Closed Bug 1148419 Opened 9 years ago Closed 9 years ago

[Shinano][Aries] Fix recovery partitions fstab

Categories

(Firefox OS Graveyard :: GonkIntegration, defect)

ARM
Gonk (Firefox OS)
defect
Not set
normal

Tracking

(Not tracked)

RESOLVED FIXED
2.2 S9 (3apr)

People

(Reporter: gerard-majax, Assigned: gerard-majax)

References

Details

(Whiteboard: [systemsfe])

Attachments

(1 file)

We are not having the proper fstab. Upstream defines: https://raw.githubusercontent.com/sonyxperiadev/device-sony-shinano/master/rootdir/fstab.shinano

> /dev/block/platform/msm_sdcc.1/by-name/system       /system      ext4    ro,barrier=1,discard                             wait
> /dev/block/platform/msm_sdcc.1/by-name/userdata     /data        ext4    nosuid,nodev,barrier=1,noauto_da_alloc,discard   wait,check,encryptable=footer
> /dev/block/platform/msm_sdcc.1/by-name/cache        /cache       ext4    nosuid,nodev,discard                             wait,check
> /dev/block/platform/msm_sdcc.1/by-name/boot         /boot        emmc    defaults                                         defaults
> /dev/block/platform/msm_sdcc.1/by-name/FOTAKernel   /recovery    emmc    defaults                                         defaults
> 
> /devices/msm_sdcc.2/mmc_host                        auto         auto    nosuid,nodev                                     voldmanaged=sdcard1:auto
> /devices/platform/xhci-hcd                          auto         auto    nosuid,nodev                                     voldmanaged=usbdisk:auto

The major differences are:
 - recovery partition label which is "FOTAKernel" and not "recovery", useful for bug 1136638
 - sdcard devices and their filesystem
Whiteboard: [systemsfe]
Attached file Shinano common PR
"auto" fstype cannot be used :(
Attachment #8585017 - Flags: review?(mwu)
Attachment #8585017 - Flags: review?(mwu) → review?(dhylands)
I put some comments on the PR which I'll repeat here:

Why remove noatime? I also noticed that the regular (non-recovery) mounts for the z3c are using relatime rather than noatime, but conventional wisdom seems to suggest that using noatime is better for performance and reduces flash writes.

And why remove barrier=1 on /cache? barrier=1 seems to be important when using a journalling filesystem (like ext4)

data=ordered is the default if not specified, so that seems ok.
Actually, looking at it some more relatime seems reasonable (not as expensive as full atime)
Well, I share your concern regarding /cache, but I just got inspired by the recovery fstab provided by Sony.

Alin, any reason barrier=1 is not set on the /cache partition of your recovery fstab?
Flags: needinfo?(alin.jerpelea)
Just set relatime instead of noatime.
I checked the kernel documentation, and it states that barrier is a default value for ext4, but it was not for ext3, https://www.kernel.org/doc/Documentation/filesystems/ext4.txt

Dave, do you want that we specify it anyway?
Flags: needinfo?(alin.jerpelea) → needinfo?(dhylands)
I'm happy to see it removed, given that it's the default, but then I would have expected it to be removed everywhere for consistency.

It seems relatime is also the default (without being specified).
Flags: needinfo?(dhylands)
Attachment #8585017 - Flags: review?(dhylands) → review+
(In reply to Dave Hylands [:dhylands] from comment #8)
> I'm happy to see it removed, given that it's the default, but then I would
> have expected it to be removed everywhere for consistency.
> 
> It seems relatime is also the default (without being specified).

Removed from everywhere
https://github.com/mozilla-b2g/device-shinano-common/commit/5acc9c51737fd08ac7b923751bb7dd3319eb1d00
Status: NEW → RESOLVED
Closed: 9 years ago
Resolution: --- → FIXED
Blocks: 1136638
Target Milestone: --- → 2.2 S9 (3apr)
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: