From mboxrd@z Thu Jan 1 00:00:00 1970 From: "heming.zhao@suse.com" Subject: Re: [PATCH v5 1/2] md-cluster: fix safemode_delay value when converting to clustered bitmap Date: Tue, 21 Jul 2020 16:58:35 +0800 Message-ID: References: <1595268533-7040-1-git-send-email-heming.zhao@suse.com> <1595268533-7040-2-git-send-email-heming.zhao@suse.com> <8f122dea-2a4a-4523-653f-fe6fc2842653@cloud.ionos.com> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable Return-path: In-Reply-To: <8f122dea-2a4a-4523-653f-fe6fc2842653@cloud.ionos.com> Content-Language: en-US Sender: linux-raid-owner@vger.kernel.org To: Guoqing Jiang , linux-raid@vger.kernel.org Cc: neilb@suse.com, song@kernel.org List-Id: linux-raid.ids On 7/21/20 4:35 PM, Guoqing Jiang wrote: >=20 >=20 > On 7/20/20 8:08 PM, Zhao Heming wrote: >> When array convert to clustered bitmap, the safe_mode_delay doesn't >> clean and vice versa. the /sys/block/mdX/md/safe_mode_delay keep origina= l >> value after changing bitmap type. In safe_delay_store(), the code forbid= s >> setting mddev->safemode_delay when array is clustered. So in cluster-md >> env, the expected safemode_delay value should be 0. >> >> Reproducible steps: >> ``` >> node1 # mdadm --zero-superblock /dev/sd{b,c,d} >> node1 # mdadm -C /dev/md0 -b internal -e 1.2 -n 2 -l mirror /dev/sdb /de= v/sdc >> node1 # cat /sys/block/md0/md/safe_mode_delay >> 0.204 >> node1 # mdadm -G /dev/md0 -b none >> node1 # mdadm --grow /dev/md0 --bitmap=3Dclustered >> node1 # cat /sys/block/md0/md/safe_mode_delay >> 0.204=C2=A0 <=3D=3D doesn't change, should ZERO for cluster-md >> >> node1 # mdadm --zero-superblock /dev/sd{b,c,d} >> node1 # mdadm -C /dev/md0 -b clustered -e 1.2 -n 2 -l mirror /dev/sdb /d= ev/sdc >> node1 # cat /sys/block/md0/md/safe_mode_delay >> 0.000 >> node1 # mdadm -G /dev/md0 -b none >> node1 # cat /sys/block/md0/md/safe_mode_delay >> 0.000=C2=A0 <=3D=3D doesn't change, should default value >> ``` >> >> Signed-off-by: Zhao Heming >> --- >> =C2=A0 drivers/md/md.c | 11 +++++++++-- >> =C2=A0 1 file changed, 9 insertions(+), 2 deletions(-) >> >> diff --git a/drivers/md/md.c b/drivers/md/md.c >> index f567f536b529..1bde3df3fa18 100644 >> --- a/drivers/md/md.c >> +++ b/drivers/md/md.c >> @@ -101,6 +101,8 @@ static void mddev_detach(struct mddev *mddev); >> =C2=A0=C2=A0 * count by 2 for every hour elapsed between read errors. >> =C2=A0=C2=A0 */ >> =C2=A0 #define MD_DEFAULT_MAX_CORRECTED_READ_ERRORS 20 >> +/* Default safemode delay: 200 msec */ >> +#define DEFAULT_SAFEMODE_DELAY ((200 * HZ)/1000 +1) >> =C2=A0 /* >> =C2=A0=C2=A0 * Current RAID-1,4,5 parallel reconstruction 'guaranteed sp= eed limit' >> =C2=A0=C2=A0 * is 1000 KB/sec, so the extra system load does not show up= that much. >> @@ -5982,7 +5984,7 @@ int md_run(struct mddev *mddev) >> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 if (mddev_is_clustered(mddev)) >> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 mddev->safemode_d= elay =3D 0; >> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 else >> -=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 mddev->safemode_delay =3D (2= 00 * HZ)/1000 +1; /* 200 msec delay */ >> +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 mddev->safemode_delay =3D DE= FAULT_SAFEMODE_DELAY; >> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 mddev->in_sync =3D 1; >> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 smp_wmb(); >> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 spin_lock(&mddev->lock); >> @@ -7361,6 +7363,7 @@ static int update_array_info(struct mddev *mddev, = mdu_array_info_t *info) >> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 mddev->bitmap_info.nodes =3D 0; >> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 md_cluster_ops->leave(mddev); >> +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0 mddev->safemode_delay =3D DEFAULT_SAFEMODE_DELAY; >> =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0 } >=20 > Not about the patch itself, just confuse about the meaning of safemode_de= lay. > As safemode_delay represents 200 ms delay, but md_write_end has this. >=20 > =C2=A0=C2=A0=C2=A0 =C2=A0=C2=A0=C2=A0 =C2=A0=C2=A0=C2=A0 =C2=A0=C2=A0=C2= =A0 /* The roundup() ensures this only performs locking once > =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0 * every ->safemode_delay *jiffies* > =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0 */ > =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0 mod_timer(&mddev->safemode_timer, > =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0 roundup(jiffies, mddev->safemode_delay) + > =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0 mddev->safemode_delay); >=20 > And the second argument in mod_timer() means "new timeout in jiffies", > does the above need to convert from ms to jiffies? Am I miss something? >=20 > Thanks, > Guoqing >=20 The safemode_delay stores jiffies not msec. it converts to ms only in safe_= delay_{store|show}. please see default value: #define DEFAULT_SAFEMODE_DELAY ((200 * HZ)/1000 +1) in safe_delay_store, user space input will multiple 10^3, which will conver= t from second to msec. then (*HZ/1000) to jiffies.