Config: need auto detection and fallback between SOF/SST/Legacy drivers #86
Assignees
Labels
Comments
|
@keyonjie could you help to take care of this? |
|
This may need refine/unified probing(especially for sof and sst) part, furthermore, sometimes using sof or sst is totally an user end choice(e.g. on those no FW authentication skl- platforms), we can't make selection for end-user in hard-coded logic. I thought of using exclusive kernel config item for SOF and SST, which means if you select one(the other one will be unselected automatically), but the cons of doing that is we can't configure e.g byt with SST + apl with SOF, I am afraid someone make want to try something like that, even though this is not important for me. @lgirdwood any comments? |
mengdonglin
added
enhancement
|
This is an enhancement actually, so set label "enhancement" |
|
@keyonjie We will have a single PCI and single ACPI entry point for all Intel SOF and legacy drivers. Drivers will default to legacy unless they are mentioned in the SOF lookup table (table of DMI names, DT names). |
lyakh
pushed a commit
to lyakh/linux
that referenced
this issue
Nov 19, 2018
Signed-off-by: Phil Elwell <[email protected]> suppress spurious messages Add #if for 3.14 kernel change (thesofproject#87) Fixes compiling after changes in torvalds@f663dd9 and torvalds@99932d4 Fixes thesofproject#86 Set dev_type to wlan Fixes thesofproject#23 Tentatively added support for more 8188CUS based devices. Add support for more 8188CUS and 8192CUS devices Add ProductId for the Netgear N150 - WNA1000M Fixes CONFIG_CONCURRENT_MODE CONFIG_MULTI_VIR_IFACES Fixes compatibility with 3.13 Enables warning in the compiler and fixes some issues, reference => https://github.com/diederikdehaas/rtl8812AU Starts device in station mode instead of monitor, fixes NetworkManager issues Enable cfg80211 support Fix cfg80211 for kernel >= 4.7 Fixes rtl8192cu for kernel >= 4.8 rtl8192: Fixup build fixup: rtl8192cu fixes from milhouse rtl8192: switch to netdev->priv_destructor() When trying to build from the rpi-4.11.y branch, I'm getting the following error : drivers/net/wireless/realtek/rtl8192cu/os_dep/linux/ioctl_cfg80211.c:3464:10: error: 'struct net_device' has no member named 'destructor' It seems to occur since this upstream commit : torvalds@cf124db [...] netdev->priv_destructor() performs all actions to free up the private resources that used to be freed by netdev->destructor(), except for free_netdev(). netdev->needs_free_netdev is a boolean that indicates whether free_netdev() should be done at the end of unregister_netdevice(). Signed-off-by: Bilal Amarni <[email protected]> ARM64: Fix build break for RTL8187/RTL8192CU wifi These drivers use an ASM function from the base system to compute the ipv6 checksum. These functions are not available on ARM64, probably because nobody has bother to write them. The base system does have a generic "C" version, so a simple fix is to include the header to use the generic version on ARM64 only. A longer term solution would be to submit the necessary ASM function to the upstream source. With this change, these drivers now compile without any errors on ARM64. Signed-off-by: Michael Zoran <[email protected]>
|
Mechanism implemented upstream (will be in 4.21), SOF needs to implement the PCI class/subclass detection as described in #468 |
fredoh9
pushed a commit
to fredoh9/linux
that referenced
this issue
Aug 5, 2019
[root@localhost ~]# ./liburing/test/link QEMU Standard PC report that: [ 29.379892] CPU: 0 PID: 84 Comm: kworker/u2:2 Not tainted 5.3.0-rc2-00051-g4010b622f1d2-dirty thesofproject#86 [ 29.379902] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.12.0-1 04/01/2014 [ 29.379913] Workqueue: io_ring-wq io_sq_wq_submit_work [ 29.379929] Call Trace: [ 29.379953] dump_stack+0xa9/0x10e [ 29.379970] ? io_sq_wq_submit_work+0xbf4/0xe90 [ 29.379986] print_address_description.cold.6+0x9/0x317 [ 29.379999] ? io_sq_wq_submit_work+0xbf4/0xe90 [ 29.380010] ? io_sq_wq_submit_work+0xbf4/0xe90 [ 29.380026] __kasan_report.cold.7+0x1a/0x34 [ 29.380044] ? io_sq_wq_submit_work+0xbf4/0xe90 [ 29.380061] kasan_report+0xe/0x12 [ 29.380076] io_sq_wq_submit_work+0xbf4/0xe90 [ 29.380104] ? io_sq_thread+0xaf0/0xaf0 [ 29.380152] process_one_work+0xb59/0x19e0 [ 29.380184] ? pwq_dec_nr_in_flight+0x2c0/0x2c0 [ 29.380221] worker_thread+0x8c/0xf40 [ 29.380248] ? __kthread_parkme+0xab/0x110 [ 29.380265] ? process_one_work+0x19e0/0x19e0 [ 29.380278] kthread+0x30b/0x3d0 [ 29.380292] ? kthread_create_on_node+0xe0/0xe0 [ 29.380311] ret_from_fork+0x3a/0x50 [ 29.380635] Allocated by task 209: [ 29.381255] save_stack+0x19/0x80 [ 29.381268] __kasan_kmalloc.constprop.6+0xc1/0xd0 [ 29.381279] kmem_cache_alloc+0xc0/0x240 [ 29.381289] io_submit_sqe+0x11bc/0x1c70 [ 29.381300] io_ring_submit+0x174/0x3c0 [ 29.381311] __x64_sys_io_uring_enter+0x601/0x780 [ 29.381322] do_syscall_64+0x9f/0x4d0 [ 29.381336] entry_SYSCALL_64_after_hwframe+0x49/0xbe [ 29.381633] Freed by task 84: [ 29.382186] save_stack+0x19/0x80 [ 29.382198] __kasan_slab_free+0x11d/0x160 [ 29.382210] kmem_cache_free+0x8c/0x2f0 [ 29.382220] io_put_req+0x22/0x30 [ 29.382230] io_sq_wq_submit_work+0x28b/0xe90 [ 29.382241] process_one_work+0xb59/0x19e0 [ 29.382251] worker_thread+0x8c/0xf40 [ 29.382262] kthread+0x30b/0x3d0 [ 29.382272] ret_from_fork+0x3a/0x50 [ 29.382569] The buggy address belongs to the object at ffff888067172140 which belongs to the cache io_kiocb of size 224 [ 29.384692] The buggy address is located 120 bytes inside of 224-byte region [ffff888067172140, ffff888067172220) [ 29.386723] The buggy address belongs to the page: [ 29.387575] page:ffffea00019c5c80 refcount:1 mapcount:0 mapping:ffff88806ace5180 index:0x0 [ 29.387587] flags: 0x100000000000200(slab) [ 29.387603] raw: 0100000000000200 dead000000000100 dead000000000122 ffff88806ace5180 [ 29.387617] raw: 0000000000000000 00000000800c000c 00000001ffffffff 0000000000000000 [ 29.387624] page dumped because: kasan: bad access detected [ 29.387920] Memory state around the buggy address: [ 29.388771] ffff888067172080: fb fb fb fb fb fb fb fb fb fb fb fb fc fc fc fc [ 29.390062] ffff888067172100: fc fc fc fc fc fc fc fc fb fb fb fb fb fb fb fb [ 29.391325] >ffff888067172180: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb [ 29.392578] ^ [ 29.393480] ffff888067172200: fb fb fb fb fc fc fc fc fc fc fc fc fc fc fc fc [ 29.394744] ffff888067172280: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc [ 29.396003] ================================================================== [ 29.397260] Disabling lock debugging due to kernel taint io_sq_wq_submit_work free and read req again. Cc: Zhengyuan Liu <[email protected]> Cc: [email protected] Cc: [email protected] Fixes: f7b76ac ("io_uring: fix counter inc/dec mismatch in async_list") Signed-off-by: Jackie Liu <[email protected]> Signed-off-by: Jens Axboe <[email protected]>
plbossart
pushed a commit
that referenced
this issue
Oct 30, 2019
Scheduled policy update work may end up racing with the freeing of the policy and unregistering the driver. One possible race is as below, where the cpufreq_driver is unregistered, but the scheduled work gets executed at later stage when, cpufreq_driver is NULL (i.e. after freeing the policy and driver). Unable to handle kernel NULL pointer dereference at virtual address 0000001c pgd = (ptrval) [0000001c] *pgd=80000080204003, *pmd=00000000 Internal error: Oops: 206 [#1] SMP THUMB2 Modules linked in: CPU: 0 PID: 34 Comm: kworker/0:1 Not tainted 5.4.0-rc3-00006-g67f5a8081a4b #86 Hardware name: ARM-Versatile Express Workqueue: events handle_update PC is at cpufreq_set_policy+0x58/0x228 LR is at dev_pm_qos_read_value+0x77/0xac Control: 70c5387d Table: 80203000 DAC: fffffffd Process kworker/0:1 (pid: 34, stack limit = 0x(ptrval)) (cpufreq_set_policy) from (refresh_frequency_limits.part.24+0x37/0x48) (refresh_frequency_limits.part.24) from (handle_update+0x2f/0x38) (handle_update) from (process_one_work+0x16d/0x3cc) (process_one_work) from (worker_thread+0xff/0x414) (worker_thread) from (kthread+0xff/0x100) (kthread) from (ret_from_fork+0x11/0x28) Fixes: 67d874c ("cpufreq: Register notifiers with the PM QoS framework") Signed-off-by: Sudeep Holla <[email protected]> [ rjw: Cancel the work before dropping the QoS requests ] Acked-by: Viresh Kumar <[email protected]> Signed-off-by: Rafael J. Wysocki <[email protected]>
plbossart
pushed a commit
that referenced
this issue
Sep 28, 2020
The commit eb1f002 ("lockdep,trace: Expose tracepoints"), started to expose us for tracepoints. This lead to the following RCU splat on an ARM64 Qcom board. [ 5.529634] WARNING: suspicious RCU usage [ 5.537307] sdhci-pltfm: SDHCI platform and OF driver helper [ 5.541092] 5.9.0-rc3 #86 Not tainted [ 5.541098] ----------------------------- [ 5.541105] ../include/trace/events/lock.h:37 suspicious rcu_dereference_check() usage! [ 5.541110] [ 5.541110] other info that might help us debug this: [ 5.541110] [ 5.541116] [ 5.541116] rcu_scheduler_active = 2, debug_locks = 1 [ 5.541122] RCU used illegally from extended quiescent state! [ 5.541129] no locks held by swapper/0/0. [ 5.541134] [ 5.541134] stack backtrace: [ 5.541143] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 5.9.0-rc3 #86 [ 5.541149] Hardware name: Qualcomm Technologies, Inc. APQ 8016 SBC (DT) [ 5.541157] Call trace: [ 5.568185] sdhci_msm 7864900.sdhci: Got CD GPIO [ 5.574186] dump_backtrace+0x0/0x1c8 [ 5.574206] show_stack+0x14/0x20 [ 5.574229] dump_stack+0xe8/0x154 [ 5.574250] lockdep_rcu_suspicious+0xd4/0xf8 [ 5.574269] lock_acquire+0x3f0/0x460 [ 5.574292] _raw_spin_lock_irqsave+0x80/0xb0 [ 5.574314] __pm_runtime_suspend+0x4c/0x188 [ 5.574341] psci_enter_domain_idle_state+0x40/0xa0 [ 5.574362] cpuidle_enter_state+0xc0/0x610 [ 5.646487] cpuidle_enter+0x38/0x50 [ 5.650651] call_cpuidle+0x18/0x40 [ 5.654467] do_idle+0x228/0x278 [ 5.657678] cpu_startup_entry+0x24/0x70 [ 5.661153] rest_init+0x1a4/0x278 [ 5.665061] arch_call_rest_init+0xc/0x14 [ 5.668272] start_kernel+0x508/0x540 Following the path in pm_runtime_put_sync_suspend() from psci_enter_domain_idle_state(), it seems like we end up using the RCU. Therefore, let's simply silence the splat by informing the RCU about it with RCU_NONIDLE. Note that, this is a temporary solution. Instead we should strive to avoid using RCU_NONIDLE (and similar), but rather push rcu_idle_enter|exit() further down, closer to the arch specific code. However, as the CPU PM notifiers are also using the RCU, additional rework is needed. Reported-by: Naresh Kamboju <[email protected]> Signed-off-by: Ulf Hansson <[email protected]> Acked-by: Paul E. McKenney <[email protected]> Signed-off-by: Rafael J. Wysocki <[email protected]>
aiChaoSONG
pushed a commit
to aiChaoSONG/linux
that referenced
this issue
May 6, 2021
…rate libkernel: remove leftover `extern crate`
bardliao
pushed a commit
to bardliao/linux
that referenced
this issue
Jul 14, 2021
Fixes checkpatch issues: ERROR: code indent should use tabs where possible thesofproject#86: FILE: drivers/reset/reset-zynqmp.c:86: + .reset_id = 0,$ WARNING: please, no spaces at the start of a line thesofproject#86: FILE: drivers/reset/reset-zynqmp.c:86: + .reset_id = 0,$ ERROR: code indent should use tabs where possible thesofproject#87: FILE: drivers/reset/reset-zynqmp.c:87: + .num_resets = VERSAL_NR_RESETS,$ WARNING: please, no spaces at the start of a line thesofproject#87: FILE: drivers/reset/reset-zynqmp.c:87: + .num_resets = VERSAL_NR_RESETS,$ Signed-off-by: Philipp Zabel <[email protected]> Reviewed-by: Michal Simek <[email protected]>
plbossart
pushed a commit
that referenced
this issue
Jul 26, 2021
Liajian reported a bug_on hit on a ThunderX2 arm64 server with FastLinQ QL41000 ethernet controller: BUG: scheduling while atomic: kworker/0:4/531/0x00000200 [qed_probe:488()]hw prepare failed kernel BUG at mm/vmalloc.c:2355! Internal error: Oops - BUG: 0 [#1] SMP CPU: 0 PID: 531 Comm: kworker/0:4 Tainted: G W 5.4.0-77-generic #86-Ubuntu pstate: 00400009 (nzcv daif +PAN -UAO) Call trace: vunmap+0x4c/0x50 iounmap+0x48/0x58 qed_free_pci+0x60/0x80 [qed] qed_probe+0x35c/0x688 [qed] __qede_probe+0x88/0x5c8 [qede] qede_probe+0x60/0xe0 [qede] local_pci_probe+0x48/0xa0 work_for_cpu_fn+0x24/0x38 process_one_work+0x1d0/0x468 worker_thread+0x238/0x4e0 kthread+0xf0/0x118 ret_from_fork+0x10/0x18 In this case, qed_hw_prepare() returns error due to hw/fw error, but in theory work queue should be in process context instead of interrupt. The root cause might be the unpaired spin_{un}lock_bh() in _qed_mcp_cmd_and_union(), which causes botton half is disabled incorrectly. Reported-by: Lijian Zhang <[email protected]> Signed-off-by: Jia He <[email protected]> Signed-off-by: David S. Miller <[email protected]>
plbossart
pushed a commit
that referenced
this issue
Jul 26, 2021
In some cases skb head could be locked and entire header data is pulled from skb. When skb_zerocopy() called in such cases, following BUG is triggered. This patch fixes it by copying entire skb in such cases. This could be optimized incase this is performance bottleneck. ---8<--- kernel BUG at net/core/skbuff.c:2961! invalid opcode: 0000 [#1] SMP PTI CPU: 2 PID: 0 Comm: swapper/2 Tainted: G OE 5.4.0-77-generic #86-Ubuntu Hardware name: OpenStack Foundation OpenStack Nova, BIOS 1.13.0-1ubuntu1.1 04/01/2014 RIP: 0010:skb_zerocopy+0x37a/0x3a0 RSP: 0018:ffffbcc70013ca38 EFLAGS: 00010246 Call Trace: <IRQ> queue_userspace_packet+0x2af/0x5e0 [openvswitch] ovs_dp_upcall+0x3d/0x60 [openvswitch] ovs_dp_process_packet+0x125/0x150 [openvswitch] ovs_vport_receive+0x77/0xd0 [openvswitch] netdev_port_receive+0x87/0x130 [openvswitch] netdev_frame_hook+0x4b/0x60 [openvswitch] __netif_receive_skb_core+0x2b4/0xc90 __netif_receive_skb_one_core+0x3f/0xa0 __netif_receive_skb+0x18/0x60 process_backlog+0xa9/0x160 net_rx_action+0x142/0x390 __do_softirq+0xe1/0x2d6 irq_exit+0xae/0xb0 do_IRQ+0x5a/0xf0 common_interrupt+0xf/0xf Code that triggered BUG: int skb_zerocopy(struct sk_buff *to, struct sk_buff *from, int len, int hlen) { int i, j = 0; int plen = 0; /* length of skb->head fragment */ int ret; struct page *page; unsigned int offset; BUG_ON(!from->head_frag && !hlen); Signed-off-by: Pravin B Shelar <[email protected]> Signed-off-by: David S. Miller <[email protected]>
plbossart
pushed a commit
that referenced
this issue
Nov 29, 2022
…ging The following bug is reported to be triggered when starting X on x86-32 system with i915: [ 225.777375] kernel BUG at mm/memory.c:2664! [ 225.777391] invalid opcode: 0000 [#1] PREEMPT SMP [ 225.777405] CPU: 0 PID: 2402 Comm: Xorg Not tainted 6.1.0-rc3-bdg+ #86 [ 225.777415] Hardware name: /8I865G775-G, BIOS F1 08/29/2006 [ 225.777421] EIP: __apply_to_page_range+0x24d/0x31c [ 225.777437] Code: ff ff 8b 55 e8 8b 45 cc e8 0a 11 ec ff 89 d8 83 c4 28 5b 5e 5f 5d c3 81 7d e0 a0 ef 96 c1 74 ad 8b 45 d0 e8 2d 83 49 00 eb a3 <0f> 0b 25 00 f0 ff ff 81 eb 00 00 00 40 01 c3 8b 45 ec 8b 00 e8 76 [ 225.777446] EAX: 00000001 EBX: c53a3b58 ECX: b5c00000 EDX: c258aa00 [ 225.777454] ESI: b5c00000 EDI: b5900000 EBP: c4b0fdb4 ESP: c4b0fd80 [ 225.777462] DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0068 EFLAGS: 00010202 [ 225.777470] CR0: 80050033 CR2: b5900000 CR3: 053a3000 CR4: 000006d0 [ 225.777479] Call Trace: [ 225.777486] ? i915_memcpy_init_early+0x63/0x63 [i915] [ 225.777684] apply_to_page_range+0x21/0x27 [ 225.777694] ? i915_memcpy_init_early+0x63/0x63 [i915] [ 225.777870] remap_io_mapping+0x49/0x75 [i915] [ 225.778046] ? i915_memcpy_init_early+0x63/0x63 [i915] [ 225.778220] ? mutex_unlock+0xb/0xd [ 225.778231] ? i915_vma_pin_fence+0x6d/0xf7 [i915] [ 225.778420] vm_fault_gtt+0x2a9/0x8f1 [i915] [ 225.778644] ? lock_is_held_type+0x56/0xe7 [ 225.778655] ? lock_is_held_type+0x7a/0xe7 [ 225.778663] ? 0xc1000000 [ 225.778670] __do_fault+0x21/0x6a [ 225.778679] handle_mm_fault+0x708/0xb21 [ 225.778686] ? mt_find+0x21e/0x5ae [ 225.778696] exc_page_fault+0x185/0x705 [ 225.778704] ? doublefault_shim+0x127/0x127 [ 225.778715] handle_exception+0x130/0x130 [ 225.778723] EIP: 0xb700468a Recently pud_huge() got aware of non-present entry by commit 3a194f3 ("mm/hugetlb: make pud_huge() and follow_huge_pud() aware of non-present pud entry") to handle some special states of gigantic page. However, it's overlooked that pud_none() always returns false when running with 2-level paging, and as a result pud_huge() can return true pointlessly. Introduce "#if CONFIG_PGTABLE_LEVELS > 2" to pud_huge() to deal with this. Link: https://lkml.kernel.org/r/[email protected] Fixes: 3a194f3 ("mm/hugetlb: make pud_huge() and follow_huge_pud() aware of non-present pud entry") Signed-off-by: Naoya Horiguchi <[email protected]> Reported-by: Ville Syrjälä <[email protected]> Tested-by: Ville Syrjälä <[email protected]> Reviewed-by: Miaohe Lin <[email protected]> Cc: David Hildenbrand <[email protected]> Cc: Liu Shixin <[email protected]> Cc: Mike Kravetz <[email protected]> Cc: Muchun Song <[email protected]> Cc: Oscar Salvador <[email protected]> Cc: Yang Shi <[email protected]> Cc: Thomas Gleixner <[email protected]> Cc: Ingo Molnar <[email protected]> Cc: Borislav Petkov <[email protected]> Cc: Dave Hansen <[email protected]> Cc: <[email protected]> Signed-off-by: Andrew Morton <[email protected]>
Vamshigopal
pushed a commit
to Vamshigopal/linux
that referenced
this issue
Jan 24, 2023
[ Upstream commit 031af50 ] The inline assembly for arm64's cmpxchg_double*() implementations use a +Q constraint to hazard against other accesses to the memory location being exchanged. However, the pointer passed to the constraint is a pointer to unsigned long, and thus the hazard only applies to the first 8 bytes of the location. GCC can take advantage of this, assuming that other portions of the location are unchanged, leading to a number of potential problems. This is similar to what we fixed back in commit: fee960b ("arm64: xchg: hazard against entire exchange variable") ... but we forgot to adjust cmpxchg_double*() similarly at the same time. The same problem applies, as demonstrated with the following test: | struct big { | u64 lo, hi; | } __aligned(128); | | unsigned long foo(struct big *b) | { | u64 hi_old, hi_new; | | hi_old = b->hi; | cmpxchg_double_local(&b->lo, &b->hi, 0x12, 0x34, 0x56, 0x78); | hi_new = b->hi; | | return hi_old ^ hi_new; | } ... which GCC 12.1.0 compiles as: | 0000000000000000 <foo>: | 0: d503233f paciasp | 4: aa0003e4 mov x4, x0 | 8: 1400000e b 40 <foo+0x40> | c: d2800240 mov x0, #0x12 // thesofproject#18 | 10: d2800681 mov x1, #0x34 // thesofproject#52 | 14: aa0003e5 mov x5, x0 | 18: aa0103e6 mov x6, x1 | 1c: d2800ac2 mov x2, #0x56 // thesofproject#86 | 20: d2800f03 mov x3, #0x78 // thesofproject#120 | 24: 48207c82 casp x0, x1, x2, x3, [x4] | 28: ca050000 eor x0, x0, x5 | 2c: ca060021 eor x1, x1, x6 | 30: aa010000 orr x0, x0, x1 | 34: d2800000 mov x0, #0x0 // #0 <--- BANG | 38: d50323bf autiasp | 3c: d65f03c0 ret | 40: d2800240 mov x0, #0x12 // thesofproject#18 | 44: d2800681 mov x1, #0x34 // thesofproject#52 | 48: d2800ac2 mov x2, #0x56 // thesofproject#86 | 4c: d2800f03 mov x3, #0x78 // thesofproject#120 | 50: f9800091 prfm pstl1strm, [x4] | 54: c87f1885 ldxp x5, x6, [x4] | 58: ca0000a5 eor x5, x5, x0 | 5c: ca0100c6 eor x6, x6, x1 | 60: aa0600a6 orr x6, x5, x6 | 64: b5000066 cbnz x6, 70 <foo+0x70> | 68: c8250c82 stxp w5, x2, x3, [x4] | 6c: 35ffff45 cbnz w5, 54 <foo+0x54> | 70: d2800000 mov x0, #0x0 // #0 <--- BANG | 74: d50323bf autiasp | 78: d65f03c0 ret Notice that at the lines with "BANG" comments, GCC has assumed that the higher 8 bytes are unchanged by the cmpxchg_double() call, and that `hi_old ^ hi_new` can be reduced to a constant zero, for both LSE and LL/SC versions of cmpxchg_double(). This patch fixes the issue by passing a pointer to __uint128_t into the +Q constraint, ensuring that the compiler hazards against the entire 16 bytes being modified. With this change, GCC 12.1.0 compiles the above test as: | 0000000000000000 <foo>: | 0: f9400407 ldr x7, [x0, thesofproject#8] | 4: d503233f paciasp | 8: aa0003e4 mov x4, x0 | c: 1400000f b 48 <foo+0x48> | 10: d2800240 mov x0, #0x12 // thesofproject#18 | 14: d2800681 mov x1, #0x34 // thesofproject#52 | 18: aa0003e5 mov x5, x0 | 1c: aa0103e6 mov x6, x1 | 20: d2800ac2 mov x2, #0x56 // thesofproject#86 | 24: d2800f03 mov x3, #0x78 // thesofproject#120 | 28: 48207c82 casp x0, x1, x2, x3, [x4] | 2c: ca050000 eor x0, x0, x5 | 30: ca060021 eor x1, x1, x6 | 34: aa010000 orr x0, x0, x1 | 38: f9400480 ldr x0, [x4, thesofproject#8] | 3c: d50323bf autiasp | 40: ca0000e0 eor x0, x7, x0 | 44: d65f03c0 ret | 48: d2800240 mov x0, #0x12 // thesofproject#18 | 4c: d2800681 mov x1, #0x34 // thesofproject#52 | 50: d2800ac2 mov x2, #0x56 // thesofproject#86 | 54: d2800f03 mov x3, #0x78 // thesofproject#120 | 58: f9800091 prfm pstl1strm, [x4] | 5c: c87f1885 ldxp x5, x6, [x4] | 60: ca0000a5 eor x5, x5, x0 | 64: ca0100c6 eor x6, x6, x1 | 68: aa0600a6 orr x6, x5, x6 | 6c: b5000066 cbnz x6, 78 <foo+0x78> | 70: c8250c82 stxp w5, x2, x3, [x4] | 74: 35ffff45 cbnz w5, 5c <foo+0x5c> | 78: f9400480 ldr x0, [x4, thesofproject#8] | 7c: d50323bf autiasp | 80: ca0000e0 eor x0, x7, x0 | 84: d65f03c0 ret ... sampling the high 8 bytes before and after the cmpxchg, and performing an EOR, as we'd expect. For backporting, I've tested this atop linux-4.9.y with GCC 5.5.0. Note that linux-4.9.y is oldest currently supported stable release, and mandates GCC 5.1+. Unfortunately I couldn't get a GCC 5.1 binary to run on my machines due to library incompatibilities. I've also used a standalone test to check that we can use a __uint128_t pointer in a +Q constraint at least as far back as GCC 4.8.5 and LLVM 3.9.1. Fixes: 5284e1b ("arm64: xchg: Implement cmpxchg_double") Fixes: e9a4b79 ("arm64: cmpxchg_dbl: patch in lse instructions when supported by the CPU") Reported-by: Boqun Feng <[email protected]> Link: https://lore.kernel.org/lkml/Y6DEfQXymYVgL3oJ@boqun-archlinux/ Reported-by: Peter Zijlstra <[email protected]> Link: https://lore.kernel.org/lkml/[email protected]/ Signed-off-by: Mark Rutland <[email protected]> Cc: [email protected] Cc: Arnd Bergmann <[email protected]> Cc: Catalin Marinas <[email protected]> Cc: Steve Capper <[email protected]> Cc: Will Deacon <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Will Deacon <[email protected]> Signed-off-by: Sasha Levin <[email protected]>
ujfalusi
pushed a commit
to ujfalusi/sof-linux
that referenced
this issue
Feb 17, 2023
The inline assembly for arm64's cmpxchg_double*() implementations use a +Q constraint to hazard against other accesses to the memory location being exchanged. However, the pointer passed to the constraint is a pointer to unsigned long, and thus the hazard only applies to the first 8 bytes of the location. GCC can take advantage of this, assuming that other portions of the location are unchanged, leading to a number of potential problems. This is similar to what we fixed back in commit: fee960b ("arm64: xchg: hazard against entire exchange variable") ... but we forgot to adjust cmpxchg_double*() similarly at the same time. The same problem applies, as demonstrated with the following test: | struct big { | u64 lo, hi; | } __aligned(128); | | unsigned long foo(struct big *b) | { | u64 hi_old, hi_new; | | hi_old = b->hi; | cmpxchg_double_local(&b->lo, &b->hi, 0x12, 0x34, 0x56, 0x78); | hi_new = b->hi; | | return hi_old ^ hi_new; | } ... which GCC 12.1.0 compiles as: | 0000000000000000 <foo>: | 0: d503233f paciasp | 4: aa0003e4 mov x4, x0 | 8: 1400000e b 40 <foo+0x40> | c: d2800240 mov x0, #0x12 // thesofproject#18 | 10: d2800681 mov x1, #0x34 // thesofproject#52 | 14: aa0003e5 mov x5, x0 | 18: aa0103e6 mov x6, x1 | 1c: d2800ac2 mov x2, #0x56 // thesofproject#86 | 20: d2800f03 mov x3, #0x78 // thesofproject#120 | 24: 48207c82 casp x0, x1, x2, x3, [x4] | 28: ca050000 eor x0, x0, x5 | 2c: ca060021 eor x1, x1, x6 | 30: aa010000 orr x0, x0, x1 | 34: d2800000 mov x0, #0x0 // #0 <--- BANG | 38: d50323bf autiasp | 3c: d65f03c0 ret | 40: d2800240 mov x0, #0x12 // thesofproject#18 | 44: d2800681 mov x1, #0x34 // thesofproject#52 | 48: d2800ac2 mov x2, #0x56 // thesofproject#86 | 4c: d2800f03 mov x3, #0x78 // thesofproject#120 | 50: f9800091 prfm pstl1strm, [x4] | 54: c87f1885 ldxp x5, x6, [x4] | 58: ca0000a5 eor x5, x5, x0 | 5c: ca0100c6 eor x6, x6, x1 | 60: aa0600a6 orr x6, x5, x6 | 64: b5000066 cbnz x6, 70 <foo+0x70> | 68: c8250c82 stxp w5, x2, x3, [x4] | 6c: 35ffff45 cbnz w5, 54 <foo+0x54> | 70: d2800000 mov x0, #0x0 // #0 <--- BANG | 74: d50323bf autiasp | 78: d65f03c0 ret Notice that at the lines with "BANG" comments, GCC has assumed that the higher 8 bytes are unchanged by the cmpxchg_double() call, and that `hi_old ^ hi_new` can be reduced to a constant zero, for both LSE and LL/SC versions of cmpxchg_double(). This patch fixes the issue by passing a pointer to __uint128_t into the +Q constraint, ensuring that the compiler hazards against the entire 16 bytes being modified. With this change, GCC 12.1.0 compiles the above test as: | 0000000000000000 <foo>: | 0: f9400407 ldr x7, [x0, thesofproject#8] | 4: d503233f paciasp | 8: aa0003e4 mov x4, x0 | c: 1400000f b 48 <foo+0x48> | 10: d2800240 mov x0, #0x12 // thesofproject#18 | 14: d2800681 mov x1, #0x34 // thesofproject#52 | 18: aa0003e5 mov x5, x0 | 1c: aa0103e6 mov x6, x1 | 20: d2800ac2 mov x2, #0x56 // thesofproject#86 | 24: d2800f03 mov x3, #0x78 // thesofproject#120 | 28: 48207c82 casp x0, x1, x2, x3, [x4] | 2c: ca050000 eor x0, x0, x5 | 30: ca060021 eor x1, x1, x6 | 34: aa010000 orr x0, x0, x1 | 38: f9400480 ldr x0, [x4, thesofproject#8] | 3c: d50323bf autiasp | 40: ca0000e0 eor x0, x7, x0 | 44: d65f03c0 ret | 48: d2800240 mov x0, #0x12 // thesofproject#18 | 4c: d2800681 mov x1, #0x34 // thesofproject#52 | 50: d2800ac2 mov x2, #0x56 // thesofproject#86 | 54: d2800f03 mov x3, #0x78 // thesofproject#120 | 58: f9800091 prfm pstl1strm, [x4] | 5c: c87f1885 ldxp x5, x6, [x4] | 60: ca0000a5 eor x5, x5, x0 | 64: ca0100c6 eor x6, x6, x1 | 68: aa0600a6 orr x6, x5, x6 | 6c: b5000066 cbnz x6, 78 <foo+0x78> | 70: c8250c82 stxp w5, x2, x3, [x4] | 74: 35ffff45 cbnz w5, 5c <foo+0x5c> | 78: f9400480 ldr x0, [x4, thesofproject#8] | 7c: d50323bf autiasp | 80: ca0000e0 eor x0, x7, x0 | 84: d65f03c0 ret ... sampling the high 8 bytes before and after the cmpxchg, and performing an EOR, as we'd expect. For backporting, I've tested this atop linux-4.9.y with GCC 5.5.0. Note that linux-4.9.y is oldest currently supported stable release, and mandates GCC 5.1+. Unfortunately I couldn't get a GCC 5.1 binary to run on my machines due to library incompatibilities. I've also used a standalone test to check that we can use a __uint128_t pointer in a +Q constraint at least as far back as GCC 4.8.5 and LLVM 3.9.1. Fixes: 5284e1b ("arm64: xchg: Implement cmpxchg_double") Fixes: e9a4b79 ("arm64: cmpxchg_dbl: patch in lse instructions when supported by the CPU") Reported-by: Boqun Feng <[email protected]> Link: https://lore.kernel.org/lkml/Y6DEfQXymYVgL3oJ@boqun-archlinux/ Reported-by: Peter Zijlstra <[email protected]> Link: https://lore.kernel.org/lkml/[email protected]/ Signed-off-by: Mark Rutland <[email protected]> Cc: [email protected] Cc: Arnd Bergmann <[email protected]> Cc: Catalin Marinas <[email protected]> Cc: Steve Capper <[email protected]> Cc: Will Deacon <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Will Deacon <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
The current code requires users to compile out two out of the three drivers for APL/GLK registered for the same PCI ID - or use mod probe blacklists. This needs to eventually move to an autodetect when the probe is handled in sequence, e.g. SOF first, then fallback to SST if firmware authentication fails, and last legacy HDaudio.
Filing it as a bug since it's really required for distributions such as ClearLinux and Ubuntu
The text was updated successfully, but these errors were encountered: