aboutsummaryrefslogtreecommitdiff
AgeCommit message (Collapse)AuthorFilesLines
2021-09-14nvme-rdma: destroy cm id before destroy qp to avoid use after freeRuozhu Li1-13/+3
We should always destroy cm_id before destroy qp to avoid to get cma event after qp was destroyed, which may lead to use after free. In RDMA connection establishment error flow, don't destroy qp in cm event handler.Just report cm_error to upper level, qp will be destroy in nvme_rdma_alloc_queue() after destroy cm id. Signed-off-by: Ruozhu Li <[email protected]> Reviewed-by: Max Gurtovoy <[email protected]> Signed-off-by: Christoph Hellwig <[email protected]>
2021-09-14nvme-multipath: fix ANA state updates when a namespace is not presentAnton Eidelman1-2/+5
nvme_update_ana_state() has a deficiency that results in a failure to properly update the ana state for a namespace in the following case: NSIDs in ctrl->namespaces: 1, 3, 4 NSIDs in desc->nsids: 1, 2, 3, 4 Loop iteration 0: ns index = 0, n = 0, ns->head->ns_id = 1, nsid = 1, MATCH. Loop iteration 1: ns index = 1, n = 1, ns->head->ns_id = 3, nsid = 2, NO MATCH. Loop iteration 2: ns index = 2, n = 2, ns->head->ns_id = 4, nsid = 4, MATCH. Where the update to the ANA state of NSID 3 is missed. To fix this increment n and retry the update with the same ns when ns->head->ns_id is higher than nsid, Signed-off-by: Anton Eidelman <[email protected]> Signed-off-by: Christoph Hellwig <[email protected]> Reviewed-by: Sagi Grimberg <[email protected]>
2021-09-14x86/mce: Avoid infinite loop for copy from user recoveryTony Luck2-11/+33
There are two cases for machine check recovery: 1) The machine check was triggered by ring3 (application) code. This is the simpler case. The machine check handler simply queues work to be executed on return to user. That code unmaps the page from all users and arranges to send a SIGBUS to the task that triggered the poison. 2) The machine check was triggered in kernel code that is covered by an exception table entry. In this case the machine check handler still queues a work entry to unmap the page, etc. but this will not be called right away because the #MC handler returns to the fix up code address in the exception table entry. Problems occur if the kernel triggers another machine check before the return to user processes the first queued work item. Specifically, the work is queued using the ->mce_kill_me callback structure in the task struct for the current thread. Attempting to queue a second work item using this same callback results in a loop in the linked list of work functions to call. So when the kernel does return to user, it enters an infinite loop processing the same entry for ever. There are some legitimate scenarios where the kernel may take a second machine check before returning to the user. 1) Some code (e.g. futex) first tries a get_user() with page faults disabled. If this fails, the code retries with page faults enabled expecting that this will resolve the page fault. 2) Copy from user code retries a copy in byte-at-time mode to check whether any additional bytes can be copied. On the other side of the fence are some bad drivers that do not check the return value from individual get_user() calls and may access multiple user addresses without noticing that some/all calls have failed. Fix by adding a counter (current->mce_count) to keep track of repeated machine checks before task_work() is called. First machine check saves the address information and calls task_work_add(). Subsequent machine checks before that task_work call back is executed check that the address is in the same page as the first machine check (since the callback will offline exactly one page). Expected worst case is four machine checks before moving on (e.g. one user access with page faults disabled, then a repeat to the same address with page faults enabled ... repeat in copy tail bytes). Just in case there is some code that loops forever enforce a limit of 10. [ bp: Massage commit message, drop noinstr, fix typo, extend panic messages. ] Fixes: 5567d11c21a1 ("x86/mce: Send #MC singal from task work") Signed-off-by: Tony Luck <[email protected]> Signed-off-by: Borislav Petkov <[email protected]> Cc: <[email protected]> Link: https://lkml.kernel.org/r/YT/[email protected]
2021-09-14rtc: cmos: Disable irq around direct invocation of cmos_interrupt()Chris Wilson1-0/+2
As previously noted in commit 66e4f4a9cc38 ("rtc: cmos: Use spin_lock_irqsave() in cmos_interrupt()"): <4>[ 254.192378] WARNING: inconsistent lock state <4>[ 254.192384] 5.12.0-rc1-CI-CI_DRM_9834+ #1 Not tainted <4>[ 254.192396] -------------------------------- <4>[ 254.192400] inconsistent {IN-HARDIRQ-W} -> {HARDIRQ-ON-W} usage. <4>[ 254.192409] rtcwake/5309 [HC0[0]:SC0[0]:HE1:SE1] takes: <4>[ 254.192429] ffffffff8263c5f8 (rtc_lock){?...}-{2:2}, at: cmos_interrupt+0x18/0x100 <4>[ 254.192481] {IN-HARDIRQ-W} state was registered at: <4>[ 254.192488] lock_acquire+0xd1/0x3d0 <4>[ 254.192504] _raw_spin_lock+0x2a/0x40 <4>[ 254.192519] cmos_interrupt+0x18/0x100 <4>[ 254.192536] rtc_handler+0x1f/0xc0 <4>[ 254.192553] acpi_ev_fixed_event_detect+0x109/0x13c <4>[ 254.192574] acpi_ev_sci_xrupt_handler+0xb/0x28 <4>[ 254.192596] acpi_irq+0x13/0x30 <4>[ 254.192620] __handle_irq_event_percpu+0x43/0x2c0 <4>[ 254.192641] handle_irq_event_percpu+0x2b/0x70 <4>[ 254.192661] handle_irq_event+0x2f/0x50 <4>[ 254.192680] handle_fasteoi_irq+0x9e/0x150 <4>[ 254.192693] __common_interrupt+0x76/0x140 <4>[ 254.192715] common_interrupt+0x96/0xc0 <4>[ 254.192732] asm_common_interrupt+0x1e/0x40 <4>[ 254.192750] _raw_spin_unlock_irqrestore+0x38/0x60 <4>[ 254.192767] resume_irqs+0xba/0xf0 <4>[ 254.192786] dpm_resume_noirq+0x245/0x3d0 <4>[ 254.192811] suspend_devices_and_enter+0x230/0xaa0 <4>[ 254.192835] pm_suspend.cold.8+0x301/0x34a <4>[ 254.192859] state_store+0x7b/0xe0 <4>[ 254.192879] kernfs_fop_write_iter+0x11d/0x1c0 <4>[ 254.192899] new_sync_write+0x11d/0x1b0 <4>[ 254.192916] vfs_write+0x265/0x390 <4>[ 254.192933] ksys_write+0x5a/0xd0 <4>[ 254.192949] do_syscall_64+0x33/0x80 <4>[ 254.192965] entry_SYSCALL_64_after_hwframe+0x44/0xae <4>[ 254.192986] irq event stamp: 43775 <4>[ 254.192994] hardirqs last enabled at (43775): [<ffffffff81c00c42>] asm_sysvec_apic_timer_interrupt+0x12/0x20 <4>[ 254.193023] hardirqs last disabled at (43774): [<ffffffff81aa691a>] sysvec_apic_timer_interrupt+0xa/0xb0 <4>[ 254.193049] softirqs last enabled at (42548): [<ffffffff81e00342>] __do_softirq+0x342/0x48e <4>[ 254.193074] softirqs last disabled at (42543): [<ffffffff810b45fd>] irq_exit_rcu+0xad/0xd0 <4>[ 254.193101] other info that might help us debug this: <4>[ 254.193107] Possible unsafe locking scenario: <4>[ 254.193112] CPU0 <4>[ 254.193117] ---- <4>[ 254.193121] lock(rtc_lock); <4>[ 254.193137] <Interrupt> <4>[ 254.193142] lock(rtc_lock); <4>[ 254.193156] *** DEADLOCK *** <4>[ 254.193161] 6 locks held by rtcwake/5309: <4>[ 254.193174] #0: ffff888104861430 (sb_writers#5){.