June 2015 - kernelnewbies@kernelnewbies.org - Kernelnewbies Mailing Lists

set_current_state() use in loops confusion
by Nicholas Mc Guire June 8, 2015

June 8, 2015

Hi ! Noticed that in a number of cases - often while(!kthread_should_stop()){ loops the call to set_current_state(...) is inside the loop - but e.g. in the below case why would that be needed ? schedule() is not changing the state visibly for the loop - so what is the point of setting the state inside the loop ? e.g. net/core/pktgen.c /* Wait for kthread_stop */ while (!kthread_should_stop()) { set_current_state(TASK_INTERRUPTIBLE); schedule(); } from code review it seems to be that this would be equivalent set_current_state(TASK_INTERRUPTIBLE); while (!kthread_should_stop()) { schedule(); } what did I miss ? thx! hofrat

1 1

how to offload NAPI poll function to workqueue
by Chan Kim June 8, 2015

June 8, 2015

Hello, I'm working with linux ver 3.3, Ethernet driver for smsc911x. Current NAPI poll function reads recevie FIFO directly which I want to do with DMA controller. For this DMA, I trigger DMA, sleep with wait_event_interruptible, and get woken up by DMA's ISR with wake_up_interruptible. As you know, NAPI poll function is in interrupt context (softirq) so I cannot sleep there for DMA completion. I want to move the NAPI poll function(reading RX FIFO) to waitqueue(process context) usnig a work_struct. The problem is, NAPI poll function is called by the kernel with two arguments : struct napi_struct *napi and int budget. I want to pass those argument to the work_struct and queue the work_struct to the workqueue (using queue_work function). the work_struct looks like below. (include/linux/workqueue.h) struct work_struct { atomic_long_t data; struct list_head entry; work_func_t func; #ifdef CONFIG_LOCKDEP struct lockdep_map lockdep_map; #endif }; I take that atomic_long_t data is for passing the argument to the work_struct. how can I pass the arguments to the work_struct? I tried this (I added in the structure for device driver struct smsc911x_data a member struct work_struct rx_work for passing the work.) : struct work_arg { // a new struct for pass the arguments struct napi_struct *napi; int budget; }; /* NAPI poll function */ static int smsc911x_poll(struct napi_struct *napi, int budget) { struct smsc911x_data *pdata = container_of(napi, struct smsc911x_data, napi); struct net_device *dev = pdata->dev; int npackets = 0; if (enable_rx_use_dma == 1) { // when using DMA for FIFO read prom_printf("moving it to workqueue\n"); struct work_arg *p; p = kzalloc(sizeof(struct work_arg), GFP_KERNEL); p->napi = napi; p->budget = budget; pdata->rx_work.data = (atomic_long_t) p; // <== THIS LINE prom_printf("queue work, with napi = %x, budget = %d\n", napi, budget); queue_work(rx_work_workqueue, &pdata->rx_work); // smsc911x_poll_work } else { -- original NAP poll function, reads FIFO until it's empty and enables the RX interrupt and -- keeps the number of processed packets to npackets. return npackets; } For "THIS LINE" above, I'm getting error during compile. with pdata->rx_work.data = p; , I get error: incompatible types when assigning to type 'atomic_long_t' from type 'struct work_arg *' with pdata->rx_work.data = (atomic_long_t) p; , I get error: conversion to non-scalar type requested. Also, in the new work function, How can I extract the original argments? I tried this below which gives me errors. /* New work function called by the default worker thread */ static int smsc911x_poll_work(struct work_struct *work) { struct smsc911x_data *pdata = container_of(work, struct smsc911x_data, rx_work); struct net_device *dev = pdata->dev; int npackets = 0; struct napi_struct *napi = (struct work_struct *)work->data.napi; // <== THIS LINE int budget = (struct work_struct *)work->data.budget; // <== THIS LINE .. } From the above 'THIS LINE's, I get erros below. error: 'atomic_long_t' has no member named 'napi' error: 'atomic_long_t' has no member named 'budget' and I don't know how to pass the return value to the original NAPI poll functino caller. So my questions are : 1. How do I pass the NAPI poll function arguments to the work_struct? 2. How can I get the NAPI poll functino arguments back from the work_struct? (related to Q.1) 3. How can I return the npackets value to the original NAPI poll function caller? I'm not sure if this kind of conversion (from NAPI poll to workqueue) is possible. Sorry for the long questions but any help will be greatly appreciated. Thanks, Chan

1 2

Guidance about minimal configurations for kernel compilation
by Mayur Patil June 7, 2015

June 7, 2015

Hi All, I am struggling to build *small size with minimum time for kernel * *compilation* with vanilla kernels of kernel.org. As mentioned in the makehelp, I have used defconfig but unable to get desirable results. I have followed linuxscratch <http://www.linuxfromscratch.org/lfs/view/development/chapter08/kernel.html> with which *make * *defconfig* is taking about 20 to 30 minutes of time only for compilation and result of that compilation the kernel is also unable to boot. Please guide which configurations should I use to get my kernel build fast and of minimum size. This is my .config file: http://fpaste.org/229718/ Seeking for guidance, Thanks !! *-- * *Regards,Mayur S Patil,A research oriented program learner,Pune, India.*

2 2

How task_struct.rt.time_slice is initialized?
by Simon Guo June 6, 2015

June 6, 2015

Dear alias, I noticed there is a time_slice variable for SCHED_RR policy in task_struct.rt member (I am using 4.1.0 kernel): struct sched_rt_entity { ... unsigned int time_slice; ... }; Per my understanding, the task_struct.rt.time_slice should be initialized to sched_rr_timeslice or RR_TIMESLICE . Like what process 0 has done: #define INIT_TASK(tsk) \ { ... .rt = { \ .run_list = LIST_HEAD_INIT(tsk.rt.run_list), \ .time_slice = RR_TIMESLICE, \ }, ... } However I didn't see somewhere in copy_process() to set time_slice value for a new forked SCHED_RR process. If time_slice is not initialized, time_slice will be with value 0 and will be overflow to a big value in following "if (--p->rt.time_slice)" statement: static void task_tick_rt(struct rq *rq, struct task_struct *p, int queued) { ... /* * RR tasks need a special form of timeslice management. * FIFO tasks have no timeslices. */ if (p->policy != SCHED_RR) return; if (--p->rt.time_slice) return; p->rt.time_slice = sched_rr_timeslice; ... } Is it a bug? Please correct me if I am wrong. Thanks, Simon

1 1

About Minimum HDD size for Kernel Compilation
by Mayur Patil June 5, 2015

June 5, 2015

Hi All, This is very beginner type question. If I have downloaded any kernel from kernel.org site, I want to install it how much *minimum size of H I require to compile and install it*? My experience: I have compiled 3.10 kernel on VMware Workstation which had taken 40 to 60 GB of Hard Disk, I want to teach someone linux kernel compilation it is not possible that each one will have so much of large amount of hard disk size. Kindly do the needful, Thanks !! *--Regards,Mayur S Patil,A research oriented program learner,Pune, India.*

3 2

RE: I want to sleep while holding a mutex
by Chan Kim June 5, 2015

June 5, 2015

> Yes that's correct. However I would recommend you look into how long the > DMA bursts are before moving to a mutex. In addition softirqs are not > internally locked and need to be implemented by the user for your > information. > Hope this Helps, > Nick Ok, thanks, I know the DMA burst can be quite long so I'll try sticking to mutex. I checked softirq and tasklet are in interrupt context, so I'll move the FIFO read stuff to workqueue. Best regards, Chan 보낸 사람 : "nick" <xerofoify(a)gmail.com> 보낸 날짜 : 2015-06-05 22:04:21 ( +09:00 ) 받는 사람 : 김찬 <ckim(a)etri.re.kr>, kernelnewbies(a)kernelnewbies.org <kernelnewbies(a)kernelnewbies.org> 참조 : 제목 : Re: I want to sleep while holding a mutex On 2015-06-05 09:01 AM, Chan Kim wrote: > >> While here is your issue your trying to sleep in an atomic/ non sleep able >> context of your driver. >> My best guess is in interrupt context but it could be somewhere else related >> to your hardware. >> Nick > > For your information, I'm using smsc911x driver in linux ver 3.3. > The FIFO burst write function is called inside ndo->start_xmit function (net device operation, I think it's process context) > and FIFO burst read function is called from NAPI poll function. (it's in softirq context, is this a problem?). > I remember somewhere reading the softirq is like interrupt context except the difference that interrupt is enabled during softirq processing. > If softirq is where you should not sleep (in my case for DMA), then, I'll have to move it to workqueue or tasklet. > Do you think my understanding sounds more like it? > Chan > Yes that's correct. However I would recommend you look into how long the DMA bursts are before moving to a mutex. In addition softirqs are not internally locked and need to be implemented by the user for your information. Hope this Helps, Nick > > > 보낸 사람 : "nick" > 보낸 날짜 : 2015-06-05 21:48:18 ( +09:00 ) > 받는 사람 : 김찬 , kernelnewbies(a)kernelnewbies.org > 참조 : > 제목 : Re: I want to sleep while holding a mutex > > > > On 2015-06-05 08:45 AM, Chan Kim wrote: >> >>>>> Are you trying to use the mutex during a interrupt handler as you >>>>> can't you need something atomic or not able to sleep like a spinlock. >>>>> Nick >>>> >>>> I use mutex during FIFO burst write or burst read. I don't want other >>> processes to mess up the FIFO access. >>>> spinlock is not for long interval I understand. >>>> Chan >>>> >>> That's during interrupt context. What you need to do is offload the data into >>> main memory during interrupt context and lock with a spinlock. Afterwards >>> you can use a mutex around the main work function called after the interrupt. >>> Nick >> >> >> I believe I lock_mutex and unlock_mutex are called in the process context. Both for transmission and for reception (rx is from softirq, it's ok.). It's only that the process is woken up by the interrupt routine. (ISR calls wake_up_interruptible). ISR doesn't lock or unlock any mutex. >> >> Chan >> > While here is your issue your trying to sleep in an atomic/ non sleep able context of your driver. > My best guess is in interrupt context but it could be somewhere else related to your hardware. > Nick >> >> >> 보낸 사람 : "nick" >> 보낸 날짜 : 2015-06-05 21:40:20 ( +09:00 ) >> 받는 사람 : 김찬 , kernelnewbies(a)kernelnewbies.org >> 참조 : >> 제목 : Re: I want to sleep while holding a mutex >> >> >> >> On 2015-06-05 08:37 AM, 김찬 wrote: >>> >>>> Are you trying to use the mutex during a interrupt handler as you can't you >>>> need something atomic or not able to sleep like a spinlock. >>>> Nick >>> >>> I use mutex during FIFO burst write or burst read. I don't want other processes to mess up the FIFO access. >>> spinlock is not for long interval I understand. >>> Chan >>> >> That's during interrupt context. What you need to do is offload the data into main memory during interrupt >> context and lock with a spinlock. Afterwards you can use a mutex around the main work function called >> after the interrupt. >> Nick >>> ________________________________ >>> 보낸 사람 : "nick" >>> 보낸 날짜 : 2015-06-05 21:34:25 ( +09:00 ) >>> 받는 사람 : 김찬 , kernelnewbies(a)kernelnewbies.org >>> 참조 : >>> 제목 : Re: I want to sleep while holding a mutex >>> >>> >>> >>> On 2015-06-05 08:30 AM, Chan Kim wrote: >>>> Hello, >>>> I need to write to and read from a Ethernet chip's FIFO and for that I want to use DMA controller. So I modified an existing driver and made the process to sleep using wait_event_interruptible after triggering DMA transfer and made it woken up by ISR using wake_up_interruptible. My problem is that the original driver was using spinlock to protect the FIFO access. But as all know, while holding spinlock, I should not sleep. So I tried replacing spinlock with mutex. (Because, I might sleep while acquiring the mutex, and after getting the mutex, I'll sleep again after DMA trigger. I get woken up by the ISR, and I'll release the mutex. No problem! I thought. ) But I still get this 'scheduling while atomic BUG'. The question is : I need to go to sleep until DMA completion while holding a lock (for FIFO read and write). Isn't it ok to sleep while holding a mutex? >>>> Best regards, >>>> Chan Kim >>>> _______________________________________________ >>>> Kernelnewbies mailing list >>>> Kernelnewbies(a)kernelnewbies.org >>>> http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies >>>> >>> Are you trying to use the mutex during a interrupt handler as you can't you need something >>> atomic or not able to sleep like a spinlock. >>> Nick

1 0

RE: I want to sleep while holding a mutex
by Chan Kim June 5, 2015

June 5, 2015

> While here is your issue your trying to sleep in an atomic/ non sleep able > context of your driver. > My best guess is in interrupt context but it could be somewhere else related > to your hardware. > Nick For your information, I'm using smsc911x driver in linux ver 3.3. The FIFO burst write function is called inside ndo->start_xmit function (net device operation, I think it's process context) and FIFO burst read function is called from NAPI poll function. (it's in softirq context, is this a problem?). I remember somewhere reading the softirq is like interrupt context except the difference that interrupt is enabled during softirq processing. If softirq is where you should not sleep (in my case for DMA), then, I'll have to move it to workqueue or tasklet. Do you think my understanding sounds more like it? Chan 보낸 사람 : "nick" <xerofoify(a)gmail.com> 보낸 날짜 : 2015-06-05 21:48:18 ( +09:00 ) 받는 사람 : 김찬 <ckim(a)etri.re.kr>, kernelnewbies(a)kernelnewbies.org <kernelnewbies(a)kernelnewbies.org> 참조 : 제목 : Re: I want to sleep while holding a mutex On 2015-06-05 08:45 AM, Chan Kim wrote: > >>>> Are you trying to use the mutex during a interrupt handler as you >>>> can't you need something atomic or not able to sleep like a spinlock. >>>> Nick >>> >>> I use mutex during FIFO burst write or burst read. I don't want other >> processes to mess up the FIFO access. >>> spinlock is not for long interval I understand. >>> Chan >>> >> That's during interrupt context. What you need to do is offload the data into >> main memory during interrupt context and lock with a spinlock. Afterwards >> you can use a mutex around the main work function called after the interrupt. >> Nick > > > I believe I lock_mutex and unlock_mutex are called in the process context. Both for transmission and for reception (rx is from softirq, it's ok.). It's only that the process is woken up by the interrupt routine. (ISR calls wake_up_interruptible). ISR doesn't lock or unlock any mutex. > > Chan > While here is your issue your trying to sleep in an atomic/ non sleep able context of your driver. My best guess is in interrupt context but it could be somewhere else related to your hardware. Nick > > > 보낸 사람 : "nick" > 보낸 날짜 : 2015-06-05 21:40:20 ( +09:00 ) > 받는 사람 : 김찬 , kernelnewbies(a)kernelnewbies.org > 참조 : > 제목 : Re: I want to sleep while holding a mutex > > > > On 2015-06-05 08:37 AM, 김찬 wrote: >> >>> Are you trying to use the mutex during a interrupt handler as you can't you >>> need something atomic or not able to sleep like a spinlock. >>> Nick >> >> I use mutex during FIFO burst write or burst read. I don't want other processes to mess up the FIFO access. >> spinlock is not for long interval I understand. >> Chan >> > That's during interrupt context. What you need to do is offload the data into main memory during interrupt > context and lock with a spinlock. Afterwards you can use a mutex around the main work function called > after the interrupt. > Nick >> ________________________________ >> 보낸 사람 : "nick" >> 보낸 날짜 : 2015-06-05 21:34:25 ( +09:00 ) >> 받는 사람 : 김찬 , kernelnewbies(a)kernelnewbies.org >> 참조 : >> 제목 : Re: I want to sleep while holding a mutex >> >> >> >> On 2015-06-05 08:30 AM, Chan Kim wrote: >>> Hello, >>> I need to write to and read from a Ethernet chip's FIFO and for that I want to use DMA controller. So I modified an existing driver and made the process to sleep using wait_event_interruptible after triggering DMA transfer and made it woken up by ISR using wake_up_interruptible. My problem is that the original driver was using spinlock to protect the FIFO access. But as all know, while holding spinlock, I should not sleep. So I tried replacing spinlock with mutex. (Because, I might sleep while acquiring the mutex, and after getting the mutex, I'll sleep again after DMA trigger. I get woken up by the ISR, and I'll release the mutex. No problem! I thought. ) But I still get this 'scheduling while atomic BUG'. The question is : I need to go to sleep until DMA completion while holding a lock (for FIFO read and write). Isn't it ok to sleep while holding a mutex? >>> Best regards, >>> Chan Kim >>> _______________________________________________ >>> Kernelnewbies mailing list >>> Kernelnewbies(a)kernelnewbies.org >>> http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies >>> >> Are you trying to use the mutex during a interrupt handler as you can't you need something >> atomic or not able to sleep like a spinlock. >> Nick

1 0

RE: I want to sleep while holding a mutex
by Chan Kim June 5, 2015

June 5, 2015

> >> Are you trying to use the mutex during a interrupt handler as you > >> can't you need something atomic or not able to sleep like a spinlock. > >> Nick > > > > I use mutex during FIFO burst write or burst read. I don't want other > processes to mess up the FIFO access. > > spinlock is not for long interval I understand. > > Chan > > > That's during interrupt context. What you need to do is offload the data into > main memory during interrupt context and lock with a spinlock. Afterwards > you can use a mutex around the main work function called after the interrupt. > Nick I believe I lock_mutex and unlock_mutex are called in the process context. Both for transmission and for reception (rx is from softirq, it's ok.). It's only that the process is woken up by the interrupt routine. (ISR calls wake_up_interruptible). ISR doesn't lock or unlock any mutex. Chan 보낸 사람 : "nick" <xerofoify(a)gmail.com> 보낸 날짜 : 2015-06-05 21:40:20 ( +09:00 ) 받는 사람 : 김찬 <ckim(a)etri.re.kr>, kernelnewbies(a)kernelnewbies.org <kernelnewbies(a)kernelnewbies.org> 참조 : 제목 : Re: I want to sleep while holding a mutex On 2015-06-05 08:37 AM, 김찬 wrote: > >> Are you trying to use the mutex during a interrupt handler as you can't you >> need something atomic or not able to sleep like a spinlock. >> Nick > > I use mutex during FIFO burst write or burst read. I don't want other processes to mess up the FIFO access. > spinlock is not for long interval I understand. > Chan > That's during interrupt context. What you need to do is offload the data into main memory during interrupt context and lock with a spinlock. Afterwards you can use a mutex around the main work function called after the interrupt. Nick > ________________________________ > 보낸 사람 : "nick" > 보낸 날짜 : 2015-06-05 21:34:25 ( +09:00 ) > 받는 사람 : 김찬 , kernelnewbies(a)kernelnewbies.org > 참조 : > 제목 : Re: I want to sleep while holding a mutex > > > > On 2015-06-05 08:30 AM, Chan Kim wrote: >> Hello, >> I need to write to and read from a Ethernet chip's FIFO and for that I want to use DMA controller. So I modified an existing driver and made the process to sleep using wait_event_interruptible after triggering DMA transfer and made it woken up by ISR using wake_up_interruptible. My problem is that the original driver was using spinlock to protect the FIFO access. But as all know, while holding spinlock, I should not sleep. So I tried replacing spinlock with mutex. (Because, I might sleep while acquiring the mutex, and after getting the mutex, I'll sleep again after DMA trigger. I get woken up by the ISR, and I'll release the mutex. No problem! I thought. ) But I still get this 'scheduling while atomic BUG'. The question is : I need to go to sleep until DMA completion while holding a lock (for FIFO read and write). Isn't it ok to sleep while holding a mutex? >> Best regards, >> Chan Kim >> _______________________________________________ >> Kernelnewbies mailing list >> Kernelnewbies(a)kernelnewbies.org >> http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies >> > Are you trying to use the mutex during a interrupt handler as you can't you need something > atomic or not able to sleep like a spinlock. > Nick >

1 0

RE: I want to sleep while holding a mutex
by 김찬 June 5, 2015

June 5, 2015

> Are you trying to use the mutex during a interrupt handler as you can't you > need something atomic or not able to sleep like a spinlock. > Nick I use mutex during FIFO burst write or burst read. I don't want other processes to mess up the FIFO access. spinlock is not for long interval I understand. Chan ________________________________ 보낸 사람 : "nick" <xerofoify(a)gmail.com> 보낸 날짜 : 2015-06-05 21:34:25 ( +09:00 ) 받는 사람 : 김찬 <ckim(a)etri.re.kr>, kernelnewbies(a)kernelnewbies.org <kernelnewbies(a)kernelnewbies.org> 참조 : 제목 : Re: I want to sleep while holding a mutex On 2015-06-05 08:30 AM, Chan Kim wrote: > Hello, > I need to write to and read from a Ethernet chip's FIFO and for that I want to use DMA controller. So I modified an existing driver and made the process to sleep using wait_event_interruptible after triggering DMA transfer and made it woken up by ISR using wake_up_interruptible. My problem is that the original driver was using spinlock to protect the FIFO access. But as all know, while holding spinlock, I should not sleep. So I tried replacing spinlock with mutex. (Because, I might sleep while acquiring the mutex, and after getting the mutex, I'll sleep again after DMA trigger. I get woken up by the ISR, and I'll release the mutex. No problem! I thought. ) But I still get this 'scheduling while atomic BUG'. The question is : I need to go to sleep until DMA completion while holding a lock (for FIFO read and write). Isn't it ok to sleep while holding a mutex? > Best regards, > Chan Kim > _______________________________________________ > Kernelnewbies mailing list > Kernelnewbies(a)kernelnewbies.org > http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies > Are you trying to use the mutex during a interrupt handler as you can't you need something atomic or not able to sleep like a spinlock. Nick

1 0

I want to sleep while holding a mutex
by Chan Kim June 5, 2015

June 5, 2015

Hello, I need to write to and read from a Ethernet chip's FIFO and for that I want to use DMA controller. So I modified an existing driver and made the process to sleep using wait_event_interruptible after triggering DMA transfer and made it woken up by ISR using wake_up_interruptible. My problem is that the original driver was using spinlock to protect the FIFO access. But as all know, while holding spinlock, I should not sleep. So I tried replacing spinlock with mutex. (Because, I might sleep while acquiring the mutex, and after getting the mutex, I'll sleep again after DMA trigger. I get woken up by the ISR, and I'll release the mutex. No problem! I thought. ) But I still get this 'scheduling while atomic BUG'. The question is : I need to go to sleep until DMA completion while holding a lock (for FIFO read and write). Isn't it ok to sleep while holding a mutex? Best regards, Chan Kim

1 0