- kernelnewbies@kernelnewbies.org - Kernelnewbies Mailing Lists

where is the module reference count incremented or decremented ?
by Jeong-Hwan Kim March 3, 2014

March 3, 2014

Hi, everyone When the module is communicating with user through /proc file system, where is the module reference count incremented or decremented in kernel source? Thanks in advance. Best Regards, J.Hwan Kim

2 2

Reset due to emergency remount r/o
by Priyaranjan Das March 3, 2014

March 3, 2014

Hi All, I am facing an kernel restart due to emergency remount. Could anybody tell me general techniques to go head with the debugging? Below are the logs which are seen during a reboot. [ 425.832395,1] SysRq : Emergency Remount R/O [ 425.876210,1] EXT4-fs (mmcblk0p36): re-mounted. Opts: (null) [ 425.925899,1] EXT4-fs (mmcblk0p33): re-mounted. Opts: (null) [ 425.928731,1] EXT4-fs (mmcblk0p29): re-mounted. Opts: (null) [ 425.955762,1] Emergency Remount complete [ 426.187136,0] Restarting system with command 'bootloader'. [ 426.187327,0] Current task:init(1) Parent task:swapper/0(0) [ 426.187430,0] Going down for restart now Regards, Priyaranjan

7 6

LF GSoC page
by Sima Baymani March 2, 2014

March 2, 2014

Hi Till, I'm guessing you're the author of this http://www.linuxfoundation.org/collaborate/workgroups/gsoc/2014-gsoc-kernel… which points to #kernelnewbies. Right now there are a lot of people coming in and asking for info or help etc. It would be helpful if you could add some more detail there. For example Greg K-H has stated he's the one to ping (or email, according to himself) - that could be something to add. Also, as far as I understand, applicants have to have an idea/suggestion of their own for projects (according to chats by Greg). Could you add information about that as well? Right now everybody else in that channel are more or less clueless, which imo isn't a good first impression for applicants =) BR, Sima

2 3

understanding register layout
by sham pavman March 2, 2014

March 2, 2014

Hi all, While going through the BCM 57712 code i found the following. /* [RW 32] reset reg#2; rite/read one = the specific block is out of reset; * write/read zero = the specific block is in reset; addr 0-wr- the write * value will be written to the register; addr 1-set - one will be written * to all the bits that have the value of one in the data written (bits that * have the value of zero will not be change) ; addr 2-clear - zero will be * written to all the bits that have the value of one in the data written * (bits that have the value of zero will not be change); addr 3-ignore; * read ignore from all addr except addr 00; inside order of the bits is: * [0] rst_bmac0; [1] rst_bmac1; [2] rst_emac0; [3] rst_emac1; [4] * rst_grc(global register); [5] rst_mcp_n_reset_reg_hard_core (global * register); [6] rst_mcp_n_hard_core_rst_b(global register); [7] * rst_mcp_n_reset_cmn_cpu(global register); [8] * rst_mcp_n_reset_cmn_core(global register); [9] rst_rbcn; [10] rst_dbg; * [11] rst_misc_core(global register); [12] rst_dbue (UART)(global * register); [13] Pci_resetmdio_n(global register); [14] * rst_emac0_hard_core; [15] rst_emac1_hard_core; 16] rst_pxp_rq_rd_wr; 17] * rst_atc; 18] rst_cnig; 19] rst_pglc (global register); 31:20] reserved */ Here is what i understand and here is what i need help with! What I understand, REG2 has 20 bits and what each of those bits signify, I understand. I also understand that if I read/write 1, it means that particular block has finished reset. If I read/write 0, it means that particular block has completed reset. But what is this addr 1-set, addr 2- clear and friends?

1 0

Completely Fair Scheduler - Min_vruntime
by Anup Buchke March 1, 2014

March 1, 2014

Hi all, I am trying to learn CFS. I came across this variable. min_vruntime : kernel documentation says it stores the total work done by the system . AND it is used to add new tasks more to the left in the rbtree so that they are scheduled quickly. But I am not able to understand why we need that variable and how it is used for 1. new added tasks 2. Tasks that wake up after sleep. I tried learning from professional linux kernel architecture+utlk and alike. But I am not able to understand. Also the key in the rbtree is used as (current_vruntime - min_vruntime) ? why are we even using such formule in which we subtract min_vruntime which is monotonically increasing?? I think it is related to some latency period and the processes which runs in one latency period. I think I may be missing some of the core concepts on scheduling process, but not able to figure out what I am missing. My understanding is pretty messed up the use of min_vruntime. I would be highly grateful if someone could explain me in bit of a detail about my answer. Thanks in Advance!!! Anup

1 0

Mapping a memory address to all the processes using it
by Sandeep K Chaudhary Feb. 28, 2014

Feb. 28, 2014

Hi guys, I am working on a hack which requires me to find all the processes using a particular memory address. Lets say I get an MCE error containing a memory address. I want to use this address to find all the processes that might access it, and kill them before they do. Please let me know if this is possible using some already implemented routine. I can think of a brute force method of looping though all the processes and looking at the VMAs they are using to see if there is a match for the address I have. But I want to do better than this brute force method. Please let me know if this is possible. Also, while looking for the same in Kernel tree, I found that it is easy to do the reverse mapping i.e. finding the memory addresses a process is using given the process, but this doesn't help with what I intend to do. Thanks and regards, Sandeep.

2 4

Re: [Printing-architecture] Google Summer of Code 2014 - IPP-over-USB printer support - Joint project idea for OpenPrinting and the kernel
by Michael Sweet Feb. 27, 2014

Feb. 27, 2014

Carlos, On Feb 26, 2014, at 9:33 PM, Carlos Rimola <crimola(a)gmail.com> wrote: > Hi All, > > I have some quick and general questions regarding IPP over USB ("IPP USB" for short). Some are related to Till's proposed project "Google Summer of Code 2014 - IPP-over-USB printer support". Any help and feedback will be greatly welcomed. I should mention that I am in favor of the proposed project for this event. > > Assumptions (please confirm or correct): > > 1) This first one may be obvious but to be sure - I am assuming that we are referring to IPP USB as defined in the "USB Print Interface Class IPP Protocol Specification Revision 1.0" dated 12/5/2012 and published by USB-IF and authored by HP's Smith Kennedy and Andrew R. Mitchell. I'll refer to it as the "IPP USB Spec". > Correct. > 2) As I understand the IPP USB Spec, there is NO network interface, NOR TCP/IP involved. The Communication Protocol to be used between Host and Device (Printer) is purely HTTP + IPP directly over USB. A place where this is noted is section 6.2 - "HTTP Headers" which states the following: > > Since there is NO network interface connection, NO DNS hostnames or IP addresses, and NO TCP port numbers associated with USB connection, the requirements of the HTTP Host field is addressed by requiring that the value of this header MUST be "localhost". > > Please correct me if I am wrong on either of these assumptions. > Correct (a port number can be passed by the "client" over USB to allow gateways/proxies to work...) > Questions: > > 1) I know the Spec is already cast in stone but I would like to understand what function HTTP serves and if it is only used for identifying "Host: localhost" and the "/ipp/printer" path? In other words, could pure IPP Requests/Responses and IPP expected format "Print Data" have sufficed? > While you might get away with that for simple IPP messages, that wouldn't work for document data since IPP by itself has no framing or other niceties - you'd never know when the data ended. Also, the HTTP portion is used for the embedded web server, doing firmware updates, and so forth. > 2) Section 3.2 "Interface Set" paragraph 2 states that "All IPP USB capable Print class interfaces provided by a device MUST be functionally equal from an IPP operation or HTTP perspective. In other words, any IPP operation or resource path that is valid over one IPP USB interface MUST be reachable via any and all of the IPP USB interfaces." > > Does this imply that, for example, a Request from the host can be sent over one interface (I/F #1) and the response received over a different interface (e.g., I/F #2)? > No, each interface is an independent channel to the printer. The reason for this requirement is to prevent having an IPP USB endpoint just for printing, and another just for scanning, and so forth. Effectively IPP USB defines an interface protocol that allows arbitrary HTTP and IPP requests to be performed. > 3) The last question is much simpler but would be helpful to an implementor - what specific printers (manufacturer, line and model) on the market support IPP USB? I have seen references to HP Photosmart and OfficeJet but no model given. Similarly, what Host(s) on the market, including OS and version, support the IPP USB protocol with these Printers? > Here is the latest list we have from HP: Deskjet 3520 Envy 120 Envy 4500 Envy 5530 Officejet 3620 Officejet 4630 Officejet 7610 Officejet Pro 276 MFP Officejet Pro X576 dw Photosmart 5520 Photosmart 6520 Photosmart 7520 I know of four other manufacturers that either have shipping products or will soon be shipping - will post here when I can do so publicly... _________________________________________________________ Michael Sweet, Senior Printing System Engineer, PWG Chair

