RK3328 : HDR media playback issues. #22
Labels
Comments
|
Using the following patch seems to make the sample linked to in original post work a little bit better. It also fixes playback of my --- a/mpp/hal/rkdec/h265d/hal_h265d_reg.c
+++ b/mpp/hal/rkdec/h265d/hal_h265d_reg.c
@@ -1468,8 +1468,6 @@ MPP_RET hal_h265d_gen_regs(void *hal, HalTaskInfo *syn)
}
hw_regs->sw_interrupt.sw_dec_e = 1;
hw_regs->sw_interrupt.sw_dec_timeout_e = 1;
- hw_regs->sw_interrupt.sw_wr_ddr_align_en = dxva_cxt->pp.tiles_enabled_flag
- ? 0 : 1;
///find s->rps_model[i] position, and set registerThere might also be an issue with the index 1bccb02..432b8db 100644
--- a/mpp/hal/rkdec/h265d/hal_h265d_reg.h
+++ b/mpp/hal/rkdec/h265d/hal_h265d_reg.h
@@ -50,7 +50,8 @@ typedef struct {
struct swreg_int {
RK_U32 sw_dec_e : 1 ;
RK_U32 sw_dec_clkgate_e : 1 ;
- RK_U32 reserve0 : 2 ;
+ RK_U32 reserve0 : 1 ;
+ RK_U32 sw_timeout_mode : 1 ;
RK_U32 sw_dec_irq_dis : 1 ;
RK_U32 sw_dec_timeout_e : 1 ;
RK_U32 sw_buf_empty_en : 1 ;
@@ -61,8 +62,9 @@ typedef struct {
RK_U32 sw_dec_rdy_sta : 1 ;
RK_U32 sw_dec_bus_sta : 1 ;
RK_U32 sw_dec_error_sta : 1 ;
+ RK_U32 sw_dec_timeout_sta : 1 ;
RK_U32 sw_dec_empty_sta : 1 ;
- RK_U32 reserve4 : 4 ;
+ RK_U32 reserve3 : 3 ;
RK_U32 sw_softrst_en_p : 1 ;
RK_U32 sw_force_softreset_valid: 1 ;
RK_U32 sw_softreset_rdy : 1 ; |
|
Looks to fix HDR playback issues for me too. |
|
Yeah, Our engineer said it would not impact to the HDR video playback. And we will check it again. |
|
I can start HDR Video but my TV did not Show HDR Content on Display. My Samsung TV say UHD 3180x2160 but not HDR UHD 3180x2160. |
|
@yanghanxing : ok here is some update on our HDR investigations and understanding The following DRM Atomic properties need to be set :
Most of this information is coming from the media itself. We can probably feed most of this information from the stream metadata information. it looks that ffmpeg grabs this from media stream and puts this in the structs below that are stored in They can be retrieved from The only field that we didn't find information for was the
Should bet set to one of the following values : is setting this property to
Presumably, we would use
should be set to the following values We could probably link that one to the ffmpeg followong enum (https://ffmpeg.org/doxygen/2.7/pixfmt_8h.html#aff71a069509a1ad3ff54d53a1c894c85) unless that's the enum we should use for the
should be set to the following values Could you please confirm that the above information is correct and give use possible values for Are all these properties required to get proper HDR display on media ? Thanks ! |
|
@LongChair ，HDR_SOURCE_METADATA is the only property required, other properties are optional, HDMI driver will automatically select the appropriate value, prefer YCbCr422 BT2020. The type field should be zero, and eotf filed should be SMPTE_ST2084. |
|
@yzheng2012 : what will happen if i set HDR_SOURCE_METADATA with an HDR movie and that the tv doesn't support it, will it fallback to a supported mode smartly and display it the best it can or will the display be broken ? Because it seems kernel is already fetching TV capabilitites before the switch to me :) |
|
Beware. Panel "COLOR_SPACE" must be set too. If "COLOR_SPACE" is leaved in default VOP makes conversions.
|
|
Beware. You must always set "hdmi_output_format" on connector before sending video (even with the same value like DRM_HDMI_OUTPUT_YCBCR_LQ). If not re-set "post CSC" chain remains in previous state (eg. when play hdr and then sdr, sdr video becomes magenta colored). |
|
Performance is not perfect. I am still slightly under 60FPS for HDR with optimization (see rockchip-linux#59). |
|
@mcerveny thanks for sharing your findings, I have also had some stability issues and seen some artifacts in decoding using the latest kernel and ddr dvfs enabled. I will probably stay on kernel commit rockchip-linux@9b04710 that seems more stable in my tests, at least until I have more time to test a more recent kernel commit. |
|
There are problems with HDR/HLG (Hybrid Log-Gamma) EOTF (Electro-Optical Transfer Function)
|
|
@yzheng2012 : could you please confirm Martin’s findings so that we have a clear view on what is required and what’s not ? |
|
HDR/HLG working. Added patch to kernel (see rockchip-linux#61). MPP patch will not be available (see rockchip-linux/mpp#38). I patched test utility to do workaround (see mcerveny/utils@12b893a). Tested (HDR/HLG recognized by LG OLED TV output limited "video=HDMI-A-1:3840x2160@30" due to lack of HDMI2 cables) and running @ 67FPS without boost CPU/DDR: |
|
@mcerveny : great ! and thanks for sharing once again :) |
|
Check the code mpi_dec.c in utils:
|
|
I looked to rockchip-linux code and some up-conversion (like SDR2HDR_FOR_HLG_HDR), down-conversion and whole HDR cross-coversion code missing (between different plane and connector EOTF>0 (PQ<->HLG<->LOG) check slide 15). (also ignored by rockchip rockchip-linux#61). |
|
HDR playback resolution 3840x2160p60 is partially solved by mcerveny/rockchip-linux@b4bc703. I am not able to decode/resolve problems in undocumented (and probably untested) rockchip registry setting part or similar. |
|
@mcerveny nice findings 👍, I will test 2160p@60hz 10-bit 4:2:0 mode with your patch ASAP. Agree, it would be great to have documentation for the inno hdmi phy. Some hw register information can be deduced from https://github.com/rockchip-linux/kernel/blob/release-3.10/drivers/video/rockchip/hdmi/rockchip-hdmiv2/rockchip_hdmiv2_hw.h#L1584-L1750, There is also two inno hdmi phy related commits in the recent |
|
4k60 HDR+HDR/HLG (10bit 4:2:0 - TMDS clock 371Mhz) with never patched rockchip-4.4 kernel is working (tested on LG TV). TMDS Clk: 92812500Hz in output is OK. |
|
@mcerveny : nice :) we're based on an older kernel still, and HDR works, but not in 60 fps. |
|
I did not make bisect. But rockchip-linux@d4953cf equals to my mcerveny@b4bc703 and add something new for faster "clock ratio". Maybe more in history sysnopsys, drm/rockchip and/or phy/rockchip (for example rockchip-linux@41ada53). |
|
@mcerveny Hi, Can you rebase your changes to latest 4.4 kernel & mpp. |
|
@Kwiboo: I have still performance issue with "LG Chess 4K Demo.mp4" in mpi_dec tester or with ffmpeg. Even if I set cpu>=1008MHz and mem==924MHz with actual rockchip-4.4 kernel and mpp+ffpeg from github I am getting only 57 FPS eg. missing additional 5-7% performance in mpp to get playable output. Are you in contact with rockchip to resolve this problem (rockchip-linux#59) ? |
|
HDR10 (SMTPE ST 2084/BT2020/YUV420 10bit) works on the RK3328 with Kodi (with @Kwiboo's sources), but there are still files that break the decoder (colors all over the place/black) or where it can't keep up (mainly 60FPS content, but also everything over 120Mbps). |
|
@mcerveny sorry for late reply, as @lorenz point out HDR can be activated with LibreELEC/LibreELEC.tv#2927, it does not set full static metadata and other limitations (60fps), hopefully it will be good enough for 24fps movies |
Kwiboo
pushed a commit
that referenced
this issue
Dec 15, 2018
Increase kasan instrumented kernel stack size from 32k to 64k. Other architectures seems to get away with just doubling kernel stack size under kasan, but on s390 this appears to be not enough due to bigger frame size. The particular pain point is kasan inlined checks (CONFIG_KASAN_INLINE vs CONFIG_KASAN_OUTLINE). With inlined checks one particular case hitting stack overflow is fs sync on xfs filesystem: #0 [9a0681e8] 704 bytes check_usage at 34b1fc #1 [9a0684a8] 432 bytes check_usage at 34c710 #2 [9a068658] 1048 bytes validate_chain at 35044a #3 [9a068a70] 312 bytes __lock_acquire at 3559fe #4 [9a068ba8] 440 bytes lock_acquire at 3576ee #5 [9a068d60] 104 bytes _raw_spin_lock at 21b44e0 #6 [9a068dc8] 1992 bytes enqueue_entity at 2dbf72 #7 [9a069590] 1496 bytes enqueue_task_fair at 2df5f0 #8 [9a069b68] 64 bytes ttwu_do_activate at 28f438 #9 [9a069ba8] 552 bytes try_to_wake_up at 298c4c #10 [9a069dd0] 168 bytes wake_up_worker at 23f97c #11 [9a069e78] 200 bytes insert_work at 23fc2e #12 [9a069f40] 648 bytes __queue_work at 2487c0 #13 [9a06a1c8] 200 bytes __queue_delayed_work at 24db28 #14 [9a06a290] 248 bytes mod_delayed_work_on at 24de84 #15 [9a06a388] 24 bytes kblockd_mod_delayed_work_on at 153e2a0 #16 [9a06a3a0] 288 bytes __blk_mq_delay_run_hw_queue at 158168c #17 [9a06a4c0] 192 bytes blk_mq_run_hw_queue at 1581a3c #18 [9a06a580] 184 bytes blk_mq_sched_insert_requests at 15a2192 #19 [9a06a638] 1024 bytes blk_mq_flush_plug_list at 1590f3a #20 [9a06aa38] 704 bytes blk_flush_plug_list at 1555028 #21 [9a06acf8] 320 bytes schedule at 219e476 #22 [9a06ae38] 760 bytes schedule_timeout at 21b0aac #23 [9a06b130] 408 bytes wait_for_common at 21a1706 #24 [9a06b2c8] 360 bytes xfs_buf_iowait at fa1540 #25 [9a06b430] 256 bytes __xfs_buf_submit at fadae6 #26 [9a06b530] 264 bytes xfs_buf_read_map at fae3f6 #27 [9a06b638] 656 bytes xfs_trans_read_buf_map at 10ac9a8 #28 [9a06b8c8] 304 bytes xfs_btree_kill_root at e72426 #29 [9a06b9f8] 288 bytes xfs_btree_lookup_get_block at e7bc5e #30 [9a06bb18] 624 bytes xfs_btree_lookup at e7e1a6 #31 [9a06bd88] 2664 bytes xfs_alloc_ag_vextent_near at dfa070 #32 [9a06c7f0] 144 bytes xfs_alloc_ag_vextent at dff3ca #33 [9a06c880] 1128 bytes xfs_alloc_vextent at e05fce #34 [9a06cce8] 584 bytes xfs_bmap_btalloc at e58342 #35 [9a06cf30] 1336 bytes xfs_bmapi_write at e618de #36 [9a06d468] 776 bytes xfs_iomap_write_allocate at ff678e #37 [9a06d770] 720 bytes xfs_map_blocks at f82af8 rockchip-linux#38 [9a06da40] 928 bytes xfs_writepage_map at f83cd6 rockchip-linux#39 [9a06dde0] 320 bytes xfs_do_writepage at f85872 rockchip-linux#40 [9a06df20] 1320 bytes write_cache_pages at 73dfe8 rockchip-linux#41 [9a06e448] 208 bytes xfs_vm_writepages at f7f892 rockchip-linux#42 [9a06e518] 88 bytes do_writepages at 73fe6a rockchip-linux#43 [9a06e570] 872 bytes __writeback_single_inode at a20cb6 rockchip-linux#44 [9a06e8d8] 664 bytes writeback_sb_inodes at a23be2 rockchip-linux#45 [9a06eb70] 296 bytes __writeback_inodes_wb at a242e0 rockchip-linux#46 [9a06ec98] 928 bytes wb_writeback at a2500e rockchip-linux#47 [9a06f038] 848 bytes wb_do_writeback at a260ae rockchip-linux#48 [9a06f388] 536 bytes wb_workfn at a28228 rockchip-linux#49 [9a06f5a0] 1088 bytes process_one_work at 24a234 rockchip-linux#50 [9a06f9e0] 1120 bytes worker_thread at 24ba26 rockchip-linux#51 [9a06fe40] 104 bytes kthread at 26545a rockchip-linux#52 [9a06fea8] kernel_thread_starter at 21b6b62 To be able to increase the stack size to 64k reuse LLILL instruction in __switch_to function to load 64k - STACK_FRAME_OVERHEAD - __PT_SIZE (65192) value as unsigned. Reported-by: Benjamin Block <bblock@linux.ibm.com> Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by: Vasily Gorbik <gor@linux.ibm.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Kwiboo
pushed a commit
that referenced
this issue
Dec 15, 2018
conn_free() holds lock with spin_lock() and it is called by both
nf_conncount_lookup() and nf_conncount_gc_list(). nf_conncount_lookup()
is called from bottom-half context and nf_conncount_gc_list() from
process context. So that spin_lock() call is not safe. Hence
conn_free() should use spin_lock_bh() instead of spin_lock().
test commands:
%nft add table ip filter
%nft add chain ip filter input { type filter hook input priority 0\; }
%nft add rule filter input meter test { ip saddr ct count over 2 } \
counter
splat looks like:
[ 461.996507] ================================
[ 461.998999] WARNING: inconsistent lock state
[ 461.998999] 4.19.0-rc6+ #22 Not tainted
[ 461.998999] --------------------------------
[ 461.998999] inconsistent {IN-SOFTIRQ-W} -> {SOFTIRQ-ON-W} usage.
[ 461.998999] kworker/0:2/134 [HC0[0]:SC0[0]:HE1:SE1] takes:
[ 461.998999] 00000000a71a559a (&(&list->list_lock)->rlock){+.?.}, at: conn_free+0x69/0x2b0 [nf_conncount]
[ 461.998999] {IN-SOFTIRQ-W} state was registered at:
[ 461.998999] _raw_spin_lock+0x30/0x70
[ 461.998999] nf_conncount_add+0x28a/0x520 [nf_conncount]
[ 461.998999] nft_connlimit_eval+0x401/0x580 [nft_connlimit]
[ 461.998999] nft_dynset_eval+0x32b/0x590 [nf_tables]
[ 461.998999] nft_do_chain+0x497/0x1430 [nf_tables]
[ 461.998999] nft_do_chain_ipv4+0x255/0x330 [nf_tables]
[ 461.998999] nf_hook_slow+0xb1/0x160
[ ... ]
[ 461.998999] other info that might help us debug this:
[ 461.998999] Possible unsafe locking scenario:
[ 461.998999]
[ 461.998999] CPU0
[ 461.998999] ----
[ 461.998999] lock(&(&list->list_lock)->rlock);
[ 461.998999] <Interrupt>
[ 461.998999] lock(&(&list->list_lock)->rlock);
[ 461.998999]
[ 461.998999] *** DEADLOCK ***
[ 461.998999]
[ ... ]
Fixes: 5c789e1 ("netfilter: nf_conncount: Add list lock and gc worker, and RCU for init tree search")
Signed-off-by: Taehee Yoo <ap420073@gmail.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Kwiboo
pushed a commit
that referenced
this issue
Dec 15, 2018
nf_conncount_tuple is an element of nft_connlimit and that is deleted by
conn_free(). Elements can be deleted by both GC routine and data path
functions (nf_conncount_lookup, nf_conncount_add) and they call
conn_free() to free elements. But conn_free() only protects lists, not
each element. So that list_del corruption could occurred.
The conn_free() doesn't check whether element is already deleted. In
order to protect elements, dead flag is added. If an element is deleted,
dead flag is set. The only conn_free() can delete elements so that both
list lock and dead flag are enough to protect it.
test commands:
%nft add table ip filter
%nft add chain ip filter input { type filter hook input priority 0\; }
%nft add rule filter input meter test { ip id ct count over 2 } counter
splat looks like:
[ 1779.495778] list_del corruption, ffff8800b6e12008->prev is LIST_POISON2 (dead000000000200)
[ 1779.505453] ------------[ cut here ]------------
[ 1779.506260] kernel BUG at lib/list_debug.c:50!
[ 1779.515831] invalid opcode: 0000 [#1] SMP DEBUG_PAGEALLOC KASAN PTI
[ 1779.516772] CPU: 0 PID: 33 Comm: kworker/0:2 Not tainted 4.19.0-rc6+ #22
[ 1779.516772] Workqueue: events_power_efficient nft_rhash_gc [nf_tables_set]
[ 1779.516772] RIP: 0010:__list_del_entry_valid+0xd8/0x150
[ 1779.516772] Code: 39 48 83 c4 08 b8 01 00 00 00 5b 5d c3 48 89 ea 48 c7 c7 00 c3 5b 98 e8 0f dc 40 ff 0f 0b 48 c7 c7 60 c3 5b 98 e8 01 dc 40 ff <0f> 0b 48 c7 c7 c0 c3 5b 98 e8 f3 db 40 ff 0f 0b 48 c7 c7 20 c4 5b
[ 1779.516772] RSP: 0018:ffff880119127420 EFLAGS: 00010286
[ 1779.516772] RAX: 000000000000004e RBX: dead000000000200 RCX: 0000000000000000
[ 1779.516772] RDX: 000000000000004e RSI: 0000000000000008 RDI: ffffed0023224e7a
[ 1779.516772] RBP: ffff88011934bc10 R08: ffffed002367cea9 R09: ffffed002367cea9
[ 1779.516772] R10: 0000000000000001 R11: ffffed002367cea8 R12: ffff8800b6e12008
[ 1779.516772] R13: ffff8800b6e12010 R14: ffff88011934bc20 R15: ffff8800b6e12008
[ 1779.516772] FS: 0000000000000000(0000) GS:ffff88011b200000(0000) knlGS:0000000000000000
[ 1779.516772] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 1779.516772] CR2: 00007fc876534010 CR3: 000000010da16000 CR4: 00000000001006f0
[ 1779.516772] Call Trace:
[ 1779.516772] conn_free+0x9f/0x2b0 [nf_conncount]
[ 1779.516772] ? nf_ct_tmpl_alloc+0x2a0/0x2a0 [nf_conntrack]
[ 1779.516772] ? nf_conncount_add+0x520/0x520 [nf_conncount]
[ 1779.516772] ? do_raw_spin_trylock+0x1a0/0x1a0
[ 1779.516772] ? do_raw_spin_trylock+0x10/0x1a0
[ 1779.516772] find_or_evict+0xe5/0x150 [nf_conncount]
[ 1779.516772] nf_conncount_gc_list+0x162/0x360 [nf_conncount]
[ 1779.516772] ? nf_conncount_lookup+0xee0/0xee0 [nf_conncount]
[ 1779.516772] ? _raw_spin_unlock_irqrestore+0x45/0x50
