April 2025 - Linux-stable-mirror

[PATCH v1] s390/virtio_ccw: don't allocate/assign airqs for non-existing queues

by David Hildenbrand

If we finds a vq without a name in our input array in virtio_ccw_find_vqs(), we treat it as "non-existing" and set the vq pointer to NULL; we will not call virtio_ccw_setup_vq() to allocate/setup a vq. Consequently, we create only a queue if it actually exists (name != NULL) and assign an incremental queue index to each such existing queue. However, in virtio_ccw_register_adapter_ind()->get_airq_indicator() we will not ignore these "non-existing queues", but instead assign an airq indicator to them. Besides never releasing them in virtio_ccw_drop_indicators() (because there is no virtqueue), the bigger issue seems to be that there will be a disagreement between the device and the Linux guest about the airq indicator to be used for notifying a queue, because the indicator bit for adapter I/O interrupt is derived from the queue index. The virtio spec states under "Setting Up Two-Stage Queue Indicators": ... indicator contains the guest address of an area wherein the indicators for the devices are contained, starting at bit_nr, one bit per virtqueue of the device. And further in "Notification via Adapter I/O Interrupts": For notifying the driver of virtqueue buffers, the device sets the bit in the guest-provided indicator area at the corresponding offset. For example, QEMU uses in virtio_ccw_notify() the queue index (passed as "vector") to select the relevant indicator bit. If a queue does not exist, it does not have a corresponding indicator bit assigned, because it effectively doesn't have a queue index. Using a virtio-balloon-ccw device under QEMU with free-page-hinting disabled ("free-page-hint=off") but free-page-reporting enabled ("free-page-reporting=on") will result in free page reporting not working as expected: in the virtio_balloon driver, we'll be stuck forever in virtballoon_free_page_report()->wait_event(), because the waitqueue will not be woken up as the notification from the device is lost: it would use the wrong indicator bit. Free page reporting stops working and we get splats (when configured to detect hung wqs) like: INFO: task kworker/1:3:463 blocked for more than 61 seconds. Not tainted 6.14.0 #4 "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message. task:kworker/1:3 [...] Workqueue: events page_reporting_process Call Trace: [<000002f404e6dfb2>] __schedule+0x402/0x1640 [<000002f404e6f22e>] schedule+0x3e/0xe0 [<000002f3846a88fa>] virtballoon_free_page_report+0xaa/0x110 [virtio_balloon] [<000002f40435c8a4>] page_reporting_process+0x2e4/0x740 [<000002f403fd3ee2>] process_one_work+0x1c2/0x400 [<000002f403fd4b96>] worker_thread+0x296/0x420 [<000002f403fe10b4>] kthread+0x124/0x290 [<000002f403f4e0dc>] __ret_from_fork+0x3c/0x60 [<000002f404e77272>] ret_from_fork+0xa/0x38 There was recently a discussion [1] whether the "holes" should be treated differently again, effectively assigning also non-existing queues a queue index: that should also fix the issue, but requires other workarounds to not break existing setups. Let's fix it without affecting existing setups for now by properly ignoring the non-existing queues, so the indicator bits will match the queue indexes. [1] https://lore.kernel.org/all/cover.1720611677.git.mst@redhat.com/ Fixes: a229989d975e ("virtio: don't allocate vqs when names[i] = NULL") Reported-by: Chandra Merla <cmerla(a)redhat.com> Cc: <Stable(a)vger.kernel.org> Cc: Cornelia Huck <cohuck(a)redhat.com> Cc: Thomas Huth <thuth(a)redhat.com> Cc: Halil Pasic <pasic(a)linux.ibm.com> Cc: Eric Farman <farman(a)linux.ibm.com> Cc: Heiko Carstens <hca(a)linux.ibm.com> Cc: Vasily Gorbik <gor(a)linux.ibm.com> Cc: Alexander Gordeev <agordeev(a)linux.ibm.com> Cc: Christian Borntraeger <borntraeger(a)linux.ibm.com> Cc: Sven Schnelle <svens(a)linux.ibm.com> Cc: "Michael S. Tsirkin" <mst(a)redhat.com> Cc: Wei Wang <wei.w.wang(a)intel.com> Signed-off-by: David Hildenbrand <david(a)redhat.com> --- drivers/s390/virtio/virtio_ccw.c | 16 ++++++++++++---- 1 file changed, 12 insertions(+), 4 deletions(-) diff --git a/drivers/s390/virtio/virtio_ccw.c b/drivers/s390/virtio/virtio_ccw.c index 21fa7ac849e5c..4904b831c0a75 100644 --- a/drivers/s390/virtio/virtio_ccw.c +++ b/drivers/s390/virtio/virtio_ccw.c @@ -302,11 +302,17 @@ static struct airq_info *new_airq_info(int index) static unsigned long *get_airq_indicator(struct virtqueue *vqs[], int nvqs, u64 *first, void **airq_info) { - int i, j; + int i, j, queue_idx, highest_queue_idx = -1; struct airq_info *info; unsigned long *indicator_addr = NULL; unsigned long bit, flags; + /* Array entries without an actual queue pointer must be ignored. */ + for (i = 0; i < nvqs; i++) { + if (vqs[i]) + highest_queue_idx++; + } + for (i = 0; i < MAX_AIRQ_AREAS && !indicator_addr; i++) { mutex_lock(&airq_areas_lock); if (!airq_areas[i]) @@ -316,7 +322,7 @@ static unsigned long *get_airq_indicator(struct virtqueue *vqs[], int nvqs, if (!info) return NULL; write_lock_irqsave(&info->lock, flags); - bit = airq_iv_alloc(info->aiv, nvqs); + bit = airq_iv_alloc(info->aiv, highest_queue_idx + 1); if (bit == -1UL) { /* Not enough vacancies. */ write_unlock_irqrestore(&info->lock, flags); @@ -325,8 +331,10 @@ static unsigned long *get_airq_indicator(struct virtqueue *vqs[], int nvqs, *first = bit; *airq_info = info; indicator_addr = info->aiv->vector; - for (j = 0; j < nvqs; j++) { - airq_iv_set_ptr(info->aiv, bit + j, + for (j = 0, queue_idx = 0; j < nvqs; j++) { + if (!vqs[j]) + continue; + airq_iv_set_ptr(info->aiv, bit + queue_idx++, (unsigned long)vqs[j]); } write_unlock_irqrestore(&info->lock, flags); -- 2.48.1

