From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: linux-kernel@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
stable@vger.kernel.org, Chao Yu <chao@kernel.org>,
Gao Xiang <hsiangkao@linux.alibaba.com>
Subject: [PATCH 5.10 18/21] erofs: fix unsafe pagevec reuse of hooked pclusters
Date: Fri, 19 Nov 2021 18:37:53 +0100 [thread overview]
Message-ID: <20211119171444.470625833@linuxfoundation.org> (raw)
In-Reply-To: <20211119171443.892729043@linuxfoundation.org>
From: Gao Xiang <hsiangkao@linux.alibaba.com>
commit 86432a6dca9bed79111990851df5756d3eb5f57c upstream.
There are pclusters in runtime marked with Z_EROFS_PCLUSTER_TAIL
before actual I/O submission. Thus, the decompression chain can be
extended if the following pcluster chain hooks such tail pcluster.
As the related comment mentioned, if some page is made of a hooked
pcluster and another followed pcluster, it can be reused for in-place
I/O (since I/O should be submitted anyway):
_______________________________________________________________
| tail (partial) page | head (partial) page |
|_____PRIMARY_HOOKED___|____________PRIMARY_FOLLOWED____________|
However, it's by no means safe to reuse as pagevec since if such
PRIMARY_HOOKED pclusters finally move into bypass chain without I/O
submission. It's somewhat hard to reproduce with LZ4 and I just found
it (general protection fault) by ro_fsstressing a LZMA image for long
time.
I'm going to actively clean up related code together with multi-page
folio adaption in the next few months. Let's address it directly for
easier backporting for now.
Call trace for reference:
z_erofs_decompress_pcluster+0x10a/0x8a0 [erofs]
z_erofs_decompress_queue.isra.36+0x3c/0x60 [erofs]
z_erofs_runqueue+0x5f3/0x840 [erofs]
z_erofs_readahead+0x1e8/0x320 [erofs]
read_pages+0x91/0x270
page_cache_ra_unbounded+0x18b/0x240
filemap_get_pages+0x10a/0x5f0
filemap_read+0xa9/0x330
new_sync_read+0x11b/0x1a0
vfs_read+0xf1/0x190
Link: https://lore.kernel.org/r/20211103182006.4040-1-xiang@kernel.org
Fixes: 3883a79abd02 ("staging: erofs: introduce VLE decompression support")
Cc: <stable@vger.kernel.org> # 4.19+
Reviewed-by: Chao Yu <chao@kernel.org>
Signed-off-by: Gao Xiang <hsiangkao@linux.alibaba.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
fs/erofs/zdata.c | 13 +++++++------
fs/erofs/zpvec.h | 13 ++++++++++---
2 files changed, 17 insertions(+), 9 deletions(-)
--- a/fs/erofs/zdata.c
+++ b/fs/erofs/zdata.c
@@ -278,8 +278,8 @@ static inline bool z_erofs_try_inplace_i
/* callers must be with collection lock held */
static int z_erofs_attach_page(struct z_erofs_collector *clt,
- struct page *page,
- enum z_erofs_page_type type)
+ struct page *page, enum z_erofs_page_type type,
+ bool pvec_safereuse)
{
int ret;
@@ -289,9 +289,9 @@ static int z_erofs_attach_page(struct z_
z_erofs_try_inplace_io(clt, page))
return 0;
- ret = z_erofs_pagevec_enqueue(&clt->vector, page, type);
+ ret = z_erofs_pagevec_enqueue(&clt->vector, page, type,
+ pvec_safereuse);
clt->cl->vcnt += (unsigned int)ret;
-
return ret ? 0 : -EAGAIN;
}
@@ -645,7 +645,8 @@ hitted:
tight &= (clt->mode >= COLLECT_PRIMARY_FOLLOWED);
retry:
- err = z_erofs_attach_page(clt, page, page_type);
+ err = z_erofs_attach_page(clt, page, page_type,
+ clt->mode >= COLLECT_PRIMARY_FOLLOWED);
/* should allocate an additional staging page for pagevec */
if (err == -EAGAIN) {
struct page *const newpage =
@@ -653,7 +654,7 @@ retry:
newpage->mapping = Z_EROFS_MAPPING_STAGING;
err = z_erofs_attach_page(clt, newpage,
- Z_EROFS_PAGE_TYPE_EXCLUSIVE);
+ Z_EROFS_PAGE_TYPE_EXCLUSIVE, true);
if (!err)
goto retry;
}
--- a/fs/erofs/zpvec.h
+++ b/fs/erofs/zpvec.h
@@ -107,11 +107,18 @@ static inline void z_erofs_pagevec_ctor_
static inline bool z_erofs_pagevec_enqueue(struct z_erofs_pagevec_ctor *ctor,
struct page *page,
- enum z_erofs_page_type type)
+ enum z_erofs_page_type type,
+ bool pvec_safereuse)
