* [drm/i915/3.17] panic in i915_digport_work_func
@ 2014-08-30 7:03 Mike Galbraith
2014-08-31 12:05 ` Mike Galbraith
0 siblings, 1 reply; 14+ messages in thread
From: Mike Galbraith @ 2014-08-30 7:03 UTC (permalink / raw)
To: LKML; +Cc: Daniel Vetter, Jani Nikula, intel-gfx
Greetings,
My little Toshiba lappy is exploding non-deterministically during boot
with master (<= 3.16 work fine). I tried to bisect it, but apparently
didn't do quite enough reboots before saying 'good', so ended up...
first bad commit: [5fa9be63a42bd4336448d27cbd1f3b993744dd88]
Merge branch 'msm-fixes-3.17' of git://people.freedesktop.org/~robclark/linux into drm-fixes
...touring bisect lala land. Setting up netconsole, I now have a
picture of the little bugger at least.
[ 4.881500] [drm] Replacing VGA console driver
[ 4.889652] checking generic (d0000000 7ff0000) vs hw (d0000000 10000000)
[ 4.897930] fb: switching to inteldrmfb from VESA VGA
[ 4.906132] Console: switching to colour dummy device 80x25
[ 4.906138] usb 7-1: new full-speed USB device number 2 using uhci_hcd
[ 4.928120] i915 0000:00:02.0: irq 29 for MSI/MSI-X
[ 4.928156] [drm] Supports vblank timestamp caching Rev 2 (21.10.2013).
[ 4.928164] [drm] Driver supports precise vblank timestamp query.
[ 4.928269] vgaarb: device changed decodes: PCI:0000:00:02.0,olddecodes=io+mem,decodes=io+mem:owns=io+mem
[ 5.017563] BUG: unable to handle kernel paging request at ffffc900108c4000
[ 5.017586] IP: [<ffffffffa018704d>] gen4_read32+0x3d/0xc0 [i915]
[ 5.017649] PGD 13f021067 PUD 13f022067 PMD 1361be067 PTE 0
[ 5.017667] Oops: 0000 [#1] SMP
[ 5.017678] Modules linked in: i915(+) drm_kms_helper drm i2c_algo_bit processor video thermal thermal_sys button scsi_dh_rdac scsi_dh_alua scsi_dh_emc scsi_dh_hp_sw scsi_dh netconsole atl1c
[ 5.017735] CPU: 0 PID: 54 Comm: kworker/u4:5 Not tainted 3.17.0-master #210
[ 5.017743] Hardware name: TOSHIBA SATELLITE T130/SATELLITE T130, BIOS V1.70 09/29/2009
[ 5.017788] Workqueue: i915-dp i915_digport_work_func [i915]
[ 5.017798] task: ffff8801360de350 ti: ffff8801360e0000 task.ti: ffff8801360e0000
[ 5.017805] RIP: 0010:[<ffffffffa018704d>] [<ffffffffa018704d>] gen4_read32+0x3d/0xc0 [i915]
[ 5.017857] RSP: 0018:ffff8801360e3d48 EFLAGS: 00010086
[ 5.017863] RAX: 0000000000000297 RBX: ffff880037290000 RCX: 00000000000000da
[ 5.017870] RDX: ffffc900108c4000 RSI: ffff8800372993c8 RDI: ffff880037290068
[ 5.017877] RBP: ffff8801360e3d70 R08: 0000000000000001 R09: a000000000000000
[ 5.017883] R10: dffeab59e7299400 R11: 00000000fffedfee R12: 00000000000c4000
[ 5.017890] R13: ffff880037290068 R14: 0000000000000001 R15: 0000000000000001
[ 5.017897] FS: 0000000000000000(0000) GS:ffff88013fc00000(0000) knlGS:0000000000000000
[ 5.017906] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[ 5.017912] CR2: ffffc900108c4000 CR3: 0000000001a15000 CR4: 00000000000407f0
[ 5.017918] Stack:
[ 5.017923] 0000000000800000 0000000000000011 ffff880037067000 0000000000000001
[ 5.017939] ffff880037dcb000 ffff8801360e3d88 ffffffffa01975c3 ffff880037290000
[ 5.017955] ffff8801360e3dc8 ffffffffa01bd883 ffff880037dcb0e0 ffff880037299400
[ 5.017970] Call Trace:
[ 5.018020] [<ffffffffa01975c3>] ibx_digital_port_connected+0x63/0xb0 [i915]
[ 5.018071] [<ffffffffa01bd883>] intel_dp_hpd_pulse+0xc3/0x1f0 [i915]
[ 5.018115] [<ffffffffa0175b0d>] i915_digport_work_func+0x9d/0x120 [i915]
[ 5.018128] [<ffffffff81060b26>] process_one_work+0x186/0x3f0
[ 5.018136] [<ffffffff81060eb1>] worker_thread+0x121/0x480
[ 5.018145] [<ffffffff81060d90>] ? process_one_work+0x3f0/0x3f0
[ 5.018154] [<ffffffff81065689>] kthread+0xc9/0xe0
[ 5.018163] [<ffffffff810655c0>] ? kthread_create_on_node+0x170/0x170
[ 5.018173] [<ffffffff8157edac>] ret_from_fork+0x7c/0xb0
[ 5.018182] [<ffffffff810655c0>] ? kthread_create_on_node+0x170/0x170
[ 5.018188] Code: 41 54 49 89 f4 53 48 8d b7 c8 93 00 00 48 89 fb 48 8b 3f 4c 8d 6b 68 e8 f2 f1 ff ff 4c 89 ef e8 3a 79 3f e1 4c 89 e2 48 03 53 60 <44> 8b 32 48 89 c6 4c 89 ef e8 95 75 3f e1 8b 05 c7 0f 08 00 85
[ 5.018329] RIP [<ffffffffa018704d>] gen4_read32+0x3d/0xc0 [i915]
[ 5.018377] RSP <ffff8801360e3d48>
[ 5.018383] CR2: ffffc900108c4000
[ 5.018390] ---[ end trace d3d1a2cea94fa73a ]---
[ 5.018397] Kernel panic - not syncing: Fatal exception
[ 5.018422] Kernel Offset: 0x0 from 0xffffffff81000000 (relocation range: 0xffffffff80000000-0xffffffff9fffffff)
crash> mod -s i915
MODULE NAME SIZE OBJECT FILE
ffffffffa01d7760 i915 935705 /lib/modules/3.17.0-master/kernel/drivers/gpu/drm/i915/i915.ko
crash> gdb list *gen4_read32+0x3d
0xffffffffa015604d is in gen4_read32 (./arch/x86/include/asm/io.h:56).
51 { asm volatile("mov" size " %0,%1": :reg (val), \
52 "m" (*(volatile type __force *)addr) barrier); }
53
54 build_mmio_read(readb, "b", unsigned char, "=q", :"memory")
55 build_mmio_read(readw, "w", unsigned short, "=r", :"memory")
56 build_mmio_read(readl, "l", unsigned int, "=r", :"memory")
57
58 build_mmio_read(__readb, "b", unsigned char, "=q", )
59 build_mmio_read(__readw, "w", unsigned short, "=r", )
60 build_mmio_read(__readl, "l", unsigned int, "=r", )
0xffffffffa01665c3 is in ibx_digital_port_connected (drivers/gpu/drm/i915/intel_display.c:1044).
1039 default:
1040 return true;
1041 }
1042 }
1043
1044 return I915_READ(SDEISR) & bit;
1045 }
1046
1047 static const char *state_string(bool enabled)
1048 {
crash> gdb list *intel_dp_hpd_pulse+0xc3/0x1f0
0xffffffffa018c7c0 is in intel_dp_hpd_pulse (drivers/gpu/drm/i915/intel_dp.c:4051).
4046 return;
4047 }
4048
4049 bool
4050 intel_dp_hpd_pulse(struct intel_digital_port *intel_dig_port, bool long_hpd)
4051 {
4052 struct intel_dp *intel_dp = &intel_dig_port->dp;
4053 struct intel_encoder *intel_encoder = &intel_dig_port->base;
4054 struct drm_device *dev = intel_dig_port->base.base.dev;
4055 struct drm_i915_private *dev_priv = dev->dev_private;
crash> gdb list *i915_digport_work_func+0x9d
0xffffffffa0144b0d is in i915_digport_work_func (drivers/gpu/drm/i915/i915_irq.c:1125).
1120 } else if (short_port_mask & (1 << i))
1121 valid = true;
1122
1123 if (valid) {
1124 ret = intel_dig_port->hpd_pulse(intel_dig_port, long_hpd);
1125 if (ret == true) {
1126 /* if we get true fallback to old school hpd */
1127 old_bits |= (1 << intel_dig_port->base.hpd_pin);
1128 }
1129 }
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [drm/i915/3.17] panic in i915_digport_work_func
2014-08-30 7:03 [drm/i915/3.17] panic in i915_digport_work_func Mike Galbraith
@ 2014-08-31 12:05 ` Mike Galbraith
2014-09-01 7:02 ` Dave Airlie
0 siblings, 1 reply; 14+ messages in thread
From: Mike Galbraith @ 2014-08-31 12:05 UTC (permalink / raw)
To: LKML; +Cc: Daniel Vetter, Jani Nikula, intel-gfx, Dave Airlie
Non-deterministic explosion culprit is the below. The commit says it
adds mysterious acronyms to Haswell, so I suspect my little core2 lappy
isn't ever supposed to see this code. The ftrace buffer is in fact
empty after a successful boot.
commit 0e32b39ceed665bfa4a77a4bc307b6652b991632
Author: Dave Airlie <airlied@redhat.com>
Date: Fri May 2 14:02:48 2014 +1000
drm/i915: add DP 1.2 MST support (v0.7)
This adds DP 1.2 MST support on Haswell systems.
[ 4.251017] **********************************************************
[ 4.251018] ** NOTICE NOTICE NOTICE NOTICE NOTICE NOTICE NOTICE **
[ 4.251019] ** **
[ 4.251020] ** trace_printk() being used. Allocating extra memory. **
[ 4.251021] ** **
[ 4.251021] ** This means that this is a DEBUG kernel and it is **
[ 4.251022] ** unsafe for produciton use. **
[ 4.251023] ** **
[ 4.251024] ** If you see this message and you are not debugging **
[ 4.251024] ** the kernel, report this immediately to your vendor! **
[ 4.251025] ** **
[ 4.251026] ** NOTICE NOTICE NOTICE NOTICE NOTICE NOTICE NOTICE **
[ 4.251027] **********************************************************
[ 4.375162] [drm] Memory usable by graphics device = 2048M
[ 4.382963] [drm] Replacing VGA console driver
[ 4.390710] checking generic (d0000000 7ff0000) vs hw (d0000000 10000000)
[ 4.398617] fb: switching to inteldrmfb from VESA VGA
[ 4.406578] Console: switching to colour dummy device 80x25
[ 4.432283] i915 0000:00:02.0: irq 29 for MSI/MSI-X
[ 4.432327] [drm] Supports vblank timestamp caching Rev 2 (21.10.2013).
[ 4.432335] [drm] Driver supports precise vblank timestamp query.
[ 4.432437] vgaarb: device changed decodes: PCI:0000:00:02.0,olddecodes=io+mem,decodes=io+mem:owns=io+mem
[ 4.521746] BUG: unable to handle kernel paging request at ffffc900108c4000
[ 4.521781] IP: [<ffffffffa014323d>] gen4_read32+0x3d/0xc0 [i915]
[ 4.521843] PGD 13f021067 PUD 13f022067 PMD 136193067 PTE 0
[ 4.521860] Oops: 0000 [#1] SMP
[ 4.521872] Dumping ftrace buffer:
[ 4.521880] ---------------------------------
[ 4.521936] udevd-120 1d.h. 4672095us : i9xx_hpd_irq_handler: PIN 1 continue
[ 4.521987] udevd-120 1d.h. 4672097us : i9xx_hpd_irq_handler: PIN 2 continue
[ 4.522036] udevd-120 1d.h. 4672098us : i9xx_hpd_irq_handler: PIN 3 continue
[ 4.522085] udevd-120 1d.h. 4672099us : i9xx_hpd_irq_handler: PIN 4 continue
[ 4.522134] udevd-120 1d.h. 4672099us : i9xx_hpd_irq_handler: PIN 5 continue
[ 4.522183] udevd-120 1d.h. 4672100us : i9xx_hpd_irq_handler: PORT 3 PIN 6 continue
[ 4.522233] udevd-120 1d.h. 4672101us : i9xx_hpd_irq_handler: PORT 3 PIN 6 sets queue_dig = true
[ 4.522283] udevd-120 1d.h. 4672102us : i9xx_hpd_irq_handler: QUEUE
[ 4.522331] <...>-6 0.... 4672125us : i915_digport_work_func: WORK
[ 4.522339] ---------------------------------
[ 4.522345] Modules linked in: i915(+) drm_kms_helper drm i2c_algo_bit thermal video processor thermal_sys button scsi_dh_rdac scsi_dh_alua scsi_dh_emc scsi_dh_hp_sw scsi_dh netconsole atl1c
[ 4.522401] CPU: 0 PID: 6 Comm: kworker/u4:0 Not tainted 3.17.0-bisect #222
[ 4.522408] Hardware name: TOSHIBA SATELLITE T130/SATELLITE T130, BIOS V1.70 09/29/2009
[ 4.522453] Workqueue: i915-dp i915_digport_work_func [i915]
[ 4.522462] task: ffff88013f166150 ti: ffff88013f168000 task.ti: ffff88013f168000
[ 4.522470] RIP: 0010:[<ffffffffa014323d>] [<ffffffffa014323d>] gen4_read32+0x3d/0xc0 [i915]
[ 4.522520] RSP: 0018:ffff88013f16bd48 EFLAGS: 00010086
[ 4.522527] RAX: 0000000000000297 RBX: ffff88013b1f0000 RCX: 00000000000000da
[ 4.522533] RDX: ffffc900108c4000 RSI: ffff88013b1f93c8 RDI: ffff88013b1f0068
[ 4.522540] RBP: ffff88013f16bd70 R08: 0000000000000001 R09: 0000000000000024
[ 4.522547] R10: 00000001167af8eb R11: 0000000000000006 R12: 00000000000c4000
[ 4.522553] R13: ffff88013b1f0068 R14: 0000000000000001 R15: 0000000000000001
[ 4.522560] FS: 0000000000000000(0000) GS:ffff88013fc00000(0000) knlGS:0000000000000000
[ 4.522568] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[ 4.522575] CR2: ffffc900108c4000 CR3: 0000000001a15000 CR4: 00000000000407f0
[ 4.522581] Stack:
[ 4.522586] 0000000000800000 0000000000000011 ffff88013584e000 0000000000000001
[ 4.522601] ffff880036c92000 ffff88013f16bd88 ffffffffa01537b3 ffff88013b1f0000
[ 4.522617] ffff88013f16bdc8 ffffffffa0179a73 ffff880036c920e0 ffff88013b1f9400
[ 4.522632] Call Trace:
[ 4.522681] [<ffffffffa01537b3>] ibx_digital_port_connected+0x63/0xb0 [i915]
[ 4.522733] [<ffffffffa0179a73>] intel_dp_hpd_pulse+0xc3/0x1f0 [i915]
[ 4.522777] [<ffffffffa0131d60>] i915_digport_work_func+0xb0/0x130 [i915]
[ 4.522789] [<ffffffff81060b26>] process_one_work+0x186/0x3f0
[ 4.522797] [<ffffffff81060eb1>] worker_thread+0x121/0x480
[ 4.522806] [<ffffffff81060d90>] ? process_one_work+0x3f0/0x3f0
[ 4.522815] [<ffffffff81065689>] kthread+0xc9/0xe0
[ 4.522824] [<ffffffff810655c0>] ? kthread_create_on_node+0x170/0x170
[ 4.522834] [<ffffffff8157edac>] ret_from_fork+0x7c/0xb0
[ 4.522843] [<ffffffff810655c0>] ? kthread_create_on_node+0x170/0x170
[ 4.522849] Code: 41 54 49 89 f4 53 48 8d b7 c8 93 00 00 48 89 fb 48 8b 3f 4c 8d 6b 68 e8 f2 f1 ff ff 4c 89 ef e8 4a b7 43 e1 4c 89 e2 48 03 53 60 <44> 8b 32 48 89 c6 4c 89 ef e8 a5 b3 43 e1 8b 05 f7 0d 08 00 85
[ 4.523001] RIP [<ffffffffa014323d>] gen4_read32+0x3d/0xc0 [i915]
[ 4.523050] RSP <ffff88013f16bd48>
[ 4.523055] CR2: ffffc900108c4000
[ 4.523064] ---[ end trace 08b04cb79531ee7e ]---
[ 4.523070] Kernel panic - not syncing: Fatal exception
[ 4.523096] Dumping ftrace buffer:
[ 4.523101] (ftrace buffer empty)
[ 4.523109] Kernel Offset: 0x0 from 0xffffffff81000000 (relocation range: 0xffffffff80000000-0xffffffff9fffffff)
[ 4.523118] Rebooting in 10 seconds..
