Re: [PATCH 02/16] vl: extract accelerator option processing to a separate function

2019-11-18 Thread Paolo Bonzini
On 18/11/19 11:58, Thomas Huth wrote: >> +static void configure_accelerators(void) >> +{ >> +qemu_opts_foreach(qemu_find_opts("icount"), >> + do_configure_icount, NULL, &error_fatal); >> + >> +qemu_opts_foreach(qemu_find_opts("accel"), >> + do_confi

Re: [PATCH 02/16] vl: extract accelerator option processing to a separate function

2019-11-18 Thread Thomas Huth
On 13/11/2019 15.38, Paolo Bonzini wrote: > As a first step towards supporting multiple "-accel" options, push -icount > and -accel semantics into a new function, and use qemu_opts_foreach to > retrieve the key/value lists instead of stashing them into globals. > > Signed-off-by: Paolo Bonzini >

Re: [PATCH 02/16] vl: extract accelerator option processing to a separate function

2019-11-14 Thread Paolo Bonzini
On 14/11/19 08:55, Marc-André Lureau wrote: >> + >> +qemu_opts_foreach(qemu_find_opts("accel"), >> + do_configure_accelerator, NULL, &error_fatal); > It used to call qemu_tcg_configure() when no -accel option given. In > this case, it still sets mttcg_enabled = default_mttc

Re: [PATCH 02/16] vl: extract accelerator option processing to a separate function

2019-11-13 Thread Marc-André Lureau
Hi On Wed, Nov 13, 2019 at 6:39 PM Paolo Bonzini wrote: > > As a first step towards supporting multiple "-accel" options, push -icount > and -accel semantics into a new function, and use qemu_opts_foreach to > retrieve the key/value lists instead of stashing them into globals. > > Signed-off-by:

[PATCH 02/16] vl: extract accelerator option processing to a separate function

2019-11-13 Thread Paolo Bonzini
As a first step towards supporting multiple "-accel" options, push -icount and -accel semantics into a new function, and use qemu_opts_foreach to retrieve the key/value lists instead of stashing them into globals. Signed-off-by: Paolo Bonzini --- vl.c | 40 ---