fwrite line buffering
Hi all, while working with stdout fwrite function, I have noticed that line buffering is not working. Is this by design or it is a bug? Regards, Jernej
Re: fwrite line buffering
Hi Jernej, Please define "not working"! :-) Are you flushing it? Also "\n" at end of line forces it to be flushed. BR, Alan On Friday, May 14, 2021, Jernej Turnsek wrote: > Hi all, > > while working with stdout fwrite function, I have noticed that line > buffering is not working. Is this by design or it is a bug? > > Regards, > Jernej >
Re: fwrite line buffering
I think his point was that it was implicitly flushed at every write, and not being buffered, but I might have misunderstood. Anyway, I wouldn't expect fwrite to buffer on a per-line basis. fwrite is basically there for doing binary I/O. If you want something buffered, line oriented and user friendly, I would suggest printf. Johnny On 2021-05-14 12:22, Alan Carvalho de Assis wrote: Hi Jernej, Please define "not working"! :-) Are you flushing it? Also "\n" at end of line forces it to be flushed. BR, Alan On Friday, May 14, 2021, Jernej Turnsek wrote: Hi all, while working with stdout fwrite function, I have noticed that line buffering is not working. Is this by design or it is a bug? Regards, Jernej -- Johnny Billquist || "I'm on a bus || on a psychedelic trip email: b...@softjar.se || Reading murder books pdp is alive! || tryin' to stay hip" - B. Idol
Re: fwrite line buffering
Hi Alan, "\n" does not explicitly force the flush. I have configured the line buffering which is actually working on printf and fputs, but not on fwrite. I am using some application code that is running ok on Linux, but on Nuttx it is not flushing properly. It flush only when the buffer is filled. I hope I make myself clear now. Jernej On Fri, May 14, 2021 at 12:22 PM Alan Carvalho de Assis wrote: > Hi Jernej, > > Please define "not working"! :-) > > Are you flushing it? Also "\n" at end of line forces it to be flushed. > > BR, > > Alan > > On Friday, May 14, 2021, Jernej Turnsek wrote: > > > Hi all, > > > > while working with stdout fwrite function, I have noticed that line > > buffering is not working. Is this by design or it is a bug? > > > > Regards, > > Jernej > > >
devif_callback_alloc() not releasing callback structure?
Hello everyone! I am still working on implementing 2 PHYs on IMXRT1064 and currently using netpkt example for transmitting the simplest packets. Here is my problem: Have 2 Ethernet interfaces (eth0 and eth1). eth0 is UP, eth1 is DOWN. Transmitting through eth1 results in packets coming out of interface that is UP (being eth0). Expected behavior when transmitting through eth0. eth0 is DOWN, eth1 is UP. Transmitting through eth0 results in packets coming out of interface that is UP (being eth1). Expected behavior when transmitting through eth1. eth0 is UP, eth1 is UP. Expected behavior when transmitting through eth0 and eth1. When going through the code in devif_callback.c, https://github.com/apache/incubator-nuttx/blob/3af0ef70ff5d5bd4d4d1cdb0152ee53182dc09dd/net/devif/devif_callback.c#L259-L270 I noticed an if condition which I am having difficulties understanding. The comment above it says something that, in my opinion, does not correspond to the code. Shouldn’t line 266 be this one instead? if (! netdev_verify (dev) || ! (dev-> d_flags & IFF_UP)) Once I made that change, the packets stopped being sent through incorrect interfaces when the specified interface is DOWN. However, now if I attempt to send packets with netpkt through an interface that is down multiple times (maybe 5-6? do not remember the precise number), at some point I will start getting ‘ERROR: Failed to allocate callback’ message. At this point, it doesn’t matter which interface I select, the error is always printed. Then I noticed that line 270: devif_callback_free ( NULL , NULL , list); Will probably not do anything because the function checks the second argument for NULL right at the very beginning. This results in function not doing anything at all. Is that the expected behavior? Please let me know your thoughts! Regards, Vadim Yastrebov P.S. in net/pkt/pkt_finddev.c In function pkt_find_device(), originally, ‘eth0’ string was hardcoded when looking up the device. This prevented me from selecting the device for transmission. Instead, I used netdev_findbyindex() which now works just fine and is able to provide the correct device.
