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

Re: Transfer Buffer in URB
by Luis de Bethencourt June 25, 2015

June 25, 2015

On Thu, Jun 25 2015 at 10:43:32AM -0400, Greg KH wrote: >> On Thu, Jun 25, 2015 at 06:16:28PM +0530, roni wrote: >> What is the transfer buffer in URB? >> >> What is the job of transfer buffer? > > The in-kernel documentation should answer these questions, which is why > we wrote it :) Just in case and for future people finding these through search engines: Documentation/usb/URB.txt https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/Docume… Luis

2 1

Poking eudyptula for status updates
by Luis de Bethencourt June 25, 2015

June 25, 2015

Hello all, I've been waiting for a week now since I submitted task 5 to Eudyptula. I understand reviewing submissions takes time, specially the tasks that are reviewed manually and there is a queue. Plus, it helps make the challege replicate the experience of contributing to an open project. Waiting is fine. What I am wondering is if my task fell through the cracks. Not sure if there is a method to poke eudyptula for a status update. To confirm the submissions is in the queue. Is there an equivalent of politely asking a project maintainer about review when a decent amount of time has passed since submission? I know that if you resubmit you get pushed to the tail of the queue. Sorry if this has been asked or explained before, I have searched and couldn't find anything about the matter. Sorry for yet another Eudyptula related email in the list. Thanks, Luis

2 3

Transfer Buffer in URB
by roni June 25, 2015

June 25, 2015

What is the transfer buffer in URB? What is the job of transfer buffer?

2 1

About guiding hello world module submission
by Mayur Patil June 25, 2015

June 25, 2015

Hi All, I am conducting one workshop at FUDCon in which I am trying to teach how to write and send your first linux kernel device driver. Could please suggest me the place where I can guide the students to send the device driver? Thanks !! -- Regards, Mayur S Patil, Looking for R&D or Soft Engg positions, Pune, India.

5 8

How to avoid TCP 3way handshare
by Ramana Reddy June 24, 2015

June 24, 2015

Hi all, Assume that the client has already ESTABLISHED state with one server, and I want to use the same state of the client with another server with out sending syn and syn+ack packets with the second server. I want to make the second server by default ESTABLISHED state. What are the modifications needed in the linux kernel code and the respective files. I am not worried about security and attacks at this moment. Regards, Ramana

2 1

Re: Kernel gets deadlocked during smp booting
by AYAN KUMAR HALDER June 21, 2015

June 21, 2015

> Could you post the full backtrace (straight from the kernel log)? The backtrace has been obtained from a debugger, not from the kernel log (as the kernel has been deadlocked). The backtrace is as follows:- arch_spin_lock _raw_spin_lock vprintk_emit vprintk_default printk smp_init do_initcall_level do_initcalls do_basic_setup kernel_init_freeable kernel_init ret_from_fork

2 3

Interrupts handling
by Alexander Kuleshov June 21, 2015

June 21, 2015

Hello, I have a question about interrupt handling in Linux kernel for x86_64. As we can read in the 64-ia-32-architectures-software-developer-vol-3a-part-1-manual, when an interurpt/exception occurs, the handler's stack contains SS, RSP, RFLAGS, CS, RIP and error code. My question is: does processor push this register on the stack when an interrupt occurs or kernel must push it? In the arch/x86/kernel/entry_64.S we can see following code in the idtentry macro: .ifeq \has_error_code pushq_cfi $-1 .endif where pushq_cfi is: .macro pushq_cfi reg pushq \reg CFI_ADJUST_CFA_OFFSET 8 .endm First of all what does it mean $-1 here? I have write simple assembly program and I see: push 0xffffffffffffffff in the gdb. The second question, if processor pushes error code and other during interrupt why we push it here? Thank you.