---
drivers/gpu/drm/i915/i915_drv.c | 2 ++
drivers/gpu/drm/i915/i915_irq.c | 11 +++++++++--
2 files changed, 11 insertions(+), 2 deletions(-)
--- a/drivers/gpu/drm/i915/i915_drv.c
+++ b/drivers/gpu/drm/i915/i915_drv.c
@@ -496,6 +496,7 @@ bool i915_semaphore_is_enabled(struct dr
void intel_hpd_cancel_work(struct drm_i915_private *dev_priv)
{
+ trace_printk("CANCEL\n");
spin_lock_irq(&dev_priv->irq_lock);
dev_priv->long_hpd_port_mask = 0;
@@ -507,6 +508,7 @@ void intel_hpd_cancel_work(struct drm_i9
cancel_work_sync(&dev_priv->dig_port_work);
cancel_work_sync(&dev_priv->hotplug_work);
cancel_delayed_work_sync(&dev_priv->hotplug_reenable_work);
+ trace_printk("CANCEL DONE\n");
}
static void intel_suspend_encoders(struct drm_i915_private *dev_priv)
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -1100,6 +1100,7 @@ static void i915_digport_work_func(struc
int i, ret;
u32 old_bits = 0;
+ trace_printk("WORK\n");
spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
long_port_mask = dev_priv->long_hpd_port_mask;
dev_priv->long_hpd_port_mask = 0;
@@ -1757,10 +1758,13 @@ static inline void intel_hpd_irq_handler
spin_lock(&dev_priv->irq_lock);
for (i = 1; i < HPD_NUM_PINS; i++) {
- if (!(hpd[i] & hotplug_trigger))
+ if (!(hpd[i] & hotplug_trigger)) {
+ trace_printk("PIN %d continue\n", i);
continue;
+ }
port = get_port_from_pin(i);
+ trace_printk("PORT %d PIN %d continue\n", port, i);
if (port && dev_priv->hpd_irq_port[port]) {
bool long_hpd;
@@ -1784,6 +1788,7 @@ static inline void intel_hpd_irq_handler
hotplug_trigger &= ~hpd[i];
}
queue_dig = true;
+ trace_printk("PORT %d PIN %d sets queue_dig = true\n", port, i);
}
}
@@ -1840,8 +1845,10 @@ static inline void intel_hpd_irq_handler
* queue for otherwise the flush_work in the pageflip code will
* deadlock.
*/
- if (queue_dig)
+ if (queue_dig) {
+ trace_printk("QUEUE\n");
queue_work(dev_priv->dp_wq, &dev_priv->dig_port_work);
+ }
if (queue_hp)
schedule_work(&dev_priv->hotplug_work);
}
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [drm/i915/3.17] panic in i915_digport_work_func
2014-08-31 12:05 ` Mike Galbraith
@ 2014-09-01 7:02 ` Dave Airlie
0 siblings, 0 replies; 14+ messages in thread
From: Dave Airlie @ 2014-09-01 7:02 UTC (permalink / raw)
To: Mike Galbraith; +Cc: LKML, Daniel Vetter, Jani Nikula, intel-gfx, Dave Airlie
[-- Attachment #1: Type: text/plain, Size: 184 bytes --]
> drm/i915: add DP 1.2 MST support (v0.7)
>
> This adds DP 1.2 MST support on Haswell systems.
>
I've attached a patch that might fix this, please test and let me know.
Dave.
[-- Attachment #2: 0001-drm-i915-handle-G45-GM45-pulse-detection-connected-s.patch --]
[-- Type: text/x-patch, Size: 3664 bytes --]
From d407c946fbf5c48f30160591f5bd71fbe158aeb4 Mon Sep 17 00:00:00 2001
From: Dave Airlie <airlied@redhat.com>
Date: Mon, 1 Sep 2014 16:58:12 +1000
Subject: [PATCH] drm/i915: handle G45/GM45 pulse detection connected state.
In the HPD pulse handler we check for long pulses if the port is actually
connected, however we do that for IBX, but we use the pulse handling code on
GM45 systems as well, so we need to use a diffent check.
This patch refactors the digital port connected check out of the g4x detection
path and reuses it in the hpd pulse path.
Should fix:
Message-ID: <1409382202.5141.36.camel@marge.simpson.net>
Reported-by: Mike Galbraith <umgwanakikbuti@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
---
drivers/gpu/drm/i915/intel_dp.c | 55 +++++++++++++++++++++++++++--------------
1 file changed, 37 insertions(+), 18 deletions(-)
diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index 67cfed6..81d7681 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -3661,24 +3661,12 @@ ironlake_dp_detect(struct intel_dp *intel_dp)
return intel_dp_detect_dpcd(intel_dp);
}
-static enum drm_connector_status
-g4x_dp_detect(struct intel_dp *intel_dp)
+static int g4x_digital_port_connected(struct drm_device *dev,
+ struct intel_digital_port *intel_dig_port)
{
- struct drm_device *dev = intel_dp_to_dev(intel_dp);
struct drm_i915_private *dev_priv = dev->dev_private;
- struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
uint32_t bit;
- /* Can't disconnect eDP, but you can close the lid... */
- if (is_edp(intel_dp)) {
- enum drm_connector_status status;
-
- status = intel_panel_detect(dev);
- if (status == connector_status_unknown)
- status = connector_status_connected;
- return status;
- }
-
if (IS_VALLEYVIEW(dev)) {
switch (intel_dig_port->port) {
case PORT_B:
@@ -3691,7 +3679,7 @@ g4x_dp_detect(struct intel_dp *intel_dp)
bit = PORTD_HOTPLUG_LIVE_STATUS_VLV;
break;
default:
- return connector_status_unknown;
+ return -EINVAL;
}
} else {
switch (intel_dig_port->port) {
@@ -3705,11 +3693,36 @@ g4x_dp_detect(struct intel_dp *intel_dp)
bit = PORTD_HOTPLUG_LIVE_STATUS_G4X;
break;
default:
- return connector_status_unknown;
+ return -EINVAL;
}
}
if ((I915_READ(PORT_HOTPLUG_STAT) & bit) == 0)
+ return 0;
+ return 1;
+}
+
+static enum drm_connector_status
+g4x_dp_detect(struct intel_dp *intel_dp)
+{
+ struct drm_device *dev = intel_dp_to_dev(intel_dp);
+ struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+ int ret;
+
+ /* Can't disconnect eDP, but you can close the lid... */
+ if (is_edp(intel_dp)) {
+ enum drm_connector_status status;
+
+ status = intel_panel_detect(dev);
+ if (status == connector_status_unknown)
+ status = connector_status_connected;
+ return status;
+ }
+
+ ret = g4x_digital_port_connected(dev, intel_dig_port);
+ if (ret == -EINVAL)
+ return connector_status_unknown;
+ else if (ret == 0)
return connector_status_disconnected;
return intel_dp_detect_dpcd(intel_dp);
@@ -4066,8 +4079,14 @@ intel_dp_hpd_pulse(struct intel_digital_port *intel_dig_port, bool long_hpd)
intel_display_power_get(dev_priv, power_domain);
if (long_hpd) {
- if (!ibx_digital_port_connected(dev_priv, intel_dig_port))
- goto mst_fail;
+
+ if (HAS_PCH_SPLIT(dev)) {
+ if (!ibx_digital_port_connected(dev_priv, intel_dig_port))
+ goto mst_fail;
+ } else {
+ if (g4x_digital_port_connected(dev, intel_dig_port) != 1)
+ goto mst_fail;
+ }
if (!intel_dp_get_dpcd(intel_dp)) {
goto mst_fail;
--
1.9.3
^ permalink raw reply related [flat|nested] 14+ messages in thread
* Re: [drm/i915/3.17] panic in i915_digport_work_func
@ 2014-09-01 7:02 ` Dave Airlie
0 siblings, 0 replies; 14+ messages in thread
From: Dave Airlie @ 2014-09-01 7:02 UTC (permalink / raw)
To: Mike Galbraith; +Cc: Daniel Vetter, intel-gfx, LKML, Dave Airlie
[-- Attachment #1: Type: text/plain, Size: 184 bytes --]
> drm/i915: add DP 1.2 MST support (v0.7)
>
> This adds DP 1.2 MST support on Haswell systems.
>
I've attached a patch that might fix this, please test and let me know.
Dave.
[-- Attachment #2: 0001-drm-i915-handle-G45-GM45-pulse-detection-connected-s.patch --]
[-- Type: text/x-patch, Size: 3664 bytes --]
From d407c946fbf5c48f30160591f5bd71fbe158aeb4 Mon Sep 17 00:00:00 2001
From: Dave Airlie <airlied@redhat.com>
Date: Mon, 1 Sep 2014 16:58:12 +1000
Subject: [PATCH] drm/i915: handle G45/GM45 pulse detection connected state.
In the HPD pulse handler we check for long pulses if the port is actually
connected, however we do that for IBX, but we use the pulse handling code on
GM45 systems as well, so we need to use a diffent check.