7 months, 3 weeks

8
62
0 0

[PATCH v2] usb: gadget: f_fs: Invalidate io_data when USB request is dequeued or interrupted

by Frode Isaksen

From: Frode Isaksen <frode(a)meta.com> Invalidate io_data by setting context to NULL when USB request is dequeued or interrupted, and check for NULL io_data in epfile_io_complete(). The invalidation of io_data in req->context is done when exiting epfile_io(), since then io_data will become invalid as it is allocated on the stack. The epfile_io_complete() may be called after ffs_epfile_io() returns in case the wait_for_completion_interruptible() is interrupted. This fixes a use-after-free error with the following call stack: Unable to handle kernel paging request at virtual address ffffffc02f7bbcc0 pc : ffs_epfile_io_complete+0x30/0x48 lr : usb_gadget_giveback_request+0x30/0xf8 Call trace: ffs_epfile_io_complete+0x30/0x48 usb_gadget_giveback_request+0x30/0xf8 dwc3_remove_requests+0x264/0x2e8 dwc3_gadget_pullup+0x1d0/0x250 kretprobe_trampoline+0x0/0xc4 usb_gadget_remove_driver+0x40/0xf4 usb_gadget_unregister_driver+0xdc/0x178 unregister_gadget_item+0x40/0x6c ffs_closed+0xd4/0x10c ffs_data_clear+0x2c/0xf0 ffs_data_closed+0x178/0x1ec ffs_ep0_release+0x24/0x38 __fput+0xe8/0x27c Signed-off-by: Frode Isaksen <frode(a)meta.com> Cc: stable(a)vger.kernel.org --- v1 -> v2: Removed WARN_ON() in ffs_epfile_io_complete(). Clarified commit message. Added stable Cc tag. drivers/usb/gadget/function/f_fs.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/drivers/usb/gadget/function/f_fs.c b/drivers/usb/gadget/function/f_fs.c index 2dea9e42a0f8..e35d32e7be58 100644 --- a/drivers/usb/gadget/function/f_fs.c +++ b/drivers/usb/gadget/function/f_fs.c @@ -738,6 +738,9 @@ static void ffs_epfile_io_complete(struct usb_ep *_ep, struct usb_request *req) { struct ffs_io_data *io_data = req->context; + if (io_data == NULL) + return; + if (req->status) io_data->status = req->status; else @@ -1126,6 +1129,7 @@ static ssize_t ffs_epfile_io(struct file *file, struct ffs_io_data *io_data) spin_lock_irq(&epfile->ffs->eps_lock); if (epfile->ep != ep) { ret = -ESHUTDOWN; + req->context = NULL; goto error_lock; } /* @@ -1140,6 +1144,7 @@ static ssize_t ffs_epfile_io(struct file *file, struct ffs_io_data *io_data) interrupted = io_data->status < 0; } + req->context = NULL; if (interrupted) ret = -EINTR; else if (io_data->read && io_data->status > 0) -- 2.49.0

7 months, 3 weeks

2
2
0 0

[PATCH v2] usb/gadget: Add NULL check in ast_vhub_init_dev()

by Henry Martin

devm_kasprintf() returns NULL when memory allocation fails. Currently, ast_vhub_init_dev() does not check for this case, which results in a NULL pointer dereference. Add NULL check after devm_kasprintf() to prevent this issue. Cc: stable(a)vger.kernel.org # v4.18 Fixes: 7ecca2a4080c ("usb/gadget: Add driver for Aspeed SoC virtual hub") Signed-off-by: Henry Martin <bsdhenrymartin(a)gmail.com> --- V1 -> V2: Add Cc: stable label and correct commit message. drivers/usb/gadget/udc/aspeed-vhub/dev.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/usb/gadget/udc/aspeed-vhub/dev.c b/drivers/usb/gadget/udc/aspeed-vhub/dev.c index 573109ca5b79..5b7d41a990d7 100644 --- a/drivers/usb/gadget/udc/aspeed-vhub/dev.c +++ b/drivers/usb/gadget/udc/aspeed-vhub/dev.c @@ -548,6 +548,8 @@ int ast_vhub_init_dev(struct ast_vhub *vhub, unsigned int idx) d->vhub = vhub; d->index = idx; d->name = devm_kasprintf(parent, GFP_KERNEL, "port%d", idx+1); + if (!d->name) + return -ENOMEM; d->regs = vhub->regs + 0x100 + 0x10 * idx; ast_vhub_init_ep0(vhub, &d->ep0, d); -- 2.34.1

7 months, 3 weeks

3
3
0 0

[PATCH V3 1/1] phy: tegra: xusb: Use a bitmask for UTMI pad power state tracking

by Wayne Chang

The current implementation uses bias_pad_enable as a reference count to manage the shared bias pad for all UTMI PHYs. However, during system suspension with connected USB devices, multiple power-down requests for the UTMI pad result in a mismatch in the reference count, which in turn produces warnings such as: [ 237.762967] WARNING: CPU: 10 PID: 1618 at tegra186_utmi_pad_power_down+0x160/0x170 [ 237.763103] Call trace: [ 237.763104] tegra186_utmi_pad_power_down+0x160/0x170 [ 237.763107] tegra186_utmi_phy_power_off+0x10/0x30 [ 237.763110] phy_power_off+0x48/0x100 [ 237.763113] tegra_xusb_enter_elpg+0x204/0x500 [ 237.763119] tegra_xusb_suspend+0x48/0x140 [ 237.763122] platform_pm_suspend+0x2c/0xb0 [ 237.763125] dpm_run_callback.isra.0+0x20/0xa0 [ 237.763127] __device_suspend+0x118/0x330 [ 237.763129] dpm_suspend+0x10c/0x1f0 [ 237.763130] dpm_suspend_start+0x88/0xb0 [ 237.763132] suspend_devices_and_enter+0x120/0x500 [ 237.763135] pm_suspend+0x1ec/0x270 The root cause was traced back to the dynamic power-down changes introduced in commit a30951d31b25 ("xhci: tegra: USB2 pad power controls"), where the UTMI pad was being powered down without verifying its current state. This unbalanced behavior led to discrepancies in the reference count. To rectify this issue, this patch replaces the single reference counter with a bitmask, renamed to utmi_pad_enabled. Each bit in the mask corresponds to one of the four USB2 PHYs, allowing us to track each pad's enablement status individually. With this change: - The bias pad is powered on only when the mask is clear. - Each UTMI pad is powered on or down based on its corresponding bit in the mask, preventing redundant operations. - The overall power state of the shared bias pad is maintained correctly during suspend/resume cycles. The mutex used to prevent race conditions during UTMI pad enable/disable operations has been moved from the tegra186_utmi_bias_pad_power_on/off functions to the parent functions tegra186_utmi_pad_power_on/down. This change ensures that there are no race conditions when updating the bitmask. Cc: stable(a)vger.kernel.org Fixes: a30951d31b25 ("xhci: tegra: USB2 pad power controls") Signed-off-by: Wayne Chang <waynec(a)nvidia.com> --- V1 -> V2: holding the padctl->lock to protect shared bitmask V2 -> V3: updating the commit message with the mutex changes drivers/phy/tegra/xusb-tegra186.c | 44 +++++++++++++++++++------------ 1 file changed, 27 insertions(+), 17 deletions(-) diff --git a/drivers/phy/tegra/xusb-tegra186.c b/drivers/phy/tegra/xusb-tegra186.c index fae6242aa730..cc7b8a6a999f 100644 --- a/drivers/phy/tegra/xusb-tegra186.c +++ b/drivers/phy/tegra/xusb-tegra186.c @@ -237,6 +237,8 @@ #define DATA0_VAL_PD BIT(1) #define USE_XUSB_AO BIT(4) +#define TEGRA_UTMI_PAD_MAX 4 + #define TEGRA186_LANE(_name, _offset, _shift, _mask, _type) \ { \ .name = _name, \ @@ -269,7 +271,7 @@ struct tegra186_xusb_padctl { /* UTMI bias and tracking */ struct clk *usb2_trk_clk; - unsigned int bias_pad_enable; + DECLARE_BITMAP(utmi_pad_enabled, TEGRA_UTMI_PAD_MAX); /* padctl context */ struct tegra186_xusb_padctl_context context; @@ -603,12 +605,8 @@ static void tegra186_utmi_bias_pad_power_on(struct tegra_xusb_padctl *padctl) u32 value; int err; - mutex_lock(&padctl->lock); - - if (priv->bias_pad_enable++ > 0) { - mutex_unlock(&padctl->lock); + if (!bitmap_empty(priv->utmi_pad_enabled, TEGRA_UTMI_PAD_MAX)) return; - } err = clk_prepare_enable(priv->usb2_trk_clk); if (err < 0) @@ -667,17 +665,8 @@ static void tegra186_utmi_bias_pad_power_off(struct tegra_xusb_padctl *padctl) struct tegra186_xusb_padctl *priv = to_tegra186_xusb_padctl(padctl); u32 value; - mutex_lock(&padctl->lock); - - if (WARN_ON(priv->bias_pad_enable == 0)) { - mutex_unlock(&padctl->lock); - return; - } - - if (--priv->bias_pad_enable > 0) { - mutex_unlock(&padctl->lock); + if (!bitmap_empty(priv->utmi_pad_enabled, TEGRA_UTMI_PAD_MAX)) return; - } value = padctl_readl(padctl, XUSB_PADCTL_USB2_BIAS_PAD_CTL1); value |= USB2_PD_TRK; @@ -690,13 +679,13 @@ static void tegra186_utmi_bias_pad_power_off(struct tegra_xusb_padctl *padctl) clk_disable_unprepare(priv->usb2_trk_clk); } - mutex_unlock(&padctl->lock); } static void tegra186_utmi_pad_power_on(struct phy *phy) { struct tegra_xusb_lane *lane = phy_get_drvdata(phy); struct tegra_xusb_padctl *padctl = lane->pad->padctl; + struct tegra186_xusb_padctl *priv = to_tegra186_xusb_padctl(padctl); struct tegra_xusb_usb2_port *port; struct device *dev = padctl->dev; unsigned int index = lane->index; @@ -705,9 +694,16 @@ static void tegra186_utmi_pad_power_on(struct phy *phy) if (!phy) return; + mutex_lock(&padctl->lock); + if (test_bit(index, priv->utmi_pad_enabled)) { + mutex_unlock(&padctl->lock); + return; + } + port = tegra_xusb_find_usb2_port(padctl, index); if (!port) { dev_err(dev, "no port found for USB2 lane %u\n", index); + mutex_unlock(&padctl->lock); return; } @@ -724,18 +720,28 @@ static void tegra186_utmi_pad_power_on(struct phy *phy) value = padctl_readl(padctl, XUSB_PADCTL_USB2_OTG_PADX_CTL1(index)); value &= ~USB2_OTG_PD_DR; padctl_writel(padctl, value, XUSB_PADCTL_USB2_OTG_PADX_CTL1(index)); + + set_bit(index, priv->utmi_pad_enabled); + mutex_unlock(&padctl->lock); } static void tegra186_utmi_pad_power_down(struct phy *phy) { struct tegra_xusb_lane *lane = phy_get_drvdata(phy); struct tegra_xusb_padctl *padctl = lane->pad->padctl; + struct tegra186_xusb_padctl *priv = to_tegra186_xusb_padctl(padctl); unsigned int index = lane->index; u32 value; if (!phy) return; + mutex_lock(&padctl->lock); + if (!test_bit(index, priv->utmi_pad_enabled)) { + mutex_unlock(&padctl->lock); + return; + } + dev_dbg(padctl->dev, "power down UTMI pad %u\n", index); value = padctl_readl(padctl, XUSB_PADCTL_USB2_OTG_PADX_CTL0(index)); @@ -748,7 +754,11 @@ static void tegra186_utmi_pad_power_down(struct phy *phy) udelay(2); + clear_bit(index, priv->utmi_pad_enabled); + tegra186_utmi_bias_pad_power_off(padctl); + + mutex_unlock(&padctl->lock); } static int tegra186_xusb_padctl_vbus_override(struct tegra_xusb_padctl *padctl, -- 2.25.1

