Attachment 41770 Details for Bug 29230 – Restore old fb after modesetting failure

[patch] Restore old fb after modesetting failure

0001-drm-Restore-the-old_fb-upon-modeset-failure.patch (text/plain), 3.42 KB, created by Chris Wilson on 2011-01-08 07:16:06 UTC

(hide)

Description:

Filename:

MIME Type:

Creator: Chris Wilson

Created: 2011-01-08 07:16:06 UTC

Size: 3.42 KB

patch

obsolete

>From 21af63db5f584a4cec8b09011457122c6ad88108 Mon Sep 17 00:00:00 2001
>From: Chris Wilson <chris@chris-wilson.co.uk>
>Date: Sat, 8 Jan 2011 15:07:18 +0000
>Subject: [PATCH] drm: Restore the old_fb upon modeset failure
>
>... or else we may end up disabling the wrong framebuffer, leading to an
>OOPS, e.g:
>
>[ 6033.229012] kernel BUG at drivers/gpu/drm/i915/i915_gem.c:3271!
>[ 6033.229012] invalid opcode: 0000 [#1] SMP
>[ 6033.229012] last sysfs file:
>/sys/devices/virtual/backlight/acpi_video0/uevent
>[ 6033.229012] Modules linked in: sunrpc cpufreq_ondemand acpi_cpufreq
>mperf snd_hda_codec_analog snd_hda_intel snd_hda_codec snd_hwdep snd_seq
>snd_seq_device snd_pcm snd_timer thinkpad_acpi ppdev snd r852 sm_common
>iTCO_wdt uvcvideo i2c_i801 iTCO_vendor_support microcode wmi nand
>videodev nand_ids nand_ecc snd_page_alloc parport_pc parport mtd
>soundcore joydev v4l1_compat pcspkr uinput ipv6 sdhci_pci sdhci mmc_core
>yenta_socket i915 drm_kms_helper drm i2c_algo_bit i2c_core video output
>[last unloaded: scsi_wait_scan]
>[ 6033.229012]
>[ 6033.229012] Pid: 4834, comm: Xorg Not tainted 2.6.37-rc8+ #25 7661BL5/7661BL5
>[ 6033.229012] EIP: 0060:[<f86fda5e>] EFLAGS: 00013246 CPU: 0
>[ 6033.229012] EIP is at i915_gem_object_unpin+0x23/0x76 [i915]
>[ 6033.229012] EAX: f68a4000 EBX: f6831f00 ECX: 000600fa EDX: f68a8000
>[ 6033.229012] ESI: f68a4014 EDI: f68a42b8 EBP: f2169c44 ESP: f2169c3c
>[ 6033.229012]  DS: 007b ES: 007b FS: 00d8 GS: 00e0 SS: 0068
>[ 6033.229012] Process Xorg (pid: 4834, ti=f2168000 task=f21c8000 task.ti=f2168000)
>[ 6033.229012] Stack:
>[ 6033.229012]  f3a84800 f68a4014 f2169c54 f87045d8 f3a84800 f872d9a8 f2169c68 f7fd8091
>[ 6033.229012]  f3b952a4 00000000 f68a414c f2169cf0 f7fd9377 00000000 00000000 f7fd98b0
>[ 6033.229012]  f7fd9f4e 0000000f f7f328a0 00000000 00000000 00000000 f2169ca4 f68a414c
>[ 6033.229012] Call Trace:
>[ 6033.229012]  [<f87045d8>] ? intel_crtc_disable+0x36/0x41 [i915]
>[ 6033.229012]  [<f7fd8091>] ?  drm_helper_disable_unused_functions+0xcd/0xf9 [drm_kms_helper]
>[ 6033.229012]  [<f7fd9377>] ? drm_crtc_helper_set_config+0x62a/0x7f7 [drm_kms_helper]
>[ 6033.229012]  [<c04daa10>] ? __slab_free+0x1b/0xa4
>[ 6033.229012]  [<f7fd7e62>] ? drm_fb_helper_initial_config+0x466/0x497 [drm_kms_helper]
>[ 6033.229012]  [<f7fd7ea3>] ? drm_fb_helper_restore+0x10/0x2a [drm_kms_helper]
>[ 6033.229012]  [<f86f2577>] ? i915_driver_lastclose+0x2a/0x57 [i915]
>[ 6033.229012]  [<f7f1989f>] ? drm_lastclose+0x45/0x23e [drm]
>[ 6033.229012]  [<f7f1a0b4>] ? drm_release+0x462/0x4d7 [drm]
>
>Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
>---
> drivers/gpu/drm/drm_crtc_helper.c |    5 ++++-
> 1 files changed, 4 insertions(+), 1 deletions(-)
>
>diff --git a/drivers/gpu/drm/drm_crtc_helper.c b/drivers/gpu/drm/drm_crtc_helper.c
>index 8de333a..952b3d4 100644
>--- a/drivers/gpu/drm/drm_crtc_helper.c
>+++ b/drivers/gpu/drm/drm_crtc_helper.c
>@@ -661,6 +661,7 @@ int drm_crtc_helper_set_config(struct drm_mode_set *set)
> 						      old_fb)) {
> 				DRM_ERROR("failed to set mode on [CRTC:%d]\n",
> 					  set->crtc->base.id);
>+				set->crtc->fb = old_fb;
> 				ret = -EINVAL;
> 				goto fail;
> 			}
>@@ -675,8 +676,10 @@ int drm_crtc_helper_set_config(struct drm_mode_set *set)
> 			set->crtc->fb = set->fb;
> 		ret = crtc_funcs->mode_set_base(set->crtc,
> 						set->x, set->y, old_fb);
>-		if (ret != 0)
>+		if (ret != 0) {
>+			set->crtc->fb = old_fb;
> 			goto fail;
>+		}
> 	}
> 	DRM_DEBUG_KMS("Setting connector DPMS state to on\n");
> 	for (i = 0; i < set->num_connectors; i++) {
>-- 
>1.7.2.3
>

Actions: View | Splinter Review

Attachments on bug 29230: 37399 | 37400 | 41770 | 42635

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.