# Bug in volk on armv6 **URL:** \ **Category:** GNU Radio\ **Created:** [July 1, 2013, 10:09am UTC](https://www.ruby-forum.com/t/bug-in-volk-on-armv6/245679 "2013-07-01T10:09:13Z")\ **Posts on this page:** 4\ **Page:** 1
**Author:** ![Sadaf\_N](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/s/858c86/32.png) [@Sadaf\_N](https://www.ruby-forum.com/u/Sadaf_N)\ **Post date:** [July 1, 2013, 10:09am UTC](https://www.ruby-forum.com/t/bug-in-volk-on-armv6/245679/1 "2013-07-01T10:09:13Z")
Hi! I would like to report a bug in volk on armv6. There is infinite “while” cycle in volk/tmpl/volk\_cpu.tmpl.c function “has\_neon” (line 111). Code “while(!found\_neon && auxvec\_f) {” will loop infinitely if there is no neon in platform. found\_neon will never be true and file descriptor auxvec\_f will always be true. – Alexey B. [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid) ---
**Author:** ![snaury](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/s/77aa72/32.png) [@snaury](https://www.ruby-forum.com/u/snaury)\ **Post date:** [July 3, 2013, 4:57pm UTC](https://www.ruby-forum.com/t/bug-in-volk-on-armv6/245679/2 "2013-07-03T16:57:43Z")
On Mon, Jul 1, 2013 at 4:08 AM, Alexey B. [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid) wrote: > – > Alexey B. [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid) Ok, I see your point. Looks like we should be testing the return value from fread, instead of auxvec\_f. Can you confirm if this patch works? diff --git a/volk/tmpl/volk\_cpu.tmpl.c b/volk/tmpl/volk\_cpu.tmpl.c index 81fc679..b1a0a4a 100644 — a/volk/tmpl/volk\_cpu.tmpl.c +++ b/volk/tmpl/volk\_cpu.tmpl.c @@ -116,10 +116,11 @@ static int has\_neon(void){ auxvec\_f = fopen(“/proc/self/auxv”, “rb”); if(!auxvec\_f) return 0; - size\_t r = 1; //so auxv is basically 32b of ID and 32b of value //so it goes like this - while(!found\_neon && auxvec\_f) { - ``` fread(auxvec, sizeof(unsigned long), 2, auxvec_f); ``` - while(!found\_neon && r) { - ``` r = fread(auxvec, sizeof(unsigned long), 2, auxvec_f); if((auxvec[0] == AT_HWCAP) && (auxvec[1] & HWCAP_NEON)) found_neon = 1; ``` } Or might be better to test ‘if(r \< 2)’ after the fread line and break in case an error occurs and we only get 1 character out and not try to read from it again. Probably not a bit deal since we’re only reading in two chars at a time. Tom ---
**Author:** ![snaury](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/s/77aa72/32.png) [@snaury](https://www.ruby-forum.com/u/snaury)\ **Post date:** [July 4, 2013, 10:15am UTC](https://www.ruby-forum.com/t/bug-in-volk-on-armv6/245679/3 "2013-07-04T10:15:09Z")
Confirming that this fix works. But now volk test fails in another place. (gdb) run Starting program: /root/gnuradio-3.6.5.1/build/volk/lib/test\_all [Thread debugging using libthread\_db enabled] Using host libthread\_db library “/lib/arm-linux-gnueabihf/libthread\_db.so.1”. Running 92 test cases… Using Volk machine: generic\_orc RUN\_VOLK\_TESTS: volk\_16ic\_s32f\_deinterleave\_real\_32f\_a no architectures to test RUN\_VOLK\_TESTS: volk\_16ic\_deinterleave\_real\_8i\_a generic completed in 0s orc completed in 0.01s Best arch: generic RUN\_VOLK\_TESTS: volk\_16ic\_deinterleave\_16i\_x2\_a generic completed in 0s orc completed in 0s Best arch: generic RUN\_VOLK\_TESTS: volk\_16ic\_s32f\_deinterleave\_32f\_x2\_a generic completed in 0.01s Program received signal SIGSEGV, Segmentation fault. 0xb6f96428 in ?? () from /usr/lib/arm-linux-gnueabihf/libvolk.so.0.0.0 (gdb) bt #0 0xb6f96428 in ?? () #from /usr/lib/arm-linux-gnueabihf/libvolk.so.0.0.0 Cannot access memory at address 0x0 #1 0xb6db3da0 in pthread\_mutex\_unlock () #from /lib/arm-linux-gnueabihf/libc.so.6 2 0x00000020 in ?? () #3 0x00000020 in ?? () Backtrace stopped: previous frame identical to this frame (corrupt stack?) P.S. I’m trying to compile gnuradio on raspbian on Raspberry Pi board. All other tests passed ok, so some broken volk function shouldn’t be a problem for me since I only want to use network sink on it. On Wed, 3 Jul 2013 10:55:06 -0400 Tom R. [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid) wrote: > > descriptor auxvec\_f will always be true. > > +++ b/volk/tmpl/volk\_cpu.tmpl.c > > - > > ``` > r = fread(auxvec, sizeof(unsigned long), 2, auxvec_f); > if((auxvec[0] == AT_HWCAP) && (auxvec[1] & HWCAP_NEON)) > found_neon = 1; > > ``` > > } > > Or might be better to test ‘if(r \< 2)’ after the fread line and break > in case an error occurs and we only get 1 character out and not try to > read from it again. Probably not a bit deal since we’re only reading > in two chars at a time. > > Tom – Alexey B. [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid) ---
**Author:** ![snaury](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/s/77aa72/32.png) [@snaury](https://www.ruby-forum.com/u/snaury)\ **Post date:** [July 16, 2013, 4:42pm UTC](https://www.ruby-forum.com/t/bug-in-volk-on-armv6/245679/4 "2013-07-16T16:42:04Z")
On Thu, Jul 4, 2013 at 4:13 AM, Alexey B. [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid) wrote: > RUN\_VOLK\_TESTS: volk\_16ic\_s32f\_deinterleave\_real\_32f\_a > generic completed in 0.01s > Backtrace stopped: previous frame identical to this frame (corrupt > stack?) > > P.S. I’m trying to compile gnuradio on raspbian on Raspberry Pi board. > All other tests passed ok, so some broken volk function shouldn’t be a > problem for me since I only want to use network sink on it. Hi Alexey, Yeah, you shouldn’t care too much about Volk on the Raspberry Pi since there’s nothing optimized for that platform. You just need to make sure all call to Volk work using the ‘generic’ proto-kernels (in other words, you need Volk because we use functions from the library in GNU Radio, but you won’t access any of the SIMD-optimized versions of the code). To learn more about the QA failure above, it’d be better to build with debug symbols on (-DCMAKE\_BUILD\_TYPE=“Debug”) to get more information out of that back trace. Tom