{
- if (!ctor->next && type)
- if (ctor->index + 1 == ctor->nr)
+ if (!ctor->next) {
+ /* some pages cannot be reused as pvec safely without I/O */
+ if (type == Z_EROFS_PAGE_TYPE_EXCLUSIVE && !pvec_safereuse)
+ type = Z_EROFS_VLE_PAGE_TYPE_TAIL_SHARED;
+
+ if (type != Z_EROFS_PAGE_TYPE_EXCLUSIVE &&
+ ctor->index + 1 == ctor->nr)
return false;
+ }
if (ctor->index >= ctor->nr)
z_erofs_pagevec_ctor_pagedown(ctor, false);
next prev parent reply other threads:[~2021-11-19 17:38 UTC|newest]
Thread overview: 39+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-11-19 17:37 [PATCH 5.10 00/21] 5.10.81-rc1 review Greg Kroah-Hartman
2021-11-19 17:37 ` [PATCH 5.10 01/21] fortify: Explicitly disable Clang support Greg Kroah-Hartman
2021-11-19 17:37 ` [PATCH 5.10 02/21] block: Add a helper to validate the block size Greg Kroah-Hartman
2021-11-19 17:37 ` [PATCH 5.10 03/21] loop: Use blk_validate_block_size() to validate " Greg Kroah-Hartman
2021-11-19 21:45 ` Pavel Machek
2021-11-20 7:39 ` Greg Kroah-Hartman
2021-11-21 16:59 ` Pavel Machek
2021-11-19 17:37 ` [PATCH 5.10 04/21] bootconfig: init: Fix memblock leak in xbc_make_cmdline() Greg Kroah-Hartman
2021-11-19 17:37 ` [PATCH 5.10 05/21] net: stmmac: add clocks management for gmac driver Greg Kroah-Hartman
2021-11-19 17:37 ` [PATCH 5.10 06/21] net: stmmac: platform: fix build error with !CONFIG_PM_SLEEP Greg Kroah-Hartman
2021-11-19 17:37 ` [PATCH 5.10 07/21] net: stmmac: fix missing unlock on error in stmmac_suspend() Greg Kroah-Hartman
2021-11-19 17:37 ` [PATCH 5.10 08/21] net: stmmac: fix system hang if change mac address after interface ifdown Greg Kroah-Hartman
2021-11-19 17:37 ` [PATCH 5.10 09/21] net: stmmac: fix issue where clk is being unprepared twice Greg Kroah-Hartman
2021-11-19 17:37 ` [PATCH 5.10 10/21] net: stmmac: dwmac-rk: fix unbalanced pm_runtime_enable warnings Greg Kroah-Hartman
2021-11-19 17:37 ` [PATCH 5.10 11/21] arm64: vdso32: suppress error message for make mrproper Greg Kroah-Hartman
2021-11-19 21:47 ` Pavel Machek
2021-11-19 22:24 ` Nick Desaulniers
2021-11-19 22:30 ` Nick Desaulniers
2021-11-21 7:48 ` Greg Kroah-Hartman
2021-11-19 17:37 ` [PATCH 5.10 12/21] x86/iopl: Fake iopl(3) CLI/STI usage Greg Kroah-Hartman
2021-11-19 17:37 ` [PATCH 5.10 13/21] parisc/entry: fix trace test in syscall exit path Greg Kroah-Hartman
2021-11-19 17:37 ` [PATCH 5.10 14/21] PCI/MSI: Destroy sysfs before freeing entries Greg Kroah-Hartman
2021-11-19 17:37 ` [PATCH 5.10 15/21] PCI/MSI: Deal with devices lying about their MSI mask capability Greg Kroah-Hartman
2021-11-19 17:37 ` [PATCH 5.10 16/21] PCI: Add MSI masking quirk for Nvidia ION AHCI Greg Kroah-Hartman
2021-11-19 17:37 ` [PATCH 5.10 17/21] erofs: remove the occupied parameter from z_erofs_pagevec_enqueue() Greg Kroah-Hartman
2021-11-19 17:37 ` Greg Kroah-Hartman [this message]
2021-11-19 17:37 ` [PATCH 5.10 19/21] scripts/lld-version.sh: Rewrite based on upstream ld-version.sh Greg Kroah-Hartman
2021-11-19 17:37 ` [PATCH 5.10 20/21] perf/core: Avoid put_page() when GUP fails Greg Kroah-Hartman
2021-11-19 17:37 ` [PATCH 5.10 21/21] thermal: Fix NULL pointer dereferences in of_thermal_ functions Greg Kroah-Hartman
2021-11-19 18:46 ` [PATCH 5.10 00/21] 5.10.81-rc1 review Florian Fainelli
2021-11-19 19:46 ` Fox Chen
2021-11-20 0:49 ` Shuah Khan
2021-11-20 7:28 ` Pavel Machek
2021-11-20 15:21 ` Sudip Mukherjee
2021-11-20 15:27 ` Daniel Díaz
2021-11-20 16:06 ` Naresh Kamboju
2021-11-21 7:49 ` Greg Kroah-Hartman
2021-11-20 16:52 ` Guenter Roeck
2021-11-22 2:48 ` Samuel Zou
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=20211119171444.470625833@linuxfoundation.org \
--to=gregkh@linuxfoundation.org \
--cc=chao@kernel.org \
--cc=hsiangkao@linux.alibaba.com \
--cc=linux-kernel@vger.kernel.org \
--cc=stable@vger.kernel.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;
as well as URLs for NNTP newsgroup(s).