+.+}-{0:0}, at: ksys_write+0x5a/0xd0 <4>[ 254.193232] #1: ffff88810f823288 (&of->mutex){+.+.}-{3:3}, at: kernfs_fop_write_iter+0xe7/0x1c0 <4>[ 254.193282] #2: ffff888100cef3c0 (kn->active#285 <7>[ 254.192706] i915 0000:00:02.0: [drm:intel_modeset_setup_hw_state [i915]] [CRTC:51:pipe A] hw state readout: disabled <4>[ 254.193307] ){.+.+}-{0:0}, at: kernfs_fop_write_iter+0xf0/0x1c0 <4>[ 254.193333] #3: ffffffff82649fa8 (system_transition_mutex){+.+.}-{3:3}, at: pm_suspend.cold.8+0xce/0x34a <4>[ 254.193387] #4: ffffffff827a2108 (acpi_scan_lock){+.+.}-{3:3}, at: acpi_suspend_begin+0x47/0x70 <4>[ 254.193433] #5: ffff8881019ea178 (&dev->mutex){....}-{3:3}, at: device_resume+0x68/0x1e0 <4>[ 254.193485] stack backtrace: <4>[ 254.193492] CPU: 1 PID: 5309 Comm: rtcwake Not tainted 5.12.0-rc1-CI-CI_DRM_9834+ #1 <4>[ 254.193514] Hardware name: Google Soraka/Soraka, BIOS MrChromebox-4.10 08/25/2019 <4>[ 254.193524] Call Trace: <4>[ 254.193536] dump_stack+0x7f/0xad <4>[ 254.193567] mark_lock.part.47+0x8ca/0xce0 <4>[ 254.193604] __lock_acquire+0x39b/0x2590 <4>[ 254.193626] ? asm_sysvec_apic_timer_interrupt+0x12/0x20 <4>[ 254.193660] lock_acquire+0xd1/0x3d0 <4>[ 254.193677] ? cmos_interrupt+0x18/0x100 <4>[ 254.193716] _raw_spin_lock+0x2a/0x40 <4>[ 254.193735] ? cmos_interrupt+0x18/0x100 <4>[ 254.193758] cmos_interrupt+0x18/0x100 <4>[ 254.193785] cmos_resume+0x2ac/0x2d0 <4>[ 254.193813] ? acpi_pm_set_device_wakeup+0x1f/0x110 <4>[ 254.193842] ? pnp_bus_suspend+0x10/0x10 <4>[ 254.193864] pnp_bus_resume+0x5e/0x90 <4>[ 254.193885] dpm_run_callback+0x5f/0x240 <4>[ 254.193914] device_resume+0xb2/0x1e0 <4>[ 254.193942] ? pm_dev_err+0x25/0x25 <4>[ 254.193974] dpm_resume+0xea/0x3f0 <4>[ 254.194005] dpm_resume_end+0x8/0x10 <4>[ 254.194030] suspend_devices_and_enter+0x29b/0xaa0 <4>[ 254.194066] pm_suspend.cold.8+0x301/0x34a <4>[ 254.194094] state_store+0x7b/0xe0 <4>[ 254.194124] kernfs_fop_write_iter+0x11d/0x1c0 <4>[ 254.194151] new_sync_write+0x11d/0x1b0 <4>[ 254.194183] vfs_write+0x265/0x390 <4>[ 254.194207] ksys_write+0x5a/0xd0 <4>[ 254.194232] do_syscall_64+0x33/0x80 <4>[ 254.194251] entry_SYSCALL_64_after_hwframe+0x44/0xae <4>[ 254.194274] RIP: 0033:0x7f07d79691e7 <4>[ 254.194293] Code: 64 89 02 48 c7 c0 ff ff ff ff eb bb 0f 1f 80 00 00 00 00 f3 0f 1e fa 64 8b 04 25 18 00 00 00 85 c0 75 10 b8 01 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 51 c3 48 83 ec 28 48 89 54 24 18 48 89 74 24 <4>[ 254.194312] RSP: 002b:00007ffd9cc2c768 EFLAGS: 00000246 ORIG_RAX: 0000000000000001 <4>[ 254.194337] RAX: ffffffffffffffda RBX: 0000000000000004 RCX: 00007f07d79691e7 <4>[ 254.194352] RDX: 0000000000000004 RSI: 0000556ebfc63590 RDI: 000000000000000b <4>[ 254.194366] RBP: 0000556ebfc63590 R08: 0000000000000000 R09: 0000000000000004 <4>[ 254.194379] R10: 0000556ebf0ec2a6 R11: 0000000000000246 R12: 0000000000000004 which breaks S3-resume on fi-kbl-soraka presumably as that's slow enough to trigger the alarm during the suspend. Fixes: 6950d046eb6e ("rtc: cmos: Replace spin_lock_irqsave with spin_lock in hard IRQ") References: 66e4f4a9cc38 ("rtc: cmos: Use spin_lock_irqsave() in cmos_interrupt()"): Signed-off-by: Chris Wilson <[email protected]> Cc: Xiaofei Tan <[email protected]> Cc: Alexandre Belloni <[email protected]> Cc: Alessandro Zummo <[email protected]> Cc: Ville Syrjälä <[email protected]> Reviewed-by: Ville Syrjälä <[email protected]> Signed-off-by: Alexandre Belloni <[email protected]> Link: https://lore.kernel.org/r/[email protected]
2021-09-14drm/i915: Get PM ref before accessing HW registerVinay Belgaumkar1-1/+7
Seeing these errors when GT is likely in suspend state- "RPM wakelock ref not held during HW access" Ensure GT is awake before trying to access HW registers. Avoid reading the register if that is not the case. Signed-off-by: Vinay Belgaumkar <[email protected]> Fixes: 41e5c17ebfc2 ("drm/i915/guc/slpc: Sysfs hooks for SLPC") Reviewed-by: Tvrtko Ursulin <[email protected]> Signed-off-by: John Harrison <[email protected]> Link: https://patchwork.freedesktop.org/patch/msgid/[email protected] (cherry picked from commit f25e3908b9cd4a3fe819e9bdcdde58f20bacb34c) Signed-off-by: Jani Nikula <[email protected]>
2021-09-14drm/i915: Release ctx->syncobj on final put, not on ctx closeDaniel Vetter1-3/+3
gem context refcounting is another exercise in least locking design it seems, where most things get destroyed upon context closure (which can race with anything really). Only the actual memory allocation and the locks survive while holding a reference. This tripped up Jason when reimplementing the single timeline feature in commit 00dae4d3d35d4f526929633b76e00b0ab4d3970d Author: Jason Ekstrand <[email protected]> Date: Thu Jul 8 10:48:12 2021 -0500 drm/i915: Implement SINGLE_TIMELINE with a syncobj (v4) We could fix the bug by holding ctx->mutex in execbuf and clear the pointer (again while holding the mutex) context_close, but it's cleaner to just make the context object actually invariant over its _entire_ lifetime. This way any other ioctl that's potentially racing, but holding a full reference, can still rely on ctx->syncobj being an immutable pointer. Which without this change, is not the case. Reviewed-by: Maarten Lankhorst <[email protected]> Signed-off-by: Daniel Vetter <[email protected]> Fixes: 00dae4d3d35d ("drm/i915: Implement SINGLE_TIMELINE with a syncobj (v4)") Cc: Jason Ekstrand <[email protected]> Cc: Chris Wilson <[email protected]> Cc: Tvrtko Ursulin <[email protected]> Cc: Joonas Lahtinen <[email protected]> Cc: Matthew Brost <[email protected]> Cc: Matthew Auld <[email protected]> Cc: Maarten Lankhorst <[email protected]> Cc: "Thomas Hellström" <[email protected]> Cc: Lionel Landwerlin <[email protected]> Cc: Dave Airlie <[email protected]> Link: https://patchwork.freedesktop.org/patch/msgid/[email protected] (cherry picked from commit c238980efd3b35af70fc926066cf7440f50a97a9) Signed-off-by: Jani Nikula <[email protected]>
2021-09-14drm/i915/gem: Fix the mman selftestThomas Hellström1-6/+20
Using the I915_MMAP_TYPE_FIXED mmap type requires the TTM backend, so for that mmap type, use __i915_gem_object_create_user() instead of i915_gem_object_create_internal(), as we really want to tests objects mmap-able by user-space. This also means that the out-of-space error happens at object creation and returns -ENXIO rather than -ENOSPC, so fix the code up to expect that on out-of-offset-space errors. Finally only use I915_MMAP_TYPE_FIXED for LMEM and SMEM for now if testing on LMEM-capable devices. For stolen LMEM, we still take the same path as for integrated, as that haven't been moved over to TTM yet, and user-space should not be able to create out of stolen LMEM anyway. v2: - Check the presence of the obj->ops->mmap_offset callback rather than hardcoding the supported mmap regions in can_mmap() (Maarten Lankhorst) Fixes: 7961c5b60f23 ("drm/i915: Add TTM offset argument to mmap.") Cc: Maarten Lankhorst <[email protected]> Signed-off-by: Thomas Hellström <[email protected]> Reviewed-by: Maarten Lankhorst <[email protected]> Link: https://patchwork.freedesktop.org/patch/msgid/[email protected] (cherry picked from commit 450cede7f3804ca7f8b3da210ebefa61c0958f22) Signed-off-by: Jani Nikula <[email protected]>
2021-09-14drm/i915/guc: drop guc_communication_enabledDaniele Ceraolo Spurio1-8/+3
The function is only used from within GEM_BUG_ON(), which is causing warnings with Wunneeded-internal-declaration in some builds. Since the function is a simple wrapper around a CT function, we can just call the CT function directly instead. Fixes: 1fb12c587152 ("drm/i915/guc: skip disabling CTBs before sanitizing the GuC") Reported-by: kernel test robot <[email protected]> Signed-off-by: Daniele Ceraolo Spurio <[email protected]> Cc: Matthew Brost <[email protected]> Cc: John Harrison <[email protected]> Reviewed-by: Matthew Brost <[email protected]> Signed-off-by: John Harrison <[email protected]> Link: https://patchwork.freedesktop.org/patch/msgid/[email protected] (cherry picked from commit 5db1856781e45c9610f7652a19cc656b984235e7) Signed-off-by: Jani Nikula <[email protected]>
2021-09-14drm/i915/dp: Use max params for panels < eDP 1.4Kai-Heng Feng1-1/+4
Users reported that after commit 2bbd6dba84d4 ("drm/i915: Try to use fast+narrow link on eDP again and fall back to the old max strategy on failure"), the screen starts to have wobbly effect. Commit a5c936add6a2 ("drm/i915/dp: Use slow and wide link training for everything") doesn't help either, that means the affected eDP 1.2 panels only work with max params. So use max params for panels < eDP 1.4 as Windows does to solve the issue. v3: - Do the eDP rev check in intel_edp_init_dpcd() v2: - Check eDP 1.4 instead of DPCD 1.1 to apply max params Cc: [email protected] Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/3714 Fixes: 2bbd6dba84d4 ("drm/i915: Try to use fast+narrow link on eDP again and fall back to the old max strategy on failure") Fixes: a5c936add6a2 ("drm/i915/dp: Use slow and wide link training for everything") Suggested-by: Ville Syrjälä <[email protected]> Signed-off-by: Kai-Heng Feng <[email protected]> Signed-off-by: Ville Syrjälä <[email protected]> Link: https://patchwork.freedesktop.org/patch/msgid/[email protected] (cherry picked from commit d7f213c131adf0bec8b731553eb82990cdac265d) Signed-off-by: Jani Nikula <[email protected]>
2021-09-14drm/i915/dp: return proper DPRX link training resultLee Shawn C1-1/+1
After DPRX link training, intel_dp_link_train_phy() did not return the training result properly. If link training failed, i915 driver would not run into link train fallback function. And no hotplug uevent would be received by user space application. Fixes: b30edfd8d0b4 ("drm/i915: Switch to LTTPR non-transparent mode link training") Cc: Ville Syrjala <[email protected]> Cc: Imre Deak <[email protected]> Cc: Jani Nikula <[email protected]> Cc: Cooper Chiou <[email protected]> Cc: William Tseng <[email protected]> Signed-off-by: Lee Shawn C <[email protected]> Reviewed-by: Imre Deak <[email protected]> Signed-off-by: Imre Deak <[email protected]> Link: https://patchwork.freedesktop.org/patch/msgid/[email protected] (cherry picked from commit dab1b47e57e053b2a02c22ead8e7449f79961335) Signed-off-by: Jani Nikula <[email protected]>
2021-09-14PM: base: power: don't try to use non-existing RTC for storing dataJuergen Gross1-0/+10
If there is no legacy RTC device, don't try to use it for storing trace data across suspend/resume. Cc: <[email protected]> Signed-off-by: Juergen Gross <[email protected]> Reviewed-by: Rafael J. Wysocki <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Juergen Gross <[email protected]>
2021-09-14xen/balloon: use a kernel thread instead a workqueueJuergen Gross1-17/+45
Today the Xen ballooning is done via delayed work in a workqueue. This might result in workqueue hangups being reported in case of large amounts of memory are being ballooned in one go (here 16GB): BUG: workqueue lockup - pool cpus=6 node=0 flags=0x0 nice=0 stuck for 64s! Showing busy workqueues and worker pools: workqueue events: flags=0x0 pwq 12: cpus=6 node=0 flags=0x0 nice=0 active=2/256 refcnt=3 in-flight: 229:balloon_process pending: cache_reap workqueue events_freezable_power_: flags=0x84 pwq 12: cpus=6 node=0 flags=0x0 nice=0 active=1/256 refcnt=2 pending: disk_events_workfn workqueue mm_percpu_wq: flags=0x8 pwq 12: cpus=6 node=0 flags=0x0 nice=0 active=1/256 refcnt=2 pending: vmstat_update pool 12: cpus=6 node=0 flags=0x0 nice=0 hung=64s workers=3 idle: 2222 43 This can easily be avoided by using a dedicated kernel thread for doing the ballooning work. Reported-by: Jan Beulich <[email protected]> Signed-off-by: Juergen Gross <[email protected]> Reviewed-by: Boris Ostrovsky <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Juergen Gross <[email protected]>
2021-09-13io_uring: pin SQPOLL data before unlocking ring lockJens Axboe1-3/+9
We need to re-check sqd->thread after we've dropped the lock. Pin the sqd before doing the lockdep lock dance, and check if the thread is alive after that. It's either NULL or alive, as the SQPOLL thread cannot exit without holding the same sqd->lock. Reported-and-tested-by: [email protected] Fixes: fa84693b3c89 ("io_uring: ensure IORING_REGISTER_IOWQ_MAX_WORKERS works with SQPOLL") Signed-off-by: Jens Axboe <[email protected]>
2021-09-13bpf, selftests: Add test case for mixed cgroup v1/v2Daniel Borkmann4-6/+127
Minimal selftest which implements a small BPF policy program to the connect(2) hook which rejects TCP connection requests to port 60123 with EPERM. This is being attached to a non-root cgroup v2 path. The test asserts that this works under cgroup v2-only and under a mixed cgroup v1/v2 environment where net_classid is set in the former case. Before fix: # ./test_progs -t cgroup_v1v2 test_cgroup_v1v2:PASS:server_fd 0 nsec test_cgroup_v1v2:PASS:client_fd 0 nsec test_cgroup_v1v2:PASS:cgroup_fd 0 nsec test_cgroup_v1v2:PASS:server_fd 0 nsec run_test:PASS:skel_open 0 nsec run_test:PASS:prog_attach 0 nsec test_cgroup_v1v2:PASS:cgroup-v2-only 0 nsec run_test:PASS:skel_open 0 nsec run_test:PASS:prog_attach 0 nsec run_test:PASS:join_classid 0 nsec (network_helpers.c:219: errno: None) Unexpected success to connect to server test_cgroup_v1v2:FAIL:cgroup-v1v2 unexpected error: -1 (errno 0) #27 cgroup_v1v2:FAIL Summary: 0/0 PASSED, 0 SKIPPED, 1 FAILED After fix: # ./test_progs -t cgroup_v1v2 #27 cgroup_v1v2:OK Summary: 1/0 PASSED, 0 SKIPPED, 0 FAILED Signed-off-by: Daniel Borkmann <[email protected]> Signed-off-by: Alexei Starovoitov <[email protected]> Acked-by: Andrii Nakryiko <[email protected]> Link: https://lore.kernel.org/bpf/[email protected]
2021-09-13bpf, selftests: Add cgroup v1 net_cls classid helpersDaniel Borkmann2-12/+141
Minimal set of helpers for net_cls classid cgroupv1 management in order to set an id, join from a process, initiate setup and teardown. cgroupv2 helpers are left as-is, but reused where possible. Signed-off-by: Daniel Borkmann <[email protected]> Signed-off-by: Alexei Starovoitov <[email protected]> Acked-by: Andrii Nakryiko <[email protected]> Link: https://lore.kernel.org/bpf/[email protected]
2021-09-13bpf, cgroups: Fix cgroup v2 fallback on v1/v2 mixed modeDaniel Borkmann5-155/+41
Fix cgroup v1 interference when non-root cgroup v2 BPF programs are used. Back in the days, commit bd1060a1d671 ("sock, cgroup: add sock->sk_cgroup") embedded per-socket cgroup information into sock->sk_cgrp_data and in order to save 8 bytes in struct sock made both mutually exclusive, that is, when cgroup v1 socket tagging (e.g. net_cls/net_prio) is used, then cgroup v2 falls back to the root cgroup in sock_cgroup_ptr() (&cgrp_dfl_root.cgrp). The assumption made was "there is no reason to mix the two and this is in line with how legacy and v2 compatibility is handled" as stated in bd1060a1d671. However, with Kubernetes more widely supporting cgroups v2 as well nowadays, this assumption no longer holds, and the possibility of the v1/v2 mixed mode with the v2 root fallback being hit becomes a real security issue. Many of the cgroup v2 BPF programs are also used for policy enforcement, just to pick _one_ example, that is, to programmatically deny socket related system calls like connect(2) or bind(2). A v2 root fallback would implicitly cause a policy bypass for the affected Pods. In production environments, we have recently seen this case due to various circumstances: i) a different 3rd party agent and/or ii) a container runtime such as [0] in the user's environment configuring legacy cgroup v1 net_cls tags, which triggered implicitly mentioned root fallback. Another case is Kubernetes projects like kind [1] which create Kubernetes nodes in a container and also add cgroup namespaces to the mix, meaning programs which are attached to the cgroup v2 root of the cgroup namespace get attached to a non-root cgroup v2 path from init namespace point of view. And the latter's root is out of reach for agents on a kind Kubernetes node to configure. Meaning, any entity on the node setting cgroup v1 net_cls tag will trigger the bypass despite cgroup v2 BPF programs attached to the namespace root. Generally, this mutual exclusiveness does not hold anymore in today's user environments and makes cgroup v2 usage from BPF side fragile and unreliable. This fix adds proper struct cgroup pointer for the cgroup v2 case to struct sock_cgroup_data in order to address these issues; this implicitly also fixes the tradeoffs being made back then with regards to races and refcount leaks as stated in bd1060a1d671, and removes the fallback, so that cgroup v2 BPF programs always operate as expected. [0] https://github.com/nestybox/sysbox/ [1] https://kind.sigs.k8s.io/ Fixes: bd1060a1d671 ("sock, cgroup: add sock->sk_cgroup") Signed-off-by: Daniel Borkmann <[email protected]> Signed-off-by: Alexei Starovoitov <[email protected]> Acked-by: Stanislav Fomichev <[email protected]> Acked-by: Tejun Heo <[email protected]> Link: https://lore.kernel.org/bpf/[email protected]
2021-09-13cifs: fix incorrect kernel doc commentsSteve French2-3/+11
Correct kernel-doc comments pointed out by the automated kernel test robot. Reported-by: kernel test robot <[email protected]> Signed-off-by: Steve French <[email protected]>
2021-09-13bpf: Add oversize check before call kvcalloc()Bixuan Cui1-0/+2
Commit 7661809d493b ("mm: don't allow oversized kvmalloc() calls") add the oversize check. When the allocation is larger than what kmalloc() supports, the following warning triggered: WARNING: CPU: 0 PID: 8408 at mm/util.c:597 kvmalloc_node+0x108/0x110 mm/util.c:597 Modules linked in: CPU: 0 PID: 8408 Comm: syz-executor221 Not tainted 5.14.0-syzkaller #0 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011 RIP: 0010:kvmalloc_node+0x108/0x110 mm/util.c:597 Call Trace: kvmalloc include/linux/mm.h:806 [inline] kvmalloc_array include/linux/mm.h:824 [inline] kvcalloc include/linux/mm.h:829 [inline] check_btf_line kernel/bpf/verifier.c:9925 [inline] check_btf_info kernel/bpf/verifier.c:10049 [inline] bpf_check+0xd634/0x150d0 kernel/bpf/verifier.c:13759 bpf_prog_load kernel/bpf/syscall.c:2301 [inline] __sys_bpf+0x11181/0x126e0 kernel/bpf/syscall.c:4587 __do_sys_bpf kernel/bpf/syscall.c:4691 [inline] __se_sys_bpf kernel/bpf/syscall.c:4689 [inline] __x64_sys_bpf+0x78/0x90 kernel/bpf/syscall.c:4689 do_syscall_x64 arch/x86/entry/common.c:50 [inline] do_syscall_64+0x3d/0xb0 arch/x86/entry/common.c:80 entry_SYSCALL_64_after_hwframe+0x44/0xae Reported-by: [email protected] Signed-off-by: Bixuan Cui <[email protected]> Signed-off-by: Alexei Starovoitov <[email protected]> Acked-by: Yonghong Song <[email protected]> Link: https://lore.kernel.org/bpf/[email protected]
2021-09-13tools: compiler-gcc.h: Guard error attribute use with __has_attributeNathan Chancellor1-1/+3
When building objtool with HOSTCC=clang, there are several errors along the lines of orc_dump.c:201:28: error: unknown attribute 'error' ignored [-Werror,-Wunknown-attributes] This occurs after commit 4e59869aa655 ("compiler-gcc.h: drop checks for older GCC versions"), which removed the GCC_VERSION gating. The removed version check just so happened to prevent __compiletime_error() from being defined with clang because it pretends to be GCC 4.2.1 for compatibility but the error attribute was not added to clang until 14.0.0. Commit 815f0ddb346c ("include/linux/compiler*.h: make compiler-*.h mutually exclusive") and commit a3f8a30f3f00 ("Compiler Attributes: use feature checks instead of version checks") refactored the handling of attributes in the main kernel to avoid situations like this but that refactoring has never been done for the tools directory. Refactoring is a rather large undertaking and this has never been an issue before so instead, just guard the definition of __compiletime_error() with __has_attribute() so that there are no more errors. Fixes: 4e59869aa655 ("compiler-gcc.h: drop checks for older GCC versions") Signed-off-by: Nathan Chancellor <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2021-09-13cifs: remove pathname for file from SPDX headerSteve French48-49/+8
checkpatch complains about source files with filenames (e.g. in these cases just below the SPDX header in comments at the top of various files in fs/cifs). It also is helpful to change this now so will be less confusing when the parent directory is renamed e.g. from fs/cifs to fs/smb_client (or fs/smbfs) Reviewed-by: Ronnie Sahlberg <[email protected]> Signed-off-by: Steve French <[email protected]>
2021-09-13Merge branch 'gcc-min-version-5.1' (make gcc-5.1 the minimum version)Linus Torvalds18-359/+16
Merge patch series from Nick Desaulniers to update the minimum gcc version to 5.1. This is some of the left-overs from the merge window that I didn't want to deal with yesterday, so it comes in after -rc1 but was sent before. Gcc-4.9 support has been an annoyance for some time, and with -Werror I had the choice of applying a fairly big patch from Kees Cook to remove a fair number of initializer warnings (still leaving some), or this patch series from Nick that just removes the source of the problem. The initializer cleanups might still be worth it regardless, but honestly, I preferred just tackling the problem with gcc-4.9 head-on. We've been more aggressiuve about no longer having to care about compilers that were released a long time ago, and I think it's been a good thing. I added a couple of patches on top to sort out a few left-overs now that we no longer support gcc-4.x. As noted by Arnd, as a result of this minimum compiler version upgrade we can probably change our use of '--std=gnu89' to '--std=gnu11', and finally start using local loop declarations etc. But this series does _not_ yet do that. Link: https://lore.kernel.org/all/[email protected]/ Link: https://lore.kernel.org/lkml/CAK7LNASs6dvU6D3jL2GG3jW58fXfaj6VNOe55NJnTB8UPuk2pA@mail.gmail.com/ Link: https://github.com/ClangBuiltLinux/linux/issues/1438 * emailed patches from Nick Desaulniers <[email protected]>: Drop some straggling mentions of gcc-4.9 as being stale compiler_attributes.h: drop __has_attribute() support for gcc4 vmlinux.lds.h: remove old check for GCC 4.9 compiler-gcc.h: drop checks for older GCC versions Makefile: drop GCC < 5 -fno-var-tracking-assignments workaround arm64: remove GCC version check for ARCH_SUPPORTS_INT128 powerpc: remove GCC version check for UPD_CONSTR riscv: remove Kconfig check for GCC version for ARCH_RV64I Kconfig.debug: drop GCC 5+ version check for DWARF5 mm/ksm: remove old GCC 4.9+ check compiler.h: drop fallback overflow checkers Documentation: raise minimum supported version of GCC to 5.1
2021-09-13Drop some straggling mentions of gcc-4.9 as being staleLinus Torvalds4-4/+3
Fix up the admin-guide README file to the new gcc-5.1 requirement, and remove a stale comment about gcc support for the __assume_aligned__ attribute. Signed-off-by: Linus Torvalds <[email protected]>
2021-09-13cpufreq: intel_pstate: Override parameters if HWP forced by BIOSDoug Smythies1-8/+14
If HWP has been already been enabled by BIOS, it may be necessary to override some kernel command line parameters. Once it has been enabled it requires a reset to be disabled. Suggested-by: Rafael J. Wysocki <[email protected]> Signed-off-by: Doug Smythies <[email protected]> Signed-off-by: Rafael J. Wysocki <[email protected]>
2021-09-13compiler_attributes.h: drop __has_attribute() support for gcc4Linus Torvalds1-20/+0
Now that GCC 5.1 is the minimally supported default, the manual workaround for older gcc versions not having __has_attribute() are no longer relevant and can be removed. Signed-off-by: Linus Torvalds <[email protected]>
