* [Patch 2.6] sd.c spinup messages
@ 2004-06-09 23:48 Chuck Ebbert
2004-06-24 22:02 ` James Bottomley
0 siblings, 1 reply; 2+ messages in thread
From: Chuck Ebbert @ 2004-06-09 23:48 UTC (permalink / raw)
To: linux-scsi
With an empty Iomega USB Powered Zip 250 drive and kernel 2.6.6,
I get this in the syslog on bootup:
Jun 9 14:38:42 d2 kernel: usb 1-1: new full speed USB device using address 2
Jun 9 14:38:42 d2 kernel: scsi1 : SCSI emulation for USB Mass Storage devices
Jun 9 14:38:42 d2 kernel: Vendor: IOMEGA Model: ZIP 250 Rev: 61.T
Jun 9 14:38:42 d2 kernel: Type: Direct-Access ANSI SCSI revision: 02
Jun 9 14:38:42 d2 kernel: sda: Spinning up disk....<6>EXT3 FS on hde3, internal journal
Jun 9 14:38:42 d2 kernel: Adding 524152k swap on /dev/hdg2. Priority:-1 extents:1
Jun 9 14:38:42 d2 kernel: Adding 524152k swap on /dev/hde2. Priority:-2 extents:1
Jun 9 14:38:42 d2 kernel: .<5>Attached scsi removable disk sda at scsi1, channel 0, id 0, lun 0
Jun 9 14:38:42 d2 kernel: Attached scsi generic sg1 at scsi1, channel 0, id 0, lun 0, type 0
Jun 9 14:38:42 d2 kernel: kjournald starting. Commit interval 5 seconds
Code in sd.c:sd_spinup_disk is looping and calling printk(".") while
waiting for the disk to spin up. Then it detects media not present
and the function just exits, leaving an incomplete line in the buffer.
Meanwhile other messages get mixed in.
I fixed it so it prints only whole-line messages and emits a
media-not-present message if it printed a spinup message earlier:
Jun 9 16:42:06 d2 kernel: usb 1-1: new full speed USB device using address 2
Jun 9 16:42:06 d2 kernel: scsi1 : SCSI emulation for USB Mass Storage devices
Jun 9 16:42:06 d2 kernel: Vendor: IOMEGA Model: ZIP 250 Rev: 61.T
Jun 9 16:42:06 d2 kernel: Type: Direct-Access ANSI SCSI revision: 02
Jun 9 16:42:06 d2 kernel: sda: Spinning up disk...
Jun 9 16:42:06 d2 kernel: EXT3 FS on hde3, internal journal
Jun 9 16:42:06 d2 kernel: Adding 524152k swap on /dev/hdg2. Priority:-1 extents:1
Jun 9 16:42:06 d2 kernel: Adding 524152k swap on /dev/hde2. Priority:-2 extents:1
Jun 9 16:42:06 d2 kernel: sda: No media present
Jun 9 16:42:06 d2 kernel: Attached scsi removable disk sda at scsi1, channel 0, id 0, lun 0
Jun 9 16:42:06 d2 kernel: Attached scsi generic sg1 at scsi1, channel 0, id 0, lun 0, type 0
Jun 9 16:42:06 d2 kernel: kjournald starting. Commit interval 5 seconds
And with media present:
Jun 9 17:52:46 d2 kernel: usb 1-1: new full speed USB device using address 2
Jun 9 17:52:46 d2 kernel: scsi1 : SCSI emulation for USB Mass Storage devices
Jun 9 17:52:46 d2 kernel: Vendor: IOMEGA Model: ZIP 250 Rev: 61.T
Jun 9 17:52:46 d2 kernel: Type: Direct-Access ANSI SCSI revision: 02
Jun 9 17:52:46 d2 kernel: sda: Spinning up disk...
Jun 9 17:52:46 d2 kernel: EXT3 FS on hde3, internal journal
Jun 9 17:52:46 d2 kernel: Adding 524152k swap on /dev/hdg2. Priority:-1 extents:1
Jun 9 17:52:46 d2 kernel: Adding 524152k swap on /dev/hde2. Priority:-2 extents:1
Jun 9 17:52:46 d2 kernel: kjournald starting. Commit interval 5 seconds
((( Snip 15 lines of ext3 mount messages )))
Jun 9 17:52:46 d2 kernel: sda: Drive ready
Jun 9 17:52:46 d2 kernel: SCSI device sda: 196608 512-byte hdwr sectors (101 MB)
Jun 9 17:52:46 d2 kernel: sda: assuming Write Enabled
Jun 9 17:52:46 d2 kernel: sda: assuming drive cache: write through
Jun 9 17:52:46 d2 kernel: sda: sda4
Jun 9 17:52:46 d2 kernel: Attached scsi removable disk sda at scsi1, channel 0, id 0, lun 0
Jun 9 17:52:46 d2 kernel: Attached scsi generic sg1 at scsi1, channel 0, id 0, lun 0, type 0
Patch:
--- 266.0/drivers/scsi/sd.c 2004-06-09 15:26:49.000000000 -0400
+++ 266.1/drivers/scsi/sd.c 2004-06-09 16:37:21.000000000 -0400
@@ -847,8 +847,11 @@ sd_spinup_disk(struct scsi_disk *sdkp, c
* any media in it, don't bother with any of the rest of
* this crap.
*/
- if (media_not_present(sdkp, SRpnt))
+ if (media_not_present(sdkp, SRpnt)) {
+ if (spintime)
+ printk(KERN_NOTICE "%s: No media present\n", diskname);
return;
+ }
if ((driver_byte(the_result) & DRIVER_SENSE) == 0) {
/* no sense, TUR either succeeded or failed
@@ -880,7 +883,7 @@ sd_spinup_disk(struct scsi_disk *sdkp, c
} else if (SRpnt->sr_sense_buffer[2] == NOT_READY) {
unsigned long time1;
if (!spintime) {
- printk(KERN_NOTICE "%s: Spinning up disk...",
+ printk(KERN_NOTICE "%s: Spinning up disk...\n",
diskname);
cmd[0] = START_STOP;
cmd[1] = 1; /* Return immediately */
@@ -903,7 +906,6 @@ sd_spinup_disk(struct scsi_disk *sdkp, c
current->state = TASK_UNINTERRUPTIBLE;
time1 = schedule_timeout(time1);
} while(time1);
- printk(".");
} else {
/* we don't understand the sense code, so it's
* probably pointless to loop */
@@ -917,12 +919,10 @@ sd_spinup_disk(struct scsi_disk *sdkp, c
} while (spintime &&
time_after(spintime_value + 100 * HZ, jiffies));
- if (spintime) {
- if (scsi_status_is_good(the_result))
- printk("ready\n");
- else
- printk("not responding...\n");
- }
+ if (spintime)
+ printk(KERN_NOTICE "%s: Drive %s\n", diskname,
+ scsi_status_is_good(the_result) ? "ready" : "not responding!");
+
}
/*
^ permalink raw reply [flat|nested] 2+ messages in thread* Re: [Patch 2.6] sd.c spinup messages
2004-06-09 23:48 [Patch 2.6] sd.c spinup messages Chuck Ebbert
@ 2004-06-24 22:02 ` James Bottomley
0 siblings, 0 replies; 2+ messages in thread
From: James Bottomley @ 2004-06-24 22:02 UTC (permalink / raw)
To: Chuck Ebbert; +Cc: linux-scsi
On Wed, 2004-06-09 at 18:48, Chuck Ebbert wrote:
> - printk(KERN_NOTICE "%s: Spinning up disk...",
> + printk(KERN_NOTICE "%s: Spinning up disk...\n",
> diskname);
> cmd[0] = START_STOP;
> cmd[1] = 1; /* Return immediately */
> @@ -903,7 +906,6 @@ sd_spinup_disk(struct scsi_disk *sdkp, c
> current->state = TASK_UNINTERRUPTIBLE;
> time1 = schedule_timeout(time1);
> } while(time1);
> - printk(".");
What this basically does is remove the indication that the drive is
being spun up. That's quite a significant user visible change from what
we do now.
Current behaviour is dictated by the fact that some of the drives
(particularly older ones) take ages to spin up. Without the visual
indication that something is going on people tend to press the reset
button and report kernel errors.
Instead of making such a global change, how about making this a
non-default configurable option within sd.
James
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2004-06-24 22:02 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2004-06-09 23:48 [Patch 2.6] sd.c spinup messages Chuck Ebbert
2004-06-24 22:02 ` James Bottomley
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox