Linux kernel source tree
Go to file
Qu Wenruo 5f50fa918f btrfs: do not use replace target device as an extra mirror
[BUG]
Currently btrfs can use dev-replace device as an extra mirror for
read-repair.  But it can lead to NODATASUM corruption in the following
case:

 There is a RAID1 data chunk, and dev-replace is running from
 dev2 to dev0.

 |//| = Replaced data
          X       X+1MB     X+2MB
  Dev 2:  |       |         |           <- Source dev
  Dev 0:  |///////|         |           <- Target dev

Then a read on dev 2 X+2MB happens.
And something wrong happened inside devid 2, causing an -EIO.

In that case, read-repair would try the next mirror, and since we can
use target device as an extra mirror, we will use that mirror instead.

But unfortunately since the read is beyond the current replace cursor,
we should not trust it at all, what we get would be just uninitialized
garbage.

But if this read is for NODATASUM range, then we just trust them and
cause data corruption.

[CAUSE]
We used to have some checks to make sure we only return such extra
mirror when the range is before our left cursor.

The first commit introducing this behavior is ad6d620e2a ("Btrfs:
allow repair code to include target disk when searching mirrors").

But later a fix, 22ab04e814 ("Btrfs: fix race between device replace
and chunk allocation") changed the behavior, to always let
btrfs_map_block() include the extra mirror to address a race in
dev-replace which can cause missing writes to target device.

This means, we lose the tracking of cursor for the extra mirror, thus
can lead to above corruption.

[FIX]
The extra mirror is never a reliable one, at the beginning of
dev-replace, the reliability is zero, while only at the end of the
replace it's a fully reliable mirror.

We either do the complex tracking, or never trust it.

IMHO it's much easier to maintain if we don't trust it at all, and the
extra mirror can only benefit for a limited period of time (during
replace).

Thus this patch would completely remove the ability to use target device
as an extra mirror.

Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
2023-04-17 18:01:16 +02:00
arch - Drop __init annotation from two rtc functions which get called after 2023-04-16 10:28:29 -07:00
block block: don't set GD_NEED_PART_SCAN if scan partition failed 2023-04-06 20:41:53 -06:00
certs Kbuild updates for v6.3 2023-02-26 11:53:25 -08:00
crypto asymmetric_keys: log on fatal failures in PE/pkcs7 2023-03-21 16:23:56 +00:00
Documentation Kbuild fixes for v6.3 (3rd) 2023-04-16 09:46:32 -07:00
drivers This pull request contains the following bug fixes for UBI and UBIFS: 2023-04-15 16:55:09 -07:00
fs btrfs: do not use replace target device as an extra mirror 2023-04-17 18:01:16 +02:00
include btrfs: replace btrfs_io_context::raid_map with a fixed u64 value 2023-04-17 18:01:14 +02:00
init init/initramfs: Fix argument forwarding to panic() in panic_show_mem() 2023-04-16 17:37:01 +09:00
io_uring io_uring-6.3-2023-04-14 2023-04-15 10:29:53 -07:00
ipc Merge branch 'work.namespace' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2023-02-24 19:20:07 -08:00
kernel - Do not pull tasks to the local scheduling group if its average load is 2023-04-16 10:33:43 -07:00
lib maple_tree: fix a potential concurrency bug in RCU mode 2023-04-05 18:06:25 -07:00
LICENSES LICENSES: Add the copyleft-next-0.3.1 license 2022-11-08 15:44:01 +01:00
mm mm/swap: fix swap_info_struct race between swapoff and get_swap_pages() 2023-04-05 18:06:24 -07:00
net Including fixes from bpf, and bluetooth. 2023-04-13 15:33:04 -07:00
rust Rust fixes for 6.3-rc1 2023-03-03 14:51:15 -08:00
samples LoongArch changes for v6.3 2023-03-01 09:27:00 -08:00
scripts kbuild: do not create intermediate *.tar for tar packages 2023-04-16 17:38:41 +09:00
security keys: Do not cache key in task struct if key is requested from kernel thread 2023-03-21 16:22:40 +00:00
sound ALSA: hda/hdmi: disable KAE for Intel DG2 2023-04-14 07:50:52 +02:00
tools Including fixes from bpf, and bluetooth. 2023-04-13 15:33:04 -07:00
usr initramfs: Check negative timestamp to prevent broken cpio archive 2023-04-16 17:37:01 +09:00
virt KVM: PPC: Make KVM_CAP_IRQFD_RESAMPLE platform dependent 2023-03-31 11:19:05 -04:00
.clang-format cpumask: re-introduce constant-sized cpumask optimizations 2023-03-05 14:30:34 -08:00
.cocciconfig
.get_maintainer.ignore get_maintainer: add Alan to .get_maintainer.ignore 2022-08-20 15:17:44 -07:00
.gitattributes .gitattributes: use 'dts' diff driver for *.dtso files 2023-02-26 15:28:23 +09:00
.gitignore kbuild: rpm-pkg: move source components to rpmbuild/SOURCES 2023-03-16 22:45:56 +09:00
.mailmap mailmap: add an entry for Leonard Crestez 2023-03-28 15:24:32 -07:00
.rustfmt.toml rust: add .rustfmt.toml 2022-09-28 09:02:20 +02:00
COPYING COPYING: state that all contributions really are covered by this file 2020-02-10 13:32:20 -08:00
CREDITS There is no particular theme here - mainly quick hits all over the tree. 2023-02-23 17:55:40 -08:00
Kbuild Kbuild updates for v6.1 2022-10-10 12:00:45 -07:00
Kconfig kbuild: ensure full rebuild when the compiler is updated 2020-05-12 13:28:33 +09:00
MAINTAINERS ata changes for 6.3-rc7 2023-04-11 09:51:45 -07:00
Makefile Linux 6.3-rc7 2023-04-16 15:23:53 -07:00
README Drop all 00-INDEX files from Documentation/ 2018-09-09 15:08:58 -06:00

Linux kernel
============

There are several guides for kernel developers and users. These guides can
be rendered in a number of formats, like HTML and PDF. Please read
Documentation/admin-guide/README.rst first.

In order to build the documentation, use ``make htmldocs`` or
``make pdfdocs``.  The formatted documentation can also be read online at:

    https://www.kernel.org/doc/html/latest/

There are various text files in the Documentation/ subdirectory,
several of them using the Restructured Text markup notation.

Please read the Documentation/process/changes.rst file, as it contains the
requirements for building and running the kernel, and information about
the problems which may result by upgrading your kernel.