2021-09-13vmlinux.lds.h: remove old check for GCC 4.9Nick Desaulniers1-4/+0
Now that GCC 5.1 is the minimally supported version of GCC, we can effectively revert commit 85c2ce9104eb ("sched, vmlinux.lds: Increase STRUCT_ALIGNMENT to 64 bytes for GCC-4.9") Cc: Peter Zijlstra <[email protected]> Signed-off-by: Nick Desaulniers <[email protected]> Acked-by: Kees Cook <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2021-09-13compiler-gcc.h: drop checks for older GCC versionsNick Desaulniers2-6/+2
Now that GCC 5.1 is the minimally supported default, drop the values we don't use. Signed-off-by: Nick Desaulniers <[email protected]> Reviewed-by: Kees Cook <[email protected]> Reviewed-by: Nathan Chancellor <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2021-09-13Makefile: drop GCC < 5 -fno-var-tracking-assignments workaroundNick Desaulniers1-6/+0
Now that GCC 5.1 is the minimally supported version, we can drop this workaround for older versions of GCC. Signed-off-by: Nick Desaulniers <[email protected]> Reviewed-by: Kees Cook <[email protected]> Reviewed-by: Nathan Chancellor <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2021-09-13arm64: remove GCC version check for ARCH_SUPPORTS_INT128Nick Desaulniers1-1/+1
Now that GCC 5.1 is the minimally supported compiler version, this Kconfig check is no longer necessary. Cc: Catalin Marinas <[email protected]> Cc: Will Deacon <[email protected]> Cc: [email protected] Signed-off-by: Nick Desaulniers <[email protected]> Reviewed-by: Kees Cook <[email protected]> Reviewed-by: Nathan Chancellor <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2021-09-13powerpc: remove GCC version check for UPD_CONSTRNick Desaulniers1-10/+0
Now that GCC 5.1 is the minimum supported version, we can drop this workaround for older versions of GCC. This adversely affected clang, too. Cc: Michael Ellerman <[email protected]> Cc: Benjamin Herrenschmidt <[email protected]> Cc: Paul Mackerras <[email protected]> Cc: Segher Boessenkool <[email protected]> Cc: Christophe Leroy <[email protected]> Cc: [email protected] Signed-off-by: Nick Desaulniers <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2021-09-13riscv: remove Kconfig check for GCC version for ARCH_RV64INick Desaulniers1-1/+1
The minimum supported version of GCC is now 5.1. The check wasn't correct as written anyways since GCC_VERSION is 0 when CC=clang. Cc: Paul Walmsley <[email protected]> Cc: Palmer Dabbelt <[email protected]> Cc: Albert Ou <[email protected]> Cc: [email protected] Signed-off-by: Nick Desaulniers <[email protected]> Reviewed-by: Kees Cook <[email protected]> Reviewed-by: Nathan Chancellor <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2021-09-13Kconfig.debug: drop GCC 5+ version check for DWARF5Nick Desaulniers1-1/+1
Now that the minimum supported version of GCC is 5.1, we no longer need this Kconfig version check for CONFIG_DEBUG_INFO_DWARF5. Signed-off-by: Nick Desaulniers <[email protected]> Reviewed-by: Kees Cook <[email protected]> Reviewed-by: Nathan Chancellor <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2021-09-13mm/ksm: remove old GCC 4.9+ checkNick Desaulniers1-2/+0
The minimum supported version of GCC has been raised to GCC 5.1. Signed-off-by: Nick Desaulniers <[email protected]> Reviewed-by: Kees Cook <[email protected]> Reviewed-by: Nathan Chancellor <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2021-09-13compiler.h: drop fallback overflow checkersNick Desaulniers5-293/+6
Once upgrading the minimum supported version of GCC to 5.1, we can drop the fallback code for !COMPILER_HAS_GENERIC_BUILTIN_OVERFLOW. This is effectively a revert of commit f0907827a8a9 ("compiler.h: enable builtin overflow checkers and add fallback code") Link: https://github.com/ClangBuiltLinux/linux/issues/1438#issuecomment-916745801 Suggested-by: Rasmus Villemoes <[email protected]> Signed-off-by: Nick Desaulniers <[email protected]> Acked-by: Kees Cook <[email protected]> Reviewed-by: Nathan Chancellor <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2021-09-13Documentation: raise minimum supported version of GCC to 5.1Nick Desaulniers2-8/+2
commit fad7cd3310db ("nbd: add the check to prevent overflow in __nbd_ioctl()") raised an issue from the fallback helpers added in commit f0907827a8a9 ("compiler.h: enable builtin overflow checkers and add fallback code") Specifically, the helpers for checking whether the results of a multiplication overflowed (__unsigned_mul_overflow, __signed_add_overflow) use the division operator when !COMPILER_HAS_GENERIC_BUILTIN_OVERFLOW. This is problematic for 64b operands on 32b hosts. Also, because the macro is type agnostic, it is very difficult to write a similarly type generic macro that dispatches to one of: * div64_s64 * div64_u64 * div_s64 * div_u64 Raising the minimum supported versions allows us to remove all of the fallback helpers for !COMPILER_HAS_GENERIC_BUILTIN_OVERFLOW, instead dispatching the compiler builtins. arm64 has already raised the minimum supported GCC version to 5.1, do this for all targets now. See the link below for the previous discussion. Link: https://lore.kernel.org/all/[email protected]/ Link: https://lore.kernel.org/lkml/CAK7LNASs6dvU6D3jL2GG3jW58fXfaj6VNOe55NJnTB8UPuk2pA@mail.gmail.com/ Link: https://github.com/ClangBuiltLinux/linux/issues/1438 Reported-by: Stephen Rothwell <[email protected]> Reported-by: Nathan Chancellor <[email protected]> Suggested-by: Rasmus Villemoes <[email protected]> Signed-off-by: Nick Desaulniers <[email protected]> Reviewed-by: Kees Cook <[email protected]> Reviewed-by: Nathan Chancellor <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2021-09-13x86/uaccess: Fix 32-bit __get_user_asm_u64() when CC_HAS_ASM_GOTO_OUTPUT=yWill Deacon1-2/+2
