iwlwifi: pcie: limit memory read spin time
Subsystems (9)
Sasha Levin
maintainerPATCH 1/9 iwlwifi: pcie: limit memory read spin time
From: Johannes Berg <[email protected]>
[ Upstream commit 04516706bb99889986ddfa3a769ed50d2dc7ac13 ]
When we read device memory, we lock a spinlock, write the address we want to read from the device and then spin in a loop reading the data in 32-bit quantities from another register.
As the description makes clear, this is rather inefficient, incurring a PCIe bus transaction for every read. In a typical device today, we want to read 786k SMEM if it crashes, leading to 192k register reads. Occasionally, we've seen the whole loop take over 20 seconds and then triggering the soft lockup detector.
Clearly, it is unreasonable to spin here for such extended periods of time.
To fix this, break the loop down into an outer and an inner loop, and break out of the inner loop if more than half a second elapsed. To avoid too much overhead, check for that only every 128 reads, though there's no particular reason for that number. Then, unlock and relock to obtain NIC access again, reprogram the start address and continue.
This will keep (interrupt) latencies on the CPU down to a reasonable time.
--- .../net/wireless/intel/iwlwifi/pcie/trans.c | 36 ++++++++++++++----- 1 file changed, 27 insertions(+), 9 deletions(-)
drivers/net/wireless/intel/iwlwifi/pcie/trans.c +27 -9
@@ -1927,18 +1927,36 @@ static int iwl_trans_pcie_read_mem(struct iwl_trans *trans, u32 addr, void *buf, int dwords) {unsigned long flags; -int offs, ret =0 ; +int offs =0 ;u32 *vals = buf; -if (iwl_trans_grab_nic_access (trans, &flags)) { -iwl_write32 (trans,HBUS_TARG_MEM_RADDR , addr); -for (offs =0 ; offs < dwords; offs++) - vals[offs] =iwl_read32 (trans,HBUS_TARG_MEM_RDAT ); -iwl_trans_release_nic_access (trans, &flags); - }else { - ret = -EBUSY ; +while (offs < dwords) { +/* limit the time we spin here under lock to 1/2s */ +ktime_t timeout =ktime_add_us (ktime_get (),500 *USEC_PER_MSEC ); + +if (iwl_trans_grab_nic_access (trans, &flags)) { +iwl_write32 (trans,HBUS_TARG_MEM_RADDR , + addr +4 * offs); + +while (offs < dwords) { + vals[offs] =iwl_read32 (trans, +HBUS_TARG_MEM_RDAT ); + offs++; + +/* calling ktime_get is expensive so +* do it once in 128 reads +*/ +if (offs %128 ==0 &&ktime_after (ktime_get (), + timeout)) +break ; + } +iwl_trans_release_nic_access (trans, &flags); + }else { +return -EBUSY ; + } } -return ret; + +return 0 ; } static int iwl_trans_pcie_write_mem(struct iwl_trans *trans, u32 addr,
Signature · 1 line
2.27.0
Sasha Levin
maintainerPATCH 2/9 arm64: dts: rockchip: Assign a fixed index to mmc devices on rk3399 boards.
From: Markus Reichl <[email protected]>
[ Upstream commit 0011c6d182774fc781fb9e115ebe8baa356029ae ]
Recently introduced async probe on mmc devices can shuffle block IDs. Pin them to fixed values to ease booting in environments where UUIDs are not practical. Use newly introduced aliases for mmcblk devices from [1].
[1]
https://patchwork.kernel.org/patch/11747669/
--- arch/arm64/boot/dts/rockchip/rk3399.dtsi | 3 +++ 1 file changed, 3 insertions(+)
arch/arm64/boot/dts/rockchip/rk3399.dtsi +3 -0
@@ -66,6 +66,9 @@ aliases { i2c6 = &i2c6; i2c7 = &i2c7; i2c8 = &i2c8; + mmc0 = &sdio0; + mmc1 = &sdmmc; + mmc2 = &sdhci; serial0 = &uart0; serial1 = &uart1; serial2 = &uart2;
Signature · 1 line
2.27.0
Sasha Levin
maintainerPATCH 3/9 iwlwifi: mvm: fix kernel panic in case of assert during CSA
In reply to Sasha LevinFrom: Sara Sharon <[email protected]>
[ Upstream commit fe56d05ee6c87f6a1a8c7267affd92c9438249cc ]
During CSA, we briefly nullify the phy context, in __iwl_mvm_unassign_vif_chanctx. In case we have a FW assert right after it, it remains NULL though. We end up running into endless loop due to mac80211 trying repeatedly to move us to ASSOC state, and we keep returning -EINVAL. Later down the road we hit a kernel panic.
Detect and avoid this endless loop.
--- drivers/net/wireless/intel/iwlwifi/mvm/mac80211.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/wireless/intel/iwlwifi/mvm/mac80211.c +1 -1
@@ -2654,7 +2654,7 @@ static int iwl_mvm_mac_sta_state(struct ieee80211_hw *hw,/* this would be a mac80211 bug ... but don't crash */ if (WARN_ON_ONCE (!mvmvif->phy_ctxt)) -return -EINVAL ; +return test_bit (IWL_MVM_STATUS_HW_RESTART_REQUESTED , &mvm->status) ?0 : -EINVAL ;/* * If we are in a STA removal flow and in DQA mode:
Signature · 1 line
2.27.0
Sasha Levin
maintainerPATCH 4/9 ARC: stack unwinding: don't assume non-current task is sleeping
In reply to Sasha LevinFrom: Vineet Gupta <[email protected]>
[ Upstream commit e42404fa10fd11fe72d0a0e149a321d10e577715 ]
To start stack unwinding (SP, PC and BLINK) are needed. When the explicit execution context (pt_regs etc) is not available, unwinder assumes the task is sleeping (in __switch_to()) and fetches SP and BLINK from kernel mode stack.
But this assumption is not true, specially in a SMP system, when top runs on 1 core, there may be active running processes on all cores.
So when unwinding non courrent tasks, ensure they are NOT running.
And while at it, handle the self unwinding case explicitly.
This came out of investigation of a customer reported hang with rcutorture+top
--- arch/arc/kernel/stacktrace.c | 23 +++++++++++++++-------- 1 file changed, 15 insertions(+), 8 deletions(-)
arch/arc/kernel/stacktrace.c +15 -8
@@ -41,15 +41,15 @@#ifdef CONFIG_ARC_DW2_UNWIND -static void seed_unwind_frame_info (struct task_struct *tsk, -struct pt_regs *regs, -struct unwind_frame_info *frame_info) +static int +seed_unwind_frame_info (struct task_struct *tsk,struct pt_regs *regs, +struct unwind_frame_info *frame_info) {/* * synchronous unwinding (e.g. dump_stack) * - uses current values of SP and friends */ -if (tsk ==NULL && regs ==NULL ) { +if (regs ==NULL && (tsk ==NULL || tsk == current)) {unsigned long fp, sp, blink, ret; frame_info->task = current;
@@ -68,11 +68,15 @@ static void seed_unwind_frame_info(struct task_struct *tsk, frame_info->call_frame =0 ; }else if (regs ==NULL ) {/* -* Asynchronous unwinding of sleeping task -* - Gets SP etc from task's pt_regs (saved bottom of kernel -* mode stack of task) +* Asynchronous unwinding of a likely sleeping task +* - first ensure it is actually sleeping +* - if so, it will be in __switch_to, kernel mode SP of task +* is safe-kept and BLINK at a well known location in there */ +if (tsk->state ==TASK_RUNNING ) +return -1 ; + frame_info->task = tsk; frame_info->regs.r27 =TSK_K_FP (tsk);
@@ -106,6 +110,8 @@ static void seed_unwind_frame_info(struct task_struct *tsk, frame_info->regs.r63 = regs->ret; frame_info->call_frame = 0; } + +return 0 ; }#endif
@@ -119,7 +125,8 @@ arc_unwind_core(struct task_struct *tsk, struct pt_regs *regs,unsigned int address;struct unwind_frame_info frame_info; -seed_unwind_frame_info (tsk, regs, &frame_info); +if (seed_unwind_frame_info (tsk, regs, &frame_info)) +return 0 ;while (1 ) { address =UNW_PC (&frame_info);
Signature · 1 line
2.27.0
Sasha Levin
maintainerPATCH 5/9 scsi: ufs: Make sure clk scaling happens only when HBA is runtime ACTIVE
In reply to Sasha LevinFrom: Can Guo <[email protected]>
[ Upstream commit 73cc291c270248567245f084dcdf5078069af6b5 ]
If someone plays with the UFS clk scaling devfreq governor through sysfs, ufshcd_devfreq_scale may be called even when HBA is not runtime ACTIVE. This can lead to unexpected error. We cannot just protect it by calling pm_runtime_get_sync() because that may cause a race condition since HBA runtime suspend ops need to suspend clk scaling. To fix this call pm_runtime_get_noresume() and check HBA's runtime status. Only proceed if HBA is runtime ACTIVE, otherwise just bail.
governor_store
devfreq_performance_handler
update_devfreq
devfreq_set_target
ufshcd_devfreq_target
ufshcd_devfreq_scale
--- drivers/scsi/ufs/ufshcd.c | 7 +++++++ 1 file changed, 7 insertions(+)
drivers/scsi/ufs/ufshcd.c +7 -0
@@ -1215,8 +1215,15 @@ static int ufshcd_devfreq_target(struct device *dev, }spin_unlock_irqrestore (hba->host->host_lock, irq_flags); +pm_runtime_get_noresume (hba->dev); +if (!pm_runtime_active (hba->dev)) { +pm_runtime_put_noidle (hba->dev); + ret = -EAGAIN ; + goto out; + } start =ktime_get (); ret =ufshcd_devfreq_scale (hba, scale_up); +pm_runtime_put (hba->dev); trace_ufshcd_profile_clk_scaling(dev_name(hba->dev), (scale_up ? "up" : "down"),
Signature · 1 line
2.27.0
Sasha Levin
maintainerPATCH 6/9 usbnet: ipheth: fix connectivity with iOS 14
In reply to Sasha LevinFrom: Yves-Alexis Perez <[email protected]>
[ Upstream commit f33d9e2b48a34e1558b67a473a1fc1d6e793f93c ]
Starting with iOS 14 released in September 2020, connectivity using the personal hotspot USB tethering function of iOS devices is broken.
Communication between the host and the device (for example ICMP traffic or DNS resolution using the DNS service running in the device itself) works fine, but communication to endpoints further away doesn't work.
Investigation on the matter shows that no UDP and ICMP traffic from the tethered host is reaching the Internet at all. For TCP traffic there are exchanges between tethered host and server but packets are modified in transit leading to impossible communication.
After some trials Matti Vuorela discovered that reducing the URB buffer size by two bytes restored the previous behavior. While a better solution might exist to fix the issue, since the protocol is not publicly documented and considering the small size of the fix, let's do that.
--- drivers/net/usb/ipheth.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/usb/ipheth.c +1 -1
@@ -70,7 +70,7 @@#define IPHETH_USBINTF_SUBCLASS 253#define IPHETH_USBINTF_PROTO 1 -#define IPHETH_BUF_SIZE 1516 +#define IPHETH_BUF_SIZE 1514#define IPHETH_IP_ALIGN 2/* padding at front of URB */ #define IPHETH_TX_TIMEOUT (5 * HZ)
Signature · 1 line
2.27.0
Sasha Levin
maintainerPATCH 7/9 tun: honor IOCB_NOWAIT flag
In reply to Sasha LevinFrom: Jens Axboe <[email protected]>
[ Upstream commit 5aac0390a63b8718237a61dd0d24a29201d1c94a ]
tun only checks the file O_NONBLOCK flag, but it should also be checking the iocb IOCB_NOWAIT flag. Any fops using ->read/write_iter() should check both, otherwise it breaks users that correctly expect O_NONBLOCK semantics if IOCB_NOWAIT is set.
--- drivers/net/tun.c | 14 +++++++++++--- 1 file changed, 11 insertions(+), 3 deletions(-)
drivers/net/tun.c +11 -3
@@ -1601,12 +1601,15 @@ static ssize_t tun_chr_write_iter(struct kiocb *iocb, struct iov_iter *from)struct tun_struct *tun =tun_get (file);struct tun_file *tfile = file->private_data;ssize_t result; +int noblock =0 ;if (!tun)return -EBADFD ; - result =tun_get_user (tun, tfile,NULL , from, - file->f_flags &O_NONBLOCK , false); +if ((file->f_flags &O_NONBLOCK ) || (iocb->ki_flags &IOCB_NOWAIT )) + noblock =1 ; + + result =tun_get_user (tun, tfile,NULL , from, noblock, false);tun_put (tun);return result;
@@ -1789,10 +1792,15 @@ static ssize_t tun_chr_read_iter(struct kiocb *iocb, struct iov_iter *to)struct tun_file *tfile = file->private_data;struct tun_struct *tun =__tun_get (tfile);ssize_t len =iov_iter_count (to), ret; +int noblock =0 ;if (!tun)return -EBADFD ; - ret =tun_do_read (tun, tfile, to, file->f_flags &O_NONBLOCK ,NULL ); + +if ((file->f_flags &O_NONBLOCK ) || (iocb->ki_flags &IOCB_NOWAIT )) + noblock =1 ; + + ret =tun_do_read (tun, tfile, to, noblock,NULL ); ret =min_t (ssize_t, ret, len);if (ret >0 ) iocb->ki_pos = ret;
Signature · 1 line
2.27.0
Sasha Levin
maintainerPATCH 8/9 soc: fsl: dpio: Get the cpumask through cpumask_of(cpu)
In reply to Sasha LevinFrom: Hao Si <[email protected]>
[ Upstream commit 2663b3388551230cbc4606a40fabf3331ceb59e4 ]
The local variable 'cpumask_t mask' is in the stack memory, and its address is assigned to 'desc->affinity' in 'irq_set_affinity_hint()'. But the memory area where this variable is located is at risk of being modified.
During LTP testing, the following error was generated:
Unable to handle kernel paging request at virtual address ffff000012e9b790
Mem abort info:
ESR = 0x96000007
Exception class = DABT (current EL), IL = 32 bits SET = 0, FnV = 0
EA = 0, S1PTW = 0
Data abort info:
ISV = 0, ISS = 0x00000007
CM = 0, WnR = 0
swapper pgtable: 4k pages, 48-bit VAs, pgdp = 0000000075ac5e07 [ffff000012e9b790] pgd=00000027dbffe003, pud=00000027dbffd003, pmd=00000027b6d61003, pte=0000000000000000
Internal error: Oops: 96000007 [#1] PREEMPT SMP Modules linked in: xt_conntrack Process read_all (pid: 20171, stack limit = 0x0000000044ea4095) CPU: 14 PID: 20171 Comm: read_all Tainted: G B W Hardware name: NXP Layerscape LX2160ARDB (DT) pstate: 80000085 (Nzcv daIf -PAN -UAO) pc : irq_affinity_hint_proc_show+0x54/0xb0 lr : irq_affinity_hint_proc_show+0x4c/0xb0
sp : ffff00001138bc10
x29: ffff00001138bc10 x28: 0000ffffd131d1e0
x27: 00000000007000c0 x26: ffff8025b9480dc0
x25: ffff8025b9480da8 x24: 00000000000003ff
x23: ffff8027334f8300 x22: ffff80272e97d000
x21: ffff80272e97d0b0 x20: ffff8025b9480d80
x19: ffff000009a49000 x18: 0000000000000000
x17: 0000000000000000 x16: 0000000000000000
x15: 0000000000000000 x14: 0000000000000000
x13: 0000000000000000 x12: 0000000000000040
x11: 0000000000000000 x10: ffff802735b79b88 x9 : 0000000000000000 x8 : 0000000000000000 x7 : ffff000009a49848 x6 : 0000000000000003 x5 : 0000000000000000 x4 : ffff000008157d6c x3 : ffff00001138bc10 x2 : ffff000012e9b790 x1 : 0000000000000000 x0 : 0000000000000000
Call trace: irq_affinity_hint_proc_show+0x54/0xb0 seq_read+0x1b0/0x440 proc_reg_read+0x80/0xd8 __vfs_read+0x60/0x178 vfs_read+0x94/0x150 ksys_read+0x74/0xf0 __arm64_sys_read+0x24/0x30 el0_svc_common.constprop.0+0xd8/0x1a0 el0_svc_handler+0x34/0x88 el0_svc+0x10/0x14 Code: f9001bbf 943e0732 f94066c2 b4000062 (f9400041) ---[ end trace b495bdcb0b3b732b ]--- Kernel panic - not syncing: Fatal exception
SMP: stopping secondary CPUs
SMP: failed to stop secondary CPUs 0,2-4,6,8,11,13-15 Kernel Offset: disabled
CPU features: 0x0,21006008
Memory Limit: none
---[ end Kernel panic - not syncing: Fatal exception ]---
Fix it by using 'cpumask_of(cpu)' to get the cpumask.
--- drivers/staging/fsl-mc/bus/dpio/dpio-driver.c | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-)
drivers/staging/fsl-mc/bus/dpio/dpio-driver.c +1 -4
@@ -77,7 +77,6 @@ static int register_dpio_irq_handlers(struct fsl_mc_device *dpio_dev, int cpu)struct dpio_priv *priv;int error;struct fsl_mc_device_irq *irq; -cpumask_t mask; priv =dev_get_drvdata (&dpio_dev->dev);
@@ -96,9 +95,7 @@ static int register_dpio_irq_handlers(struct fsl_mc_device *dpio_dev, int cpu) }/* set the affinity hint */ -cpumask_clear (&mask); -cpumask_set_cpu (cpu, &mask); -if (irq_set_affinity_hint (irq->msi_desc->irq, &mask)) +if (irq_set_affinity_hint (irq->msi_desc->irq,cpumask_of (cpu)))dev_err (&dpio_dev->dev,"irq_set_affinity failed irq %d cpu %d\n" , irq->msi_desc->irq, cpu);
Signature · 1 line
2.27.0
Sasha Levin
maintainerPATCH 9/9 platform/x86: acer-wmi: add automatic keyboard background light toggle key as KEY_LIGHTS_TOGGLE
In reply to Sasha LevinFrom: Timo Witte <[email protected]>
[ Upstream commit 9e7a005ad56aa7d6ea5830c5ffcc60bf35de380b ]
Got a dmesg message on my AMD Renoir based Acer laptop:
"acer_wmi: Unknown key number - 0x84" when toggling keyboard background light
--- drivers/platform/x86/acer-wmi.c | 1 + 1 file changed, 1 insertion(+)
drivers/platform/x86/acer-wmi.c +1 -0
@@ -124,6 +124,7 @@ static const struct key_entry acer_wmi_keymap[] __initconst = { {KE_KEY, 0x64, {KEY_SWITCHVIDEOMODE} }, /* Display Switch */ {KE_IGNORE, 0x81, {KEY_SLEEP} }, {KE_KEY, 0x82, {KEY_TOUCHPAD_TOGGLE} }, /* Touch Pad Toggle */ + {KE_IGNORE, 0x84, {KEY_KBDILLUMTOGGLE} }, /* Automatic Keyboard background light toggle */ {KE_KEY, KEY_TOUCHPAD_ON, {KEY_TOUCHPAD_ON} }, {KE_KEY, KEY_TOUCHPAD_OFF, {KEY_TOUCHPAD_OFF} }, {KE_IGNORE, 0x83, {KEY_TOUCHPAD_TOGGLE} },
Signature · 1 line
2.27.0