June 2019 - Linux-stable-mirror

[PATCH] [media] media: mtk-mdp: fix reference count on old device tree

by Matthias Brugger

of_get_next_child() increments the reference count of the returning device_node. Decrement it in the check if we are using the old or the new DTB. Fixes: ba1f1f70c2c0 ("[media] media: mtk-mdp: Fix mdp device tree") Signed-off-by: Matthias Brugger <matthias.bgg(a)gmail.com> --- drivers/media/platform/mtk-mdp/mtk_mdp_core.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/drivers/media/platform/mtk-mdp/mtk_mdp_core.c b/drivers/media/platform/mtk-mdp/mtk_mdp_core.c index bbb24fb95b95..bafe53c5d54a 100644 --- a/drivers/media/platform/mtk-mdp/mtk_mdp_core.c +++ b/drivers/media/platform/mtk-mdp/mtk_mdp_core.c @@ -118,7 +118,9 @@ static int mtk_mdp_probe(struct platform_device *pdev) mutex_init(&mdp->vpulock); /* Old dts had the components as child nodes */ - if (of_get_next_child(dev->of_node, NULL)) { + parent = of_get_next_child(dev->of_node, NULL); + if (parent) { + of_node_put(parent); parent = dev->of_node; dev_warn(dev, "device tree is out of date\n"); } else { -- 2.21.0

4 years, 9 months

3
3
0 0

[PATCH v4.4 00/45] V4.4 backport of arm64 Spectre patches

by Viresh Kumar

Hello, Here is an attempt to backport arm64 spectre patches to v4.4 stable tree. I have started this backport with Mark Rutland's backport of Spectre to 4.9 [1] and tried applying the upstream version of them over 4.4 and resolved conflicts by checking how they have been resolved in 4.9. I had to pick few extra upstream patches to avoid unnecessary conflicts (upstream commit ids mentioned): a842789837c0 arm64: remove duplicate macro __KERNEL__ check 64f8ebaf115b mm/kasan: add API to check memory regions bffe1baff5d5 arm64: kasan: instrument user memory access API 92406f0cc9e3 arm64: cpufeature: Add scope for capability check 9eb8a2cdf65c arm64: cputype info for Broadcom Vulcan 0d90718871fe arm64: cputype: Add MIDR values for Cavium ThunderX2 CPUs 98dd64f34f47 ARM: 8478/2: arm/arm64: add arm-smccc I had to drop few patches as well as they weren't getting applied properly due to missing files/features (upstream commit id mentioned): 93f339ef4175 arm64: cpufeature: __this_cpu_has_cap() shouldn't stop early 3c31fa5a06b4 arm64: Run enable method for errata work arounds on late CPUs 6840bdd73d07 arm64: KVM: Use per-CPU vector when BP hardening is enabled 90348689d500 arm64: KVM: Make PSCI_VERSION a fast path Since v4.4 doesn't contain arch/arm/kvm/hyp/switch.c file, changes for it are dropped from some of the patches. The commit log of specific patches are updated with this information. Also for commit id (from 4.9 stable): c24c205d2528 arm64: Add ARM_SMCCC_ARCH_WORKAROUND_1 BP hardening support I have dropped arch/arm64/crypto/sha256-core.S and sha512-core.S files as they weren't part of the upstream commit. Not sure why it was included by Mark as the commit log doesn't provide any reasoning for it. The patches in this series are pushed here [2]. This is only build/boot tested by me as I don't have access to the required test-suite which can verify spectre mitigations. @Julien: Can you please help reviewing / testing them ? Thanks. -- viresh [1] https://patches.linaro.org/cover/133195/ with top commit in 4.9 stable tree: a3b292fe0560 arm64: futex: Mask __user pointers prior to dereference [2] https://git.kernel.org/pub/scm/linux/kernel/git/vireshk/linux.git stable/v4.4.y/spectre Andrey Ryabinin (1): mm/kasan: add API to check memory regions Catalin Marinas (1): arm64: Factor out TTBR0_EL1 post-update workaround into a specific asm macro Jayachandran C (3): arm64: cputype info for Broadcom Vulcan arm64: cputype: Add MIDR values for Cavium ThunderX2 CPUs arm64: Branch predictor hardening for Cavium ThunderX2 Jens Wiklander (1): ARM: 8478/2: arm/arm64: add arm-smccc Laura Abbott (1): mm: Introduce lm_alias Marc Zyngier (14): arm64: Move post_ttbr_update_workaround to C code arm64: Move BP hardening to check_and_switch_context arm64: cpu_errata: Allow an erratum to be match for all revisions of a core arm64: KVM: Increment PC after handling an SMC trap arm/arm64: KVM: Add PSCI_VERSION helper arm/arm64: KVM: Add smccc accessors to PSCI code arm/arm64: KVM: Implement PSCI 1.0 support arm64: KVM: Add SMCCC_ARCH_WORKAROUND_1 fast handling firmware/psci: Expose PSCI conduit firmware/psci: Expose SMCCC version through psci_ops arm/arm64: smccc: Make function identifiers an unsigned quantity arm/arm64: smccc: Implement SMCCC v1.1 inline primitive arm64: Add ARM_SMCCC_ARCH_WORKAROUND_1 BP hardening support arm64: Kill PSCI_GET_VERSION as a variant-2 workaround Mark Rutland (4): arm/arm64: KVM: Consolidate the PSCI include files arm/arm64: KVM: Advertise SMCCC v1.1 arm/arm64: KVM: Turn kvm_psci_version into a static inline arm64: KVM: Report SMCCC_ARCH_WORKAROUND_1 BP hardening support Robin Murphy (3): arm64: Implement array_index_mask_nospec() arm64: Make USER_DS an inclusive limit arm64: Use pointer masking to limit uaccess speculation Suzuki K Poulose (1): arm64: cpufeature: Add scope for capability check Will Deacon (13): arm64: barrier: Add CSDB macros to control data-value prediction arm64: entry: Ensure branch through syscall table is bounded under speculation arm64: uaccess: Prevent speculative use of the current addr_limit arm64: uaccess: Don't bother eliding access_ok checks in __{get, put}_user arm64: uaccess: Mask __user pointers for __arch_{clear, copy_*}_user arm64: cpufeature: Pass capability structure to ->enable callback drivers/firmware: Expose psci_get_version through psci_ops structure arm64: Add skeleton to harden the branch predictor against aliasing attacks arm64: entry: Apply BP hardening for high-priority synchronous exceptions arm64: entry: Apply BP hardening for suspicious interrupts from EL0 arm64: cputype: Add missing MIDR values for Cortex-A72 and Cortex-A75 arm64: Implement branch predictor hardening for affected Cortex-A CPUs arm64: futex: Mask __user pointers prior to dereference Yang Shi (1): arm64: kasan: instrument user memory access API Yury Norov (1): arm64: move TASK_* definitions to <asm/processor.h> zijun_hu (1): arm64: remove duplicate macro __KERNEL__ check MAINTAINERS | 14 ++ arch/arm/include/asm/kvm_host.h | 6 + arch/arm/include/asm/kvm_psci.h | 27 --- arch/arm/kvm/arm.c | 2 +- arch/arm/kvm/handle_exit.c | 4 +- arch/arm/kvm/psci.c | 143 ++++++++++++--- arch/arm64/Kconfig | 17 ++ arch/arm64/include/asm/assembler.h | 18 ++ arch/arm64/include/asm/barrier.h | 23 +++ arch/arm64/include/asm/cpufeature.h | 12 +- arch/arm64/include/asm/cputype.h | 12 ++ arch/arm64/include/asm/futex.h | 9 +- arch/arm64/include/asm/kvm_host.h | 5 + arch/arm64/include/asm/kvm_psci.h | 27 --- arch/arm64/include/asm/memory.h | 15 -- arch/arm64/include/asm/mmu.h | 39 ++++ arch/arm64/include/asm/processor.h | 26 ++- arch/arm64/include/asm/sysreg.h | 2 + arch/arm64/include/asm/uaccess.h | 175 ++++++++++++------ arch/arm64/kernel/Makefile | 5 + arch/arm64/kernel/arm64ksyms.c | 8 +- arch/arm64/kernel/bpi.S | 75 ++++++++ arch/arm64/kernel/cpu_errata.c | 185 ++++++++++++++++++- arch/arm64/kernel/cpufeature.c | 112 ++++++------ arch/arm64/kernel/entry.S | 26 ++- arch/arm64/kvm/handle_exit.c | 16 +- arch/arm64/kvm/hyp.S | 20 ++- arch/arm64/lib/clear_user.S | 6 +- arch/arm64/lib/copy_from_user.S | 4 +- arch/arm64/lib/copy_in_user.S | 4 +- arch/arm64/lib/copy_to_user.S | 4 +- arch/arm64/mm/context.c | 12 ++ arch/arm64/mm/fault.c | 31 ++++ arch/arm64/mm/proc.S | 12 +- drivers/firmware/Kconfig | 3 + drivers/firmware/psci.c | 58 +++++- include/kvm/arm_psci.h | 51 ++++++ include/linux/arm-smccc.h | 267 ++++++++++++++++++++++++++++ include/linux/kasan-checks.h | 12 ++ include/linux/mm.h | 4 + include/linux/psci.h | 14 ++ include/uapi/linux/psci.h | 3 + mm/kasan/kasan.c | 12 ++ 43 files changed, 1270 insertions(+), 250 deletions(-) delete mode 100644 arch/arm/include/asm/kvm_psci.h delete mode 100644 arch/arm64/include/asm/kvm_psci.h create mode 100644 arch/arm64/kernel/bpi.S create mode 100644 include/kvm/arm_psci.h create mode 100644 include/linux/arm-smccc.h create mode 100644 include/linux/kasan-checks.h -- 2.21.0.rc0.269.g1a574e7a288b

4 years, 9 months

3
56
0 0

[PATCH AUTOSEL 4.19 01/60] soc: brcmstb: Fix error path for unsupported CPUs

by Sasha Levin

From: Florian Fainelli <f.fainelli(a)gmail.com> [ Upstream commit 490cad5a3ad6ef0bfd3168a5063140b982f3b22a ] In case setup_hifcpubiuctrl_regs() returns an error, because of e.g: an unsupported CPU type, just catch that error and return instead of blindly continuing with the initialization. This fixes a NULL pointer de-reference with the code continuing without having a proper array of registers to use. Fixes: 22f7a9116eba ("soc: brcmstb: Correct CPU_CREDIT_REG offset for Brahma-B53 CPUs") Signed-off-by: Florian Fainelli <f.fainelli(a)gmail.com> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- drivers/soc/bcm/brcmstb/biuctrl.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/drivers/soc/bcm/brcmstb/biuctrl.c b/drivers/soc/bcm/brcmstb/biuctrl.c index 6d89ebf13b8a..c16273b31b94 100644 --- a/drivers/soc/bcm/brcmstb/biuctrl.c +++ b/drivers/soc/bcm/brcmstb/biuctrl.c @@ -246,7 +246,9 @@ static int __init brcmstb_biuctrl_init(void) if (!np) return 0; - setup_hifcpubiuctrl_regs(np); + ret = setup_hifcpubiuctrl_regs(np); + if (ret) + return ret; ret = mcp_write_pairing_set(); if (ret) { -- 2.20.1

4 years, 9 months

2
62
0 0

[4.4.y PATCH 0/4] Backported fixes for 4.4 stable tree

by Srivatsa S. Bhat

Hi, This patchset includes a few backported fixes for the 4.4 stable tree. I would appreciate if you could kindly consider including them in the next release. Thank you! Regards, Srivatsa --- Gen Zhang (2): ip_sockglue: Fix missing-check bug in ip_ra_control() ipv6_sockglue: Fix a missing-check bug in ip6_ra_control() Vivek Goyal (1): ovl: modify ovl_permission() to do checks on two inodes Wanpeng Li (1): KVM: X86: Fix scan ioapic use-before-initialization arch/x86/kvm/x86.c | 3 ++- fs/overlayfs/inode.c | 13 +++++++++++++ net/ipv4/ip_sockglue.c | 2 ++ net/ipv6/ipv6_sockglue.c | 2 ++ 4 files changed, 19 insertions(+), 1 deletion(-)

4 years, 9 months

3
8
0 0

[PATCH 3.16 0/2] Fix FUSE read/write deadlock on stream-like files

by Kirill Smelkov

Hello stable team, Please consider applying the following 2 patches to Linux-3.16 stable tree. The patches fix regression introduced in 3.14 where both read and write started to run under lock taken, which resulted in FUSE (and many other drivers) deadlocks for cases where stream-like files are used with read and write being run simultaneously. Please see complete problem description in upstream commit 10dce8af3422 ("fs: stream_open - opener for stream-like files so that read and write can run simultaneously without deadlock"). The actual FUSE fix (upstream commit bbd84f33652f "fuse: Add FOPEN_STREAM to use stream_open()") was merged into 5.2 with `Cc: stable(a)vger.kernel.org # v3.14+` mark and is already included into 5.1, 5.0 and 4.19 stable trees. However for some reason it is not (yet ?) included into 4.14, 4.9, 4.4, 3.18 and 3.16 trees. The patches fix a real problem into which my FUSE filesystem ran, and which also likely affects OSSPD (full details are in the patches description). Please consider including the fixes into 3.16 (as well as into other stable trees - I'm sending corresponding series separately - - one per tree). Thanks beforehand, Kirill P.S. the patches have been already a bit discussed in stable context some time ago: https://lore.kernel.org/linux-fsdevel/CAHk-=wgh234SyBG810=vB360PCzVkAhQRqGg… https://lore.kernel.org/linux-fsdevel/20190424183012.GB3798@deco.navytux.sp… https://lore.kernel.org/linux-fsdevel/20190424191652.GE3798@deco.navytux.sp… ... Kirill Smelkov (2): fs: stream_open - opener for stream-like files so that read and write can run simultaneously without deadlock fuse: Add FOPEN_STREAM to use stream_open() drivers/xen/xenbus/xenbus_dev_frontend.c | 2 +- fs/fuse/file.c | 4 +- fs/open.c | 18 ++ fs/read_write.c | 5 +- include/linux/fs.h | 4 + include/uapi/linux/fuse.h | 2 + scripts/coccinelle/api/stream_open.cocci | 363 +++++++++++++++++++++++ 7 files changed, 394 insertions(+), 4 deletions(-) create mode 100644 scripts/coccinelle/api/stream_open.cocci -- 2.20.1

4 years, 9 months

2
4
0 0

[PATCH] perf: assign proper ff->ph in perf_event__synthesize_features()

by Song Liu

bpf/btf write_* functions need ff->ph->env. With this missing, pipe-mode (perf record -o -) would crash like: Program terminated with signal SIGSEGV, Segmentation fault. This patch assign proper ph value to ff. Cc: stable(a)vger.kernel.org #v5.1+ Fixes: 606f972b1361 ("perf bpf: Save bpf_prog_info information as headers to perf.data") Reported-by: David Carrillo Cisneros <davidca(a)fb.com> Signed-off-by: Song Liu <songliubraving(a)fb.com> --- tools/perf/util/header.c | 1 + 1 file changed, 1 insertion(+) diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c index 06ddb6618ef3..5f1aa0284e1b 100644 --- a/tools/perf/util/header.c +++ b/tools/perf/util/header.c @@ -3684,6 +3684,7 @@ int perf_event__synthesize_features(struct perf_tool *tool, return -ENOMEM; ff.size = sz - sz_hdr; + ff.ph = &session->header; for_each_set_bit(feat, header->adds_features, HEADER_FEAT_BITS) { if (!feat_ops[feat].synthesize) { -- 2.17.1

4 years, 9 months

2
1
0 0

[PATCH] Bluetooth: hci_ldisc: check for missing tty operations

by Vladis Dronov

Certain ttys lack operations (eg: pty_unix98_ops) which still can be called by certain hci uart proto (eg: mrvl, ath). Currently this leads to execution at address 0x0. Fix this by adding checks for missing tty operations. Link: https://syzkaller.appspot.com/bug?id=1b42faa2848963564a5b1b7f8c837ea7b55ffa… Reported-by: syzbot+79337b501d6aa974d0f6(a)syzkaller.appspotmail.com Cc: stable(a)vger.kernel.org # v2.6.39+ Signed-off-by: Vladis Dronov <vdronov(a)redhat.com> --- drivers/bluetooth/hci_ath.c | 7 ++++- drivers/bluetooth/hci_ldisc.c | 58 ++++++++++++++++++++--------------- 2 files changed, 39 insertions(+), 26 deletions(-) diff --git a/drivers/bluetooth/hci_ath.c b/drivers/bluetooth/hci_ath.c index a55be205b91a..92d9bded5880 100644 --- a/drivers/bluetooth/hci_ath.c +++ b/drivers/bluetooth/hci_ath.c @@ -49,7 +49,12 @@ struct ath_vendor_cmd { static int ath_wakeup_ar3k(struct tty_struct *tty) { - int status = tty->driver->ops->tiocmget(tty); + int status; + + if (!(tty->driver->ops->tiocmget && tty->driver->ops->tiocmset)) + return TIOCM_CTS; + + status = tty->driver->ops->tiocmget(tty); if (status & TIOCM_CTS) return status; diff --git a/drivers/bluetooth/hci_ldisc.c b/drivers/bluetooth/hci_ldisc.c index c84f985f348d..29b4970f9bca 100644 --- a/drivers/bluetooth/hci_ldisc.c +++ b/drivers/bluetooth/hci_ldisc.c @@ -307,32 +307,40 @@ void hci_uart_set_flow_control(struct hci_uart *hu, bool enable) BT_DBG("Disabling hardware flow control: %s", status ? "failed" : "success"); - /* Clear RTS to prevent the device from sending */ - /* Most UARTs need OUT2 to enable interrupts */ - status = tty->driver->ops->tiocmget(tty); - BT_DBG("Current tiocm 0x%x", status); - - set &= ~(TIOCM_OUT2 | TIOCM_RTS); - clear = ~set; - set &= TIOCM_DTR | TIOCM_RTS | TIOCM_OUT1 | - TIOCM_OUT2 | TIOCM_LOOP; - clear &= TIOCM_DTR | TIOCM_RTS | TIOCM_OUT1 | - TIOCM_OUT2 | TIOCM_LOOP; - status = tty->driver->ops->tiocmset(tty, set, clear); - BT_DBG("Clearing RTS: %s", status ? "failed" : "success"); + if (tty->driver->ops->tiocmget && tty->driver->ops->tiocmset) { + /* Clear RTS to prevent the device from sending */ + /* Most UARTs need OUT2 to enable interrupts */ + status = tty->driver->ops->tiocmget(tty); + BT_DBG("Current tiocm 0x%x", status); + + set &= ~(TIOCM_OUT2 | TIOCM_RTS); + clear = ~set; + set &= TIOCM_DTR | TIOCM_RTS | TIOCM_OUT1 | + TIOCM_OUT2 | TIOCM_LOOP; + clear &= TIOCM_DTR | TIOCM_RTS | TIOCM_OUT1 | + TIOCM_OUT2 | TIOCM_LOOP; + status = tty->driver->ops->tiocmset(tty, set, clear); + BT_DBG("Clearing RTS: %s", + status ? "failed" : "success"); + } else + BT_DBG("Terminal RTS control is not present"); } else { - /* Set RTS to allow the device to send again */ - status = tty->driver->ops->tiocmget(tty); - BT_DBG("Current tiocm 0x%x", status); - - set |= (TIOCM_OUT2 | TIOCM_RTS); - clear = ~set; - set &= TIOCM_DTR | TIOCM_RTS | TIOCM_OUT1 | - TIOCM_OUT2 | TIOCM_LOOP; - clear &= TIOCM_DTR | TIOCM_RTS | TIOCM_OUT1 | - TIOCM_OUT2 | TIOCM_LOOP; - status = tty->driver->ops->tiocmset(tty, set, clear); - BT_DBG("Setting RTS: %s", status ? "failed" : "success"); + if (tty->driver->ops->tiocmget && tty->driver->ops->tiocmset) { + /* Set RTS to allow the device to send again */ + status = tty->driver->ops->tiocmget(tty); + BT_DBG("Current tiocm 0x%x", status); + + set |= (TIOCM_OUT2 | TIOCM_RTS); + clear = ~set; + set &= TIOCM_DTR | TIOCM_RTS | TIOCM_OUT1 | + TIOCM_OUT2 | TIOCM_LOOP; + clear &= TIOCM_DTR | TIOCM_RTS | TIOCM_OUT1 | + TIOCM_OUT2 | TIOCM_LOOP; + status = tty->driver->ops->tiocmset(tty, set, clear); + BT_DBG("Setting RTS: %s", + status ? "failed" : "success"); + } else + BT_DBG("Terminal RTS control is not present"); /* Re-enable hardware flow control */ ktermios = tty->termios; -- 2.20.1

4 years, 9 months

2
2
0 0

[PATCH] Bluetooth: SMP: Workaround Microsoft Surface Precision Mouse bug

by Szymon Janc

Microsoft Surface Precision Mouse provides bogus identity address when pairing. It connects with Static Random address but provides Public Address in SMP Identity Address Information PDU. Address has same value but type is different. Workaround this by dropping IRK if ID address discrepancy is detected. > HCI Event: LE Meta Event (0x3e) plen 19 LE Connection Complete (0x01) Status: Success (0x00) Handle: 75 Role: Master (0x00) Peer address type: Random (0x01) Peer address: E0:52:33:93:3B:21 (Static) Connection interval: 50.00 msec (0x0028) Connection latency: 0 (0x0000) Supervision timeout: 420 msec (0x002a) Master clock accuracy: 0x00 .... > ACL Data RX: Handle 75 flags 0x02 dlen 12 SMP: Identity Address Information (0x09) len 7 Address type: Public (0x00) Address: E0:52:33:93:3B:21 Signed-off-by: Szymon Janc <szymon.janc(a)codecoup.pl> Tested-by: Maarten Fonville <maarten.fonville(a)gmail.com> Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=199461 Cc: stable(a)vger.kernel.org --- net/bluetooth/smp.c | 14 ++++++++++++++ 1 file changed, 14 insertions(+) diff --git a/net/bluetooth/smp.c b/net/bluetooth/smp.c index e68c715f8d37..d528d95a2c27 100644 --- a/net/bluetooth/smp.c +++ b/net/bluetooth/smp.c @@ -2579,6 +2579,20 @@ static int smp_cmd_ident_addr_info(struct l2cap_conn *conn, goto distribute; } + /* + * Drop IRK if peer is using identity address during pairing but is + * providing different address as identity information. + * + * Microsoft Surface Precision Mouse is known to have this bug. + */ + if (hci_is_identity_address(&hcon->dst, hcon->dst_type) && + (bacmp(&info->bdaddr, &hcon->dst) || + info->addr_type != hcon->dst_type)) { + bt_dev_err(hcon->hdev, + "ignoring IRK with invalid identity address"); + goto distribute; + } + bacpy(&smp->id_addr, &info->bdaddr); smp->id_addr_type = info->addr_type; -- 2.21.0

4 years, 9 months

2
2
0 0

[PATCH v3 2/2] Bluetooth: hci_ldisc: Add NULL check for tiocmget() and tiocmset() in hci_uart_set_flow_control()

by Myungho Jung

tiocmget() or tiocmset() operations are optional. Just return from hci_uart_set_flow_control() if tiocmget() or tiocmset() operation is NULL. Fixes: 2a973dfada2b ("hci_uart: Add new line discipline enhancements") Cc: <stable(a)vger.kernel.org> # 4.2 Signed-off-by: Myungho Jung <mhjungk(a)gmail.com> --- Changes in v2: - Remove braces in if statment Changes in v3: - Split into 2 patches - Add stable CC and fixes tags drivers/bluetooth/hci_ldisc.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/drivers/bluetooth/hci_ldisc.c b/drivers/bluetooth/hci_ldisc.c index fbf7b4df23ab..cb31c2d8d826 100644 --- a/drivers/bluetooth/hci_ldisc.c +++ b/drivers/bluetooth/hci_ldisc.c @@ -314,6 +314,10 @@ void hci_uart_set_flow_control(struct hci_uart *hu, bool enable) return; } + /* tiocmget() and tiocmset() operations are optional */ + if (!tty->driver->ops->tiocmget || !tty->driver->ops->tiocmset) + return; + if (enable) { /* Disable hardware flow control */ ktermios = tty->termios; -- 2.17.1

4 years, 9 months

2
1
0 0

[PATCH] IB/hfi1: Close PSM sdma_progress sleep window

by Mike Marciniszyn

commit da9de5f8527f4b9efc82f967d29a583318c034c7 upstream. The call to sdma_progress() is called outside the wait lock. In this case, there is a race condition where sdma_progress() can return false and the sdma_engine can idle. If that happens, there will be no more sdma interrupts to cause the wakeup and the user_sdma xmit will hang. Fix by moving the lock to enclose the sdma_progress() call. Also, delete busycount. The need for this was removed by: commit bcad29137a97 ("IB/hfi1: Serve the most starved iowait entry first") Ported to linux-4.9.y. Cc: <stable(a)vger.kernel.org> Fixes: 7724105686e7 ("IB/hfi1: add driver files") Reviewed-by: Gary Leshner <Gary.S.Leshner(a)intel.com> Signed-off-by: Mike Marciniszyn <mike.marciniszyn(a)intel.com> Signed-off-by: Dennis Dalessandro <dennis.dalessandro(a)intel.com> Signed-off-by: Jason Gunthorpe <jgg(a)mellanox.com> --- drivers/infiniband/hw/hfi1/user_sdma.c | 13 ++++--------- 1 file changed, 4 insertions(+), 9 deletions(-) diff --git a/drivers/infiniband/hw/hfi1/user_sdma.c b/drivers/infiniband/hw/hfi1/user_sdma.c index 4c11116..098296a 100644 --- a/drivers/infiniband/hw/hfi1/user_sdma.c +++ b/drivers/infiniband/hw/hfi1/user_sdma.c @@ -260,7 +260,6 @@ struct user_sdma_txreq { struct list_head list; struct user_sdma_request *req; u16 flags; - unsigned busycount; u64 seqnum; }; @@ -323,25 +322,22 @@ static int defer_packet_queue( struct hfi1_user_sdma_pkt_q *pq = container_of(wait, struct hfi1_user_sdma_pkt_q, busy); struct hfi1_ibdev *dev = &pq->dd->verbs_dev; - struct user_sdma_txreq *tx = - container_of(txreq, struct user_sdma_txreq, txreq); - if (sdma_progress(sde, seq, txreq)) { - if (tx->busycount++ < MAX_DEFER_RETRY_COUNT) - goto eagain; - } + write_seqlock(&dev->iowait_lock); + if (sdma_progress(sde, seq, txreq)) + goto eagain; /* * We are assuming that if the list is enqueued somewhere, it * is to the dmawait list since that is the only place where * it is supposed to be enqueued. */ xchg(&pq->state, SDMA_PKT_Q_DEFERRED); - write_seqlock(&dev->iowait_lock); if (list_empty(&pq->busy.list)) list_add_tail(&pq->busy.list, &sde->dmawait); write_sequnlock(&dev->iowait_lock); return -EBUSY; eagain: + write_sequnlock(&dev->iowait_lock); return -EAGAIN; } @@ -925,7 +921,6 @@ static int user_sdma_send_pkts(struct user_sdma_request *req, unsigned maxpkts) tx->flags = 0; tx->req = req; - tx->busycount = 0; INIT_LIST_HEAD(&tx->list); if (req->seqnum == req->info.npkts - 1)

4 years, 9 months

2
1
0 0

2024

2023

2022

2021

2020

2019

2018

2017

Linux-stable-mirror June 2019