Re: [Spice-devel] usbredirparser: fix EAGAIN error in *_do_write()

2016-10-15 Thread Victor Toso
Hi, On Sat, Oct 15, 2016 at 12:28:21PM +, Dmitriy Samborskiy wrote: > > From your description looks like the patch is reversed. > > Yes, it's reversed, sorry. > > > I'm not that familiar with usbredir code but looking at the code > > if usbredirparser_do_write returns 0 the server should just

Re: [Spice-devel] usbredirparser: fix EAGAIN error in *_do_write()

2016-10-15 Thread Dmitriy Samborskiy
> From your description looks like the patch is reversed. Yes, it's reversed, sorry. > I'm not that familiar with usbredir code but looking at the code > if usbredirparser_do_write returns 0 the server should just loop > again. You are right. I found this bug when I debugged the usbredir sources

Re: [Spice-devel] usbredirparser: fix EAGAIN error in *_do_write()

2016-10-14 Thread Frediano Ziglio
> > Hello, > > I'd like to propose a bugfix for usbredirparser.c usbredirparser_do_write() > procedure, see below. > > > Issue description: > > usbredirparser_do_write() does not handle EAGAIN properly. > In case of large write parser->callb.write_func() returns 0 with > errno=EAGAIN. > > The

[Spice-devel] usbredirparser: fix EAGAIN error in *_do_write()

2016-10-14 Thread Dmitriy Samborskiy
Hello, I'd like to propose a bugfix for usbredirparser.c usbredirparser_do_write() procedure, see below. Issue description: usbredirparser_do_write() does not handle EAGAIN properly. In case of large write parser->callb.write_func() returns 0 with errno=EAGAIN. Then usbredirparser_do_write() c