On Thu, Jan 10, 2019 at 04:43:31PM +0000, Elie Roudninski wrote: [ ... ] > Your patch seems to work. I still got a stacktrace but not a segfault > anymore, and it appears to be related to something else: > [ 148.133895] WARNING: CPU: 1 PID: 580 at > drivers/mmc/host/meson-gx-mmc.c:1027 meson_mmc_irq+0x210/0x2a0 > [ 148.137535] Modules linked in: cfg80211 rfkill 8021q garp mrp stp > llc dw_hdmi_cec meson_dw_hdmi dw_hdmi meson_drm dwmac_generic > drm_kms_helper dwmac_meson8b drm crct10dif_ce stmmac_platform stmmac > drm_panel_orientation_quirks meson_ir ptp syscopyarea rc_core > sysfillrect sysimgblt fb_sys_fops pps_core meson_canvas meson_gxbb_wdt > [ 148.166345] CPU: 1 PID: 580 Comm: bash Tainted: G W > 4.20.0_2 #1 > [ 148.173498] Hardware name: amlogic p212/p212, BIOS > 2019.01-rc2-00176-gf97c49d6a2 01/10/2019 > [ 148.181780] pstate: 60400085 (nZCv daIf +PAN -UAO) > [ 148.186527] pc : meson_mmc_irq+0x210/0x2a0 > [ 148.190586] lr : __free_irq+0x184/0x318 > [ 148.194370] sp : ffff0000108efac0 > [ 148.197647] x29: ffff0000108efac0 x28: ffff800005cfb600 > [ 148.202908] x27: 0000000000000000 x26: 0000000000000000 > [ 148.208169] x25: 0000000000000000 x24: ffff8000105e8228 > [ 148.213431] x23: 0000000000000013 x22: ffff8000105e8358 > [ 148.218692] x21: ffff800005e61f00 x20: ffff8000105e8200 > [ 148.223953] x19: ffff800005e61f00 x18: 0000000000000000 > [ 148.229214] x17: 0000000000000000 x16: 0000000000000000 > [ 148.234476] x15: ffffffffffffffff x14: 6d6d2e3030303437 > [ 148.239737] x13: 3030642f6270612e x12: 0000000000000040 > [ 148.244998] x11: 0000000000000228 x10: 0000000000000040 > [ 148.250259] x9 : ffff000009fa9de8 x8 : ffff000009fa9de0 > [ 148.255521] x7 : ffff800007e7e050 x6 : 0000000000000000 > [ 148.260782] x5 : ffff800007e7df90 x4 : 0000000000000000 > [ 148.266043] x3 : 0000000000000000 x2 : ffff000008b90080 > [ 148.271304] x1 : ffff800005e61f00 x0 : 0000000000000000 > [ 148.276566] Call trace: > [ 148.278984] meson_mmc_irq+0x210/0x2a0 > [ 148.282692] __free_irq+0x184/0x318 > [ 148.286142] free_irq+0x40/0x80 > [ 148.289247] meson_mmc_remove+0x48/0x190 > [ 148.293130] platform_drv_remove+0x30/0x50 > [ 148.297190] device_release_driver_internal+0x1b0/0x248 > [ 148.302357] device_release_driver+0x28/0x38 > [ 148.306583] unbind_store+0xdc/0x148 > [ 148.310119] drv_attr_store+0x40/0x58 > [ 148.313743] sysfs_kf_write+0x5c/0x78 > [ 148.317364] kernfs_fop_write+0xe8/0x1e0 > [ 148.321247] __vfs_write+0x60/0x190 > [ 148.324695] vfs_write+0xac/0x1b0 > [ 148.327972] ksys_write+0x6c/0xd0 > [ 148.331250] __arm64_sys_write+0x24/0x30 > [ 148.335133] el0_svc_common+0x94/0xe8 > [ 148.338754] el0_svc_handler+0x38/0x80 > [ 148.342463] el0_svc+0x8/0xc > [ 148.345308] ---[ end trace eb849c1a635abc72 ]--- > [ 148.350408] meson-gx-mmc d0074000.mmc: Dropping the link to regulator.3 > [ 148.356575] meson-gx-mmc d0074000.mmc: Dropping the link to regulator.1 > Yes should be fixed in 4.21 by https://lkml.org/lkml/2018/12/6/748. > Moreover when I bind the eMMC after the unbind, Linux detects it > properly, its a huge improvement! > Thank you very much! > Good, so I have submitted a slightly modified version of this patch https://patchwork.kernel.org/patch/10756449/. -- Remi
