| Message ID | 20210927081402.191717-1-wangkefeng.wang@huawei.com (mailing list archive) |
|---|---|
| Headers | show |
| Series | Cleanup MAY_HAVE_SPARSE_IRQ | expand |
I didn't see the patch delete:
#else /* !CONFIG_SPARSE_IRQ */
struct irq_desc irq_desc[NR_IRQS] __cacheline_aligned_in_smp = {
[0 ... NR_IRQS-1] = {
.handle_irq = handle_bad_irq,
.depth = 1,
.lock = __RAW_SPIN_LOCK_UNLOCKED(irq_desc->lock),
}
};
...
Flat irq_desc[] is simple and easy for debugging. We do want to del it?
On Mon, Sep 27, 2021 at 4:11 PM Kefeng Wang <wangkefeng.wang@huawei.com> wrote:
>
> Most ARCHs support SPARSE_IRQ, and MAY_HAVE_SPARSE_IRQ is useless, and
> only sh and csky select it, but the could use SPARSE_IRQ too, let's
> kill MAY_HAVE_SPARSE_IRQ, also cleanup the kernel/irq/Kconfig a little.
>
> Kefeng Wang (3):
> sh: Cleanup about SPARSE_IRQ
> csky: Use SPARSE_IRQ
> genirq: Cleanup Kconfig
>
> arch/csky/Kconfig | 2 +-
> arch/sh/Kconfig | 1 -
> arch/sh/include/asm/irq.h | 9 -------
> kernel/irq/Kconfig | 50 ++++++++++++++++-----------------------
> 4 files changed, 21 insertions(+), 41 deletions(-)
>
> --
> 2.26.2
>
On 2021/9/27 20:09, Guo Ren wrote: > I didn't see the patch delete: > #else /* !CONFIG_SPARSE_IRQ */ > struct irq_desc irq_desc[NR_IRQS] __cacheline_aligned_in_smp = { > [0 ... NR_IRQS-1] = { > .handle_irq = handle_bad_irq, > .depth = 1, > .lock = __RAW_SPIN_LOCK_UNLOCKED(irq_desc->lock), > } > }; > ... > > Flat irq_desc[] is simple and easy for debugging. We do want to del it? This patches want to kill MAY_HAVE_SPARSE_IRQ, not !SPARSE_IRQ. so I won't delete above parts(eg, ARM could use both SPARSE_IRQ and !SPARSE_IRQ via different config,) > > On Mon, Sep 27, 2021 at 4:11 PM Kefeng Wang <wangkefeng.wang@huawei.com> wrote: >> Most ARCHs support SPARSE_IRQ, and MAY_HAVE_SPARSE_IRQ is useless, and >> only sh and csky select it, but the could use SPARSE_IRQ too, let's >> kill MAY_HAVE_SPARSE_IRQ, also cleanup the kernel/irq/Kconfig a little. >> >> Kefeng Wang (3): >> sh: Cleanup about SPARSE_IRQ >> csky: Use SPARSE_IRQ >> genirq: Cleanup Kconfig >> >> arch/csky/Kconfig | 2 +- >> arch/sh/Kconfig | 1 - >> arch/sh/include/asm/irq.h | 9 ------- >> kernel/irq/Kconfig | 50 ++++++++++++++++----------------------- >> 4 files changed, 21 insertions(+), 41 deletions(-) >> >> -- >> 2.26.2 >> >
On Mon, Sep 27, 2021 at 8:35 PM Kefeng Wang <wangkefeng.wang@huawei.com> wrote: > > > On 2021/9/27 20:09, Guo Ren wrote: > > I didn't see the patch delete: > > #else /* !CONFIG_SPARSE_IRQ */ > > struct irq_desc irq_desc[NR_IRQS] __cacheline_aligned_in_smp = { > > [0 ... NR_IRQS-1] = { > > .handle_irq = handle_bad_irq, > > .depth = 1, > > .lock = __RAW_SPIN_LOCK_UNLOCKED(irq_desc->lock), > > } > > }; > > ... > > > > Flat irq_desc[] is simple and easy for debugging. We do want to del it? > > This patches want to kill MAY_HAVE_SPARSE_IRQ, not !SPARSE_IRQ. > > so I won't delete above parts(eg, ARM could use both SPARSE_IRQ and > > !SPARSE_IRQ via different config,) > > > > > On Mon, Sep 27, 2021 at 4:11 PM Kefeng Wang <wangkefeng.wang@huawei.com> wrote: > >> Most ARCHs support SPARSE_IRQ, and MAY_HAVE_SPARSE_IRQ is useless, and > >> only sh and csky select it, but the could use SPARSE_IRQ too, let's > >> kill MAY_HAVE_SPARSE_IRQ, also cleanup the kernel/irq/Kconfig a little. Why couldn't choice SPARSE in menuconfig? > >> > >> Kefeng Wang (3): > >> sh: Cleanup about SPARSE_IRQ > >> csky: Use SPARSE_IRQ > >> genirq: Cleanup Kconfig > >> > >> arch/csky/Kconfig | 2 +- > >> arch/sh/Kconfig | 1 - > >> arch/sh/include/asm/irq.h | 9 ------- > >> kernel/irq/Kconfig | 50 ++++++++++++++++----------------------- > >> 4 files changed, 21 insertions(+), 41 deletions(-) > >> > >> -- > >> 2.26.2 > >> > >
On Mon, Sep 27, 2021 at 4:11 PM Kefeng Wang <wangkefeng.wang@huawei.com> wrote: > > Most ARCHs support SPARSE_IRQ, and MAY_HAVE_SPARSE_IRQ is useless, and > only sh and csky select it, but the could use SPARSE_IRQ too, let's > kill MAY_HAVE_SPARSE_IRQ, also cleanup the kernel/irq/Kconfig a little. