firefix core dump with linux 5.0.3_rt1-2

firefix core dump with linux 5.0.3_rt1-2

[Bernhard Landauer]

The following happens only with Manjaro's rt patched version of linux50.
No issue with the regular version.

When firefox tries to recover tabs of a previous session, they crash.
Normal operation of firefox is not affected. Only these specific tabs
are broken.

Full journal with core dump here:

https://pastebin.com/W1mTnQ6N

kind gregards
Bernhard

Re: firefix core dump with linux 5.0.3_rt1-2

[Sebastian Andrzej Siewior]

On 2019-03-28 10:29:54 [+0100], Bernhard Landauer wrote:
> The following happens only with Manjaro's rt patched version of linux50.
> No issue with the regular version.
> 
> When firefox tries to recover tabs of a previous session, they crash.
> Normal operation of firefox is not affected. Only these specific tabs
> are broken.

Is this only related to 5.0-RT or are also previous RT versions
affected?

> Full journal with core dump here:
> 
> https://pastebin.com/W1mTnQ6N

This is not helpful. Do you have an reproducer?

> kind gregards
> Bernhard

Sebastian

Re: firefix core dump with linux 5.0.3_rt1-2

[Ralf Mardorf]

On Thu, 28 Mar 2019 10:29:54 +0100, Bernhard Landauer wrote:
>The following happens only with Manjaro's rt patched version of
>linux50. No issue with the regular version.
>
>When firefox tries to recover tabs of a previous session, they crash.
>Normal operation of firefox is not affected. Only these specific tabs
>are broken.
>
>Full journal with core dump here:
>
>https://pastebin.com/W1mTnQ6N

Hi,

does Manjaro use additional patches, maybe those used by linux-rt from
AUR?

I build 5.0.3_rt1-0 without the nf_tables-fix patch and using a
different config than linux-rt from AUR.

Apart from the name 5.0.3_rt1-1 is equal with linux-rt from AUR.

I needed to downgrade, because a windows guest in virtualbox became
unresponsive. I tried several times to use the guest, when booted into
the AUR kernel 5.0.3_rt1-1. There is no issue when using 5.0.3_rt1-0
without one patch and a different config.

[rocketmouse@archlinux ~]$ grep pussytoes\  /var/log/pacman.log | grep downgrade
[2019-03-27 06:33] [ALPM] downgraded linux-rt-pussytoes (5.0.3_rt1-1 -> 5.0.3_rt1-0)

Regards,
Ralf

-- 
pacman -Q linux{,-rt{-pussytoes,-securityink,-cornflower,}}|cut -d\  -f2
5.0.4.arch1-1
5.0.3_rt1-0
4.19.31_rt18-0
4.19.25_rt16-0
4.19.23_rt13-0.1

Re: firefix core dump with linux 5.0.3_rt1-2

[Bernhard Landauer]

@Ralf
> does Manjaro use additional patches, maybe those used by linux-rt from
> AUR?
Our PKGBUILD can be seen here:
https://gitlab.manjaro.org/packages/community/realtime-kernels/linux50-rt/raw/master/PKGBUILD


As you can see we did not (yet) apply the nf-tables patch
Apart from that Manjaro only uses the bootsplash patches and the same
add-sysctl patch as the AUR

Our config can be found here:
https://gitlab.manjaro.org/packages/community/realtime-kernels/linux50-rt/raw/master/config.rt

@Sebastian

> Is this only related to 5.0-RT or are also previous RT versions
> affected?
>
My alternative is 4.19.31_rt18. No such issue with this one.
> This is not helpful. Do you have an reproducer?
>
Afraid not. Could you instruct what to get, how to use?

regards
Bernhard

Re: firefix core dump with linux 5.0.3_rt1-2

[Sebastian Andrzej Siewior]

On 2019-03-28 12:30:30 [+0100], Bernhard Landauer wrote:
> @Ralf
> > does Manjaro use additional patches, maybe those used by linux-rt from
> > AUR?
> Our PKGBUILD can be seen here:
> https://gitlab.manjaro.org/packages/community/realtime-kernels/linux50-rt/raw/master/PKGBUILD
> 
> 
> As you can see we did not (yet) apply the nf-tables patch
> Apart from that Manjaro only uses the bootsplash patches and the same
> add-sysctl patch as the AUR
> 
> Our config can be found here:
> https://gitlab.manjaro.org/packages/community/realtime-kernels/linux50-rt/raw/master/config.rt

"You need to sign in or sign up before continuing."

> @Sebastian
> 
> > Is this only related to 5.0-RT or are also previous RT versions
> > affected?
> >
> My alternative is 4.19.31_rt18. No such issue with this one.
> > This is not helpful. Do you have an reproducer?
> >
> Afraid not. Could you instruct what to get, how to use?

Can you compile with:
| CONFIG_DEBUG_PREEMPT=y
| CONFIG_LOCK_DEBUGGING_SUPPORT=y
| CONFIG_PROVE_LOCKING=y
| # CONFIG_LOCK_STAT is not set
| CONFIG_DEBUG_RT_MUTEXES=y
| CONFIG_DEBUG_SPINLOCK=y
| CONFIG_DEBUG_MUTEXES=y
| CONFIG_DEBUG_WW_MUTEX_SLOWPATH=y
| CONFIG_DEBUG_LOCK_ALLOC=y
| CONFIG_LOCKDEP=y
| # CONFIG_DEBUG_LOCKDEP is not set
| CONFIG_DEBUG_ATOMIC_SLEEP=y
| # CONFIG_LOCK_TORTURE_TEST is not set
| # CONFIG_WW_MUTEX_SELFTEST is not set
| CONFIG_TRACE_IRQFLAGS=y
| CONFIG_STACKTRACE=y
| # CONFIG_WARN_ALL_UNSEEDED_RANDOM is not set
| # CONFIG_DEBUG_KOBJECT is not set
| # CONFIG_DEBUG_KOBJECT_RELEASE is not set
| CONFIG_DEBUG_BUGVERBOSE=y
| CONFIG_DEBUG_LIST=y

and check dmesg if the kernel says anything out of the ordinary after the
crash?

> regards
> Bernhard

Sebastian

Re: firefix core dump with linux 5.0.3_rt1-2

[Ralf Mardorf]

Hi,

I don't have access to gitlab.manjaro, so I make a diff from the AUR
config (note, it still needs to pass make olddefconfig, so it might not
be the final config from AUR) and the config used by my kernel.

[rocketmouse@archlinux tmp]$ wget -q https://aur.archlinux.org/cgit/aur.git/snapshot/linux-rt.tar.gz
[rocketmouse@archlinux tmp]$ tar xf linux-rt.tar.gz
[rocketmouse@archlinux tmp]$ diff linux-rt/config /lib/modules/5.0.3-rt1-0-rt-pussytoes-0-rt/build/.config 
22,23c22,23
< CONFIG_LOCALVERSION=""
< CONFIG_LOCALVERSION_AUTO=y
---
> CONFIG_LOCALVERSION="-0-rt"
> # CONFIG_LOCALVERSION_AUTO is not set
37c37
< CONFIG_DEFAULT_HOSTNAME="archlinux"
---
> CONFIG_DEFAULT_HOSTNAME="archlinux-rt"
112,113c112
< CONFIG_PSI=y
< # CONFIG_PSI_DEFAULT_DISABLED is not set
---
> # CONFIG_PSI is not set
301c300
< CONFIG_X86_CPU_RESCTRL=y
---
> # CONFIG_X86_CPU_RESCTRL is not set
495c494
< CONFIG_ENERGY_MODEL=y
---
> # CONFIG_ENERGY_MODEL is not set
803c802
< # CONFIG_MODVERSIONS is not set
---
> CONFIG_MODVERSIONS=y
1160c1159
< CONFIG_NFT_XFRM=m
---
> # CONFIG_NFT_XFRM is not set
1494,1496d1492
< CONFIG_NET_DSA_TAG_GSWIP=y
< CONFIG_NET_DSA_TAG_KSZ=y
< CONFIG_NET_DSA_TAG_KSZ9477=y
1549c1545
< CONFIG_NET_SCH_TAPRIO=m
---
> # CONFIG_NET_SCH_TAPRIO is not set
1656c1652
< CONFIG_NCSI_OEM_CMD_GET_MAC=y
---
> # CONFIG_NCSI_OEM_CMD_GET_MAC is not set
1705c1701
< CONFIG_CAN_FLEXCAN=m
---
> # CONFIG_CAN_FLEXCAN is not set
1953c1949
< CONFIG_PCI_P2PDMA=y
---
> # CONFIG_PCI_P2PDMA is not set
1989c1985
< CONFIG_PCI_MESON=y
---
> # CONFIG_PCI_MESON is not set
2142,2145c2138,2139
< CONFIG_MTD_PHYSMAP_OF=y
< CONFIG_MTD_PHYSMAP_VERSATILE=y
< CONFIG_MTD_PHYSMAP_GEMINI=y
< CONFIG_MTD_PHYSMAP_GPIO_ADDR=y
---
> # CONFIG_MTD_PHYSMAP_OF is not set
> # CONFIG_MTD_PHYSMAP_GPIO_ADDR is not set
2298c2292
< CONFIG_NVME_TCP=m
---
> # CONFIG_NVME_TCP is not set
2304c2298
< CONFIG_NVME_TARGET_TCP=m
---
> # CONFIG_NVME_TARGET_TCP is not set
2350c2344
< CONFIG_EEPROM_EE1004=m
---
> # CONFIG_EEPROM_EE1004 is not set
2413c2407
< CONFIG_MISC_ALCOR_PCI=m
---
> # CONFIG_MISC_ALCOR_PCI is not set
2505c2499
< CONFIG_SCSI_UFS_CDNS_PLATFORM=m
---
> # CONFIG_SCSI_UFS_CDNS_PLATFORM is not set
2507c2501
< CONFIG_SCSI_UFS_BSG=y
---
> # CONFIG_SCSI_UFS_BSG is not set
2511,2512c2505,2506
< CONFIG_SCSI_MYRB=m
< CONFIG_SCSI_MYRS=m
---
> # CONFIG_SCSI_MYRB is not set
> # CONFIG_SCSI_MYRS is not set
2829c2823
< CONFIG_B53_SERDES=m
---
> # CONFIG_B53_SERDES is not set
2832c2826
< CONFIG_NET_DSA_LANTIQ_GSWIP=m
---
> # CONFIG_NET_DSA_LANTIQ_GSWIP is not set
2834,2836c2828
< CONFIG_NET_DSA_MICROCHIP_KSZ_COMMON=m
< CONFIG_NET_DSA_MICROCHIP_KSZ9477=m
< CONFIG_NET_DSA_MICROCHIP_KSZ9477_SPI=m
---
> # CONFIG_NET_DSA_MICROCHIP_KSZ9477 is not set
2987c2979
< CONFIG_IGC=m
---
> # CONFIG_IGC is not set
3055c3047
< CONFIG_NI_XGE_MANAGEMENT_ENET=m
---
> # CONFIG_NI_XGE_MANAGEMENT_ENET is not set
3285c3277
< CONFIG_USB_NET_AQC111=m
---
> # CONFIG_USB_NET_AQC111 is not set
3474c3466
< CONFIG_MT76x0E=m
---
> # CONFIG_MT76x0E is not set
3556,3557c3548
< CONFIG_QTNFMAC=m
< CONFIG_QTNFMAC_PCIE=m
---
> # CONFIG_QTNFMAC_PCIE is not set
3562c3553
< CONFIG_VIRT_WIFI=m
---
> # CONFIG_VIRT_WIFI is not set
4039c4030
< CONFIG_SERIO_OLPC_APSP=m
---
> # CONFIG_SERIO_OLPC_APSP is not set
4252c4243
< CONFIG_I2C_NVIDIA_GPU=m
---
> # CONFIG_I2C_NVIDIA_GPU is not set
4307,4309c4298
< CONFIG_I3C=m
< CONFIG_CDNS_I3C_MASTER=m
< CONFIG_DW_I3C_MASTER=m
---
> # CONFIG_I3C is not set
4337c4326
< CONFIG_SPI_MXIC=m
---
> # CONFIG_SPI_MXIC is not set
4400c4389
< CONFIG_PINCTRL_OCELOT=y
---
> # CONFIG_PINCTRL_OCELOT is not set
4432c4421
< CONFIG_GPIO_CADENCE=m
---
> # CONFIG_GPIO_CADENCE is not set
4444,4445c4433,4434
< CONFIG_GPIO_SAMA5D2_PIOBU=m
< CONFIG_GPIO_SIOX=m
---
> # CONFIG_GPIO_SAMA5D2_PIOBU is not set
> # CONFIG_GPIO_SIOX is not set
4748,4750c4737
< CONFIG_SENSORS_OCC_P8_I2C=m
< CONFIG_SENSORS_OCC_P9_SBE=m
< CONFIG_SENSORS_OCC=y
---
> # CONFIG_SENSORS_OCC_P8_I2C is not set
4927c4914
< CONFIG_TQMX86_WDT=m
---
> # CONFIG_TQMX86_WDT is not set
5186c5173
< CONFIG_REGULATOR_MCP16502=m
---
> # CONFIG_REGULATOR_MCP16502 is not set
5230a5218
> CONFIG_CEC_PIN=y
5269c5257
< CONFIG_RC_XBOX_DVD=m
---
> # CONFIG_RC_XBOX_DVD is not set
5280a5269
> # CONFIG_CEC_PIN_ERROR_INJ is not set
5556c5545
< CONFIG_VIDEO_ASPEED=m
---
> # CONFIG_VIDEO_ASPEED is not set
5576,5578c5565,5566
< # CONFIG_CEC_GPIO is not set
< CONFIG_VIDEO_SECO_CEC=m
< CONFIG_VIDEO_SECO_RC=y
---
> CONFIG_CEC_GPIO=m
> # CONFIG_VIDEO_SECO_CEC is not set
6012c6000
< CONFIG_HSA_AMD=y
---
> # CONFIG_HSA_AMD is not set
6075c6063
< CONFIG_DRM_PANEL_OLIMEX_LCD_OLINUXINO=m
---
> # CONFIG_DRM_PANEL_OLIMEX_LCD_OLINUXINO is not set
6080c6068
< CONFIG_DRM_PANEL_SAMSUNG_S6D16D0=m
---
> # CONFIG_DRM_PANEL_SAMSUNG_S6D16D0 is not set
6088c6076
< CONFIG_DRM_PANEL_TRULY_NT35597_WQXGA=m
---
> # CONFIG_DRM_PANEL_TRULY_NT35597_WQXGA is not set
6106c6094
< CONFIG_DRM_TOSHIBA_TC358764=m
---
> # CONFIG_DRM_TOSHIBA_TC358764 is not set
6109c6097
< CONFIG_DRM_TI_SN65DSI86=m
---
> # CONFIG_DRM_TI_SN65DSI86 is not set
6124c6112
< CONFIG_TINYDRM_HX8357D=m
---
> # CONFIG_TINYDRM_HX8357D is not set
6471c6459
< CONFIG_SND_SOC_AMD_ACP3x=m
---
> # CONFIG_SND_SOC_AMD_ACP3x is not set
6473c6461
< CONFIG_SND_SOC_MIKROE_PROTO=m
---
> # CONFIG_SND_SOC_MIKROE_PROTO is not set
6518c6506
< CONFIG_SND_SOC_INTEL_SKYLAKE_HDAUDIO_CODEC=y
---
> # CONFIG_SND_SOC_INTEL_SKYLAKE_HDAUDIO_CODEC is not set
6542,6543c6530,6531
< CONFIG_SND_SOC_INTEL_KBL_DA7219_MAX98927_MACH=m
< CONFIG_SND_SOC_INTEL_KBL_RT5660_MACH=m
---
> # CONFIG_SND_SOC_INTEL_KBL_DA7219_MAX98927_MACH is not set
> # CONFIG_SND_SOC_INTEL_KBL_RT5660_MACH is not set
6545d6532
< CONFIG_SND_SOC_INTEL_SKL_HDA_DSP_GENERIC_MACH=m
6550c6537
< CONFIG_SND_SOC_XILINX_I2S=m
---
> # CONFIG_SND_SOC_XILINX_I2S is not set
6567c6554
< CONFIG_SND_SOC_AK4118=m
---
> # CONFIG_SND_SOC_AK4118 is not set
6610d6596
< CONFIG_SND_SOC_HDAC_HDA=m
6612c6598
< CONFIG_SND_SOC_MAX98088=m
---
> # CONFIG_SND_SOC_MAX98088 is not set
6631,6633c6617,6618
< CONFIG_SND_SOC_PCM3060=m
< CONFIG_SND_SOC_PCM3060_I2C=m
< CONFIG_SND_SOC_PCM3060_SPI=m
---
> # CONFIG_SND_SOC_PCM3060_I2C is not set
> # CONFIG_SND_SOC_PCM3060_SPI is not set
6651d6635
< CONFIG_SND_SOC_RT5660=m
6719c6703
< CONFIG_SND_SOC_NAU8822=m
---
> # CONFIG_SND_SOC_NAU8822 is not set
6756c6740
< CONFIG_HID_BIGBEN_FF=m
---
> # CONFIG_HID_BIGBEN_FF is not set
7242c7226
< CONFIG_UCSI_CCG=m
---
> # CONFIG_UCSI_CCG is not set
7287d7270
< CONFIG_MMC_ALCOR=m
7303c7286
< CONFIG_MMC_SDHCI_AM654=m
---
> # CONFIG_MMC_SDHCI_AM654 is not set
7332c7315
< CONFIG_LEDS_AN30259A=m
---
> # CONFIG_LEDS_AN30259A is not set
7414c7397
< CONFIG_LEDS_TRIGGER_PATTERN=m
---
> # CONFIG_LEDS_TRIGGER_PATTERN is not set
7660c7643
< CONFIG_UDMABUF=y
---
> # CONFIG_UDMABUF is not set
7948c7931
< CONFIG_VIDEO_IPU3_IMGU=m
---
> # CONFIG_VIDEO_IPU3_IMGU is not set
8026c8009
< CONFIG_LG_LAPTOP=m
---
> # CONFIG_LG_LAPTOP is not set
8084,8085c8067,8068
< CONFIG_INTEL_ATOMISP2_PM=m
< CONFIG_HUAWEI_WMI=m
---
> # CONFIG_INTEL_ATOMISP2_PM is not set
> # CONFIG_HUAWEI_WMI is not set
8125c8108
< CONFIG_COMMON_CLK_BD718XX=m
---
> # CONFIG_COMMON_CLK_BD718XX is not set
8264,8266c8247,8248
< CONFIG_ADXL372=m
< CONFIG_ADXL372_SPI=m
< CONFIG_ADXL372_I2C=m
---
> # CONFIG_ADXL372_SPI is not set
> # CONFIG_ADXL372_I2C is not set
8305c8287
< CONFIG_AD7124=m
---
> # CONFIG_AD7124 is not set
8315c8297
< CONFIG_AD7949=m
---
> # CONFIG_AD7949 is not set
8338c8320
< CONFIG_MCP3911=m
---
> # CONFIG_MCP3911 is not set
8345c8327
< CONFIG_QCOM_SPMI_ADC5=m
---
> # CONFIG_QCOM_SPMI_ADC5 is not set
8421c8403
< CONFIG_LTC1660=m
---
> # CONFIG_LTC1660 is not set
8442c8424
< CONFIG_TI_DAC7311=m
---
> # CONFIG_TI_DAC7311 is not set
8571c8553
< CONFIG_VCNL4035=m
---
> # CONFIG_VCNL4035 is not set
8594,8596c8576,8577
< CONFIG_SENSORS_RM3100=m
< CONFIG_SENSORS_RM3100_I2C=m
< CONFIG_SENSORS_RM3100_SPI=m
---
> # CONFIG_SENSORS_RM3100_I2C is not set
> # CONFIG_SENSORS_RM3100_SPI is not set
8627c8608
< CONFIG_MCP41010=m
---
> # CONFIG_MCP41010 is not set
8676c8657
< CONFIG_VL53L0X_I2C=m
---
> # CONFIG_VL53L0X_I2C is not set
8759,8761c8740,8742
< CONFIG_PHY_CADENCE_DP=m
< CONFIG_PHY_CADENCE_SIERRA=m
< CONFIG_PHY_FSL_IMX8MQ_USB=m
---
> # CONFIG_PHY_CADENCE_DP is not set
> # CONFIG_PHY_CADENCE_SIERRA is not set
> # CONFIG_PHY_FSL_IMX8MQ_USB is not set
8766c8747
< CONFIG_PHY_OCELOT_SERDES=m
---
> # CONFIG_PHY_OCELOT_SERDES is not set
8808,8809c8789,8790
< CONFIG_STM_PROTO_BASIC=m
< CONFIG_STM_PROTO_SYS_T=m
---
> # CONFIG_STM_PROTO_BASIC is not set
> # CONFIG_STM_PROTO_SYS_T is not set
8847c8828
< CONFIG_FSI_OCC=m
---
> # CONFIG_FSI_OCC is not set
9042c9023
< CONFIG_UBIFS_FS_AUTHENTICATION=y
---
> # CONFIG_UBIFS_FS_AUTHENTICATION is not set
9345c9326
< CONFIG_CRYPTO_ECB=m
---
> CONFIG_CRYPTO_ECB=y
9347c9328
< CONFIG_CRYPTO_OFB=m
---
> # CONFIG_CRYPTO_OFB is not set
9351,9354c9332,9334
< CONFIG_CRYPTO_NHPOLY1305=m
< CONFIG_CRYPTO_NHPOLY1305_SSE2=m
< CONFIG_CRYPTO_NHPOLY1305_AVX2=m
< CONFIG_CRYPTO_ADIANTUM=m
---
> # CONFIG_CRYPTO_NHPOLY1305_SSE2 is not set
> # CONFIG_CRYPTO_NHPOLY1305_AVX2 is not set
> # CONFIG_CRYPTO_ADIANTUM is not set
9391c9371
< CONFIG_CRYPTO_STREEBOG=m
---
> # CONFIG_CRYPTO_STREEBOG is not set
9489c9469
< CONFIG_ASYMMETRIC_TPM_KEY_SUBTYPE=m
---
> # CONFIG_ASYMMETRIC_TPM_KEY_SUBTYPE is not set
9491,9492c9471
< CONFIG_PKCS8_PRIVATE_KEY_PARSER=m
< CONFIG_TPM_KEY_PARSER=m
---
> # CONFIG_PKCS8_PRIVATE_KEY_PARSER is not set
9737a9717
> CONFIG_TRACE_IRQFLAGS=y
9777a9758
> CONFIG_PREEMPTIRQ_TRACEPOINTS=y
9784c9765,9766
< # CONFIG_PREEMPTIRQ_EVENTS is not set
---
> CONFIG_TRACE_PREEMPT_TOGGLE=y
> CONFIG_PREEMPTIRQ_EVENTS=y

FWIW I didn't use this kernel for real-time audio work, it was only
used for averaged desktop work.

At the moment I'm using

[rocketmouse@archlinux tmp]$ uname -rm
4.19.31-rt18-0-rt-securityink x86_64

also not for real-time related work, just for office stuff.

Regards,
Ralf

-- 
pacman -Q linux{,-rt{-pussytoes,-securityink,-cornflower,}}|cut -d\  -f2
5.0.4.arch1-1
5.0.3_rt1-0
4.19.31_rt18-0
4.19.25_rt16-0
4.19.23_rt13-0.1

Re: firefix core dump with linux 5.0.3_rt1-2

[Bernhard Landauer]

> "You need to sign in or sign up before continuing."

meh... sorry!

PKGBUILD:
https://pastebin.com/1N39b4E0

config:
https://pastebin.com/uE4xqyj1

Re: firefix core dump with linux 5.0.3_rt1-2

[Bernhard Landauer]

> Can you compile with:

I'll try that tonight. Thanks.

Re: firefix core dump with linux 5.0.3_rt1-2

[Bernhard Landauer]

> and check dmesg if the kernel says anything out of the ordinary after
the crash?

This is all I got. Not sure that this will help in any way ... ;)

[   38.970923] audit: type=1131 audit(1553811118.623:61): pid=1 uid=0
auid=4294967295 ses=4294967295 msg='unit=systemd-hostnamed
comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=?
terminal=? res=success'
[   71.377402] audit: type=1701 audit(1553811151.165:62): auid=1000
uid=1000 gid=1000 ses=2 pid=1692 comm=57656220436F6E74656E74
exe="/usr/lib/firefox/firefox" sig=8 res=1
[   71.426246] audit: type=1130 audit(1553811151.215:63): pid=1 uid=0
auid=4294967295 ses=4294967295 msg='unit=systemd-coredump@0-1936-0
comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=?
terminal=? res=success'
[   72.892044] audit: type=1131 audit(1553811152.679:64): pid=1 uid=0
auid=4294967295 ses=4294967295 msg='unit=systemd-coredump@0-1936-0
comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=?
terminal=? res=success'
[   74.147791] audit: type=1701 audit(1553811153.937:65): auid=1000
uid=1000 gid=1000 ses=2 pid=1902 comm=57656220436F6E74656E74
exe="/usr/lib/firefox/firefox" sig=8 res=1
[   74.164974] audit: type=1130 audit(1553811153.953:66): pid=1 uid=0
auid=4294967295 ses=4294967295 msg='unit=systemd-coredump@1-2016-0
comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=?
terminal=? res=success'
[   75.691743] audit: type=1131 audit(1553811155.481:67): pid=1 uid=0
auid=4294967295 ses=4294967295 msg='unit=systemd-coredump@1-2016-0
comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=?
terminal=? res=success'
[   82.813829] audit: type=1701 audit(1553811162.608:68): auid=1000
uid=1000 gid=1000 ses=2 pid=1981 comm=57656220436F6E74656E74
exe="/usr/lib/firefox/firefox" sig=8 res=1
[   82.830137] audit: type=1130 audit(1553811162.624:69): pid=1 uid=0
auid=4294967295 ses=4294967295 msg='unit=systemd-coredump@2-2096-0
comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=?
terminal=? res=success'
[   84.831242] audit: type=1131 audit(1553811164.622:70): pid=1 uid=0
auid=4294967295 ses=4294967295 msg='unit=systemd-coredump@2-2096-0
comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=?
terminal=? res=success'
[   90.422327] audit: type=1701 audit(1553811170.217:71): auid=1000
uid=1000 gid=1000 ses=2 pid=2038 comm=57656220436F6E74656E74
exe="/usr/lib/firefox/firefox" sig=8 res=1
[   90.439593] audit: type=1130 audit(1553811170.237:72): pid=1 uid=0
auid=4294967295 ses=4294967295 msg='unit=systemd-coredump@3-2147-0
comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=?
terminal=? res=success'
[   91.920716] audit: type=1131 audit(1553811171.717:73): pid=1 uid=0
auid=4294967295 ses=4294967295 msg='unit=systemd-coredump@3-2147-0
comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=?
terminal=? res=success'

Recently there were several systemd updates via Archlinux - maybe this
could be related. However, as I said before, this issue is not present
with 4.19-rt

regards
Bernhard

Re: firefix core dump with linux 5.0.3_rt1-2

[Bernhard Landauer]

I'm building 5.0.5-rt3 now and check if things have changed anyway.
Also, a new version of firefox just arrived ... ;)

will report shortly.

Re: firefix core dump with linux 5.0.3_rt1-2

[Bernhard Landauer]

Unfortunately the issue is still the same with 5.0.5-rt3 (and with
firefox 66.0.2)

Re: firefix core dump with linux 5.0.3_rt1-2

[Sebastian Andrzej Siewior]

On 2019-03-29 00:56:01 [+0100], Bernhard Landauer wrote:
> Unfortunately the issue is still the same with 5.0.5-rt3 (and with
> firefox 66.0.2)

okay. How do I reproduce this? What do I need to do with firefox in
order to trigger this?

Sebastian

Re: firefix core dump with linux 5.0.3_rt1-2

[Bernhard Landauer]

> How do I reproduce this?

For me this happens reliably when I launch a fresh instance of ff and it
automatically tries to restore tabs of the previous session.
Alternatively, when I select 'Restore Previous Session' from the menu.

Re: firefix core dump with linux 5.0.3_rt1-2

[Sebastian Andrzej Siewior]

On 2019-03-29 09:38:33 [+0100], Bernhard Landauer wrote:
> 
> > How do I reproduce this?
> 
> For me this happens reliably when I launch a fresh instance of ff and it
> automatically tries to restore tabs of the previous session.
> Alternatively, when I select 'Restore Previous Session' from the menu.

Does this help?

diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c
index ca95ab2f4cfa3..8744d20ac1681 100644
--- a/drivers/gpu/drm/i915/i915_request.c
+++ b/drivers/gpu/drm/i915/i915_request.c
@@ -278,9 +278,7 @@ static void __retire_engine_request(struct intel_engine_cs *engine,
 
 	GEM_BUG_ON(!i915_request_completed(rq));
 
-	local_irq_disable();
-
-	spin_lock(&engine->timeline.lock);
+	spin_lock_irq(&engine->timeline.lock);
 	GEM_BUG_ON(!list_is_first(&rq->link, &engine->timeline.requests));
 	list_del_init(&rq->link);
 	spin_unlock(&engine->timeline.lock);
@@ -294,9 +292,7 @@ static void __retire_engine_request(struct intel_engine_cs *engine,
 		GEM_BUG_ON(!atomic_read(&rq->i915->gt_pm.rps.num_waiters));
 		atomic_dec(&rq->i915->gt_pm.rps.num_waiters);
 	}
-	spin_unlock(&rq->lock);
-
-	local_irq_enable();
+	spin_unlock_irq(&rq->lock);
 
 	/*
 	 * The backing object for the context is done after switching to the

Sebastian

Re: firefix core dump with linux 5.0.3_rt1-2

[Daniel Vacek]

On Mon, Apr 8, 2019 at 7:31 PM Sebastian Andrzej Siewior
<bigeasy@linutronix.de> wrote:
>
> On 2019-03-29 09:38:33 [+0100], Bernhard Landauer wrote:
> >
> > > How do I reproduce this?
> >
> > For me this happens reliably when I launch a fresh instance of ff and it
> > automatically tries to restore tabs of the previous session.
> > Alternatively, when I select 'Restore Previous Session' from the menu.
>
> Does this help?

From an uninvolved lurker, how is this supposed to change a thing?
Moreover, I'd say the original version is better readable.

--nX

> diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c
> index ca95ab2f4cfa3..8744d20ac1681 100644
> --- a/drivers/gpu/drm/i915/i915_request.c
> +++ b/drivers/gpu/drm/i915/i915_request.c
> @@ -278,9 +278,7 @@ static void __retire_engine_request(struct intel_engine_cs *engine,
>
>         GEM_BUG_ON(!i915_request_completed(rq));
>
> -       local_irq_disable();
> -
> -       spin_lock(&engine->timeline.lock);
> +       spin_lock_irq(&engine->timeline.lock);
>         GEM_BUG_ON(!list_is_first(&rq->link, &engine->timeline.requests));
>         list_del_init(&rq->link);
>         spin_unlock(&engine->timeline.lock);
> @@ -294,9 +292,7 @@ static void __retire_engine_request(struct intel_engine_cs *engine,
>                 GEM_BUG_ON(!atomic_read(&rq->i915->gt_pm.rps.num_waiters));
>                 atomic_dec(&rq->i915->gt_pm.rps.num_waiters);
>         }
> -       spin_unlock(&rq->lock);
> -
> -       local_irq_enable();
> +       spin_unlock_irq(&rq->lock);
>
>         /*
>          * The backing object for the context is done after switching to the
>
> Sebastian

Re: firefix core dump with linux 5.0.3_rt1-2

[Sebastian Andrzej Siewior]

On 2019-04-08 21:33:55 [+0200], Daniel Vacek wrote:
> On Mon, Apr 8, 2019 at 7:31 PM Sebastian Andrzej Siewior
> <bigeasy@linutronix.de> wrote:
> >
> > On 2019-03-29 09:38:33 [+0100], Bernhard Landauer wrote:
> > >
> > > > How do I reproduce this?
> > >
> > > For me this happens reliably when I launch a fresh instance of ff and it
> > > automatically tries to restore tabs of the previous session.
> > > Alternatively, when I select 'Restore Previous Session' from the menu.
> >
> > Does this help?
> 
> >From an uninvolved lurker, how is this supposed to change a thing?

on RT
  local_irq_disable() + spin_lock() != spin_lock_irq()

> Moreover, I'd say the original version is better readable.

how so? You remove the irq/irq_save() from a lock and put a local_irq_()
in front of it. Is the only purpose is to avoid enabling/disabling twice
in a row?

> --nX

Sebastian

Re: firefix core dump with linux 5.0.3_rt1-2

[Daniel Vacek]

On Mon, Apr 8, 2019 at 9:57 PM Sebastian Andrzej Siewior
<bigeasy@linutronix.de> wrote:
>
> On 2019-04-08 21:33:55 [+0200], Daniel Vacek wrote:
> > On Mon, Apr 8, 2019 at 7:31 PM Sebastian Andrzej Siewior
> > <bigeasy@linutronix.de> wrote:
> > >
> > > On 2019-03-29 09:38:33 [+0100], Bernhard Landauer wrote:
> > > >
> > > > > How do I reproduce this?
> > > >
> > > > For me this happens reliably when I launch a fresh instance of ff and it
> > > > automatically tries to restore tabs of the previous session.
> > > > Alternatively, when I select 'Restore Previous Session' from the menu.
> > >
> > > Does this help?
> >
> > >From an uninvolved lurker, how is this supposed to change a thing?
>
> on RT
>   local_irq_disable() + spin_lock() != spin_lock_irq()

Nah, the rt_mutex... I always fall into this trap :-/

> > Moreover, I'd say the original version is better readable.
>
> how so? You remove the irq/irq_save() from a lock and put a local_irq_()
> in front of it. Is the only purpose is to avoid enabling/disabling twice
> in a row?

What I meant is that it's two different locks so it's easier to see
that you a) disable interrupts and then b) do the work with the locks.
But yeah, you would have to change this to
local_irq_{dis,en}able_nort() to achieve what you did here. Still I'd
argue that doing this would be clearer to document the intention.
Anyways, please disregard. I'll back to lurking...

--nX

> > --nX
>
> Sebastian

Re: firefix core dump with linux 5.0.3_rt1-2

[Bernhard Landauer]

> Does this help?
>
I'm afraid not. The issue is still unchanged for me.

regards
Bernhard

Re: firefix core dump with linux 5.0.3_rt1-2

[Bernhard Landauer]

Oh dang! Appologies. I just realize I didn't actually apply the patch ... ;)
Stay tuned please. Have to rebuild once more. ...

On 10.04.19 09:23, Bernhard Landauer wrote:
>> Does this help?
>>
> I'm afraid not. The issue is still unchanged for me.
>
> regards
> Bernhard
>

Re: firefix core dump with linux 5.0.3_rt1-2

[Bernhard Landauer]

So no. The patch definitely doesn't solve the issue I'm afraid. Just
tested now and situation is still the same.

On 10.04.19 09:27, Bernhard Landauer wrote:
> Oh dang! Appologies. I just realize I didn't actually apply the patch ... ;)
> Stay tuned please. Have to rebuild once more. ...
>
> On 10.04.19 09:23, Bernhard Landauer wrote:
>>> Does this help?
>>>
>> I'm afraid not. The issue is still unchanged for me.
>>
>> regards
>> Bernhard
>>

Re: firefix core dump with linux 5.0.3_rt1-2

[Sebastian Andrzej Siewior]

On 2019-04-10 10:55:24 [+0200], Bernhard Landauer wrote:
> So no. The patch definitely doesn't solve the issue I'm afraid. Just
> tested now and situation is still the same.

Do you have another box to cross check that? I have here an i7 with i915
and I don't have the problem. I killed firefox in the middle of the
operation and then restoring last session just went fine.
So whatever it is, I can't reproduce it.

Can you enable lockdep / debugging and check what dmesg says?

Sebastian

Re: firefix core dump with linux 5.0.3_rt1-2

[Sebastian Andrzej Siewior]

On 2019-04-10 10:59:59 [+0200], To Bernhard Landauer wrote:
> On 2019-04-10 10:55:24 [+0200], Bernhard Landauer wrote:
> > So no. The patch definitely doesn't solve the issue I'm afraid. Just
> > tested now and situation is still the same.
> 
> Do you have another box to cross check that? I have here an i7 with i915
> and I don't have the problem. I killed firefox in the middle of the
> operation and then restoring last session just went fine.
> So whatever it is, I can't reproduce it.
> 
> Can you enable lockdep / debugging and check what dmesg says?

did the issue vanish in v5.0.7-rt5?

Sebastian