This patch refactors the digital port connected check out of the g4x detection
path and reuses it in the hpd pulse path.
Should fix:
Message-ID: <1409382202.5141.36.camel@marge.simpson.net>
Reported-by: Mike Galbraith <umgwanakikbuti@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
---
drivers/gpu/drm/i915/intel_dp.c | 55 +++++++++++++++++++++++++++--------------
1 file changed, 37 insertions(+), 18 deletions(-)
diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index 67cfed6..81d7681 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -3661,24 +3661,12 @@ ironlake_dp_detect(struct intel_dp *intel_dp)
return intel_dp_detect_dpcd(intel_dp);
}
-static enum drm_connector_status
-g4x_dp_detect(struct intel_dp *intel_dp)
+static int g4x_digital_port_connected(struct drm_device *dev,
+ struct intel_digital_port *intel_dig_port)
{
- struct drm_device *dev = intel_dp_to_dev(intel_dp);
struct drm_i915_private *dev_priv = dev->dev_private;
- struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
uint32_t bit;
- /* Can't disconnect eDP, but you can close the lid... */
- if (is_edp(intel_dp)) {
- enum drm_connector_status status;
-
- status = intel_panel_detect(dev);
- if (status == connector_status_unknown)
- status = connector_status_connected;
- return status;
- }
-
if (IS_VALLEYVIEW(dev)) {
switch (intel_dig_port->port) {
case PORT_B:
@@ -3691,7 +3679,7 @@ g4x_dp_detect(struct intel_dp *intel_dp)
bit = PORTD_HOTPLUG_LIVE_STATUS_VLV;
break;
default:
- return connector_status_unknown;
+ return -EINVAL;
}
} else {
switch (intel_dig_port->port) {
@@ -3705,11 +3693,36 @@ g4x_dp_detect(struct intel_dp *intel_dp)
bit = PORTD_HOTPLUG_LIVE_STATUS_G4X;
break;
default:
- return connector_status_unknown;
+ return -EINVAL;
}
}
if ((I915_READ(PORT_HOTPLUG_STAT) & bit) == 0)
+ return 0;
+ return 1;
+}
+
+static enum drm_connector_status
+g4x_dp_detect(struct intel_dp *intel_dp)
+{
+ struct drm_device *dev = intel_dp_to_dev(intel_dp);
+ struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+ int ret;
+
+ /* Can't disconnect eDP, but you can close the lid... */
+ if (is_edp(intel_dp)) {
+ enum drm_connector_status status;
+
+ status = intel_panel_detect(dev);
+ if (status == connector_status_unknown)
+ status = connector_status_connected;
+ return status;
+ }
+
+ ret = g4x_digital_port_connected(dev, intel_dig_port);
+ if (ret == -EINVAL)
+ return connector_status_unknown;
+ else if (ret == 0)
return connector_status_disconnected;
return intel_dp_detect_dpcd(intel_dp);
@@ -4066,8 +4079,14 @@ intel_dp_hpd_pulse(struct intel_digital_port *intel_dig_port, bool long_hpd)
intel_display_power_get(dev_priv, power_domain);
if (long_hpd) {
- if (!ibx_digital_port_connected(dev_priv, intel_dig_port))
- goto mst_fail;
+
+ if (HAS_PCH_SPLIT(dev)) {
+ if (!ibx_digital_port_connected(dev_priv, intel_dig_port))
+ goto mst_fail;
+ } else {
+ if (g4x_digital_port_connected(dev, intel_dig_port) != 1)
+ goto mst_fail;
+ }
if (!intel_dp_get_dpcd(intel_dp)) {
goto mst_fail;
--
1.9.3
[-- Attachment #3: Type: text/plain, Size: 159 bytes --]
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply related [flat|nested] 14+ messages in thread
* Re: [drm/i915/3.17] panic in i915_digport_work_func
2014-09-01 7:02 ` Dave Airlie
@ 2014-09-01 7:52 ` Mike Galbraith
-1 siblings, 0 replies; 14+ messages in thread
From: Mike Galbraith @ 2014-09-01 7:52 UTC (permalink / raw)
To: Dave Airlie; +Cc: LKML, Daniel Vetter, Jani Nikula, intel-gfx, Dave Airlie
On Mon, 2014-09-01 at 17:02 +1000, Dave Airlie wrote:
> > drm/i915: add DP 1.2 MST support (v0.7)
> >
> > This adds DP 1.2 MST support on Haswell systems.
> >
> I've attached a patch that might fix this, please test and let me know.
Lappy hasn't exploded in 20 boot cycles, which judging by previous
behavior should mean all is well (knocks anti-deterministic wood).
Thanks.
-Mike
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [drm/i915/3.17] panic in i915_digport_work_func
@ 2014-09-01 7:52 ` Mike Galbraith
0 siblings, 0 replies; 14+ messages in thread
From: Mike Galbraith @ 2014-09-01 7:52 UTC (permalink / raw)
To: Dave Airlie; +Cc: Daniel Vetter, intel-gfx, LKML, Dave Airlie
On Mon, 2014-09-01 at 17:02 +1000, Dave Airlie wrote:
> > drm/i915: add DP 1.2 MST support (v0.7)
> >
> > This adds DP 1.2 MST support on Haswell systems.
> >
> I've attached a patch that might fix this, please test and let me know.
Lappy hasn't exploded in 20 boot cycles, which judging by previous
behavior should mean all is well (knocks anti-deterministic wood).
Thanks.
-Mike
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [drm/i915/3.17] panic in i915_digport_work_func
2014-09-01 7:02 ` Dave Airlie
@ 2014-09-01 9:14 ` Daniel Vetter
-1 siblings, 0 replies; 14+ messages in thread
From: Daniel Vetter @ 2014-09-01 9:14 UTC (permalink / raw)
To: Dave Airlie
Cc: Mike Galbraith, LKML, Daniel Vetter, Jani Nikula, intel-gfx, Dave Airlie
On Mon, Sep 01, 2014 at 05:02:51PM +1000, Dave Airlie wrote:
> > drm/i915: add DP 1.2 MST support (v0.7)
> >
> > This adds DP 1.2 MST support on Haswell systems.
> >
> I've attached a patch that might fix this, please test and let me know.
>
> Dave.
> From d407c946fbf5c48f30160591f5bd71fbe158aeb4 Mon Sep 17 00:00:00 2001
> From: Dave Airlie <airlied@redhat.com>
> Date: Mon, 1 Sep 2014 16:58:12 +1000
> Subject: [PATCH] drm/i915: handle G45/GM45 pulse detection connected state.
>
> In the HPD pulse handler we check for long pulses if the port is actually
> connected, however we do that for IBX, but we use the pulse handling code on
> GM45 systems as well, so we need to use a diffent check.
>
> This patch refactors the digital port connected check out of the g4x detection
> path and reuses it in the hpd pulse path.
>
> Should fix:
> Message-ID: <1409382202.5141.36.camel@marge.simpson.net>
>
> Reported-by: Mike Galbraith <umgwanakikbuti@gmail.com>
> Signed-off-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
> ---
> drivers/gpu/drm/i915/intel_dp.c | 55 +++++++++++++++++++++++++++--------------
> 1 file changed, 37 insertions(+), 18 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
> index 67cfed6..81d7681 100644
> --- a/drivers/gpu/drm/i915/intel_dp.c
> +++ b/drivers/gpu/drm/i915/intel_dp.c
> @@ -3661,24 +3661,12 @@ ironlake_dp_detect(struct intel_dp *intel_dp)
> return intel_dp_detect_dpcd(intel_dp);
> }
>
> -static enum drm_connector_status
> -g4x_dp_detect(struct intel_dp *intel_dp)
> +static int g4x_digital_port_connected(struct drm_device *dev,
> + struct intel_digital_port *intel_dig_port)
> {
> - struct drm_device *dev = intel_dp_to_dev(intel_dp);
> struct drm_i915_private *dev_priv = dev->dev_private;
> - struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
> uint32_t bit;
>
> - /* Can't disconnect eDP, but you can close the lid... */
> - if (is_edp(intel_dp)) {
> - enum drm_connector_status status;
> -
> - status = intel_panel_detect(dev);
> - if (status == connector_status_unknown)
> - status = connector_status_connected;
> - return status;
> - }
> -
> if (IS_VALLEYVIEW(dev)) {
> switch (intel_dig_port->port) {
> case PORT_B:
> @@ -3691,7 +3679,7 @@ g4x_dp_detect(struct intel_dp *intel_dp)
> bit = PORTD_HOTPLUG_LIVE_STATUS_VLV;
> break;
> default:
> - return connector_status_unknown;
> + return -EINVAL;
> }
> } else {
> switch (intel_dig_port->port) {
> @@ -3705,11 +3693,36 @@ g4x_dp_detect(struct intel_dp *intel_dp)
> bit = PORTD_HOTPLUG_LIVE_STATUS_G4X;
> break;
> default:
> - return connector_status_unknown;
> + return -EINVAL;
> }
> }
>
> if ((I915_READ(PORT_HOTPLUG_STAT) & bit) == 0)
> + return 0;
> + return 1;
> +}
> +
> +static enum drm_connector_status
> +g4x_dp_detect(struct intel_dp *intel_dp)
> +{
> + struct drm_device *dev = intel_dp_to_dev(intel_dp);
> + struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
> + int ret;
> +
> + /* Can't disconnect eDP, but you can close the lid... */
> + if (is_edp(intel_dp)) {
> + enum drm_connector_status status;
> +
> + status = intel_panel_detect(dev);
> + if (status == connector_status_unknown)
> + status = connector_status_connected;
> + return status;
> + }
> +
> + ret = g4x_digital_port_connected(dev, intel_dig_port);
> + if (ret == -EINVAL)
> + return connector_status_unknown;
> + else if (ret == 0)
> return connector_status_disconnected;
>
> return intel_dp_detect_dpcd(intel_dp);
> @@ -4066,8 +4079,14 @@ intel_dp_hpd_pulse(struct intel_digital_port *intel_dig_port, bool long_hpd)
> intel_display_power_get(dev_priv, power_domain);
>
> if (long_hpd) {
> - if (!ibx_digital_port_connected(dev_priv, intel_dig_port))
> - goto mst_fail;
> +
> + if (HAS_PCH_SPLIT(dev)) {
> + if (!ibx_digital_port_connected(dev_priv, intel_dig_port))
> + goto mst_fail;
> + } else {
> + if (g4x_digital_port_connected(dev, intel_dig_port) != 1)
> + goto mst_fail;
> + }
>
> if (!intel_dp_get_dpcd(intel_dp)) {
> goto mst_fail;
> --
> 1.9.3
>
--
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [drm/i915/3.17] panic in i915_digport_work_func
@ 2014-09-01 9:14 ` Daniel Vetter
0 siblings, 0 replies; 14+ messages in thread
From: Daniel Vetter @ 2014-09-01 9:14 UTC (permalink / raw)
To: Dave Airlie; +Cc: Daniel Vetter, intel-gfx, LKML, Mike Galbraith, Dave Airlie
On Mon, Sep 01, 2014 at 05:02:51PM +1000, Dave Airlie wrote:
> > drm/i915: add DP 1.2 MST support (v0.7)
> >
> > This adds DP 1.2 MST support on Haswell systems.
> >
> I've attached a patch that might fix this, please test and let me know.
>
> Dave.
> From d407c946fbf5c48f30160591f5bd71fbe158aeb4 Mon Sep 17 00:00:00 2001
> From: Dave Airlie <airlied@redhat.com>
> Date: Mon, 1 Sep 2014 16:58:12 +1000
> Subject: [PATCH] drm/i915: handle G45/GM45 pulse detection connected state.
>
> In the HPD pulse handler we check for long pulses if the port is actually
> connected, however we do that for IBX, but we use the pulse handling code on
> GM45 systems as well, so we need to use a diffent check.
>
> This patch refactors the digital port connected check out of the g4x detection
> path and reuses it in the hpd pulse path.
>
> Should fix:
> Message-ID: <1409382202.5141.36.camel@marge.simpson.net>
>
> Reported-by: Mike Galbraith <umgwanakikbuti@gmail.com>
> Signed-off-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
> ---
> drivers/gpu/drm/i915/intel_dp.c | 55 +++++++++++++++++++++++++++--------------
> 1 file changed, 37 insertions(+), 18 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
> index 67cfed6..81d7681 100644
> --- a/drivers/gpu/drm/i915/intel_dp.c
> +++ b/drivers/gpu/drm/i915/intel_dp.c
> @@ -3661,24 +3661,12 @@ ironlake_dp_detect(struct intel_dp *intel_dp)
> return intel_dp_detect_dpcd(intel_dp);
> }
>
> -static enum drm_connector_status
> -g4x_dp_detect(struct intel_dp *intel_dp)
> +static int g4x_digital_port_connected(struct drm_device *dev,
> + struct intel_digital_port *intel_dig_port)
> {
> - struct drm_device *dev = intel_dp_to_dev(intel_dp);
> struct drm_i915_private *dev_priv = dev->dev_private;
> - struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
> uint32_t bit;
>
> - /* Can't disconnect eDP, but you can close the lid... */
> - if (is_edp(intel_dp)) {
> - enum drm_connector_status status;
> -
> - status = intel_panel_detect(dev);
> - if (status == connector_status_unknown)
> - status = connector_status_connected;
> - return status;
> - }
> -
> if (IS_VALLEYVIEW(dev)) {
> switch (intel_dig_port->port) {
> case PORT_B:
> @@ -3691,7 +3679,7 @@ g4x_dp_detect(struct intel_dp *intel_dp)
> bit = PORTD_HOTPLUG_LIVE_STATUS_VLV;
> break;
> default:
> - return connector_status_unknown;
> + return -EINVAL;
> }
> } else {
> switch (intel_dig_port->port) {
> @@ -3705,11 +3693,36 @@ g4x_dp_detect(struct intel_dp *intel_dp)
> bit = PORTD_HOTPLUG_LIVE_STATUS_G4X;
> break;
> default:
> - return connector_status_unknown;
> + return -EINVAL;
> }
> }
>
> if ((I915_READ(PORT_HOTPLUG_STAT) & bit) == 0)
> + return 0;
> + return 1;
> +}
> +
> +static enum drm_connector_status
> +g4x_dp_detect(struct intel_dp *intel_dp)
> +{
> + struct drm_device *dev = intel_dp_to_dev(intel_dp);
> + struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
> + int ret;
> +
> + /* Can't disconnect eDP, but you can close the lid... */
> + if (is_edp(intel_dp)) {
> + enum drm_connector_status status;
> +
> + status = intel_panel_detect(dev);
> + if (status == connector_status_unknown)
> + status = connector_status_connected;
> + return status;
> + }
> +
> + ret = g4x_digital_port_connected(dev, intel_dig_port);
> + if (ret == -EINVAL)
> + return connector_status_unknown;
> + else if (ret == 0)
> return connector_status_disconnected;
>
> return intel_dp_detect_dpcd(intel_dp);
> @@ -4066,8 +4079,14 @@ intel_dp_hpd_pulse(struct intel_digital_port *intel_dig_port, bool long_hpd)
> intel_display_power_get(dev_priv, power_domain);
>
> if (long_hpd) {
> - if (!ibx_digital_port_connected(dev_priv, intel_dig_port))
> - goto mst_fail;
> +
> + if (HAS_PCH_SPLIT(dev)) {
> + if (!ibx_digital_port_connected(dev_priv, intel_dig_port))
> + goto mst_fail;
> + } else {
> + if (g4x_digital_port_connected(dev, intel_dig_port) != 1)
> + goto mst_fail;
> + }
>
> if (!intel_dp_get_dpcd(intel_dp)) {
> goto mst_fail;
> --
> 1.9.3
>
--
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [Intel-gfx] [drm/i915/3.17] panic in i915_digport_work_func
2014-09-01 7:02 ` Dave Airlie
@ 2014-09-01 13:28 ` Imre Deak
-1 siblings, 0 replies; 14+ messages in thread
From: Imre Deak @ 2014-09-01 13:28 UTC (permalink / raw)
To: Dave Airlie
Cc: Mike Galbraith, Daniel Vetter, intel-gfx, LKML, Dave Airlie,
Nikula, Jani
[-- Attachment #1: Type: text/plain, Size: 5117 bytes --]
On Mon, 2014-09-01 at 17:02 +1000, Dave Airlie wrote:
> From: Dave Airlie <airlied@redhat.com>
> Date: Mon, 1 Sep 2014 16:58:12 +1000
> Subject: [PATCH] drm/i915: handle G45/GM45 pulse detection connected
> state.
>
> In the HPD pulse handler we check for long pulses if the port is
> actually
> connected, however we do that for IBX, but we use the pulse handling
> code on
> GM45 systems as well, so we need to use a diffent check.
>
> This patch refactors the digital port connected check out of the g4x
> detection
> path and reuses it in the hpd pulse path.
>
> Should fix:
> Message-ID: <1409382202.5141.36.camel@marge.simpson.net>
>
> Reported-by: Mike Galbraith <umgwanakikbuti@gmail.com>
> Signed-off-by: Dave Airlie <airlied@redhat.com>
Daniel reviewed this already, but Jani asked me to take a look, so:
Acked-by: Imre Deak <imre.deak@intel.com>
One thing for the future is to move ibx_digital_port_connected() to
intel_dp.c too and make its return value match that of
g4x_digital_port_connected().
