February 2017 - kernelnewbies@kernelnewbies.org

How to dump user space stack from kernel space
by Abhay Koralli Feb. 22, 2017

Feb. 22, 2017

Hi, We want to get the user space stack trace from the kernel space. Is there any easy way to dump the same ? Warm Regards,-Abhay.

2 1

question about nopage sample code
by Ali Aminian Feb. 22, 2017

Feb. 22, 2017

in ldd3 no page is implemented as the code posted below (page 428). can someone explain why should we subtract vma->vm_start? struct page *simple_vma_nopage(struct vm_area_struct *vma, unsigned long address, int *type) { struct page *pageptr; unsigned long offset = vma->vm_pgoff << PAGE_SHIFT; unsigned long physaddr = address - vma->vm_start + offset; // why should we subtract vma->vm_start ? unsigned long pageframe = physaddr >> PAGE_SHIFT; if (!pfn_valid(pageframe)) return NOPAGE_SIGBUS; pageptr = pfn_to_page(pageframe); get_page(pageptr); if (type) *type = VM_FAULT_MINOR; return pageptr; } thank you Ali Aminian.

1 0

powerpc build fail
by Tobin C. Harding Feb. 21, 2017

Feb. 21, 2017

The current (2bfe01e) torvalds git tree fails to build with the following error arch/powerpc/kernel/time.c: In function ‘running_clock’: arch/powerpc/kernel/time.c:712:25: error: implicit declaration of function ‘cputime_to_nsecs’ [-Werror=implicit-function-declaration] return local_clock() - cputime_to_nsecs(kcpustat_this_cpu->cpustat[CPUTIME_STEAL]); ^ I would like to do more than just file a bug report. Further investigation (cputime.h) shows that cputime_to_nsecs has a preprocessor guard on the config option (currently enabled) CONFIG_VIRT_CPU_ACCOUNTING_NATIVE. CONFIG_PPC_PSERIES is also enabled. The offending code in arch/powerpc/kernel/time.c is #ifdef CONFIG_PPC_PSERIES /* * Running clock - attempts to give a view of time passing for a virtualised * kernels. * Uses the VTB register if available otherwise a next best guess. */ unsigned long long running_clock(void) { /* * Don't read the VTB as a host since KVM does not switch in host * timebase into the VTB when it takes a guest off the CPU, reading the * VTB would result in reading 'last switched out' guest VTB. * * Host kernels are often compiled with CONFIG_PPC_PSERIES checked, it * would be unsafe to rely only on the #ifdef above. */ if (firmware_has_feature(FW_FEATURE_LPAR) && cpu_has_feature(CPU_FTR_ARCH_207S)) return mulhdu(get_vtb() - boot_tb, tb_to_ns_scale) << tb_to_ns_shift; /* * This is a next best approximation without a VTB. * On a host which is running bare metal there should never be any stolen * time and on a host which doesn't do any virtualisation TB *should* equal * VTB so it makes no difference anyway. */ return local_clock() - cputime_to_nsecs(kcpustat_this_cpu->cpustat[CPUTIME_STEAL]); } #endif Can anyone give me a hint on this one? Where to start reading? thanks, Tobin.

1 1

Re: kernel development process question, patch review
by Ozgur Karatas Feb. 21, 2017

Feb. 21, 2017

Hello, I think the list idea is no need for confusion, it is desirable to provide people with a new kernel newbies update to this list with a help and a quick solution to their problems. This is like pre-kernel training. TO and CC already come with the get_maintainer.pl script. TO: Primary responsible CC: development group + kernel mailing list Kernelnewbies = kernelnewbies help Regards, Ozgur > 21.02.2017, 04:44, "Tobin C. Harding" <me(a)tobin.cc>: >> If a reviewer makes a suggestion and one intends on making the change >> as suggested is it required (normal protocol) to reply stating that >> the you understand their suggestion and intend on implementing it or >> is this just noise. Should one simply re send the next version of the >> patch? >> >> Continuing on this theme, should one add a Suggested-by: tag in the >> previous situation. Some where Documentation/process it states that no >> tag except the CC tag should be added without the authority of the >> person being named? >> >> thanks, >> Tobin.

1 0

kernel development process question, patch review
by Tobin C. Harding Feb. 21, 2017

Feb. 21, 2017

If a reviewer makes a suggestion and one intends on making the change as suggested is it required (normal protocol) to reply stating that the you understand their suggestion and intend on implementing it or is this just noise. Should one simply re send the next version of the patch? Continuing on this theme, should one add a Suggested-by: tag in the previous situation. Some where Documentation/process it states that no tag except the CC tag should be added without the authority of the person being named? thanks, Tobin.

4 3

Help needed in kernel boot
by Er Krishna Feb. 21, 2017

Feb. 21, 2017

Hi Guys, This question is not directly related with Linux Kernel, but more or less with U boot and Kernel booting. So apologies. When we do cpu release <Cpu no> the satndalone application (may be kernel or so) starts running on particular processor. How does it happens ? Can anyone pls let me know this logic and point me in U -boot code for this function ? BR/Krish

1 0

kernel development process
by Tobin Harding Feb. 19, 2017

Feb. 19, 2017

I am posting this to kernelnewbies because I do not know the correct way to broach the issue elsewhere. Code audit in drivers/staging on calls to strncpy(). Tree: gregKH/staging commit d5adbfcd5f7bcc6fa58a41c5c5ada0e5c826ce2c File: drivers/staging/rtl8192u/ieee80211/ieee80211_softmac_wx.c Function: int ieee80211_wx_set_essid(struct ieee80211_device *ieee, struct iw_request_info *a, union iwreq_data *wrqu, char *extra) Lines: 415 - 429 spin_lock_irqsave(&ieee->lock, flags); if (wrqu->essid.flags && wrqu->essid.length) { /* first flush current network.ssid */ len = ((wrqu->essid.length-1) < IW_ESSID_MAX_SIZE) ? (wrqu->essid.length-1) : IW_ESSID_MAX_SIZE; strncpy(ieee->current_network.ssid, extra, len+1); ieee->current_network.ssid_len = len+1; ieee->ssid_set = 1; } else { ieee->ssid_set = 0; ieee->current_network.ssid[0] = '\0'; ieee->current_network.ssid_len = 0; } spin_unlock_irqrestore(&ieee->lock, flags); Control can reach this snippet in two states: 1. wrqu->essid.length *can* be larger than IW_ESSID_MAX_SIZE 2. wrqu->essid.length *can not* be larger than IW_ESSID_MAX_SIZE (likely) If wrqu->essid.length *can* be larger than IW_ESSID_MAX_SIZE: - For the case when wrqu->essid.length > IW_ESSID_MAX_SIZE code sets len to IW_ESSID_MAX_SIZE. If then, source string for call to strncpy() (extra) is >= len+1 (IW_ESSID_MAX_SIZE+1) this results in a non null terminated buffer. If wrqu->essid.length *can not* be larger than IW_ESSID_MAX_SIZE: - Setting of 'len' is overly complicated - The code could be simplified by getting rid of the +1's and -1's - The string would be null terminated without additional code (in line with usage throughout the file). There may not be an issue here, the code may be correct. It seems overly complicated to prove that there is not an off by one error or a buffer overflow vulnerability. The code does not seem to conform to the rest of the file. *If* there is a problem adding a null terminating byte would solve it for both possible states. Comments/questions ------------------ 1. How would a more experienced kernel developer approach an issue like this? With an initial patch and the discussion in a cover letter? 2. General kernel question: Using compiler constants to define buffer storage in two differing manners (i.e constant includes space for null byte vs constant does not include space for null byte) especially when done using the *same* constant seems to be asking for trouble. I'm sure this has been discussed before at length, is there a consensus on this? Is there a per subsystem consensus? thanks, Tobin.

1 0

driver todo's
by srishti sharma Feb. 14, 2017

Feb. 14, 2017

when driver has checkpatch cleanups in it's todo , does that also including fixing warnings , should one try to fix them ?

3 7

recovery from fail
by Tobin Harding Feb. 14, 2017

Feb. 14, 2017

I have made and epic fail and am asking for advice on how to proceed to cause the least amount of upset. I submitted a simple patch series to LKML. A reviewer made a suggestion. I submitted v2 - without building it :( kbuild test robot picked up that it doesn't compile. The problem is that I now cannot get the second patch of the series to compile. What is the correct protocol to follow? I don't want to make any more noise than I already have but I also don't want to ignore the reviewer by not implementing the suggested changes. Is it rude to reply to the original review email for further discussion having already botched the patch? Thanks in advance, Tobin.

3 2

Sleeping in workqueues
by Tobias FItschen Feb. 13, 2017

Feb. 13, 2017

Hello people, I have a question regarding workqueues. I know it's possible to sleep in workqueue routines. Does that mean I can use: 1. "set_current_state(TASK_INTERRUPTIBLE);" get the the task with "current" and pass it to a other task and then call "schedule();" in a workqueue routine and at a different place (perhaps another workqueue routine) use wake_up_process(task)" ? 2. initialize completions in a workqueue routine and call complete() in a different workqueue routine? I tried it in a simple example setup and both worked. I just want to make sure that was not just by accident. As I understand it a work item is executed on a worker, which is a unique process, but I also know that there are exceptions, e.g. I can not copy to or from user space in a workqueue routine. Also I don't know if a worker "sticks" with the current work item until it is completed or can execute multiple work items at the "same" time. 3. In case completions and schedule()/wakeup() work, I would also like to know, which has the better performance. Thanks a lot already! Tobias

2 2