On Tue, Feb 15, 2022 at 03:45:00PM +0100, Jan Beulich wrote:
> On 15.02.2022 15:16, Roger Pau Monné wrote:
> > I could add to the commit message:
> >
> > "Originally _QR_BIAS was subtracted from the result of
> > atomic_add_return_acquire in order to prevent GCC from emitting an
> > unneeded ADD i
On 15.02.2022 15:16, Roger Pau Monné wrote:
> On Tue, Feb 15, 2022 at 02:22:02PM +0100, Jan Beulich wrote:
>> On 15.02.2022 14:13, Julien Grall wrote:
>>> On 15/02/2022 09:39, Roger Pau Monne wrote:
There's no need to subtract _QR_BIAS from the lock value for storing
in the local cnts var
On Tue, Feb 15, 2022 at 02:22:02PM +0100, Jan Beulich wrote:
> On 15.02.2022 14:13, Julien Grall wrote:
> > On 15/02/2022 09:39, Roger Pau Monne wrote:
> >> There's no need to subtract _QR_BIAS from the lock value for storing
> >> in the local cnts variable in the read lock slow path: the users of
On 15.02.2022 14:13, Julien Grall wrote:
> On 15/02/2022 09:39, Roger Pau Monne wrote:
>> There's no need to subtract _QR_BIAS from the lock value for storing
>> in the local cnts variable in the read lock slow path: the users of
>> the value in cnts only care about the writer-related bits and use
Hi,
On 15/02/2022 09:39, Roger Pau Monne wrote:
There's no need to subtract _QR_BIAS from the lock value for storing
in the local cnts variable in the read lock slow path: the users of
the value in cnts only care about the writer-related bits and use a
mask to get the value.
Note that further s
> On 15 Feb 2022, at 09:39, Roger Pau Monne wrote:
>
> There's no need to subtract _QR_BIAS from the lock value for storing
> in the local cnts variable in the read lock slow path: the users of
> the value in cnts only care about the writer-related bits and use a
> mask to get the value.
>
>
On 15.02.2022 10:39, Roger Pau Monne wrote:
> There's no need to subtract _QR_BIAS from the lock value for storing
> in the local cnts variable in the read lock slow path: the users of
> the value in cnts only care about the writer-related bits and use a
> mask to get the value.
>
> Note that furt
There's no need to subtract _QR_BIAS from the lock value for storing
in the local cnts variable in the read lock slow path: the users of
the value in cnts only care about the writer-related bits and use a
mask to get the value.
Note that further setting of cnts in rspin_until_writer_unlock already