--Imre
> ---
> drivers/gpu/drm/i915/intel_dp.c | 55
> +++++++++++++++++++++++++++--------------
> 1 file changed, 37 insertions(+), 18 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/intel_dp.c
> b/drivers/gpu/drm/i915/intel_dp.c
> index 67cfed6..81d7681 100644
> --- a/drivers/gpu/drm/i915/intel_dp.c
> +++ b/drivers/gpu/drm/i915/intel_dp.c
> @@ -3661,24 +3661,12 @@ ironlake_dp_detect(struct intel_dp *intel_dp)
> return intel_dp_detect_dpcd(intel_dp);
> }
>
> -static enum drm_connector_status
> -g4x_dp_detect(struct intel_dp *intel_dp)
> +static int g4x_digital_port_connected(struct drm_device *dev,
> + struct intel_digital_port
> *intel_dig_port)
> {
> - struct drm_device *dev = intel_dp_to_dev(intel_dp);
> struct drm_i915_private *dev_priv = dev->dev_private;
> - struct intel_digital_port *intel_dig_port =
> dp_to_dig_port(intel_dp);
> uint32_t bit;
>
> - /* Can't disconnect eDP, but you can close the lid... */
> - if (is_edp(intel_dp)) {
> - enum drm_connector_status status;
> -
> - status = intel_panel_detect(dev);
> - if (status == connector_status_unknown)
> - status = connector_status_connected;
> - return status;
> - }
> -
> if (IS_VALLEYVIEW(dev)) {
> switch (intel_dig_port->port) {
> case PORT_B:
> @@ -3691,7 +3679,7 @@ g4x_dp_detect(struct intel_dp *intel_dp)
> bit = PORTD_HOTPLUG_LIVE_STATUS_VLV;
> break;
> default:
> - return connector_status_unknown;
> + return -EINVAL;
> }
> } else {
> switch (intel_dig_port->port) {
> @@ -3705,11 +3693,36 @@ g4x_dp_detect(struct intel_dp *intel_dp)
> bit = PORTD_HOTPLUG_LIVE_STATUS_G4X;
> break;
> default:
> - return connector_status_unknown;
> + return -EINVAL;
> }
> }
>
> if ((I915_READ(PORT_HOTPLUG_STAT) & bit) == 0)
> + return 0;
> + return 1;
> +}
> +
> +static enum drm_connector_status
> +g4x_dp_detect(struct intel_dp *intel_dp)
> +{
> + struct drm_device *dev = intel_dp_to_dev(intel_dp);
> + struct intel_digital_port *intel_dig_port =
> dp_to_dig_port(intel_dp);
> + int ret;
> +
> + /* Can't disconnect eDP, but you can close the lid... */
> + if (is_edp(intel_dp)) {
> + enum drm_connector_status status;
> +
> + status = intel_panel_detect(dev);
> + if (status == connector_status_unknown)
> + status = connector_status_connected;
> + return status;
> + }
> +
> + ret = g4x_digital_port_connected(dev, intel_dig_port);
> + if (ret == -EINVAL)
> + return connector_status_unknown;
> + else if (ret == 0)
> return connector_status_disconnected;
>
> return intel_dp_detect_dpcd(intel_dp);
> @@ -4066,8 +4079,14 @@ intel_dp_hpd_pulse(struct intel_digital_port
> *intel_dig_port, bool long_hpd)
> intel_display_power_get(dev_priv, power_domain);
>
> if (long_hpd) {
> - if (!ibx_digital_port_connected(dev_priv,
> intel_dig_port))
> - goto mst_fail;
> +
> + if (HAS_PCH_SPLIT(dev)) {
> + if (!ibx_digital_port_connected(dev_priv,
> intel_dig_port))
> + goto mst_fail;
> + } else {
> + if (g4x_digital_port_connected(dev,
> intel_dig_port) != 1)
> + goto mst_fail;
> + }
>
> if (!intel_dp_get_dpcd(intel_dp)) {
> goto mst_fail;
[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 490 bytes --]
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [drm/i915/3.17] panic in i915_digport_work_func
@ 2014-09-01 13:28 ` Imre Deak
0 siblings, 0 replies; 14+ messages in thread
From: Imre Deak @ 2014-09-01 13:28 UTC (permalink / raw)
To: Dave Airlie
Cc: Nikula, Jani, Daniel Vetter, intel-gfx, LKML, Mike Galbraith,
Dave Airlie
[-- Attachment #1.1: Type: text/plain, Size: 5117 bytes --]
On Mon, 2014-09-01 at 17:02 +1000, Dave Airlie wrote:
> From: Dave Airlie <airlied@redhat.com>
> Date: Mon, 1 Sep 2014 16:58:12 +1000
> Subject: [PATCH] drm/i915: handle G45/GM45 pulse detection connected
> state.
>
> In the HPD pulse handler we check for long pulses if the port is
> actually
> connected, however we do that for IBX, but we use the pulse handling
> code on
> GM45 systems as well, so we need to use a diffent check.
>
> This patch refactors the digital port connected check out of the g4x
> detection
> path and reuses it in the hpd pulse path.
>
> Should fix:
> Message-ID: <1409382202.5141.36.camel@marge.simpson.net>
>
> Reported-by: Mike Galbraith <umgwanakikbuti@gmail.com>
> Signed-off-by: Dave Airlie <airlied@redhat.com>
Daniel reviewed this already, but Jani asked me to take a look, so:
Acked-by: Imre Deak <imre.deak@intel.com>
One thing for the future is to move ibx_digital_port_connected() to
intel_dp.c too and make its return value match that of
g4x_digital_port_connected().
--Imre
> ---
> drivers/gpu/drm/i915/intel_dp.c | 55
> +++++++++++++++++++++++++++--------------
> 1 file changed, 37 insertions(+), 18 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/intel_dp.c
> b/drivers/gpu/drm/i915/intel_dp.c
> index 67cfed6..81d7681 100644
> --- a/drivers/gpu/drm/i915/intel_dp.c
> +++ b/drivers/gpu/drm/i915/intel_dp.c
> @@ -3661,24 +3661,12 @@ ironlake_dp_detect(struct intel_dp *intel_dp)
> return intel_dp_detect_dpcd(intel_dp);
> }
>
> -static enum drm_connector_status
> -g4x_dp_detect(struct intel_dp *intel_dp)
> +static int g4x_digital_port_connected(struct drm_device *dev,
> + struct intel_digital_port
> *intel_dig_port)
> {
> - struct drm_device *dev = intel_dp_to_dev(intel_dp);
> struct drm_i915_private *dev_priv = dev->dev_private;
> - struct intel_digital_port *intel_dig_port =
> dp_to_dig_port(intel_dp);
> uint32_t bit;
>
> - /* Can't disconnect eDP, but you can close the lid... */
> - if (is_edp(intel_dp)) {
> - enum drm_connector_status status;
> -
> - status = intel_panel_detect(dev);
> - if (status == connector_status_unknown)
> - status = connector_status_connected;
> - return status;
> - }
> -
> if (IS_VALLEYVIEW(dev)) {
> switch (intel_dig_port->port) {
> case PORT_B:
> @@ -3691,7 +3679,7 @@ g4x_dp_detect(struct intel_dp *intel_dp)
> bit = PORTD_HOTPLUG_LIVE_STATUS_VLV;
> break;
> default:
> - return connector_status_unknown;
> + return -EINVAL;
> }
> } else {
> switch (intel_dig_port->port) {
> @@ -3705,11 +3693,36 @@ g4x_dp_detect(struct intel_dp *intel_dp)
> bit = PORTD_HOTPLUG_LIVE_STATUS_G4X;
> break;
> default:
> - return connector_status_unknown;
> + return -EINVAL;
> }
> }
>
> if ((I915_READ(PORT_HOTPLUG_STAT) & bit) == 0)
> + return 0;
> + return 1;
> +}
> +
> +static enum drm_connector_status
> +g4x_dp_detect(struct intel_dp *intel_dp)
> +{
> + struct drm_device *dev = intel_dp_to_dev(intel_dp);
> + struct intel_digital_port *intel_dig_port =
> dp_to_dig_port(intel_dp);
> + int ret;
> +
> + /* Can't disconnect eDP, but you can close the lid... */
> + if (is_edp(intel_dp)) {
> + enum drm_connector_status status;
> +
> + status = intel_panel_detect(dev);
> + if (status == connector_status_unknown)
> + status = connector_status_connected;
> + return status;
> + }
> +
> + ret = g4x_digital_port_connected(dev, intel_dig_port);
> + if (ret == -EINVAL)
> + return connector_status_unknown;
> + else if (ret == 0)
> return connector_status_disconnected;
>
> return intel_dp_detect_dpcd(intel_dp);
> @@ -4066,8 +4079,14 @@ intel_dp_hpd_pulse(struct intel_digital_port
> *intel_dig_port, bool long_hpd)
> intel_display_power_get(dev_priv, power_domain);
>
> if (long_hpd) {
> - if (!ibx_digital_port_connected(dev_priv,
> intel_dig_port))
> - goto mst_fail;
> +
> + if (HAS_PCH_SPLIT(dev)) {
> + if (!ibx_digital_port_connected(dev_priv,
> intel_dig_port))
> + goto mst_fail;
> + } else {
> + if (g4x_digital_port_connected(dev,
> intel_dig_port) != 1)
> + goto mst_fail;
> + }
>
> if (!intel_dp_get_dpcd(intel_dp)) {
> goto mst_fail;
[-- Attachment #1.2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 490 bytes --]
[-- Attachment #2: Type: text/plain, Size: 159 bytes --]
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [Intel-gfx] [drm/i915/3.17] panic in i915_digport_work_func
2014-09-01 13:28 ` Imre Deak
@ 2014-09-01 14:11 ` Jani Nikula
-1 siblings, 0 replies; 14+ messages in thread
From: Jani Nikula @ 2014-09-01 14:11 UTC (permalink / raw)
To: imre.deak, Dave Airlie
Cc: Mike Galbraith, Daniel Vetter, intel-gfx, LKML, Dave Airlie
On Mon, 01 Sep 2014, Imre Deak <imre.deak@intel.com> wrote:
> On Mon, 2014-09-01 at 17:02 +1000, Dave Airlie wrote:
>> From: Dave Airlie <airlied@redhat.com>
>> Date: Mon, 1 Sep 2014 16:58:12 +1000
>> Subject: [PATCH] drm/i915: handle G45/GM45 pulse detection connected
>> state.
>>
>> In the HPD pulse handler we check for long pulses if the port is
>> actually
>> connected, however we do that for IBX, but we use the pulse handling
>> code on
>> GM45 systems as well, so we need to use a diffent check.
>>
>> This patch refactors the digital port connected check out of the g4x
>> detection
>> path and reuses it in the hpd pulse path.
>>
>> Should fix:
>> Message-ID: <1409382202.5141.36.camel@marge.simpson.net>
>>
>> Reported-by: Mike Galbraith <umgwanakikbuti@gmail.com>
>> Signed-off-by: Dave Airlie <airlied@redhat.com>
>
> Daniel reviewed this already, but Jani asked me to take a look, so:
> Acked-by: Imre Deak <imre.deak@intel.com>
Pushed to drm-intel-fixes, thanks for the patch, review, and testing.
BR,
Jani.
>
> One thing for the future is to move ibx_digital_port_connected() to
> intel_dp.c too and make its return value match that of
> g4x_digital_port_connected().
>
> --Imre
>
>> ---
>> drivers/gpu/drm/i915/intel_dp.c | 55
>> +++++++++++++++++++++++++++--------------
>> 1 file changed, 37 insertions(+), 18 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/i915/intel_dp.c
>> b/drivers/gpu/drm/i915/intel_dp.c
>> index 67cfed6..81d7681 100644
>> --- a/drivers/gpu/drm/i915/intel_dp.c
>> +++ b/drivers/gpu/drm/i915/intel_dp.c
>> @@ -3661,24 +3661,12 @@ ironlake_dp_detect(struct intel_dp *intel_dp)
>> return intel_dp_detect_dpcd(intel_dp);
>> }
>>
>> -static enum drm_connector_status
>> -g4x_dp_detect(struct intel_dp *intel_dp)
>> +static int g4x_digital_port_connected(struct drm_device *dev,
>> + struct intel_digital_port
>> *intel_dig_port)
>> {
>> - struct drm_device *dev = intel_dp_to_dev(intel_dp);
>> struct drm_i915_private *dev_priv = dev->dev_private;
>> - struct intel_digital_port *intel_dig_port =
>> dp_to_dig_port(intel_dp);
>> uint32_t bit;
>>
>> - /* Can't disconnect eDP, but you can close the lid... */
>> - if (is_edp(intel_dp)) {
>> - enum drm_connector_status status;
>> -
>> - status = intel_panel_detect(dev);
>> - if (status == connector_status_unknown)
>> - status = connector_status_connected;
>> - return status;
>> - }
>> -
>> if (IS_VALLEYVIEW(dev)) {
>> switch (intel_dig_port->port) {
>> case PORT_B:
>> @@ -3691,7 +3679,7 @@ g4x_dp_detect(struct intel_dp *intel_dp)
>> bit = PORTD_HOTPLUG_LIVE_STATUS_VLV;
>> break;
>> default:
>> - return connector_status_unknown;
>> + return -EINVAL;
>> }
>> } else {
>> switch (intel_dig_port->port) {
>> @@ -3705,11 +3693,36 @@ g4x_dp_detect(struct intel_dp *intel_dp)
>> bit = PORTD_HOTPLUG_LIVE_STATUS_G4X;
>> break;
>> default:
>> - return connector_status_unknown;
>> + return -EINVAL;
>> }
>> }
>>
>> if ((I915_READ(PORT_HOTPLUG_STAT) & bit) == 0)
>> + return 0;
>> + return 1;
>> +}
>> +
>> +static enum drm_connector_status
>> +g4x_dp_detect(struct intel_dp *intel_dp)
>> +{
>> + struct drm_device *dev = intel_dp_to_dev(intel_dp);
>> + struct intel_digital_port *intel_dig_port =
>> dp_to_dig_port(intel_dp);
>> + int ret;
>> +
>> + /* Can't disconnect eDP, but you can close the lid... */
>> + if (is_edp(intel_dp)) {
>> + enum drm_connector_status status;
>> +
>> + status = intel_panel_detect(dev);
>> + if (status == connector_status_unknown)
>> + status = connector_status_connected;
>> + return status;
>> + }
>> +
>> + ret = g4x_digital_port_connected(dev, intel_dig_port);
>> + if (ret == -EINVAL)
>> + return connector_status_unknown;
>> + else if (ret == 0)
>> return connector_status_disconnected;
>>
>> return intel_dp_detect_dpcd(intel_dp);
>> @@ -4066,8 +4079,14 @@ intel_dp_hpd_pulse(struct intel_digital_port
>> *intel_dig_port, bool long_hpd)
>> intel_display_power_get(dev_priv, power_domain);
>>
>> if (long_hpd) {
>> - if (!ibx_digital_port_connected(dev_priv,
>> intel_dig_port))
>> - goto mst_fail;
>> +
>> + if (HAS_PCH_SPLIT(dev)) {
>> + if (!ibx_digital_port_connected(dev_priv,
>> intel_dig_port))
>> + goto mst_fail;
>> + } else {
>> + if (g4x_digital_port_connected(dev,
>> intel_dig_port) != 1)
>> + goto mst_fail;
>> + }
>>
>> if (!intel_dp_get_dpcd(intel_dp)) {
>> goto mst_fail;
>
--
Jani Nikula, Intel Open Source Technology Center
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [drm/i915/3.17] panic in i915_digport_work_func
@ 2014-09-01 14:11 ` Jani Nikula
0 siblings, 0 replies; 14+ messages in thread
From: Jani Nikula @ 2014-09-01 14:11 UTC (permalink / raw)
To: imre.deak, Dave Airlie
Cc: Daniel Vetter, intel-gfx, Mike Galbraith, LKML, Dave Airlie
On Mon, 01 Sep 2014, Imre Deak <imre.deak@intel.com> wrote:
> On Mon, 2014-09-01 at 17:02 +1000, Dave Airlie wrote:
>> From: Dave Airlie <airlied@redhat.com>
>> Date: Mon, 1 Sep 2014 16:58:12 +1000
>> Subject: [PATCH] drm/i915: handle G45/GM45 pulse detection connected
>> state.
>>
>> In the HPD pulse handler we check for long pulses if the port is
>> actually
>> connected, however we do that for IBX, but we use the pulse handling
>> code on
>> GM45 systems as well, so we need to use a diffent check.
>>
>> This patch refactors the digital port connected check out of the g4x
>> detection
>> path and reuses it in the hpd pulse path.
>>
>> Should fix:
>> Message-ID: <1409382202.5141.36.camel@marge.simpson.net>
>>
>> Reported-by: Mike Galbraith <umgwanakikbuti@gmail.com>
>> Signed-off-by: Dave Airlie <airlied@redhat.com>
>
> Daniel reviewed this already, but Jani asked me to take a look, so:
> Acked-by: Imre Deak <imre.deak@intel.com>
Pushed to drm-intel-fixes, thanks for the patch, review, and testing.
BR,
Jani.
>
> One thing for the future is to move ibx_digital_port_connected() to
> intel_dp.c too and make its return value match that of
> g4x_digital_port_connected().
>
> --Imre
>
>> ---
>> drivers/gpu/drm/i915/intel_dp.c | 55
>> +++++++++++++++++++++++++++--------------
>> 1 file changed, 37 insertions(+), 18 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/i915/intel_dp.c
>> b/drivers/gpu/drm/i915/intel_dp.c
>> index 67cfed6..81d7681 100644
>> --- a/drivers/gpu/drm/i915/intel_dp.c
>> +++ b/drivers/gpu/drm/i915/intel_dp.c
>> @@ -3661,24 +3661,12 @@ ironlake_dp_detect(struct intel_dp *intel_dp)
>> return intel_dp_detect_dpcd(intel_dp);
>> }
>>
>> -static enum drm_connector_status
>> -g4x_dp_detect(struct intel_dp *intel_dp)
>> +static int g4x_digital_port_connected(struct drm_device *dev,
>> + struct intel_digital_port
>> *intel_dig_port)
>> {
>> - struct drm_device *dev = intel_dp_to_dev(intel_dp);
>> struct drm_i915_private *dev_priv = dev->dev_private;
>> - struct intel_digital_port *intel_dig_port =
>> dp_to_dig_port(intel_dp);
>> uint32_t bit;
>>
>> - /* Can't disconnect eDP, but you can close the lid... */
>> - if (is_edp(intel_dp)) {
>> - enum drm_connector_status status;
>> -
>> - status = intel_panel_detect(dev);
>> - if (status == connector_status_unknown)
>> - status = connector_status_connected;
>> - return status;
>> - }
>> -
>> if (IS_VALLEYVIEW(dev)) {
>> switch (intel_dig_port->port) {
>> case PORT_B:
>> @@ -3691,7 +3679,7 @@ g4x_dp_detect(struct intel_dp *intel_dp)
>> bit = PORTD_HOTPLUG_LIVE_STATUS_VLV;
>> break;
>> default:
>> - return connector_status_unknown;
>> + return -EINVAL;
>> }
>> } else {
>> switch (intel_dig_port->port) {
>> @@ -3705,11 +3693,36 @@ g4x_dp_detect(struct intel_dp *intel_dp)
>> bit = PORTD_HOTPLUG_LIVE_STATUS_G4X;
>> break;
>> default:
>> - return connector_status_unknown;
>> + return -EINVAL;
>> }
>> }
>>
>> if ((I915_READ(PORT_HOTPLUG_STAT) & bit) == 0)
>> + return 0;
>> + return 1;
>> +}
>> +
>> +static enum drm_connector_status
>> +g4x_dp_detect(struct intel_dp *intel_dp)
>> +{
>> + struct drm_device *dev = intel_dp_to_dev(intel_dp);
>> + struct intel_digital_port *intel_dig_port =
>> dp_to_dig_port(intel_dp);
>> + int ret;
>> +
>> + /* Can't disconnect eDP, but you can close the lid... */
>> + if (is_edp(intel_dp)) {
>> + enum drm_connector_status status;
>> +
>> + status = intel_panel_detect(dev);
>> + if (status == connector_status_unknown)
>> + status = connector_status_connected;
>> + return status;
>> + }
>> +
>> + ret = g4x_digital_port_connected(dev, intel_dig_port);
>> + if (ret == -EINVAL)
>> + return connector_status_unknown;
>> + else if (ret == 0)
>> return connector_status_disconnected;
>>
>> return intel_dp_detect_dpcd(intel_dp);
>> @@ -4066,8 +4079,14 @@ intel_dp_hpd_pulse(struct intel_digital_port
>> *intel_dig_port, bool long_hpd)
>> intel_display_power_get(dev_priv, power_domain);
>>
>> if (long_hpd) {
>> - if (!ibx_digital_port_connected(dev_priv,
>> intel_dig_port))
>> - goto mst_fail;
>> +
>> + if (HAS_PCH_SPLIT(dev)) {
>> + if (!ibx_digital_port_connected(dev_priv,
>> intel_dig_port))
>> + goto mst_fail;
>> + } else {
>> + if (g4x_digital_port_connected(dev,
>> intel_dig_port) != 1)
>> + goto mst_fail;
>> + }
>>
>> if (!intel_dp_get_dpcd(intel_dp)) {
>> goto mst_fail;
>
--
Jani Nikula, Intel Open Source Technology Center
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [drm/i915/3.17] panic in i915_digport_work_func
2014-09-01 9:14 ` Daniel Vetter
@ 2014-09-03 8:08 ` Jiri Kosina
-1 siblings, 0 replies; 14+ messages in thread
From: Jiri Kosina @ 2014-09-03 8:08 UTC (permalink / raw)
To: Daniel Vetter
Cc: Dave Airlie, Mike Galbraith, LKML, Daniel Vetter, Jani Nikula,
intel-gfx, Dave Airlie
On Mon, 1 Sep 2014, Daniel Vetter wrote:
> > From d407c946fbf5c48f30160591f5bd71fbe158aeb4 Mon Sep 17 00:00:00 2001
> > From: Dave Airlie <airlied@redhat.com>
> > Date: Mon, 1 Sep 2014 16:58:12 +1000
> > Subject: [PATCH] drm/i915: handle G45/GM45 pulse detection connected state.
> >
> > In the HPD pulse handler we check for long pulses if the port is actually
> > connected, however we do that for IBX, but we use the pulse handling code on
> > GM45 systems as well, so we need to use a diffent check.
> >
> > This patch refactors the digital port connected check out of the g4x detection
> > path and reuses it in the hpd pulse path.
> >
> > Should fix:
> > Message-ID: <1409382202.5141.36.camel@marge.simpson.net>
> >
> > Reported-by: Mike Galbraith <umgwanakikbuti@gmail.com>
> > Signed-off-by: Dave Airlie <airlied@redhat.com>
>
> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
I was just about to start bisecting, as my thinkpad x200s with Gen4 chip
wouldn't ocassionally boot with -rc3, but this patch seems to be the cure
for me as well.
FWIW
Tested-by: Jiri Kosina <jkosina@suse.cz>
--
Jiri Kosina
SUSE Labs
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [drm/i915/3.17] panic in i915_digport_work_func
@ 2014-09-03 8:08 ` Jiri Kosina
0 siblings, 0 replies; 14+ messages in thread
From: Jiri Kosina @ 2014-09-03 8:08 UTC (permalink / raw)
To: Daniel Vetter; +Cc: Daniel Vetter, intel-gfx, LKML, Mike Galbraith, Dave Airlie
On Mon, 1 Sep 2014, Daniel Vetter wrote:
> > From d407c946fbf5c48f30160591f5bd71fbe158aeb4 Mon Sep 17 00:00:00 2001
> > From: Dave Airlie <airlied@redhat.com>
> > Date: Mon, 1 Sep 2014 16:58:12 +1000
> > Subject: [PATCH] drm/i915: handle G45/GM45 pulse detection connected state.
> >
> > In the HPD pulse handler we check for long pulses if the port is actually
> > connected, however we do that for IBX, but we use the pulse handling code on
> > GM45 systems as well, so we need to use a diffent check.
> >
> > This patch refactors the digital port connected check out of the g4x detection
> > path and reuses it in the hpd pulse path.
> >
> > Should fix:
> > Message-ID: <1409382202.5141.36.camel@marge.simpson.net>
> >
> > Reported-by: Mike Galbraith <umgwanakikbuti@gmail.com>
> > Signed-off-by: Dave Airlie <airlied@redhat.com>
>
> Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
I was just about to start bisecting, as my thinkpad x200s with Gen4 chip
wouldn't ocassionally boot with -rc3, but this patch seems to be the cure
for me as well.
FWIW
Tested-by: Jiri Kosina <jkosina@suse.cz>
--
Jiri Kosina
SUSE Labs
^ permalink raw reply [flat|nested] 14+ messages in thread
end of thread, other threads:[~2014-09-03 8:09 UTC | newest]
Thread overview: 14+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2014-08-30 7:03 [drm/i915/3.17] panic in i915_digport_work_func Mike Galbraith
2014-08-31 12:05 ` Mike Galbraith
2014-09-01 7:02 ` Dave Airlie
2014-09-01 7:02 ` Dave Airlie
2014-09-01 7:52 ` Mike Galbraith
2014-09-01 7:52 ` Mike Galbraith
2014-09-01 9:14 ` Daniel Vetter
2014-09-01 9:14 ` Daniel Vetter
2014-09-03 8:08 ` Jiri Kosina
2014-09-03 8:08 ` Jiri Kosina
2014-09-01 13:28 ` [Intel-gfx] " Imre Deak
2014-09-01 13:28 ` Imre Deak
2014-09-01 14:11 ` [Intel-gfx] " Jani Nikula
2014-09-01 14:11 ` Jani Nikula
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.