Hi Hillf, [FYI, it's a private test report for your RFC patch.] [auto build test ERROR on linus/master] [cannot apply to v5.4-rc2 next-20191011] [if your patch is applied to the wrong git tree, please drop us a note to help improve the system. BTW, we also suggest to use '--base' option to specify the base tree in git format-patch, please see https://stackoverflow.com/a/37406982] url: https://github.com/0day-ci/linux/commits/Hillf-Danton/writeback-add-elast... config: i386-tinyconfig (attached as .config) compiler: gcc-7 (Debian 7.4.0-13) 7.4.0 reproduce: # save the attached .config to linux build tree make ARCH=i386 If you fix the issue, kindly add following tag Reported-by: kbuild test robot <lkp(a)intel.com&gt; All error/warnings (new ones prefixed by >>): mm/page-writeback.c: In function 'balance_dirty_pages_ratelimited': cgwb_bdp(wb); ^~~~~~~~ if (unlikely(current->nr_dirtied >= ratelimit)) ^ cc1: some warnings being treated as errors -- In file included from mm/backing-dev.c:3:0: mm/backing-dev.c: In function 'wb_init': init_waitqueue_head(&wb->bdp_waitq); ^ include/linux/wait.h:67:26: note: in definition of macro 'init_waitqueue_head' __init_waitqueue_head((wq_head), #wq_head, &__key); \ ^~~~~~~ vim +/cgwb_bdp +1954 mm/page-writeback.c 1889 1890 /** 1891 * balance_dirty_pages_ratelimited - balance dirty memory state 1892 * @mapping: address_space which was dirtied 1893 * 1894 * Processes which are dirtying memory should call in here once for each page 1895 * which was newly dirtied. The function will periodically check the system's 1896 * dirty state and will initiate writeback if needed. 1897 * 1898 * On really big machines, get_writeback_state is expensive, so try to avoid 1899 * calling it too often (ratelimiting). But once we're over the dirty memory 1900 * limit we decrease the ratelimiting by a lot, to prevent individual processes 1901 * from overshooting the limit by (ratelimit_pages) each. 1902 */ 1903 void balance_dirty_pages_ratelimited(struct address_space *mapping) 1904 { 1905 struct inode *inode = mapping->host; 1906 struct backing_dev_info *bdi = inode_to_bdi(inode); 1907 struct bdi_writeback *wb = NULL; 1908 int ratelimit; 1909 int *p; 1910 1911 if (!bdi_cap_account_dirty(bdi)) 1912 return; 1913 1914 if (inode_cgwb_enabled(inode)) 1915 wb = wb_get_create_current(bdi, GFP_KERNEL); 1916 if (!wb) 1917 wb = &bdi->wb; 1918 1919 ratelimit = current->nr_dirtied_pause; 1920 if (wb->dirty_exceeded) 1921 ratelimit = min(ratelimit, 32 >> (PAGE_SHIFT - 10)); 1922 1923 preempt_disable(); 1924 /* 1925 * This prevents one CPU to accumulate too many dirtied pages without 1926 * calling into balance_dirty_pages(), which can happen when there are 1927 * 1000+ tasks, all of them start dirtying pages at exactly the same 1928 * time, hence all honoured too large initial task->nr_dirtied_pause. 1929 */ 1930 p = this_cpu_ptr(&bdp_ratelimits); 1931 if (unlikely(current->nr_dirtied >= ratelimit)) 1932 *p = 0; 1933 else if (unlikely(*p >= ratelimit_pages)) { 1934 *p = 0; 1935 ratelimit = 0; 1936 } 1937 /* 1938 * Pick up the dirtied pages by the exited tasks. This avoids lots of 1939 * short-lived tasks (eg. gcc invocations in a kernel build) escaping 1940 * the dirty throttling and livelock other long-run dirtiers. 1941 */ 1942 p = this_cpu_ptr(&dirty_throttle_leaks); 1943 if (*p > 0 && current->nr_dirtied < ratelimit) { 1944 unsigned long nr_pages_dirtied; 1945 nr_pages_dirtied = min(*p, ratelimit - current->nr_dirtied); 1946 *p -= nr_pages_dirtied; 1947 current->nr_dirtied += nr_pages_dirtied; 1948 } 1949 preempt_enable(); 1950 1952 if (IS_ENABLED(CONFIG_CGROUP_WRITEBACK) && 1953 IS_ENABLED(CONFIG_CGWB_BDP_WITH_EBDI)) 1955 else 1956 balance_dirty_pages(wb, current->nr_dirtied); 1957 1958 wb_put(wb); 1959 } 1960 EXPORT_SYMBOL(balance_dirty_pages_ratelimited); 1961 --- 0-DAY kernel test infrastructure Open Source Technology Center https://lists.01.org/pipermail/kbuild-all Intel Corporation