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 --- crypto/842.c | 87 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 87 insertions(+) create mode 100644 crypto/842.c (limited to 'crypto/842.c') diff --git a/crypto/842.c b/crypto/842.c new file mode 100644 index 000000000..4007e87be --- /dev/null +++ b/crypto/842.c @@ -0,0 +1,87 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Cryptographic API for the 842 software compression algorithm. + * + * Copyright (C) IBM Corporation, 2011-2015 + * + * Original Authors: Robert Jennings + * Seth Jennings + * + * Rewrite: Dan Streetman + * + * This is the software implementation of compression and decompression using + * the 842 format. This uses the software 842 library at lib/842/ which is + * only a reference implementation, and is very, very slow as compared to other + * software compressors. You probably do not want to use this software + * compression. If you have access to the PowerPC 842 compression hardware, you + * want to use the 842 hardware compression interface, which is at: + * drivers/crypto/nx/nx-842-crypto.c + */ + +#include +#include +#include +#include + +static void *crypto842_alloc_ctx(void) +{ + void *ctx; + + ctx = kmalloc(SW842_MEM_COMPRESS, GFP_KERNEL); + if (!ctx) + return ERR_PTR(-ENOMEM); + + return ctx; +} + +static void crypto842_free_ctx(void *ctx) +{ + kfree(ctx); +} + +static int crypto842_scompress(struct crypto_scomp *tfm, + const u8 *src, unsigned int slen, + u8 *dst, unsigned int *dlen, void *ctx) +{ + return sw842_compress(src, slen, dst, dlen, ctx); +} + +static int crypto842_sdecompress(struct crypto_scomp *tfm, + const u8 *src, unsigned int slen, + u8 *dst, unsigned int *dlen, void *ctx) +{ + return sw842_decompress(src, slen, dst, dlen); +} + +static struct scomp_alg scomp = { + .streams = { + .alloc_ctx = crypto842_alloc_ctx, + .free_ctx = crypto842_free_ctx, + }, + .compress = crypto842_scompress, + .decompress = crypto842_sdecompress, + .base = { + .cra_name = "842", + .cra_driver_name = "842-scomp", + .cra_priority = 100, + .cra_module = THIS_MODULE, + } +}; + +static int __init crypto842_mod_init(void) +{ + return crypto_register_scomp(&scomp); +} +module_init(crypto842_mod_init); + +static void __exit crypto842_mod_exit(void) +{ + crypto_unregister_scomp(&scomp); +} +module_exit(crypto842_mod_exit); + +MODULE_LICENSE("GPL"); +MODULE_DESCRIPTION("842 Software Compression Algorithm"); +MODULE_ALIAS_CRYPTO("842"); +MODULE_ALIAS_CRYPTO("842-generic"); +MODULE_AUTHOR("Dan Streetman "); -- cgit v1.3.1