August 2020 - Linux-stable-mirror

[PATCH 1/2] drm/nouveau/kms/nv50-: Program notifier offset before requesting disp caps

by Lyude Paul

Not entirely sure why this never came up when I originally tested this (maybe some BIOSes already have this setup?) but the ->caps_init vfunc appears to cause the display engine to throw an exception on driver init, at least on my ThinkPad P72: nouveau 0000:01:00.0: disp: chid 0 mthd 008c data 00000000 0000508c 0000102b This is magic nvidia speak for "You need to have the DMA notifier offset programmed before you can call NV507D_GET_CAPABILITIES." So, let's fix this by doing that, and also perform an update afterwards to prevent racing with the GPU when reading capabilities. Changes since v1: * Don't just program the DMA notifier offset, make sure to actually perform an update Signed-off-by: Lyude Paul <lyude(a)redhat.com> Fixes: 4a2cb4181b07 ("drm/nouveau/kms/nv50-: Probe SOR and PIOR caps for DP interlacing support") Cc: <stable(a)vger.kernel.org> # v5.8+ --- drivers/gpu/drm/nouveau/dispnv50/core507d.c | 25 ++++++++++++++++----- 1 file changed, 19 insertions(+), 6 deletions(-) diff --git a/drivers/gpu/drm/nouveau/dispnv50/core507d.c b/drivers/gpu/drm/nouveau/dispnv50/core507d.c index e341f572c2696..5e86feec3b720 100644 --- a/drivers/gpu/drm/nouveau/dispnv50/core507d.c +++ b/drivers/gpu/drm/nouveau/dispnv50/core507d.c @@ -65,13 +65,26 @@ core507d_ntfy_init(struct nouveau_bo *bo, u32 offset) int core507d_caps_init(struct nouveau_drm *drm, struct nv50_disp *disp) { - u32 *push = evo_wait(&disp->core->chan, 2); + struct nv50_core *core = disp->core; + u32 interlock[NV50_DISP_INTERLOCK__SIZE] = {0}; + u32 *push; - if (push) { - evo_mthd(push, 0x008c, 1); - evo_data(push, 0x0); - evo_kick(push, &disp->core->chan); - } + core->func->ntfy_init(disp->sync, NV50_DISP_CORE_NTFY); + + push = evo_wait(&core->chan, 4); + if (!push) + return 0; + + evo_mthd(push, 0x0084, 1); + evo_data(push, 0x80000000 | NV50_DISP_CORE_NTFY); + evo_mthd(push, 0x008c, 1); + evo_data(push, 0x0); + evo_kick(push, &core->chan); + + core->func->update(core, interlock, false); + if (core->func->ntfy_wait_done(disp->sync, NV50_DISP_CORE_NTFY, + core->chan.base.device)) + NV_ERROR(drm, "core notifier timeout\n"); return 0; } -- 2.26.2

3 years, 8 months

3
8
0 0

[PATCH] scsi: target: iscsi: Fix data digest calculation

by Varun Prakash

Current code does not consider 'page_off' in data digest calculation, to fix this add a local variable 'first_sg' and set first_sg.offset to sg->offset + page_off. Fixes: e48354ce078c ("iscsi-target: Add iSCSI fabric support for target v4.1") Cc: <stable(a)vger.kernel.org> Signed-off-by: Varun Prakash <varun(a)chelsio.com> --- drivers/target/iscsi/iscsi_target.c | 17 +++++++++++++++-- 1 file changed, 15 insertions(+), 2 deletions(-) diff --git a/drivers/target/iscsi/iscsi_target.c b/drivers/target/iscsi/iscsi_target.c index c968961..2ec778e 100644 --- a/drivers/target/iscsi/iscsi_target.c +++ b/drivers/target/iscsi/iscsi_target.c @@ -1389,14 +1389,27 @@ static u32 iscsit_do_crypto_hash_sg( sg = cmd->first_data_sg; page_off = cmd->first_data_sg_off; + if (data_length && page_off) { + struct scatterlist first_sg; + u32 len = min_t(u32, data_length, sg->length - page_off); + + sg_init_table(&first_sg, 1); + sg_set_page(&first_sg, sg_page(sg), len, sg->offset + page_off); + + ahash_request_set_crypt(hash, &first_sg, NULL, len); + crypto_ahash_update(hash); + + data_length -= len; + sg = sg_next(sg); + } + while (data_length) { - u32 cur_len = min_t(u32, data_length, (sg->length - page_off)); + u32 cur_len = min_t(u32, data_length, sg->length); ahash_request_set_crypt(hash, sg, NULL, cur_len); crypto_ahash_update(hash); data_length -= cur_len; - page_off = 0; /* iscsit_map_iovec has already checked for invalid sg pointers */ sg = sg_next(sg); } -- 2.0.2

3 years, 8 months

3
2
0 0

[PATCH] iscsi-target: fix hang in iscsit_access_np() when getting tpg->np_login_sem

by Hou Pu

The iscsi target login thread might stuck in following stack: cat /proc/`pidof iscsi_np`/stack [<0>] down_interruptible+0x42/0x50 [<0>] iscsit_access_np+0xe3/0x167 [<0>] iscsi_target_locate_portal+0x695/0x8ac [<0>] __iscsi_target_login_thread+0x855/0xb82 [<0>] iscsi_target_login_thread+0x2f/0x5a [<0>] kthread+0xfa/0x130 [<0>] ret_from_fork+0x1f/0x30 This could be reproduced by following steps: 1. Initiator A try to login iqn1-tpg1 on port 3260. After finishing PDU exchange in the login thread and before the negotiation is finished, at this time the network link is down. In a production environment, this could happen. I could emulated it by bring the network card down in the initiator node by ifconfig eth0 down. (Now A could never finish this login. And tpg->np_login_sem is hold by it). 2. Initiator B try to login iqn2-tpg1 on port 3260. After finishing PDU exchange in the login thread. The target expect to process remaining login PDUs in workqueue context. 3. Initiator A' try to re-login to iqn1-tpg1 on port 3260 from a new socket. It will wait for tpg->np_login_sem with np->np_login_timer loaded to wait for at most 15 second. (Because the lock is held by A. A never gets a change to release tpg->np_login_sem. so A' should finally get timeout). 4. Before A' got timeout. Initiator B gets negotiation failed and calls iscsi_target_login_drop()->iscsi_target_login_sess_out(). The np->np_login_timer is canceled. And initiator A' will hang there forever. Because A' is now in the login thread. All other login requests could not be serviced. Fix this by moving iscsi_stop_login_thread_timer() out of iscsi_target_login_sess_out(). Also remove iscsi_np parameter from iscsi_target_login_sess_out(). Cc: stable(a)vger.kernel.org Signed-off-by: Hou Pu <houpu(a)bytedance.com> --- drivers/target/iscsi/iscsi_target_login.c | 6 +++--- drivers/target/iscsi/iscsi_target_login.h | 3 +-- drivers/target/iscsi/iscsi_target_nego.c | 3 +-- 3 files changed, 5 insertions(+), 7 deletions(-) diff --git a/drivers/target/iscsi/iscsi_target_login.c b/drivers/target/iscsi/iscsi_target_login.c index 85748e338858..893d1b406c29 100644 --- a/drivers/target/iscsi/iscsi_target_login.c +++ b/drivers/target/iscsi/iscsi_target_login.c @@ -1149,7 +1149,7 @@ void iscsit_free_conn(struct iscsi_conn *conn) } void iscsi_target_login_sess_out(struct iscsi_conn *conn, - struct iscsi_np *np, bool zero_tsih, bool new_sess) + bool zero_tsih, bool new_sess) { if (!new_sess) goto old_sess_out; @@ -1167,7 +1167,6 @@ void iscsi_target_login_sess_out(struct iscsi_conn *conn, conn->sess = NULL; old_sess_out: - iscsi_stop_login_thread_timer(np); /* * If login negotiation fails check if the Time2Retain timer * needs to be restarted. @@ -1407,8 +1406,9 @@ static int __iscsi_target_login_thread(struct iscsi_np *np) new_sess_out: new_sess = true; old_sess_out: + iscsi_stop_login_thread_timer(np); tpg_np = conn->tpg_np; - iscsi_target_login_sess_out(conn, np, zero_tsih, new_sess); + iscsi_target_login_sess_out(conn, zero_tsih, new_sess); new_sess = false; if (tpg) { diff --git a/drivers/target/iscsi/iscsi_target_login.h b/drivers/target/iscsi/iscsi_target_login.h index 3b8e3639ff5d..fc95e6150253 100644 --- a/drivers/target/iscsi/iscsi_target_login.h +++ b/drivers/target/iscsi/iscsi_target_login.h @@ -22,8 +22,7 @@ extern int iscsit_put_login_tx(struct iscsi_conn *, struct iscsi_login *, u32); extern void iscsit_free_conn(struct iscsi_conn *); extern int iscsit_start_kthreads(struct iscsi_conn *); extern void iscsi_post_login_handler(struct iscsi_np *, struct iscsi_conn *, u8); -extern void iscsi_target_login_sess_out(struct iscsi_conn *, struct iscsi_np *, - bool, bool); +extern void iscsi_target_login_sess_out(struct iscsi_conn *, bool, bool); extern int iscsi_target_login_thread(void *); extern void iscsi_handle_login_thread_timeout(struct timer_list *t); diff --git a/drivers/target/iscsi/iscsi_target_nego.c b/drivers/target/iscsi/iscsi_target_nego.c index 685d771b51d4..e32d93b92742 100644 --- a/drivers/target/iscsi/iscsi_target_nego.c +++ b/drivers/target/iscsi/iscsi_target_nego.c @@ -535,12 +535,11 @@ static bool iscsi_target_sk_check_and_clear(struct iscsi_conn *conn, unsigned in static void iscsi_target_login_drop(struct iscsi_conn *conn, struct iscsi_login *login) { - struct iscsi_np *np = login->np; bool zero_tsih = login->zero_tsih; iscsi_remove_failed_auth_entry(conn); iscsi_target_nego_release(conn); - iscsi_target_login_sess_out(conn, np, zero_tsih, true); + iscsi_target_login_sess_out(conn, zero_tsih, true); } struct conn_timeout { -- 2.11.0

3 years, 8 months

4
5
0 0

[PATCH] Makefile: add -fuse-ld=lld to KBUILD_HOSTLDFLAGS when LLVM=1

by Nick Desaulniers

While moving Android kernels over to use LLVM=1, we observe the failure when building in a hermetic docker image: HOSTCC scripts/basic/fixdep clang: error: unable to execute command: Executable "ld" doesn't exist! The is because the build of the host utility fixdep builds the fixdep executable in one step by invoking the compiler as the driver, rather than individual compile then link steps. Clang when configured from source defaults to use the system's linker, and not LLVM's own LLD, unless the CMake config -DCLANG_DEFAULT_LINKER='lld' is set when configuring a build of clang itself. Don't rely on the compiler's implicit default linker; be explicit. Cc: stable(a)vger.kernel.org Fixes: commit a0d1c951ef08 ("kbuild: support LLVM=1 to switch the default tools to Clang/LLVM") Reported-by: Matthias Maennich <maennich(a)google.com> Signed-off-by: Nick Desaulniers <ndesaulniers(a)google.com> --- Makefile | 1 + 1 file changed, 1 insertion(+) diff --git a/Makefile b/Makefile index def590b743a9..b4e93b228a26 100644 --- a/Makefile +++ b/Makefile @@ -436,6 +436,7 @@ OBJDUMP = llvm-objdump READELF = llvm-readelf OBJSIZE = llvm-size STRIP = llvm-strip +KBUILD_HOSTLDFLAGS += -fuse-ld=lld else CC = $(CROSS_COMPILE)gcc LD = $(CROSS_COMPILE)ld -- 2.28.0.297.g1956fa8f8d-goog

3 years, 8 months

3
4
0 0

[PATCH V2] block: allow for_each_bvec to support zero len bvec

by Ming Lei

Block layer usually doesn't support or allow zero-length bvec. Since commit 1bdc76aea115 ("iov_iter: use bvec iterator to implement iterate_bvec()"), iterate_bvec() switches to bvec iterator. However, Al mentioned that 'Zero-length segments are not disallowed' in iov_iter. Fixes for_each_bvec() so that it can move on after seeing one zero length bvec. Signed-off-by: Ming Lei <ming.lei(a)redhat.com> Link: https://www.mail-archive.com/linux-kernel@vger.kernel.org/msg2262077.html Fixes: 1bdc76aea115 ("iov_iter: use bvec iterator to implement iterate_bvec()") Reported-by: syzbot <syzbot+61acc40a49a3e46e25ea(a)syzkaller.appspotmail.com> Tested-by: Tetsuo Handa <penguin-kernel(a)i-love.sakura.ne.jp> Cc: Al Viro <viro(a)zeniv.linux.org.uk> Cc: Matthew Wilcox <willy(a)infradead.org> Cc: <stable(a)vger.kernel.org> --- V2: - fix reported-by tag include/linux/bvec.h | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) diff --git a/include/linux/bvec.h b/include/linux/bvec.h index ac0c7299d5b8..9c4fab5f22a7 100644 --- a/include/linux/bvec.h +++ b/include/linux/bvec.h @@ -117,11 +117,18 @@ static inline bool bvec_iter_advance(const struct bio_vec *bv, return true; } +static inline void bvec_iter_skip_zero_bvec(struct bvec_iter *iter) +{ + iter->bi_bvec_done = 0; + iter->bi_idx++; +} + #define for_each_bvec(bvl, bio_vec, iter, start) \ for (iter = (start); \ (iter).bi_size && \ ((bvl = bvec_iter_bvec((bio_vec), (iter))), 1); \ - bvec_iter_advance((bio_vec), &(iter), (bvl).bv_len)) + (bvl).bv_len ? bvec_iter_advance((bio_vec), &(iter), \ + (bvl).bv_len) : bvec_iter_skip_zero_bvec(&(iter))) /* for iterating one bio from start to end */ #define BVEC_ITER_ALL_INIT (struct bvec_iter) \ -- 2.25.2

3 years, 8 months

2
2
0 0

[PATCH 01/11] evm: Execute evm_inode_init_security() only when the HMAC key is loaded

by Roberto Sassu

evm_inode_init_security() requires the HMAC key to calculate the HMAC on initial xattrs provided by LSMs. Unfortunately, with the evm_key_loaded() check, the function continues even if the HMAC key is not loaded (evm_key_loaded() returns true also if EVM has been initialized only with a public key). If the HMAC key is not loaded, evm_inode_init_security() returns an error later when it calls evm_init_hmac(). Thus, this patch replaces the evm_key_loaded() check with a check of the EVM_INIT_HMAC flag in evm_initialized, so that evm_inode_init_security() returns 0 instead of an error. Cc: stable(a)vger.kernel.org # 4.5.x Fixes: 26ddabfe96b ("evm: enable EVM when X509 certificate is loaded") Signed-off-by: Roberto Sassu <roberto.sassu(a)huawei.com> --- security/integrity/evm/evm_main.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/security/integrity/evm/evm_main.c b/security/integrity/evm/evm_main.c index 0d36259b690d..744c105b48d1 100644 --- a/security/integrity/evm/evm_main.c +++ b/security/integrity/evm/evm_main.c @@ -521,7 +521,8 @@ int evm_inode_init_security(struct inode *inode, struct evm_xattr *xattr_data; int rc; - if (!evm_key_loaded() || !evm_protected_xattr(lsm_xattr->name)) + if (!(evm_initialized & EVM_INIT_HMAC) || + !evm_protected_xattr(lsm_xattr->name)) return 0; xattr_data = kzalloc(sizeof(*xattr_data), GFP_NOFS); -- 2.17.1

3 years, 8 months

2
10
0 0

[PATCH] drm/tve200: Stabilize enable/disable

by Linus Walleij

The TVE200 will occasionally print a bunch of lost interrupts and similar dmesg messages, sometimes during boot and sometimes after disabling and coming back to enablement. This is probably because the hardware is left in an unknown state by the boot loader that displays a logo. This can be fixed by bringing the controller into a known state by resetting the controller while enabling it. We retry reset 5 times like the vendor driver does. We also put the controller into reset before de-clocking it and clear all interrupts before enabling the vblank IRQ. This makes the video enable/disable/enable cycle rock solid on the D-Link DIR-685. Tested extensively. Cc: stable(a)vger.kernel.org Signed-off-by: Linus Walleij <linus.walleij(a)linaro.org> --- drivers/gpu/drm/tve200/tve200_display.c | 22 +++++++++++++++++++++- 1 file changed, 21 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/tve200/tve200_display.c b/drivers/gpu/drm/tve200/tve200_display.c index d733bbc4ac0e..17ff24d999d1 100644 --- a/drivers/gpu/drm/tve200/tve200_display.c +++ b/drivers/gpu/drm/tve200/tve200_display.c @@ -14,6 +14,7 @@ #include <linux/version.h> #include <linux/dma-buf.h> #include <linux/of_graph.h> +#include <linux/delay.h> #include <drm/drm_fb_cma_helper.h> #include <drm/drm_fourcc.h> @@ -130,9 +131,25 @@ static void tve200_display_enable(struct drm_simple_display_pipe *pipe, struct drm_connector *connector = priv->connector; u32 format = fb->format->format; u32 ctrl1 = 0; + int retries; clk_prepare_enable(priv->clk); + /* Reset the TVE200 and wait for it to come back online */ + writel(TVE200_CTRL_4_RESET, priv->regs + TVE200_CTRL_4); + for (retries = 0; retries < 5; retries++) { + usleep_range(30000, 50000); + if (readl(priv->regs + TVE200_CTRL_4) & TVE200_CTRL_4_RESET) + continue; + else + break; + } + if (retries == 5 && + readl(priv->regs + TVE200_CTRL_4) & TVE200_CTRL_4_RESET) { + dev_err(drm->dev, "can't get hardware out of reset\n"); + return; + } + /* Function 1 */ ctrl1 |= TVE200_CTRL_CSMODE; /* Interlace mode for CCIR656: parameterize? */ @@ -230,8 +247,9 @@ static void tve200_display_disable(struct drm_simple_display_pipe *pipe) drm_crtc_vblank_off(crtc); - /* Disable and Power Down */ + /* Disable put into reset and Power Down */ writel(0, priv->regs + TVE200_CTRL); + writel(TVE200_CTRL_4_RESET, priv->regs + TVE200_CTRL_4); clk_disable_unprepare(priv->clk); } @@ -279,6 +297,8 @@ static int tve200_display_enable_vblank(struct drm_simple_display_pipe *pipe) struct drm_device *drm = crtc->dev; struct tve200_drm_dev_private *priv = drm->dev_private; + /* Clear any IRQs and enable */ + writel(0xFF, priv->regs + TVE200_INT_CLR); writel(TVE200_INT_V_STATUS, priv->regs + TVE200_INT_EN); return 0; } -- 2.26.2

3 years, 8 months

2
2
0 0

Re: [PATCH v2 1/1] netfilter: nat: add a range check for l3/l4 protonum

by William Mcvicker

Hi, I have attached the follow up fix that checks for the proto index during conntrack creation. Thanks, Will On 07/31/2020, William Mcvicker wrote: > Hi Pablo, > > > Note that this code does not exist in the tree anymore. I'm not sure > > if this problem still exists upstream, this patch does not apply to > > nf.git. This fix should only go for -stable maintainers. > > Right, the vulnerability has been fixed by the refactor commit fe2d0020994cd > ("netfilter: nat: remove l4proto->in_range"), but this patch is a part of > a full re-work of the code and doesn't backport very cleanly to the LTS > branches. So this fix is only applicable to the 4.19, 4.14, 4.9, and 4.4 LTS > branches. I missed the -stable email, but will re-add it to this thread with > the re-worked patch. > > Thanks, > Will > > On 07/31/2020, Pablo Neira Ayuso wrote: > > Hi William, > > > > On Fri, Jul 31, 2020 at 12:26:11AM +0000, William Mcvicker wrote: > > > Hi Pablo, > > > > > > Yes, I believe this oops is only triggered by userspace when the user > > > specifically passes in an invalid nf_nat_l3protos index. I'm happy to re-work > > > the patch to check for this in ctnetlink_create_conntrack(). > > > > Great. > > > > Note that this code does not exist in the tree anymore. I'm not sure > > if this problem still exists upstream, this patch does not apply to > > nf.git. This fix should only go for -stable maintainers. > > > > > > BTW, do you have a Fixes: tag for this? This will be useful for > > > > -stable maintainer to pick up this fix. > > > > > > Regarding the Fixes: tag, I don't have one offhand since this bug was reported > > > to me, but I can search through the code history to find the commit that > > > exposed this vulnerability. > > > > That would be great. > > > > Thank you.

3 years, 8 months

5
8
0 0

Re: [PATCH 1/2] mm: memcontrol: flush percpu vmstats before releasing memcg

by Roman Gushchin

On Tue, Aug 13, 2019 at 02:27:52PM -0700, Andrew Morton wrote: > On Mon, 12 Aug 2019 15:29:10 -0700 Roman Gushchin <guro(a)fb.com> wrote: > > > Percpu caching of local vmstats with the conditional propagation > > by the cgroup tree leads to an accumulation of errors on non-leaf > > levels. > > > > Let's imagine two nested memory cgroups A and A/B. Say, a process > > belonging to A/B allocates 100 pagecache pages on the CPU 0. > > The percpu cache will spill 3 times, so that 32*3=96 pages will be > > accounted to A/B and A atomic vmstat counters, 4 pages will remain > > in the percpu cache. > > > > Imagine A/B is nearby memory.max, so that every following allocation > > triggers a direct reclaim on the local CPU. Say, each such attempt > > will free 16 pages on a new cpu. That means every percpu cache will > > have -16 pages, except the first one, which will have 4 - 16 = -12. > > A/B and A atomic counters will not be touched at all. > > > > Now a user removes A/B. All percpu caches are freed and corresponding > > vmstat numbers are forgotten. A has 96 pages more than expected. > > > > As memory cgroups are created and destroyed, errors do accumulate. > > Even 1-2 pages differences can accumulate into large numbers. > > > > To fix this issue let's accumulate and propagate percpu vmstat > > values before releasing the memory cgroup. At this point these > > numbers are stable and cannot be changed. > > > > Since on cpu hotplug we do flush percpu vmstats anyway, we can > > iterate only over online cpus. > > > > Fixes: 42a300353577 ("mm: memcontrol: fix recursive statistics correctness & scalabilty") > > Is this not serious enough for a cc:stable? I hope the "Fixes" tag will work, but yeah, my bad, cc:stable is definitely a good idea here. Added stable@ to cc. Thanks!

3 years, 8 months

2
1
0 0

[PATCH] gfs2: Make sure we don't miss any delayed withdraws

by Andreas Gruenbacher

Commit ca399c96e96e changes gfs2_log_flush to not withdraw the filesystem while holding the log flush lock, but it fails to check if the filesystem needs to be withdrawn once the log flush lock has been released. Likewise, commit f05b86db314d depends on gfs2_log_flush to trigger for delayed withdraws. Add that and clean up the code flow somewhat. In gfs2_put_super, add a check for delayed withdraws that have been missed to prevent these kinds of bugs in the future. Fixes: ca399c96e96e ("gfs2: flesh out delayed withdraw for gfs2_log_flush") Fixes: f05b86db314d ("gfs2: Prepare to withdraw as soon as an IO error occurs in log write") Cc: stable(a)vger.kernel.org # v5.7+ Signed-off-by: Andreas Gruenbacher <agruenba(a)redhat.com> --- fs/gfs2/log.c | 61 +++++++++++++++++++++++++------------------------ fs/gfs2/super.c | 2 ++ fs/gfs2/util.h | 10 ++++++++ 3 files changed, 43 insertions(+), 30 deletions(-) diff --git a/fs/gfs2/log.c b/fs/gfs2/log.c index 3763c9ff1406..93032feb5159 100644 --- a/fs/gfs2/log.c +++ b/fs/gfs2/log.c @@ -954,10 +954,8 @@ void gfs2_log_flush(struct gfs2_sbd *sdp, struct gfs2_glock *gl, u32 flags) goto out; /* Log might have been flushed while we waited for the flush lock */ - if (gl && !test_bit(GLF_LFLUSH, &gl->gl_flags)) { - up_write(&sdp->sd_log_flush_lock); - return; - } + if (gl && !test_bit(GLF_LFLUSH, &gl->gl_flags)) + goto out; trace_gfs2_log_flush(sdp, 1, flags); if (flags & GFS2_LOG_HEAD_FLUSH_SHUTDOWN) @@ -971,25 +969,25 @@ void gfs2_log_flush(struct gfs2_sbd *sdp, struct gfs2_glock *gl, u32 flags) if (unlikely (state == SFS_FROZEN)) if (gfs2_assert_withdraw_delayed(sdp, !tr->tr_num_buf_new && !tr->tr_num_databuf_new)) - goto out; + goto out_withdraw; } if (unlikely(state == SFS_FROZEN)) if (gfs2_assert_withdraw_delayed(sdp, !sdp->sd_log_num_revoke)) - goto out; + goto out_withdraw; if (gfs2_assert_withdraw_delayed(sdp, sdp->sd_log_num_revoke == sdp->sd_log_committed_revoke)) - goto out; + goto out_withdraw; gfs2_ordered_write(sdp); if (gfs2_withdrawn(sdp)) - goto out; + goto out_withdraw; lops_before_commit(sdp, tr); if (gfs2_withdrawn(sdp)) - goto out; + goto out_withdraw; gfs2_log_submit_bio(&sdp->sd_log_bio, REQ_OP_WRITE); if (gfs2_withdrawn(sdp)) - goto out; + goto out_withdraw; if (sdp->sd_log_head != sdp->sd_log_flush_head) { log_flush_wait(sdp); @@ -1000,7 +998,7 @@ void gfs2_log_flush(struct gfs2_sbd *sdp, struct gfs2_glock *gl, u32 flags) log_write_header(sdp, flags); } if (gfs2_withdrawn(sdp)) - goto out; + goto out_withdraw; lops_after_commit(sdp, tr); gfs2_log_lock(sdp); @@ -1020,7 +1018,7 @@ void gfs2_log_flush(struct gfs2_sbd *sdp, struct gfs2_glock *gl, u32 flags) if (!sdp->sd_log_idle) { empty_ail1_list(sdp); if (gfs2_withdrawn(sdp)) - goto out; + goto out_withdraw; atomic_dec(&sdp->sd_log_blks_free); /* Adjust for unreserved buffer */ trace_gfs2_log_blocks(sdp, -1); log_write_header(sdp, flags); @@ -1033,27 +1031,30 @@ void gfs2_log_flush(struct gfs2_sbd *sdp, struct gfs2_glock *gl, u32 flags) atomic_set(&sdp->sd_freeze_state, SFS_FROZEN); } -out: - if (gfs2_withdrawn(sdp)) { - trans_drain(tr); - /** - * If the tr_list is empty, we're withdrawing during a log - * flush that targets a transaction, but the transaction was - * never queued onto any of the ail lists. Here we add it to - * ail1 just so that ail_drain() will find and free it. - */ - spin_lock(&sdp->sd_ail_lock); - if (tr && list_empty(&tr->tr_list)) - list_add(&tr->tr_list, &sdp->sd_ail1_list); - spin_unlock(&sdp->sd_ail_lock); - ail_drain(sdp); /* frees all transactions */ - tr = NULL; - } - +out_end: trace_gfs2_log_flush(sdp, 0, flags); +out: up_write(&sdp->sd_log_flush_lock); - gfs2_trans_free(sdp, tr); + if (gfs2_withdrawing(sdp)) + gfs2_withdraw(sdp); + return; + +out_withdraw: + trans_drain(tr); + /** + * If the tr_list is empty, we're withdrawing during a log + * flush that targets a transaction, but the transaction was + * never queued onto any of the ail lists. Here we add it to + * ail1 just so that ail_drain() will find and free it. + */ + spin_lock(&sdp->sd_ail_lock); + if (tr && list_empty(&tr->tr_list)) + list_add(&tr->tr_list, &sdp->sd_ail1_list); + spin_unlock(&sdp->sd_ail_lock); + ail_drain(sdp); /* frees all transactions */ + tr = NULL; + goto out_end; } /** diff --git a/fs/gfs2/super.c b/fs/gfs2/super.c index 9f4d9e7be839..19add2da1013 100644 --- a/fs/gfs2/super.c +++ b/fs/gfs2/super.c @@ -702,6 +702,8 @@ static void gfs2_put_super(struct super_block *sb) if (error) gfs2_io_error(sdp); } + WARN_ON(gfs2_withdrawing(sdp)); + /* At this point, we're through modifying the disk */ /* Release stuff */ diff --git a/fs/gfs2/util.h b/fs/gfs2/util.h index 6d9157efe16c..d7562981b3a0 100644 --- a/fs/gfs2/util.h +++ b/fs/gfs2/util.h @@ -205,6 +205,16 @@ static inline bool gfs2_withdrawn(struct gfs2_sbd *sdp) test_bit(SDF_WITHDRAWING, &sdp->sd_flags); } +/** + * gfs2_withdrawing - check if a withdraw is pending + * @sdp: the superblock + */ +static inline bool gfs2_withdrawing(struct gfs2_sbd *sdp) +{ + return test_bit(SDF_WITHDRAWING, &sdp->sd_flags) && + !test_bit(SDF_WITHDRAWN, &sdp->sd_flags); +} + #define gfs2_tune_get(sdp, field) \ gfs2_tune_get_i(&(sdp)->sd_tune, &(sdp)->sd_tune.field) -- 2.26.2

3 years, 8 months

3
2
0 0

2024

2023

2022

2021

2020

2019

2018

2017

Linux-stable-mirror August 2020