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

Re: patch "acpi: typec: ucsi: Introduce a ->poll_cci method" added to usb-linus

by Fedor Pchelkin

On Wed, 19. Feb 15:20, gregkh(a)linuxfoundation.org wrote: > > This is a note to let you know that I've just added the patch titled > > acpi: typec: ucsi: Introduce a ->poll_cci method > > to my usb git tree which can be found at > git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb.git > in the usb-linus branch. > > The patch will show up in the next release of the linux-next tree > (usually sometime within the next 24 hours during the week.) > > The patch will hopefully also be merged in Linus's tree for the > next -rc kernel release. > > If you have any questions about this process, please let me know. > > > From 976e7e9bdc7719a023a4ecccd2e3daec9ab20a40 Mon Sep 17 00:00:00 2001 > From: "Christian A. Ehrhardt" <lk(a)c--e.de> > Date: Mon, 17 Feb 2025 13:54:39 +0300 > Subject: acpi: typec: ucsi: Introduce a ->poll_cci method > > For the ACPI backend of UCSI the UCSI "registers" are just a memory copy > of the register values in an opregion. The ACPI implementation in the > BIOS ensures that the opregion contents are synced to the embedded > controller and it ensures that the registers (in particular CCI) are > synced back to the opregion on notifications. While there is an ACPI call > that syncs the actual registers to the opregion there is rarely a need to > do this and on some ACPI implementations it actually breaks in various > interesting ways. > > The only reason to force a sync from the embedded controller is to poll > CCI while notifications are disabled. Only the ucsi core knows if this > is the case and guessing based on the current command is suboptimal, i.e. > leading to the following spurious assertion splat: > > WARNING: CPU: 3 PID: 76 at drivers/usb/typec/ucsi/ucsi.c:1388 ucsi_reset_ppm+0x1b4/0x1c0 [typec_ucsi] > CPU: 3 UID: 0 PID: 76 Comm: kworker/3:0 Not tainted 6.12.11-200.fc41.x86_64 #1 > Hardware name: LENOVO 21D0/LNVNB161216, BIOS J6CN45WW 03/17/2023 > Workqueue: events_long ucsi_init_work [typec_ucsi] > RIP: 0010:ucsi_reset_ppm+0x1b4/0x1c0 [typec_ucsi] > Call Trace: > <TASK> > ucsi_init_work+0x3c/0xac0 [typec_ucsi] > process_one_work+0x179/0x330 > worker_thread+0x252/0x390 > kthread+0xd2/0x100 > ret_from_fork+0x34/0x50 > ret_from_fork_asm+0x1a/0x30 > </TASK> > > Thus introduce a ->poll_cci() method that works like ->read_cci() with an > additional forced sync and document that this should be used when polling > with notifications disabled. For all other backends that presumably don't > have this issue use the same implementation for both methods. > > Fixes: fa48d7e81624 ("usb: typec: ucsi: Do not call ACPI _DSM method for UCSI read operations") > Cc: stable <stable(a)kernel.org> Oh, the stable tag has been mangled here.. I didn't notice this, sorry. Now Cc'ing the appropriate list. Could you apply the patch to stables based on Fixes tag then, please? They are 6.12 and 6.13. Thanks! > Signed-off-by: Christian A. Ehrhardt <lk(a)c--e.de> > Tested-by: Fedor Pchelkin <boddah8794(a)gmail.com> > Signed-off-by: Fedor Pchelkin <boddah8794(a)gmail.com> > Reviewed-by: Heikki Krogerus <heikki.krogerus(a)linux.intel.com> > Link: https://lore.kernel.org/r/20250217105442.113486-2-boddah8794@gmail.com > Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> > --- > drivers/usb/typec/ucsi/ucsi.c | 10 +++++----- > drivers/usb/typec/ucsi/ucsi.h | 2 ++ > drivers/usb/typec/ucsi/ucsi_acpi.c | 21 ++++++++++++++------- > drivers/usb/typec/ucsi/ucsi_ccg.c | 1 + > drivers/usb/typec/ucsi/ucsi_glink.c | 1 + > drivers/usb/typec/ucsi/ucsi_stm32g0.c | 1 + > drivers/usb/typec/ucsi/ucsi_yoga_c630.c | 1 + > 7 files changed, 25 insertions(+), 12 deletions(-) > > diff --git a/drivers/usb/typec/ucsi/ucsi.c b/drivers/usb/typec/ucsi/ucsi.c > index fcf499cc9458..0fe1476f4c29 100644 > --- a/drivers/usb/typec/ucsi/ucsi.c > +++ b/drivers/usb/typec/ucsi/ucsi.c > @@ -1346,7 +1346,7 @@ static int ucsi_reset_ppm(struct ucsi *ucsi) > > mutex_lock(&ucsi->ppm_lock); > > - ret = ucsi->ops->read_cci(ucsi, &cci); > + ret = ucsi->ops->poll_cci(ucsi, &cci); > if (ret < 0) > goto out; > > @@ -1364,7 +1364,7 @@ static int ucsi_reset_ppm(struct ucsi *ucsi) > > tmo = jiffies + msecs_to_jiffies(UCSI_TIMEOUT_MS); > do { > - ret = ucsi->ops->read_cci(ucsi, &cci); > + ret = ucsi->ops->poll_cci(ucsi, &cci); > if (ret < 0) > goto out; > if (cci & UCSI_CCI_COMMAND_COMPLETE) > @@ -1393,7 +1393,7 @@ static int ucsi_reset_ppm(struct ucsi *ucsi) > /* Give the PPM time to process a reset before reading CCI */ > msleep(20); > > - ret = ucsi->ops->read_cci(ucsi, &cci); > + ret = ucsi->ops->poll_cci(ucsi, &cci); > if (ret) > goto out; > > @@ -1929,8 +1929,8 @@ struct ucsi *ucsi_create(struct device *dev, const struct ucsi_operations *ops) > struct ucsi *ucsi; > > if (!ops || > - !ops->read_version || !ops->read_cci || !ops->read_message_in || > - !ops->sync_control || !ops->async_control) > + !ops->read_version || !ops->read_cci || !ops->poll_cci || > + !ops->read_message_in || !ops->sync_control || !ops->async_control) > return ERR_PTR(-EINVAL); > > ucsi = kzalloc(sizeof(*ucsi), GFP_KERNEL); > diff --git a/drivers/usb/typec/ucsi/ucsi.h b/drivers/usb/typec/ucsi/ucsi.h > index 82735eb34f0e..28780acc4af2 100644 > --- a/drivers/usb/typec/ucsi/ucsi.h > +++ b/drivers/usb/typec/ucsi/ucsi.h > @@ -62,6 +62,7 @@ struct dentry; > * struct ucsi_operations - UCSI I/O operations > * @read_version: Read implemented UCSI version > * @read_cci: Read CCI register > + * @poll_cci: Read CCI register while polling with notifications disabled > * @read_message_in: Read message data from UCSI > * @sync_control: Blocking control operation > * @async_control: Non-blocking control operation > @@ -76,6 +77,7 @@ struct dentry; > struct ucsi_operations { > int (*read_version)(struct ucsi *ucsi, u16 *version); > int (*read_cci)(struct ucsi *ucsi, u32 *cci); > + int (*poll_cci)(struct ucsi *ucsi, u32 *cci); > int (*read_message_in)(struct ucsi *ucsi, void *val, size_t val_len); > int (*sync_control)(struct ucsi *ucsi, u64 command); > int (*async_control)(struct ucsi *ucsi, u64 command); > diff --git a/drivers/usb/typec/ucsi/ucsi_acpi.c b/drivers/usb/typec/ucsi/ucsi_acpi.c > index 5c5515551963..ac1ebb5d9527 100644 > --- a/drivers/usb/typec/ucsi/ucsi_acpi.c > +++ b/drivers/usb/typec/ucsi/ucsi_acpi.c > @@ -59,19 +59,24 @@ static int ucsi_acpi_read_version(struct ucsi *ucsi, u16 *version) > static int ucsi_acpi_read_cci(struct ucsi *ucsi, u32 *cci) > { > struct ucsi_acpi *ua = ucsi_get_drvdata(ucsi); > - int ret; > - > - if (UCSI_COMMAND(ua->cmd) == UCSI_PPM_RESET) { > - ret = ucsi_acpi_dsm(ua, UCSI_DSM_FUNC_READ); > - if (ret) > - return ret; > - } > > memcpy(cci, ua->base + UCSI_CCI, sizeof(*cci)); > > return 0; > } > > +static int ucsi_acpi_poll_cci(struct ucsi *ucsi, u32 *cci) > +{ > + struct ucsi_acpi *ua = ucsi_get_drvdata(ucsi); > + int ret; > + > + ret = ucsi_acpi_dsm(ua, UCSI_DSM_FUNC_READ); > + if (ret) > + return ret; > + > + return ucsi_acpi_read_cci(ucsi, cci); > +} > + > static int ucsi_acpi_read_message_in(struct ucsi *ucsi, void *val, size_t val_len) > { > struct ucsi_acpi *ua = ucsi_get_drvdata(ucsi); > @@ -94,6 +99,7 @@ static int ucsi_acpi_async_control(struct ucsi *ucsi, u64 command) > static const struct ucsi_operations ucsi_acpi_ops = { > .read_version = ucsi_acpi_read_version, > .read_cci = ucsi_acpi_read_cci, > + .poll_cci = ucsi_acpi_poll_cci, > .read_message_in = ucsi_acpi_read_message_in, > .sync_control = ucsi_sync_control_common, > .async_control = ucsi_acpi_async_control > @@ -142,6 +148,7 @@ static int ucsi_gram_sync_control(struct ucsi *ucsi, u64 command) > static const struct ucsi_operations ucsi_gram_ops = { > .read_version = ucsi_acpi_read_version, > .read_cci = ucsi_acpi_read_cci, > + .poll_cci = ucsi_acpi_poll_cci, > .read_message_in = ucsi_gram_read_message_in, > .sync_control = ucsi_gram_sync_control, > .async_control = ucsi_acpi_async_control > diff --git a/drivers/usb/typec/ucsi/ucsi_ccg.c b/drivers/usb/typec/ucsi/ucsi_ccg.c > index 740171f24ef9..4b1668733a4b 100644 > --- a/drivers/usb/typec/ucsi/ucsi_ccg.c > +++ b/drivers/usb/typec/ucsi/ucsi_ccg.c > @@ -664,6 +664,7 @@ static int ucsi_ccg_sync_control(struct ucsi *ucsi, u64 command) > static const struct ucsi_operations ucsi_ccg_ops = { > .read_version = ucsi_ccg_read_version, > .read_cci = ucsi_ccg_read_cci, > + .poll_cci = ucsi_ccg_read_cci, > .read_message_in = ucsi_ccg_read_message_in, > .sync_control = ucsi_ccg_sync_control, > .async_control = ucsi_ccg_async_control, > diff --git a/drivers/usb/typec/ucsi/ucsi_glink.c b/drivers/usb/typec/ucsi/ucsi_glink.c > index fed39d458090..8af79101a2fc 100644 > --- a/drivers/usb/typec/ucsi/ucsi_glink.c > +++ b/drivers/usb/typec/ucsi/ucsi_glink.c > @@ -206,6 +206,7 @@ static void pmic_glink_ucsi_connector_status(struct ucsi_connector *con) > static const struct ucsi_operations pmic_glink_ucsi_ops = { > .read_version = pmic_glink_ucsi_read_version, > .read_cci = pmic_glink_ucsi_read_cci, > + .poll_cci = pmic_glink_ucsi_read_cci, > .read_message_in = pmic_glink_ucsi_read_message_in, > .sync_control = ucsi_sync_control_common, > .async_control = pmic_glink_ucsi_async_control, > diff --git a/drivers/usb/typec/ucsi/ucsi_stm32g0.c b/drivers/usb/typec/ucsi/ucsi_stm32g0.c > index 6923fad31d79..57ef7d83a412 100644 > --- a/drivers/usb/typec/ucsi/ucsi_stm32g0.c > +++ b/drivers/usb/typec/ucsi/ucsi_stm32g0.c > @@ -424,6 +424,7 @@ static irqreturn_t ucsi_stm32g0_irq_handler(int irq, void *data) > static const struct ucsi_operations ucsi_stm32g0_ops = { > .read_version = ucsi_stm32g0_read_version, > .read_cci = ucsi_stm32g0_read_cci, > + .poll_cci = ucsi_stm32g0_read_cci, > .read_message_in = ucsi_stm32g0_read_message_in, > .sync_control = ucsi_sync_control_common, > .async_control = ucsi_stm32g0_async_control, > diff --git a/drivers/usb/typec/ucsi/ucsi_yoga_c630.c b/drivers/usb/typec/ucsi/ucsi_yoga_c630.c > index 4cae85c0dc12..d33e3f2dd1d8 100644 > --- a/drivers/usb/typec/ucsi/ucsi_yoga_c630.c > +++ b/drivers/usb/typec/ucsi/ucsi_yoga_c630.c > @@ -74,6 +74,7 @@ static int yoga_c630_ucsi_async_control(struct ucsi *ucsi, u64 command) > static const struct ucsi_operations yoga_c630_ucsi_ops = { > .read_version = yoga_c630_ucsi_read_version, > .read_cci = yoga_c630_ucsi_read_cci, > + .poll_cci = yoga_c630_ucsi_read_cci, > .read_message_in = yoga_c630_ucsi_read_message_in, > .sync_control = ucsi_sync_control_common, > .async_control = yoga_c630_ucsi_async_control, > -- > 2.48.1 > >

4 months

2
1
0 0

FAILED: patch "[PATCH] usb: dwc3: gadget: Prevent irq storm when TH re-executes" failed to apply to 5.4-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 5.4-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. To reproduce the conflict and resubmit, you may use the following commands: git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-5.4.y git checkout FETCH_HEAD git cherry-pick -x 69c58deec19628c8a686030102176484eb94fed4 # <resolve conflicts, build, test, etc.> git commit -s git send-email --to '<stable(a)vger.kernel.org>' --in-reply-to '2025030917-porridge-retired-1abd@gregkh' --subject-prefix 'PATCH 5.4.y' HEAD^.. Possible dependencies: thanks, greg k-h ------------------ original commit in Linus's tree ------------------ From 69c58deec19628c8a686030102176484eb94fed4 Mon Sep 17 00:00:00 2001 From: Badhri Jagan Sridharan <badhri(a)google.com> Date: Sun, 16 Feb 2025 22:30:02 +0000 Subject: [PATCH] usb: dwc3: gadget: Prevent irq storm when TH re-executes While commit d325a1de49d6 ("usb: dwc3: gadget: Prevent losing events in event cache") makes sure that top half(TH) does not end up overwriting the cached events before processing them when the TH gets invoked more than one time, returning IRQ_HANDLED results in occasional irq storm where the TH hogs the CPU. The irq storm can be prevented by the flag before event handler busy is cleared. Default enable interrupt moderation in all versions which support them. ftrace event stub during dwc3 irq storm: irq/504_dwc3-1111 ( 1111) [000] .... 70.000866: irq_handler_exit: irq=14 ret=handled irq/504_dwc3-1111 ( 1111) [000] .... 70.000872: irq_handler_entry: irq=504 name=dwc3 irq/504_dwc3-1111 ( 1111) [000] .... 70.000874: irq_handler_exit: irq=504 ret=handled irq/504_dwc3-1111 ( 1111) [000] .... 70.000881: irq_handler_entry: irq=504 name=dwc3 irq/504_dwc3-1111 ( 1111) [000] .... 70.000883: irq_handler_exit: irq=504 ret=handled irq/504_dwc3-1111 ( 1111) [000] .... 70.000889: irq_handler_entry: irq=504 name=dwc3 irq/504_dwc3-1111 ( 1111) [000] .... 70.000892: irq_handler_exit: irq=504 ret=handled irq/504_dwc3-1111 ( 1111) [000] .... 70.000898: irq_handler_entry: irq=504 name=dwc3 irq/504_dwc3-1111 ( 1111) [000] .... 70.000901: irq_handler_exit: irq=504 ret=handled irq/504_dwc3-1111 ( 1111) [000] .... 70.000907: irq_handler_entry: irq=504 name=dwc3 irq/504_dwc3-1111 ( 1111) [000] .... 70.000909: irq_handler_exit: irq=504 ret=handled irq/504_dwc3-1111 ( 1111) [000] .... 70.000915: irq_handler_entry: irq=504 name=dwc3 irq/504_dwc3-1111 ( 1111) [000] .... 70.000918: irq_handler_exit: irq=504 ret=handled irq/504_dwc3-1111 ( 1111) [000] .... 70.000924: irq_handler_entry: irq=504 name=dwc3 irq/504_dwc3-1111 ( 1111) [000] .... 70.000927: irq_handler_exit: irq=504 ret=handled irq/504_dwc3-1111 ( 1111) [000] .... 70.000933: irq_handler_entry: irq=504 name=dwc3 irq/504_dwc3-1111 ( 1111) [000] .... 70.000935: irq_handler_exit: irq=504 ret=handled .... Cc: stable <stable(a)kernel.org> Suggested-by: Thinh Nguyen <Thinh.Nguyen(a)synopsys.com> Fixes: d325a1de49d6 ("usb: dwc3: gadget: Prevent losing events in event cache") Signed-off-by: Badhri Jagan Sridharan <badhri(a)google.com> Acked-by: Thinh Nguyen <Thinh.Nguyen(a)synopsys.com> Link: https://lore.kernel.org/r/20250216223003.3568039-1-badhri@google.com Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> diff --git a/drivers/usb/dwc3/core.c b/drivers/usb/dwc3/core.c index dfa1b5fe48dc..2c472cb97f6c 100644 --- a/drivers/usb/dwc3/core.c +++ b/drivers/usb/dwc3/core.c @@ -1835,8 +1835,6 @@ static void dwc3_get_properties(struct dwc3 *dwc) dwc->tx_thr_num_pkt_prd = tx_thr_num_pkt_prd; dwc->tx_max_burst_prd = tx_max_burst_prd; - dwc->imod_interval = 0; - dwc->tx_fifo_resize_max_num = tx_fifo_resize_max_num; } @@ -1854,21 +1852,19 @@ static void dwc3_check_params(struct dwc3 *dwc) unsigned int hwparam_gen = DWC3_GHWPARAMS3_SSPHY_IFC(dwc->hwparams.hwparams3); - /* Check for proper value of imod_interval */ - if (dwc->imod_interval && !dwc3_has_imod(dwc)) { - dev_warn(dwc->dev, "Interrupt moderation not supported\n"); - dwc->imod_interval = 0; - } - /* + * Enable IMOD for all supporting controllers. + * + * Particularly, DWC_usb3 v3.00a must enable this feature for + * the following reason: + * * Workaround for STAR 9000961433 which affects only version * 3.00a of the DWC_usb3 core. This prevents the controller * interrupt from being masked while handling events. IMOD * allows us to work around this issue. Enable it for the * affected version. */ - if (!dwc->imod_interval && - DWC3_VER_IS(DWC3, 300A)) + if (dwc3_has_imod((dwc))) dwc->imod_interval = 1; /* Check the maximum_speed parameter */ diff --git a/drivers/usb/dwc3/gadget.c b/drivers/usb/dwc3/gadget.c index ddd6b2ce5710..89a4dc8ebf94 100644 --- a/drivers/usb/dwc3/gadget.c +++ b/drivers/usb/dwc3/gadget.c @@ -4501,14 +4501,18 @@ static irqreturn_t dwc3_process_event_buf(struct dwc3_event_buffer *evt) dwc3_writel(dwc->regs, DWC3_GEVNTSIZ(0), DWC3_GEVNTSIZ_SIZE(evt->length)); + evt->flags &= ~DWC3_EVENT_PENDING; + /* + * Add an explicit write memory barrier to make sure that the update of + * clearing DWC3_EVENT_PENDING is observed in dwc3_check_event_buf() + */ + wmb(); + if (dwc->imod_interval) { dwc3_writel(dwc->regs, DWC3_GEVNTCOUNT(0), DWC3_GEVNTCOUNT_EHB); dwc3_writel(dwc->regs, DWC3_DEV_IMOD(0), dwc->imod_interval); } - /* Keep the clearing of DWC3_EVENT_PENDING at the end */ - evt->flags &= ~DWC3_EVENT_PENDING; - return ret; }

4 months

1
0
0 0

FAILED: patch "[PATCH] usb: dwc3: gadget: Prevent irq storm when TH re-executes" failed to apply to 5.10-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 5.10-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. To reproduce the conflict and resubmit, you may use the following commands: git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-5.10.y git checkout FETCH_HEAD git cherry-pick -x 69c58deec19628c8a686030102176484eb94fed4 # <resolve conflicts, build, test, etc.> git commit -s git send-email --to '<stable(a)vger.kernel.org>' --in-reply-to '2025030916-expenses-extended-20cc@gregkh' --subject-prefix 'PATCH 5.10.y' HEAD^.. Possible dependencies: thanks, greg k-h ------------------ original commit in Linus's tree ------------------ From 69c58deec19628c8a686030102176484eb94fed4 Mon Sep 17 00:00:00 2001 From: Badhri Jagan Sridharan <badhri(a)google.com> Date: Sun, 16 Feb 2025 22:30:02 +0000 Subject: [PATCH] usb: dwc3: gadget: Prevent irq storm when TH re-executes While commit d325a1de49d6 ("usb: dwc3: gadget: Prevent losing events in event cache") makes sure that top half(TH) does not end up overwriting the cached events before processing them when the TH gets invoked more than one time, returning IRQ_HANDLED results in occasional irq storm where the TH hogs the CPU. The irq storm can be prevented by the flag before event handler busy is cleared. Default enable interrupt moderation in all versions which support them. ftrace event stub during dwc3 irq storm: irq/504_dwc3-1111 ( 1111) [000] .... 70.000866: irq_handler_exit: irq=14 ret=handled irq/504_dwc3-1111 ( 1111) [000] .... 70.000872: irq_handler_entry: irq=504 name=dwc3 irq/504_dwc3-1111 ( 1111) [000] .... 70.000874: irq_handler_exit: irq=504 ret=handled irq/504_dwc3-1111 ( 1111) [000] .... 70.000881: irq_handler_entry: irq=504 name=dwc3 irq/504_dwc3-1111 ( 1111) [000] .... 70.000883: irq_handler_exit: irq=504 ret=handled irq/504_dwc3-1111 ( 1111) [000] .... 70.000889: irq_handler_entry: irq=504 name=dwc3 irq/504_dwc3-1111 ( 1111) [000] .... 70.000892: irq_handler_exit: irq=504 ret=handled irq/504_dwc3-1111 ( 1111) [000] .... 70.000898: irq_handler_entry: irq=504 name=dwc3 irq/504_dwc3-1111 ( 1111) [000] .... 70.000901: irq_handler_exit: irq=504 ret=handled irq/504_dwc3-1111 ( 1111) [000] .... 70.000907: irq_handler_entry: irq=504 name=dwc3 irq/504_dwc3-1111 ( 1111) [000] .... 70.000909: irq_handler_exit: irq=504 ret=handled irq/504_dwc3-1111 ( 1111) [000] .... 70.000915: irq_handler_entry: irq=504 name=dwc3 irq/504_dwc3-1111 ( 1111) [000] .... 70.000918: irq_handler_exit: irq=504 ret=handled irq/504_dwc3-1111 ( 1111) [000] .... 70.000924: irq_handler_entry: irq=504 name=dwc3 irq/504_dwc3-1111 ( 1111) [000] .... 70.000927: irq_handler_exit: irq=504 ret=handled irq/504_dwc3-1111 ( 1111) [000] .... 70.000933: irq_handler_entry: irq=504 name=dwc3 irq/504_dwc3-1111 ( 1111) [000] .... 70.000935: irq_handler_exit: irq=504 ret=handled .... Cc: stable <stable(a)kernel.org> Suggested-by: Thinh Nguyen <Thinh.Nguyen(a)synopsys.com> Fixes: d325a1de49d6 ("usb: dwc3: gadget: Prevent losing events in event cache") Signed-off-by: Badhri Jagan Sridharan <badhri(a)google.com> Acked-by: Thinh Nguyen <Thinh.Nguyen(a)synopsys.com> Link: https://lore.kernel.org/r/20250216223003.3568039-1-badhri@google.com Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> diff --git a/drivers/usb/dwc3/core.c b/drivers/usb/dwc3/core.c index dfa1b5fe48dc..2c472cb97f6c 100644 --- a/drivers/usb/dwc3/core.c +++ b/drivers/usb/dwc3/core.c @@ -1835,8 +1835,6 @@ static void dwc3_get_properties(struct dwc3 *dwc) dwc->tx_thr_num_pkt_prd = tx_thr_num_pkt_prd; dwc->tx_max_burst_prd = tx_max_burst_prd; - dwc->imod_interval = 0; - dwc->tx_fifo_resize_max_num = tx_fifo_resize_max_num; } @@ -1854,21 +1852,19 @@ static void dwc3_check_params(struct dwc3 *dwc) unsigned int hwparam_gen = DWC3_GHWPARAMS3_SSPHY_IFC(dwc->hwparams.hwparams3); - /* Check for proper value of imod_interval */ - if (dwc->imod_interval && !dwc3_has_imod(dwc)) { - dev_warn(dwc->dev, "Interrupt moderation not supported\n"); - dwc->imod_interval = 0; - } - /* + * Enable IMOD for all supporting controllers. + * + * Particularly, DWC_usb3 v3.00a must enable this feature for + * the following reason: + * * Workaround for STAR 9000961433 which affects only version * 3.00a of the DWC_usb3 core. This prevents the controller * interrupt from being masked while handling events. IMOD * allows us to work around this issue. Enable it for the * affected version. */ - if (!dwc->imod_interval && - DWC3_VER_IS(DWC3, 300A)) + if (dwc3_has_imod((dwc))) dwc->imod_interval = 1; /* Check the maximum_speed parameter */ diff --git a/drivers/usb/dwc3/gadget.c b/drivers/usb/dwc3/gadget.c index ddd6b2ce5710..89a4dc8ebf94 100644 --- a/drivers/usb/dwc3/gadget.c +++ b/drivers/usb/dwc3/gadget.c @@ -4501,14 +4501,18 @@ static irqreturn_t dwc3_process_event_buf(struct dwc3_event_buffer *evt) dwc3_writel(dwc->regs, DWC3_GEVNTSIZ(0), DWC3_GEVNTSIZ_SIZE(evt->length)); + evt->flags &= ~DWC3_EVENT_PENDING; + /* + * Add an explicit write memory barrier to make sure that the update of + * clearing DWC3_EVENT_PENDING is observed in dwc3_check_event_buf() + */ + wmb(); + if (dwc->imod_interval) { dwc3_writel(dwc->regs, DWC3_GEVNTCOUNT(0), DWC3_GEVNTCOUNT_EHB); dwc3_writel(dwc->regs, DWC3_DEV_IMOD(0), dwc->imod_interval); } - /* Keep the clearing of DWC3_EVENT_PENDING at the end */ - evt->flags &= ~DWC3_EVENT_PENDING; - return ret; }

4 months

1
0
0 0

FAILED: patch "[PATCH] usb: dwc3: Set SUSPENDENABLE soon after phy init" failed to apply to 5.10-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 5.10-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. To reproduce the conflict and resubmit, you may use the following commands: git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-5.10.y git checkout FETCH_HEAD git cherry-pick -x cc5bfc4e16fc1d1c520cd7bb28646e82b6e69217 # <resolve conflicts, build, test, etc.> git commit -s git send-email --to '<stable(a)vger.kernel.org>' --in-reply-to '2025030953-washboard-overcrowd-fed5@gregkh' --subject-prefix 'PATCH 5.10.y' HEAD^.. Possible dependencies: thanks, greg k-h ------------------ original commit in Linus's tree ------------------ From cc5bfc4e16fc1d1c520cd7bb28646e82b6e69217 Mon Sep 17 00:00:00 2001 From: Thinh Nguyen <Thinh.Nguyen(a)synopsys.com> Date: Thu, 30 Jan 2025 23:49:31 +0000 Subject: [PATCH] usb: dwc3: Set SUSPENDENABLE soon after phy init After phy initialization, some phy operations can only be executed while in lower P states. Ensure GUSB3PIPECTL.SUSPENDENABLE and GUSB2PHYCFG.SUSPHY are set soon after initialization to avoid blocking phy ops. Previously the SUSPENDENABLE bits are only set after the controller initialization, which may not happen right away if there's no gadget driver or xhci driver bound. Revise this to clear SUSPENDENABLE bits only when there's mode switching (change in GCTL.PRTCAPDIR). Fixes: 6d735722063a ("usb: dwc3: core: Prevent phy suspend during init") Cc: stable <stable(a)kernel.org> Signed-off-by: Thinh Nguyen <Thinh.Nguyen(a)synopsys.com> Link: https://lore.kernel.org/r/633aef0afee7d56d2316f7cc3e1b2a6d518a8cc9.17382809… Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> diff --git a/drivers/usb/dwc3/core.c b/drivers/usb/dwc3/core.c index 2c472cb97f6c..66a08b527165 100644 --- a/drivers/usb/dwc3/core.c +++ b/drivers/usb/dwc3/core.c @@ -131,11 +131,24 @@ void dwc3_enable_susphy(struct dwc3 *dwc, bool enable) } } -void dwc3_set_prtcap(struct dwc3 *dwc, u32 mode) +void dwc3_set_prtcap(struct dwc3 *dwc, u32 mode, bool ignore_susphy) { + unsigned int hw_mode; u32 reg; reg = dwc3_readl(dwc->regs, DWC3_GCTL); + + /* + * For DRD controllers, GUSB3PIPECTL.SUSPENDENABLE and + * GUSB2PHYCFG.SUSPHY should be cleared during mode switching, + * and they can be set after core initialization. + */ + hw_mode = DWC3_GHWPARAMS0_MODE(dwc->hwparams.hwparams0); + if (hw_mode == DWC3_GHWPARAMS0_MODE_DRD && !ignore_susphy) { + if (DWC3_GCTL_PRTCAP(reg) != mode) + dwc3_enable_susphy(dwc, false); + } + reg &= ~(DWC3_GCTL_PRTCAPDIR(DWC3_GCTL_PRTCAP_OTG)); reg |= DWC3_GCTL_PRTCAPDIR(mode); dwc3_writel(dwc->regs, DWC3_GCTL, reg); @@ -216,7 +229,7 @@ static void __dwc3_set_mode(struct work_struct *work) spin_lock_irqsave(&dwc->lock, flags); - dwc3_set_prtcap(dwc, desired_dr_role); + dwc3_set_prtcap(dwc, desired_dr_role, false); spin_unlock_irqrestore(&dwc->lock, flags); @@ -658,16 +671,7 @@ static int dwc3_ss_phy_setup(struct dwc3 *dwc, int index) */ reg &= ~DWC3_GUSB3PIPECTL_UX_EXIT_PX; - /* - * Above DWC_usb3.0 1.94a, it is recommended to set - * DWC3_GUSB3PIPECTL_SUSPHY to '0' during coreConsultant configuration. - * So default value will be '0' when the core is reset. Application - * needs to set it to '1' after the core initialization is completed. - * - * Similarly for DRD controllers, GUSB3PIPECTL.SUSPENDENABLE must be - * cleared after power-on reset, and it can be set after core - * initialization. - */ + /* Ensure the GUSB3PIPECTL.SUSPENDENABLE is cleared prior to phy init. */ reg &= ~DWC3_GUSB3PIPECTL_SUSPHY; if (dwc->u2ss_inp3_quirk) @@ -747,15 +751,7 @@ static int dwc3_hs_phy_setup(struct dwc3 *dwc, int index) break; } - /* - * Above DWC_usb3.0 1.94a, it is recommended to set - * DWC3_GUSB2PHYCFG_SUSPHY to '0' during coreConsultant configuration. - * So default value will be '0' when the core is reset. Application - * needs to set it to '1' after the core initialization is completed. - * - * Similarly for DRD controllers, GUSB2PHYCFG.SUSPHY must be cleared - * after power-on reset, and it can be set after core initialization. - */ + /* Ensure the GUSB2PHYCFG.SUSPHY is cleared prior to phy init. */ reg &= ~DWC3_GUSB2PHYCFG_SUSPHY; if (dwc->dis_enblslpm_quirk) @@ -830,6 +826,25 @@ static int dwc3_phy_init(struct dwc3 *dwc) goto err_exit_usb3_phy; } + /* + * Above DWC_usb3.0 1.94a, it is recommended to set + * DWC3_GUSB3PIPECTL_SUSPHY and DWC3_GUSB2PHYCFG_SUSPHY to '0' during + * coreConsultant configuration. So default value will be '0' when the + * core is reset. Application needs to set it to '1' after the core + * initialization is completed. + * + * Certain phy requires to be in P0 power state during initialization. + * Make sure GUSB3PIPECTL.SUSPENDENABLE and GUSB2PHYCFG.SUSPHY are clear + * prior to phy init to maintain in the P0 state. + * + * After phy initialization, some phy operations can only be executed + * while in lower P states. Ensure GUSB3PIPECTL.SUSPENDENABLE and + * GUSB2PHYCFG.SUSPHY are set soon after initialization to avoid + * blocking phy ops. + */ + if (!DWC3_VER_IS_WITHIN(DWC3, ANY, 194A)) + dwc3_enable_susphy(dwc, true); + return 0; err_exit_usb3_phy: @@ -1588,7 +1603,7 @@ static int dwc3_core_init_mode(struct dwc3 *dwc) switch (dwc->dr_mode) { case USB_DR_MODE_PERIPHERAL: - dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE); + dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE, false); if (dwc->usb2_phy) otg_set_vbus(dwc->usb2_phy->otg, false); @@ -1600,7 +1615,7 @@ static int dwc3_core_init_mode(struct dwc3 *dwc) return dev_err_probe(dev, ret, "failed to initialize gadget\n"); break; case USB_DR_MODE_HOST: - dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_HOST); + dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_HOST, false); if (dwc->usb2_phy) otg_set_vbus(dwc->usb2_phy->otg, true); @@ -1645,7 +1660,7 @@ static void dwc3_core_exit_mode(struct dwc3 *dwc) } /* de-assert DRVVBUS for HOST and OTG mode */ - dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE); + dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE, true); } static void dwc3_get_software_properties(struct dwc3 *dwc) @@ -2453,7 +2468,7 @@ static int dwc3_resume_common(struct dwc3 *dwc, pm_message_t msg) if (ret) return ret; - dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE); + dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_DEVICE, true); dwc3_gadget_resume(dwc); break; case DWC3_GCTL_PRTCAP_HOST: @@ -2461,7 +2476,7 @@ static int dwc3_resume_common(struct dwc3 *dwc, pm_message_t msg) ret = dwc3_core_init_for_resume(dwc); if (ret) return ret; - dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_HOST); + dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_HOST, true); break; } /* Restore GUSB2PHYCFG bits that were modified in suspend */ @@ -2490,7 +2505,7 @@ static int dwc3_resume_common(struct dwc3 *dwc, pm_message_t msg) if (ret) return ret; - dwc3_set_prtcap(dwc, dwc->current_dr_role); + dwc3_set_prtcap(dwc, dwc->current_dr_role, true); dwc3_otg_init(dwc); if (dwc->current_otg_role == DWC3_OTG_ROLE_HOST) { diff --git a/drivers/usb/dwc3/core.h b/drivers/usb/dwc3/core.h index c955039bb4f6..aaa39e663f60 100644 --- a/drivers/usb/dwc3/core.h +++ b/drivers/usb/dwc3/core.h @@ -1558,7 +1558,7 @@ struct dwc3_gadget_ep_cmd_params { #define DWC3_HAS_OTG BIT(3) /* prototypes */ -void dwc3_set_prtcap(struct dwc3 *dwc, u32 mode); +void dwc3_set_prtcap(struct dwc3 *dwc, u32 mode, bool ignore_susphy); void dwc3_set_mode(struct dwc3 *dwc, u32 mode); u32 dwc3_core_fifo_space(struct dwc3_ep *dep, u8 type); diff --git a/drivers/usb/dwc3/drd.c b/drivers/usb/dwc3/drd.c index d76ae676783c..7977860932b1 100644 --- a/drivers/usb/dwc3/drd.c +++ b/drivers/usb/dwc3/drd.c @@ -173,7 +173,7 @@ void dwc3_otg_init(struct dwc3 *dwc) * block "Initialize GCTL for OTG operation". */ /* GCTL.PrtCapDir=2'b11 */ - dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_OTG); + dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_OTG, true); /* GUSB2PHYCFG0.SusPHY=0 */ reg = dwc3_readl(dwc->regs, DWC3_GUSB2PHYCFG(0)); reg &= ~DWC3_GUSB2PHYCFG_SUSPHY; @@ -556,7 +556,7 @@ int dwc3_drd_init(struct dwc3 *dwc) dwc3_drd_update(dwc); } else { - dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_OTG); + dwc3_set_prtcap(dwc, DWC3_GCTL_PRTCAP_OTG, true); /* use OTG block to get ID event */ irq = dwc3_otg_get_irq(dwc);

4 months

1
0
0 0

FAILED: patch "[PATCH] usb: xhci: Enable the TRB overfetch quirk on VIA VL805" failed to apply to 5.4-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 5.4-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. To reproduce the conflict and resubmit, you may use the following commands: git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-5.4.y git checkout FETCH_HEAD git cherry-pick -x c133ec0e5717868c9967fa3df92a55e537b1aead # <resolve conflicts, build, test, etc.> git commit -s git send-email --to '<stable(a)vger.kernel.org>' --in-reply-to '2025030902-fernlike-flashback-65c0@gregkh' --subject-prefix 'PATCH 5.4.y' HEAD^.. Possible dependencies: thanks, greg k-h ------------------ original commit in Linus's tree ------------------ From c133ec0e5717868c9967fa3df92a55e537b1aead Mon Sep 17 00:00:00 2001 From: Michal Pecio <michal.pecio(a)gmail.com> Date: Tue, 25 Feb 2025 11:59:27 +0200 Subject: [PATCH] usb: xhci: Enable the TRB overfetch quirk on VIA VL805 Raspberry Pi is a major user of those chips and they discovered a bug - when the end of a transfer ring segment is reached, up to four TRBs can be prefetched from the next page even if the segment ends with link TRB and on page boundary (the chip claims to support standard 4KB pages). It also appears that if the prefetched TRBs belong to a different ring whose doorbell is later rung, they may be used without refreshing from system RAM and the endpoint will stay idle if their cycle bit is stale. Other users complain about IOMMU faults on x86 systems, unsurprisingly. Deal with it by using existing quirk which allocates a dummy page after each transfer ring segment. This was seen to resolve both problems. RPi came up with a more efficient solution, shortening each segment by four TRBs, but it complicated the driver and they ditched it for this quirk. Also rename the quirk and add VL805 device ID macro. Signed-off-by: Michal Pecio <michal.pecio(a)gmail.com> Link: https://github.com/raspberrypi/linux/issues/4685 Closes: https://bugzilla.kernel.org/show_bug.cgi?id=215906 CC: stable(a)vger.kernel.org Signed-off-by: Mathias Nyman <mathias.nyman(a)linux.intel.com> Link: https://lore.kernel.org/r/20250225095927.2512358-2-mathias.nyman@linux.inte… Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> diff --git a/drivers/usb/host/xhci-mem.c b/drivers/usb/host/xhci-mem.c index 92703efda1f7..fdf0c1008225 100644 --- a/drivers/usb/host/xhci-mem.c +++ b/drivers/usb/host/xhci-mem.c @@ -2437,7 +2437,8 @@ int xhci_mem_init(struct xhci_hcd *xhci, gfp_t flags) * and our use of dma addresses in the trb_address_map radix tree needs * TRB_SEGMENT_SIZE alignment, so we pick the greater alignment need. */ - if (xhci->quirks & XHCI_ZHAOXIN_TRB_FETCH) + if (xhci->quirks & XHCI_TRB_OVERFETCH) + /* Buggy HC prefetches beyond segment bounds - allocate dummy space at the end */ xhci->segment_pool = dma_pool_create("xHCI ring segments", dev, TRB_SEGMENT_SIZE * 2, TRB_SEGMENT_SIZE * 2, xhci->page_size * 2); else diff --git a/drivers/usb/host/xhci-pci.c b/drivers/usb/host/xhci-pci.c index ad0ff356f6fa..54460d11f7ee 100644 --- a/drivers/usb/host/xhci-pci.c +++ b/drivers/usb/host/xhci-pci.c @@ -38,6 +38,8 @@ #define PCI_DEVICE_ID_ETRON_EJ168 0x7023 #define PCI_DEVICE_ID_ETRON_EJ188 0x7052 +#define PCI_DEVICE_ID_VIA_VL805 0x3483 + #define PCI_DEVICE_ID_INTEL_LYNXPOINT_XHCI 0x8c31 #define PCI_DEVICE_ID_INTEL_LYNXPOINT_LP_XHCI 0x9c31 #define PCI_DEVICE_ID_INTEL_WILDCATPOINT_LP_XHCI 0x9cb1 @@ -418,8 +420,10 @@ static void xhci_pci_quirks(struct device *dev, struct xhci_hcd *xhci) pdev->device == 0x3432) xhci->quirks |= XHCI_BROKEN_STREAMS; - if (pdev->vendor == PCI_VENDOR_ID_VIA && pdev->device == 0x3483) + if (pdev->vendor == PCI_VENDOR_ID_VIA && pdev->device == PCI_DEVICE_ID_VIA_VL805) { xhci->quirks |= XHCI_LPM_SUPPORT; + xhci->quirks |= XHCI_TRB_OVERFETCH; + } if (pdev->vendor == PCI_VENDOR_ID_ASMEDIA && pdev->device == PCI_DEVICE_ID_ASMEDIA_1042_XHCI) { @@ -467,11 +471,11 @@ static void xhci_pci_quirks(struct device *dev, struct xhci_hcd *xhci) if (pdev->device == 0x9202) { xhci->quirks |= XHCI_RESET_ON_RESUME; - xhci->quirks |= XHCI_ZHAOXIN_TRB_FETCH; + xhci->quirks |= XHCI_TRB_OVERFETCH; } if (pdev->device == 0x9203) - xhci->quirks |= XHCI_ZHAOXIN_TRB_FETCH; + xhci->quirks |= XHCI_TRB_OVERFETCH; } if (pdev->vendor == PCI_VENDOR_ID_CDNS && diff --git a/drivers/usb/host/xhci.h b/drivers/usb/host/xhci.h index 8c164340a2c3..779b01dee068 100644 --- a/drivers/usb/host/xhci.h +++ b/drivers/usb/host/xhci.h @@ -1632,7 +1632,7 @@ struct xhci_hcd { #define XHCI_EP_CTX_BROKEN_DCS BIT_ULL(42) #define XHCI_SUSPEND_RESUME_CLKS BIT_ULL(43) #define XHCI_RESET_TO_DEFAULT BIT_ULL(44) -#define XHCI_ZHAOXIN_TRB_FETCH BIT_ULL(45) +#define XHCI_TRB_OVERFETCH BIT_ULL(45) #define XHCI_ZHAOXIN_HOST BIT_ULL(46) #define XHCI_WRITE_64_HI_LO BIT_ULL(47) #define XHCI_CDNS_SCTX_QUIRK BIT_ULL(48)

4 months

1
0
0 0

FAILED: patch "[PATCH] usb: xhci: Enable the TRB overfetch quirk on VIA VL805" failed to apply to 6.1-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 6.1-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. To reproduce the conflict and resubmit, you may use the following commands: git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-6.1.y git checkout FETCH_HEAD git cherry-pick -x c133ec0e5717868c9967fa3df92a55e537b1aead # <resolve conflicts, build, test, etc.> git commit -s git send-email --to '<stable(a)vger.kernel.org>' --in-reply-to '2025030959-character-delouse-db17@gregkh' --subject-prefix 'PATCH 6.1.y' HEAD^.. Possible dependencies: thanks, greg k-h ------------------ original commit in Linus's tree ------------------ From c133ec0e5717868c9967fa3df92a55e537b1aead Mon Sep 17 00:00:00 2001 From: Michal Pecio <michal.pecio(a)gmail.com> Date: Tue, 25 Feb 2025 11:59:27 +0200 Subject: [PATCH] usb: xhci: Enable the TRB overfetch quirk on VIA VL805 Raspberry Pi is a major user of those chips and they discovered a bug - when the end of a transfer ring segment is reached, up to four TRBs can be prefetched from the next page even if the segment ends with link TRB and on page boundary (the chip claims to support standard 4KB pages). It also appears that if the prefetched TRBs belong to a different ring whose doorbell is later rung, they may be used without refreshing from system RAM and the endpoint will stay idle if their cycle bit is stale. Other users complain about IOMMU faults on x86 systems, unsurprisingly. Deal with it by using existing quirk which allocates a dummy page after each transfer ring segment. This was seen to resolve both problems. RPi came up with a more efficient solution, shortening each segment by four TRBs, but it complicated the driver and they ditched it for this quirk. Also rename the quirk and add VL805 device ID macro. Signed-off-by: Michal Pecio <michal.pecio(a)gmail.com> Link: https://github.com/raspberrypi/linux/issues/4685 Closes: https://bugzilla.kernel.org/show_bug.cgi?id=215906 CC: stable(a)vger.kernel.org Signed-off-by: Mathias Nyman <mathias.nyman(a)linux.intel.com> Link: https://lore.kernel.org/r/20250225095927.2512358-2-mathias.nyman@linux.inte… Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> diff --git a/drivers/usb/host/xhci-mem.c b/drivers/usb/host/xhci-mem.c index 92703efda1f7..fdf0c1008225 100644 --- a/drivers/usb/host/xhci-mem.c +++ b/drivers/usb/host/xhci-mem.c @@ -2437,7 +2437,8 @@ int xhci_mem_init(struct xhci_hcd *xhci, gfp_t flags) * and our use of dma addresses in the trb_address_map radix tree needs * TRB_SEGMENT_SIZE alignment, so we pick the greater alignment need. */ - if (xhci->quirks & XHCI_ZHAOXIN_TRB_FETCH) + if (xhci->quirks & XHCI_TRB_OVERFETCH) + /* Buggy HC prefetches beyond segment bounds - allocate dummy space at the end */ xhci->segment_pool = dma_pool_create("xHCI ring segments", dev, TRB_SEGMENT_SIZE * 2, TRB_SEGMENT_SIZE * 2, xhci->page_size * 2); else diff --git a/drivers/usb/host/xhci-pci.c b/drivers/usb/host/xhci-pci.c index ad0ff356f6fa..54460d11f7ee 100644 --- a/drivers/usb/host/xhci-pci.c +++ b/drivers/usb/host/xhci-pci.c @@ -38,6 +38,8 @@ #define PCI_DEVICE_ID_ETRON_EJ168 0x7023 #define PCI_DEVICE_ID_ETRON_EJ188 0x7052 +#define PCI_DEVICE_ID_VIA_VL805 0x3483 + #define PCI_DEVICE_ID_INTEL_LYNXPOINT_XHCI 0x8c31 #define PCI_DEVICE_ID_INTEL_LYNXPOINT_LP_XHCI 0x9c31 #define PCI_DEVICE_ID_INTEL_WILDCATPOINT_LP_XHCI 0x9cb1 @@ -418,8 +420,10 @@ static void xhci_pci_quirks(struct device *dev, struct xhci_hcd *xhci) pdev->device == 0x3432) xhci->quirks |= XHCI_BROKEN_STREAMS; - if (pdev->vendor == PCI_VENDOR_ID_VIA && pdev->device == 0x3483) + if (pdev->vendor == PCI_VENDOR_ID_VIA && pdev->device == PCI_DEVICE_ID_VIA_VL805) { xhci->quirks |= XHCI_LPM_SUPPORT; + xhci->quirks |= XHCI_TRB_OVERFETCH; + } if (pdev->vendor == PCI_VENDOR_ID_ASMEDIA && pdev->device == PCI_DEVICE_ID_ASMEDIA_1042_XHCI) { @@ -467,11 +471,11 @@ static void xhci_pci_quirks(struct device *dev, struct xhci_hcd *xhci) if (pdev->device == 0x9202) { xhci->quirks |= XHCI_RESET_ON_RESUME; - xhci->quirks |= XHCI_ZHAOXIN_TRB_FETCH; + xhci->quirks |= XHCI_TRB_OVERFETCH; } if (pdev->device == 0x9203) - xhci->quirks |= XHCI_ZHAOXIN_TRB_FETCH; + xhci->quirks |= XHCI_TRB_OVERFETCH; } if (pdev->vendor == PCI_VENDOR_ID_CDNS && diff --git a/drivers/usb/host/xhci.h b/drivers/usb/host/xhci.h index 8c164340a2c3..779b01dee068 100644 --- a/drivers/usb/host/xhci.h +++ b/drivers/usb/host/xhci.h @@ -1632,7 +1632,7 @@ struct xhci_hcd { #define XHCI_EP_CTX_BROKEN_DCS BIT_ULL(42) #define XHCI_SUSPEND_RESUME_CLKS BIT_ULL(43) #define XHCI_RESET_TO_DEFAULT BIT_ULL(44) -#define XHCI_ZHAOXIN_TRB_FETCH BIT_ULL(45) +#define XHCI_TRB_OVERFETCH BIT_ULL(45) #define XHCI_ZHAOXIN_HOST BIT_ULL(46) #define XHCI_WRITE_64_HI_LO BIT_ULL(47) #define XHCI_CDNS_SCTX_QUIRK BIT_ULL(48)

4 months

1
0
0 0

FAILED: patch "[PATCH] usb: xhci: Enable the TRB overfetch quirk on VIA VL805" failed to apply to 6.6-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 6.6-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. To reproduce the conflict and resubmit, you may use the following commands: git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-6.6.y git checkout FETCH_HEAD git cherry-pick -x c133ec0e5717868c9967fa3df92a55e537b1aead # <resolve conflicts, build, test, etc.> git commit -s git send-email --to '<stable(a)vger.kernel.org>' --in-reply-to '2025030959-thee-uniformed-b4eb@gregkh' --subject-prefix 'PATCH 6.6.y' HEAD^.. Possible dependencies: thanks, greg k-h ------------------ original commit in Linus's tree ------------------ From c133ec0e5717868c9967fa3df92a55e537b1aead Mon Sep 17 00:00:00 2001 From: Michal Pecio <michal.pecio(a)gmail.com> Date: Tue, 25 Feb 2025 11:59:27 +0200 Subject: [PATCH] usb: xhci: Enable the TRB overfetch quirk on VIA VL805 Raspberry Pi is a major user of those chips and they discovered a bug - when the end of a transfer ring segment is reached, up to four TRBs can be prefetched from the next page even if the segment ends with link TRB and on page boundary (the chip claims to support standard 4KB pages). It also appears that if the prefetched TRBs belong to a different ring whose doorbell is later rung, they may be used without refreshing from system RAM and the endpoint will stay idle if their cycle bit is stale. Other users complain about IOMMU faults on x86 systems, unsurprisingly. Deal with it by using existing quirk which allocates a dummy page after each transfer ring segment. This was seen to resolve both problems. RPi came up with a more efficient solution, shortening each segment by four TRBs, but it complicated the driver and they ditched it for this quirk. Also rename the quirk and add VL805 device ID macro. Signed-off-by: Michal Pecio <michal.pecio(a)gmail.com> Link: https://github.com/raspberrypi/linux/issues/4685 Closes: https://bugzilla.kernel.org/show_bug.cgi?id=215906 CC: stable(a)vger.kernel.org Signed-off-by: Mathias Nyman <mathias.nyman(a)linux.intel.com> Link: https://lore.kernel.org/r/20250225095927.2512358-2-mathias.nyman@linux.inte… Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> diff --git a/drivers/usb/host/xhci-mem.c b/drivers/usb/host/xhci-mem.c index 92703efda1f7..fdf0c1008225 100644 --- a/drivers/usb/host/xhci-mem.c +++ b/drivers/usb/host/xhci-mem.c @@ -2437,7 +2437,8 @@ int xhci_mem_init(struct xhci_hcd *xhci, gfp_t flags) * and our use of dma addresses in the trb_address_map radix tree needs * TRB_SEGMENT_SIZE alignment, so we pick the greater alignment need. */ - if (xhci->quirks & XHCI_ZHAOXIN_TRB_FETCH) + if (xhci->quirks & XHCI_TRB_OVERFETCH) + /* Buggy HC prefetches beyond segment bounds - allocate dummy space at the end */ xhci->segment_pool = dma_pool_create("xHCI ring segments", dev, TRB_SEGMENT_SIZE * 2, TRB_SEGMENT_SIZE * 2, xhci->page_size * 2); else diff --git a/drivers/usb/host/xhci-pci.c b/drivers/usb/host/xhci-pci.c index ad0ff356f6fa..54460d11f7ee 100644 --- a/drivers/usb/host/xhci-pci.c +++ b/drivers/usb/host/xhci-pci.c @@ -38,6 +38,8 @@ #define PCI_DEVICE_ID_ETRON_EJ168 0x7023 #define PCI_DEVICE_ID_ETRON_EJ188 0x7052 +#define PCI_DEVICE_ID_VIA_VL805 0x3483 + #define PCI_DEVICE_ID_INTEL_LYNXPOINT_XHCI 0x8c31 #define PCI_DEVICE_ID_INTEL_LYNXPOINT_LP_XHCI 0x9c31 #define PCI_DEVICE_ID_INTEL_WILDCATPOINT_LP_XHCI 0x9cb1 @@ -418,8 +420,10 @@ static void xhci_pci_quirks(struct device *dev, struct xhci_hcd *xhci) pdev->device == 0x3432) xhci->quirks |= XHCI_BROKEN_STREAMS; - if (pdev->vendor == PCI_VENDOR_ID_VIA && pdev->device == 0x3483) + if (pdev->vendor == PCI_VENDOR_ID_VIA && pdev->device == PCI_DEVICE_ID_VIA_VL805) { xhci->quirks |= XHCI_LPM_SUPPORT; + xhci->quirks |= XHCI_TRB_OVERFETCH; + } if (pdev->vendor == PCI_VENDOR_ID_ASMEDIA && pdev->device == PCI_DEVICE_ID_ASMEDIA_1042_XHCI) { @@ -467,11 +471,11 @@ static void xhci_pci_quirks(struct device *dev, struct xhci_hcd *xhci) if (pdev->device == 0x9202) { xhci->quirks |= XHCI_RESET_ON_RESUME; - xhci->quirks |= XHCI_ZHAOXIN_TRB_FETCH; + xhci->quirks |= XHCI_TRB_OVERFETCH; } if (pdev->device == 0x9203) - xhci->quirks |= XHCI_ZHAOXIN_TRB_FETCH; + xhci->quirks |= XHCI_TRB_OVERFETCH; } if (pdev->vendor == PCI_VENDOR_ID_CDNS && diff --git a/drivers/usb/host/xhci.h b/drivers/usb/host/xhci.h index 8c164340a2c3..779b01dee068 100644 --- a/drivers/usb/host/xhci.h +++ b/drivers/usb/host/xhci.h @@ -1632,7 +1632,7 @@ struct xhci_hcd { #define XHCI_EP_CTX_BROKEN_DCS BIT_ULL(42) #define XHCI_SUSPEND_RESUME_CLKS BIT_ULL(43) #define XHCI_RESET_TO_DEFAULT BIT_ULL(44) -#define XHCI_ZHAOXIN_TRB_FETCH BIT_ULL(45) +#define XHCI_TRB_OVERFETCH BIT_ULL(45) #define XHCI_ZHAOXIN_HOST BIT_ULL(46) #define XHCI_WRITE_64_HI_LO BIT_ULL(47) #define XHCI_CDNS_SCTX_QUIRK BIT_ULL(48)

4 months

1
0
0 0

FAILED: patch "[PATCH] usb: xhci: Enable the TRB overfetch quirk on VIA VL805" failed to apply to 6.12-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 6.12-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. To reproduce the conflict and resubmit, you may use the following commands: git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-6.12.y git checkout FETCH_HEAD git cherry-pick -x c133ec0e5717868c9967fa3df92a55e537b1aead # <resolve conflicts, build, test, etc.> git commit -s git send-email --to '<stable(a)vger.kernel.org>' --in-reply-to '2025030958-june-lard-2d9f@gregkh' --subject-prefix 'PATCH 6.12.y' HEAD^.. Possible dependencies: thanks, greg k-h ------------------ original commit in Linus's tree ------------------ From c133ec0e5717868c9967fa3df92a55e537b1aead Mon Sep 17 00:00:00 2001 From: Michal Pecio <michal.pecio(a)gmail.com> Date: Tue, 25 Feb 2025 11:59:27 +0200 Subject: [PATCH] usb: xhci: Enable the TRB overfetch quirk on VIA VL805 Raspberry Pi is a major user of those chips and they discovered a bug - when the end of a transfer ring segment is reached, up to four TRBs can be prefetched from the next page even if the segment ends with link TRB and on page boundary (the chip claims to support standard 4KB pages). It also appears that if the prefetched TRBs belong to a different ring whose doorbell is later rung, they may be used without refreshing from system RAM and the endpoint will stay idle if their cycle bit is stale. Other users complain about IOMMU faults on x86 systems, unsurprisingly. Deal with it by using existing quirk which allocates a dummy page after each transfer ring segment. This was seen to resolve both problems. RPi came up with a more efficient solution, shortening each segment by four TRBs, but it complicated the driver and they ditched it for this quirk. Also rename the quirk and add VL805 device ID macro. Signed-off-by: Michal Pecio <michal.pecio(a)gmail.com> Link: https://github.com/raspberrypi/linux/issues/4685 Closes: https://bugzilla.kernel.org/show_bug.cgi?id=215906 CC: stable(a)vger.kernel.org Signed-off-by: Mathias Nyman <mathias.nyman(a)linux.intel.com> Link: https://lore.kernel.org/r/20250225095927.2512358-2-mathias.nyman@linux.inte… Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> diff --git a/drivers/usb/host/xhci-mem.c b/drivers/usb/host/xhci-mem.c index 92703efda1f7..fdf0c1008225 100644 --- a/drivers/usb/host/xhci-mem.c +++ b/drivers/usb/host/xhci-mem.c @@ -2437,7 +2437,8 @@ int xhci_mem_init(struct xhci_hcd *xhci, gfp_t flags) * and our use of dma addresses in the trb_address_map radix tree needs * TRB_SEGMENT_SIZE alignment, so we pick the greater alignment need. */ - if (xhci->quirks & XHCI_ZHAOXIN_TRB_FETCH) + if (xhci->quirks & XHCI_TRB_OVERFETCH) + /* Buggy HC prefetches beyond segment bounds - allocate dummy space at the end */ xhci->segment_pool = dma_pool_create("xHCI ring segments", dev, TRB_SEGMENT_SIZE * 2, TRB_SEGMENT_SIZE * 2, xhci->page_size * 2); else diff --git a/drivers/usb/host/xhci-pci.c b/drivers/usb/host/xhci-pci.c index ad0ff356f6fa..54460d11f7ee 100644 --- a/drivers/usb/host/xhci-pci.c +++ b/drivers/usb/host/xhci-pci.c @@ -38,6 +38,8 @@ #define PCI_DEVICE_ID_ETRON_EJ168 0x7023 #define PCI_DEVICE_ID_ETRON_EJ188 0x7052 +#define PCI_DEVICE_ID_VIA_VL805 0x3483 + #define PCI_DEVICE_ID_INTEL_LYNXPOINT_XHCI 0x8c31 #define PCI_DEVICE_ID_INTEL_LYNXPOINT_LP_XHCI 0x9c31 #define PCI_DEVICE_ID_INTEL_WILDCATPOINT_LP_XHCI 0x9cb1 @@ -418,8 +420,10 @@ static void xhci_pci_quirks(struct device *dev, struct xhci_hcd *xhci) pdev->device == 0x3432) xhci->quirks |= XHCI_BROKEN_STREAMS; - if (pdev->vendor == PCI_VENDOR_ID_VIA && pdev->device == 0x3483) + if (pdev->vendor == PCI_VENDOR_ID_VIA && pdev->device == PCI_DEVICE_ID_VIA_VL805) { xhci->quirks |= XHCI_LPM_SUPPORT; + xhci->quirks |= XHCI_TRB_OVERFETCH; + } if (pdev->vendor == PCI_VENDOR_ID_ASMEDIA && pdev->device == PCI_DEVICE_ID_ASMEDIA_1042_XHCI) { @@ -467,11 +471,11 @@ static void xhci_pci_quirks(struct device *dev, struct xhci_hcd *xhci) if (pdev->device == 0x9202) { xhci->quirks |= XHCI_RESET_ON_RESUME; - xhci->quirks |= XHCI_ZHAOXIN_TRB_FETCH; + xhci->quirks |= XHCI_TRB_OVERFETCH; } if (pdev->device == 0x9203) - xhci->quirks |= XHCI_ZHAOXIN_TRB_FETCH; + xhci->quirks |= XHCI_TRB_OVERFETCH; } if (pdev->vendor == PCI_VENDOR_ID_CDNS && diff --git a/drivers/usb/host/xhci.h b/drivers/usb/host/xhci.h index 8c164340a2c3..779b01dee068 100644 --- a/drivers/usb/host/xhci.h +++ b/drivers/usb/host/xhci.h @@ -1632,7 +1632,7 @@ struct xhci_hcd { #define XHCI_EP_CTX_BROKEN_DCS BIT_ULL(42) #define XHCI_SUSPEND_RESUME_CLKS BIT_ULL(43) #define XHCI_RESET_TO_DEFAULT BIT_ULL(44) -#define XHCI_ZHAOXIN_TRB_FETCH BIT_ULL(45) +#define XHCI_TRB_OVERFETCH BIT_ULL(45) #define XHCI_ZHAOXIN_HOST BIT_ULL(46) #define XHCI_WRITE_64_HI_LO BIT_ULL(47) #define XHCI_CDNS_SCTX_QUIRK BIT_ULL(48)

4 months

1
0
0 0

FAILED: patch "[PATCH] mm/slab/kvfree_rcu: Switch to WQ_MEM_RECLAIM wq" failed to apply to 6.13-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 6.13-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. To reproduce the conflict and resubmit, you may use the following commands: git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-6.13.y git checkout FETCH_HEAD git cherry-pick -x dfd3df31c9db752234d7d2e09bef2aeabb643ce4 # <resolve conflicts, build, test, etc.> git commit -s git send-email --to '<stable(a)vger.kernel.org>' --in-reply-to '2025030913-contact-unit-647c@gregkh' --subject-prefix 'PATCH 6.13.y' HEAD^.. Possible dependencies: thanks, greg k-h ------------------ original commit in Linus's tree ------------------ From dfd3df31c9db752234d7d2e09bef2aeabb643ce4 Mon Sep 17 00:00:00 2001 From: "Uladzislau Rezki (Sony)" <urezki(a)gmail.com> Date: Fri, 28 Feb 2025 13:13:56 +0100 Subject: [PATCH] mm/slab/kvfree_rcu: Switch to WQ_MEM_RECLAIM wq Currently kvfree_rcu() APIs use a system workqueue which is "system_unbound_wq" to driver RCU machinery to reclaim a memory. Recently, it has been noted that the following kernel warning can be observed: <snip> workqueue: WQ_MEM_RECLAIM nvme-wq:nvme_scan_work is flushing !WQ_MEM_RECLAIM events_unbound:kfree_rcu_work WARNING: CPU: 21 PID: 330 at kernel/workqueue.c:3719 check_flush_dependency+0x112/0x120 Modules linked in: intel_uncore_frequency(E) intel_uncore_frequency_common(E) skx_edac(E) ... CPU: 21 UID: 0 PID: 330 Comm: kworker/u144:6 Tainted: G E 6.13.2-0_g925d379822da #1 Hardware name: Wiwynn Twin Lakes MP/Twin Lakes Passive MP, BIOS YMM20 02/01/2023 Workqueue: nvme-wq nvme_scan_work RIP: 0010:check_flush_dependency+0x112/0x120 Code: 05 9a 40 14 02 01 48 81 c6 c0 00 00 00 48 8b 50 18 48 81 c7 c0 00 00 00 48 89 f9 48 ... RSP: 0018:ffffc90000df7bd8 EFLAGS: 00010082 RAX: 000000000000006a RBX: ffffffff81622390 RCX: 0000000000000027 RDX: 00000000fffeffff RSI: 000000000057ffa8 RDI: ffff88907f960c88 RBP: 0000000000000000 R08: ffffffff83068e50 R09: 000000000002fffd R10: 0000000000000004 R11: 0000000000000000 R12: ffff8881001a4400 R13: 0000000000000000 R14: ffff88907f420fb8 R15: 0000000000000000 FS: 0000000000000000(0000) GS:ffff88907f940000(0000) knlGS:0000000000000000 CR2: 00007f60c3001000 CR3: 000000107d010005 CR4: 00000000007726f0 PKRU: 55555554 Call Trace: <TASK> ? __warn+0xa4/0x140 ? check_flush_dependency+0x112/0x120 ? report_bug+0xe1/0x140 ? check_flush_dependency+0x112/0x120 ? handle_bug+0x5e/0x90 ? exc_invalid_op+0x16/0x40 ? asm_exc_invalid_op+0x16/0x20 ? timer_recalc_next_expiry+0x190/0x190 ? check_flush_dependency+0x112/0x120 ? check_flush_dependency+0x112/0x120 __flush_work.llvm.1643880146586177030+0x174/0x2c0 flush_rcu_work+0x28/0x30 kvfree_rcu_barrier+0x12f/0x160 kmem_cache_destroy+0x18/0x120 bioset_exit+0x10c/0x150 disk_release.llvm.6740012984264378178+0x61/0xd0 device_release+0x4f/0x90 kobject_put+0x95/0x180 nvme_put_ns+0x23/0xc0 nvme_remove_invalid_namespaces+0xb3/0xd0 nvme_scan_work+0x342/0x490 process_scheduled_works+0x1a2/0x370 worker_thread+0x2ff/0x390 ? pwq_release_workfn+0x1e0/0x1e0 kthread+0xb1/0xe0 ? __kthread_parkme+0x70/0x70 ret_from_fork+0x30/0x40 ? __kthread_parkme+0x70/0x70 ret_from_fork_asm+0x11/0x20 </TASK> ---[ end trace 0000000000000000 ]--- <snip> To address this switch to use of independent WQ_MEM_RECLAIM workqueue, so the rules are not violated from workqueue framework point of view. Apart of that, since kvfree_rcu() does reclaim memory it is worth to go with WQ_MEM_RECLAIM type of wq because it is designed for this purpose. Fixes: 6c6c47b063b5 ("mm, slab: call kvfree_rcu_barrier() from kmem_cache_destroy()"), Reported-by: Keith Busch <kbusch(a)kernel.org> Closes: https://lore.kernel.org/all/Z7iqJtCjHKfo8Kho@kbusch-mbp/ Cc: stable(a)vger.kernel.org Signed-off-by: Uladzislau Rezki (Sony) <urezki(a)gmail.com> Reviewed-by: Joel Fernandes <joelagnelf(a)nvidia.com> Signed-off-by: Vlastimil Babka <vbabka(a)suse.cz> diff --git a/mm/slab_common.c b/mm/slab_common.c index 4030907b6b7d..4c9f0a87f733 100644 --- a/mm/slab_common.c +++ b/mm/slab_common.c @@ -1304,6 +1304,8 @@ module_param(rcu_min_cached_objs, int, 0444); static int rcu_delay_page_cache_fill_msec = 5000; module_param(rcu_delay_page_cache_fill_msec, int, 0444); +static struct workqueue_struct *rcu_reclaim_wq; + /* Maximum number of jiffies to wait before draining a batch. */ #define KFREE_DRAIN_JIFFIES (5 * HZ) #define KFREE_N_BATCHES 2 @@ -1632,10 +1634,10 @@ __schedule_delayed_monitor_work(struct kfree_rcu_cpu *krcp) if (delayed_work_pending(&krcp->monitor_work)) { delay_left = krcp->monitor_work.timer.expires - jiffies; if (delay < delay_left) - mod_delayed_work(system_unbound_wq, &krcp->monitor_work, delay); + mod_delayed_work(rcu_reclaim_wq, &krcp->monitor_work, delay); return; } - queue_delayed_work(system_unbound_wq, &krcp->monitor_work, delay); + queue_delayed_work(rcu_reclaim_wq, &krcp->monitor_work, delay); } static void @@ -1733,7 +1735,7 @@ kvfree_rcu_queue_batch(struct kfree_rcu_cpu *krcp) // "free channels", the batch can handle. Break // the loop since it is done with this CPU thus // queuing an RCU work is _always_ success here. - queued = queue_rcu_work(system_unbound_wq, &krwp->rcu_work); + queued = queue_rcu_work(rcu_reclaim_wq, &krwp->rcu_work); WARN_ON_ONCE(!queued); break; } @@ -1883,7 +1885,7 @@ run_page_cache_worker(struct kfree_rcu_cpu *krcp) if (rcu_scheduler_active == RCU_SCHEDULER_RUNNING && !atomic_xchg(&krcp->work_in_progress, 1)) { if (atomic_read(&krcp->backoff_page_cache_fill)) { - queue_delayed_work(system_unbound_wq, + queue_delayed_work(rcu_reclaim_wq, &krcp->page_cache_work, msecs_to_jiffies(rcu_delay_page_cache_fill_msec)); } else { @@ -2120,6 +2122,10 @@ void __init kvfree_rcu_init(void) int i, j; struct shrinker *kfree_rcu_shrinker; + rcu_reclaim_wq = alloc_workqueue("kvfree_rcu_reclaim", + WQ_UNBOUND | WQ_MEM_RECLAIM, 0); + WARN_ON(!rcu_reclaim_wq); + /* Clamp it to [0:100] seconds interval. */ if (rcu_delay_page_cache_fill_msec < 0 || rcu_delay_page_cache_fill_msec > 100 * MSEC_PER_SEC) {

4 months

1
0
0 0

FAILED: patch "[PATCH] mm/hugetlb: wait for hugetlb folios to be freed" failed to apply to 5.4-stable tree

by gregkh＠linuxfoundation.org

The patch below does not apply to the 5.4-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <stable(a)vger.kernel.org>. To reproduce the conflict and resubmit, you may use the following commands: git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-5.4.y git checkout FETCH_HEAD git cherry-pick -x 67bab13307c83fb742c2556b06cdc39dbad27f07 # <resolve conflicts, build, test, etc.> git commit -s git send-email --to '<stable(a)vger.kernel.org>' --in-reply-to '2025030908-defacing-rumor-448c@gregkh' --subject-prefix 'PATCH 5.4.y' HEAD^.. Possible dependencies: thanks, greg k-h ------------------ original commit in Linus's tree ------------------ From 67bab13307c83fb742c2556b06cdc39dbad27f07 Mon Sep 17 00:00:00 2001 From: Ge Yang <yangge1116(a)126.com> Date: Wed, 19 Feb 2025 11:46:44 +0800 Subject: [PATCH] mm/hugetlb: wait for hugetlb folios to be freed Since the introduction of commit c77c0a8ac4c52 ("mm/hugetlb: defer freeing of huge pages if in non-task context"), which supports deferring the freeing of hugetlb pages, the allocation of contiguous memory through cma_alloc() may fail probabilistically. In the CMA allocation process, if it is found that the CMA area is occupied by in-use hugetlb folios, these in-use hugetlb folios need to be migrated to another location. When there are no available hugetlb folios in the free hugetlb pool during the migration of in-use hugetlb folios, new folios are allocated from the buddy system. A temporary state is set on the newly allocated folio. Upon completion of the hugetlb folio migration, the temporary state is transferred from the new folios to the old folios. Normally, when the old folios with the temporary state are freed, it is directly released back to the buddy system. However, due to the deferred freeing of hugetlb pages, the PageBuddy() check fails, ultimately leading to the failure of cma_alloc(). Here is a simplified call trace illustrating the process: cma_alloc() ->__alloc_contig_migrate_range() // Migrate in-use hugetlb folios ->unmap_and_move_huge_page() ->folio_putback_hugetlb() // Free old folios ->test_pages_isolated() ->__test_page_isolated_in_pageblock() ->PageBuddy(page) // Check if the page is in buddy To resolve this issue, we have implemented a function named wait_for_freed_hugetlb_folios(). This function ensures that the hugetlb folios are properly released back to the buddy system after their migration is completed. By invoking wait_for_freed_hugetlb_folios() before calling PageBuddy(), we ensure that PageBuddy() will succeed. Link: https://lkml.kernel.org/r/1739936804-18199-1-git-send-email-yangge1116@126.… Fixes: c77c0a8ac4c5 ("mm/hugetlb: defer freeing of huge pages if in non-task context") Signed-off-by: Ge Yang <yangge1116(a)126.com> Reviewed-by: Muchun Song <muchun.song(a)linux.dev> Acked-by: David Hildenbrand <david(a)redhat.com> Cc: Baolin Wang <baolin.wang(a)linux.alibaba.com> Cc: Barry Song <21cnbao(a)gmail.com> Cc: Oscar Salvador <osalvador(a)suse.de> Cc: <stable(a)vger.kernel.org> Signed-off-by: Andrew Morton <akpm(a)linux-foundation.org> diff --git a/include/linux/hugetlb.h b/include/linux/hugetlb.h index ec8c0ccc8f95..dbe76d4f1bfc 100644 --- a/include/linux/hugetlb.h +++ b/include/linux/hugetlb.h @@ -682,6 +682,7 @@ struct huge_bootmem_page { int isolate_or_dissolve_huge_page(struct page *page, struct list_head *list); int replace_free_hugepage_folios(unsigned long start_pfn, unsigned long end_pfn); +void wait_for_freed_hugetlb_folios(void); struct folio *alloc_hugetlb_folio(struct vm_area_struct *vma, unsigned long addr, bool cow_from_owner); struct folio *alloc_hugetlb_folio_nodemask(struct hstate *h, int preferred_nid, @@ -1066,6 +1067,10 @@ static inline int replace_free_hugepage_folios(unsigned long start_pfn, return 0; } +static inline void wait_for_freed_hugetlb_folios(void) +{ +} + static inline struct folio *alloc_hugetlb_folio(struct vm_area_struct *vma, unsigned long addr, bool cow_from_owner) diff --git a/mm/hugetlb.c b/mm/hugetlb.c index 163190e89ea1..811b29f77abf 100644 --- a/mm/hugetlb.c +++ b/mm/hugetlb.c @@ -2943,6 +2943,14 @@ int replace_free_hugepage_folios(unsigned long start_pfn, unsigned long end_pfn) return ret; } +void wait_for_freed_hugetlb_folios(void) +{ + if (llist_empty(&hpage_freelist)) + return; + + flush_work(&free_hpage_work); +} + typedef enum { /* * For either 0/1: we checked the per-vma resv map, and one resv diff --git a/mm/page_isolation.c b/mm/page_isolation.c index c608e9d72865..a051a29e95ad 100644 --- a/mm/page_isolation.c +++ b/mm/page_isolation.c @@ -607,6 +607,16 @@ int test_pages_isolated(unsigned long start_pfn, unsigned long end_pfn, struct zone *zone; int ret; + /* + * Due to the deferred freeing of hugetlb folios, the hugepage folios may + * not immediately release to the buddy system. This can cause PageBuddy() + * to fail in __test_page_isolated_in_pageblock(). To ensure that the + * hugetlb folios are properly released back to the buddy system, we + * invoke the wait_for_freed_hugetlb_folios() function to wait for the + * release to complete. + */ + wait_for_freed_hugetlb_folios(); + /* * Note: pageblock_nr_pages != MAX_PAGE_ORDER. Then, chunks of free * pages are not aligned to pageblock_nr_pages.

4 months

1
0
0 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

Linux-stable-mirror