2 2

Correctness of compaction_suitable condition
by Alexander June 20, 2015

June 20, 2015

Hello. Actually this question is about kernel 3.10, but as i can see the situation didn't change in future versions. I would like to mention 2 places in the code: 1.) First one is an internals of buddy allocator subsystem function __rmqueue As i can see, kernel tries to allocate smallest suitable free block for given order and migratetype here __rmqueue_smallest. If it fails, then it could try other migration types here __rmqueue_fallback. The suitable migration types are in fallback list. For example, if we have MIGRATE_UNMOVABLE allocation, we can try: [MIGRATE_UNMOVABLE] = { MIGRATE_RECLAIMABLE, MIGRATE_MOVABLE, MIGRATE_RESERVE} but as i can see we can't try MIGRATE_CMA free blocks. 2.) Now let's consider compaction_suitable function. This one is used in compact_zone, to check if the compaction is suitable. This function uses fill_contig_page_info, where number of suitable blocks for current order allocation is calculated like: 570 /* Count number of free blocks */ 571 blocks = zone->free_area[order].nr_free; 572 info->free_blocks_total += blocks; 573 574 /* Count free base pages */ 575 info->free_pages += blocks << order; 576 577 /* Count the suitable free blocks */ 578 if (order >= suitable_order) 579 info->free_blocks_suitable += blocks << 580 (order - suitable_order); So as i can see, zone->free_area[order].nr_free includes all MIGRATE_TYPES for given order. But if we allocate unmovable blocks, in fact we can't use free blocks with MIGRATE_CMA. As i think this could lead to the fact, that on slowpath __alloc_pages_direct_compact wouldn't invoke compaction, because it will think that we have suitable blocks for current order allocation, but in fact we haven't. Please, could you explain is this a problem, or i just don't understand some logic behind this? Thank you.

1 0

Getting absolute path of a dentry
by Nikolay Borisov June 20, 2015

June 20, 2015

Hello, I'm developing some custom functionality in the kernel in the flavour of inotify, basically creating a linked list (in the kernel) with all changed dentries. Then I want to consume that list (just print the file names from userspace). I've done all of this and I'm using dentry_path_raw do get the path, this works well for mounts which originate from the root mount point. However, in the case I have a dentry which is not under the root mount dentry_path_raw returns the path under the mount point but not the mount point itself. Concretely, If I have my /home as a separate mount and something is changed under /home/dir1/dir2/changed_file and I do the dentry_path_raw(changed_file), the path returned is /dir1/dir2/changed_file. So mu question is how do I also get the path of the mount point included as well? I did try including more context, than the dentry like the get_fs_root of the process which triggered the change and then feeding the dentry + the vfsmount from the changing process' root to d_absolute_path to no avail. Any ideas or pointers would be greatly appreciated.

1 0

spi_async not working, spi_sync working fine
by Schrey, Moritz June 19, 2015

June 19, 2015

Dear All, I'm having trouble understanding the difference between spi_sync and spi_async: Theoretically, spi_sync waits for completion while spi_async provides completion callbacks. However, on my platform spi_async does not do anything! I have the following code that is supposed to write a single byte. Using spi_sync works fine, calling the completion callback with spi_async also works. I just do not see anything on the connected oscilloscope screen. What am I missing? Thanks in advance Moritz int lprf_set_WREN(void) { int status, ret; int sync = 0; struct spi_message m; struct spi_transfer t; u8 *buffer = kmalloc(sizeof(u8), GFP_KERNEL); memset(buffer, 0, sizeof(u8)); buffer[0] = 0x06; t.tx_buf = buffer; t.rx_buf = NULL; printk(KERN_DEBUG "lprf: buffer=%p\n", buffer); t.len = 1; t.bits_per_word = 0; spi_message_init(&m); spi_message_add_tail(&t, &m); printk(KERN_DEBUG "lprf: lprf_set_WREN start\n"); if (sync) { /* Variante A: sync */ status = spi_sync(lp->spi, &m); printk(KERN_DEBUG "lprf: lprf_set_WREN sync end. status=%d \n", status); } else { /* Variante B: async */ init_completion(&lp->state_complete); m.complete = lprf_mycomplete; m.context = &lp->state_complete; ret = spi_myvalidate(lp->spi, &m); if (ret) { printk(KERN_DEBUG "spi_myvalidate returned %d\n", ret); return ret; } status = spi_async_locked(lp->spi, &m); if (status == 0) { wait_for_completion(&lp->state_complete); status = m.status; } printk(KERN_DEBUG "lprf: lprf_set_WREN async end status=%d m.status=%d\n", status, m.status); m.context = NULL; } if (status) printk(KERN_ERR "lprf: spi_async failed at %s, %i\n", __FILE__, __LINE__); return status; }

1 0