2 1

Fwd: zero copy with PACKET_MMAP
by ratheesh kannoth Feb. 27, 2014

Feb. 27, 2014

Hi , The link https://www.kernel.org/doc/Documentation/networking/packet_mmap.txt talks about achieving zero copy. But i find a deep packet copy ( of length snaplen= skb->len ) here http://lxr.free-electrons.com/source/net/packet/af_packet.c#L1841. Is there anything missing/wrong in my understanding ? Thanks, Ratheesh

2 1

Re: Blocked I/O in read() and mmap()
by Rajat Sharma Feb. 27, 2014

Feb. 27, 2014

Why do you need to block in mmap()? mmap is supposed to create a mapping area in virtual address space for the process. Actual transfer happens later through page fault handlers on demand basis. look at vm_operations fault/readpage etc methods, these might be the places you want to wait for the data. On Wed, Feb 26, 2014 at 4:14 PM, Le Tan <tamlokveer(a)gmail.com> wrote: > So what should I do if I want the mmap() not to return right now? Is > it strange to block in mmap() and few people will do this? Thanks for > your help! > > 2014-02-27 4:45 GMT+08:00 Rajat Sharma <fs.rajat(a)gmail.com>: > > It seems this task "landscape-sysin" is trying to peek into virtual > memory > > of your processes and the process within mmap call is holding its > > mm->mmap_sem semaphore which grants access to its address space. > > landscape-sysin is trying to grab this semaphore to poke into address > space > > of your mmap process address space. As from your description, it might be > > invoked everytime you are opening a new shell. Not sure why this process > > bother's about other process address space. Little googling shows this as > > relevant to your case: > > > > > http://www.techques.com/question/2-66765/Disable-usage-of-console-kit-daemo… > > > > Your read process is innocent and not involved in this deadlock. > > > > -Rajat > > > > > > On Wed, Feb 26, 2014 at 4:13 AM, Le Tan <tamlokveer(a)gmail.com> wrote: > >> > >> Hi, I am writing a driver module. Now I have some questions about > blocked > >> I/O. > >> my_read() is the read function in the file_operations struct in my > >> module. my_read() is just as simple as this: > >> ssize_t my_read(....) > >> { > >> if(wait_event_interruptible(dev->queue, a == b)) > >> return -ERESTARTSYS; > >> return count; > >> } > >> Then I write a simple program to open and read the device. Obviously > >> the program will be blocked. Now I still can open a new shell window > >> and log in ( I use xshell). > >> > >> However, then I implement my_mmap(), the mmap function in the > >> file_operations struct in my module, like this: > >> int my_mmap(....) > >> { > >> if(wait_event_interruptible(dev->queue, a == b)) > >> return -ERESTARTSYS; > >> return 0; > >> } > >> Then I write a simple program to open and mmap() the device. Obviously > >> the program will be blocked again. However, when I open a new shell > >> window in xshell and try to connect to the linux, it displays like > >> this: > >> > >> Connecting to 192.168.146.118:22... > >> Connection established. > >> To escape to local shell, press 'Ctrl+Alt+]'. > >> > >> And I can't log in! Then after a while, in the syslog, there is one > >> message like this: > >> [38306.614103] INFO: task landscape-sysin:17616 blocked for more than > >> 120 seconds. > >> [38306.614114] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" > >> disables this message. > >> [38306.614120] landscape-sysin D ffffffff8180fb60 0 17616 17609 > >> 0x00000000 > >> [38306.614125] ffff88031d609c90 0000000000000082 ffff88032fffdb08 > >> 0000000000000000 > >> [38306.614130] ffff8803130bdc40 ffff88031d609fd8 ffff88031d609fd8 > >> ffff88031d609fd8 > >> [38306.614133] ffff88062150c530 ffff8803130bdc40 0000004100000000 > >> ffff8803130bdc40 > >> [38306.614137] Call Trace: > >> [38306.614147] [<ffffffff816b2c49>] schedule+0x29/0x70 > >> [38306.614151] [<ffffffff816b3acd>] rwsem_down_read_failed+0x9d/0xf0 > >> [38306.614157] [<ffffffff81341824>] > call_rwsem_down_read_failed+0x14/0x30 > >> [38306.614160] [<ffffffff816b1644>] ? down_read+0x24/0x2b > >> [38306.614166] [<ffffffff81153661>] __access_remote_vm+0x41/0x1f0 > >> [38306.614170] [<ffffffff81153ddb>] access_process_vm+0x5b/0x80 > >> [38306.614175] [<ffffffff811ea423>] proc_pid_cmdline+0x93/0x120 > >> [38306.614178] [<ffffffff811eb425>] proc_info_read+0xa5/0xf0 > >> [38306.614182] [<ffffffff81186e84>] vfs_read+0xb4/0x180 > >> [38306.614185] [<ffffffff81187102>] SyS_read+0x52/0xa0 > >> [38306.614189] [<ffffffff816bc8c2>] system_call_fastpath+0x16/0x1b > >> > >> If I terminate the program by force, then I can log in right now. > >> So, are there any differences between the read and the mmap function > >> to the wait_event_interruptible()? Why? If I want to block mmap() just > >> like blocking read(), what should I do? Or it is impossible? > >> Thanks! > >> > >> _______________________________________________ > >> Kernelnewbies mailing list > >> Kernelnewbies(a)kernelnewbies.org > >> http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies > > > > >

2 2

Google Summer of Code 2014 - IPP-over-USB printer support - Joint project idea for OpenPrinting and the kernel
by Till Kamppeter Feb. 27, 2014

Feb. 27, 2014

Hi, a new standard not yet supported under Linux but starting to penetrate the market is IPP-over-USB (Internet Printing Protocol over USB). IPP, the Internet Printing Protocol from the Printing Working Group (PWG, http://www.pwg.org/) is a standard protocol for network printers (and also used by CUPS, the standard printing environment on practically all non-Windows operating systems). IPP network printers have a lot of advantages compared to USB printers (letting the ability of several computers on a network being able to access them aside): - Encrypted job transfer - Possibility to request printer status and printer capabilities - Web interface to configure the printer All this works with standard protocols and without any requirement of printer manufacturer/model specific software. For printers which also understand standard languages for the jobs themselves (PostScript, PDF, PWG Raster, PCL, JPG, TIFF) this means completely driverless printing (IPP Everywhere). Unfortunately, this is a network protocol for network printers. Fortunately, the PWG has added a standard to make it also go into USB printers, IPP-over-USB. Problem is that there is no Linux support for that. First, I want to make a feature request to the kernel to add it. Second, I want to suggest this as a Google Summer of Code project, asking for mentors on the kernel side. Mentoring Organization will be the Linux Foundation, hosting projects for both OpenPrinting and the kernel. It should not be too complex. Probably one can start on the driver for USB Ethernet or WLAN sticks, as they are also USB devices which introduce a network interface to the system. What one has to do is to create a driver for another, probably similar device, the IPP-over-USB printer. The driver should not be specific to the printer model (it is an open standard protocol) and it also should provide a network interface to the system under which there is only found the printer. The printer should be accessible under this interface via port 80 (web interface), 631 (IPP), and 443 (encrypted). There are already HP printers available which do IPP-over-USB. I will try to make arrangements for developers/students to get samples. WDYT? Is this a viable project? Should we post it on the ideas lists? Who would mentor it? Till

5 9