From: Teresa Remmet <t.remmet@phytec.de>
To: barebox@lists.infradead.org
Subject: [PATCH 07/56] UBI: Fix static volume checks when Fastmap is used
Date: Wed, 29 Aug 2018 14:19:23 +0200 [thread overview]
Message-ID: <1535545212-18871-8-git-send-email-t.remmet@phytec.de> (raw)
In-Reply-To: <1535545212-18871-1-git-send-email-t.remmet@phytec.de>
From: Richard Weinberger <richard@nod.at>
Ezequiel reported that he's facing UBI going into read-only
mode after power cut. It turned out that this behavior happens
only when updating a static volume is interrupted and Fastmap is
used.
A possible trace can look like:
ubi0 warning: ubi_io_read_vid_hdr [ubi]: no VID header found at PEB 2323, only 0xFF bytes
ubi0 warning: ubi_eba_read_leb [ubi]: switch to read-only mode
CPU: 0 PID: 833 Comm: ubiupdatevol Not tainted 4.6.0-rc2-ARCH #4
Hardware name: SAMSUNG ELECTRONICS CO., LTD. 300E4C/300E5C/300E7C/NP300E5C-AD8AR, BIOS P04RAP 10/15/2012
0000000000000286 00000000eba949bd ffff8800c45a7b38 ffffffff8140d841
ffff8801964be000 ffff88018eaa4800 ffff8800c45a7bb8 ffffffffa003abf6
ffffffff850e2ac0 8000000000000163 ffff8801850e2ac0 ffff8801850e2ac0
Call Trace:
[<ffffffff8140d841>] dump_stack+0x63/0x82
[<ffffffffa003abf6>] ubi_eba_read_leb+0x486/0x4a0 [ubi]
[<ffffffffa00453b3>] ubi_check_volume+0x83/0xf0 [ubi]
[<ffffffffa0039d97>] ubi_open_volume+0x177/0x350 [ubi]
[<ffffffffa00375d8>] vol_cdev_open+0x58/0xb0 [ubi]
[<ffffffff8124b08e>] chrdev_open+0xae/0x1d0
[<ffffffff81243bcf>] do_dentry_open+0x1ff/0x300
[<ffffffff8124afe0>] ? cdev_put+0x30/0x30
[<ffffffff81244d36>] vfs_open+0x56/0x60
[<ffffffff812545f4>] path_openat+0x4f4/0x1190
[<ffffffff81256621>] do_filp_open+0x91/0x100
[<ffffffff81263547>] ? __alloc_fd+0xc7/0x190
[<ffffffff812450df>] do_sys_open+0x13f/0x210
[<ffffffff812451ce>] SyS_open+0x1e/0x20
[<ffffffff81a99e32>] entry_SYSCALL_64_fastpath+0x1a/0xa4
UBI checks static volumes for data consistency and reads the
whole volume upon first open. If the volume is found erroneous
users of UBI cannot read from it, but another volume update is
possible to fix it. The check is performed by running
ubi_eba_read_leb() on every allocated LEB of the volume.
For static volumes ubi_eba_read_leb() computes the checksum of all
data stored in a LEB. To verify the computed checksum it has to read
the LEB's volume header which stores the original checksum.
If the volume header is not found UBI treats this as fatal internal
error and switches to RO mode. If the UBI device was attached via a
full scan the assumption is correct, the volume header has to be
present as it had to be there while scanning to get known as mapped.
If the attach operation happened via Fastmap the assumption is no
longer correct. When attaching via Fastmap UBI learns the mapping
table from Fastmap's snapshot of the system state and not via a full
scan. It can happen that a LEB got unmapped after a Fastmap was
written to the flash. Then UBI can learn the LEB still as mapped and
accessing it returns only 0xFF bytes. As UBI is not a FTL it is
allowed to have mappings to empty PEBs, it assumes that the layer
above takes care of LEB accounting and referencing.
UBIFS does so using the LEB property tree (LPT).
For static volumes UBI blindly assumes that all LEBs are present and
therefore special actions have to be taken.
The described situation can happen when updating a static volume is
interrupted, either by a user or a power cut.
The volume update code first unmaps all LEBs of a volume and then
writes LEB by LEB. If the sequence of operations is interrupted UBI
detects this either by the absence of LEBs, no volume header present
at scan time, or corrupted payload, detected via checksum.
In the Fastmap case the former method won't trigger as no scan
happened and UBI automatically thinks all LEBs are present.
Only by reading data from a LEB it detects that the volume header is
missing and incorrectly treats this as fatal error.
To deal with the situation ubi_eba_read_leb() from now on checks
whether we attached via Fastmap and handles the absence of a
volume header like a data corruption error.
This way interrupted static volume updates will correctly get detected
also when Fastmap is used.
Cc: <stable@vger.kernel.org>
Reported-by: Ezequiel Garcia <ezequiel@vanguardiasur.com.ar>
Tested-by: Ezequiel Garcia <ezequiel@vanguardiasur.com.ar>
Signed-off-by: Richard Weinberger <richard@nod.at>
[Fixed conflict]
Signed-off-by: Teresa Remmet <t.remmet@phytec.de>
---
drivers/mtd/ubi/eba.c | 21 +++++++++++++++++++--
drivers/mtd/ubi/fastmap.c | 1 +
drivers/mtd/ubi/ubi.h | 2 ++
3 files changed, 22 insertions(+), 2 deletions(-)
diff --git a/drivers/mtd/ubi/eba.c b/drivers/mtd/ubi/eba.c
index c242c460e849..6e09cfb0a119 100644
--- a/drivers/mtd/ubi/eba.c
+++ b/drivers/mtd/ubi/eba.c
@@ -392,8 +392,25 @@ retry:
pnum, vol_id, lnum);
err = -EBADMSG;
} else {
- err = -EINVAL;
- ubi_ro_mode(ubi);
+ /*
+ * Ending up here in the non-Fastmap case
+ * is a clear bug as the VID header had to
+ * be present at scan time to have it referenced.
+ * With fastmap the story is more complicated.
+ * Fastmap has the mapping info without the need
+ * of a full scan. So the LEB could have been
+ * unmapped, Fastmap cannot know this and keeps
+ * the LEB referenced.
+ * This is valid and works as the layer above UBI
+ * has to do bookkeeping about used/referenced
+ * LEBs in any case.
+ */
+ if (ubi->fast_attach) {
+ err = -EBADMSG;
+ } else {
+ err = -EINVAL;
+ ubi_ro_mode(ubi);
+ }
}
}
goto out_free;
diff --git a/drivers/mtd/ubi/fastmap.c b/drivers/mtd/ubi/fastmap.c
index 2e0fdb7583e4..5bbbd7c6ae2a 100644
--- a/drivers/mtd/ubi/fastmap.c
+++ b/drivers/mtd/ubi/fastmap.c
@@ -991,6 +991,7 @@ int ubi_scan_fastmap(struct ubi_device *ubi, struct ubi_attach_info *ai,
ubi_msg(ubi, "fastmap WL pool size: %d",
ubi->fm_wl_pool.max_size);
ubi->fm_disabled = 0;
+ ubi->fast_attach = 1;
ubi_free_vid_hdr(ubi, vh);
kfree(ech);
diff --git a/drivers/mtd/ubi/ubi.h b/drivers/mtd/ubi/ubi.h
index 1f0ad386b45e..368b829fa17f 100644
--- a/drivers/mtd/ubi/ubi.h
+++ b/drivers/mtd/ubi/ubi.h
@@ -416,6 +416,7 @@ struct ubi_debug_info {
* @fm_size: fastmap size in bytes
* @fm_sem: allows ubi_update_fastmap() to block EBA table changes
* @fm_work: fastmap work queue
+ * @fast_attach: non-zero if UBI was attached by fastmap
*
* @used: RB-tree of used physical eraseblocks
* @erroneous: RB-tree of erroneous used physical eraseblocks
@@ -515,6 +516,7 @@ struct ubi_device {
struct ubi_fm_pool fm_wl_pool;
void *fm_buf;
size_t fm_size;
+ int fast_attach;
/* Wear-leveling sub-system's stuff */
struct rb_root used;
--
2.7.4
_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox
next prev parent reply other threads:[~2018-08-29 12:20 UTC|newest]
Thread overview: 58+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-08-29 12:19 [PATCH 00/56] Update Barebox UBI Teresa Remmet
2018-08-29 12:19 ` [PATCH 01/56] UBI: Fastmap: Fix memory leak while attaching Teresa Remmet
2018-08-29 12:19 ` [PATCH 02/56] UBI: Remove alloc_ai() slab name from parameter list Teresa Remmet
2018-08-29 12:19 ` [PATCH 03/56] UBI: Don't read back all data in ubi_eba_copy_leb() Teresa Remmet
2018-08-29 12:19 ` [PATCH 04/56] UBI: Modify wrong comment in ubi_leb_map function Teresa Remmet
2018-08-29 12:19 ` [PATCH 05/56] UBI: Clean up return in ubi_remove_volume() Teresa Remmet
2018-08-29 12:19 ` [PATCH 06/56] UBI: Set free_count to zero before walking through erase list Teresa Remmet
2018-08-29 12:19 ` Teresa Remmet [this message]
2018-08-29 12:19 ` [PATCH 08/56] ubi: Make recover_peb power cut aware Teresa Remmet
2018-08-29 12:19 ` [PATCH 09/56] ubi: Fix race condition between ubi device creation and udev Teresa Remmet
2018-08-29 12:19 ` [PATCH 10/56] ubi: Fix early logging Teresa Remmet
2018-08-29 12:19 ` [PATCH 11/56] ubi: Make volume resize power cut aware Teresa Remmet
2018-08-29 12:19 ` [PATCH 12/56] ubi: Fix scan_fast() comment Teresa Remmet
2018-08-29 12:19 ` [PATCH 13/56] ubi: Introduce vol_ignored() Teresa Remmet
2018-08-29 12:19 ` [PATCH 14/56] ubi: Fix whitespace issue in count_fastmap_pebs() Teresa Remmet
2018-08-29 12:19 ` [PATCH 15/56] ubi: Rework Fastmap attach base code Teresa Remmet
2018-08-29 12:19 ` [PATCH 16/56] ubi: Check whether the Fastmap anchor matches the super block Teresa Remmet
2018-08-29 12:19 ` [PATCH 17/56] ubi: Be more paranoid while seaching for the most recent Fastmap Teresa Remmet
2018-08-29 12:19 ` [PATCH 18/56] UBI: fastmap: use ubi_find_volume() instead of open coding it Teresa Remmet
2018-08-29 12:19 ` [PATCH 19/56] UBI: fix add_fastmap() to use the vid_hdr passed in argument Teresa Remmet
2018-08-29 12:19 ` [PATCH 20/56] UBI: fastmap: avoid multiple be32_to_cpu() when unneccesary Teresa Remmet
2018-08-29 12:19 ` [PATCH 21/56] UBI: fastmap: scrub PEB when bitflips are detected in a free PEB EC header Teresa Remmet
2018-08-29 12:19 ` [PATCH 22/56] UBI: factorize code used to manipulate volumes at attach time Teresa Remmet
2018-08-29 12:19 ` [PATCH 23/56] UBI: factorize destroy_av() and ubi_remove_av() code Teresa Remmet
2018-08-29 12:19 ` [PATCH 24/56] UBI: fastmap: use ubi_rb_for_each_entry() in unmap_peb() Teresa Remmet
2018-08-29 12:19 ` [PATCH 25/56] UBI: fastmap: use ubi_io_{read, write}_data() instead of ubi_io_{read, write}() Teresa Remmet
2018-08-29 12:19 ` [PATCH 26/56] UBI: provide helpers to allocate and free aeb elements Teresa Remmet
2018-08-29 12:19 ` [PATCH 27/56] UBI: move the global ech and vidh variables into struct ubi_attach_info Teresa Remmet
2018-08-29 12:19 ` [PATCH 28/56] UBI: simplify recover_peb() code Teresa Remmet
2018-08-29 12:19 ` [PATCH 29/56] UBI: simplify LEB write and atomic LEB change code Teresa Remmet
2018-08-29 12:19 ` [PATCH 30/56] UBI: add an helper to check lnum validity Teresa Remmet
2018-08-29 12:19 ` [PATCH 31/56] UBI: provide an helper to check whether a LEB is mapped or not Teresa Remmet
2018-08-29 12:19 ` [PATCH 32/56] UBI: provide an helper to query LEB information Teresa Remmet
2018-08-29 12:19 ` [PATCH 33/56] UBI: hide EBA internals Teresa Remmet
2018-08-29 12:19 ` [PATCH 34/56] UBI: introduce the VID buffer concept Teresa Remmet
2018-08-29 12:19 ` [PATCH 35/56] ubi: Deal with interrupted erasures in WL Teresa Remmet
2018-08-29 12:19 ` [PATCH 36/56] ubi: Fix races around ubi_refill_pools() Teresa Remmet
2018-08-29 12:19 ` [PATCH 37/56] ubi: Fix Fastmap's update_vol() Teresa Remmet
2018-08-29 12:19 ` [PATCH 38/56] ubi: fix swapped arguments to call to ubi_alloc_aeb Teresa Remmet
2018-08-29 12:19 ` [PATCH 39/56] UBI: Fix crash in try_recover_peb() Teresa Remmet
2018-08-29 12:19 ` [PATCH 40/56] ubi: fastmap: Fix add_vol() return value test in ubi_attach_fastmap() Teresa Remmet
2018-08-29 12:19 ` [PATCH 41/56] UBI: Fix typos Teresa Remmet
2018-08-29 12:19 ` [PATCH 42/56] ubi/upd: Always flush after prepared for an update Teresa Remmet
2018-08-29 12:19 ` [PATCH 43/56] ubi: fastmap: Fix slab corruption Teresa Remmet
2018-08-29 12:20 ` [PATCH 44/56] ubi: pr_err() strings should end with newlines Teresa Remmet
2018-08-29 12:20 ` [PATCH 45/56] ubi: fastmap: fix spelling mistake: "invalidiate" -> "invalidate" Teresa Remmet
2018-08-29 12:20 ` [PATCH 46/56] UBI: Fix two typos in comments Teresa Remmet
2018-08-29 12:20 ` [PATCH 47/56] ubi: fastmap: Clean up the initialization of pointer p Teresa Remmet
2018-08-29 12:20 ` [PATCH 48/56] ubi: fastmap: Erase outdated anchor PEBs during attach Teresa Remmet
2018-08-29 12:20 ` [PATCH 49/56] ubi: Fastmap: Fix typo Teresa Remmet
2018-08-29 12:20 ` [PATCH 50/56] ubi: Fix copy/paste error in function documentation Teresa Remmet
2018-08-29 12:20 ` [PATCH 51/56] mtd: ubi: wl: Fix error return code in ubi_wl_init() Teresa Remmet
2018-08-29 12:20 ` [PATCH 52/56] ubi: fastmap: Correctly handle interrupted erasures in EBA Teresa Remmet
2018-08-29 12:20 ` [PATCH 53/56] ubi: fastmap: Check each mapping only once Teresa Remmet
2018-08-29 12:20 ` [PATCH 54/56] ubi: fastmap: Detect EBA mismatches on-the-fly Teresa Remmet
2018-08-29 12:20 ` [PATCH 55/56] mtd: ubi: Update ubi-media.h to dual license Teresa Remmet
2018-08-29 12:20 ` [PATCH 56/56] ubi: Initialize Fastmap checkmapping correctly Teresa Remmet
2018-08-31 6:25 ` [PATCH 00/56] Update Barebox UBI Sascha Hauer
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1535545212-18871-8-git-send-email-t.remmet@phytec.de \
--to=t.remmet@phytec.de \
--cc=barebox@lists.infradead.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox