- Linux-stable-mirror - lists.linaro.org

[PATCH v2] io_uring: fix filename leak in __io_openat_prep()

by Prithvi Tambewagh

__io_openat_prep() allocates a struct filename using getname(). However, for the condition of the file being installed in the fixed file table as well as having O_CLOEXEC flag set, the function returns early. At that point, the request doesn't have REQ_F_NEED_CLEANUP flag set. Due to this, the memory for the newly allocated struct filename is not cleaned up, causing a memory leak. Fix this by setting the REQ_F_NEED_CLEANUP for the request just after the successful getname() call, so that when the request is torn down, the filename will be cleaned up, along with other resources needing cleanup. Reported-by: syzbot+00e61c43eb5e4740438f(a)syzkaller.appspotmail.com Closes: https://syzkaller.appspot.com/bug?extid=00e61c43eb5e4740438f Tested-by: syzbot+00e61c43eb5e4740438f(a)syzkaller.appspotmail.com Cc: stable(a)vger.kernel.org Signed-off-by: Prithvi Tambewagh <activprithvi(a)gmail.com> --- io_uring/openclose.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/io_uring/openclose.c b/io_uring/openclose.c index bfeb91b31bba..15dde9bd6ff6 100644 --- a/io_uring/openclose.c +++ b/io_uring/openclose.c @@ -73,13 +73,13 @@ static int __io_openat_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe open->filename = NULL; return ret; } + req->flags |= REQ_F_NEED_CLEANUP; open->file_slot = READ_ONCE(sqe->file_index); if (open->file_slot && (open->how.flags & O_CLOEXEC)) return -EINVAL; open->nofile = rlimit(RLIMIT_NOFILE); - req->flags |= REQ_F_NEED_CLEANUP; if (io_openat_force_async(open)) req->flags |= REQ_F_FORCE_ASYNC; return 0; base-commit: b927546677c876e26eba308550207c2ddf812a43 -- 2.34.1

1 hour, 39 minutes

1
0
0 0

[PATCH] io_uring: fix filename leak in __io_openat_prep()

by Prithvi Tambewagh

__io_openat_prep() allocates a struct filename using getname(), but it isn't freed in case the present file is installed in the fixed file table and simultaneously, it has the flag O_CLOEXEC set in the open->how.flags field. This is an erroneous condition, since for a file installed in the fixed file table, it won't be installed in the normal file table, due to which the file cannot support close on exec. Earlier, the code just returned -EINVAL error code for this condition, however, the memory allocated for that struct filename wasn't freed, resulting in a memory leak. Hence, the case of file being installed in the fixed file table as well as having O_CLOEXEC flag in open->how.flags set, is adressed by using putname() to release the memory allocated to the struct filename, then setting the field open->filename to NULL, and after that, returning -EINVAL. Reported-by: syzbot+00e61c43eb5e4740438f(a)syzkaller.appspotmail.com Closes: https://syzkaller.appspot.com/bug?extid=00e61c43eb5e4740438f Tested-by: syzbot+00e61c43eb5e4740438f(a)syzkaller.appspotmail.com Cc: stable(a)vger.kernel.org Signed-off-by: Prithvi Tambewagh <activprithvi(a)gmail.com> --- io_uring/openclose.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/io_uring/openclose.c b/io_uring/openclose.c index bfeb91b31bba..fc190a3d8112 100644 --- a/io_uring/openclose.c +++ b/io_uring/openclose.c @@ -75,8 +75,11 @@ static int __io_openat_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe } open->file_slot = READ_ONCE(sqe->file_index); - if (open->file_slot && (open->how.flags & O_CLOEXEC)) + if (open->file_slot && (open->how.flags & O_CLOEXEC)) { + putname(open->filename); + open->filename = NULL; return -EINVAL; + } open->nofile = rlimit(RLIMIT_NOFILE); req->flags |= REQ_F_NEED_CLEANUP; base-commit: b927546677c876e26eba308550207c2ddf812a43 -- 2.34.1

1 hour, 59 minutes

3
2
0 0

[PATCH v2 0/2] hfsplus: avoid re-allocating header node 0 and stop on already-hashed nodes

by Shardul Bankar

Hi, syzbot reported a warning and crash when mounting a corrupted HFS+ image where the on-disk B-tree bitmap has node 0 (header node) marked free. In that case hfs_bmap_alloc() can try to allocate node 0 and reach hfs_bnode_create() with an already-hashed node number. Patch 1 prevents allocating the reserved header node (node 0) even if the bitmap is corrupted. Patch 2 follows Slava's review suggestion and changes the "already hashed" path in hfs_bnode_create() to return ERR_PTR(-EEXIST) instead of returning the existing node pointer, so we don't continue in a non-"business as usual" situation. v2 changes: - Implement Slava's suggestion: return ERR_PTR(-EEXIST) for already-hashed nodes. - Keep the node-0 allocation guard as a minimal, targeted hardening measure. Reported-by: syzbot+1c8ff72d0cd8a50dfeaa(a)syzkaller.appspotmail.com Link: https://syzkaller.appspot.com/bug?extid=1c8ff72d0cd8a50dfeaa Link: https://lore.kernel.org/all/20251213233215.368558-1-shardul.b@mpiricsoftwar… Shardul Bankar (2): hfsplus: skip node 0 in hfs_bmap_alloc hfsplus: return error when node already exists in hfs_bnode_create fs/hfsplus/bnode.c | 2 +- fs/hfsplus/btree.c | 3 +++ 2 files changed, 4 insertions(+), 1 deletion(-) -- 2.34.1

2 hours, 34 minutes

2
4
0 0

[PATCH 0/4] mm/damon/sysfs: free setup failures generated zombie sub-sub dirs

by SeongJae Park

Some DAMON sysfs directory setup functions generates its sub and sub-sub directories. For example, 'monitoring_attrs/' directory setup creates 'intervals/' and 'intervals/intervals_goal/' directories under 'monitoring_attrs/' directory. When such sub-sub directories are successfully made but followup setup is failed, the setup function should recursively clean up the subdirectories. However, such setup functions are only dereferencing sub directory reference counters. As a result, under certain setup failures, the sub-sub directories keep having non-zero reference counters. It means the directories cannot be removed like zombies, and the memory for the directories cannot be freed. The user impact of this issue is limited due to the following reasons. When the issue happens, the zombie directories are still taking the path. Hence attempts to generate the directories again will fail, without additional memory leak. This means the upper bound memory leak is limited. Nonetheless this also implies controlling DAMON with a feature that requires the setup-failed sysfs files will be impossible until the system reboots. Also, the setup operations are quite simple. The certain failures would hence only rarely happen, and are difficult to artificially trigger. SeongJae Park (4): mm/damon/sysfs: cleanup intervals subdirs on attrs dir setup failure mm/damon/sysfs: cleanup attrs subdirs on context dir setup failure mm/damon/sysfs-scheme: cleanup quotas subdirs on scheme dir setup failure mm/damon/sysfs-scheme: cleanup access_pattern subdirs on scheme dir setup failure mm/damon/sysfs-schemes.c | 10 ++++++---- mm/damon/sysfs.c | 9 ++++++--- 2 files changed, 12 insertions(+), 7 deletions(-) base-commit: 6d039da6a260dd7919bebc70ebb65d250bb9c24e -- 2.47.3

6 hours, 37 minutes

1
4
0 0

Fwd: Request to add mainline merged patch to stable kernels

by JP Dehollain

Hi Greg, thanks for looking into this.. The full commit hash is 807221d3c5ff6e3c91ff57bc82a0b7a541462e20 Note: apologies if you received this multiple times, the previous one got bounced due to html Cheers, JP ________________________________ From: Greg KH <gregkh(a)linuxfoundation.org> Sent: Tuesday, December 23, 2025 6:39:22 PM To: JP Dehollain <jpdehollain(a)gmail.com> Cc: stable(a)vger.kernel.org <stable(a)vger.kernel.org> Subject: Re: Request to add mainline merged patch to stable kernels On Tue, Dec 23, 2025 at 04:05:24PM +1100, JP Dehollain wrote: > Hello, > I recently used the patch misc: rtsx_pci: Add separate CD/WP pin > polarity reversal support with commit ID 807221d, to fix a bug causing > the cardreader driver to always load sd cards in read-only mode. > On the suggestion of the driver maintainer, I am requesting that this > patch be applied to all stable kernel versions, as it is currently > only applied to >=6.18. > Thanks, > JP > What is the git id of the commit you are looking to have backported? thanks, greg k-h

11 hours, 12 minutes

1
0
0 0

[PATCH] wifi: ath11k: fix qmi memory allocation logic for CALDB region

by Alexandru Gagniuc

Memory region assignment in ath11k_qmi_assign_target_mem_chunk() assumes that: 1. firmware will make a HOST_DDR_REGION_TYPE request, and 2. this request is processed before CALDB_MEM_REGION_TYPE In this case CALDB_MEM_REGION_TYPE, can safely be assigned immediately after the host region. However, if the HOST_DDR_REGION_TYPE request is not made, or the reserved-memory node is not present, then res.start and res.end are 0, and host_ddr_sz remains uninitialized. The physical address should fall back to ATH11K_QMI_CALDB_ADDRESS. That doesn't happen: resource_size(&res) returns 1 for an empty resource, and thus the if clause never takes the fallback path. ab->qmi.target_mem[idx].paddr is assigned the uninitialized value of host_ddr_sz + 0 (res.start). Use "if (res.end > res.start)" for the predicate, which correctly falls back to ATH11K_QMI_CALDB_ADDRESS. Fixes: 900730dc4705 ("wifi: ath: Use of_reserved_mem_region_to_resource() for "memory-region"") Cc: stable(a)vger.kernel.org # v6.18 Signed-off-by: Alexandru Gagniuc <mr.nuke.me(a)gmail.com> --- drivers/net/wireless/ath/ath11k/qmi.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/net/wireless/ath/ath11k/qmi.c b/drivers/net/wireless/ath/ath11k/qmi.c index aea56c38bf8f3..6cc26d1c1e2a4 100644 --- a/drivers/net/wireless/ath/ath11k/qmi.c +++ b/drivers/net/wireless/ath/ath11k/qmi.c @@ -2054,7 +2054,7 @@ static int ath11k_qmi_assign_target_mem_chunk(struct ath11k_base *ab) return ret; } - if (res.end - res.start + 1 < ab->qmi.target_mem[i].size) { + if (resource_size(&res) < ab->qmi.target_mem[i].size) { ath11k_dbg(ab, ATH11K_DBG_QMI, "fail to assign memory of sz\n"); return -EINVAL; @@ -2086,7 +2086,7 @@ static int ath11k_qmi_assign_target_mem_chunk(struct ath11k_base *ab) } if (ath11k_core_coldboot_cal_support(ab)) { - if (resource_size(&res)) { + if (res.end > res.start) { ab->qmi.target_mem[idx].paddr = res.start + host_ddr_sz; ab->qmi.target_mem[idx].iaddr = -- 2.45.1

11 hours, 19 minutes

4
5
0 0

[PATCH] ARM: dts: integrator: Fix DMA ranges mismatch warning on IM-PD1

by Kuan-Wei Chiu

When compiling the device tree for the Integrator/AP with IM-PD1, the following warning is observed regarding the display controller node: arch/arm/boot/dts/arm/integratorap-im-pd1.dts:251.3-14: Warning (dma_ranges_format): /bus@c0000000/bus@c0000000/display@1000000:dma-ranges: empty "dma-ranges" property but its #address-cells (2) differs from /bus@c0000000/bus@c0000000 (1) The display node specifies an empty "dma-ranges" property, intended to describe a 1:1 identity mapping. However, the node lacks explicit "#address-cells" and "#size-cells" properties. In this case, the device tree compiler defaults the address cells to 2 (64-bit), which conflicts with the parent bus configuration (32-bit, 1 cell). Fix this by explicitly defining "#address-cells" and "#size-cells" as 1. This matches the 32-bit architecture of the Integrator platform and ensures the address translation range is correctly parsed by the compiler. Fixes: 7bea67a99430 ("ARM: dts: integrator: Fix DMA ranges") Cc: stable(a)vger.kernel.org Signed-off-by: Kuan-Wei Chiu <visitorckw(a)gmail.com> --- arch/arm/boot/dts/arm/integratorap-im-pd1.dts | 2 ++ 1 file changed, 2 insertions(+) diff --git a/arch/arm/boot/dts/arm/integratorap-im-pd1.dts b/arch/arm/boot/dts/arm/integratorap-im-pd1.dts index db13e09f2fab..6d90d9a42dc9 100644 --- a/arch/arm/boot/dts/arm/integratorap-im-pd1.dts +++ b/arch/arm/boot/dts/arm/integratorap-im-pd1.dts @@ -248,6 +248,8 @@ display@1000000 { /* 640x480 16bpp @ 25.175MHz is 36827428 bytes/s */ max-memory-bandwidth = <40000000>; memory-region = <&impd1_ram>; + #address-cells = <1>; + #size-cells = <1>; dma-ranges; port@0 { -- 2.52.0.177.g9f829587af-goog

12 hours, 10 minutes

3
2
0 0

[PATCH v1 0/3] mm: mm_cid static initialization fixes

by Mathieu Desnoyers

Hi Andrew, Here are 2 fixes for missing mm_cid fields for init_mm and efi_mm static initialization. The renaming of cpu_bitmap to flexible_array (patch 2) is needed for patch 3. Those are relevant for mainline, with CC stable. They are based on v6.19-rc2. Cc: Andrew Morton <akpm(a)linux-foundation.org> Cc: Thomas Gleixner <tglx(a)linutronix.de> Cc: stable(a)vger.kernel.org Cc: linux-mm(a)kvack.org Mathieu Desnoyers (3): mm: Add missing static initializer for init_mm::mm_cid.lock mm: Rename cpu_bitmap field to flexible_array mm: Take into account mm_cid size for mm_struct static definitions drivers/firmware/efi/efi.c | 2 +- include/linux/mm_types.h | 18 +++++++++++++----- mm/init-mm.c | 5 ++++- 3 files changed, 18 insertions(+), 7 deletions(-) -- 2.39.5

15 hours, 33 minutes

1
3
0 0

[report] Performance regressions introduced via "cpuidle: menu: Remove iowait influence" on 6.12.y

by ALOK TIWARI

Hi, I’m reporting a performance regression of up to 6% sequential I/O vdbench regression observed on 6.12.y kernel. While running performance benchmarks on v6.12.60 kernel the sequential I/O vdbench metrics are showing a 5-6% performance regression when compared to v6.12.48 Bisect root cause commit ======================== - commit b39b62075ab4 ("cpuidle: menu: Remove iowait influence") Things work fine again when the previously removed performance-multiplier code is added back. Test details ============ The system is connected to a number of disks in disk array using multipathing and directio configuration in the vdbench profile. wd=wd1,sd=sd*,rdpct=0,seekpct=sequential,xfersize=128k rd=128k64T,wd=wd1,iorate=max,elapsed=600,interval=1,warmup=300,threads=64 Thanks, Alok

19 hours, 41 minutes

2
2
0 0

[PATCH v3] virtio: console: fix lost wakeup when device is written and polled

by Lorenz Bauer

A process issuing blocking writes to a virtio console may get stuck indefinitely if another thread polls the device. Here is how to trigger the bug: - Thread A writes to the port until the virtqueue is full. - Thread A calls wait_port_writable() and goes to sleep, waiting on port->waitqueue. - The host processes some of the write, marks buffers as used and raises an interrupt. - Before the interrupt is serviced, thread B executes port_fops_poll(). This calls reclaim_consumed_buffers() via will_write_block() and consumes all used buffers. - The interrupt is serviced. vring_interrupt() finds no used buffers via more_used() and returns without waking port->waitqueue. - Thread A is still in wait_event(port->waitqueue), waiting for a wakeup that never arrives. The crux is that invoking reclaim_consumed_buffers() may cause vring_interrupt() to omit wakeups. Fix this by calling reclaim_consumed_buffers() in out_int() before waking. This is similar to the call to discard_port_data() in in_intr() which also frees buffer from a non-sleepable context. This in turn guarantees that port->outvq_full is up to date when handling polling. Since in_intr() already populates port->inbuf we use that to avoid changing reader state. Cc: stable(a)vger.kernel.org Signed-off-by: Lorenz Bauer <lmb(a)isovalent.com> --- As far as I can tell all currently maintained stable series kernels need this commit. Applies and builds cleanly on 5.10.247, verified to fix the issue. --- Changes in v3: - Use spin_lock_irq in port_fops_poll (Arnd) - Use spin_lock in out_intr (Arnd) - Link to v2: https://lore.kernel.org/r/20251222-virtio-console-lost-wakeup-v2-1-5de93cb3… Changes in v2: - Call reclaim_consumed_buffers() in out_intr instead of issuing another wake. - Link to v1: https://lore.kernel.org/r/20251215-virtio-console-lost-wakeup-v1-1-79a5c578… --- drivers/char/virtio_console.c | 15 +++++++++++++-- 1 file changed, 13 insertions(+), 2 deletions(-) diff --git a/drivers/char/virtio_console.c b/drivers/char/virtio_console.c index 088182e54debd6029ea2c2a5542d7a28500e67b8..e6048e04c3b23d008caa2a1d31d4ac6b2841045f 100644 --- a/drivers/char/virtio_console.c +++ b/drivers/char/virtio_console.c @@ -971,10 +971,17 @@ static __poll_t port_fops_poll(struct file *filp, poll_table *wait) return EPOLLHUP; } ret = 0; - if (!will_read_block(port)) + + spin_lock_irq(&port->inbuf_lock); + if (port->inbuf) ret |= EPOLLIN | EPOLLRDNORM; - if (!will_write_block(port)) + spin_unlock_irq(&port->inbuf_lock); + + spin_lock_irq(&port->outvq_lock); + if (!port->outvq_full) ret |= EPOLLOUT; + spin_unlock_irq(&port->outvq_lock); + if (!port->host_connected) ret |= EPOLLHUP; @@ -1705,6 +1712,10 @@ static void out_intr(struct virtqueue *vq) return; } + spin_lock(&port->outvq_lock); + reclaim_consumed_buffers(port); + spin_unlock(&port->outvq_lock); + wake_up_interruptible(&port->waitqueue); } --- base-commit: d358e5254674b70f34c847715ca509e46eb81e6f change-id: 20251215-virtio-console-lost-wakeup-0f566c5cd35f Best regards, -- Lorenz Bauer <lmb(a)isovalent.com>

22 hours, 12 minutes

1
0
0 0

[PATCH RESEND v5 0/3] phy: qcom: edp: Add missing ref clock to x1e80100

by Abel Vesa

According to documentation, the DP PHY on x1e80100 has another clock called ref. The current X Elite devices supported upstream work fine without this clock, because the boot firmware leaves this clock enabled. But we should not rely on that. Also, when it comes to power management, this clock needs to be also disabled on suspend. So even though this change breaks the ABI, it is needed in order to make we disable this clock on runtime PM, when that is going to be enabled in the driver. So rework the driver to allow different number of clocks, fix the dt-bindings schema and add the clock to the DT node as well. Signed-off-by: Abel Vesa <abel.vesa(a)linaro.org> --- Changes in v5: - Picked-up Bjorn's R-b tags. - Replaced "parse" with "get" on clocks acquiring failure. - Link to v4: https://lore.kernel.org/r/20251029-phy-qcom-edp-add-missing-refclk-v4-0-adb… Changes in v4: - Picked Dmitry's R-b tag for the driver patch - Added x1e80100 substring to subject of dts patch - Link to v3 (resend): https://lore.kernel.org/r/20251014-phy-qcom-edp-add-missing-refclk-v3-0-078… Changes in v3 (resend) - picked-up Krzysztof's R-b tag for bindings patch - Link to v3: https://lore.kernel.org/r/20250909-phy-qcom-edp-add-missing-refclk-v3-0-4ec… Changes in v3: - Use dev_err_probe() on clocks parsing failure. - Explain why the ABI break is necessary. - Drop the extra 'clk' suffix from the clock name. So ref instead of refclk. - Link to v2: https://lore.kernel.org/r/20250903-phy-qcom-edp-add-missing-refclk-v2-0-d88… Changes in v2: - Fix schema by adding the minItems, as suggested by Krzysztof. - Use devm_clk_bulk_get_all, as suggested by Konrad. - Rephrase the commit messages to reflect the flexible number of clocks. - Link to v1: https://lore.kernel.org/r/20250730-phy-qcom-edp-add-missing-refclk-v1-0-6f7… --- Abel Vesa (3): dt-bindings: phy: qcom-edp: Add missing clock for X Elite phy: qcom: edp: Make the number of clocks flexible arm64: dts: qcom: x1e80100: Add missing TCSR ref clock to the DP PHYs .../devicetree/bindings/phy/qcom,edp-phy.yaml | 28 +++++++++++++++++++++- arch/arm64/boot/dts/qcom/hamoa.dtsi | 12 ++++++---- drivers/phy/qualcomm/phy-qcom-edp.c | 16 ++++++------- 3 files changed, 43 insertions(+), 13 deletions(-) --- base-commit: 131f3d9446a6075192cdd91f197989d98302faa6 change-id: 20250730-phy-qcom-edp-add-missing-refclk-5ab82828f8e7 Best regards, -- Abel Vesa <abel.vesa(a)oss.qualcomm.com>

22 hours, 14 minutes

1
3
0 0

[PATCH v6.1 0/2] x86/mm/pat: fix VM_PAT handling

by Ajay Kaher

[PATCH v6.1 1/2]: required to backport [PATCH v6.1 2/2]. [PATCH v6.1 2/2]: backport to fix VM_PAT handling when fork() fails in copy_page_range(). arch/x86/mm/pat/memtype.c | 50 +++++++++++++++++++++++---------------- include/linux/pgtable.h | 31 ++++++++++++++++++------ kernel/fork.c | 4 ++++ mm/memory.c | 10 ++++---- mm/mremap.c | 2 +- 5 files changed, 62 insertions(+), 35 deletions(-) -- 2.40.4

22 hours, 43 minutes

1
2
0 0

[PATCH] usb: cdns3: fix a null pointer dereference in cdns3_gadget_ep_queue()

by Haoxiang Li

If cdns3_gadget_ep_alloc_request() fails, a null pointer dereference occurs. Add a check to prevent it. Found by code review and compiled on ubuntu 20.04. Fixes: 7733f6c32e36 ("usb: cdns3: Add Cadence USB3 DRD Driver") Cc: stable(a)vger.kernel.org Signed-off-by: Haoxiang Li <lihaoxiang(a)isrc.iscas.ac.cn> --- drivers/usb/cdns3/cdns3-gadget.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/usb/cdns3/cdns3-gadget.c b/drivers/usb/cdns3/cdns3-gadget.c index 168707213ed9..cc2421a34588 100644 --- a/drivers/usb/cdns3/cdns3-gadget.c +++ b/drivers/usb/cdns3/cdns3-gadget.c @@ -2659,6 +2659,8 @@ static int cdns3_gadget_ep_queue(struct usb_ep *ep, struct usb_request *request, struct cdns3_request *priv_req; zlp_request = cdns3_gadget_ep_alloc_request(ep, GFP_ATOMIC); + if (!zlp_request) + return -ENOMEM; zlp_request->buf = priv_dev->zlp_buf; zlp_request->length = 0; -- 2.25.1

23 hours, 21 minutes

2
1
0 0

[PATCH] usb: cdns2: fix a null pointer dereference in cdns2_gadget_ep_queue()

by Haoxiang Li

If cdns2_gadget_ep_alloc_request() fails, a null pointer dereference occurs. Add a check to prevent it. Fixes: 3eb1f1efe204 ("usb: cdns2: Add main part of Cadence USBHS driver") Cc: stable(a)vger.kernel.org Signed-off-by: Haoxiang Li <lihaoxiang(a)isrc.iscas.ac.cn> --- drivers/usb/gadget/udc/cdns2/cdns2-gadget.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/usb/gadget/udc/cdns2/cdns2-gadget.c b/drivers/usb/gadget/udc/cdns2/cdns2-gadget.c index 9b53daf76583..c5b9dae743d8 100644 --- a/drivers/usb/gadget/udc/cdns2/cdns2-gadget.c +++ b/drivers/usb/gadget/udc/cdns2/cdns2-gadget.c @@ -1725,6 +1725,8 @@ static int cdns2_gadget_ep_queue(struct usb_ep *ep, struct usb_request *request, struct cdns2_request *preq; zlp_request = cdns2_gadget_ep_alloc_request(ep, GFP_ATOMIC); + if (!zlp_request) + return -ENOMEM; zlp_request->buf = pdev->zlp_buf; zlp_request->length = 0; -- 2.25.1

23 hours, 22 minutes

2
1
0 0

[PATCH] drm/msm/dpu: Add missing NULL pointer check for pingpong interface

by Nikolay Kuratov

It is checked almost always in dpu_encoder_phys_wb_setup_ctl(), but in a single place the check is missing. Also use convenient locals instead of phys_enc->* where available. Cc: stable(a)vger.kernel.org Fixes: d7d0e73f7de33 ("drm/msm/dpu: introduce the dpu_encoder_phys_* for writeback") Signed-off-by: Nikolay Kuratov <kniv(a)yandex-team.ru> --- drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_wb.c | 10 ++++------ 1 file changed, 4 insertions(+), 6 deletions(-) diff --git a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_wb.c b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_wb.c index 46f348972a97..6d28f2281c76 100644 --- a/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_wb.c +++ b/drivers/gpu/drm/msm/disp/dpu1/dpu_encoder_phys_wb.c @@ -247,14 +247,12 @@ static void dpu_encoder_phys_wb_setup_ctl(struct dpu_encoder_phys *phys_enc) if (hw_cdm) intf_cfg.cdm = hw_cdm->idx; - if (phys_enc->hw_pp->merge_3d && phys_enc->hw_pp->merge_3d->ops.setup_3d_mode) - phys_enc->hw_pp->merge_3d->ops.setup_3d_mode(phys_enc->hw_pp->merge_3d, - mode_3d); + if (hw_pp && hw_pp->merge_3d && hw_pp->merge_3d->ops.setup_3d_mode) + hw_pp->merge_3d->ops.setup_3d_mode(hw_pp->merge_3d, mode_3d); /* setup which pp blk will connect to this wb */ - if (hw_pp && phys_enc->hw_wb->ops.bind_pingpong_blk) - phys_enc->hw_wb->ops.bind_pingpong_blk(phys_enc->hw_wb, - phys_enc->hw_pp->idx); + if (hw_pp && hw_wb->ops.bind_pingpong_blk) + hw_wb->ops.bind_pingpong_blk(hw_wb, hw_pp->idx); phys_enc->hw_ctl->ops.setup_intf_cfg(phys_enc->hw_ctl, &intf_cfg); } else if (phys_enc->hw_ctl && phys_enc->hw_ctl->ops.setup_intf_cfg) { -- 2.34.1

23 hours, 40 minutes

2
3
0 0

[PATCH V2 0/2] LoongArch: KVM: fix "unreliable stack" issue

by Xianglai Li

When starting multi-core loongarch virtualization on loongarch physical machine, loading livepatch on the physical machine will cause an error similar to the following: [ 411.686289] livepatch: klp_try_switch_task: CPU 31/KVM:3116 has an unreliable stack The specific test steps are as follows: 1.Start a multi-core virtual machine on a physical machine 2.Enter the following command on the physical machine to turn on the debug switch: echo "file kernel/livepatch/transition.c +p" > /sys/kernel/debug/\ dynamic_debug/control 3.Load livepatch: modprobe livepatch-sample Through the above steps, similar prints can be viewed in dmesg. The reason for this issue is that the code of the kvm_exc_entry function was copied in the function kvm_loongarch_env_init. When the cpu needs to execute kvm_exc_entry, it will switch to the copied address for execution. The new address of the kvm_exc_entry function cannot be recognized in ORC, which eventually leads to the arch_stack_walk_reliable function returning an error and printing an exception message. To solve the above problems, we directly compile the switch.S file into the kernel instead of the module. In this way, the function kvm_exc_entry will no longer need to be copied. changlog: V2<-V1: 1.Rollback the modification of function parameter types such as kvm_save_fpu. In the asm-prototypes.h header file, only the parameter types it depends on are included Cc: Huacai Chen <chenhuacai(a)kernel.org> Cc: WANG Xuerui <kernel(a)xen0n.name> Cc: Tianrui Zhao <zhaotianrui(a)loongson.cn> Cc: Bibo Mao <maobibo(a)loongson.cn> Cc: Charlie Jenkins <charlie(a)rivosinc.com> Cc: Xianglai Li <lixianglai(a)loongson.cn> Cc: Thomas Gleixner <tglx(a)linutronix.de> Cc: stable(a)vger.kernel.org Xianglai Li (2): LoongArch: KVM: Compile the switch.S file directly into the kernel LoongArch: KVM: fix "unreliable stack" issue arch/loongarch/Kbuild | 2 +- arch/loongarch/include/asm/asm-prototypes.h | 21 +++++++++++++ arch/loongarch/include/asm/kvm_host.h | 3 -- arch/loongarch/kvm/Makefile | 2 +- arch/loongarch/kvm/main.c | 35 ++------------------- arch/loongarch/kvm/switch.S | 24 +++++++++++--- 6 files changed, 45 insertions(+), 42 deletions(-) base-commit: 8f0b4cce4481fb22653697cced8d0d04027cb1e8 -- 2.39.1

23 hours, 54 minutes

4
12
0 0

[PATCH] drm/vmwgfx: Fix an error return check in vmw_compat_shader_add()

by Haoxiang Li

In vmw_compat_shader_add(), the return value check of vmw_shader_alloc() is not proper. Modify the check for the return pointer 'res'. Found by code review and compiled on ubuntu 20.04. Fixes: 18e4a4669c50 ("drm/vmwgfx: Fix compat shader namespace") Cc: stable(a)vger.kernel.org Signed-off-by: Haoxiang Li <lihaoxiang(a)isrc.iscas.ac.cn> --- drivers/gpu/drm/vmwgfx/vmwgfx_shader.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_shader.c b/drivers/gpu/drm/vmwgfx/vmwgfx_shader.c index 69dfe69ce0f8..7ed938710342 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_shader.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_shader.c @@ -923,8 +923,10 @@ int vmw_compat_shader_add(struct vmw_private *dev_priv, ttm_bo_unreserve(&buf->tbo); res = vmw_shader_alloc(dev_priv, buf, size, 0, shader_type); - if (unlikely(ret != 0)) + if (IS_ERR(res)) { + ret = PTR_ERR(res); goto no_reserve; + } ret = vmw_cmdbuf_res_add(man, vmw_cmdbuf_res_shader, vmw_shader_key(user_key, shader_type), -- 2.25.1

23 hours, 56 minutes

1
0
0 0

[PATCH] media: iris: Add missing platform data entries for SM8750

by Dikshita Agarwal

Two platform-data fields for SM8750 were missed: - get_vpu_buffer_size = iris_vpu33_buf_size Without this, the driver fails to allocate the required internal buffers, leading to basic decode/encode failures during session bring-up. - max_core_mbps = ((7680 * 4320) / 256) * 60 Without this capability exposed, capability checks are incomplete and v4l2-compliance for encoder fails. Fixes: a5925a2ce077 ("media: iris: add VPU33 specific encoding buffer calculation") Fixes: a6882431a138 ("media: iris: Add support for ENUM_FRAMESIZES/FRAMEINTERVALS for encoder") Cc: stable(a)vger.kernel.org Signed-off-by: Dikshita Agarwal <dikshita.agarwal(a)oss.qualcomm.com> --- drivers/media/platform/qcom/iris/iris_platform_gen2.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/media/platform/qcom/iris/iris_platform_gen2.c b/drivers/media/platform/qcom/iris/iris_platform_gen2.c index c1989240c248601c34b84f508f1b72d72f81260a..00d1d55463179429f2ae7554546dcbe4fb1431cc 100644 --- a/drivers/media/platform/qcom/iris/iris_platform_gen2.c +++ b/drivers/media/platform/qcom/iris/iris_platform_gen2.c @@ -915,6 +915,7 @@ const struct iris_platform_data sm8750_data = { .get_instance = iris_hfi_gen2_get_instance, .init_hfi_command_ops = iris_hfi_gen2_command_ops_init, .init_hfi_response_ops = iris_hfi_gen2_response_ops_init, + .get_vpu_buffer_size = iris_vpu33_buf_size, .vpu_ops = &iris_vpu35_ops, .set_preset_registers = iris_set_sm8550_preset_registers, .icc_tbl = sm8550_icc_table, @@ -945,6 +946,7 @@ const struct iris_platform_data sm8750_data = { .num_vpp_pipe = 4, .max_session_count = 16, .max_core_mbpf = NUM_MBS_8K * 2, + .max_core_mbps = ((7680 * 4320) / 256) * 60, .dec_input_config_params_default = sm8550_vdec_input_config_params_default, .dec_input_config_params_default_size = --- base-commit: 8f0b4cce4481fb22653697cced8d0d04027cb1e8 change-id: 20251217-iris-sm8750-fix-2d42a133058d Best regards, -- Dikshita Agarwal <dikshita.agarwal(a)oss.qualcomm.com>

1 day

3
2
0 0

[PATCH v5.10-v6.1] drm/vmwgfx: Fix a null-ptr access in the cursor snooper

by Shivani Agarwal

From: Zack Rusin <zack.rusin(a)broadcom.com> [ Upstream commit 5ac2c0279053a2c5265d46903432fb26ae2d0da2 ] Check that the resource which is converted to a surface exists before trying to use the cursor snooper on it. vmw_cmd_res_check allows explicit invalid (SVGA3D_INVALID_ID) identifiers because some svga commands accept SVGA3D_INVALID_ID to mean "no surface", unfortunately functions that accept the actual surfaces as objects might (and in case of the cursor snooper, do not) be able to handle null objects. Make sure that we validate not only the identifier (via the vmw_cmd_res_check) but also check that the actual resource exists before trying to do something with it. Fixes unchecked null-ptr reference in the snooping code. Signed-off-by: Zack Rusin <zack.rusin(a)broadcom.com> Fixes: c0951b797e7d ("drm/vmwgfx: Refactor resource management") Reported-by: Kuzey Arda Bulut <kuzeyardabulut(a)gmail.com> Cc: Broadcom internal kernel review list <bcm-kernel-feedback-list(a)broadcom.com> Cc: dri-devel(a)lists.freedesktop.org Reviewed-by: Ian Forbes <ian.forbes(a)broadcom.com> Link: https://lore.kernel.org/r/20250917153655.1968583-1-zack.rusin@broadcom.com Signed-off-by: Sasha Levin <sashal(a)kernel.org> [Shivani: Modified to apply on v5.10.y-v6.1.y] Signed-off-by: Shivani Agarwal <shivani.agarwal(a)broadcom.com> --- drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c | 17 ++++++++++++----- 1 file changed, 12 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c index 0d12d6af6..5d3827b5d 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c @@ -1507,6 +1507,7 @@ static int vmw_cmd_dma(struct vmw_private *dev_priv, SVGA3dCmdHeader *header) { struct vmw_buffer_object *vmw_bo = NULL; + struct vmw_resource *res; struct vmw_surface *srf = NULL; VMW_DECLARE_CMD_VAR(*cmd, SVGA3dCmdSurfaceDMA); int ret; @@ -1542,18 +1543,24 @@ static int vmw_cmd_dma(struct vmw_private *dev_priv, dirty = (cmd->body.transfer == SVGA3D_WRITE_HOST_VRAM) ? VMW_RES_DIRTY_SET : 0; - ret = vmw_cmd_res_check(dev_priv, sw_context, vmw_res_surface, - dirty, user_surface_converter, - &cmd->body.host.sid, NULL); + ret = vmw_cmd_res_check(dev_priv, sw_context, vmw_res_surface, dirty, + user_surface_converter, &cmd->body.host.sid, + NULL); if (unlikely(ret != 0)) { if (unlikely(ret != -ERESTARTSYS)) VMW_DEBUG_USER("could not find surface for DMA.\n"); return ret; } - srf = vmw_res_to_srf(sw_context->res_cache[vmw_res_surface].res); + res = sw_context->res_cache[vmw_res_surface].res; + if (!res) { + VMW_DEBUG_USER("Invalid DMA surface.\n"); + return -EINVAL; + } - vmw_kms_cursor_snoop(srf, sw_context->fp->tfile, &vmw_bo->base, header); + srf = vmw_res_to_srf(res); + vmw_kms_cursor_snoop(srf, sw_context->fp->tfile, &vmw_bo->base, + header); return 0; } -- 2.40.4

1 day

1
0
0 0

[PATCH 1/1] phy: tegra: xusb: Explicitly configure HS_DISCON_LEVEL to 0x7

by Wayne Chang

The USB2 Bias Pad Control register manages analog parameters for signal detection. Previously, the HS_DISCON_LEVEL relied on hardware reset values, which may lead to the detection failure. Explicitly configure HS_DISCON_LEVEL to 0x7. This ensures the disconnect threshold is sufficient to guarantee reliable detection. Fixes: bbf711682cd5 ("phy: tegra: xusb: Add Tegra186 support") Cc: stable(a)vger.kernel.org Signed-off-by: Wayne Chang <waynec(a)nvidia.com> --- drivers/phy/tegra/xusb-tegra186.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/phy/tegra/xusb-tegra186.c b/drivers/phy/tegra/xusb-tegra186.c index e818f6c3980e..bec9616c4a2e 100644 --- a/drivers/phy/tegra/xusb-tegra186.c +++ b/drivers/phy/tegra/xusb-tegra186.c @@ -84,6 +84,7 @@ #define XUSB_PADCTL_USB2_BIAS_PAD_CTL0 0x284 #define BIAS_PAD_PD BIT(11) #define HS_SQUELCH_LEVEL(x) (((x) & 0x7) << 0) +#define HS_DISCON_LEVEL(x) (((x) & 0x7) << 3) #define XUSB_PADCTL_USB2_BIAS_PAD_CTL1 0x288 #define USB2_TRK_START_TIMER(x) (((x) & 0x7f) << 12) @@ -623,6 +624,8 @@ static void tegra186_utmi_bias_pad_power_on(struct tegra_xusb_padctl *padctl) value &= ~BIAS_PAD_PD; value &= ~HS_SQUELCH_LEVEL(~0); value |= HS_SQUELCH_LEVEL(priv->calib.hs_squelch); + value &= ~HS_DISCON_LEVEL(~0); + value |= HS_DISCON_LEVEL(0x7); padctl_writel(padctl, value, XUSB_PADCTL_USB2_BIAS_PAD_CTL0); udelay(1); -- 2.25.1

1 day, 1 hour

2
1
0 0

[PATCH] ocfs2: Add check for total number of chains in chain list

by Prithvi Tambewagh

The functions ocfs2_reserve_suballoc_bits(), ocfs2_block_group_alloc(), ocfs2_block_group_alloc_contig() and ocfs2_find_smallest_chain() trust the on-disk values related to the allocation chain. However, KASAN bug was triggered in these functions, and the kernel panicked when accessing redzoned memory. This occurred due to the corrupted value of `cl_count` field of `struct ocfs2_chain_list`. Upon analysis, the value of `cl_count` was observed to be overwhemingly large, due to which the code accessed redzoned memory. The fix introduces an if statement which validates value of `cl_count` (both lower and upper bounds). Lower bound check ensures the value of `cl_count` is not zero and upper bound check ensures that the value of `cl_count` is in the range such that it has a value less than the total size of struct ocfs2_chain_list and maximum number of chains that can be present, so as to fill one block. Reported-by: syzbot+af14efe17dfa46173239(a)syzkaller.appspotmail.com Closes: https://syzkaller.appspot.com/bug?extid=af14efe17dfa46173239 Tested-by: syzbot+af14efe17dfa46173239(a)syzkaller.appspotmail.com Cc: stable(a)vger.kernel.org Signed-off-by: Prithvi Tambewagh <activprithvi(a)gmail.com> --- fs/ocfs2/suballoc.c | 15 +++++++++++++++ 1 file changed, 15 insertions(+) diff --git a/fs/ocfs2/suballoc.c b/fs/ocfs2/suballoc.c index f7b483f0de2a..7ea63e9cc4f8 100644 --- a/fs/ocfs2/suballoc.c +++ b/fs/ocfs2/suballoc.c @@ -671,6 +671,21 @@ static int ocfs2_block_group_alloc(struct ocfs2_super *osb, BUG_ON(ocfs2_is_cluster_bitmap(alloc_inode)); cl = &fe->id2.i_chain; + unsigned int block_size = osb->sb->s_blocksize; + unsigned int max_cl_count = + (block_size - offsetof(struct ocfs2_chain_list, cl_recs)) / + sizeof(struct ocfs2_chain_rec); + + if (!le16_to_cpu(cl->cl_count) || + le16_to_cpu(cl->cl_count) > max_cl_count) { + ocfs2_error(osb->sb, + "Invalid chain list: cl_count %u " + "exceeds max %u", + le16_to_cpu(cl->cl_count), max_cl_count); + status = -EIO; + goto bail; + } + status = ocfs2_reserve_clusters_with_limit(osb, le16_to_cpu(cl->cl_cpg), max_block, flags, &ac); base-commit: 36c254515dc6592c44db77b84908358979dd6b50 -- 2.34.1

1 day, 2 hours

3
2
0 0

[PATCH] platform/x86: intel_telemetry: Fix swapped arrays in PSS output

by Kaushlendra Kumar

The LTR blocking statistics and wakeup event counters are incorrectly cross-referenced during debugfs output rendering. The code populates pss_ltr_blkd[] with LTR blocking data and pss_s0ix_wakeup[] with wakeup data, but the display loops reference the wrong arrays. This causes the "LTR Blocking Status" section to print wakeup events and the "Wakes Status" section to print LTR blockers, misleading power management analysis and S0ix residency debugging. Fix by aligning array usage with the intended output section labels. Fixes: 87bee290998d ("platform:x86: Add Intel Telemetry Debugfs interfaces") Cc: stable(a)vger.kernel.org Signed-off-by: Kaushlendra Kumar <kaushlendra.kumar(a)intel.com> --- drivers/platform/x86/intel/telemetry/debugfs.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/platform/x86/intel/telemetry/debugfs.c b/drivers/platform/x86/intel/telemetry/debugfs.c index 70e5736c44c7..189c61ff7ff0 100644 --- a/drivers/platform/x86/intel/telemetry/debugfs.c +++ b/drivers/platform/x86/intel/telemetry/debugfs.c @@ -449,7 +449,7 @@ static int telem_pss_states_show(struct seq_file *s, void *unused) for (index = 0; index < debugfs_conf->pss_ltr_evts; index++) { seq_printf(s, "%-32s\t%u\n", debugfs_conf->pss_ltr_data[index].name, - pss_s0ix_wakeup[index]); + pss_ltr_blkd[index]); } seq_puts(s, "\n--------------------------------------\n"); @@ -459,7 +459,7 @@ static int telem_pss_states_show(struct seq_file *s, void *unused) for (index = 0; index < debugfs_conf->pss_wakeup_evts; index++) { seq_printf(s, "%-32s\t%u\n", debugfs_conf->pss_wakeup[index].name, - pss_ltr_blkd[index]); + pss_s0ix_wakeup[index]); } return 0; -- 2.34.1

1 day, 5 hours

1
0
0 0

[PATCH v2 0/2] iommu/vt-d: Skip dev-iotlb flush for inaccessible PCIe device

by Jinhui Guo

Hi, all We hit hard-lockups when the Intel IOMMU waits indefinitely for an ATS invalidation that cannot complete, especially under GDR high-load conditions. 1. Hard-lock when a passthrough PCIe NIC with ATS enabled link-down in Intel IOMMU non-scalable mode. Two scenarios exist: NIC link-down with an explicit link-down event and link-down without any event. a) NIC link-down with an explicit link-dow event. Call Trace: qi_submit_sync qi_flush_dev_iotlb __context_flush_dev_iotlb.part.0 domain_context_clear_one_cb pci_for_each_dma_alias device_block_translation blocking_domain_attach_dev iommu_deinit_device __iommu_group_remove_device iommu_release_device iommu_bus_notifier blocking_notifier_call_chain bus_notify device_del pci_remove_bus_device pci_stop_and_remove_bus_device pciehp_unconfigure_device pciehp_disable_slot pciehp_handle_presence_or_link_change pciehp_ist b) NIC link-down without an event - hard-lock on VM destroy. Call Trace: qi_submit_sync qi_flush_dev_iotlb __context_flush_dev_iotlb.part.0 domain_context_clear_one_cb pci_for_each_dma_alias device_block_translation blocking_domain_attach_dev __iommu_attach_device __iommu_device_set_domain __iommu_group_set_domain_internal iommu_detach_group vfio_iommu_type1_detach_group vfio_group_detach_container vfio_group_fops_release __fput 2. Hard-lock when a passthrough PCIe NIC with ATS enabled link-down in Intel IOMMU scalable mode; NIC link-down without an event hard-locks on VM destroy. Call Trace: qi_submit_sync qi_flush_dev_iotlb intel_pasid_tear_down_entry device_block_translation blocking_domain_attach_dev __iommu_attach_device __iommu_device_set_domain __iommu_group_set_domain_internal iommu_detach_group vfio_iommu_type1_detach_group vfio_group_detach_container vfio_group_fops_release __fput Fix both issues with two patches: 1. Skip dev-IOTLB flush for inaccessible devices in __context_flush_dev_iotlb() using pci_device_is_present(). 2. Use pci_device_is_present() instead of pci_dev_is_disconnected() to decide when to skip ATS invalidation in devtlb_invalidation_with_pasid(). Best Regards, Jinhui --- v1: https://lore.kernel.org/all/20251210171431.1589-1-guojinhui.liam@bytedance.… Changelog in v1 -> v2 (suggested by Baolu Lu) - Simplify the pci_device_is_present() check in __context_flush_dev_iotlb(). - Add Cc: stable(a)vger.kernel.org to both patches. Jinhui Guo (2): iommu/vt-d: Skip dev-iotlb flush for inaccessible PCIe device without scalable mode iommu/vt-d: Flush dev-IOTLB only when PCIe device is accessible in scalable mode drivers/iommu/intel/pasid.c | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-) -- 2.20.1

1 day, 5 hours

3
7
0 0

[PATCH 2/5] x86, fpu: separate fpstate->xfd and guest XFD

by Paolo Bonzini

Until now, fpstate->xfd has acted as both the guest value and the value that the host used when executing XSAVES and XRSTORS. This is wrong: the data in the guest's FPU might not be initialized even if a bit is set in XFD and, when that happens, XRSTORing the guest FPU will fail with a #NM exception *on the host*. Instead, store the value of XFD together with XFD_ERR in struct fpu_guest; it will still be synchronized in fpu_load_guest_fpstate(), but the XRSTOR(S) operation will be able to load any valid state of the FPU independent of the XFD value. Cc: stable(a)vger.kernel.org Fixes: 820a6ee944e7 ("kvm: x86: Add emulation for IA32_XFD", 2022-01-14) Signed-off-by: Paolo Bonzini <pbonzini(a)redhat.com> --- arch/x86/include/asm/fpu/api.h | 6 ++---- arch/x86/include/asm/fpu/types.h | 7 +++++++ arch/x86/kernel/fpu/core.c | 19 ++++--------------- arch/x86/kernel/fpu/xstate.h | 18 ++++++++++-------- arch/x86/kvm/x86.c | 6 +++--- 5 files changed, 26 insertions(+), 30 deletions(-) diff --git a/arch/x86/include/asm/fpu/api.h b/arch/x86/include/asm/fpu/api.h index 0820b2621416..ee9ba06b7dbe 100644 --- a/arch/x86/include/asm/fpu/api.h +++ b/arch/x86/include/asm/fpu/api.h @@ -152,11 +152,9 @@ extern int fpu_swap_kvm_fpstate(struct fpu_guest *gfpu, bool enter_guest); extern int fpu_enable_guest_xfd_features(struct fpu_guest *guest_fpu, u64 xfeatures); #ifdef CONFIG_X86_64 -extern void fpu_update_guest_xfd(struct fpu_guest *guest_fpu, u64 xfd); -extern void fpu_sync_guest_vmexit_xfd_state(void); +extern void fpu_sync_guest_vmexit_xfd_state(struct fpu_guest *gfpu); #else -static inline void fpu_update_guest_xfd(struct fpu_guest *guest_fpu, u64 xfd) { } -static inline void fpu_sync_guest_vmexit_xfd_state(void) { } +static inline void fpu_sync_guest_vmexit_xfd_state(struct fpu_guest *gfpu) { } #endif extern void fpu_copy_guest_fpstate_to_uabi(struct fpu_guest *gfpu, void *buf, diff --git a/arch/x86/include/asm/fpu/types.h b/arch/x86/include/asm/fpu/types.h index 93e99d2583d6..7abe231e2ffe 100644 --- a/arch/x86/include/asm/fpu/types.h +++ b/arch/x86/include/asm/fpu/types.h @@ -545,6 +545,13 @@ struct fpu_guest { */ u64 xfeatures; + /* + * @xfd: Save the guest value. Note that this is + * *not* fpstate->xfd, which is the value + * the host uses when doing XSAVE/XRSTOR. + */ + u64 xfd; + /* * @xfd_err: Save the guest value. */ diff --git a/arch/x86/kernel/fpu/core.c b/arch/x86/kernel/fpu/core.c index a480fa8c65d5..ff17c96d290a 100644 --- a/arch/x86/kernel/fpu/core.c +++ b/arch/x86/kernel/fpu/core.c @@ -317,16 +317,6 @@ int fpu_enable_guest_xfd_features(struct fpu_guest *guest_fpu, u64 xfeatures) EXPORT_SYMBOL_FOR_KVM(fpu_enable_guest_xfd_features); #ifdef CONFIG_X86_64 -void fpu_update_guest_xfd(struct fpu_guest *guest_fpu, u64 xfd) -{ - fpregs_lock(); - guest_fpu->fpstate->xfd = xfd; - if (guest_fpu->fpstate->in_use) - xfd_update_state(guest_fpu->fpstate); - fpregs_unlock(); -} -EXPORT_SYMBOL_FOR_KVM(fpu_update_guest_xfd); - /** * fpu_sync_guest_vmexit_xfd_state - Synchronize XFD MSR and software state * @@ -339,14 +329,12 @@ EXPORT_SYMBOL_FOR_KVM(fpu_update_guest_xfd); * Note: It can be invoked unconditionally even when write emulation is * enabled for the price of a then pointless MSR read. */ -void fpu_sync_guest_vmexit_xfd_state(void) +void fpu_sync_guest_vmexit_xfd_state(struct fpu_guest *gfpu) { - struct fpstate *fpstate = x86_task_fpu(current)->fpstate; - lockdep_assert_irqs_disabled(); if (fpu_state_size_dynamic()) { - rdmsrq(MSR_IA32_XFD, fpstate->xfd); - __this_cpu_write(xfd_state, fpstate->xfd); + rdmsrq(MSR_IA32_XFD, gfpu->xfd); + __this_cpu_write(xfd_state, gfpu->xfd); } } EXPORT_SYMBOL_FOR_KVM(fpu_sync_guest_vmexit_xfd_state); @@ -890,6 +878,7 @@ void fpu_load_guest_fpstate(struct fpu_guest *gfpu) fpregs_restore_userregs(); fpregs_assert_state_consistent(); + xfd_set_state(gfpu->xfd); if (gfpu->xfd_err) wrmsrq(MSR_IA32_XFD_ERR, gfpu->xfd_err); } diff --git a/arch/x86/kernel/fpu/xstate.h b/arch/x86/kernel/fpu/xstate.h index 52ce19289989..c0ce05bee637 100644 --- a/arch/x86/kernel/fpu/xstate.h +++ b/arch/x86/kernel/fpu/xstate.h @@ -180,26 +180,28 @@ static inline void xfd_validate_state(struct fpstate *fpstate, u64 mask, bool rs #endif #ifdef CONFIG_X86_64 -static inline void xfd_set_state(u64 xfd) +static inline void __xfd_set_state(u64 xfd) { wrmsrq(MSR_IA32_XFD, xfd); __this_cpu_write(xfd_state, xfd); } +static inline void xfd_set_state(u64 xfd) +{ + if (__this_cpu_read(xfd_state) != xfd) + __xfd_set_state(xfd); +} + static inline void xfd_update_state(struct fpstate *fpstate) { - if (fpu_state_size_dynamic()) { - u64 xfd = fpstate->xfd; - - if (__this_cpu_read(xfd_state) != xfd) - xfd_set_state(xfd); - } + if (fpu_state_size_dynamic()) + xfd_set_state(fpstate->xfd); } extern int __xfd_enable_feature(u64 which, struct fpu_guest *guest_fpu); #else static inline void xfd_set_state(u64 xfd) { } - +static inline void __xfd_set_state(u64 xfd) { } static inline void xfd_update_state(struct fpstate *fpstate) { } static inline int __xfd_enable_feature(u64 which, struct fpu_guest *guest_fpu) { diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index 01d95192dfc5..56fd082859bc 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -4261,7 +4261,7 @@ int kvm_set_msr_common(struct kvm_vcpu *vcpu, struct msr_data *msr_info) if (data & ~kvm_guest_supported_xfd(vcpu)) return 1; - fpu_update_guest_xfd(&vcpu->arch.guest_fpu, data); + vcpu->arch.guest_fpu.xfd = data; break; case MSR_IA32_XFD_ERR: if (!msr_info->host_initiated && @@ -4617,7 +4617,7 @@ int kvm_get_msr_common(struct kvm_vcpu *vcpu, struct msr_data *msr_info) !guest_cpu_cap_has(vcpu, X86_FEATURE_XFD)) return 1; - msr_info->data = vcpu->arch.guest_fpu.fpstate->xfd; + msr_info->data = vcpu->arch.guest_fpu.xfd; break; case MSR_IA32_XFD_ERR: if (!msr_info->host_initiated && @@ -11405,7 +11405,7 @@ static int vcpu_enter_guest(struct kvm_vcpu *vcpu) * in #NM irqoff handler). */ if (vcpu->arch.xfd_no_write_intercept) - fpu_sync_guest_vmexit_xfd_state(); + fpu_sync_guest_vmexit_xfd_state(&vcpu->arch.guest_fpu); kvm_x86_call(handle_exit_irqoff)(vcpu); -- 2.52.0

1 day, 8 hours

1
0
0 0

[PATCH 1/5] x86, fpu: introduce fpu_load_guest_fpstate()

by Paolo Bonzini

Create a variant of fpregs_lock_and_load() that KVM can use in its vCPU entry code after preemption has been disabled. While basing it on the existing logic in vcpu_enter_guest(), ensure that fpregs_assert_state_consistent() always runs and sprinkle a few more assertions. Cc: stable(a)vger.kernel.org Fixes: 820a6ee944e7 ("kvm: x86: Add emulation for IA32_XFD", 2022-01-14) Signed-off-by: Paolo Bonzini <pbonzini(a)redhat.com> --- arch/x86/include/asm/fpu/api.h | 1 + arch/x86/kernel/fpu/core.c | 17 +++++++++++++++++ arch/x86/kvm/x86.c | 8 +------- 3 files changed, 19 insertions(+), 7 deletions(-) diff --git a/arch/x86/include/asm/fpu/api.h b/arch/x86/include/asm/fpu/api.h index cd6f194a912b..0820b2621416 100644 --- a/arch/x86/include/asm/fpu/api.h +++ b/arch/x86/include/asm/fpu/api.h @@ -147,6 +147,7 @@ extern void *get_xsave_addr(struct xregs_state *xsave, int xfeature_nr); /* KVM specific functions */ extern bool fpu_alloc_guest_fpstate(struct fpu_guest *gfpu); extern void fpu_free_guest_fpstate(struct fpu_guest *gfpu); +extern void fpu_load_guest_fpstate(struct fpu_guest *gfpu); extern int fpu_swap_kvm_fpstate(struct fpu_guest *gfpu, bool enter_guest); extern int fpu_enable_guest_xfd_features(struct fpu_guest *guest_fpu, u64 xfeatures); diff --git a/arch/x86/kernel/fpu/core.c b/arch/x86/kernel/fpu/core.c index 3ab27fb86618..a480fa8c65d5 100644 --- a/arch/x86/kernel/fpu/core.c +++ b/arch/x86/kernel/fpu/core.c @@ -878,6 +878,23 @@ void fpregs_lock_and_load(void) fpregs_assert_state_consistent(); } +void fpu_load_guest_fpstate(struct fpu_guest *gfpu) +{ +#ifdef CONFIG_X86_DEBUG_FPU + struct fpu *fpu = x86_task_fpu(current); + WARN_ON_ONCE(gfpu->fpstate != fpu->fpstate); +#endif + + lockdep_assert_preemption_disabled(); + if (test_thread_flag(TIF_NEED_FPU_LOAD)) + fpregs_restore_userregs(); + + fpregs_assert_state_consistent(); + if (gfpu->xfd_err) + wrmsrq(MSR_IA32_XFD_ERR, gfpu->xfd_err); +} +EXPORT_SYMBOL_FOR_KVM(fpu_load_guest_fpstate); + #ifdef CONFIG_X86_DEBUG_FPU /* * If current FPU state according to its tracking (loaded FPU context on this diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index ff8812f3a129..01d95192dfc5 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -11300,13 +11300,7 @@ static int vcpu_enter_guest(struct kvm_vcpu *vcpu) kvm_make_request(KVM_REQ_EVENT, vcpu); } - fpregs_assert_state_consistent(); - if (test_thread_flag(TIF_NEED_FPU_LOAD)) - switch_fpu_return(); - - if (vcpu->arch.guest_fpu.xfd_err) - wrmsrq(MSR_IA32_XFD_ERR, vcpu->arch.guest_fpu.xfd_err); - + fpu_load_guest_fpstate(&vcpu->arch.guest_fpu); kvm_load_xfeatures(vcpu, true); if (unlikely(vcpu->arch.switch_db_regs && -- 2.52.0

1 day, 8 hours

1
0
0 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

Linux-stable-mirror