amdgpu_ring *ring, uint32_t handle,
ib = &job->ibs[0];
ib->ptr[0] = PACKETJ(SOC15_REG_OFFSET(UVD, 0, mmUVD_JPEG_PITCH), 0, 0, PACKETJ_TYPE0);
ib->ptr[0] = PACKETJ(SOC15_REG_OFFSET(UVD, 0, mmUVD_SCRATCH9), 0, 0, PACKETJ_TYPE0);
ib->ptr[1] = 0xDEADBEEF;
for (i = 2; i < 16; i += 2) {
ib->ptr[i] = PACKETJ(0, 0, 0, PACKETJ_TYPE6);
@ -714,7 +873,7 @@ int amdgpu_vcn_jpeg_ring_test_ib(struct amdgpu_ring *ring, long timeout)
r = 0;
for (i = 0; i < adev->usec_timeout; i++) {
tmp = RREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_JPEG_PITCH));
tmp = RREG32(SOC15_REG_OFFSET(UVD, 0, mmUVD_SCRATCH9));
if (tmp == 0xDEADBEEF)
break;
DRM_UDELAY(1);

18
drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.h
View File

@ -24,9 +24,9 @@
#ifndef __AMDGPU_VCN_H__
#define __AMDGPU_VCN_H__
#define AMDGPU_VCN_STACK_SIZE (200*1024)
#define AMDGPU_VCN_HEAP_SIZE (256*1024)
#define AMDGPU_VCN_SESSION_SIZE (50*1024)
#define AMDGPU_VCN_STACK_SIZE (128*1024)
#define AMDGPU_VCN_CONTEXT_SIZE (512*1024)
#define AMDGPU_VCN_FIRMWARE_OFFSET 256
#define AMDGPU_VCN_MAX_ENC_RINGS 3
@ -56,6 +56,16 @@ enum engine_status_constants {
UVD_STATUS__RBC_BUSY = 0x1,
};
enum internal_dpg_state {
VCN_DPG_STATE__UNPAUSE = 0,
VCN_DPG_STATE__PAUSE,
};
struct dpg_pause_state {
enum internal_dpg_state fw_based;
enum internal_dpg_state jpeg;
};
struct amdgpu_vcn {
struct amdgpu_bo *vcpu_bo;
void *cpu_addr;
@ -69,6 +79,8 @@ struct amdgpu_vcn {
struct amdgpu_ring ring_jpeg;
struct amdgpu_irq_src irq;
unsigned num_enc_rings;
enum amd_powergating_state cur_state;
struct dpg_pause_state pause_state;
};
int amdgpu_vcn_sw_init(struct amdgpu_device *adev);

4
drivers/gpu/drm/amd/amdgpu/ci_dpm.c
View File

@ -6277,12 +6277,12 @@ static int ci_dpm_sw_init(void *handle)
int ret;
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
ret = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 230,
ret = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 230,
&adev->pm.dpm.thermal.irq);
if (ret)
return ret;
ret = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 231,
ret = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 231,
&adev->pm.dpm.thermal.irq);
if (ret)
return ret;

17
drivers/gpu/drm/amd/amdgpu/cik.c
View File

@ -2002,6 +2002,8 @@ int cik_set_ip_blocks(struct amdgpu_device *adev)
amdgpu_device_ip_block_add(adev, &cik_common_ip_block);
amdgpu_device_ip_block_add(adev, &gmc_v7_0_ip_block);
amdgpu_device_ip_block_add(adev, &cik_ih_ip_block);
amdgpu_device_ip_block_add(adev, &gfx_v7_2_ip_block);
amdgpu_device_ip_block_add(adev, &cik_sdma_ip_block);
if (amdgpu_dpm == -1)
amdgpu_device_ip_block_add(adev, &pp_smu_ip_block);
else
@ -2014,8 +2016,6 @@ int cik_set_ip_blocks(struct amdgpu_device *adev)
#endif
else
amdgpu_device_ip_block_add(adev, &dce_v8_2_ip_block);
amdgpu_device_ip_block_add(adev, &gfx_v7_2_ip_block);
amdgpu_device_ip_block_add(adev, &cik_sdma_ip_block);
amdgpu_device_ip_block_add(adev, &uvd_v4_2_ip_block);
amdgpu_device_ip_block_add(adev, &vce_v2_0_ip_block);
break;
@ -2023,6 +2023,8 @@ int cik_set_ip_blocks(struct amdgpu_device *adev)
amdgpu_device_ip_block_add(adev, &cik_common_ip_block);
amdgpu_device_ip_block_add(adev, &gmc_v7_0_ip_block);
amdgpu_device_ip_block_add(adev, &cik_ih_ip_block);
amdgpu_device_ip_block_add(adev, &gfx_v7_3_ip_block);
amdgpu_device_ip_block_add(adev, &cik_sdma_ip_block);
if (amdgpu_dpm == -1)
amdgpu_device_ip_block_add(adev, &pp_smu_ip_block);
else
@ -2035,8 +2037,6 @@ int cik_set_ip_blocks(struct amdgpu_device *adev)
#endif
else
amdgpu_device_ip_block_add(adev, &dce_v8_5_ip_block);
amdgpu_device_ip_block_add(adev, &gfx_v7_3_ip_block);
amdgpu_device_ip_block_add(adev, &cik_sdma_ip_block);
amdgpu_device_ip_block_add(adev, &uvd_v4_2_ip_block);
amdgpu_device_ip_block_add(adev, &vce_v2_0_ip_block);
break;
@ -2044,6 +2044,8 @@ int cik_set_ip_blocks(struct amdgpu_device *adev)
amdgpu_device_ip_block_add(adev, &cik_common_ip_block);
amdgpu_device_ip_block_add(adev, &gmc_v7_0_ip_block);
amdgpu_device_ip_block_add(adev, &cik_ih_ip_block);
amdgpu_device_ip_block_add(adev, &gfx_v7_1_ip_block);
amdgpu_device_ip_block_add(adev, &cik_sdma_ip_block);
amdgpu_device_ip_block_add(adev, &kv_smu_ip_block);
if (adev->enable_virtual_display)
amdgpu_device_ip_block_add(adev, &dce_virtual_ip_block);
@ -2053,8 +2055,7 @@ int cik_set_ip_blocks(struct amdgpu_device *adev)
#endif
else
amdgpu_device_ip_block_add(adev, &dce_v8_1_ip_block);
amdgpu_device_ip_block_add(adev, &gfx_v7_1_ip_block);
amdgpu_device_ip_block_add(adev, &cik_sdma_ip_block);
amdgpu_device_ip_block_add(adev, &uvd_v4_2_ip_block);
amdgpu_device_ip_block_add(adev, &vce_v2_0_ip_block);
break;
@ -2063,6 +2064,8 @@ int cik_set_ip_blocks(struct amdgpu_device *adev)
amdgpu_device_ip_block_add(adev, &cik_common_ip_block);
amdgpu_device_ip_block_add(adev, &gmc_v7_0_ip_block);
amdgpu_device_ip_block_add(adev, &cik_ih_ip_block);
amdgpu_device_ip_block_add(adev, &gfx_v7_2_ip_block);
amdgpu_device_ip_block_add(adev, &cik_sdma_ip_block);
amdgpu_device_ip_block_add(adev, &kv_smu_ip_block);
if (adev->enable_virtual_display)
amdgpu_device_ip_block_add(adev, &dce_virtual_ip_block);
@ -2072,8 +2075,6 @@ int cik_set_ip_blocks(struct amdgpu_device *adev)
#endif
else
amdgpu_device_ip_block_add(adev, &dce_v8_3_ip_block);
amdgpu_device_ip_block_add(adev, &gfx_v7_2_ip_block);
amdgpu_device_ip_block_add(adev, &cik_sdma_ip_block);
amdgpu_device_ip_block_add(adev, &uvd_v4_2_ip_block);
amdgpu_device_ip_block_add(adev, &vce_v2_0_ip_block);
break;

9
drivers/gpu/drm/amd/amdgpu/cik_ih.c
View File

@ -276,7 +276,7 @@ static void cik_ih_decode_iv(struct amdgpu_device *adev,
dw[2] = le32_to_cpu(adev->irq.ih.ring[ring_index + 2]);
dw[3] = le32_to_cpu(adev->irq.ih.ring[ring_index + 3]);
entry->client_id = AMDGPU_IH_CLIENTID_LEGACY;
entry->client_id = AMDGPU_IRQ_CLIENTID_LEGACY;
entry->src_id = dw[0] & 0xff;
entry->src_data[0] = dw[1] & 0xfffffff;
entry->ring_id = dw[2] & 0xff;
@ -318,7 +318,7 @@ static int cik_ih_sw_init(void *handle)
int r;
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
r = amdgpu_ih_ring_init(adev, 64 * 1024, false);
r = amdgpu_ih_ring_init(adev, &adev->irq.ih, 64 * 1024, false);
if (r)
return r;
@ -332,7 +332,7 @@ static int cik_ih_sw_fini(void *handle)
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
amdgpu_irq_fini(adev);
amdgpu_ih_ring_fini(adev);
amdgpu_ih_ring_fini(adev, &adev->irq.ih);
amdgpu_irq_remove_domain(adev);
return 0;
@ -468,8 +468,7 @@ static const struct amdgpu_ih_funcs cik_ih_funcs = {
static void cik_ih_set_interrupt_funcs(struct amdgpu_device *adev)
{
if (adev->irq.ih_funcs == NULL)
adev->irq.ih_funcs = &cik_ih_funcs;
adev->irq.ih_funcs = &cik_ih_funcs;
}
const struct amdgpu_ip_block_version cik_ih_ip_block =

26
drivers/gpu/drm/amd/amdgpu/cik_sdma.c
View File

@ -970,19 +970,19 @@ static int cik_sdma_sw_init(void *handle)
}
/* SDMA trap event */
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 224,
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 224,
&adev->sdma.trap_irq);
if (r)
return r;
/* SDMA Privileged inst */
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 241,
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 241,
&adev->sdma.illegal_inst_irq);
if (r)
return r;
/* SDMA Privileged inst */
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 247,
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 247,
&adev->sdma.illegal_inst_irq);
if (r)
return r;
@ -1370,10 +1370,8 @@ static const struct amdgpu_buffer_funcs cik_sdma_buffer_funcs = {
static void cik_sdma_set_buffer_funcs(struct amdgpu_device *adev)
{
if (adev->mman.buffer_funcs == NULL) {
adev->mman.buffer_funcs = &cik_sdma_buffer_funcs;
adev->mman.buffer_funcs_ring = &adev->sdma.instance[0].ring;
}
adev->mman.buffer_funcs = &cik_sdma_buffer_funcs;
adev->mman.buffer_funcs_ring = &adev->sdma.instance[0].ring;
}
static const struct amdgpu_vm_pte_funcs cik_sdma_vm_pte_funcs = {
@ -1389,15 +1387,13 @@ static void cik_sdma_set_vm_pte_funcs(struct amdgpu_device *adev)
struct drm_gpu_scheduler *sched;
unsigned i;
if (adev->vm_manager.vm_pte_funcs == NULL) {
adev->vm_manager.vm_pte_funcs = &cik_sdma_vm_pte_funcs;
for (i = 0; i < adev->sdma.num_instances; i++) {
sched = &adev->sdma.instance[i].ring.sched;
adev->vm_manager.vm_pte_rqs[i] =
&sched->sched_rq[DRM_SCHED_PRIORITY_KERNEL];
}
adev->vm_manager.vm_pte_num_rqs = adev->sdma.num_instances;
adev->vm_manager.vm_pte_funcs = &cik_sdma_vm_pte_funcs;
for (i = 0; i < adev->sdma.num_instances; i++) {
sched = &adev->sdma.instance[i].ring.sched;
adev->vm_manager.vm_pte_rqs[i] =
&sched->sched_rq[DRM_SCHED_PRIORITY_KERNEL];
}
adev->vm_manager.vm_pte_num_rqs = adev->sdma.num_instances;
}
const struct amdgpu_ip_block_version cik_sdma_ip_block =

9
drivers/gpu/drm/amd/amdgpu/cz_ih.c
View File

@ -255,7 +255,7 @@ static void cz_ih_decode_iv(struct amdgpu_device *adev,
dw[2] = le32_to_cpu(adev->irq.ih.ring[ring_index + 2]);
dw[3] = le32_to_cpu(adev->irq.ih.ring[ring_index + 3]);
entry->client_id = AMDGPU_IH_CLIENTID_LEGACY;
entry->client_id = AMDGPU_IRQ_CLIENTID_LEGACY;
entry->src_id = dw[0] & 0xff;
entry->src_data[0] = dw[1] & 0xfffffff;
entry->ring_id = dw[2] & 0xff;
@ -297,7 +297,7 @@ static int cz_ih_sw_init(void *handle)
int r;
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
r = amdgpu_ih_ring_init(adev, 64 * 1024, false);
r = amdgpu_ih_ring_init(adev, &adev->irq.ih, 64 * 1024, false);
if (r)
return r;
@ -311,7 +311,7 @@ static int cz_ih_sw_fini(void *handle)
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
amdgpu_irq_fini(adev);
amdgpu_ih_ring_fini(adev);
amdgpu_ih_ring_fini(adev, &adev->irq.ih);
amdgpu_irq_remove_domain(adev);
return 0;
@ -449,8 +449,7 @@ static const struct amdgpu_ih_funcs cz_ih_funcs = {
static void cz_ih_set_interrupt_funcs(struct amdgpu_device *adev)
{
if (adev->irq.ih_funcs == NULL)
adev->irq.ih_funcs = &cz_ih_funcs;
adev->irq.ih_funcs = &cz_ih_funcs;
}
const struct amdgpu_ip_block_version cz_ih_ip_block =

9
drivers/gpu/drm/amd/amdgpu/dce_v10_0.c
View File

@ -2746,19 +2746,19 @@ static int dce_v10_0_sw_init(void *handle)
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
for (i = 0; i < adev->mode_info.num_crtc; i++) {
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, i + 1, &adev->crtc_irq);
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, i + 1, &adev->crtc_irq);
if (r)
return r;
}
for (i = VISLANDS30_IV_SRCID_D1_GRPH_PFLIP; i < 20; i += 2) {
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, i, &adev->pageflip_irq);
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, i, &adev->pageflip_irq);
if (r)
return r;
}
/* HPD hotplug */
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_HOTPLUG_DETECT_A, &adev->hpd_irq);
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_HOTPLUG_DETECT_A, &adev->hpd_irq);
if (r)
return r;
@ -3570,8 +3570,7 @@ static const struct amdgpu_display_funcs dce_v10_0_display_funcs = {
static void dce_v10_0_set_display_funcs(struct amdgpu_device *adev)
{
if (adev->mode_info.funcs == NULL)
adev->mode_info.funcs = &dce_v10_0_display_funcs;
adev->mode_info.funcs = &dce_v10_0_display_funcs;
}
static const struct amdgpu_irq_src_funcs dce_v10_0_crtc_irq_funcs = {

9
drivers/gpu/drm/amd/amdgpu/dce_v11_0.c
View File

@ -2867,19 +2867,19 @@ static int dce_v11_0_sw_init(void *handle)
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
for (i = 0; i < adev->mode_info.num_crtc; i++) {
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, i + 1, &adev->crtc_irq);
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, i + 1, &adev->crtc_irq);
if (r)
return r;
}
for (i = VISLANDS30_IV_SRCID_D1_GRPH_PFLIP; i < 20; i += 2) {
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, i, &adev->pageflip_irq);
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, i, &adev->pageflip_irq);
if (r)
return r;
}
/* HPD hotplug */
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_HOTPLUG_DETECT_A, &adev->hpd_irq);
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_HOTPLUG_DETECT_A, &adev->hpd_irq);
if (r)
return r;
@ -3702,8 +3702,7 @@ static const struct amdgpu_display_funcs dce_v11_0_display_funcs = {
static void dce_v11_0_set_display_funcs(struct amdgpu_device *adev)
{
if (adev->mode_info.funcs == NULL)
adev->mode_info.funcs = &dce_v11_0_display_funcs;
adev->mode_info.funcs = &dce_v11_0_display_funcs;
}
static const struct amdgpu_irq_src_funcs dce_v11_0_crtc_irq_funcs = {

9
drivers/gpu/drm/amd/amdgpu/dce_v6_0.c
View File

@ -2616,19 +2616,19 @@ static int dce_v6_0_sw_init(void *handle)
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
for (i = 0; i < adev->mode_info.num_crtc; i++) {
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, i + 1, &adev->crtc_irq);
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, i + 1, &adev->crtc_irq);
if (r)
return r;
}
for (i = 8; i < 20; i += 2) {
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, i, &adev->pageflip_irq);
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, i, &adev->pageflip_irq);
if (r)
return r;
}
/* HPD hotplug */
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 42, &adev->hpd_irq);
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 42, &adev->hpd_irq);
if (r)
return r;
@ -3376,8 +3376,7 @@ static const struct amdgpu_display_funcs dce_v6_0_display_funcs = {
static void dce_v6_0_set_display_funcs(struct amdgpu_device *adev)
{
if (adev->mode_info.funcs == NULL)
adev->mode_info.funcs = &dce_v6_0_display_funcs;
adev->mode_info.funcs = &dce_v6_0_display_funcs;
}
static const struct amdgpu_irq_src_funcs dce_v6_0_crtc_irq_funcs = {

9
drivers/gpu/drm/amd/amdgpu/dce_v8_0.c
View File

@ -2643,19 +2643,19 @@ static int dce_v8_0_sw_init(void *handle)
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
for (i = 0; i < adev->mode_info.num_crtc; i++) {
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, i + 1, &adev->crtc_irq);
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, i + 1, &adev->crtc_irq);
if (r)
return r;
}
for (i = 8; i < 20; i += 2) {
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, i, &adev->pageflip_irq);
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, i, &adev->pageflip_irq);
if (r)
return r;
}
/* HPD hotplug */
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 42, &adev->hpd_irq);
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 42, &adev->hpd_irq);
if (r)
return r;
@ -3458,8 +3458,7 @@ static const struct amdgpu_display_funcs dce_v8_0_display_funcs = {
static void dce_v8_0_set_display_funcs(struct amdgpu_device *adev)
{
if (adev->mode_info.funcs == NULL)
adev->mode_info.funcs = &dce_v8_0_display_funcs;
adev->mode_info.funcs = &dce_v8_0_display_funcs;
}
static const struct amdgpu_irq_src_funcs dce_v8_0_crtc_irq_funcs = {

