[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] [PATCH v5 12/16] x86/hvm: remove HVMIO_dispatched I/O state

On 30/06/15 14:05, Paul Durrant wrote:
> By removing the HVMIO_dispatched state and making all pending emulations
> (i.e. all those not handled by the hypervisor) use HVMIO_awating_completion,
> various code-paths can be simplified.
> The completion case for HVMIO_dispatched can also be trivally removed
> from hvmemul_do_io() as it was already unreachable. This is because that
> state was only ever used for writes or I/O to/from a guest page and
> hvmemul_do_io() is never called to complete such I/O.
> NOTE: There is one sublety in handle_pio()... The only case when
>       handle_pio() got a return code of X86EMUL_RETRY back from
>       hvmemul_do_pio_buffer() and found io_state was not
>       HVMIO_awaiting_completion was in the case where the domain is
>       shutting down. This is because all writes normally yield a return
>       of HVMEMUL_OKAY and all reads put io_state into
>       HVMIO_awaiting_completion. Hence the io_state check there is
>       replaced with a check of the is_shutting_down flag on the domain.
> Signed-off-by: Paul Durrant <paul.durrant@xxxxxxxxxx>
> Cc: Keir Fraser <keir@xxxxxxx>
> Cc: Jan Beulich <jbeulich@xxxxxxxx>
> Cc: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>

Reviewed-by: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>

Xen-devel mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.