7 months, 3 weeks

3
2
0 0

[PATCH v1 2/2] mpi3mr: resets the pending interrupt flag

by Ranjan Kumar

If an admin interrupt is missed, admin_pend_isr may stay set and trigger admin reply processing even when no admin IOs are pending. Clearing/Resetting it in the admin completion path prevents this. Fixes: ca41929b2ed5 ("scsi: mpi3mr: Check admin reply queue from Watchdog") Cc: stable(a)vger.kernel.org Signed-off-by: Sathya Prakash <sathya.prakash(a)broadcom.com> Signed-off-by: Ranjan Kumar <ranjan.kumar(a)broadcom.com> --- drivers/scsi/mpi3mr/mpi3mr_fw.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/scsi/mpi3mr/mpi3mr_fw.c b/drivers/scsi/mpi3mr/mpi3mr_fw.c index d6e402aacb2a..003e1f7005c4 100644 --- a/drivers/scsi/mpi3mr/mpi3mr_fw.c +++ b/drivers/scsi/mpi3mr/mpi3mr_fw.c @@ -451,6 +451,7 @@ int mpi3mr_process_admin_reply_q(struct mpi3mr_ioc *mrioc) return 0; } + atomic_set(&mrioc->admin_pend_isr, 0); reply_desc = (struct mpi3_default_reply_descriptor *)mrioc->admin_reply_base + admin_reply_ci; @@ -2925,6 +2926,7 @@ static int mpi3mr_setup_admin_qpair(struct mpi3mr_ioc *mrioc) mrioc->admin_reply_ci = 0; mrioc->admin_reply_ephase = 1; atomic_set(&mrioc->admin_reply_q_in_use, 0); + atomic_set(&mrioc->admin_pend_isr, 0); if (!mrioc->admin_req_base) { mrioc->admin_req_base = dma_alloc_coherent(&mrioc->pdev->dev, @@ -4653,6 +4655,7 @@ void mpi3mr_memset_buffers(struct mpi3mr_ioc *mrioc) if (mrioc->admin_reply_base) memset(mrioc->admin_reply_base, 0, mrioc->admin_reply_q_sz); atomic_set(&mrioc->admin_reply_q_in_use, 0); + atomic_set(&mrioc->admin_pend_isr, 0); if (mrioc->init_cmds.reply) { memset(mrioc->init_cmds.reply, 0, sizeof(*mrioc->init_cmds.reply)); -- 2.31.1

7 months, 3 weeks

1
0
0 0

[PATCH v1 1/2] Regression fix: Fix pending IOs counter per reply queue

by Ranjan Kumar

Commit 199510e33dea ("scsi: mpi3mr: Update consumer index of reply queues after every 100 replies") introduced a regression with Per reply queue pending IOs counter was wrongly decremented leading to counter getting negative. Fixed the issue by dropping the extra atomic decrement for pending IOs counter. Fixes: 199510e33dea ("scsi: mpi3mr: Update consumer index of reply queues after every 100 replies") Cc: stable(a)vger.kernel.org Signed-off-by: Sathya Prakash <sathya.prakash(a)broadcom.com> Signed-off-by: Ranjan Kumar <ranjan.kumar(a)broadcom.com> --- drivers/scsi/mpi3mr/mpi3mr_fw.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/scsi/mpi3mr/mpi3mr_fw.c b/drivers/scsi/mpi3mr/mpi3mr_fw.c index 3fcb1ad3b070..d6e402aacb2a 100644 --- a/drivers/scsi/mpi3mr/mpi3mr_fw.c +++ b/drivers/scsi/mpi3mr/mpi3mr_fw.c @@ -565,7 +565,7 @@ int mpi3mr_process_op_reply_q(struct mpi3mr_ioc *mrioc, WRITE_ONCE(op_req_q->ci, le16_to_cpu(reply_desc->request_queue_ci)); mpi3mr_process_op_reply_desc(mrioc, reply_desc, &reply_dma, reply_qidx); - atomic_dec(&op_reply_q->pend_ios); + if (reply_dma) mpi3mr_repost_reply_buf(mrioc, reply_dma); num_op_reply++; -- 2.31.1

7 months, 3 weeks

1
0
0 0

[PATCH v2] asus-laptop: Fix an uninitialized variable

by Denis Arefev

The value returned by acpi_evaluate_integer() is not checked, but the result is not always successful, so it is necessary to add a check of the returned value. If the result remains negative during three iterations of the loop, then the uninitialized variable 'val' will be used in the clamp_val() macro, so it must be initialized with the current value of the 'curr' variable. In this case, the algorithm should be less noisy. Found by Linux Verification Center (linuxtesting.org) with SVACE. Fixes: b23910c2194e ("asus-laptop: Pegatron Lucid accelerometer") Cc: stable(a)vger.kernel.org Signed-off-by: Denis Arefev <arefev(a)swemel.ru> --- V1 -> V2: Added check of the return value it as Ilpo Järvinen <ilpo.jarvinen(a)linux.intel.com> suggested. Changed initialization of 'val' variable it as Ilpo Järvinen <ilpo.jarvinen(a)linux.intel.com> suggested. drivers/platform/x86/asus-laptop.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/drivers/platform/x86/asus-laptop.c b/drivers/platform/x86/asus-laptop.c index d460dd194f19..ff674c6d0bbb 100644 --- a/drivers/platform/x86/asus-laptop.c +++ b/drivers/platform/x86/asus-laptop.c @@ -427,10 +427,12 @@ static int asus_pega_lucid_set(struct asus_laptop *asus, int unit, bool enable) static int pega_acc_axis(struct asus_laptop *asus, int curr, char *method) { int i, delta; - unsigned long long val; + acpi_status status; + unsigned long long val = (unsigned long long)curr; for (i = 0; i < PEGA_ACC_RETRIES; i++) { - acpi_evaluate_integer(asus->handle, method, NULL, &val); - + status = acpi_evaluate_integer(asus->handle, method, NULL, &val); + if (ACPI_FAILURE(status)) + continue; /* The output is noisy. From reading the ASL * dissassembly, timeout errors are returned with 1's * in the high word, and the lack of locking around -- 2.43.0

7 months, 3 weeks

2
1
0 0

[PATCH v3] platform/x86: amd: pmf: Fix STT limits

by Mario Limonciello

From: Mario Limonciello <mario.limonciello(a)amd.com> On some platforms it has been observed that STT limits are not being applied properly causing poor performance as power limits are set too low. STT limits that are sent to the platform are supposed to be in Q8.8 format. Convert them before sending. Reported-by: Yijun Shen <Yijun.Shen(a)dell.com> Fixes: 7c45534afa443 ("platform/x86/amd/pmf: Add support for PMF Policy Binary") Cc: stable(a)vger.kernel.org Tested-by: Yijun Shen <Yijun_Shen(a)Dell.com> Signed-off-by: Mario Limonciello <mario.limonciello(a)amd.com> --- v3: * Add a helper with a generic name (so it can be easily be moved to library code in the future) --- drivers/platform/x86/amd/pmf/auto-mode.c | 4 ++-- drivers/platform/x86/amd/pmf/cnqf.c | 8 ++++---- drivers/platform/x86/amd/pmf/core.c | 14 ++++++++++++++ drivers/platform/x86/amd/pmf/pmf.h | 1 + drivers/platform/x86/amd/pmf/sps.c | 12 ++++++++---- drivers/platform/x86/amd/pmf/tee-if.c | 6 ++++-- 6 files changed, 33 insertions(+), 12 deletions(-) diff --git a/drivers/platform/x86/amd/pmf/auto-mode.c b/drivers/platform/x86/amd/pmf/auto-mode.c index 02ff68be10d01..1400ac70c52d1 100644 --- a/drivers/platform/x86/amd/pmf/auto-mode.c +++ b/drivers/platform/x86/amd/pmf/auto-mode.c @@ -120,9 +120,9 @@ static void amd_pmf_set_automode(struct amd_pmf_dev *dev, int idx, amd_pmf_send_cmd(dev, SET_SPPT_APU_ONLY, false, pwr_ctrl->sppt_apu_only, NULL); amd_pmf_send_cmd(dev, SET_STT_MIN_LIMIT, false, pwr_ctrl->stt_min, NULL); amd_pmf_send_cmd(dev, SET_STT_LIMIT_APU, false, - pwr_ctrl->stt_skin_temp[STT_TEMP_APU], NULL); + fixp_q88_from_integer(pwr_ctrl->stt_skin_temp[STT_TEMP_APU]), NULL); amd_pmf_send_cmd(dev, SET_STT_LIMIT_HS2, false, - pwr_ctrl->stt_skin_temp[STT_TEMP_HS2], NULL); + fixp_q88_from_integer(pwr_ctrl->stt_skin_temp[STT_TEMP_HS2]), NULL); if (is_apmf_func_supported(dev, APMF_FUNC_SET_FAN_IDX)) apmf_update_fan_idx(dev, config_store.mode_set[idx].fan_control.manual, diff --git a/drivers/platform/x86/amd/pmf/cnqf.c b/drivers/platform/x86/amd/pmf/cnqf.c index bc8899e15c914..3cde8a5de64a9 100644 --- a/drivers/platform/x86/amd/pmf/cnqf.c +++ b/drivers/platform/x86/amd/pmf/cnqf.c @@ -81,10 +81,10 @@ static int amd_pmf_set_cnqf(struct amd_pmf_dev *dev, int src, int idx, amd_pmf_send_cmd(dev, SET_SPPT, false, pc->sppt, NULL); amd_pmf_send_cmd(dev, SET_SPPT_APU_ONLY, false, pc->sppt_apu_only, NULL); amd_pmf_send_cmd(dev, SET_STT_MIN_LIMIT, false, pc->stt_min, NULL); - amd_pmf_send_cmd(dev, SET_STT_LIMIT_APU, false, pc->stt_skin_temp[STT_TEMP_APU], - NULL); - amd_pmf_send_cmd(dev, SET_STT_LIMIT_HS2, false, pc->stt_skin_temp[STT_TEMP_HS2], - NULL); + amd_pmf_send_cmd(dev, SET_STT_LIMIT_APU, false, + fixp_q88_from_integer(pc->stt_skin_temp[STT_TEMP_APU]), NULL); + amd_pmf_send_cmd(dev, SET_STT_LIMIT_HS2, false, + fixp_q88_from_integer(pc->stt_skin_temp[STT_TEMP_HS2]), NULL); if (is_apmf_func_supported(dev, APMF_FUNC_SET_FAN_IDX)) apmf_update_fan_idx(dev, diff --git a/drivers/platform/x86/amd/pmf/core.c b/drivers/platform/x86/amd/pmf/core.c index a2cb2d5544f5b..5209996eba674 100644 --- a/drivers/platform/x86/amd/pmf/core.c +++ b/drivers/platform/x86/amd/pmf/core.c @@ -176,6 +176,20 @@ static void __maybe_unused amd_pmf_dump_registers(struct amd_pmf_dev *dev) dev_dbg(dev->dev, "AMD_PMF_REGISTER_MESSAGE:%x\n", value); } +/** + * fixp_q88_from_integer: Convert integer to Q8.8 + * @val: input value + * + * Converts an integer into binary fixed point format where 8 bits + * are used for integer and 8 bits are used for the decimal. + * + * Return: unsigned integer converted to Q8.8 format + */ +u32 fixp_q88_from_integer(u32 val) +{ + return val << 8; +} + int amd_pmf_send_cmd(struct amd_pmf_dev *dev, u8 message, bool get, u32 arg, u32 *data) { int rc; diff --git a/drivers/platform/x86/amd/pmf/pmf.h b/drivers/platform/x86/amd/pmf/pmf.h index e6bdee68ccf34..2865e0a70b43d 100644 --- a/drivers/platform/x86/amd/pmf/pmf.h +++ b/drivers/platform/x86/amd/pmf/pmf.h @@ -777,6 +777,7 @@ int apmf_install_handler(struct amd_pmf_dev *pmf_dev); int apmf_os_power_slider_update(struct amd_pmf_dev *dev, u8 flag); int amd_pmf_set_dram_addr(struct amd_pmf_dev *dev, bool alloc_buffer); int amd_pmf_notify_sbios_heartbeat_event_v2(struct amd_pmf_dev *dev, u8 flag); +u32 fixp_q88_from_integer(u32 val); /* SPS Layer */ int amd_pmf_get_pprof_modes(struct amd_pmf_dev *pmf); diff --git a/drivers/platform/x86/amd/pmf/sps.c b/drivers/platform/x86/amd/pmf/sps.c index d3083383f11fb..dfc5285b681f7 100644 --- a/drivers/platform/x86/amd/pmf/sps.c +++ b/drivers/platform/x86/amd/pmf/sps.c @@ -198,9 +198,11 @@ static void amd_pmf_update_slider_v2(struct amd_pmf_dev *dev, int idx) amd_pmf_send_cmd(dev, SET_STT_MIN_LIMIT, false, apts_config_store.val[idx].stt_min_limit, NULL); amd_pmf_send_cmd(dev, SET_STT_LIMIT_APU, false, - apts_config_store.val[idx].stt_skin_temp_limit_apu, NULL); + fixp_q88_from_integer(apts_config_store.val[idx].stt_skin_temp_limit_apu), + NULL); amd_pmf_send_cmd(dev, SET_STT_LIMIT_HS2, false, - apts_config_store.val[idx].stt_skin_temp_limit_hs2, NULL); + fixp_q88_from_integer(apts_config_store.val[idx].stt_skin_temp_limit_hs2), + NULL); } void amd_pmf_update_slider(struct amd_pmf_dev *dev, bool op, int idx, @@ -217,9 +219,11 @@ void amd_pmf_update_slider(struct amd_pmf_dev *dev, bool op, int idx, amd_pmf_send_cmd(dev, SET_STT_MIN_LIMIT, false, config_store.prop[src][idx].stt_min, NULL); amd_pmf_send_cmd(dev, SET_STT_LIMIT_APU, false, - config_store.prop[src][idx].stt_skin_temp[STT_TEMP_APU], NULL); + fixp_q88_from_integer(config_store.prop[src][idx].stt_skin_temp[STT_TEMP_APU]), + NULL); amd_pmf_send_cmd(dev, SET_STT_LIMIT_HS2, false, - config_store.prop[src][idx].stt_skin_temp[STT_TEMP_HS2], NULL); + fixp_q88_from_integer(config_store.prop[src][idx].stt_skin_temp[STT_TEMP_HS2]), + NULL); } else if (op == SLIDER_OP_GET) { amd_pmf_send_cmd(dev, GET_SPL, true, ARG_NONE, &table->prop[src][idx].spl); amd_pmf_send_cmd(dev, GET_FPPT, true, ARG_NONE, &table->prop[src][idx].fppt); diff --git a/drivers/platform/x86/amd/pmf/tee-if.c b/drivers/platform/x86/amd/pmf/tee-if.c index a1e43873a07b0..22d48048f9d01 100644 --- a/drivers/platform/x86/amd/pmf/tee-if.c +++ b/drivers/platform/x86/amd/pmf/tee-if.c @@ -123,7 +123,8 @@ static void amd_pmf_apply_policies(struct amd_pmf_dev *dev, struct ta_pmf_enact_ case PMF_POLICY_STT_SKINTEMP_APU: if (dev->prev_data->stt_skintemp_apu != val) { - amd_pmf_send_cmd(dev, SET_STT_LIMIT_APU, false, val, NULL); + amd_pmf_send_cmd(dev, SET_STT_LIMIT_APU, false, + fixp_q88_from_integer(val), NULL); dev_dbg(dev->dev, "update STT_SKINTEMP_APU: %u\n", val); dev->prev_data->stt_skintemp_apu = val; } @@ -131,7 +132,8 @@ static void amd_pmf_apply_policies(struct amd_pmf_dev *dev, struct ta_pmf_enact_ case PMF_POLICY_STT_SKINTEMP_HS2: if (dev->prev_data->stt_skintemp_hs2 != val) { - amd_pmf_send_cmd(dev, SET_STT_LIMIT_HS2, false, val, NULL); + amd_pmf_send_cmd(dev, SET_STT_LIMIT_HS2, false, + fixp_q88_from_integer(val), NULL); dev_dbg(dev->dev, "update STT_SKINTEMP_HS2: %u\n", val); dev->prev_data->stt_skintemp_hs2 = val; } -- 2.43.0

7 months, 3 weeks

3
3
0 0

[PATCH] usb: cdnsp: Fix issue with resuming from L1

by Pawel Laszczak

Subject: [PATCH] usb: cdnsp: Fix issue with resuming from L1 In very rare cases after resuming controller from L1 to L0 it reads registers before the clock has been enabled and as the result driver reads incorrect value. To fix this issue driver increases APB timeout value. Probably this issue occurs only on Cadence platform but fix should have no impact for other existing platforms. Fixes: 3d82904559f4 ("usb: cdnsp: cdns3 Add main part of Cadence USBSSP DRD Driver") cc: stable(a)vger.kernel.org Signed-off-by: Pawel Laszczak <pawell(a)cadence.com> --- drivers/usb/cdns3/cdnsp-gadget.c | 22 ++++++++++++++++++++++ drivers/usb/cdns3/cdnsp-gadget.h | 4 ++++ 2 files changed, 26 insertions(+) diff --git a/drivers/usb/cdns3/cdnsp-gadget.c b/drivers/usb/cdns3/cdnsp-gadget.c index 87f310841735..b12581b94567 100644 --- a/drivers/usb/cdns3/cdnsp-gadget.c +++ b/drivers/usb/cdns3/cdnsp-gadget.c @@ -139,6 +139,21 @@ static void cdnsp_clear_port_change_bit(struct cdnsp_device *pdev, (portsc & PORT_CHANGE_BITS), port_regs); } +static void cdnsp_set_apb_timeout_value(struct cdnsp_device *pdev) +{ + __le32 __iomem *reg; + void __iomem *base; + u32 offset = 0; + u32 val; + + base = &pdev->cap_regs->hc_capbase; + offset = cdnsp_find_next_ext_cap(base, offset, D_XEC_PRE_REGS_CAP); + reg = base + offset + REG_CHICKEN_BITS_3_OFFSET; + + val = le32_to_cpu(readl(reg)); + writel(cpu_to_le32(CHICKEN_APB_TIMEOUT_SET(val)), reg); +} + static void cdnsp_set_chicken_bits_2(struct cdnsp_device *pdev, u32 bit) { __le32 __iomem *reg; @@ -1798,6 +1813,13 @@ static int cdnsp_gen_setup(struct cdnsp_device *pdev) pdev->hci_version = HC_VERSION(pdev->hcc_params); pdev->hcc_params = readl(&pdev->cap_regs->hcc_params); + /* In very rare cases after resuming controller from L1 to L0 it reads + * registers before the clock has been enabled and as the result driver + * reads incorrect value. + * To fix this issue driver increases APB timeout value. + */ + cdnsp_set_apb_timeout_value(pdev); + cdnsp_get_rev_cap(pdev); /* Make sure the Device Controller is halted. */ diff --git a/drivers/usb/cdns3/cdnsp-gadget.h b/drivers/usb/cdns3/cdnsp-gadget.h index 84887dfea763..a4d678fba005 100644 --- a/drivers/usb/cdns3/cdnsp-gadget.h +++ b/drivers/usb/cdns3/cdnsp-gadget.h @@ -520,6 +520,10 @@ struct cdnsp_rev_cap { #define REG_CHICKEN_BITS_2_OFFSET 0x48 #define CHICKEN_XDMA_2_TP_CACHE_DIS BIT(28) +#define REG_CHICKEN_BITS_3_OFFSET 0x4C +#define CHICKEN_APB_TIMEOUT_VALUE 0x1C20 +#define CHICKEN_APB_TIMEOUT_SET(p) (((p) & ~GENMASK(21, 0)) | CHICKEN_APB_TIMEOUT_VALUE) + /* XBUF Extended Capability ID. */ #define XBUF_CAP_ID 0xCB #define XBUF_RX_TAG_MASK_0_OFFSET 0x1C -- 2.43.0

7 months, 3 weeks

3
5
0 0

[PATCH 1/3] drm/nouveau: Prevent signaled fences in pending list

by Philipp Stanner

Nouveau currently relies on the assumption that dma_fences will only ever get signaled through nouveau_fence_signal(), which takes care of removing a signaled fence from the list nouveau_fence_chan.pending. This self-imposed rule is violated in nouveau_fence_done(), where dma_fence_is_signaled() (somewhat surprisingly, considering its name) can signal the fence without removing it from the list. This enables accesses to already signaled fences through the list, which is a bug. In particular, it can race with nouveau_fence_context_kill(), which would then attempt to set an error code on an already signaled fence, which is illegal. In nouveau_fence_done(), the call to nouveau_fence_update() already ensures to signal all ready fences. Thus, the signaling potentially performed by dma_fence_is_signaled() is actually not necessary. Replace the call to dma_fence_is_signaled() with nouveau_fence_base_is_signaled(). Cc: <stable(a)vger.kernel.org> # 4.10+, precise commit not to be determined Signed-off-by: Philipp Stanner <phasta(a)kernel.org> --- drivers/gpu/drm/nouveau/nouveau_fence.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/nouveau/nouveau_fence.c b/drivers/gpu/drm/nouveau/nouveau_fence.c index 7cc84472cece..33535987d8ed 100644 --- a/drivers/gpu/drm/nouveau/nouveau_fence.c +++ b/drivers/gpu/drm/nouveau/nouveau_fence.c @@ -274,7 +274,7 @@ nouveau_fence_done(struct nouveau_fence *fence) nvif_event_block(&fctx->event); spin_unlock_irqrestore(&fctx->lock, flags); } - return dma_fence_is_signaled(&fence->base); + return test_bit(DMA_FENCE_FLAG_SIGNALED_BIT, &fence->base.flags); } static long -- 2.48.1

7 months, 3 weeks

4
8
0 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

Linux-stable-mirror April 2025