5
drivers/gpu/drm/amd/amdgpu/dce_virtual.c
View File

@ -372,7 +372,7 @@ static int dce_virtual_sw_init(void *handle)
int r, i;
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_SMU_DISP_TIMER2_TRIGGER, &adev->crtc_irq);
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_SMU_DISP_TIMER2_TRIGGER, &adev->crtc_irq);
if (r)
return r;
@ -649,8 +649,7 @@ static const struct amdgpu_display_funcs dce_virtual_display_funcs = {
static void dce_virtual_set_display_funcs(struct amdgpu_device *adev)
{
if (adev->mode_info.funcs == NULL)
adev->mode_info.funcs = &dce_virtual_display_funcs;
adev->mode_info.funcs = &dce_virtual_display_funcs;
}
static int dce_virtual_pageflip(struct amdgpu_device *adev,

10
drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c
View File

@ -1552,7 +1552,7 @@ static void gfx_v6_0_config_init(struct amdgpu_device *adev)
adev->gfx.config.double_offchip_lds_buf = 0;
}
static void gfx_v6_0_gpu_init(struct amdgpu_device *adev)
static void gfx_v6_0_constants_init(struct amdgpu_device *adev)
{
u32 gb_addr_config = 0;
u32 mc_shared_chmap, mc_arb_ramcfg;
@ -3094,15 +3094,15 @@ static int gfx_v6_0_sw_init(void *handle)
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
int i, r;
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 181, &adev->gfx.eop_irq);
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 181, &adev->gfx.eop_irq);
if (r)
return r;
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 184, &adev->gfx.priv_reg_irq);
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 184, &adev->gfx.priv_reg_irq);
if (r)
return r;
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 185, &adev->gfx.priv_inst_irq);
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 185, &adev->gfx.priv_inst_irq);
if (r)
return r;
@ -3175,7 +3175,7 @@ static int gfx_v6_0_hw_init(void *handle)
int r;
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
gfx_v6_0_gpu_init(adev);
gfx_v6_0_constants_init(adev);
r = gfx_v6_0_rlc_resume(adev);
if (r)

16
drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c
View File

@ -1886,14 +1886,14 @@ static void gfx_v7_0_config_init(struct amdgpu_device *adev)
}
/**
* gfx_v7_0_gpu_init - setup the 3D engine
* gfx_v7_0_constants_init - setup the 3D engine
*
* @adev: amdgpu_device pointer
*
* Configures the 3D engine and tiling configuration
* registers so that the 3D engine is usable.
* init the gfx constants such as the 3D engine, tiling configuration
* registers, maximum number of quad pipes, render backends...
*/
static void gfx_v7_0_gpu_init(struct amdgpu_device *adev)
static void gfx_v7_0_constants_init(struct amdgpu_device *adev)
{
u32 sh_mem_cfg, sh_static_mem_cfg, sh_mem_base;
u32 tmp;
@ -4516,18 +4516,18 @@ static int gfx_v7_0_sw_init(void *handle)
adev->gfx.mec.num_queue_per_pipe = 8;
/* EOP Event */
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 181, &adev->gfx.eop_irq);
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 181, &adev->gfx.eop_irq);
if (r)
return r;
/* Privileged reg */
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 184,
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 184,
&adev->gfx.priv_reg_irq);
if (r)
return r;
/* Privileged inst */
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 185,
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 185,
&adev->gfx.priv_inst_irq);
if (r)
return r;
@ -4624,7 +4624,7 @@ static int gfx_v7_0_hw_init(void *handle)
int r;
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
gfx_v7_0_gpu_init(adev);
gfx_v7_0_constants_init(adev);
/* init rlc */
r = gfx_v7_0_rlc_resume(adev);

437
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
View File

@ -1173,64 +1173,61 @@ static int gfx_v8_0_init_microcode(struct amdgpu_device *adev)
}
}
if (adev->firmware.load_type == AMDGPU_FW_LOAD_SMU) {
info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_PFP];
info->ucode_id = AMDGPU_UCODE_ID_CP_PFP;
info->fw = adev->gfx.pfp_fw;
header = (const struct common_firmware_header *)info->fw->data;
adev->firmware.fw_size +=
ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE);
info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_PFP];
info->ucode_id = AMDGPU_UCODE_ID_CP_PFP;
info->fw = adev->gfx.pfp_fw;
header = (const struct common_firmware_header *)info->fw->data;
adev->firmware.fw_size +=
ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE);
info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_ME];
info->ucode_id = AMDGPU_UCODE_ID_CP_ME;
info->fw = adev->gfx.me_fw;
header = (const struct common_firmware_header *)info->fw->data;
adev->firmware.fw_size +=
ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE);
info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_ME];
info->ucode_id = AMDGPU_UCODE_ID_CP_ME;
info->fw = adev->gfx.me_fw;
header = (const struct common_firmware_header *)info->fw->data;
adev->firmware.fw_size +=
ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE);
info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_CE];
info->ucode_id = AMDGPU_UCODE_ID_CP_CE;
info->fw = adev->gfx.ce_fw;
header = (const struct common_firmware_header *)info->fw->data;
adev->firmware.fw_size +=
ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE);
info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_CE];
info->ucode_id = AMDGPU_UCODE_ID_CP_CE;
info->fw = adev->gfx.ce_fw;
header = (const struct common_firmware_header *)info->fw->data;
adev->firmware.fw_size +=
ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE);
info = &adev->firmware.ucode[AMDGPU_UCODE_ID_RLC_G];
info->ucode_id = AMDGPU_UCODE_ID_RLC_G;
info->fw = adev->gfx.rlc_fw;
header = (const struct common_firmware_header *)info->fw->data;
adev->firmware.fw_size +=
ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE);
info = &adev->firmware.ucode[AMDGPU_UCODE_ID_RLC_G];
info->ucode_id = AMDGPU_UCODE_ID_RLC_G;
info->fw = adev->gfx.rlc_fw;
header = (const struct common_firmware_header *)info->fw->data;
adev->firmware.fw_size +=
ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE);
info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_MEC1];
info->ucode_id = AMDGPU_UCODE_ID_CP_MEC1;
info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_MEC1];
info->ucode_id = AMDGPU_UCODE_ID_CP_MEC1;
info->fw = adev->gfx.mec_fw;
header = (const struct common_firmware_header *)info->fw->data;
adev->firmware.fw_size +=
ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE);
/* we need account JT in */
cp_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.mec_fw->data;
adev->firmware.fw_size +=
ALIGN(le32_to_cpu(cp_hdr->jt_size) << 2, PAGE_SIZE);
if (amdgpu_sriov_vf(adev)) {
info = &adev->firmware.ucode[AMDGPU_UCODE_ID_STORAGE];
info->ucode_id = AMDGPU_UCODE_ID_STORAGE;
info->fw = adev->gfx.mec_fw;
adev->firmware.fw_size +=
ALIGN(le32_to_cpu(64 * PAGE_SIZE), PAGE_SIZE);
}
if (adev->gfx.mec2_fw) {
info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_MEC2];
info->ucode_id = AMDGPU_UCODE_ID_CP_MEC2;
info->fw = adev->gfx.mec2_fw;
header = (const struct common_firmware_header *)info->fw->data;
adev->firmware.fw_size +=
ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE);
/* we need account JT in */
cp_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.mec_fw->data;
adev->firmware.fw_size +=
ALIGN(le32_to_cpu(cp_hdr->jt_size) << 2, PAGE_SIZE);
if (amdgpu_sriov_vf(adev)) {
info = &adev->firmware.ucode[AMDGPU_UCODE_ID_STORAGE];
info->ucode_id = AMDGPU_UCODE_ID_STORAGE;
info->fw = adev->gfx.mec_fw;
adev->firmware.fw_size +=
ALIGN(le32_to_cpu(64 * PAGE_SIZE), PAGE_SIZE);
}
if (adev->gfx.mec2_fw) {
info = &adev->firmware.ucode[AMDGPU_UCODE_ID_CP_MEC2];
info->ucode_id = AMDGPU_UCODE_ID_CP_MEC2;
info->fw = adev->gfx.mec2_fw;
header = (const struct common_firmware_header *)info->fw->data;
adev->firmware.fw_size +=
ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE);
}
}
out:
@ -2048,36 +2045,31 @@ static int gfx_v8_0_sw_init(void *handle)
adev->gfx.mec.num_pipe_per_mec = 4;
adev->gfx.mec.num_queue_per_pipe = 8;
/* KIQ event */
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_CP_INT_IB2, &adev->gfx.kiq.irq);
if (r)
return r;
/* EOP Event */
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_CP_END_OF_PIPE, &adev->gfx.eop_irq);
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_CP_END_OF_PIPE, &adev->gfx.eop_irq);
if (r)
return r;
/* Privileged reg */
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_CP_PRIV_REG_FAULT,
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_CP_PRIV_REG_FAULT,
&adev->gfx.priv_reg_irq);
if (r)
return r;
/* Privileged inst */
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_CP_PRIV_INSTR_FAULT,
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_CP_PRIV_INSTR_FAULT,
&adev->gfx.priv_inst_irq);
if (r)
return r;
/* Add CP EDC/ECC irq */
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_CP_ECC_ERROR,
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_CP_ECC_ERROR,
&adev->gfx.cp_ecc_error_irq);
if (r)
return r;
/* SQ interrupts. */
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_SQ_INTERRUPT_MSG,
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_SQ_INTERRUPT_MSG,
&adev->gfx.sq_irq);
if (r) {
DRM_ERROR("amdgpu_irq_add() for SQ failed: %d\n", r);
@ -3835,7 +3827,7 @@ static void gfx_v8_0_config_init(struct amdgpu_device *adev)
}
}
static void gfx_v8_0_gpu_init(struct amdgpu_device *adev)
static void gfx_v8_0_constants_init(struct amdgpu_device *adev)
{
u32 tmp, sh_static_mem_cfg;
int i;
@ -4181,65 +4173,11 @@ static void gfx_v8_0_rlc_start(struct amdgpu_device *adev)
udelay(50);
}
static int gfx_v8_0_rlc_load_microcode(struct amdgpu_device *adev)
{
const struct rlc_firmware_header_v2_0 *hdr;
const __le32 *fw_data;
unsigned i, fw_size;
if (!adev->gfx.rlc_fw)
return -EINVAL;
hdr = (const struct rlc_firmware_header_v2_0 *)adev->gfx.rlc_fw->data;
amdgpu_ucode_print_rlc_hdr(&hdr->header);
fw_data = (const __le32 *)(adev->gfx.rlc_fw->data +
le32_to_cpu(hdr->header.ucode_array_offset_bytes));
fw_size = le32_to_cpu(hdr->header.ucode_size_bytes) / 4;
WREG32(mmRLC_GPM_UCODE_ADDR, 0);
for (i = 0; i < fw_size; i++)
WREG32(mmRLC_GPM_UCODE_DATA, le32_to_cpup(fw_data++));
WREG32(mmRLC_GPM_UCODE_ADDR, adev->gfx.rlc_fw_version);
return 0;
}
static int gfx_v8_0_rlc_resume(struct amdgpu_device *adev)
{
int r;
u32 tmp;
gfx_v8_0_rlc_stop(adev);
/* disable CG */
tmp = RREG32(mmRLC_CGCG_CGLS_CTRL);
tmp &= ~(RLC_CGCG_CGLS_CTRL__CGCG_EN_MASK |
RLC_CGCG_CGLS_CTRL__CGLS_EN_MASK);
WREG32(mmRLC_CGCG_CGLS_CTRL, tmp);
if (adev->asic_type == CHIP_POLARIS11 ||
adev->asic_type == CHIP_POLARIS10 ||
adev->asic_type == CHIP_POLARIS12 ||
adev->asic_type == CHIP_VEGAM) {
tmp = RREG32(mmRLC_CGCG_CGLS_CTRL_3D);
tmp &= ~0x3;
WREG32(mmRLC_CGCG_CGLS_CTRL_3D, tmp);
}
/* disable PG */
WREG32(mmRLC_PG_CNTL, 0);
gfx_v8_0_rlc_reset(adev);
gfx_v8_0_init_pg(adev);
if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) {
/* legacy rlc firmware loading */
r = gfx_v8_0_rlc_load_microcode(adev);
if (r)
return r;
}
gfx_v8_0_rlc_start(adev);
return 0;
@ -4265,63 +4203,6 @@ static void gfx_v8_0_cp_gfx_enable(struct amdgpu_device *adev, bool enable)
udelay(50);
}
static int gfx_v8_0_cp_gfx_load_microcode(struct amdgpu_device *adev)
{
const struct gfx_firmware_header_v1_0 *pfp_hdr;
const struct gfx_firmware_header_v1_0 *ce_hdr;
const struct gfx_firmware_header_v1_0 *me_hdr;
const __le32 *fw_data;
unsigned i, fw_size;
if (!adev->gfx.me_fw || !adev->gfx.pfp_fw || !adev->gfx.ce_fw)
return -EINVAL;
pfp_hdr = (const struct gfx_firmware_header_v1_0 *)
adev->gfx.pfp_fw->data;
ce_hdr = (const struct gfx_firmware_header_v1_0 *)
adev->gfx.ce_fw->data;
me_hdr = (const struct gfx_firmware_header_v1_0 *)
adev->gfx.me_fw->data;
amdgpu_ucode_print_gfx_hdr(&pfp_hdr->header);
amdgpu_ucode_print_gfx_hdr(&ce_hdr->header);
amdgpu_ucode_print_gfx_hdr(&me_hdr->header);
gfx_v8_0_cp_gfx_enable(adev, false);
/* PFP */
fw_data = (const __le32 *)
(adev->gfx.pfp_fw->data +
le32_to_cpu(pfp_hdr->header.ucode_array_offset_bytes));
fw_size = le32_to_cpu(pfp_hdr->header.ucode_size_bytes) / 4;
WREG32(mmCP_PFP_UCODE_ADDR, 0);
for (i = 0; i < fw_size; i++)
WREG32(mmCP_PFP_UCODE_DATA, le32_to_cpup(fw_data++));
WREG32(mmCP_PFP_UCODE_ADDR, adev->gfx.pfp_fw_version);
/* CE */
fw_data = (const __le32 *)
(adev->gfx.ce_fw->data +
le32_to_cpu(ce_hdr->header.ucode_array_offset_bytes));
fw_size = le32_to_cpu(ce_hdr->header.ucode_size_bytes) / 4;
WREG32(mmCP_CE_UCODE_ADDR, 0);
for (i = 0; i < fw_size; i++)
WREG32(mmCP_CE_UCODE_DATA, le32_to_cpup(fw_data++));
WREG32(mmCP_CE_UCODE_ADDR, adev->gfx.ce_fw_version);
/* ME */
fw_data = (const __le32 *)
(adev->gfx.me_fw->data +
le32_to_cpu(me_hdr->header.ucode_array_offset_bytes));
fw_size = le32_to_cpu(me_hdr->header.ucode_size_bytes) / 4;
WREG32(mmCP_ME_RAM_WADDR, 0);
for (i = 0; i < fw_size; i++)
WREG32(mmCP_ME_RAM_DATA, le32_to_cpup(fw_data++));
WREG32(mmCP_ME_RAM_WADDR, adev->gfx.me_fw_version);
return 0;
}
static u32 gfx_v8_0_get_csb_size(struct amdgpu_device *adev)
{
u32 count = 0;
@ -4521,52 +4402,6 @@ static void gfx_v8_0_cp_compute_enable(struct amdgpu_device *adev, bool enable)
udelay(50);
}
static int gfx_v8_0_cp_compute_load_microcode(struct amdgpu_device *adev)
{
const struct gfx_firmware_header_v1_0 *mec_hdr;
const __le32 *fw_data;
unsigned i, fw_size;
if (!adev->gfx.mec_fw)
return -EINVAL;
gfx_v8_0_cp_compute_enable(adev, false);
mec_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.mec_fw->data;
amdgpu_ucode_print_gfx_hdr(&mec_hdr->header);
fw_data = (const __le32 *)
(adev->gfx.mec_fw->data +
le32_to_cpu(mec_hdr->header.ucode_array_offset_bytes));
fw_size = le32_to_cpu(mec_hdr->header.ucode_size_bytes) / 4;
/* MEC1 */
WREG32(mmCP_MEC_ME1_UCODE_ADDR, 0);
for (i = 0; i < fw_size; i++)
WREG32(mmCP_MEC_ME1_UCODE_DATA, le32_to_cpup(fw_data+i));
WREG32(mmCP_MEC_ME1_UCODE_ADDR, adev->gfx.mec_fw_version);
/* Loading MEC2 firmware is only necessary if MEC2 should run different microcode than MEC1. */
if (adev->gfx.mec2_fw) {
const struct gfx_firmware_header_v1_0 *mec2_hdr;
mec2_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.mec2_fw->data;
amdgpu_ucode_print_gfx_hdr(&mec2_hdr->header);
fw_data = (const __le32 *)
(adev->gfx.mec2_fw->data +
le32_to_cpu(mec2_hdr->header.ucode_array_offset_bytes));
fw_size = le32_to_cpu(mec2_hdr->header.ucode_size_bytes) / 4;
WREG32(mmCP_MEC_ME2_UCODE_ADDR, 0);
for (i = 0; i < fw_size; i++)
WREG32(mmCP_MEC_ME2_UCODE_DATA, le32_to_cpup(fw_data+i));
WREG32(mmCP_MEC_ME2_UCODE_ADDR, adev->gfx.mec2_fw_version);
}
return 0;
}
/* KIQ functions */
static void gfx_v8_0_kiq_setting(struct amdgpu_ring *ring)
{
@ -4892,7 +4727,7 @@ static int gfx_v8_0_kcq_init_queue(struct amdgpu_ring *ring)
struct vi_mqd *mqd = ring->mqd_ptr;
int mqd_idx = ring - &adev->gfx.compute_ring[0];
if (!adev->in_gpu_reset && !adev->gfx.in_suspend) {
if (!adev->in_gpu_reset && !adev->in_suspend) {
memset((void *)mqd, 0, sizeof(struct vi_mqd_allocation));
((struct vi_mqd_allocation *)mqd)->dynamic_cu_mask = 0xFFFFFFFF;
((struct vi_mqd_allocation *)mqd)->dynamic_rb_mask = 0xFFFFFFFF;
@ -5000,17 +4835,6 @@ static int gfx_v8_0_cp_resume(struct amdgpu_device *adev)
if (!(adev->flags & AMD_IS_APU))
gfx_v8_0_enable_gui_idle_interrupt(adev, false);
if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) {
/* legacy firmware loading */
r = gfx_v8_0_cp_gfx_load_microcode(adev);
if (r)
return r;
r = gfx_v8_0_cp_compute_load_microcode(adev);
if (r)
return r;
}
r = gfx_v8_0_kiq_resume(adev);
if (r)
return r;
@ -5039,7 +4863,7 @@ static int gfx_v8_0_hw_init(void *handle)
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
gfx_v8_0_init_golden_registers(adev);
gfx_v8_0_gpu_init(adev);
gfx_v8_0_constants_init(adev);
r = gfx_v8_0_rlc_resume(adev);
if (r)
@ -5080,6 +4904,55 @@ static int gfx_v8_0_kcq_disable(struct amdgpu_device *adev)
return r;
}
static bool gfx_v8_0_is_idle(void *handle)
{
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
if (REG_GET_FIELD(RREG32(mmGRBM_STATUS), GRBM_STATUS, GUI_ACTIVE)
|| RREG32(mmGRBM_STATUS2) != 0x8)
return false;
else
return true;
}
static bool gfx_v8_0_rlc_is_idle(void *handle)
{
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
if (RREG32(mmGRBM_STATUS2) != 0x8)
return false;
else
return true;
}
static int gfx_v8_0_wait_for_rlc_idle(void *handle)
{
unsigned int i;
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
for (i = 0; i < adev->usec_timeout; i++) {
if (gfx_v8_0_rlc_is_idle(handle))
return 0;
udelay(1);
}
return -ETIMEDOUT;
}
static int gfx_v8_0_wait_for_idle(void *handle)
{
unsigned int i;
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
for (i = 0; i < adev->usec_timeout; i++) {
if (gfx_v8_0_is_idle(handle))
return 0;
udelay(1);
}
return -ETIMEDOUT;
}
static int gfx_v8_0_hw_fini(void *handle)
{
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
@ -5098,51 +4971,27 @@ static int gfx_v8_0_hw_fini(void *handle)
pr_debug("For SRIOV client, shouldn't do anything.\n");
return 0;
}
gfx_v8_0_cp_enable(adev, false);
gfx_v8_0_rlc_stop(adev);
adev->gfx.rlc.funcs->enter_safe_mode(adev);
if (!gfx_v8_0_wait_for_idle(adev))
gfx_v8_0_cp_enable(adev, false);
else
pr_err("cp is busy, skip halt cp\n");
if (!gfx_v8_0_wait_for_rlc_idle(adev))
gfx_v8_0_rlc_stop(adev);
else
pr_err("rlc is busy, skip halt rlc\n");
adev->gfx.rlc.funcs->exit_safe_mode(adev);
return 0;
}
static int gfx_v8_0_suspend(void *handle)
{
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
adev->gfx.in_suspend = true;
return gfx_v8_0_hw_fini(adev);
return gfx_v8_0_hw_fini(handle);
}
static int gfx_v8_0_resume(void *handle)
{
int r;
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
r = gfx_v8_0_hw_init(adev);
adev->gfx.in_suspend = false;
return r;
}
static bool gfx_v8_0_is_idle(void *handle)
{
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
if (REG_GET_FIELD(RREG32(mmGRBM_STATUS), GRBM_STATUS, GUI_ACTIVE))
return false;
else
return true;
}
static int gfx_v8_0_wait_for_idle(void *handle)
{
unsigned i;
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
for (i = 0; i < adev->usec_timeout; i++) {
if (gfx_v8_0_is_idle(handle))
return 0;
udelay(1);
}
return -ETIMEDOUT;
return gfx_v8_0_hw_init(handle);
}
static bool gfx_v8_0_check_soft_reset(void *handle)
@ -7013,52 +6862,6 @@ static int gfx_v8_0_sq_irq(struct amdgpu_device *adev,
return 0;
}
static int gfx_v8_0_kiq_set_interrupt_state(struct amdgpu_device *adev,
struct amdgpu_irq_src *src,
unsigned int type,
enum amdgpu_interrupt_state state)
{
struct amdgpu_ring *ring = &(adev->gfx.kiq.ring);
switch (type) {
case AMDGPU_CP_KIQ_IRQ_DRIVER0:
WREG32_FIELD(CPC_INT_CNTL, GENERIC2_INT_ENABLE,
state == AMDGPU_IRQ_STATE_DISABLE ? 0 : 1);
if (ring->me == 1)
WREG32_FIELD_OFFSET(CP_ME1_PIPE0_INT_CNTL,
ring->pipe,
GENERIC2_INT_ENABLE,
state == AMDGPU_IRQ_STATE_DISABLE ? 0 : 1);
else
WREG32_FIELD_OFFSET(CP_ME2_PIPE0_INT_CNTL,
ring->pipe,
GENERIC2_INT_ENABLE,
state == AMDGPU_IRQ_STATE_DISABLE ? 0 : 1);
break;
default:
BUG(); /* kiq only support GENERIC2_INT now */
break;
}
return 0;
}
static int gfx_v8_0_kiq_irq(struct amdgpu_device *adev,
struct amdgpu_irq_src *source,
struct amdgpu_iv_entry *entry)
{
u8 me_id, pipe_id, queue_id;
struct amdgpu_ring *ring = &(adev->gfx.kiq.ring);
me_id = (entry->ring_id & 0x0c) >> 2;
pipe_id = (entry->ring_id & 0x03) >> 0;
queue_id = (entry->ring_id & 0x70) >> 4;
DRM_DEBUG("IH: CPC GENERIC2_INT, me:%d, pipe:%d, queue:%d\n",
me_id, pipe_id, queue_id);
amdgpu_fence_process(ring);
return 0;
}
static const struct amd_ip_funcs gfx_v8_0_ip_funcs = {
.name = "gfx_v8_0",
.early_init = gfx_v8_0_early_init,
@ -7209,11 +7012,6 @@ static const struct amdgpu_irq_src_funcs gfx_v8_0_priv_inst_irq_funcs = {
.process = gfx_v8_0_priv_inst_irq,
};
static const struct amdgpu_irq_src_funcs gfx_v8_0_kiq_irq_funcs = {
.set = gfx_v8_0_kiq_set_interrupt_state,
.process = gfx_v8_0_kiq_irq,
};
static const struct amdgpu_irq_src_funcs gfx_v8_0_cp_ecc_error_irq_funcs = {
.set = gfx_v8_0_set_cp_ecc_int_state,
.process = gfx_v8_0_cp_ecc_error_irq,
@ -7235,9 +7033,6 @@ static void gfx_v8_0_set_irq_funcs(struct amdgpu_device *adev)
adev->gfx.priv_inst_irq.num_types = 1;
adev->gfx.priv_inst_irq.funcs = &gfx_v8_0_priv_inst_irq_funcs;
adev->gfx.kiq.irq.num_types = AMDGPU_CP_KIQ_IRQ_LAST;
adev->gfx.kiq.irq.funcs = &gfx_v8_0_kiq_irq_funcs;
adev->gfx.cp_ecc_error_irq.num_types = 1;
adev->gfx.cp_ecc_error_irq.funcs = &gfx_v8_0_cp_ecc_error_irq_funcs;

