From: kernel test robot <lkp@intel.com>
To: 'Guanjun' <guanjun@linux.alibaba.com>
Cc: llvm@lists.linux.dev, oe-kbuild-all@lists.linux.dev
Subject: Re: [PATCH RFC v1 1/2] genirq/affinity: add support for limiting managed interrupts
Date: Sun, 3 Nov 2024 00:41:00 +0800 [thread overview]
Message-ID: <202411030024.eIDPlX3p-lkp@intel.com> (raw)
In-Reply-To: <20241031074618.3585491-2-guanjun@linux.alibaba.com>
Hi 'Guanjun',
[This is a private test report for your RFC patch.]
kernel test robot noticed the following build errors:
[auto build test ERROR on tip/irq/core]
[also build test ERROR on axboe-block/for-next mszeredi-fuse/for-next tip/smp/core linus/master v6.12-rc5 next-20241101]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]
url: https://github.com/intel-lab-lkp/linux/commits/Guanjun/genirq-affinity-add-support-for-limiting-managed-interrupts/20241031-154824
base: tip/irq/core
patch link: https://lore.kernel.org/r/20241031074618.3585491-2-guanjun%40linux.alibaba.com
patch subject: [PATCH RFC v1 1/2] genirq/affinity: add support for limiting managed interrupts
config: arm-allnoconfig (https://download.01.org/0day-ci/archive/20241103/202411030024.eIDPlX3p-lkp@intel.com/config)
compiler: clang version 20.0.0git (https://github.com/llvm/llvm-project 639a7ac648f1e50ccd2556e17d401c04f9cce625)
reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20241103/202411030024.eIDPlX3p-lkp@intel.com/reproduce)
If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <lkp@intel.com>
| Closes: https://lore.kernel.org/oe-kbuild-all/202411030024.eIDPlX3p-lkp@intel.com/
All errors (new ones prefixed by >>):
>> lib/group_cpus.c:480:17: error: conflicting types for 'group_cpus_evenly'
480 | struct cpumask *group_cpus_evenly(unsigned int numgrps)
| ^
include/linux/group_cpus.h:12:17: note: previous declaration is here
12 | struct cpumask *group_cpus_evenly(unsigned int numgrps, bool is_managed);
| ^
1 error generated.
vim +/group_cpus_evenly +480 lib/group_cpus.c
f7b3ea8cf72f3d Ming Lei 2022-12-27 379
f7b3ea8cf72f3d Ming Lei 2022-12-27 380 /**
f7b3ea8cf72f3d Ming Lei 2022-12-27 381 * group_cpus_evenly - Group all CPUs evenly per NUMA/CPU locality
f7b3ea8cf72f3d Ming Lei 2022-12-27 382 * @numgrps: number of groups
59cfa36232b76f Guanjun 2024-10-31 383 * @is_managed: if these groups managed by kernel
f7b3ea8cf72f3d Ming Lei 2022-12-27 384 *
f7b3ea8cf72f3d Ming Lei 2022-12-27 385 * Return: cpumask array if successful, NULL otherwise. And each element
f7b3ea8cf72f3d Ming Lei 2022-12-27 386 * includes CPUs assigned to this group
f7b3ea8cf72f3d Ming Lei 2022-12-27 387 *
f7b3ea8cf72f3d Ming Lei 2022-12-27 388 * Try to put close CPUs from viewpoint of CPU and NUMA locality into
f7b3ea8cf72f3d Ming Lei 2022-12-27 389 * same group, and run two-stage grouping:
f7b3ea8cf72f3d Ming Lei 2022-12-27 390 * 1) allocate present CPUs on these groups evenly first
f7b3ea8cf72f3d Ming Lei 2022-12-27 391 * 2) allocate other possible CPUs on these groups evenly
f7b3ea8cf72f3d Ming Lei 2022-12-27 392 *
f7b3ea8cf72f3d Ming Lei 2022-12-27 393 * We guarantee in the resulted grouping that all CPUs are covered, and
f7b3ea8cf72f3d Ming Lei 2022-12-27 394 * no same CPU is assigned to multiple groups
f7b3ea8cf72f3d Ming Lei 2022-12-27 395 */
59cfa36232b76f Guanjun 2024-10-31 396 struct cpumask *group_cpus_evenly(unsigned int numgrps, bool is_managed)
f7b3ea8cf72f3d Ming Lei 2022-12-27 397 {
f7b3ea8cf72f3d Ming Lei 2022-12-27 398 unsigned int curgrp = 0, nr_present = 0, nr_others = 0;
f7b3ea8cf72f3d Ming Lei 2022-12-27 399 cpumask_var_t *node_to_cpumask;
f7b3ea8cf72f3d Ming Lei 2022-12-27 400 cpumask_var_t nmsk, npresmsk;
f7b3ea8cf72f3d Ming Lei 2022-12-27 401 int ret = -ENOMEM;
f7b3ea8cf72f3d Ming Lei 2022-12-27 402 struct cpumask *masks = NULL;
f7b3ea8cf72f3d Ming Lei 2022-12-27 403
f7b3ea8cf72f3d Ming Lei 2022-12-27 404 if (!zalloc_cpumask_var(&nmsk, GFP_KERNEL))
f7b3ea8cf72f3d Ming Lei 2022-12-27 405 return NULL;
f7b3ea8cf72f3d Ming Lei 2022-12-27 406
f7b3ea8cf72f3d Ming Lei 2022-12-27 407 if (!zalloc_cpumask_var(&npresmsk, GFP_KERNEL))
f7b3ea8cf72f3d Ming Lei 2022-12-27 408 goto fail_nmsk;
f7b3ea8cf72f3d Ming Lei 2022-12-27 409
f7b3ea8cf72f3d Ming Lei 2022-12-27 410 node_to_cpumask = alloc_node_to_cpumask();
f7b3ea8cf72f3d Ming Lei 2022-12-27 411 if (!node_to_cpumask)
f7b3ea8cf72f3d Ming Lei 2022-12-27 412 goto fail_npresmsk;
f7b3ea8cf72f3d Ming Lei 2022-12-27 413
f7b3ea8cf72f3d Ming Lei 2022-12-27 414 masks = kcalloc(numgrps, sizeof(*masks), GFP_KERNEL);
f7b3ea8cf72f3d Ming Lei 2022-12-27 415 if (!masks)
f7b3ea8cf72f3d Ming Lei 2022-12-27 416 goto fail_node_to_cpumask;
f7b3ea8cf72f3d Ming Lei 2022-12-27 417
f7b3ea8cf72f3d Ming Lei 2022-12-27 418 build_node_to_cpumask(node_to_cpumask);
f7b3ea8cf72f3d Ming Lei 2022-12-27 419
0263f92fadbb9d Ming Lei 2023-11-20 420 /*
0263f92fadbb9d Ming Lei 2023-11-20 421 * Make a local cache of 'cpu_present_mask', so the two stages
0263f92fadbb9d Ming Lei 2023-11-20 422 * spread can observe consistent 'cpu_present_mask' without holding
0263f92fadbb9d Ming Lei 2023-11-20 423 * cpu hotplug lock, then we can reduce deadlock risk with cpu
0263f92fadbb9d Ming Lei 2023-11-20 424 * hotplug code.
0263f92fadbb9d Ming Lei 2023-11-20 425 *
0263f92fadbb9d Ming Lei 2023-11-20 426 * Here CPU hotplug may happen when reading `cpu_present_mask`, and
0263f92fadbb9d Ming Lei 2023-11-20 427 * we can live with the case because it only affects that hotplug
0263f92fadbb9d Ming Lei 2023-11-20 428 * CPU is handled in the 1st or 2nd stage, and either way is correct
0263f92fadbb9d Ming Lei 2023-11-20 429 * from API user viewpoint since 2-stage spread is sort of
0263f92fadbb9d Ming Lei 2023-11-20 430 * optimization.
0263f92fadbb9d Ming Lei 2023-11-20 431 */
0263f92fadbb9d Ming Lei 2023-11-20 432 cpumask_copy(npresmsk, data_race(cpu_present_mask));
0263f92fadbb9d Ming Lei 2023-11-20 433
59cfa36232b76f Guanjun 2024-10-31 434 /* Limit the count of managed interrupts on every node */
59cfa36232b76f Guanjun 2024-10-31 435 if (is_managed && managed_irqs_per_node)
59cfa36232b76f Guanjun 2024-10-31 436 __group_prepare_affinity(npresmsk, node_to_cpumask);
59cfa36232b76f Guanjun 2024-10-31 437
f7b3ea8cf72f3d Ming Lei 2022-12-27 438 /* grouping present CPUs first */
f7b3ea8cf72f3d Ming Lei 2022-12-27 439 ret = __group_cpus_evenly(curgrp, numgrps, node_to_cpumask,
0263f92fadbb9d Ming Lei 2023-11-20 440 npresmsk, nmsk, masks);
f7b3ea8cf72f3d Ming Lei 2022-12-27 441 if (ret < 0)
f7b3ea8cf72f3d Ming Lei 2022-12-27 442 goto fail_build_affinity;
f7b3ea8cf72f3d Ming Lei 2022-12-27 443 nr_present = ret;
f7b3ea8cf72f3d Ming Lei 2022-12-27 444
f7b3ea8cf72f3d Ming Lei 2022-12-27 445 /*
f7b3ea8cf72f3d Ming Lei 2022-12-27 446 * Allocate non present CPUs starting from the next group to be
f7b3ea8cf72f3d Ming Lei 2022-12-27 447 * handled. If the grouping of present CPUs already exhausted the
f7b3ea8cf72f3d Ming Lei 2022-12-27 448 * group space, assign the non present CPUs to the already
f7b3ea8cf72f3d Ming Lei 2022-12-27 449 * allocated out groups.
f7b3ea8cf72f3d Ming Lei 2022-12-27 450 */
f7b3ea8cf72f3d Ming Lei 2022-12-27 451 if (nr_present >= numgrps)
f7b3ea8cf72f3d Ming Lei 2022-12-27 452 curgrp = 0;
f7b3ea8cf72f3d Ming Lei 2022-12-27 453 else
f7b3ea8cf72f3d Ming Lei 2022-12-27 454 curgrp = nr_present;
0263f92fadbb9d Ming Lei 2023-11-20 455 cpumask_andnot(npresmsk, cpu_possible_mask, npresmsk);
f7b3ea8cf72f3d Ming Lei 2022-12-27 456 ret = __group_cpus_evenly(curgrp, numgrps, node_to_cpumask,
f7b3ea8cf72f3d Ming Lei 2022-12-27 457 npresmsk, nmsk, masks);
f7b3ea8cf72f3d Ming Lei 2022-12-27 458 if (ret >= 0)
f7b3ea8cf72f3d Ming Lei 2022-12-27 459 nr_others = ret;
f7b3ea8cf72f3d Ming Lei 2022-12-27 460
f7b3ea8cf72f3d Ming Lei 2022-12-27 461 fail_build_affinity:
f7b3ea8cf72f3d Ming Lei 2022-12-27 462 if (ret >= 0)
f7b3ea8cf72f3d Ming Lei 2022-12-27 463 WARN_ON(nr_present + nr_others < numgrps);
f7b3ea8cf72f3d Ming Lei 2022-12-27 464
f7b3ea8cf72f3d Ming Lei 2022-12-27 465 fail_node_to_cpumask:
f7b3ea8cf72f3d Ming Lei 2022-12-27 466 free_node_to_cpumask(node_to_cpumask);
f7b3ea8cf72f3d Ming Lei 2022-12-27 467
f7b3ea8cf72f3d Ming Lei 2022-12-27 468 fail_npresmsk:
f7b3ea8cf72f3d Ming Lei 2022-12-27 469 free_cpumask_var(npresmsk);
f7b3ea8cf72f3d Ming Lei 2022-12-27 470
f7b3ea8cf72f3d Ming Lei 2022-12-27 471 fail_nmsk:
f7b3ea8cf72f3d Ming Lei 2022-12-27 472 free_cpumask_var(nmsk);
f7b3ea8cf72f3d Ming Lei 2022-12-27 473 if (ret < 0) {
f7b3ea8cf72f3d Ming Lei 2022-12-27 474 kfree(masks);
f7b3ea8cf72f3d Ming Lei 2022-12-27 475 return NULL;
f7b3ea8cf72f3d Ming Lei 2022-12-27 476 }
f7b3ea8cf72f3d Ming Lei 2022-12-27 477 return masks;
f7b3ea8cf72f3d Ming Lei 2022-12-27 478 }
188a569658584e Ingo Molnar 2023-01-18 479 #else /* CONFIG_SMP */
f7b3ea8cf72f3d Ming Lei 2022-12-27 @480 struct cpumask *group_cpus_evenly(unsigned int numgrps)
--
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki
next prev parent reply other threads:[~2024-11-02 16:41 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-10-31 7:46 [PATCH RFC v1 0/2] Support for limiting the number of managed interrupts on every node per allocation 'Guanjun'
2024-10-31 7:46 ` [PATCH RFC v1 1/2] genirq/affinity: add support for limiting managed interrupts 'Guanjun'
2024-10-31 10:35 ` Thomas Gleixner
2024-10-31 10:50 ` Ming Lei
[not found] ` <43FD1116-C188-4729-A3AB-C2A0F5A087D2@linux.alibaba.com>
2024-11-01 3:34 ` Jason Wang
2024-11-01 3:03 ` mapicccy
2024-11-01 23:37 ` Thomas Gleixner
2024-11-01 7:06 ` Jiri Slaby
2024-11-02 16:30 ` kernel test robot
2024-11-02 16:41 ` kernel test robot [this message]
2024-10-31 7:46 ` [PATCH RFC v1 2/2] genirq/cpuhotplug: Handle managed IRQs when the last CPU hotplug out in the affinity 'Guanjun'
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=202411030024.eIDPlX3p-lkp@intel.com \
--to=lkp@intel.com \
--cc=guanjun@linux.alibaba.com \
--cc=llvm@lists.linux.dev \
--cc=oe-kbuild-all@lists.linux.dev \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.