Commit 865c50e1d279 ("x86/uaccess: utilize CONFIG_CC_HAS_ASM_GOTO_OUTPUT") added an optimised version of __get_user_asm() for x86 using 'asm goto'. Like the non-optimised code, the 32-bit implementation of 64-bit get_user() expands to a pair of 32-bit accesses. Unlike the non-optimised code, the _original_ pointer is incremented to copy the high word instead of loading through a new pointer explicitly constructed to point at a 32-bit type. Consequently, if the pointer points at a 64-bit type then we end up loading the wrong data for the upper 32-bits. This was observed as a mount() failure in Android targeting i686 after b0cfcdd9b967 ("d_path: make 'prepend()' fill up the buffer exactly on overflow") because the call to copy_from_kernel_nofault() from prepend_copy() ends up in __get_kernel_nofault() and casts the source pointer to a 'u64 __user *'. An attempt to mount at "/debug_ramdisk" therefore ends up failing trying to mount "/debumdismdisk". Use the existing '__gu_ptr' source pointer to unsigned int for 32-bit __get_user_asm_u64() instead of the original pointer. Cc: Bill Wendling <[email protected]> Cc: Thomas Gleixner <[email protected]> Cc: Ingo Molnar <[email protected]> Cc: Borislav Petkov <[email protected]> Cc: Peter Zijlstra <[email protected]> Reported-by: Greg Kroah-Hartman <[email protected]> Fixes: 865c50e1d279 ("x86/uaccess: utilize CONFIG_CC_HAS_ASM_GOTO_OUTPUT") Signed-off-by: Will Deacon <[email protected]> Reviewed-by: Nick Desaulniers <[email protected]> Tested-by: Greg Kroah-Hartman <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2021-09-13io-wq: provide IO_WQ_* constants for IORING_REGISTER_IOWQ_MAX_WORKERS arg itemsEugene Syromiatnikov2-1/+12
The items passed in the array pointed by the arg parameter of IORING_REGISTER_IOWQ_MAX_WORKERS io_uring_register operation carry certain semantics: they refer to different io-wq worker categories; provide IO_WQ_* constants in the UAPI, so these categories can be referenced in the user space code. Suggested-by: Jens Axboe <[email protected]> Complements: 2e480058ddc21ec5 ("io-wq: provide a way to limit max number of workers") Signed-off-by: Eugene Syromiatnikov <[email protected]> Link: https://lore.kernel.org/r/[email protected] Signed-off-by: Jens Axboe <[email protected]>
2021-09-13dma-debug: prevent an error message from causing runtime problemsHamza Mahfooz1-1/+2
For some drivers, that use the DMA API. This error message can be reached several millions of times per second, causing spam to the kernel's printk buffer and bringing the CPU usage up to 100% (so, it should be rate limited). However, since there is at least one driver that is in the mainline and suffers from the error condition, it is more useful to err_printk() here instead of just rate limiting the error message (in hopes that it will make it easier for other drivers that suffer from this issue to be spotted). Link: https://lkml.kernel.org/r/[email protected] Reported-by: Jeremy Linton <[email protected]> Signed-off-by: Hamza Mahfooz <[email protected]> Signed-off-by: Christoph Hellwig <[email protected]>
2021-09-13nvme: avoid race in shutdown namespace removalDaniel Wagner1-8/+7
When we remove the siblings entry, we update ns->head->list, hence we can't separate the removal and test for being empty. They have to be in the same critical section to avoid a race. To avoid breaking the refcounting imbalance again, add a list empty check to nvme_find_ns_head. Fixes: 5396fdac56d8 ("nvme: fix refcounting imbalance when all paths are down") Signed-off-by: Daniel Wagner <[email protected]> Reviewed-by: Hannes Reinecke <[email protected]> Tested-by: Hannes Reinecke <[email protected]> Signed-off-by: Christoph Hellwig <[email protected]>
2021-09-13nvmet: fix a width vs precision bug in nvmet_subsys_attr_serial_show()Dan Carpenter1-1/+1
This was intended to limit the number of characters printed from "subsys->serial" to NVMET_SN_MAX_SIZE. But accidentally the width specifier was used instead of the precision specifier so it only affects the alignment and not the number of characters printed. Fixes: f04064814c2a ("nvmet: fixup buffer overrun in nvmet_subsys_attr_serial()") Signed-off-by: Dan Carpenter <[email protected]> Signed-off-by: Christoph Hellwig <[email protected]>
2021-09-13Merge branch 'hns3-fixes'David S. Miller4-14/+29
Guangbin Huang says: ==================== net: hns3: add some fixes for -net This series adds some fixes for the HNS3 ethernet driver. ==================== Signed-off-by: David S. Miller <[email protected]>
2021-09-13net: hns3: fix the timing issue of VF clearing interrupt sourcesJiaran Zhang1-3/+3
Currently, the VF does not clear the interrupt source immediately after receiving the interrupt. As a result, if the second interrupt task is triggered when processing the first interrupt task, clearing the interrupt source before exiting will clear the interrupt sources of the two tasks at the same time. As a result, no interrupt is triggered for the second task. The VF detects the missed message only when the next interrupt is generated. Clearing it immediately after executing check_evt_cause ensures that: 1. Even if two interrupt tasks are triggered at the same time, they can be processed. 2. If the second task is triggered during the processing of the first task and the interrupt source is not cleared, the interrupt is reported after vector0 is enabled. Fixes: b90fcc5bd904 ("net: hns3: add reset handling for VF when doing Core/Global/IMP reset") Signed-off-by: Jiaran Zhang <[email protected]> Signed-off-by: Guangbin Huang <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2021-09-13net: hns3: fix the exception when query imp infoJiaran Zhang1-0/+4
When the command for querying imp info is issued to the firmware, if the firmware does not support the command, the returned value of bd num is 0. Add protection mechanism before alloc memory to prevent apply for 0-length memory. Fixes: 0b198b0d80ea ("net: hns3: refactor dump m7 info of debugfs") Signed-off-by: Jiaran Zhang <[email protected]> Signed-off-by: Guangbin Huang <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2021-09-13net: hns3: disable mac in flr processYufeng Mo1-2/+3
The firmware will not disable mac in flr process. Therefore, the driver needs to proactively disable mac during flr, which is the same as the function reset. Fixes: 35d93a30040c ("net: hns3: adjust the process of PF reset") Signed-off-by: Yufeng Mo <[email protected]> Signed-off-by: Guangbin Huang <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2021-09-13net: hns3: change affinity_mask to numa node rangeYufeng Mo1-6/+8
Currently, affinity_mask is set to a single cpu. As a result, irqbalance becomes invalid in SUBSET or EXACT mode. To solve this problem, change affinity_mask to numa node range. In this way, irqbalance can be performed on the cpu of the numa node. Fixes: 0812545487ec ("net: hns3: add interrupt affinity support for misc interrupt") Signed-off-by: Yufeng Mo <[email protected]> Signed-off-by: Guangbin Huang <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2021-09-13net: hns3: pad the short tunnel frame before sending to hardwareYufeng Mo1-2/+6
The hardware cannot handle short tunnel frames below 65 bytes, and will cause vlan tag missing problem. So pads packet size to 65 bytes for tunnel frames to fix this bug. Fixes: 3db084d28dc0("net: hns3: Fix for vxlan tx checksum bug") Signed-off-by: Yufeng Mo <[email protected]> Signed-off-by: Guangbin Huang <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2021-09-13net: hns3: add option to turn off page pool featureYunsheng Lin1-1/+5
When page pool is added to the hns3 driver, it is always enabled unconditionally, which means spilt page handling in the hns3 driver is dead code. As there is a requirement to test the performance between spilt page handling in driver and page pool, so add a module param to support disabling the page pool. When the page pool is proved to perform better in most case, the spilt page handling in driver can be removed. Fixes: 93188e9642c3 ("net: hns3: support skb's frag page recycling based on page pool") Signed-off-by: Yunsheng Lin <[email protected]> Signed-off-by: Guangbin Huang <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2021-09-13powerpc/mce: Fix access error in mce handlerGanesh Goudar1-2/+15
We queue an irq work for deferred processing of mce event in realmode mce handler, where translation is disabled. Queuing of the work may result in accessing memory outside RMO region, such access needs the translation to be enabled for an LPAR running with hash mmu else the kernel crashes. After enabling translation in mce_handle_error() we used to leave it enabled to avoid crashing here, but now with the commit 74c3354bc1d89 ("powerpc/pseries/mce: restore msr before returning from handler") we are restoring the MSR to disable translation. Hence to fix this enable the translation before queuing the work. Without this change following trace is seen on injecting SLB multihit in an LPAR running with hash mmu. Oops: Kernel access of bad area, sig: 11 [#1] LE PAGE_SIZE=64K MMU=Hash SMP NR_CPUS=2048 NUMA pSeries CPU: 5 PID: 1883 Comm: insmod Tainted: G OE 5.14.0-mce+ #137 NIP: c000000000735d60 LR: c000000000318640 CTR: 0000000000000000 REGS: c00000001ebff9a0 TRAP: 0300 Tainted: G OE (5.14.0-mce+) MSR: 8000000000001003 <SF,ME,RI,LE> CR: 28008228 XER: 00000001 CFAR: c00000000031863c DAR: c00000027fa8fe08 DSISR: 40000000 IRQMASK: 0 ... NIP llist_add_batch+0x0/0x40 LR __irq_work_queue_local+0x70/0xc0 Call Trace: 0xc00000001ebffc0c (unreliable) irq_work_queue+0x40/0x70 machine_check_queue_event+0xbc/0xd0 machine_check_early_common+0x16c/0x1f4 Fixes: 74c3354bc1d89 ("powerpc/pseries/mce: restore msr before returning from handler") Signed-off-by: Ganesh Goudar <[email protected]> [mpe: Fix comment formatting, trim oops in change log for readability] Signed-off-by: Michael Ellerman <[email protected]> Link: https://lore.kernel.org/r/[email protected]
2021-09-13KVM: PPC: Book3S HV: Tolerate treclaim. in fake-suspend mode changing registersNicholas Piggin1-2/+34
POWER9 DD2.2 and 2.3 hardware implements a "fake-suspend" mode where certain TM instructions executed in HV=0 mode cause softpatch interrupts so the hypervisor can emulate them and prevent problematic processor conditions. In this fake-suspend mode, the treclaim. instruction does not modify registers. Unfortunately the rfscv instruction executed by the guest do not generate softpatch interrupts, which can cause the hypervisor to lose track of the fake-suspend mode, and it can execute this treclaim. while not in fake-suspend mode. This modifies GPRs and crashes the hypervisor. It's not trivial to disable scv in the guest with HFSCR now, because they assume a POWER9 has scv available. So this fix saves and restores checkpointed registers across the treclaim. Fixes: 7854f7545bff ("KVM: PPC: Book3S: Rework TM save/restore code and make it C-callable") Signed-off-by: Nicholas Piggin <[email protected]> Signed-off-by: Michael Ellerman <[email protected]> Link: https://lore.kernel.org/r/[email protected]
2021-09-13powerpc/64s: system call rfscv workaround for TM bugsNicholas Piggin1-0/+13
The rfscv instruction does not work correctly with the fake-suspend mode in POWER9, which can end up with the hypervisor restoring an incorrect checkpoint. Work around this by setting the _TIF_RESTOREALL flag if a system call returns to a transaction active state, causing rfid to be used instead of rfscv to return, which will do the right thing. The contents of the registers are irrelevant because they will be overwritten in this case anyway. Fixes: 7fa95f9adaee7 ("powerpc/64s: system call support for scv/rfscv instructions") Reported-by: Eirik Fuller <[email protected]> Signed-off-by: Nicholas Piggin <[email protected]> Signed-off-by: Michael Ellerman <[email protected]> Link: https://lore.kernel.org/r/[email protected]
2021-09-13selftests/powerpc: Add scv versions of the basic TM syscall testsNicholas Piggin2-8/+65
The basic TM vs syscall test code hard codes an sc instruction for the system call, which fails to cover scv even when the userspace libc has support for it. Duplicate the tests with hard coded scv variants so both are tested when possible. Signed-off-by: Nicholas Piggin <[email protected]> [mpe: Fix build on old toolchains by using .long for scv] Signed-off-by: Michael Ellerman <[email protected]> Link: https://lore.kernel.org/r/[email protected]
Powered by cgit, Gruvboxed by Blaster4385.