239
drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
View File

@ -97,6 +97,7 @@ MODULE_FIRMWARE("amdgpu/raven2_rlc.bin");
static const struct soc15_reg_golden golden_settings_gc_9_0[] =
{
SOC15_REG_GOLDEN_VALUE(GC, 0, mmDB_DEBUG2, 0xf00fffff, 0x00000400),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmDB_DEBUG3, 0x80000000, 0x80000000),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmGB_GPU_ID, 0x0000000f, 0x00000000),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmPA_SC_BINNER_EVENT_CNTL_3, 0x00000003, 0x82400024),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmPA_SC_ENHANCE, 0x3fffffff, 0x00000001),
@ -133,7 +134,10 @@ static const struct soc15_reg_golden golden_settings_gc_9_0_vg10[] =
SOC15_REG_GOLDEN_VALUE(GC, 0, mmRMI_UTCL1_CNTL2, 0x00030000, 0x00020000),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmSPI_CONFIG_CNTL_1, 0x0000000f, 0x01000107),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmTD_CNTL, 0x00001800, 0x00000800),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmWD_UTCL1_CNTL, 0x08000000, 0x08000080)
SOC15_REG_GOLDEN_VALUE(GC, 0, mmWD_UTCL1_CNTL, 0x08000000, 0x08000080),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_MEC1_F32_INT_DIS, 0x00000000, 0x00000800),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_MEC2_F32_INT_DIS, 0x00000000, 0x00000800),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_DEBUG, 0x00000000, 0x00008000)
};
static const struct soc15_reg_golden golden_settings_gc_9_0_vg20[] =
@ -173,7 +177,10 @@ static const struct soc15_reg_golden golden_settings_gc_9_1[] =
SOC15_REG_GOLDEN_VALUE(GC, 0, mmTCP_CHAN_STEER_LO, 0xffffffff, 0x00003120),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmVGT_CACHE_INVALIDATION, 0x3fff3af3, 0x19200000),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmVGT_GS_MAX_WAVE_ID, 0x00000fff, 0x000000ff),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmWD_UTCL1_CNTL, 0x08000000, 0x08000080)
SOC15_REG_GOLDEN_VALUE(GC, 0, mmWD_UTCL1_CNTL, 0x08000000, 0x08000080),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_MEC1_F32_INT_DIS, 0x00000000, 0x00000800),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_MEC2_F32_INT_DIS, 0x00000000, 0x00000800),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_DEBUG, 0x00000000, 0x00008000)
};
static const struct soc15_reg_golden golden_settings_gc_9_1_rv1[] =
@ -247,7 +254,10 @@ static const struct soc15_reg_golden golden_settings_gc_9_2_1_vg12[] =
SOC15_REG_GOLDEN_VALUE(GC, 0, mmSPI_CONFIG_CNTL_1, 0xffff03ff, 0x01000107),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmTCP_CHAN_STEER_HI, 0xffffffff, 0x00000000),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmTCP_CHAN_STEER_LO, 0xffffffff, 0x76325410),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmTD_CNTL, 0x01bd9f33, 0x01000000)
SOC15_REG_GOLDEN_VALUE(GC, 0, mmTD_CNTL, 0x01bd9f33, 0x01000000),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_MEC1_F32_INT_DIS, 0x00000000, 0x00000800),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_MEC2_F32_INT_DIS, 0x00000000, 0x00000800),
SOC15_REG_GOLDEN_VALUE(GC, 0, mmCP_DEBUG, 0x00000000, 0x00008000)
};
static const u32 GFX_RLC_SRM_INDEX_CNTL_ADDR_OFFSETS[] =
@ -908,6 +918,50 @@ static void gfx_v9_0_get_csb_buffer(struct amdgpu_device *adev,
buffer[count++] = cpu_to_le32(0);
}
static void gfx_v9_0_init_always_on_cu_mask(struct amdgpu_device *adev)
{
struct amdgpu_cu_info *cu_info = &adev->gfx.cu_info;
uint32_t pg_always_on_cu_num = 2;
uint32_t always_on_cu_num;
uint32_t i, j, k;
uint32_t mask, cu_bitmap, counter;
if (adev->flags & AMD_IS_APU)
always_on_cu_num = 4;
else if (adev->asic_type == CHIP_VEGA12)
always_on_cu_num = 8;
else
always_on_cu_num = 12;
mutex_lock(&adev->grbm_idx_mutex);
for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
mask = 1;
cu_bitmap = 0;
counter = 0;
gfx_v9_0_select_se_sh(adev, i, j, 0xffffffff);
for (k = 0; k < adev->gfx.config.max_cu_per_sh; k ++) {
if (cu_info->bitmap[i][j] & mask) {
if (counter == pg_always_on_cu_num)
WREG32_SOC15(GC, 0, mmRLC_PG_ALWAYS_ON_CU_MASK, cu_bitmap);
if (counter < always_on_cu_num)
cu_bitmap |= mask;
else
break;
counter++;
}
mask <<= 1;
}
WREG32_SOC15(GC, 0, mmRLC_LB_ALWAYS_ACTIVE_CU_MASK, cu_bitmap);
cu_info->ao_cu_bitmap[i][j] = cu_bitmap;
}
}
gfx_v9_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
mutex_unlock(&adev->grbm_idx_mutex);
}
static void gfx_v9_0_init_lbpw(struct amdgpu_device *adev)
{
uint32_t data;
@ -941,8 +995,10 @@ static void gfx_v9_0_init_lbpw(struct amdgpu_device *adev)
data |= 0x00C00000;
WREG32_SOC15(GC, 0, mmRLC_GPM_GENERAL_7, data);
/* set RLC_LB_ALWAYS_ACTIVE_CU_MASK = 0xFFF */
WREG32_SOC15(GC, 0, mmRLC_LB_ALWAYS_ACTIVE_CU_MASK, 0xFFF);
/*
* RLC_LB_ALWAYS_ACTIVE_CU_MASK = 0xF (4 CUs AON for Raven),
* programmed in gfx_v9_0_init_always_on_cu_mask()
*/
/* set RLC_LB_CNTL = 0x8000_0095, 31 bit is reserved,
* but used for RLC_LB_CNTL configuration */
@ -951,6 +1007,57 @@ static void gfx_v9_0_init_lbpw(struct amdgpu_device *adev)
data |= REG_SET_FIELD(data, RLC_LB_CNTL, RESERVED, 0x80000);
WREG32_SOC15(GC, 0, mmRLC_LB_CNTL, data);
mutex_unlock(&adev->grbm_idx_mutex);
gfx_v9_0_init_always_on_cu_mask(adev);
}
static void gfx_v9_4_init_lbpw(struct amdgpu_device *adev)
{
uint32_t data;
/* set mmRLC_LB_THR_CONFIG_1/2/3/4 */
WREG32_SOC15(GC, 0, mmRLC_LB_THR_CONFIG_1, 0x0000007F);
WREG32_SOC15(GC, 0, mmRLC_LB_THR_CONFIG_2, 0x033388F8);
WREG32_SOC15(GC, 0, mmRLC_LB_THR_CONFIG_3, 0x00000077);
WREG32_SOC15(GC, 0, mmRLC_LB_THR_CONFIG_4, (0x10 | 0x27 << 8 | 0x02FA << 16));
/* set mmRLC_LB_CNTR_INIT = 0x0000_0000 */
WREG32_SOC15(GC, 0, mmRLC_LB_CNTR_INIT, 0x00000000);
/* set mmRLC_LB_CNTR_MAX = 0x0000_0500 */
WREG32_SOC15(GC, 0, mmRLC_LB_CNTR_MAX, 0x00000800);
mutex_lock(&adev->grbm_idx_mutex);
/* set mmRLC_LB_INIT_CU_MASK thru broadcast mode to enable all SE/SH*/
gfx_v9_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
WREG32_SOC15(GC, 0, mmRLC_LB_INIT_CU_MASK, 0xffffffff);
/* set mmRLC_LB_PARAMS = 0x003F_1006 */
data = REG_SET_FIELD(0, RLC_LB_PARAMS, FIFO_SAMPLES, 0x0003);
data |= REG_SET_FIELD(data, RLC_LB_PARAMS, PG_IDLE_SAMPLES, 0x0010);
data |= REG_SET_FIELD(data, RLC_LB_PARAMS, PG_IDLE_SAMPLE_INTERVAL, 0x033F);
WREG32_SOC15(GC, 0, mmRLC_LB_PARAMS, data);
/* set mmRLC_GPM_GENERAL_7[31-16] = 0x00C0 */
data = RREG32_SOC15(GC, 0, mmRLC_GPM_GENERAL_7);
data &= 0x0000FFFF;
data |= 0x00C00000;
WREG32_SOC15(GC, 0, mmRLC_GPM_GENERAL_7, data);
/*
* RLC_LB_ALWAYS_ACTIVE_CU_MASK = 0xFFF (12 CUs AON),
* programmed in gfx_v9_0_init_always_on_cu_mask()
*/
/* set RLC_LB_CNTL = 0x8000_0095, 31 bit is reserved,
* but used for RLC_LB_CNTL configuration */
data = RLC_LB_CNTL__LB_CNT_SPIM_ACTIVE_MASK;
data |= REG_SET_FIELD(data, RLC_LB_CNTL, CU_MASK_USED_OFF_HYST, 0x09);
data |= REG_SET_FIELD(data, RLC_LB_CNTL, RESERVED, 0x80000);
WREG32_SOC15(GC, 0, mmRLC_LB_CNTL, data);
mutex_unlock(&adev->grbm_idx_mutex);
gfx_v9_0_init_always_on_cu_mask(adev);
}
static void gfx_v9_0_enable_lbpw(struct amdgpu_device *adev, bool enable)
@ -1084,8 +1191,17 @@ static int gfx_v9_0_rlc_init(struct amdgpu_device *adev)
rv_init_cp_jump_table(adev);
amdgpu_bo_kunmap(adev->gfx.rlc.cp_table_obj);
amdgpu_bo_unreserve(adev->gfx.rlc.cp_table_obj);
}
switch (adev->asic_type) {
case CHIP_RAVEN:
gfx_v9_0_init_lbpw(adev);
break;
case CHIP_VEGA20:
gfx_v9_4_init_lbpw(adev);
break;
default:
break;
}
return 0;
@ -1605,11 +1721,6 @@ static int gfx_v9_0_sw_init(void *handle)
adev->gfx.mec.num_pipe_per_mec = 4;
adev->gfx.mec.num_queue_per_pipe = 8;
/* KIQ event */
r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_GRBM_CP, GFX_9_0__SRCID__CP_IB2_INTERRUPT_PKT, &adev->gfx.kiq.irq);
if (r)
return r;
/* EOP Event */
r = amdgpu_irq_add_id(adev, SOC15_IH_CLIENTID_GRBM_CP, GFX_9_0__SRCID__CP_EOP_INTERRUPT, &adev->gfx.eop_irq);
if (r)
@ -1847,7 +1958,7 @@ static void gfx_v9_0_init_compute_vmid(struct amdgpu_device *adev)
mutex_unlock(&adev->srbm_mutex);
}
static void gfx_v9_0_gpu_init(struct amdgpu_device *adev)
static void gfx_v9_0_constants_init(struct amdgpu_device *adev)
{
u32 tmp;
int i;
@ -2403,7 +2514,8 @@ static int gfx_v9_0_rlc_resume(struct amdgpu_device *adev)
return r;
}
if (adev->asic_type == CHIP_RAVEN) {
if (adev->asic_type == CHIP_RAVEN ||
adev->asic_type == CHIP_VEGA20) {
if (amdgpu_lbpw != 0)
gfx_v9_0_enable_lbpw(adev, true);
else
@ -3091,7 +3203,7 @@ static int gfx_v9_0_kcq_init_queue(struct amdgpu_ring *ring)
struct v9_mqd *mqd = ring->mqd_ptr;
int mqd_idx = ring - &adev->gfx.compute_ring[0];
if (!adev->in_gpu_reset && !adev->gfx.in_suspend) {
if (!adev->in_gpu_reset && !adev->in_suspend) {
memset((void *)mqd, 0, sizeof(struct v9_mqd_allocation));
((struct v9_mqd_allocation *)mqd)->dynamic_cu_mask = 0xFFFFFFFF;
((struct v9_mqd_allocation *)mqd)->dynamic_rb_mask = 0xFFFFFFFF;
@ -3235,7 +3347,7 @@ static int gfx_v9_0_hw_init(void *handle)
gfx_v9_0_init_golden_registers(adev);
gfx_v9_0_gpu_init(adev);
gfx_v9_0_constants_init(adev);
r = gfx_v9_0_csb_vram_pin(adev);
if (r)
@ -3310,7 +3422,7 @@ static int gfx_v9_0_hw_fini(void *handle)
/* Use deinitialize sequence from CAIL when unbinding device from driver,
* otherwise KIQ is hanging when binding back
*/
if (!adev->in_gpu_reset && !adev->gfx.in_suspend) {
if (!adev->in_gpu_reset && !adev->in_suspend) {
mutex_lock(&adev->srbm_mutex);
soc15_grbm_select(adev, adev->gfx.kiq.ring.me,
adev->gfx.kiq.ring.pipe,
@ -3330,20 +3442,12 @@ static int gfx_v9_0_hw_fini(void *handle)
static int gfx_v9_0_suspend(void *handle)
{
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
adev->gfx.in_suspend = true;
return gfx_v9_0_hw_fini(adev);
return gfx_v9_0_hw_fini(handle);
}
static int gfx_v9_0_resume(void *handle)
{
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
int r;
r = gfx_v9_0_hw_init(adev);
adev->gfx.in_suspend = false;
return r;
return gfx_v9_0_hw_init(handle);
}
static bool gfx_v9_0_is_idle(void *handle)
@ -4609,68 +4713,6 @@ static int gfx_v9_0_priv_inst_irq(struct amdgpu_device *adev,
return 0;
}
static int gfx_v9_0_kiq_set_interrupt_state(struct amdgpu_device *adev,
struct amdgpu_irq_src *src,
unsigned int type,
enum amdgpu_interrupt_state state)
{
uint32_t tmp, target;
struct amdgpu_ring *ring = &(adev->gfx.kiq.ring);
if (ring->me == 1)
target = SOC15_REG_OFFSET(GC, 0, mmCP_ME1_PIPE0_INT_CNTL);
else
target = SOC15_REG_OFFSET(GC, 0, mmCP_ME2_PIPE0_INT_CNTL);
target += ring->pipe;
switch (type) {
case AMDGPU_CP_KIQ_IRQ_DRIVER0:
if (state == AMDGPU_IRQ_STATE_DISABLE) {
tmp = RREG32_SOC15(GC, 0, mmCPC_INT_CNTL);
tmp = REG_SET_FIELD(tmp, CPC_INT_CNTL,
GENERIC2_INT_ENABLE, 0);
WREG32_SOC15(GC, 0, mmCPC_INT_CNTL, tmp);
tmp = RREG32(target);
tmp = REG_SET_FIELD(tmp, CP_ME2_PIPE0_INT_CNTL,
GENERIC2_INT_ENABLE, 0);
WREG32(target, tmp);
} else {
tmp = RREG32_SOC15(GC, 0, mmCPC_INT_CNTL);
tmp = REG_SET_FIELD(tmp, CPC_INT_CNTL,
GENERIC2_INT_ENABLE, 1);
WREG32_SOC15(GC, 0, mmCPC_INT_CNTL, tmp);
tmp = RREG32(target);
tmp = REG_SET_FIELD(tmp, CP_ME2_PIPE0_INT_CNTL,
GENERIC2_INT_ENABLE, 1);
WREG32(target, tmp);
}
break;
default:
BUG(); /* kiq only support GENERIC2_INT now */
break;
}
return 0;
}
static int gfx_v9_0_kiq_irq(struct amdgpu_device *adev,
struct amdgpu_irq_src *source,
struct amdgpu_iv_entry *entry)
{
u8 me_id, pipe_id, queue_id;
struct amdgpu_ring *ring = &(adev->gfx.kiq.ring);
me_id = (entry->ring_id & 0x0c) >> 2;
pipe_id = (entry->ring_id & 0x03) >> 0;
queue_id = (entry->ring_id & 0x70) >> 4;
DRM_DEBUG("IH: CPC GENERIC2_INT, me:%d, pipe:%d, queue:%d\n",
me_id, pipe_id, queue_id);
amdgpu_fence_process(ring);
return 0;
}
static const struct amd_ip_funcs gfx_v9_0_ip_funcs = {
.name = "gfx_v9_0",
.early_init = gfx_v9_0_early_init,
@ -4819,11 +4861,6 @@ static void gfx_v9_0_set_ring_funcs(struct amdgpu_device *adev)
adev->gfx.compute_ring[i].funcs = &gfx_v9_0_ring_funcs_compute;
}
static const struct amdgpu_irq_src_funcs gfx_v9_0_kiq_irq_funcs = {
.set = gfx_v9_0_kiq_set_interrupt_state,
.process = gfx_v9_0_kiq_irq,
};
static const struct amdgpu_irq_src_funcs gfx_v9_0_eop_irq_funcs = {
.set = gfx_v9_0_set_eop_interrupt_state,
.process = gfx_v9_0_eop_irq,
@ -4849,9 +4886,6 @@ static void gfx_v9_0_set_irq_funcs(struct amdgpu_device *adev)
adev->gfx.priv_inst_irq.num_types = 1;
adev->gfx.priv_inst_irq.funcs = &gfx_v9_0_priv_inst_irq_funcs;
adev->gfx.kiq.irq.num_types = AMDGPU_CP_KIQ_IRQ_LAST;
adev->gfx.kiq.irq.funcs = &gfx_v9_0_kiq_irq_funcs;
}
static void gfx_v9_0_set_rlc_funcs(struct amdgpu_device *adev)
@ -4871,7 +4905,20 @@ static void gfx_v9_0_set_rlc_funcs(struct amdgpu_device *adev)
static void gfx_v9_0_set_gds_init(struct amdgpu_device *adev)
{
/* init asci gds info */
adev->gds.mem.total_size = RREG32_SOC15(GC, 0, mmGDS_VMID0_SIZE);
switch (adev->asic_type) {
case CHIP_VEGA10:
case CHIP_VEGA12:
case CHIP_VEGA20:
adev->gds.mem.total_size = 0x10000;
break;
case CHIP_RAVEN:
adev->gds.mem.total_size = 0x1000;
break;
default:
adev->gds.mem.total_size = 0x10000;
break;
}
adev->gds.gws.total_size = 64;
adev->gds.oa.total_size = 16;

3
drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c
View File

@ -82,7 +82,8 @@ static void gfxhub_v1_0_init_system_aperture_regs(struct amdgpu_device *adev)
* to get rid of the VM fault and hardware hang.
*/
WREG32_SOC15(GC, 0, mmMC_VM_SYSTEM_APERTURE_HIGH_ADDR,
(max(adev->gmc.vram_end, adev->gmc.agp_end) >> 18) + 0x1);
max((adev->gmc.vram_end >> 18) + 0x1,
adev->gmc.agp_end >> 18));
else
WREG32_SOC15(GC, 0, mmMC_VM_SYSTEM_APERTURE_HIGH_ADDR,
max(adev->gmc.vram_end, adev->gmc.agp_end) >> 18);

7
drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c
View File

@ -859,11 +859,11 @@ static int gmc_v6_0_sw_init(void *handle)
adev->gmc.vram_type = gmc_v6_0_convert_vram_type(tmp);
}
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 146, &adev->gmc.vm_fault);
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 146, &adev->gmc.vm_fault);
if (r)
return r;
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 147, &adev->gmc.vm_fault);
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 147, &adev->gmc.vm_fault);
if (r)
return r;
@ -1180,8 +1180,7 @@ static const struct amdgpu_irq_src_funcs gmc_v6_0_irq_funcs = {
static void gmc_v6_0_set_gmc_funcs(struct amdgpu_device *adev)
{
if (adev->gmc.gmc_funcs == NULL)
adev->gmc.gmc_funcs = &gmc_v6_0_gmc_funcs;
adev->gmc.gmc_funcs = &gmc_v6_0_gmc_funcs;
}
static void gmc_v6_0_set_irq_funcs(struct amdgpu_device *adev)

7
drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c
View File

@ -991,11 +991,11 @@ static int gmc_v7_0_sw_init(void *handle)
adev->gmc.vram_type = gmc_v7_0_convert_vram_type(tmp);
}
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_GFX_PAGE_INV_FAULT, &adev->gmc.vm_fault);
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_GFX_PAGE_INV_FAULT, &adev->gmc.vm_fault);
if (r)
return r;
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_GFX_MEM_PROT_FAULT, &adev->gmc.vm_fault);
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_GFX_MEM_PROT_FAULT, &adev->gmc.vm_fault);
if (r)
return r;
@ -1388,8 +1388,7 @@ static const struct amdgpu_irq_src_funcs gmc_v7_0_irq_funcs = {
static void gmc_v7_0_set_gmc_funcs(struct amdgpu_device *adev)
{
if (adev->gmc.gmc_funcs == NULL)
adev->gmc.gmc_funcs = &gmc_v7_0_gmc_funcs;
adev->gmc.gmc_funcs = &gmc_v7_0_gmc_funcs;
}
static void gmc_v7_0_set_irq_funcs(struct amdgpu_device *adev)

7
drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c
View File

@ -1095,11 +1095,11 @@ static int gmc_v8_0_sw_init(void *handle)
adev->gmc.vram_type = gmc_v8_0_convert_vram_type(tmp);
}
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_GFX_PAGE_INV_FAULT, &adev->gmc.vm_fault);
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_GFX_PAGE_INV_FAULT, &adev->gmc.vm_fault);
if (r)
return r;
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_GFX_MEM_PROT_FAULT, &adev->gmc.vm_fault);
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_GFX_MEM_PROT_FAULT, &adev->gmc.vm_fault);
if (r)
return r;
@ -1733,8 +1733,7 @@ static const struct amdgpu_irq_src_funcs gmc_v8_0_irq_funcs = {
static void gmc_v8_0_set_gmc_funcs(struct amdgpu_device *adev)
{
if (adev->gmc.gmc_funcs == NULL)
adev->gmc.gmc_funcs = &gmc_v8_0_gmc_funcs;
adev->gmc.gmc_funcs = &gmc_v8_0_gmc_funcs;
}
static void gmc_v8_0_set_irq_funcs(struct amdgpu_device *adev)

3
drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
View File

@ -593,8 +593,7 @@ static const struct amdgpu_gmc_funcs gmc_v9_0_gmc_funcs = {
static void gmc_v9_0_set_gmc_funcs(struct amdgpu_device *adev)
{
if (adev->gmc.gmc_funcs == NULL)
adev->gmc.gmc_funcs = &gmc_v9_0_gmc_funcs;
adev->gmc.gmc_funcs = &gmc_v9_0_gmc_funcs;
}
static int gmc_v9_0_early_init(void *handle)

9
drivers/gpu/drm/amd/amdgpu/iceland_ih.c
View File

@ -255,7 +255,7 @@ static void iceland_ih_decode_iv(struct amdgpu_device *adev,
dw[2] = le32_to_cpu(adev->irq.ih.ring[ring_index + 2]);
dw[3] = le32_to_cpu(adev->irq.ih.ring[ring_index + 3]);
entry->client_id = AMDGPU_IH_CLIENTID_LEGACY;
entry->client_id = AMDGPU_IRQ_CLIENTID_LEGACY;
entry->src_id = dw[0] & 0xff;
entry->src_data[0] = dw[1] & 0xfffffff;
entry->ring_id = dw[2] & 0xff;
@ -297,7 +297,7 @@ static int iceland_ih_sw_init(void *handle)
int r;
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
r = amdgpu_ih_ring_init(adev, 64 * 1024, false);
r = amdgpu_ih_ring_init(adev, &adev->irq.ih, 64 * 1024, false);
if (r)
return r;
@ -311,7 +311,7 @@ static int iceland_ih_sw_fini(void *handle)
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
amdgpu_irq_fini(adev);
amdgpu_ih_ring_fini(adev);
amdgpu_ih_ring_fini(adev, &adev->irq.ih);
amdgpu_irq_remove_domain(adev);
return 0;
@ -447,8 +447,7 @@ static const struct amdgpu_ih_funcs iceland_ih_funcs = {
static void iceland_ih_set_interrupt_funcs(struct amdgpu_device *adev)
{
if (adev->irq.ih_funcs == NULL)
adev->irq.ih_funcs = &iceland_ih_funcs;
adev->irq.ih_funcs = &iceland_ih_funcs;
}
const struct amdgpu_ip_block_version iceland_ih_ip_block =

4
drivers/gpu/drm/amd/amdgpu/kv_dpm.c
View File

@ -2995,12 +2995,12 @@ static int kv_dpm_sw_init(void *handle)
int ret;
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
ret = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 230,
ret = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 230,
&adev->pm.dpm.thermal.irq);
if (ret)
return ret;
ret = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 231,
ret = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 231,
&adev->pm.dpm.thermal.irq);
if (ret)
return ret;

3
drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c
View File

@ -100,7 +100,8 @@ static void mmhub_v1_0_init_system_aperture_regs(struct amdgpu_device *adev)
* to get rid of the VM fault and hardware hang.
*/
WREG32_SOC15(MMHUB, 0, mmMC_VM_SYSTEM_APERTURE_HIGH_ADDR,
(max(adev->gmc.vram_end, adev->gmc.agp_end) >> 18) + 0x1);
max((adev->gmc.vram_end >> 18) + 0x1,
adev->gmc.agp_end >> 18));
else
WREG32_SOC15(MMHUB, 0, mmMC_VM_SYSTEM_APERTURE_HIGH_ADDR,
max(adev->gmc.vram_end, adev->gmc.agp_end) >> 18);

4
drivers/gpu/drm/amd/amdgpu/mxgpu_vi.c
View File

@ -580,11 +580,11 @@ int xgpu_vi_mailbox_add_irq_id(struct amdgpu_device *adev)
{
int r;
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 135, &adev->virt.rcv_irq);
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 135, &adev->virt.rcv_irq);
if (r)
return r;
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 138, &adev->virt.ack_irq);
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 138, &adev->virt.ack_irq);
if (r) {
amdgpu_irq_put(adev, &adev->virt.rcv_irq, 0);
return r;

3
drivers/gpu/drm/amd/amdgpu/psp_v11_0.c
View File

@ -306,11 +306,8 @@ static int psp_v11_0_ring_stop(struct psp_context *psp,
enum psp_ring_type ring_type)
{
int ret = 0;
struct psp_ring *ring;
struct amdgpu_device *adev = psp->adev;
ring = &psp->km_ring;
/* Write the ring destroy command to C2PMSG_64 */
WREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_64, GFX_CTRL_CMD_ID_DESTROY_RINGS);

68
drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c
View File

@ -504,41 +504,6 @@ static int sdma_v2_4_rlc_resume(struct amdgpu_device *adev)
return 0;
}
/**
* sdma_v2_4_load_microcode - load the sDMA ME ucode
*
* @adev: amdgpu_device pointer
*
* Loads the sDMA0/1 ucode.
* Returns 0 for success, -EINVAL if the ucode is not available.
*/
static int sdma_v2_4_load_microcode(struct amdgpu_device *adev)
{
const struct sdma_firmware_header_v1_0 *hdr;
const __le32 *fw_data;
u32 fw_size;
int i, j;
/* halt the MEs */
sdma_v2_4_enable(adev, false);
for (i = 0; i < adev->sdma.num_instances; i++) {
if (!adev->sdma.instance[i].fw)
return -EINVAL;
hdr = (const struct sdma_firmware_header_v1_0 *)adev->sdma.instance[i].fw->data;
amdgpu_ucode_print_sdma_hdr(&hdr->header);
fw_size = le32_to_cpu(hdr->header.ucode_size_bytes) / 4;
fw_data = (const __le32 *)
(adev->sdma.instance[i].fw->data +
le32_to_cpu(hdr->header.ucode_array_offset_bytes));
WREG32(mmSDMA0_UCODE_ADDR + sdma_offsets[i], 0);
for (j = 0; j < fw_size; j++)
WREG32(mmSDMA0_UCODE_DATA + sdma_offsets[i], le32_to_cpup(fw_data++));
WREG32(mmSDMA0_UCODE_ADDR + sdma_offsets[i], adev->sdma.instance[i].fw_version);
}
return 0;
}
/**
* sdma_v2_4_start - setup and start the async dma engines
@ -552,13 +517,6 @@ static int sdma_v2_4_start(struct amdgpu_device *adev)
{
int r;
if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) {
r = sdma_v2_4_load_microcode(adev);
if (r)
return r;
}
/* halt the engine before programing */
sdma_v2_4_enable(adev, false);
@ -898,19 +856,19 @@ static int sdma_v2_4_sw_init(void *handle)
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
/* SDMA trap event */
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_SDMA_TRAP,
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_SDMA_TRAP,
&adev->sdma.trap_irq);
if (r)
return r;
/* SDMA Privileged inst */
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 241,
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 241,
&adev->sdma.illegal_inst_irq);
if (r)
return r;
/* SDMA Privileged inst */
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_SDMA_SRBM_WRITE,
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_SDMA_SRBM_WRITE,
&adev->sdma.illegal_inst_irq);
if (r)
return r;
@ -1296,10 +1254,8 @@ static const struct amdgpu_buffer_funcs sdma_v2_4_buffer_funcs = {
static void sdma_v2_4_set_buffer_funcs(struct amdgpu_device *adev)
{
if (adev->mman.buffer_funcs == NULL) {
adev->mman.buffer_funcs = &sdma_v2_4_buffer_funcs;
adev->mman.buffer_funcs_ring = &adev->sdma.instance[0].ring;
}
adev->mman.buffer_funcs = &sdma_v2_4_buffer_funcs;
adev->mman.buffer_funcs_ring = &adev->sdma.instance[0].ring;
}
static const struct amdgpu_vm_pte_funcs sdma_v2_4_vm_pte_funcs = {
@ -1315,15 +1271,13 @@ static void sdma_v2_4_set_vm_pte_funcs(struct amdgpu_device *adev)
struct drm_gpu_scheduler *sched;
unsigned i;
if (adev->vm_manager.vm_pte_funcs == NULL) {
adev->vm_manager.vm_pte_funcs = &sdma_v2_4_vm_pte_funcs;
for (i = 0; i < adev->sdma.num_instances; i++) {
sched = &adev->sdma.instance[i].ring.sched;
adev->vm_manager.vm_pte_rqs[i] =
&sched->sched_rq[DRM_SCHED_PRIORITY_KERNEL];
}
adev->vm_manager.vm_pte_num_rqs = adev->sdma.num_instances;
adev->vm_manager.vm_pte_funcs = &sdma_v2_4_vm_pte_funcs;
for (i = 0; i < adev->sdma.num_instances; i++) {
sched = &adev->sdma.instance[i].ring.sched;
adev->vm_manager.vm_pte_rqs[i] =
&sched->sched_rq[DRM_SCHED_PRIORITY_KERNEL];
}
adev->vm_manager.vm_pte_num_rqs = adev->sdma.num_instances;
}
const struct amdgpu_ip_block_version sdma_v2_4_ip_block =

83
drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c
View File

@ -318,14 +318,13 @@ static int sdma_v3_0_init_microcode(struct amdgpu_device *adev)
if (adev->sdma.instance[i].feature_version >= 20)
adev->sdma.instance[i].burst_nop = true;
if (adev->firmware.load_type == AMDGPU_FW_LOAD_SMU) {
info = &adev->firmware.ucode[AMDGPU_UCODE_ID_SDMA0 + i];
info->ucode_id = AMDGPU_UCODE_ID_SDMA0 + i;
info->fw = adev->sdma.instance[i].fw;
header = (const struct common_firmware_header *)info->fw->data;
adev->firmware.fw_size +=
ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE);
}
info = &adev->firmware.ucode[AMDGPU_UCODE_ID_SDMA0 + i];
info->ucode_id = AMDGPU_UCODE_ID_SDMA0 + i;
info->fw = adev->sdma.instance[i].fw;
header = (const struct common_firmware_header *)info->fw->data;
adev->firmware.fw_size +=
ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE);
}
out:
if (err) {
@ -777,42 +776,6 @@ static int sdma_v3_0_rlc_resume(struct amdgpu_device *adev)
return 0;
}
/**
* sdma_v3_0_load_microcode - load the sDMA ME ucode
*
* @adev: amdgpu_device pointer
*
* Loads the sDMA0/1 ucode.
* Returns 0 for success, -EINVAL if the ucode is not available.
*/
static int sdma_v3_0_load_microcode(struct amdgpu_device *adev)
{
const struct sdma_firmware_header_v1_0 *hdr;
const __le32 *fw_data;
u32 fw_size;
int i, j;
/* halt the MEs */
sdma_v3_0_enable(adev, false);
for (i = 0; i < adev->sdma.num_instances; i++) {
if (!adev->sdma.instance[i].fw)
return -EINVAL;
hdr = (const struct sdma_firmware_header_v1_0 *)adev->sdma.instance[i].fw->data;
amdgpu_ucode_print_sdma_hdr(&hdr->header);
fw_size = le32_to_cpu(hdr->header.ucode_size_bytes) / 4;
fw_data = (const __le32 *)
(adev->sdma.instance[i].fw->data +
le32_to_cpu(hdr->header.ucode_array_offset_bytes));
WREG32(mmSDMA0_UCODE_ADDR + sdma_offsets[i], 0);
for (j = 0; j < fw_size; j++)
WREG32(mmSDMA0_UCODE_DATA + sdma_offsets[i], le32_to_cpup(fw_data++));
WREG32(mmSDMA0_UCODE_ADDR + sdma_offsets[i], adev->sdma.instance[i].fw_version);
}
return 0;
}
/**
* sdma_v3_0_start - setup and start the async dma engines
*
@ -825,12 +788,6 @@ static int sdma_v3_0_start(struct amdgpu_device *adev)
{
int r;
if (adev->firmware.load_type == AMDGPU_FW_LOAD_DIRECT) {
r = sdma_v3_0_load_microcode(adev);
if (r)
return r;
}
/* disable sdma engine before programing it */
sdma_v3_0_ctx_switch_enable(adev, false);
sdma_v3_0_enable(adev, false);
@ -1177,19 +1134,19 @@ static int sdma_v3_0_sw_init(void *handle)
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
/* SDMA trap event */
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_SDMA_TRAP,
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_SDMA_TRAP,
&adev->sdma.trap_irq);
if (r)
return r;
/* SDMA Privileged inst */
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 241,
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 241,
&adev->sdma.illegal_inst_irq);
if (r)
return r;
/* SDMA Privileged inst */
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_SDMA_SRBM_WRITE,
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_SDMA_SRBM_WRITE,
&adev->sdma.illegal_inst_irq);
if (r)
return r;
@ -1736,10 +1693,8 @@ static const struct amdgpu_buffer_funcs sdma_v3_0_buffer_funcs = {
static void sdma_v3_0_set_buffer_funcs(struct amdgpu_device *adev)
{
if (adev->mman.buffer_funcs == NULL) {
adev->mman.buffer_funcs = &sdma_v3_0_buffer_funcs;
adev->mman.buffer_funcs_ring = &adev->sdma.instance[0].ring;
}
adev->mman.buffer_funcs = &sdma_v3_0_buffer_funcs;
adev->mman.buffer_funcs_ring = &adev->sdma.instance[0].ring;
}
static const struct amdgpu_vm_pte_funcs sdma_v3_0_vm_pte_funcs = {
@ -1755,15 +1710,13 @@ static void sdma_v3_0_set_vm_pte_funcs(struct amdgpu_device *adev)
struct drm_gpu_scheduler *sched;
unsigned i;
if (adev->vm_manager.vm_pte_funcs == NULL) {
adev->vm_manager.vm_pte_funcs = &sdma_v3_0_vm_pte_funcs;
for (i = 0; i < adev->sdma.num_instances; i++) {
sched = &adev->sdma.instance[i].ring.sched;
adev->vm_manager.vm_pte_rqs[i] =
&sched->sched_rq[DRM_SCHED_PRIORITY_KERNEL];
}
adev->vm_manager.vm_pte_num_rqs = adev->sdma.num_instances;
adev->vm_manager.vm_pte_funcs = &sdma_v3_0_vm_pte_funcs;
for (i = 0; i < adev->sdma.num_instances; i++) {
sched = &adev->sdma.instance[i].ring.sched;
adev->vm_manager.vm_pte_rqs[i] =
&sched->sched_rq[DRM_SCHED_PRIORITY_KERNEL];
}
adev->vm_manager.vm_pte_num_rqs = adev->sdma.num_instances;
}
const struct amdgpu_ip_block_version sdma_v3_0_ip_block =

42
drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c
View File

@ -148,6 +148,7 @@ static const struct soc15_reg_golden golden_settings_sdma0_4_2[] =
SOC15_REG_GOLDEN_VALUE(SDMA0, 0, mmSDMA0_RLC7_RB_RPTR_ADDR_LO, 0xfffffffd, 0x00000001),
SOC15_REG_GOLDEN_VALUE(SDMA0, 0, mmSDMA0_RLC7_RB_WPTR_POLL_CNTL, 0xfffffff7, 0x00403000),
SOC15_REG_GOLDEN_VALUE(SDMA0, 0, mmSDMA0_UTCL1_PAGE, 0x000003ff, 0x000003c0),
SOC15_REG_GOLDEN_VALUE(SDMA0, 0, mmSDMA0_UTCL1_WATERMK, 0xFE000000, 0x00000000),
};
static const struct soc15_reg_golden golden_settings_sdma1_4_2[] = {
@ -177,6 +178,7 @@ static const struct soc15_reg_golden golden_settings_sdma1_4_2[] = {
SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_RLC7_RB_RPTR_ADDR_LO, 0xfffffffd, 0x00000001),
SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_RLC7_RB_WPTR_POLL_CNTL, 0xfffffff7, 0x00403000),
SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_UTCL1_PAGE, 0x000003ff, 0x000003c0),
SOC15_REG_GOLDEN_VALUE(SDMA1, 0, mmSDMA1_UTCL1_WATERMK, 0xFE000000, 0x00000000),
};
static const struct soc15_reg_golden golden_settings_sdma_rv1[] =
@ -818,7 +820,7 @@ sdma_v4_1_update_power_gating(struct amdgpu_device *adev, bool enable)
uint32_t def, data;
if (enable && (adev->pg_flags & AMD_PG_SUPPORT_SDMA)) {
/* disable idle interrupt */
/* enable idle interrupt */
def = data = RREG32(SOC15_REG_OFFSET(SDMA0, 0, mmSDMA0_CNTL));
data |= SDMA0_CNTL__CTXEMPTY_INT_ENABLE_MASK;
@ -1320,9 +1322,15 @@ static int sdma_v4_0_sw_init(void *handle)
DRM_INFO("use_doorbell being set to: [%s]\n",
ring->use_doorbell?"true":"false");
ring->doorbell_index = (i == 0) ?
(AMDGPU_DOORBELL64_sDMA_ENGINE0 << 1) //get DWORD offset
: (AMDGPU_DOORBELL64_sDMA_ENGINE1 << 1); // get DWORD offset
if (adev->asic_type == CHIP_VEGA10)
ring->doorbell_index = (i == 0) ?
(AMDGPU_VEGA10_DOORBELL64_sDMA_ENGINE0 << 1) //get DWORD offset
: (AMDGPU_VEGA10_DOORBELL64_sDMA_ENGINE1 << 1); // get DWORD offset
else
ring->doorbell_index = (i == 0) ?
(AMDGPU_DOORBELL64_sDMA_ENGINE0 << 1) //get DWORD offset
: (AMDGPU_DOORBELL64_sDMA_ENGINE1 << 1); // get DWORD offset
sprintf(ring->name, "sdma%d", i);
r = amdgpu_ring_init(adev, ring, 1024,
@ -1358,6 +1366,9 @@ static int sdma_v4_0_hw_init(void *handle)
int r;
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
if (adev->asic_type == CHIP_RAVEN && adev->powerplay.pp_funcs->set_powergating_by_smu)
amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_SDMA, false);
sdma_v4_0_init_golden_registers(adev);
r = sdma_v4_0_start(adev);
@ -1375,6 +1386,9 @@ static int sdma_v4_0_hw_fini(void *handle)
sdma_v4_0_ctx_switch_enable(adev, false);
sdma_v4_0_enable(adev, false);
if (adev->asic_type == CHIP_RAVEN && adev->powerplay.pp_funcs->set_powergating_by_smu)
amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_SDMA, true);
return 0;
}
@ -1801,10 +1815,8 @@ static const struct amdgpu_buffer_funcs sdma_v4_0_buffer_funcs = {
static void sdma_v4_0_set_buffer_funcs(struct amdgpu_device *adev)
{
if (adev->mman.buffer_funcs == NULL) {
adev->mman.buffer_funcs = &sdma_v4_0_buffer_funcs;
adev->mman.buffer_funcs_ring = &adev->sdma.instance[0].ring;
}
adev->mman.buffer_funcs = &sdma_v4_0_buffer_funcs;
adev->mman.buffer_funcs_ring = &adev->sdma.instance[0].ring;
}
static const struct amdgpu_vm_pte_funcs sdma_v4_0_vm_pte_funcs = {
@ -1820,15 +1832,13 @@ static void sdma_v4_0_set_vm_pte_funcs(struct amdgpu_device *adev)
struct drm_gpu_scheduler *sched;
unsigned i;
if (adev->vm_manager.vm_pte_funcs == NULL) {
adev->vm_manager.vm_pte_funcs = &sdma_v4_0_vm_pte_funcs;
for (i = 0; i < adev->sdma.num_instances; i++) {
sched = &adev->sdma.instance[i].ring.sched;
adev->vm_manager.vm_pte_rqs[i] =
&sched->sched_rq[DRM_SCHED_PRIORITY_KERNEL];
}
adev->vm_manager.vm_pte_num_rqs = adev->sdma.num_instances;
adev->vm_manager.vm_pte_funcs = &sdma_v4_0_vm_pte_funcs;
for (i = 0; i < adev->sdma.num_instances; i++) {
sched = &adev->sdma.instance[i].ring.sched;
adev->vm_manager.vm_pte_rqs[i] =
&sched->sched_rq[DRM_SCHED_PRIORITY_KERNEL];
}
adev->vm_manager.vm_pte_num_rqs = adev->sdma.num_instances;
}
const struct amdgpu_ip_block_version sdma_v4_0_ip_block = {

13
drivers/gpu/drm/amd/amdgpu/si.c
View File

@ -2057,13 +2057,13 @@ int si_set_ip_blocks(struct amdgpu_device *adev)
amdgpu_device_ip_block_add(adev, &si_common_ip_block);
amdgpu_device_ip_block_add(adev, &gmc_v6_0_ip_block);
amdgpu_device_ip_block_add(adev, &si_ih_ip_block);
amdgpu_device_ip_block_add(adev, &gfx_v6_0_ip_block);
amdgpu_device_ip_block_add(adev, &si_dma_ip_block);
amdgpu_device_ip_block_add(adev, &si_smu_ip_block);
if (adev->enable_virtual_display)
amdgpu_device_ip_block_add(adev, &dce_virtual_ip_block);
else
amdgpu_device_ip_block_add(adev, &dce_v6_0_ip_block);
amdgpu_device_ip_block_add(adev, &gfx_v6_0_ip_block);
amdgpu_device_ip_block_add(adev, &si_dma_ip_block);
/* amdgpu_device_ip_block_add(adev, &uvd_v3_1_ip_block); */
/* amdgpu_device_ip_block_add(adev, &vce_v1_0_ip_block); */
break;
@ -2071,13 +2071,14 @@ int si_set_ip_blocks(struct amdgpu_device *adev)
amdgpu_device_ip_block_add(adev, &si_common_ip_block);
amdgpu_device_ip_block_add(adev, &gmc_v6_0_ip_block);
amdgpu_device_ip_block_add(adev, &si_ih_ip_block);
amdgpu_device_ip_block_add(adev, &gfx_v6_0_ip_block);
amdgpu_device_ip_block_add(adev, &si_dma_ip_block);
amdgpu_device_ip_block_add(adev, &si_smu_ip_block);
if (adev->enable_virtual_display)
amdgpu_device_ip_block_add(adev, &dce_virtual_ip_block);
else
amdgpu_device_ip_block_add(adev, &dce_v6_4_ip_block);
amdgpu_device_ip_block_add(adev, &gfx_v6_0_ip_block);
amdgpu_device_ip_block_add(adev, &si_dma_ip_block);
/* amdgpu_device_ip_block_add(adev, &uvd_v3_1_ip_block); */
/* amdgpu_device_ip_block_add(adev, &vce_v1_0_ip_block); */
break;
@ -2085,11 +2086,11 @@ int si_set_ip_blocks(struct amdgpu_device *adev)
amdgpu_device_ip_block_add(adev, &si_common_ip_block);
amdgpu_device_ip_block_add(adev, &gmc_v6_0_ip_block);
amdgpu_device_ip_block_add(adev, &si_ih_ip_block);
amdgpu_device_ip_block_add(adev, &gfx_v6_0_ip_block);
amdgpu_device_ip_block_add(adev, &si_dma_ip_block);
amdgpu_device_ip_block_add(adev, &si_smu_ip_block);
if (adev->enable_virtual_display)
amdgpu_device_ip_block_add(adev, &dce_virtual_ip_block);
amdgpu_device_ip_block_add(adev, &gfx_v6_0_ip_block);
amdgpu_device_ip_block_add(adev, &si_dma_ip_block);
break;
default:
BUG();

47
drivers/gpu/drm/amd/amdgpu/si_dma.c
View File

@ -502,12 +502,14 @@ static int si_dma_sw_init(void *handle)
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
/* DMA0 trap event */
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 224, &adev->sdma.trap_irq);
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 224,
&adev->sdma.trap_irq);
if (r)
return r;
/* DMA1 trap event */
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 244, &adev->sdma.trap_irq_1);
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 244,
&adev->sdma.trap_irq);
if (r)
return r;
@ -649,17 +651,10 @@ static int si_dma_process_trap_irq(struct amdgpu_device *adev,
struct amdgpu_irq_src *source,
struct amdgpu_iv_entry *entry)
{
amdgpu_fence_process(&adev->sdma.instance[0].ring);
return 0;
}
static int si_dma_process_trap_irq_1(struct amdgpu_device *adev,
struct amdgpu_irq_src *source,
struct amdgpu_iv_entry *entry)
{
amdgpu_fence_process(&adev->sdma.instance[1].ring);
if (entry->src_id == 224)
amdgpu_fence_process(&adev->sdma.instance[0].ring);
else
amdgpu_fence_process(&adev->sdma.instance[1].ring);
return 0;
}
@ -786,11 +781,6 @@ static const struct amdgpu_irq_src_funcs si_dma_trap_irq_funcs = {
.process = si_dma_process_trap_irq,
};
static const struct amdgpu_irq_src_funcs si_dma_trap_irq_funcs_1 = {
.set = si_dma_set_trap_irq_state,
.process = si_dma_process_trap_irq_1,
};
static const struct amdgpu_irq_src_funcs si_dma_illegal_inst_irq_funcs = {
.process = si_dma_process_illegal_inst_irq,
};
@ -799,7 +789,6 @@ static void si_dma_set_irq_funcs(struct amdgpu_device *adev)
{
adev->sdma.trap_irq.num_types = AMDGPU_SDMA_IRQ_LAST;
adev->sdma.trap_irq.funcs = &si_dma_trap_irq_funcs;
adev->sdma.trap_irq_1.funcs = &si_dma_trap_irq_funcs_1;
adev->sdma.illegal_inst_irq.funcs = &si_dma_illegal_inst_irq_funcs;
}
@ -863,10 +852,8 @@ static const struct amdgpu_buffer_funcs si_dma_buffer_funcs = {
static void si_dma_set_buffer_funcs(struct amdgpu_device *adev)
{
if (adev->mman.buffer_funcs == NULL) {
adev->mman.buffer_funcs = &si_dma_buffer_funcs;
adev->mman.buffer_funcs_ring = &adev->sdma.instance[0].ring;
}
adev->mman.buffer_funcs = &si_dma_buffer_funcs;
adev->mman.buffer_funcs_ring = &adev->sdma.instance[0].ring;
}
static const struct amdgpu_vm_pte_funcs si_dma_vm_pte_funcs = {
@ -882,15 +869,13 @@ static void si_dma_set_vm_pte_funcs(struct amdgpu_device *adev)
struct drm_gpu_scheduler *sched;
unsigned i;
if (adev->vm_manager.vm_pte_funcs == NULL) {
adev->vm_manager.vm_pte_funcs = &si_dma_vm_pte_funcs;
for (i = 0; i < adev->sdma.num_instances; i++) {
sched = &adev->sdma.instance[i].ring.sched;
adev->vm_manager.vm_pte_rqs[i] =
&sched->sched_rq[DRM_SCHED_PRIORITY_KERNEL];
}
adev->vm_manager.vm_pte_num_rqs = adev->sdma.num_instances;
adev->vm_manager.vm_pte_funcs = &si_dma_vm_pte_funcs;
for (i = 0; i < adev->sdma.num_instances; i++) {
sched = &adev->sdma.instance[i].ring.sched;
adev->vm_manager.vm_pte_rqs[i] =
&sched->sched_rq[DRM_SCHED_PRIORITY_KERNEL];
}
adev->vm_manager.vm_pte_num_rqs = adev->sdma.num_instances;
}
const struct amdgpu_ip_block_version si_dma_ip_block =

4
drivers/gpu/drm/amd/amdgpu/si_dpm.c
View File

@ -7687,11 +7687,11 @@ static int si_dpm_sw_init(void *handle)
int ret;
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
ret = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 230, &adev->pm.dpm.thermal.irq);
ret = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 230, &adev->pm.dpm.thermal.irq);
if (ret)
return ret;
ret = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 231, &adev->pm.dpm.thermal.irq);
ret = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 231, &adev->pm.dpm.thermal.irq);
if (ret)
return ret;

9
drivers/gpu/drm/amd/amdgpu/si_ih.c
View File

@ -142,7 +142,7 @@ static void si_ih_decode_iv(struct amdgpu_device *adev,
dw[2] = le32_to_cpu(adev->irq.ih.ring[ring_index + 2]);
dw[3] = le32_to_cpu(adev->irq.ih.ring[ring_index + 3]);
entry->client_id = AMDGPU_IH_CLIENTID_LEGACY;
entry->client_id = AMDGPU_IRQ_CLIENTID_LEGACY;
entry->src_id = dw[0] & 0xff;
entry->src_data[0] = dw[1] & 0xfffffff;
entry->ring_id = dw[2] & 0xff;
@ -170,7 +170,7 @@ static int si_ih_sw_init(void *handle)
int r;
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
r = amdgpu_ih_ring_init(adev, 64 * 1024, false);
r = amdgpu_ih_ring_init(adev, &adev->irq.ih, 64 * 1024, false);
if (r)
return r;
@ -182,7 +182,7 @@ static int si_ih_sw_fini(void *handle)
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
amdgpu_irq_fini(adev);
amdgpu_ih_ring_fini(adev);
amdgpu_ih_ring_fini(adev, &adev->irq.ih);
return 0;
}
@ -308,8 +308,7 @@ static const struct amdgpu_ih_funcs si_ih_funcs = {
static void si_ih_set_interrupt_funcs(struct amdgpu_device *adev)
{
if (adev->irq.ih_funcs == NULL)
adev->irq.ih_funcs = &si_ih_funcs;
adev->irq.ih_funcs = &si_ih_funcs;
}
const struct amdgpu_ip_block_version si_ih_ip_block =

11
drivers/gpu/drm/amd/amdgpu/soc15.c
View File

@ -529,6 +529,8 @@ int soc15_set_ip_blocks(struct amdgpu_device *adev)
amdgpu_device_ip_block_add(adev, &psp_v11_0_ip_block);
else
amdgpu_device_ip_block_add(adev, &psp_v3_1_ip_block);
amdgpu_device_ip_block_add(adev, &gfx_v9_0_ip_block);
amdgpu_device_ip_block_add(adev, &sdma_v4_0_ip_block);
if (!amdgpu_sriov_vf(adev))
amdgpu_device_ip_block_add(adev, &pp_smu_ip_block);
if (adev->enable_virtual_display || amdgpu_sriov_vf(adev))
@ -539,8 +541,6 @@ int soc15_set_ip_blocks(struct amdgpu_device *adev)
#else
# warning "Enable CONFIG_DRM_AMD_DC for display support on SOC15."
#endif
amdgpu_device_ip_block_add(adev, &gfx_v9_0_ip_block);
amdgpu_device_ip_block_add(adev, &sdma_v4_0_ip_block);
if (!(adev->asic_type == CHIP_VEGA20 && amdgpu_sriov_vf(adev))) {
amdgpu_device_ip_block_add(adev, &uvd_v7_0_ip_block);
amdgpu_device_ip_block_add(adev, &vce_v4_0_ip_block);
@ -551,6 +551,8 @@ int soc15_set_ip_blocks(struct amdgpu_device *adev)
amdgpu_device_ip_block_add(adev, &gmc_v9_0_ip_block);
amdgpu_device_ip_block_add(adev, &vega10_ih_ip_block);
amdgpu_device_ip_block_add(adev, &psp_v10_0_ip_block);
amdgpu_device_ip_block_add(adev, &gfx_v9_0_ip_block);
amdgpu_device_ip_block_add(adev, &sdma_v4_0_ip_block);
amdgpu_device_ip_block_add(adev, &pp_smu_ip_block);
if (adev->enable_virtual_display || amdgpu_sriov_vf(adev))
amdgpu_device_ip_block_add(adev, &dce_virtual_ip_block);
@ -560,8 +562,6 @@ int soc15_set_ip_blocks(struct amdgpu_device *adev)
#else
# warning "Enable CONFIG_DRM_AMD_DC for display support on SOC15."
#endif
amdgpu_device_ip_block_add(adev, &gfx_v9_0_ip_block);
amdgpu_device_ip_block_add(adev, &sdma_v4_0_ip_block);
amdgpu_device_ip_block_add(adev, &vcn_v1_0_ip_block);
break;
default:
@ -739,7 +739,8 @@ static int soc15_common_early_init(void *handle)
adev->pg_flags = AMD_PG_SUPPORT_SDMA |
AMD_PG_SUPPORT_MMHUB |
AMD_PG_SUPPORT_VCN;
AMD_PG_SUPPORT_VCN |
AMD_PG_SUPPORT_VCN_DPG;
} else {
adev->cg_flags = AMD_CG_SUPPORT_GFX_MGCG |
AMD_CG_SUPPORT_GFX_MGLS |

22
drivers/gpu/drm/amd/amdgpu/soc15_common.h
View File

@ -57,13 +57,33 @@
loop--; \
if (!loop) { \
DRM_ERROR("Register(%d) [%s] failed to reach value 0x%08x != 0x%08x\n", \
inst, #reg, expected_value, (tmp_ & (mask))); \
inst, #reg, (unsigned)expected_value, (unsigned)(tmp_ & (mask))); \
ret = -ETIMEDOUT; \
break; \
} \
} \
} while (0)
#define RREG32_SOC15_DPG_MODE(ip, inst, reg, mask, sram_sel) \
({ WREG32_SOC15(ip, inst, mmUVD_DPG_LMA_MASK, mask); \
WREG32_SOC15(ip, inst, mmUVD_DPG_LMA_CTL, \
UVD_DPG_LMA_CTL__MASK_EN_MASK | \
((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg) \
<< UVD_DPG_LMA_CTL__READ_WRITE_ADDR__SHIFT) | \
(sram_sel << UVD_DPG_LMA_CTL__SRAM_SEL__SHIFT)); \
RREG32_SOC15(ip, inst, mmUVD_DPG_LMA_DATA); })
#define WREG32_SOC15_DPG_MODE(ip, inst, reg, value, mask, sram_sel) \
do { \
WREG32_SOC15(ip, inst, mmUVD_DPG_LMA_DATA, value); \
WREG32_SOC15(ip, inst, mmUVD_DPG_LMA_MASK, mask); \
WREG32_SOC15(ip, inst, mmUVD_DPG_LMA_CTL, \
UVD_DPG_LMA_CTL__READ_WRITE_MASK | \
((adev->reg_offset[ip##_HWIP][inst][reg##_BASE_IDX] + reg) \
<< UVD_DPG_LMA_CTL__READ_WRITE_ADDR__SHIFT) | \
(sram_sel << UVD_DPG_LMA_CTL__SRAM_SEL__SHIFT)); \
} while (0)
#endif

9
drivers/gpu/drm/amd/amdgpu/tonga_ih.c
View File

@ -266,7 +266,7 @@ static void tonga_ih_decode_iv(struct amdgpu_device *adev,
dw[2] = le32_to_cpu(adev->irq.ih.ring[ring_index + 2]);
dw[3] = le32_to_cpu(adev->irq.ih.ring[ring_index + 3]);
entry->client_id = AMDGPU_IH_CLIENTID_LEGACY;
entry->client_id = AMDGPU_IRQ_CLIENTID_LEGACY;
entry->src_id = dw[0] & 0xff;
entry->src_data[0] = dw[1] & 0xfffffff;
entry->ring_id = dw[2] & 0xff;
@ -317,7 +317,7 @@ static int tonga_ih_sw_init(void *handle)
int r;
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
r = amdgpu_ih_ring_init(adev, 64 * 1024, true);
r = amdgpu_ih_ring_init(adev, &adev->irq.ih, 64 * 1024, true);
if (r)
return r;
@ -334,7 +334,7 @@ static int tonga_ih_sw_fini(void *handle)
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
amdgpu_irq_fini(adev);
amdgpu_ih_ring_fini(adev);
amdgpu_ih_ring_fini(adev, &adev->irq.ih);
amdgpu_irq_remove_domain(adev);
return 0;
@ -513,8 +513,7 @@ static const struct amdgpu_ih_funcs tonga_ih_funcs = {
static void tonga_ih_set_interrupt_funcs(struct amdgpu_device *adev)
{
if (adev->irq.ih_funcs == NULL)
adev->irq.ih_funcs = &tonga_ih_funcs;
adev->irq.ih_funcs = &tonga_ih_funcs;
}
const struct amdgpu_ip_block_version tonga_ih_ip_block =

2
drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c
View File

@ -108,7 +108,7 @@ static int uvd_v4_2_sw_init(void *handle)
int r;
/* UVD TRAP */
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 124, &adev->uvd.inst->irq);
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 124, &adev->uvd.inst->irq);
if (r)
return r;

2
drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c
View File

@ -105,7 +105,7 @@ static int uvd_v5_0_sw_init(void *handle)
int r;
/* UVD TRAP */
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_UVD_SYSTEM_MESSAGE, &adev->uvd.inst->irq);
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_UVD_SYSTEM_MESSAGE, &adev->uvd.inst->irq);
if (r)
return r;

14
drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c
View File

@ -274,7 +274,7 @@ static int uvd_v6_0_enc_get_create_msg(struct amdgpu_ring *ring, uint32_t handle
*/
static int uvd_v6_0_enc_get_destroy_msg(struct amdgpu_ring *ring,
uint32_t handle,
bool direct, struct dma_fence **fence)
struct dma_fence **fence)
{
const unsigned ib_size_dw = 16;
struct amdgpu_job *job;
@ -310,11 +310,7 @@ static int uvd_v6_0_enc_get_destroy_msg(struct amdgpu_ring *ring,
for (i = ib->length_dw; i < ib_size_dw; ++i)
ib->ptr[i] = 0x0;
if (direct)
r = amdgpu_job_submit_direct(job, ring, &f);
else
r = amdgpu_job_submit(job, &ring->adev->vce.entity,
AMDGPU_FENCE_OWNER_UNDEFINED, &f);
r = amdgpu_job_submit_direct(job, ring, &f);
if (r)
goto err;
@ -345,7 +341,7 @@ static int uvd_v6_0_enc_ring_test_ib(struct amdgpu_ring *ring, long timeout)
goto error;
}
r = uvd_v6_0_enc_get_destroy_msg(ring, 1, true, &fence);
r = uvd_v6_0_enc_get_destroy_msg(ring, 1, &fence);
if (r) {
DRM_ERROR("amdgpu: failed to get destroy ib (%ld).\n", r);
goto error;
@ -393,14 +389,14 @@ static int uvd_v6_0_sw_init(void *handle)
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
/* UVD TRAP */
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_UVD_SYSTEM_MESSAGE, &adev->uvd.inst->irq);
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_UVD_SYSTEM_MESSAGE, &adev->uvd.inst->irq);
if (r)
return r;
/* UVD ENC TRAP */
if (uvd_v6_0_enc_support(adev)) {
for (i = 0; i < adev->uvd.num_enc_rings; ++i) {
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, i + VISLANDS30_IV_SRCID_UVD_ENC_GEN_PURP, &adev->uvd.inst->irq);
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, i + VISLANDS30_IV_SRCID_UVD_ENC_GEN_PURP, &adev->uvd.inst->irq);
if (r)
return r;
}

12
drivers/gpu/drm/amd/amdgpu/uvd_v7_0.c
View File

@ -280,8 +280,8 @@ static int uvd_v7_0_enc_get_create_msg(struct amdgpu_ring *ring, uint32_t handle
*
* Close up a stream for HW test or if userspace failed to do so
*/
int uvd_v7_0_enc_get_destroy_msg(struct amdgpu_ring *ring, uint32_t handle,
bool direct, struct dma_fence **fence)
static int uvd_v7_0_enc_get_destroy_msg(struct amdgpu_ring *ring, uint32_t handle,
struct dma_fence **fence)
{
const unsigned ib_size_dw = 16;
struct amdgpu_job *job;
@ -317,11 +317,7 @@ int uvd_v7_0_enc_get_destroy_msg(struct amdgpu_ring *ring, uint32_t handle,
for (i = ib->length_dw; i < ib_size_dw; ++i)
ib->ptr[i] = 0x0;
if (direct)
r = amdgpu_job_submit_direct(job, ring, &f);
else
r = amdgpu_job_submit(job, &ring->adev->vce.entity,
AMDGPU_FENCE_OWNER_UNDEFINED, &f);
r = amdgpu_job_submit_direct(job, ring, &f);
if (r)
goto err;
@ -352,7 +348,7 @@ static int uvd_v7_0_enc_ring_test_ib(struct amdgpu_ring *ring, long timeout)
goto error;
}
r = uvd_v7_0_enc_get_destroy_msg(ring, 1, true, &fence);
r = uvd_v7_0_enc_get_destroy_msg(ring, 1, &fence);
if (r) {
DRM_ERROR("amdgpu: (%d)failed to get destroy ib (%ld).\n", ring->me, r);
goto error;

2
drivers/gpu/drm/amd/amdgpu/vce_v2_0.c
View File

@ -417,7 +417,7 @@ static int vce_v2_0_sw_init(void *handle)
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
/* VCE */
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, 167, &adev->vce.irq);
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, 167, &adev->vce.irq);
if (r)
return r;

2
drivers/gpu/drm/amd/amdgpu/vce_v3_0.c
View File

@ -423,7 +423,7 @@ static int vce_v3_0_sw_init(void *handle)
int r, i;
/* VCE */
r = amdgpu_irq_add_id(adev, AMDGPU_IH_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_VCE_TRAP, &adev->vce.irq);
r = amdgpu_irq_add_id(adev, AMDGPU_IRQ_CLIENTID_LEGACY, VISLANDS30_IV_SRCID_VCE_TRAP, &adev->vce.irq);
if (r)
return r;

540
drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c
View File

@ -37,6 +37,11 @@
#include "ivsrcid/vcn/irqsrcs_vcn_1_0.h"
#define mmUVD_RBC_XX_IB_REG_CHECK 0x05ab
#define mmUVD_RBC_XX_IB_REG_CHECK_BASE_IDX 1
#define mmUVD_REG_XX_MASK 0x05ac
#define mmUVD_REG_XX_MASK_BASE_IDX 1
static int vcn_v1_0_stop(struct amdgpu_device *adev);
static void vcn_v1_0_set_dec_ring_funcs(struct amdgpu_device *adev);
static void vcn_v1_0_set_enc_ring_funcs(struct amdgpu_device *adev);
@ -198,7 +203,8 @@ static int vcn_v1_0_hw_init(void *handle)
done:
if (!r)
DRM_INFO("VCN decode and encode initialized successfully.\n");
DRM_INFO("VCN decode and encode initialized successfully(under %s).\n",
(adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)?"DPG Mode":"SPG Mode");
return r;
}
@ -266,17 +272,18 @@ static int vcn_v1_0_resume(void *handle)
}
/**
* vcn_v1_0_mc_resume - memory controller programming
* vcn_v1_0_mc_resume_spg_mode - memory controller programming
*
* @adev: amdgpu_device pointer
*
* Let the VCN memory controller know it's offsets
*/
static void vcn_v1_0_mc_resume(struct amdgpu_device *adev)
static void vcn_v1_0_mc_resume_spg_mode(struct amdgpu_device *adev)
{
uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.fw->size + 4);
uint32_t offset;
/* cache window 0: fw */
if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
WREG32_SOC15(UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW,
(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN].tmr_mc_addr_lo));
@ -296,20 +303,21 @@ static void vcn_v1_0_mc_resume(struct amdgpu_device *adev)
WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_SIZE0, size);
/* cache window 1: stack */
WREG32_SOC15(UVD, 0, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW,
lower_32_bits(adev->vcn.gpu_addr + offset));
WREG32_SOC15(UVD, 0, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH,
upper_32_bits(adev->vcn.gpu_addr + offset));
WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_OFFSET1, 0);
WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_SIZE1, AMDGPU_VCN_HEAP_SIZE);
WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_SIZE1, AMDGPU_VCN_STACK_SIZE);
/* cache window 2: context */
WREG32_SOC15(UVD, 0, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW,
lower_32_bits(adev->vcn.gpu_addr + offset + AMDGPU_VCN_HEAP_SIZE));
lower_32_bits(adev->vcn.gpu_addr + offset + AMDGPU_VCN_STACK_SIZE));
WREG32_SOC15(UVD, 0, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH,
upper_32_bits(adev->vcn.gpu_addr + offset + AMDGPU_VCN_HEAP_SIZE));
upper_32_bits(adev->vcn.gpu_addr + offset + AMDGPU_VCN_STACK_SIZE));
WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_OFFSET2, 0);
WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_SIZE2,
AMDGPU_VCN_STACK_SIZE + (AMDGPU_VCN_SESSION_SIZE * 40));
WREG32_SOC15(UVD, 0, mmUVD_VCPU_CACHE_SIZE2, AMDGPU_VCN_CONTEXT_SIZE);
WREG32_SOC15(UVD, 0, mmUVD_UDEC_ADDR_CONFIG,
adev->gfx.config.gb_addr_config);
@ -317,6 +325,96 @@ static void vcn_v1_0_mc_resume(struct amdgpu_device *adev)
adev->gfx.config.gb_addr_config);
WREG32_SOC15(UVD, 0, mmUVD_UDEC_DBW_ADDR_CONFIG,
adev->gfx.config.gb_addr_config);
WREG32_SOC15(UVD, 0, mmUVD_UDEC_DBW_UV_ADDR_CONFIG,
adev->gfx.config.gb_addr_config);
WREG32_SOC15(UVD, 0, mmUVD_MIF_CURR_ADDR_CONFIG,
adev->gfx.config.gb_addr_config);
WREG32_SOC15(UVD, 0, mmUVD_MIF_CURR_UV_ADDR_CONFIG,
adev->gfx.config.gb_addr_config);
WREG32_SOC15(UVD, 0, mmUVD_MIF_RECON1_ADDR_CONFIG,
adev->gfx.config.gb_addr_config);
WREG32_SOC15(UVD, 0, mmUVD_MIF_RECON1_UV_ADDR_CONFIG,
adev->gfx.config.gb_addr_config);
WREG32_SOC15(UVD, 0, mmUVD_MIF_REF_ADDR_CONFIG,
adev->gfx.config.gb_addr_config);
WREG32_SOC15(UVD, 0, mmUVD_MIF_REF_UV_ADDR_CONFIG,
adev->gfx.config.gb_addr_config);
WREG32_SOC15(UVD, 0, mmUVD_JPEG_ADDR_CONFIG,
adev->gfx.config.gb_addr_config);
WREG32_SOC15(UVD, 0, mmUVD_JPEG_UV_ADDR_CONFIG,
adev->gfx.config.gb_addr_config);
}
static void vcn_v1_0_mc_resume_dpg_mode(struct amdgpu_device *adev)
{
uint32_t size = AMDGPU_GPU_PAGE_ALIGN(adev->vcn.fw->size + 4);
uint32_t offset;
/* cache window 0: fw */
if (adev->firmware.load_type == AMDGPU_FW_LOAD_PSP) {
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW,
(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN].tmr_mc_addr_lo),
0xFFFFFFFF, 0);
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH,
(adev->firmware.ucode[AMDGPU_UCODE_ID_VCN].tmr_mc_addr_hi),
0xFFFFFFFF, 0);
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_VCPU_CACHE_OFFSET0, 0,
0xFFFFFFFF, 0);
offset = 0;
} else {
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_LOW,
lower_32_bits(adev->vcn.gpu_addr), 0xFFFFFFFF, 0);
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_VCPU_CACHE_64BIT_BAR_HIGH,
upper_32_bits(adev->vcn.gpu_addr), 0xFFFFFFFF, 0);
offset = size;
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_VCPU_CACHE_OFFSET0,
AMDGPU_UVD_FIRMWARE_OFFSET >> 3, 0xFFFFFFFF, 0);
}
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_VCPU_CACHE_SIZE0, size, 0xFFFFFFFF, 0);
/* cache window 1: stack */
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_LOW,
lower_32_bits(adev->vcn.gpu_addr + offset), 0xFFFFFFFF, 0);
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_VCPU_CACHE1_64BIT_BAR_HIGH,
upper_32_bits(adev->vcn.gpu_addr + offset), 0xFFFFFFFF, 0);
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_VCPU_CACHE_OFFSET1, 0,
0xFFFFFFFF, 0);
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_VCPU_CACHE_SIZE1, AMDGPU_VCN_STACK_SIZE,
0xFFFFFFFF, 0);
/* cache window 2: context */
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_LOW,
lower_32_bits(adev->vcn.gpu_addr + offset + AMDGPU_VCN_STACK_SIZE),
0xFFFFFFFF, 0);
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_VCPU_CACHE2_64BIT_BAR_HIGH,
upper_32_bits(adev->vcn.gpu_addr + offset + AMDGPU_VCN_STACK_SIZE),
0xFFFFFFFF, 0);
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_VCPU_CACHE_OFFSET2, 0, 0xFFFFFFFF, 0);
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_VCPU_CACHE_SIZE2, AMDGPU_VCN_CONTEXT_SIZE,
0xFFFFFFFF, 0);
/* VCN global tiling registers */
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_UDEC_ADDR_CONFIG,
adev->gfx.config.gb_addr_config, 0xFFFFFFFF, 0);
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_UDEC_DB_ADDR_CONFIG,
adev->gfx.config.gb_addr_config, 0xFFFFFFFF, 0);
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_UDEC_DBW_ADDR_CONFIG,
adev->gfx.config.gb_addr_config, 0xFFFFFFFF, 0);
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_UDEC_DBW_UV_ADDR_CONFIG,
adev->gfx.config.gb_addr_config, 0xFFFFFFFF, 0);
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_MIF_CURR_ADDR_CONFIG,
adev->gfx.config.gb_addr_config, 0xFFFFFFFF, 0);
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_MIF_CURR_UV_ADDR_CONFIG,
adev->gfx.config.gb_addr_config, 0xFFFFFFFF, 0);
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_MIF_RECON1_ADDR_CONFIG,
adev->gfx.config.gb_addr_config, 0xFFFFFFFF, 0);
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_MIF_RECON1_UV_ADDR_CONFIG,
adev->gfx.config.gb_addr_config, 0xFFFFFFFF, 0);
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_MIF_REF_ADDR_CONFIG,
adev->gfx.config.gb_addr_config, 0xFFFFFFFF, 0);
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_MIF_REF_UV_ADDR_CONFIG,
adev->gfx.config.gb_addr_config, 0xFFFFFFFF, 0);
}
/**
@ -519,6 +617,60 @@ static void vcn_v1_0_enable_clock_gating(struct amdgpu_device *adev)
WREG32_SOC15(VCN, 0, mmUVD_SUVD_CGC_CTRL, data);
}
static void vcn_v1_0_clock_gating_dpg_mode(struct amdgpu_device *adev, uint8_t sram_sel)
{
uint32_t reg_data = 0;
/* disable JPEG CGC */
if (adev->cg_flags & AMD_CG_SUPPORT_VCN_MGCG)
reg_data = 1 << JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
else
reg_data = 0 << JPEG_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
reg_data |= 1 << JPEG_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
reg_data |= 4 << JPEG_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
WREG32_SOC15_DPG_MODE(UVD, 0, mmJPEG_CGC_CTRL, reg_data, 0xFFFFFFFF, sram_sel);
WREG32_SOC15_DPG_MODE(UVD, 0, mmJPEG_CGC_GATE, 0, 0xFFFFFFFF, sram_sel);
/* enable sw clock gating control */
if (adev->cg_flags & AMD_CG_SUPPORT_VCN_MGCG)
reg_data = 1 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
else
reg_data = 0 << UVD_CGC_CTRL__DYN_CLOCK_MODE__SHIFT;
reg_data |= 1 << UVD_CGC_CTRL__CLK_GATE_DLY_TIMER__SHIFT;
reg_data |= 4 << UVD_CGC_CTRL__CLK_OFF_DELAY__SHIFT;
reg_data &= ~(UVD_CGC_CTRL__UDEC_RE_MODE_MASK |
UVD_CGC_CTRL__UDEC_CM_MODE_MASK |
UVD_CGC_CTRL__UDEC_IT_MODE_MASK |
UVD_CGC_CTRL__UDEC_DB_MODE_MASK |
UVD_CGC_CTRL__UDEC_MP_MODE_MASK |
UVD_CGC_CTRL__SYS_MODE_MASK |
UVD_CGC_CTRL__UDEC_MODE_MASK |
UVD_CGC_CTRL__MPEG2_MODE_MASK |
UVD_CGC_CTRL__REGS_MODE_MASK |
UVD_CGC_CTRL__RBC_MODE_MASK |
UVD_CGC_CTRL__LMI_MC_MODE_MASK |
UVD_CGC_CTRL__LMI_UMC_MODE_MASK |
UVD_CGC_CTRL__IDCT_MODE_MASK |
UVD_CGC_CTRL__MPRD_MODE_MASK |
UVD_CGC_CTRL__MPC_MODE_MASK |
UVD_CGC_CTRL__LBSI_MODE_MASK |
UVD_CGC_CTRL__LRBBM_MODE_MASK |
UVD_CGC_CTRL__WCB_MODE_MASK |
UVD_CGC_CTRL__VCPU_MODE_MASK |
UVD_CGC_CTRL__SCPU_MODE_MASK);
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_CGC_CTRL, reg_data, 0xFFFFFFFF, sram_sel);
/* turn off clock gating */
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_CGC_GATE, 0, 0xFFFFFFFF, sram_sel);
/* turn on SUVD clock gating */
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_SUVD_CGC_GATE, 1, 0xFFFFFFFF, sram_sel);
/* turn on sw mode in UVD_SUVD_CGC_CTRL */
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_SUVD_CGC_CTRL, 0, 0xFFFFFFFF, sram_sel);
}
static void vcn_1_0_disable_static_power_gating(struct amdgpu_device *adev)
{
uint32_t data = 0;
@ -614,7 +766,7 @@ static void vcn_1_0_enable_static_power_gating(struct amdgpu_device *adev)
*
* Setup and start the VCN block
*/
static int vcn_v1_0_start(struct amdgpu_device *adev)
static int vcn_v1_0_start_spg_mode(struct amdgpu_device *adev)
{
struct amdgpu_ring *ring = &adev->vcn.ring_dec;
uint32_t rb_bufsz, tmp;
@ -625,41 +777,24 @@ static int vcn_v1_0_start(struct amdgpu_device *adev)
lmi_swap_cntl = 0;
vcn_1_0_disable_static_power_gating(adev);
tmp = RREG32_SOC15(UVD, 0, mmUVD_STATUS) | UVD_STATUS__UVD_BUSY;
WREG32_SOC15(UVD, 0, mmUVD_STATUS, tmp);
/* disable clock gating */
vcn_v1_0_disable_clock_gating(adev);
vcn_v1_0_mc_resume(adev);
/* disable interupt */
WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_MASTINT_EN), 0,
~UVD_MASTINT_EN__VCPU_EN_MASK);
/* stall UMC and register bus before resetting VCPU */
WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_CTRL2),
UVD_LMI_CTRL2__STALL_ARB_UMC_MASK,
~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK);
mdelay(1);
/* put LMI, VCPU, RBC etc... into reset */
WREG32_SOC15(UVD, 0, mmUVD_SOFT_RESET,
UVD_SOFT_RESET__LMI_SOFT_RESET_MASK |
UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK |
UVD_SOFT_RESET__LBSI_SOFT_RESET_MASK |
UVD_SOFT_RESET__RBC_SOFT_RESET_MASK |
UVD_SOFT_RESET__CSM_SOFT_RESET_MASK |
UVD_SOFT_RESET__CXW_SOFT_RESET_MASK |
UVD_SOFT_RESET__TAP_SOFT_RESET_MASK |
UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK);
mdelay(5);
/* initialize VCN memory controller */
WREG32_SOC15(UVD, 0, mmUVD_LMI_CTRL,
(0x40 << UVD_LMI_CTRL__WRITE_CLEAN_TIMER__SHIFT) |
UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK |
UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK |
UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK |
UVD_LMI_CTRL__REQ_MODE_MASK |
0x00100000L);
tmp = RREG32_SOC15(UVD, 0, mmUVD_LMI_CTRL);
WREG32_SOC15(UVD, 0, mmUVD_LMI_CTRL, tmp |
UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK |
UVD_LMI_CTRL__MASK_MC_URGENT_MASK |
UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK |
UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK);
#ifdef __BIG_ENDIAN
/* swap (8 in 32) RB and IB */
@ -667,41 +802,61 @@ static int vcn_v1_0_start(struct amdgpu_device *adev)
#endif
WREG32_SOC15(UVD, 0, mmUVD_LMI_SWAP_CNTL, lmi_swap_cntl);
WREG32_SOC15(UVD, 0, mmUVD_MPC_SET_MUXA0, 0x40c2040);
WREG32_SOC15(UVD, 0, mmUVD_MPC_SET_MUXA1, 0x0);
WREG32_SOC15(UVD, 0, mmUVD_MPC_SET_MUXB0, 0x40c2040);
WREG32_SOC15(UVD, 0, mmUVD_MPC_SET_MUXB1, 0x0);
WREG32_SOC15(UVD, 0, mmUVD_MPC_SET_ALU, 0);
WREG32_SOC15(UVD, 0, mmUVD_MPC_SET_MUX, 0x88);
tmp = RREG32_SOC15(UVD, 0, mmUVD_MPC_CNTL);
tmp &= ~UVD_MPC_CNTL__REPLACEMENT_MODE_MASK;
tmp |= 0x2 << UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT;
WREG32_SOC15(UVD, 0, mmUVD_MPC_CNTL, tmp);
/* take all subblocks out of reset, except VCPU */
WREG32_SOC15(UVD, 0, mmUVD_SOFT_RESET,
UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK);
mdelay(5);
WREG32_SOC15(UVD, 0, mmUVD_MPC_SET_MUXA0,
((0x1 << UVD_MPC_SET_MUXA0__VARA_1__SHIFT) |
(0x2 << UVD_MPC_SET_MUXA0__VARA_2__SHIFT) |
(0x3 << UVD_MPC_SET_MUXA0__VARA_3__SHIFT) |
(0x4 << UVD_MPC_SET_MUXA0__VARA_4__SHIFT)));
WREG32_SOC15(UVD, 0, mmUVD_MPC_SET_MUXB0,
((0x1 << UVD_MPC_SET_MUXB0__VARB_1__SHIFT) |
(0x2 << UVD_MPC_SET_MUXB0__VARB_2__SHIFT) |
(0x3 << UVD_MPC_SET_MUXB0__VARB_3__SHIFT) |
(0x4 << UVD_MPC_SET_MUXB0__VARB_4__SHIFT)));
WREG32_SOC15(UVD, 0, mmUVD_MPC_SET_MUX,
((0x0 << UVD_MPC_SET_MUX__SET_0__SHIFT) |
(0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) |
(0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT)));
vcn_v1_0_mc_resume_spg_mode(adev);
WREG32_SOC15(UVD, 0, mmUVD_REG_XX_MASK, 0x10);
WREG32_SOC15(UVD, 0, mmUVD_RBC_XX_IB_REG_CHECK,
RREG32_SOC15(UVD, 0, mmUVD_RBC_XX_IB_REG_CHECK) | 0x3);
/* enable VCPU clock */
WREG32_SOC15(UVD, 0, mmUVD_VCPU_CNTL,
UVD_VCPU_CNTL__CLK_EN_MASK);
WREG32_SOC15(UVD, 0, mmUVD_VCPU_CNTL, UVD_VCPU_CNTL__CLK_EN_MASK);
/* boot up the VCPU */
WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_SOFT_RESET), 0,
~UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK);
/* enable UMC */
WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_CTRL2), 0,
~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK);
/* boot up the VCPU */
WREG32_SOC15(UVD, 0, mmUVD_SOFT_RESET, 0);
mdelay(10);
tmp = RREG32_SOC15(UVD, 0, mmUVD_SOFT_RESET);
tmp &= ~UVD_SOFT_RESET__LMI_SOFT_RESET_MASK;
tmp &= ~UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK;
WREG32_SOC15(UVD, 0, mmUVD_SOFT_RESET, tmp);
for (i = 0; i < 10; ++i) {
uint32_t status;
for (j = 0; j < 100; ++j) {
status = RREG32_SOC15(UVD, 0, mmUVD_STATUS);
if (status & 2)
if (status & UVD_STATUS__IDLE)
break;
mdelay(10);
}
r = 0;
if (status & 2)
if (status & UVD_STATUS__IDLE)
break;
DRM_ERROR("VCN decode not responding, trying to reset the VCPU!!!\n");
@ -721,24 +876,22 @@ static int vcn_v1_0_start(struct amdgpu_device *adev)
}
/* enable master interrupt */
WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_MASTINT_EN),
(UVD_MASTINT_EN__VCPU_EN_MASK|UVD_MASTINT_EN__SYS_EN_MASK),
~(UVD_MASTINT_EN__VCPU_EN_MASK|UVD_MASTINT_EN__SYS_EN_MASK));
UVD_MASTINT_EN__VCPU_EN_MASK, ~UVD_MASTINT_EN__VCPU_EN_MASK);
/* enable system interrupt for JRBC, TODO: move to set interrupt*/
WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_SYS_INT_EN),
UVD_SYS_INT_EN__UVD_JRBC_EN_MASK,
~UVD_SYS_INT_EN__UVD_JRBC_EN_MASK);
/* clear the bit 4 of VCN_STATUS */
WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_STATUS), 0,
~(2 << UVD_STATUS__VCPU_REPORT__SHIFT));
/* clear the busy bit of UVD_STATUS */
tmp = RREG32_SOC15(UVD, 0, mmUVD_STATUS) & ~UVD_STATUS__UVD_BUSY;
WREG32_SOC15(UVD, 0, mmUVD_STATUS, tmp);
/* force RBC into idle state */
rb_bufsz = order_base_2(ring->ring_size);
tmp = REG_SET_FIELD(0, UVD_RBC_RB_CNTL, RB_BUFSZ, rb_bufsz);
tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_BLKSZ, 1);
tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_NO_FETCH, 1);
tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_WPTR_POLL_EN, 0);
tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_NO_UPDATE, 1);
tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_RPTR_WR_EN, 1);
WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_CNTL, tmp);
@ -759,6 +912,8 @@ static int vcn_v1_0_start(struct amdgpu_device *adev)
/* Initialize the ring buffer's read and write pointers */
WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_RPTR, 0);
WREG32_SOC15(UVD, 0, mmUVD_SCRATCH2, 0);
ring->wptr = RREG32_SOC15(UVD, 0, mmUVD_RBC_RB_RPTR);
WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR,
lower_32_bits(ring->wptr));
@ -782,12 +937,13 @@ static int vcn_v1_0_start(struct amdgpu_device *adev)
ring = &adev->vcn.ring_jpeg;
WREG32_SOC15(UVD, 0, mmUVD_LMI_JRBC_RB_VMID, 0);
WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_CNTL, (0x00000001L | 0x00000002L));
WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_CNTL, UVD_JRBC_RB_CNTL__RB_NO_FETCH_MASK |
UVD_JRBC_RB_CNTL__RB_RPTR_WR_EN_MASK);
WREG32_SOC15(UVD, 0, mmUVD_LMI_JRBC_RB_64BIT_BAR_LOW, lower_32_bits(ring->gpu_addr));
WREG32_SOC15(UVD, 0, mmUVD_LMI_JRBC_RB_64BIT_BAR_HIGH, upper_32_bits(ring->gpu_addr));
WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_RPTR, 0);
WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_WPTR, 0);
WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_CNTL, 0x00000002L);
WREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_CNTL, UVD_JRBC_RB_CNTL__RB_RPTR_WR_EN_MASK);
/* initialize wptr */
ring->wptr = RREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_WPTR);
@ -799,6 +955,166 @@ static int vcn_v1_0_start(struct amdgpu_device *adev)
return 0;
}
static int vcn_v1_0_start_dpg_mode(struct amdgpu_device *adev)
{
struct amdgpu_ring *ring = &adev->vcn.ring_dec;
uint32_t rb_bufsz, tmp;
uint32_t lmi_swap_cntl;
/* disable byte swapping */
lmi_swap_cntl = 0;
vcn_1_0_enable_static_power_gating(adev);
/* enable dynamic power gating mode */
tmp = RREG32_SOC15(UVD, 0, mmUVD_POWER_STATUS);
tmp |= UVD_POWER_STATUS__UVD_PG_MODE_MASK;
tmp |= UVD_POWER_STATUS__UVD_PG_EN_MASK;
WREG32_SOC15(UVD, 0, mmUVD_POWER_STATUS, tmp);
/* enable clock gating */
vcn_v1_0_clock_gating_dpg_mode(adev, 0);
/* enable VCPU clock */
tmp = (0xFF << UVD_VCPU_CNTL__PRB_TIMEOUT_VAL__SHIFT);
tmp |= UVD_VCPU_CNTL__CLK_EN_MASK;
tmp |= UVD_VCPU_CNTL__MIF_WR_LOW_THRESHOLD_BP_MASK;
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_VCPU_CNTL, tmp, 0xFFFFFFFF, 0);
/* disable interupt */
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_MASTINT_EN,
0, UVD_MASTINT_EN__VCPU_EN_MASK, 0);
/* initialize VCN memory controller */
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_CTRL,
(8 << UVD_LMI_CTRL__WRITE_CLEAN_TIMER__SHIFT) |
UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK |
UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK |
UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK |
UVD_LMI_CTRL__REQ_MODE_MASK |
UVD_LMI_CTRL__CRC_RESET_MASK |
UVD_LMI_CTRL__MASK_MC_URGENT_MASK |
0x00100000L, 0xFFFFFFFF, 0);
#ifdef __BIG_ENDIAN
/* swap (8 in 32) RB and IB */
lmi_swap_cntl = 0xa;
#endif
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_SWAP_CNTL, lmi_swap_cntl, 0xFFFFFFFF, 0);
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_MPC_CNTL,
0x2 << UVD_MPC_CNTL__REPLACEMENT_MODE__SHIFT, 0xFFFFFFFF, 0);
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_MPC_SET_MUXA0,
((0x1 << UVD_MPC_SET_MUXA0__VARA_1__SHIFT) |
(0x2 << UVD_MPC_SET_MUXA0__VARA_2__SHIFT) |
(0x3 << UVD_MPC_SET_MUXA0__VARA_3__SHIFT) |
(0x4 << UVD_MPC_SET_MUXA0__VARA_4__SHIFT)), 0xFFFFFFFF, 0);
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_MPC_SET_MUXB0,
((0x1 << UVD_MPC_SET_MUXB0__VARB_1__SHIFT) |
(0x2 << UVD_MPC_SET_MUXB0__VARB_2__SHIFT) |
(0x3 << UVD_MPC_SET_MUXB0__VARB_3__SHIFT) |
(0x4 << UVD_MPC_SET_MUXB0__VARB_4__SHIFT)), 0xFFFFFFFF, 0);
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_MPC_SET_MUX,
((0x0 << UVD_MPC_SET_MUX__SET_0__SHIFT) |
(0x1 << UVD_MPC_SET_MUX__SET_1__SHIFT) |
(0x2 << UVD_MPC_SET_MUX__SET_2__SHIFT)), 0xFFFFFFFF, 0);
vcn_v1_0_mc_resume_dpg_mode(adev);
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_REG_XX_MASK, 0x10, 0xFFFFFFFF, 0);
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_RBC_XX_IB_REG_CHECK, 0x3, 0xFFFFFFFF, 0);
/* boot up the VCPU */
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_SOFT_RESET, 0, 0xFFFFFFFF, 0);
/* enable UMC */
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_CTRL2,
0x1F << UVD_LMI_CTRL2__RE_OFLD_MIF_WR_REQ_NUM__SHIFT,
0xFFFFFFFF, 0);
/* enable master interrupt */
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_MASTINT_EN,
UVD_MASTINT_EN__VCPU_EN_MASK, UVD_MASTINT_EN__VCPU_EN_MASK, 0);
vcn_v1_0_clock_gating_dpg_mode(adev, 1);
/* setup mmUVD_LMI_CTRL */
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_LMI_CTRL,
(8 << UVD_LMI_CTRL__WRITE_CLEAN_TIMER__SHIFT) |
UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK |
UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK |
UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK |
UVD_LMI_CTRL__REQ_MODE_MASK |
UVD_LMI_CTRL__CRC_RESET_MASK |
UVD_LMI_CTRL__MASK_MC_URGENT_MASK |
0x00100000L, 0xFFFFFFFF, 1);
tmp = adev->gfx.config.gb_addr_config;
/* setup VCN global tiling registers */
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_JPEG_ADDR_CONFIG, tmp, 0xFFFFFFFF, 1);
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_JPEG_UV_ADDR_CONFIG, tmp, 0xFFFFFFFF, 1);
/* enable System Interrupt for JRBC */
WREG32_SOC15_DPG_MODE(UVD, 0, mmUVD_SYS_INT_EN,
UVD_SYS_INT_EN__UVD_JRBC_EN_MASK, 0xFFFFFFFF, 1);
/* force RBC into idle state */
rb_bufsz = order_base_2(ring->ring_size);
tmp = REG_SET_FIELD(0, UVD_RBC_RB_CNTL, RB_BUFSZ, rb_bufsz);
tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_BLKSZ, 1);
tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_NO_FETCH, 1);
tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_NO_UPDATE, 1);
tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_RPTR_WR_EN, 1);
WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_CNTL, tmp);
/* set the write pointer delay */
WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR_CNTL, 0);
/* set the wb address */
WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_RPTR_ADDR,
(upper_32_bits(ring->gpu_addr) >> 2));
/* programm the RB_BASE for ring buffer */
WREG32_SOC15(UVD, 0, mmUVD_LMI_RBC_RB_64BIT_BAR_LOW,
lower_32_bits(ring->gpu_addr));
WREG32_SOC15(UVD, 0, mmUVD_LMI_RBC_RB_64BIT_BAR_HIGH,
upper_32_bits(ring->gpu_addr));
/* Initialize the ring buffer's read and write pointers */
WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_RPTR, 0);
WREG32_SOC15(UVD, 0, mmUVD_SCRATCH2, 0);
ring->wptr = RREG32_SOC15(UVD, 0, mmUVD_RBC_RB_RPTR);
WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR,
lower_32_bits(ring->wptr));
WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_RBC_RB_CNTL), 0,
~UVD_RBC_RB_CNTL__RB_NO_FETCH_MASK);
/* initialize wptr */
ring->wptr = RREG32_SOC15(UVD, 0, mmUVD_JRBC_RB_WPTR);
/* copy patch commands to the jpeg ring */
vcn_v1_0_jpeg_ring_set_patch_ring(ring,
(ring->wptr + ring->max_dw * amdgpu_sched_hw_submission));
return 0;
}
static int vcn_v1_0_start(struct amdgpu_device *adev)
{
int r;
if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)
r = vcn_v1_0_start_dpg_mode(adev);
else
r = vcn_v1_0_start_spg_mode(adev);
return r;
}
/**
* vcn_v1_0_stop - stop VCN block
*
@ -806,41 +1122,90 @@ static int vcn_v1_0_start(struct amdgpu_device *adev)
*
* stop the VCN block
*/
static int vcn_v1_0_stop(struct amdgpu_device *adev)
static int vcn_v1_0_stop_spg_mode(struct amdgpu_device *adev)
{
/* force RBC into idle state */
WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_CNTL, 0x11010101);
int ret_code, tmp;
/* Stall UMC and register bus before resetting VCPU */
WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_CTRL2),
UVD_LMI_CTRL2__STALL_ARB_UMC_MASK,
~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK);
mdelay(1);
SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_STATUS, UVD_STATUS__IDLE, 0x7, ret_code);
tmp = UVD_LMI_STATUS__VCPU_LMI_WRITE_CLEAN_MASK |
UVD_LMI_STATUS__READ_CLEAN_MASK |
UVD_LMI_STATUS__WRITE_CLEAN_MASK |
UVD_LMI_STATUS__WRITE_CLEAN_RAW_MASK;
SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_LMI_STATUS, tmp, tmp, ret_code);
/* put VCPU into reset */
WREG32_SOC15(UVD, 0, mmUVD_SOFT_RESET,
UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK);
mdelay(5);
WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_SOFT_RESET),
UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK,
~UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK);
tmp = UVD_LMI_STATUS__UMC_READ_CLEAN_RAW_MASK |
UVD_LMI_STATUS__UMC_WRITE_CLEAN_RAW_MASK;
SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_LMI_STATUS, tmp, tmp, ret_code);
/* disable VCPU clock */
WREG32_SOC15(UVD, 0, mmUVD_VCPU_CNTL, 0x0);
WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_VCPU_CNTL), 0,
~UVD_VCPU_CNTL__CLK_EN_MASK);
/* Unstall UMC and register bus */
WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_LMI_CTRL2), 0,
~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK);
/* reset LMI UMC/LMI */
WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_SOFT_RESET),
UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK,
~UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK);
WREG32_SOC15(VCN, 0, mmUVD_STATUS, 0);
WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_SOFT_RESET),
UVD_SOFT_RESET__LMI_SOFT_RESET_MASK,
~UVD_SOFT_RESET__LMI_SOFT_RESET_MASK);
WREG32_SOC15(UVD, 0, mmUVD_STATUS, 0);
vcn_v1_0_enable_clock_gating(adev);
vcn_1_0_enable_static_power_gating(adev);
return 0;
}
static int vcn_v1_0_stop_dpg_mode(struct amdgpu_device *adev)
{
int ret_code = 0;
/* Wait for power status to be UVD_POWER_STATUS__UVD_POWER_STATUS_TILES_OFF */
SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_POWER_STATUS,
UVD_POWER_STATUS__UVD_POWER_STATUS_TILES_OFF,
UVD_POWER_STATUS__UVD_POWER_STATUS_MASK, ret_code);
if (!ret_code) {
int tmp = RREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR) & 0x7FFFFFFF;
/* wait for read ptr to be equal to write ptr */
SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_RBC_RB_RPTR, tmp, 0xFFFFFFFF, ret_code);
SOC15_WAIT_ON_RREG(UVD, 0, mmUVD_POWER_STATUS,
UVD_POWER_STATUS__UVD_POWER_STATUS_TILES_OFF,
UVD_POWER_STATUS__UVD_POWER_STATUS_MASK, ret_code);
}
/* disable dynamic power gating mode */
WREG32_P(SOC15_REG_OFFSET(UVD, 0, mmUVD_POWER_STATUS), 0,
~UVD_POWER_STATUS__UVD_PG_MODE_MASK);
return 0;
}
static int vcn_v1_0_stop(struct amdgpu_device *adev)
{
int r;
if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)
r = vcn_v1_0_stop_dpg_mode(adev);
else
r = vcn_v1_0_stop_spg_mode(adev);
return r;
}
static bool vcn_v1_0_is_idle(void *handle)
{
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
return (RREG32_SOC15(VCN, 0, mmUVD_STATUS) == 0x2);
return (RREG32_SOC15(VCN, 0, mmUVD_STATUS) == UVD_STATUS__IDLE);
}
static int vcn_v1_0_wait_for_idle(void *handle)
@ -848,7 +1213,8 @@ static int vcn_v1_0_wait_for_idle(void *handle)
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
int ret = 0;
SOC15_WAIT_ON_RREG(VCN, 0, mmUVD_STATUS, 0x2, 0x2, ret);
SOC15_WAIT_ON_RREG(VCN, 0, mmUVD_STATUS, UVD_STATUS__IDLE,
UVD_STATUS__IDLE, ret);
return ret;
}
@ -910,6 +1276,10 @@ static void vcn_v1_0_dec_ring_set_wptr(struct amdgpu_ring *ring)
{
struct amdgpu_device *adev = ring->adev;
if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)
WREG32_SOC15(UVD, 0, mmUVD_SCRATCH2,
lower_32_bits(ring->wptr) | 0x80000000);
WREG32_SOC15(UVD, 0, mmUVD_RBC_RB_WPTR, lower_32_bits(ring->wptr));
}
@ -1633,12 +2003,20 @@ static int vcn_v1_0_set_powergating_state(void *handle,
* revisit this when there is a cleaner line between
* the smc and the hw blocks
*/
int ret;
struct amdgpu_device *adev = (struct amdgpu_device *)handle;
if(state == adev->vcn.cur_state)
return 0;
if (state == AMD_PG_STATE_GATE)
return vcn_v1_0_stop(adev);
ret = vcn_v1_0_stop(adev);
else
return vcn_v1_0_start(adev);
ret = vcn_v1_0_start(adev);
if(!ret)
adev->vcn.cur_state = state;
return ret;
}
static const struct amd_ip_funcs vcn_v1_0_ip_funcs = {

Some files were not shown because too many files have changed in this diff Show More