After updating Raspbian to 4.14.98-v7+ #1200 SMP Tue Feb 12 20:27:48 GMT
2019 armv7l  the experimental GL desktop has started locking up and then
crashing when I open a particular website. An error is reported in dmesg:

[ 4364.372304] vc4-drm soc:gpu: failed to allocate buffer with size 9805824
[ 4364.372369] vc4-drm soc:gpu: failed to allocate buffer with size 9805824
[ 4364.372427] vc4-drm soc:gpu: failed to allocate buffer with size 9805824
[ 4364.372475] [drm:vc4_bo_create [vc4]] *ERROR* Failed to allocate from CMA:
[ 4364.372479] [drm]                         kernel:   8100kb BOs (1)
[ 4364.372482] [drm]                            V3D: 218012kb BOs (182)
[ 4364.372486] [drm]                     V3D shader:    228kb BOs (55)
[ 4364.372489] [drm]                           dumb:     48kb BOs (3)
[ 4371.277991] ------------[ cut here ]------------
[ 4371.278108] WARNING: CPU: 0 PID: 805 at drivers/gpu/drm/vc4/vc4_plane.c:782
vc4_plane_async_set_fb+0xa4/0xa8 [vc4]
[ 4371.278117] Modules linked in: fuse rfcomm bnep hci_uart btbcm serdev
bluetooth ecdh_generic sg vc4 brcmfmac drm_kms_helper brcmutil drm cfg80211
snd_soc_core snd_compress rfkill snd_pcm_dmaengine syscopyarea snd_bcm2835(C)
sysfillrect sysimgblt fb_
sys_fops snd_pcm snd_timer snd i2c_bcm2835 uio_pdrv_genirq uio fixed evdev uas
i2c_dev ip_tables x_tables ipv6
[ 4371.278364] CPU: 0 PID: 805 Comm: Xorg Tainted: G         C      4.14.98-v7+
#1200
[ 4371.278370] Hardware name: BCM2835
[ 4371.278418] [] (unwind_backtrace) from []
(show_stack+0x20/0x24)
[ 4371.278442] [] (show_stack) from []
(dump_stack+0xd4/0x118)
[ 4371.278464] [] (dump_stack) from [] (__warn+0xf8/0x110)
[ 4371.278487] [] (__warn) from []
(warn_slowpath_null+0x30/0x38)
[ 4371.278585] [] (warn_slowpath_null) from []
(vc4_plane_async_set_fb+0xa4/0xa8 [vc4])
[ 4371.278742] [] (vc4_plane_async_set_fb [vc4]) from []
(vc4_update_plane+0xd4/0x15c [vc4])
[ 4371.279109] [] (vc4_update_plane [vc4]) from []
(__setplane_internal+0x194/0x240 [drm])
[ 4371.279613] [] (__setplane_internal [drm]) from []
(drm_mode_cursor_universal+0x128/0x1c4 [drm])
[ 4371.280066] [] (drm_mode_cursor_universal [drm]) from []
(drm_mode_cursor_common+0x174/0x200 [drm])
[ 4371.280518] [] (drm_mode_cursor_common [drm]) from []
(drm_mode_cursor2_ioctl+0x18/0x1c [drm])
[ 4371.280968] [] (drm_mode_cursor2_ioctl [drm]) from []
(drm_ioctl_kernel+0x78/0xb8 [drm])
[ 4371.281420] [] (drm_ioctl_kernel [drm]) from []
(drm_ioctl+0x21c/0x398 [drm])
[ 4371.281657] [] (drm_ioctl [drm]) from []
(do_vfs_ioctl+0xac/0x7c4)
[ 4371.281674] [] (do_vfs_ioctl) from []
(SyS_ioctl+0x44/0x6c)
[ 4371.281696] [] (SyS_ioctl) from []
(ret_fast_syscall+0x0/0x28)
[ 4371.281830] ---[ end trace bc3d7a848566b9d3 ]---

Curiously, backing down to the "old" display driver causes X to not
start, seemingly with no explicit error message.

If anybody knows a workaround, or a better place to grumble about this,
please let me know.

Thanks for reading,

bob prohaska

--- SoupGate-Win32 v1.05