From: Jan Stancek Date: Fri, 8 Mar 2019 06:21:31 -0500 (EST) Subject: [LTP] [PATCH v3] mm: rewrite mtest01 with new API In-Reply-To: <20190308101049.21201-1-liwang@redhat.com> References: <20190308101049.21201-1-liwang@redhat.com> Message-ID: <2064253320.6099818.1552044091281.JavaMail.zimbra@redhat.com> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: ltp@lists.linux.it ----- Original Message ----- > Test issue: > mtest01 start many children to alloc chunck of memory and do write > page(with -w option), but occasionally some children were killed by > oom-killer and exit with SIGCHLD signal sending. After the parent > reciving this SIGCHLD signal it will report FAIL as a test result. > > It seems not a real kernel bug if something just like that, it's > trying to use 80% of memory and swap. Once it uses most of memory, > system starts swapping, but the test is likely consuming memory at > greater rate than kswapd can provide, which eventually triggers OOM. > > ---- FAIL LOG ---- > mtest01 0 TINFO : Total memory already used on system = 1027392 > kbytes > mtest01 0 TINFO : Total memory used needed to reach maximum = > 12715520 kbytes > mtest01 0 TINFO : Filling up 80% of ram which is 11688128 kbytes > mtest01 1 TFAIL : mtest01.c:314: child process exited unexpectedly > ------------------- > > Rewrite changes: > To make mtest01 more easier to understand, I just rewrite it into > LTP new API and make a little changes in children behavior. > > * decrease the pressure to 80% of free memory for testing > * drop the signal SIGCHLD action becasue new API help to > check_child_status > * make child pause itself after finishing their memory allocating/writing > * parent sends SIGCONT to make children continue and exit > * use TST_PROCESS_STATE_WAIT to wait child changes to 'T' state > * involve ALLOC_THRESHOLD to rework same code in defines > * to make mtest01 support running with -i N > 1 > > Signed-off-by: Li Wang > --- > > Notes: > v2 --> v3 > * modify test description for new version > * move some global variables to local > * remove SIGCONT sending from cleanup > * involve ALLOC_THRESHOLD to avoid same code in defines Hi, looks lot more readable than original version. Some nit-picks below / no need to re-post for that typo, I can fix it before commit. > +#if defined (_s390_) > +#define ALLOC_THRESHOLD FIVE_HUNDRED_MB > +#elif __WORDSIZE == 32 > +#define ALLOC_THRESHOL (unsigned long long)(2*FIVE_HUNDRED_MB) typo here, missing 'D' > +#elif __WORDSIZE == 64 > +#define ALLOC_THRESHOLD (unsigned long long)(6*FIVE_HUNDRED_MB) > +#endif > > -static void handler(int signo) > +#define STOP_THRESHOLD 15 /* seconds remaining before reaching timeout */ > + > +static pid_t *pid_list; > +static sig_atomic_t pid_count; > +static int max_pids; > +static unsigned long long alloc_maxbytes; > +static unsigned long long original_maxbytes; We could use "original_maxbytes" in setup() and options(), then alloc_maxbytes could be local variable of mem_test(). > +static void do_write_page(char *mem, int chunksize) > +{ > + int i; > > - free(pid_list); > + for (i = 0; i < chunksize; i++) > + *(mem + i) = 'a'; We could do i+=pagesize to make it slightly faster. > +static void child_loop_alloc(unsigned long long alloc_bytes) > +{ > + unsigned long bytecount = (unsigned long)chunksize; Shouldn't this variable start at 0? > + char *mem; > + > + while (1) { > + mem = SAFE_MALLOC(chunksize); > + if (dowrite) > + do_write_page(mem, chunksize); > + > + if (verbose) > + tst_res(TINFO, > + "child %d allocated %lu bytes chunksize is %d", > + getpid(), bytecount, chunksize); > + bytecount += chunksize; > + if (bytecount >= alloc_bytes) > + break; > } Overall, the rewrite looks good to me. Thanks, Jan