Re: [uml-devel] /sys/class/tty/tty0/active?

2012-01-27 Thread richard -rw- weinberger
On Fri, Jan 27, 2012 at 3:02 PM, Alan Cox wrote: > On Fri, 27 Jan 2012 13:04:37 +0100 > richard -rw- weinberger wrote: > >> On Fri, Jan 27, 2012 at 12:51 PM, Alan Cox wrote: >> >> UML's console driver (arch/um/drivers/line.c) implements tty_operations. >> >> The crash happens because the tty sub

Re: [uml-devel] [PATCH] uml/hostfs: Propagate dirent.d_type to filldir()

2012-01-27 Thread Geert Uytterhoeven
Hi Christoph, On Fri, Jan 27, 2012 at 19:20, Christoph Hellwig wrote: > On Fri, Jan 27, 2012 at 07:14:58PM +0100, Geert Uytterhoeven wrote: >> Currently the (optional) d_type member in struct dirent is always >> DT_UNKNOWN on hostfs, which may confuse buggy software using readdir(). >> Make sure

Re: [uml-devel] [PATCH] uml/hostfs: Propagate dirent.d_type to filldir()

2012-01-27 Thread Christoph Hellwig
On Fri, Jan 27, 2012 at 07:14:58PM +0100, Geert Uytterhoeven wrote: > Currently the (optional) d_type member in struct dirent is always > DT_UNKNOWN on hostfs, which may confuse buggy software using readdir(). > Make sure to propagate its value from the underlying filesystem if it's > available the

[uml-devel] [PATCH] uml/hostfs: Propagate dirent.d_type to filldir()

2012-01-27 Thread Geert Uytterhoeven
Currently the (optional) d_type member in struct dirent is always DT_UNKNOWN on hostfs, which may confuse buggy software using readdir(). Make sure to propagate its value from the underlying filesystem if it's available there. Signed-off-by: Geert Uytterhoeven --- fs/hostfs/hostfs.h |3

Re: [uml-devel] rw semaphore lost wakeup in x86_64

2012-01-27 Thread richard -rw- weinberger
Hi, On Fri, Jan 27, 2012 at 5:07 AM, SeongHan Seo wrote: > I discovered a bug in your  X86_64 configuration system a while back which > involves lost wakeups for rw semaphores > > In what follows is an analysis of the problem > > and I'd be much obliged if you'd commit it to the mainline kernel >

[uml-devel] rw semaphore lost wakeup in x86_64

2012-01-27 Thread SeongHan Seo
Hi My name is Sunghan Suh and I'm a software engineer for a software company in South Korea I'd like to express my gratitude for your contributions to UML, which is helping me tremendously I discovered a bug in your X86_64 configuration system a while back which involves lost wakeups for rw se

Re: [uml-devel] /sys/class/tty/tty0/active?

2012-01-27 Thread Alan Cox
On Fri, 27 Jan 2012 13:04:37 +0100 richard -rw- weinberger wrote: > On Fri, Jan 27, 2012 at 12:51 PM, Alan Cox wrote: > >> UML's console driver (arch/um/drivers/line.c) implements tty_operations. > >> The crash happens because the tty subsystem calls the driver's close() > >> function and later

Re: [uml-devel] /sys/class/tty/tty0/active?

2012-01-27 Thread richard -rw- weinberger
On Fri, Jan 27, 2012 at 12:51 PM, Alan Cox wrote: >> UML's console driver (arch/um/drivers/line.c) implements tty_operations. >> The crash happens because the tty subsystem calls the driver's close() >> function and later >> write_room() or chars_in_buffer(). >> >> write_room() and chars_in_buffer

Re: [uml-devel] /sys/class/tty/tty0/active?

2012-01-27 Thread Alan Cox
> UML's console driver (arch/um/drivers/line.c) implements tty_operations. > The crash happens because the tty subsystem calls the driver's close() > function and later > write_room() or chars_in_buffer(). > > write_room() and chars_in_buffer() fail badly because close() already > cleaned up the d

Re: [uml-devel] /sys/class/tty/tty0/active?

2012-01-27 Thread richard -rw- weinberger
Greg, Kay, >> Is that a problem for UML somehow? The login on ttyX takes 30+ seconds. This very ugly and mostly wrong hack makes the login fast again: mount -o bind /sys/class/tty/console/ /sys/class/tty/tty0/ /sys/class/tty/console/active exists and contains "tty0"... > In general, the "crippl