[ 1779.516772] ? trace_hardirqs_off+0x6b/0x220
[ 1779.516772] ? trace_hardirqs_on_caller+0x220/0x220
[ 1779.516772] nft_rhash_gc+0x16b/0x540 [nf_tables_set]
[ ... ]
Fixes: 5c789e1 ("netfilter: nf_conncount: Add list lock and gc worker, and RCU for init tree search")
Signed-off-by: Taehee Yoo <ap420073@gmail.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Kwiboo
pushed a commit
that referenced
this issue
Jan 8, 2020
'chrdev_open()' calls 'cdev_get()' to obtain a reference to the 'struct cdev *' stashed in the 'i_cdev' field of the target inode structure. If the pointer is NULL, then it is initialised lazily by looking up the kobject in the 'cdev_map' and so the whole procedure is protected by the 'cdev_lock' spinlock to serialise initialisation of the shared pointer. Unfortunately, it is possible for the initialising thread to fail *after* installing the new pointer, for example if the subsequent '->open()' call on the file fails. In this case, 'cdev_put()' is called, the reference count on the kobject is dropped and, if nobody else has taken a reference, the release function is called which finally clears 'inode->i_cdev' from 'cdev_purge()' before potentially freeing the object. The problem here is that a racing thread can happily take the 'cdev_lock' and see the non-NULL pointer in the inode, which can result in a refcount increment from zero and a warning: | ------------[ cut here ]------------ | refcount_t: addition on 0; use-after-free. | WARNING: CPU: 2 PID: 6385 at lib/refcount.c:25 refcount_warn_saturate+0x6d/0xf0 | Modules linked in: | CPU: 2 PID: 6385 Comm: repro Not tainted 5.5.0-rc2+ #22 | Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.12.0-1 04/01/2014 | RIP: 0010:refcount_warn_saturate+0x6d/0xf0 | Code: 05 55 9a 15 01 01 e8 9d aa c8 ff 0f 0b c3 80 3d 45 9a 15 01 00 75 ce 48 c7 c7 00 9c 62 b3 c6 08 | RSP: 0018:ffffb524c1b9bc70 EFLAGS: 00010282 | RAX: 0000000000000000 RBX: ffff9e9da1f71390 RCX: 0000000000000000 | RDX: ffff9e9dbbd27618 RSI: ffff9e9dbbd18798 RDI: ffff9e9dbbd18798 | RBP: 0000000000000000 R08: 000000000000095f R09: 0000000000000039 | R10: 0000000000000000 R11: ffffb524c1b9bb20 R12: ffff9e9da1e8c700 | R13: ffffffffb25ee8b0 R14: 0000000000000000 R15: ffff9e9da1e8c700 | FS: 00007f3b87d26700(0000) GS:ffff9e9dbbd00000(0000) knlGS:0000000000000000 | CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 | CR2: 00007fc16909c000 CR3: 000000012df9c000 CR4: 00000000000006e0 | DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 | DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 | Call Trace: | kobject_get+0x5c/0x60 | cdev_get+0x2b/0x60 | chrdev_open+0x55/0x220 | ? cdev_put.part.3+0x20/0x20 | do_dentry_open+0x13a/0x390 | path_openat+0x2c8/0x1470 | do_filp_open+0x93/0x100 | ? selinux_file_ioctl+0x17f/0x220 | do_sys_open+0x186/0x220 | do_syscall_64+0x48/0x150 | entry_SYSCALL_64_after_hwframe+0x44/0xa9 | RIP: 0033:0x7f3b87efcd0e | Code: 89 54 24 08 e8 a3 f4 ff ff 8b 74 24 0c 48 8b 3c 24 41 89 c0 44 8b 54 24 08 b8 01 01 00 00 89 f4 | RSP: 002b:00007f3b87d259f0 EFLAGS: 00000293 ORIG_RAX: 0000000000000101 | RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007f3b87efcd0e | RDX: 0000000000000000 RSI: 00007f3b87d25a80 RDI: 00000000ffffff9c | RBP: 00007f3b87d25e90 R08: 0000000000000000 R09: 0000000000000000 | R10: 0000000000000000 R11: 0000000000000293 R12: 00007ffe188f504e | R13: 00007ffe188f504f R14: 00007f3b87d26700 R15: 0000000000000000 | ---[ end trace 24f53ca58db8180a ]--- Since 'cdev_get()' can already fail to obtain a reference, simply move it over to use 'kobject_get_unless_zero()' instead of 'kobject_get()', which will cause the racing thread to return -ENXIO if the initialising thread fails unexpectedly. Cc: Hillf Danton <hdanton@sina.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Al Viro <viro@zeniv.linux.org.uk> Reported-by: syzbot+82defefbbd8527e1c2cb@syzkaller.appspotmail.com Signed-off-by: Will Deacon <will@kernel.org> Cc: stable <stable@vger.kernel.org> Link: https://lore.kernel.org/r/20191219120203.32691-1-will@kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Kwiboo
pushed a commit
that referenced
this issue
Jul 20, 2020
The following deadlock was captured. The first process is holding 'kernfs_mutex' and hung by io. The io was staging in 'r1conf.pending_bio_list' of raid1 device, this pending bio list would be flushed by second process 'md127_raid1', but it was hung by 'kernfs_mutex'. Using sysfs_notify_dirent_safe() to replace sysfs_notify() can fix it. There were other sysfs_notify() invoked from io path, removed all of them. PID: 40430 TASK: ffff8ee9c8c65c40 CPU: 29 COMMAND: "probe_file" #0 [ffffb87c4df37260] __schedule at ffffffff9a8678ec #1 [ffffb87c4df372f8] schedule at ffffffff9a867f06 #2 [ffffb87c4df37310] io_schedule at ffffffff9a0c73e6 #3 [ffffb87c4df37328] __dta___xfs_iunpin_wait_3443 at ffffffffc03a4057 [xfs] #4 [ffffb87c4df373a0] xfs_iunpin_wait at ffffffffc03a6c79 [xfs] #5 [ffffb87c4df373b0] __dta_xfs_reclaim_inode_3357 at ffffffffc039a46c [xfs] #6 [ffffb87c4df37400] xfs_reclaim_inodes_ag at ffffffffc039a8b6 [xfs] #7 [ffffb87c4df37590] xfs_reclaim_inodes_nr at ffffffffc039bb33 [xfs] #8 [ffffb87c4df375b0] xfs_fs_free_cached_objects at ffffffffc03af0e9 [xfs] #9 [ffffb87c4df375c0] super_cache_scan at ffffffff9a287ec7 #10 [ffffb87c4df37618] shrink_slab at ffffffff9a1efd93 #11 [ffffb87c4df37700] shrink_node at ffffffff9a1f5968 #12 [ffffb87c4df37788] do_try_to_free_pages at ffffffff9a1f5ea2 #13 [ffffb87c4df377f0] try_to_free_mem_cgroup_pages at ffffffff9a1f6445 #14 [ffffb87c4df37880] try_charge at ffffffff9a26cc5f #15 [ffffb87c4df37920] memcg_kmem_charge_memcg at ffffffff9a270f6a #16 [ffffb87c4df37958] new_slab at ffffffff9a251430 #17 [ffffb87c4df379c0] ___slab_alloc at ffffffff9a251c85 #18 [ffffb87c4df37a80] __slab_alloc at ffffffff9a25635d #19 [ffffb87c4df37ac0] kmem_cache_alloc at ffffffff9a251f89 #20 [ffffb87c4df37b00] alloc_inode at ffffffff9a2a2b10 #21 [ffffb87c4df37b20] iget_locked at ffffffff9a2a4854 #22 [ffffb87c4df37b60] kernfs_get_inode at ffffffff9a311377 #23 [ffffb87c4df37b80] kernfs_iop_lookup at ffffffff9a311e2b #24 [ffffb87c4df37ba8] lookup_slow at ffffffff9a290118 #25 [ffffb87c4df37c10] walk_component at ffffffff9a291e83 #26 [ffffb87c4df37c78] path_lookupat at ffffffff9a293619 #27 [ffffb87c4df37cd8] filename_lookup at ffffffff9a2953af #28 [ffffb87c4df37de8] user_path_at_empty at ffffffff9a295566 #29 [ffffb87c4df37e10] vfs_statx at ffffffff9a289787 #30 [ffffb87c4df37e70] SYSC_newlstat at ffffffff9a289d5d #31 [ffffb87c4df37f18] sys_newlstat at ffffffff9a28a60e #32 [ffffb87c4df37f28] do_syscall_64 at ffffffff9a003949 #33 [ffffb87c4df37f50] entry_SYSCALL_64_after_hwframe at ffffffff9aa001ad RIP: 00007f617a5f2905 RSP: 00007f607334f838 RFLAGS: 00000246 RAX: ffffffffffffffda RBX: 00007f6064044b20 RCX: 00007f617a5f2905 RDX: 00007f6064044b20 RSI: 00007f6064044b20 RDI: 00007f6064005890 RBP: 00007f6064044aa0 R8: 0000000000000030 R9: 000000000000011c R10: 0000000000000013 R11: 0000000000000246 R12: 00007f606417e6d0 R13: 00007f6064044aa0 R14: 00007f6064044b10 R15: 00000000ffffffff ORIG_RAX: 0000000000000006 CS: 0033 SS: 002b PID: 927 TASK: ffff8f15ac5dbd80 CPU: 42 COMMAND: "md127_raid1" #0 [ffffb87c4df07b28] __schedule at ffffffff9a8678ec #1 [ffffb87c4df07bc0] schedule at ffffffff9a867f06 #2 [ffffb87c4df07bd8] schedule_preempt_disabled at ffffffff9a86825e #3 [ffffb87c4df07be8] __mutex_lock at ffffffff9a869bcc #4 [ffffb87c4df07ca0] __mutex_lock_slowpath at ffffffff9a86a013 #5 [ffffb87c4df07cb0] mutex_lock at ffffffff9a86a04f #6 [ffffb87c4df07cc8] kernfs_find_and_get_ns at ffffffff9a311d83 #7 [ffffb87c4df07cf0] sysfs_notify at ffffffff9a314b3a #8 [ffffb87c4df07d18] md_update_sb at ffffffff9a688696 #9 [ffffb87c4df07d98] md_update_sb at ffffffff9a6886d5 #10 [ffffb87c4df07da8] md_check_recovery at ffffffff9a68ad9c #11 [ffffb87c4df07dd0] raid1d at ffffffffc01f0375 [raid1] #12 [ffffb87c4df07ea0] md_thread at ffffffff9a680348 #13 [ffffb87c4df07f08] kthread at ffffffff9a0b8005 #14 [ffffb87c4df07f50] ret_from_fork at ffffffff9aa00344 Signed-off-by: Junxiao Bi <junxiao.bi@oracle.com> Signed-off-by: Song Liu <songliubraving@fb.com>
Kwiboo
pushed a commit
that referenced
this issue
Nov 10, 2020
'chrdev_open()' calls 'cdev_get()' to obtain a reference to the 'struct cdev *' stashed in the 'i_cdev' field of the target inode structure. If the pointer is NULL, then it is initialised lazily by looking up the kobject in the 'cdev_map' and so the whole procedure is protected by the 'cdev_lock' spinlock to serialise initialisation of the shared pointer. Unfortunately, it is possible for the initialising thread to fail *after* installing the new pointer, for example if the subsequent '->open()' call on the file fails. In this case, 'cdev_put()' is called, the reference count on the kobject is dropped and, if nobody else has taken a reference, the release function is called which finally clears 'inode->i_cdev' from 'cdev_purge()' before potentially freeing the object. The problem here is that a racing thread can happily take the 'cdev_lock' and see the non-NULL pointer in the inode, which can result in a refcount increment from zero and a warning: | ------------[ cut here ]------------ | refcount_t: addition on 0; use-after-free. | WARNING: CPU: 2 PID: 6385 at lib/refcount.c:25 refcount_warn_saturate+0x6d/0xf0 | Modules linked in: | CPU: 2 PID: 6385 Comm: repro Not tainted 5.5.0-rc2+ #22 | Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.12.0-1 04/01/2014 | RIP: 0010:refcount_warn_saturate+0x6d/0xf0 | Code: 05 55 9a 15 01 01 e8 9d aa c8 ff 0f 0b c3 80 3d 45 9a 15 01 00 75 ce 48 c7 c7 00 9c 62 b3 c6 08 | RSP: 0018:ffffb524c1b9bc70 EFLAGS: 00010282 | RAX: 0000000000000000 RBX: ffff9e9da1f71390 RCX: 0000000000000000 | RDX: ffff9e9dbbd27618 RSI: ffff9e9dbbd18798 RDI: ffff9e9dbbd18798 | RBP: 0000000000000000 R08: 000000000000095f R09: 0000000000000039 | R10: 0000000000000000 R11: ffffb524c1b9bb20 R12: ffff9e9da1e8c700 | R13: ffffffffb25ee8b0 R14: 0000000000000000 R15: ffff9e9da1e8c700 | FS: 00007f3b87d26700(0000) GS:ffff9e9dbbd00000(0000) knlGS:0000000000000000 | CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 | CR2: 00007fc16909c000 CR3: 000000012df9c000 CR4: 00000000000006e0 | DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 | DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 | Call Trace: | kobject_get+0x5c/0x60 | cdev_get+0x2b/0x60 | chrdev_open+0x55/0x220 | ? cdev_put.part.3+0x20/0x20 | do_dentry_open+0x13a/0x390 | path_openat+0x2c8/0x1470 | do_filp_open+0x93/0x100 | ? selinux_file_ioctl+0x17f/0x220 | do_sys_open+0x186/0x220 | do_syscall_64+0x48/0x150 | entry_SYSCALL_64_after_hwframe+0x44/0xa9 | RIP: 0033:0x7f3b87efcd0e | Code: 89 54 24 08 e8 a3 f4 ff ff 8b 74 24 0c 48 8b 3c 24 41 89 c0 44 8b 54 24 08 b8 01 01 00 00 89 f4 | RSP: 002b:00007f3b87d259f0 EFLAGS: 00000293 ORIG_RAX: 0000000000000101 | RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007f3b87efcd0e | RDX: 0000000000000000 RSI: 00007f3b87d25a80 RDI: 00000000ffffff9c | RBP: 00007f3b87d25e90 R08: 0000000000000000 R09: 0000000000000000 | R10: 0000000000000000 R11: 0000000000000293 R12: 00007ffe188f504e | R13: 00007ffe188f504f R14: 00007f3b87d26700 R15: 0000000000000000 | ---[ end trace 24f53ca58db8180a ]--- Since 'cdev_get()' can already fail to obtain a reference, simply move it over to use 'kobject_get_unless_zero()' instead of 'kobject_get()', which will cause the racing thread to return -ENXIO if the initialising thread fails unexpectedly. Cc: Hillf Danton <hdanton@sina.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Al Viro <viro@zeniv.linux.org.uk> Reported-by: syzbot+82defefbbd8527e1c2cb@syzkaller.appspotmail.com Signed-off-by: Will Deacon <will@kernel.org> Cc: stable <stable@vger.kernel.org> Link: https://lore.kernel.org/r/20191219120203.32691-1-will@kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Wenping Zhang <wenping.zhang@rock-chips.com> Change-Id: Idc4211025f422fdc4440a23ad491b5ccc459d4bd (cherry picked from commit 68faa67)
Kwiboo
pushed a commit
that referenced
this issue
Nov 29, 2020
[ Upstream commit 4ff753f ] When an UE or memory error exception is encountered the MCE handler tries to find the pfn using addr_to_pfn() which takes effective address as an argument, later pfn is used to poison the page where memory error occurred, recent rework in this area made addr_to_pfn to run in real mode, which can be fatal as it may try to access memory outside RMO region. Have two helper functions to separate things to be done in real mode and virtual mode without changing any functionality. This also fixes the following error as the use of addr_to_pfn is now moved to virtual mode. Without this change following kernel crash is seen on hitting UE. [ 485.128036] Oops: Kernel access of bad area, sig: 11 [#1] [ 485.128040] LE SMP NR_CPUS=2048 NUMA pSeries [ 485.128047] Modules linked in: [ 485.128067] CPU: 15 PID: 6536 Comm: insmod Kdump: loaded Tainted: G OE 5.7.0 #22 [ 485.128074] NIP: c00000000009b24c LR: c0000000000398d8 CTR: c000000000cd57c0 [ 485.128078] REGS: c000000003f1f970 TRAP: 0300 Tainted: G OE (5.7.0) [ 485.128082] MSR: 8000000000001003 <SF,ME,RI,LE> CR: 28008284 XER: 00000001 [ 485.128088] CFAR: c00000000009b190 DAR: c0000001fab00000 DSISR: 40000000 IRQMASK: 1 [ 485.128088] GPR00: 0000000000000001 c000000003f1fbf0 c000000001634300 0000b0fa01000000 [ 485.128088] GPR04: d000000002220000 0000000000000000 00000000fab00000 0000000000000022 [ 485.128088] GPR08: c0000001fab00000 0000000000000000 c0000001fab00000 c000000003f1fc14 [ 485.128088] GPR12: 0000000000000008 c000000003ff5880 d000000002100008 0000000000000000 [ 485.128088] GPR16: 000000000000ff20 000000000000fff1 000000000000fff2 d0000000021a1100 [ 485.128088] GPR20: d000000002200000 c00000015c893c50 c000000000d49b28 c00000015c893c50 [ 485.128088] GPR24: d0000000021a0d08 c0000000014e5da8 d0000000021a0818 000000000000000a [ 485.128088] GPR28: 0000000000000008 000000000000000a c0000000017e2970 000000000000000a [ 485.128125] NIP [c00000000009b24c] __find_linux_pte+0x11c/0x310 [ 485.128130] LR [c0000000000398d8] addr_to_pfn+0x138/0x170 [ 485.128133] Call Trace: [ 485.128135] Instruction dump: [ 485.128138] 3929ffff 7d4a3378 7c883c36 7d2907b4 794a1564 7d294038 794af082 3900ffff [ 485.128144] 79291f24 790af00e 78e70020 7d095214 <7c69502a> 2fa30000 419e011c 70690040 [ 485.128152] ---[ end trace d34b27e29ae0e340 ]--- Fixes: 9ca766f ("powerpc/64s/pseries: machine check convert to use common event code") Signed-off-by: Ganesh Goudar <ganeshgr@linux.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200724063946.21378-1-ganeshgr@linux.ibm.com Signed-off-by: Sasha Levin <sashal@kernel.org>
rubenvb
pushed a commit
to rubenvb/linux-rockchip
that referenced
this issue
Jan 3, 2021
[ Upstream commit 4e79f02 ] When running in BE mode on LPAE hardware with a PA-to-VA translation that exceeds 4 GB, we patch bits 39:32 of the offset into the wrong byte of the opcode. So fix that, by rotating the offset in r0 to the right by 8 bits, which will put the 8-bit immediate in bits 31:24. Note that this will also move bit Kwiboo#22 in its correct place when applying the rotation to the constant #0x400000. Fixes: d9a790d ("ARM: 7883/1: fix mov to mvn conversion in case of 64 bit phys_addr_t and BE") Acked-by: Nicolas Pitre <nico@fluxnic.net> Reviewed-by: Linus Walleij <linus.walleij@linaro.org> Signed-off-by: Ard Biesheuvel <ardb@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
We are having issues to play our HDR samples on RK3328.
One of the samples that shows that issue is http://demo-uhd3d.com/fiche.php?cat=uhd&id=159
From what we see, MPP seems to error while parsing the data, this is the log we have been able to retrieve :
The text was updated successfully, but these errors were encountered: