- kernelnewbies@kernelnewbies.org - Kernelnewbies Mailing Lists

Understanding pipes and blocking calls
by Kristian Evensen Sept. 15, 2011

Sept. 15, 2011

Hi, I am working on a simple command line application for Linux, in order to properly understand how pipes and blocking calls interact. The application writes some data to stdout, before sleeping and then writing some more. I have set stdout to be unbuffered, and do an fflush before the sleep to be absolutely sure that there is no buffered data left. When I try to pipe this data to another application, the receiving application does not progress before the sleep is over. Based on my understanding of pipes, the two applications are independent and, thus, the receiving application should not be affected by any sleeps in the producer of data. Is this incorrect, and is there a way to fix this (i.e., having blocking calls in the producer)? Thanks in advance, Kristian

2 1

CPU usage accounting
by Vimal Sept. 15, 2011

Sept. 15, 2011

Hi, In the following cases, how does the CPU work done by the kernel on behalf of the application get accounted for? 1. When an application writes to a TCP/UDP socket, the networking stack does transmit side processing. 2. When an application receives data on a TCP/UDP socket, the networking stack does receive side processing. >From what I understand: In case (1), the kernel code executes in the context of the application, so the CPU cycles are accounted directly to process that called write() (or send/sendto). But in case (2), unless some processing is done, it is impossible to know which application is going to receive this packet. Whom do these cycles get charged to? Thanks, -- Vimal

2 3

writing on mmap() ?
by Andrea Gasparini Sept. 15, 2011

Sept. 15, 2011

Hi, i've got a working version of an mmap()'ed I/O, though read-only. A device puts its data on RAM through dma and userspace can read it (and consequentely dispatch it to other devices). Now, I need to write some data on the buffer, before the data is sent out. So, I added PROT_WRITE to the userspace mmap() and <bang>. :) Data went written on the virtual address, by the userspace, and indeed the userspace app can read it correcly, but... the physical memory seems untouched. Briefly, my driver allocate a page of memory this way: buffer = __get_free_pages(GFP_KERNEL | GFP_DMA, get_order(4K)); and fill it with a pattern: for(i=0; i<maxframebuffer; i++) buffer[i] = (char)(maxframebuffer - i); That works fine. (I can inspect the physical memory through a JTAG debugger). Then I define the device_mmap(), nothing more that the usual mmap stub: int mmaptest_mmap(struct file *filp, struct vm_area_struct *vma) { unsigned int buffersize = 0; buffersize = (4K >> PAGE_SHIFT)+1; if ( (vma->vm_end - vma->vm_start)>>PAGE_SHIFT > buffersize ) return -ENOMEM; vma->vm_flags |= VM_IO | VM_RESERVED; vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot); if (remap_pfn_range(vma, vma->vm_start, virt_to_phys(buffer) >> PAGE_SHIFT, vma->vm_end - vma->vm_start, vma->vm_page_prot)){ return -EAGAIN; } vma->vm_ops = &simple_remap_vm_ops; simple_vma_open(vma); return 0; } Finally, the userspace program do simply: pointer = mmap(blablabla); //with PROT_WRITE and MAP_PRIVATE contained *pointer = 0xA5; // just a random constant. After this, the userspace pointer contains actually 0xA5, while both the virtual and physical address of kernel side still contain the values initialized before. ( a snapshot of both memories, user side (left) and kernel (right) could be found here: http://imageshack.us/f/18/screenshot1fo.png/ ) Some ideas of what happen here? Seems there is some kind of COW on pages used with mmap()... is there a way to disable it? Thanks in advance. bye! -- - Andrea Gasparini - ----------------------------------------------- -------- https://launchpad.net/~gaspa --------- ----- HomePage: http://gaspa.yattaweb.it ------

1 0

perf hardware events
by Andreas Leppert Sept. 14, 2011

Sept. 14, 2011

Hello, I am using the 'perf' tool for evaluation purposes. I know that there are some predefined "Hardware Events" for my platform (Intel Core 2 Duo), e.g. cycles and instructions. My question is: which internal event number is used to calculate the stats for cycles / instructions etc.? Could you give me a hint where in the source code these alias names are translated into the real hardware event numbers? I have already tried to find it, but I didn't succeed. Is there a docu/homepage where the several events are described in more detail? E.g. what does "instructions" stand for? The number of CPU instructions executed while the given program is running? Am I correct that on a platform with Level1 and Level2 cache, LLC (last level cache) = L2 ? Thanks for your efforts, Andreas Leppert

1 0

error : no symbol version for "function"
by kashish bhatia Sept. 14, 2011

Sept. 14, 2011

Hi all, I am trying to port a kernel module from 2.6.23 kernel to 2.6.32 kernel. While inserting the module , it shows an error : insmod : error inserting 'module.ko' : -1 Unknown symbol in module I checked in /var/log/messages for the unknown symbol and it shows, module : no symbol version for function_name Can anybody explain what is this error and how to resolve this error ? -- Thanks & Regards, Kashish Bhatia

4 6

Watchdog timer expired, but panic logs not seen
by sandeep kumar Sept. 14, 2011

Sept. 14, 2011

Hi all, I wanted to manipulate the watchdog timer driver to see how it works in deadlocks. In our target qualcom ARM board we use msm_watchdog hardware timer. It has one register WDT0_BARK_TIME, if you set some time, when the timeout occurs it will send and interrupt. In the same driver the interrupt handler is also present, which dumps all the application stacks and calls panic("watchdog bark recieved"); But when is see the logs, immediately after the interrupt generated the device is turning off. It is not even entering the bark_interrupt_handler. (i did put logs in the starting of intr handler and i put another log just before the interrupt occured) My question is this, When watch dog timer expires(hardware watchdog), its interrupt directly resets the system (or) it is treated as an interrupt and a handler is executed. THanks, Sandeep.

3 3

Paging of Kernel Memory
by ashish anand Sept. 13, 2011

Sept. 13, 2011

Hi on wed 7th sep Christopher Harvey wrote >>It means that it can't be swapped to your swap partition, even if >>you're not using it. this thing I understood it pretty well but what about the line "Therefore, every byte of memory you consume is one less byte of available physical memory".What is the meaning of this line and why it is so. On Tue, Sep 6, 2011 at 9:30 PM, <kernelnewbies-request(a)kernelnewbies.org>wrote: > Send Kernelnewbies mailing list submissions to > kernelnewbies(a)kernelnewbies.org > > To subscribe or unsubscribe via the World Wide Web, visit > http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies > or, via email, send a message with subject or body 'help' to > kernelnewbies-request(a)kernelnewbies.org > > You can reach the person managing the list at > kernelnewbies-owner(a)kernelnewbies.org > > When replying, please edit your Subject line so it is more specific > than "Re: Contents of Kernelnewbies digest..." > > > Today's Topics: > > 1. Re: Re: Regarding threaded irq and normal irq (anish singh) > 2. CRC calculation in Module.symvers (Shakthi Kannan) > 3. Re: Finding physical and logical core number in a kernel > module ? (Alexandre Courbot) > 4. How to identify Connected USB Storage device (Harddisk, flash > Memory or SSD) (mani) > 5. Paging of Kernel Memory (Vijay Chauhan) > 6. Re: Paging of Kernel Memory (Christopher Harvey) > > > ---------------------------------------------------------------------- > > Message: 1 > Date: Tue, 6 Sep 2011 13:57:22 +0530 > From: anish singh <anish198519851985(a)gmail.com> > Subject: Re: Re: Regarding threaded irq and normal irq > To: Mulyadi Santosa <mulyadi.santosa(a)gmail.com> > Cc: Greg KH <greg(a)kroah.com>, tglx(a)linutronix.de, sandeep kumar > <coolsandyforyou(a)gmail.com>, kernelnewbies(a)kernelnewbies.org > Message-ID: > <CAK7N6voUN%2Bb8nj%3DnQNoU1wGVr%3Db9jCm%2B0Q43eHg7bPLcgB4Nzw(a)mail.gmail.com > > > Content-Type: text/plain; charset=UTF-8 > > On Tue, Sep 6, 2011 at 10:59 AM, anish singh > <anish198519851985(a)gmail.com> wrote: > > On Tue, Sep 6, 2011 at 10:51 AM, Mulyadi Santosa > > <mulyadi.santosa(a)gmail.com> wrote: > >> Hi :) > >> > >> On 06/09/2011, sandeep kumar <coolsandyforyou(a)gmail.com> wrote: > >>> Hmm..But when i tried to take a mutex lock in threaded_irq, it is > throwing a > >>> warning message > >>> "BUG: sleeping function called from invalid context"... So i was > wondering > >>> which way it is.. > >> > >> Please don't top post :) > >> > >> Anyway, regarding that "BUG:" message, could it be that mutex function > >> mistakenly assume it's called inside top half, thus bails out since > > How can mutex assume? IMHO the message would have printed > > because sleeping would be forbidden there but again AFAIK in the threaded > > handler we can do I2C transactions which can sleep.So kind of > > paradox there. > > >>something like using in_irq() inside the threaded irq handler? > I think after going thorough the code of mutex_lock we can find out the > reason why it is printing the warning eventhough we are calling it in a > threaded handler i.e. kernel thread. > > >From the code: > 126 /** > 127 * might_sleep - annotation for functions that can sleep > 128 * > 129 * this macro will print a stack trace if it is executed in an atomic > 130 * context (spinlock, irq-handler, ...). > > 129 line no is important. > This might_sleep is called by mutex_lock function and in the defintion > of __might_sleep i found below code.I think this below if condition is > false because irqs_disabled is true.This causes the warning. > > if ((preempt_count_equals(preempt_offset) && !irqs_disabled()) || > system_state != SYSTEM_RUNNING || oops_in_progress) > return; > > Hope i am right? > > > ps:I have seen the code where sandeep is doing the taking the mutex lock. > > He is taking mutex lock at the beginging of the handler and releasing > > at the end. > >> sleeping is forbidden there? which mutex function do you use? (sorry I > >> don't follow every locking functions exist now) > >> > >> -- > >> regards, > >> > >> Mulyadi Santosa > >> Freelance Linux trainer and consultant > >> > >> blog: the-hydra.blogspot.com > >> training: mulyaditraining.blogspot.com > >> > >> _______________________________________________ > >> Kernelnewbies mailing list > >> Kernelnewbies(a)kernelnewbies.org > >> http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies > >> > > > > > > ------------------------------ > > Message: 2 > Date: Tue, 6 Sep 2011 12:41:24 +0530 > From: Shakthi Kannan <shakthimaan(a)gmail.com> > Subject: CRC calculation in Module.symvers > To: kernelnewbies(a)kernelnewbies.org > Message-ID: > <CABG-yt2W02ToqJJ%3D-n773Lg-JCRnDYc7qzA0HsCw%2BmorgnSx8Q(a)mail.gmail.com > > > Content-Type: text/plain; charset=UTF-8 > > Greetings! > > I would like to know as to: > > 1. How the CRC is computed for an exported symbol in the Linux kernel, and > 2. Given a kernel .c file, how can I obtain the CRC checksums for a symbol. > > For linux-3.0.2 (example), I compiled with CONFIG_MODVERSIONS=y, and > running make built vmlinux, and generated the Module.symvers file. I > believe genksyms is used to generate the CRC information. > > I had referred the "Kernel Symbols and CONFIG_MODVERSIONS" > documentation [1] written by Mark McLoughlin. I tried the following in > the top-level kernel sources: > > $ gcc -E -D__GENKSYMS__ -DCONFIG_MODVERSIONS -DEXPORT_SYMTAB > crypto/api.c -I/tmp/linux-3.0.2/include > -I/tmp/linux-3.0.2/arch/x86/include | ./scripts/genksyms/genksyms > > and I get the following error output: > > === Output === > > /tmp/linux-3.0.2/include/linux/jiffies.h:43:42: error: division by zero in > #if > In file included from /tmp/linux-3.0.2/include/linux/tracepoint.h:19:0, > from /tmp/linux-3.0.2/include/linux/module.h:18, > from crypto/api.c:22: > /tmp/linux-3.0.2/include/linux/rcupdate.h:160:2: error: #error > "Unknown RCU implementation specified to kernel configuration" > In file included from /tmp/linux-3.0.2/include/linux/module.h:21:0, > from crypto/api.c:22: > /tmp/linux-3.0.2/arch/x86/include/asm/module.h:59:2: error: #error > unknown processor family > > === End === > > The documentation for genksyms is old, and now there is no -k option? > > Appreciate any inputs in this regard, > > Thanks! > > SK > > [1] Kernel Symbols and CONFIG_MODVERSIONS. > http://www.skynet.ie/~mark/pub/symbols.txt > > -- > Shakthi Kannan > http://www.shakthimaan.com > > > > ------------------------------ > > Message: 3 > Date: Tue, 6 Sep 2011 16:14:19 +0900 > From: Alexandre Courbot <gnurou(a)gmail.com> > Subject: Re: Finding physical and logical core number in a kernel > module ? > To: Vaibhav Jain <vjoss197(a)gmail.com> > Cc: kernelnewbies(a)kernelnewbies.org > Message-ID: > <CAAVeFu%2Bh9%3D-roQcJaoh_d4XNbm8BtqtbQRR7qrWgx0B-k27dHA(a)mail.gmail.com > > > Content-Type: text/plain; charset=UTF-8 > > > I agree that the kernel module is not restricted to any core but is it > > possible that whenever it is executed > > it gives the core number whichever is executing it ? > > Have a look at the end of include/linux/smp.h: smp_processor_id() > gives you the number of the current executing CPU. get_cpu() will do > the same and will also disable preemption so that you will stay on > that CPU until put_cpu() is called. > > >From user-space, you can use sched_getcpu() to obtain the same > information. > > Hope this helps, > Alex. > > > > ------------------------------ > > Message: 4 > Date: Tue, 6 Sep 2011 16:02:33 +0530 > From: mani <manishrma(a)gmail.com> > Subject: How to identify Connected USB Storage device (Harddisk, flash > Memory or SSD) > To: kernelnewbies(a)kernelnewbies.org > Message-ID: > <CAB%2BTZU93_edNTfbNVKvQdgWiJxDTNHWWunNT6VNbHP%2B%3DRTixfw(a)mail.gmail.com > > > Content-Type: text/plain; charset="iso-8859-1" > > Hi all, > > Is there any way to find out the type of the connected USB disk (Hard > disk/SDD)? > > AFAIK there is a significant performance difference in Harddisk/SDD > devices. > > Previously i though that the rotational flag in the sysfs will distinguish > them but it is same for both the types > > #cat /sys/block/sda/queue/rotational > 1 for both flash & HDD > > Is there any way to distinguish the USB Harddisk or SDD drives ? > > Thanks & Regards, > Manish >

3 2

kthread Working
by Gaurav Mahajan Sept. 13, 2011

Sept. 13, 2011

Hi, I have a douby regarding kthread working. How can we check that a kthread is working or exited. It may exit beacuse of call to do_exit in threadfn or callling kthread_stop function on task_struct (correct me if i am weong). I want to check that if kthread has finished its work or not. If not then what should I call so that kthread gets killed. If it is finshhed then which function I should call so that as to free kthread resources. Regards, Gaurav Mahajan.

2 1

maintenance time and git error
by esmaeil mirzaee Sept. 13, 2011

Sept. 13, 2011

Hi apologize in weak English and unrelated send. does anyone know when the maintaining of the web page www.kernel.org will be finish? and when I try this "git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux2.6.git linux I got below error "Initialized empty Git repository in /home/esmaeil/fixbug/krl/linux/.git/" "fatal: Unable to look up git.kernel.org (port 9418) (Name or service not known)" -- Best Esmaeil https://sites.google.com/site/esmaeilmirzaee

2 3

Invitation to connect on LinkedIn
by Venugopal Rao via LinkedIn Sept. 12, 2011

Sept. 12, 2011

LinkedIn ------------ Venugopal Rao requested to add you as a connection on LinkedIn: ------------------------------------------ Deepak, I'd like to add you to my professional network on LinkedIn. - Venugopal Accept invitation from Venugopal Rao http://www.linkedin.com/e/8z2206-gshpcgof-4y/L90NvsYL3GREvJdnM0f-HnfkWNbSD_… View invitation from Venugopal Rao http://www.linkedin.com/e/8z2206-gshpcgof-4y/L90NvsYL3GREvJdnM0f-HnfkWNbSD_… ------------------------------------------ DID YOU KNOW you can use your LinkedIn profile as your website? Select a vanity URL and then promote this address on your business cards, email signatures, website, etc http://www.linkedin.com/e/8z2206-gshpcgof-4y/ewp/inv-21/?hs=false&tok=0b4Vb… -- (c) 2011, LinkedIn Corporation

1 0