From 0d6526f82c3cdefcca47f73f5fc08dc6f335eac6 Mon Sep 17 00:00:00 2001 From: Tim Chen Date: Mon, 21 Sep 2026 17:37:22 -0700 Subject: sched/cache: Keep nr_pref_llc_running in the runnable domain, to fix LLC mis-scheduling bug alb_break_llc() decides whether to break LLC preference during active load balance. It does so by testing that every runnable fair task on the source rq prefers its LLC: env->src_rq->nr_pref_llc_running == env->src_rq->cfs.h_nr_runnable But the two counters cover different sets. nr_pref_llc_running is updated in account_llc_enqueue()/account_llc_dequeue(), next to cfs_rq->nr_queued, so it follows queued tasks. h_nr_runnable is updated in set_delayed()/ clear_delayed() and drops delay-dequeued tasks. So under DELAY_DEQUEUE, a preferring task that goes to sleep stays counted in nr_pref_llc_running while h_nr_runnable falls. The equality then breaks, alb_break_llc() returns false, and active balance is free to pull a task off its preferred LLC. Active balance only moves runnable tasks, and this is the only LLC check it consults: once the stopper runs, LBF_ACTIVE_LB skips the per-task test in can_migrate_task(). The runnable set is the one we want. Fix it on the counter side. A task should be counted in nr_pref_llc_running exactly while it is both queued on its preferred LLC (pref_llc_queued) and runnable (!sched_delayed). Define that membership once in task_pref_llc_runnable(), and adjust the counter only through pref_llc_running_inc()/pref_llc_running_dec() from the four sites that change either input: account_llc_enqueue(), account_llc_dequeue(), set_delayed() and clear_delayed(). Gating every update on the same predicate keeps the delay, wake and dequeue paths from double-counting or underflowing; see the comments at those sites for the ordering. nr_llc_running and sd->llc_counts are not touched and stay on queued semantics. Fixes: 714059f79ff0 ("sched/cache: Handle moving single tasks to/from their preferred LLC") Closes: https://lore.kernel.org/lkml/20260827135000.735138-1-zhanxusheng@xiaomi.com/ Reported-by: Zhan Xusheng Suggested-by: Chen Yu Signed-off-by: Tim Chen Signed-off-by: Peter Zijlstra (Intel) Signed-off-by: Ingo Molnar Reviewed-by: Kayra Cizmeci Cc: # v7.2.x Link: https://patch.msgid.link/06af61afedac32e6477f57feb4d658f6c411c3af.1790035273.git.tim.c.chen@linux.intel.com --- Documentation/ABI/testing/sysfs-driver-uacce | 57 ++++++++++++++++++++++++++++ 1 file changed, 57 insertions(+) create mode 100644 Documentation/ABI/testing/sysfs-driver-uacce (limited to 'Documentation/ABI/testing/sysfs-driver-uacce') diff --git a/Documentation/ABI/testing/sysfs-driver-uacce b/Documentation/ABI/testing/sysfs-driver-uacce new file mode 100644 index 000000000..d3f0b8f3c --- /dev/null +++ b/Documentation/ABI/testing/sysfs-driver-uacce @@ -0,0 +1,57 @@ +What: /sys/class/uacce//api +Date: Feb 2020 +KernelVersion: 5.7 +Contact: linux-accelerators@lists.ozlabs.org +Description: Api of the device + Can be any string and up to userspace to parse. + Application use the api to match the correct driver + +What: /sys/class/uacce//flags +Date: Feb 2020 +KernelVersion: 5.7 +Contact: linux-accelerators@lists.ozlabs.org +Description: Attributes of the device, see UACCE_DEV_xxx flag defined in uacce.h + +What: /sys/class/uacce//available_instances +Date: Feb 2020 +KernelVersion: 5.7 +Contact: linux-accelerators@lists.ozlabs.org +Description: Available instances left of the device + Return -ENODEV if uacce_ops get_available_instances is not provided + +What: /sys/class/uacce//isolate_strategy +Date: Nov 2022 +KernelVersion: 6.1 +Contact: linux-accelerators@lists.ozlabs.org +Description: (RW) A sysfs node that configure the error threshold for the hardware + isolation strategy. This size is a configured integer value, which is the + number of threshold for hardware errors occurred in one hour. The default is 0. + 0 means never isolate the device. The maximum value is 65535. You can write + a number of threshold based on your hardware. + +What: /sys/class/uacce//isolate +Date: Nov 2022 +KernelVersion: 6.1 +Contact: linux-accelerators@lists.ozlabs.org +Description: (R) A sysfs node that read the device isolated state. The value 1 + means the device is unavailable. The 0 means the device is + available. + +What: /sys/class/uacce//algorithms +Date: Feb 2020 +KernelVersion: 5.7 +Contact: linux-accelerators@lists.ozlabs.org +Description: Algorithms supported by this accelerator, separated by new line. + Can be any string and up to userspace to parse. + +What: /sys/class/uacce//region_mmio_size +Date: Feb 2020 +KernelVersion: 5.7 +Contact: linux-accelerators@lists.ozlabs.org +Description: Size (bytes) of mmio region queue file + +What: /sys/class/uacce//region_dus_size +Date: Feb 2020 +KernelVersion: 5.7 +Contact: linux-accelerators@lists.ozlabs.org +Description: Size (bytes) of dus region queue file -- cgit v1.3.1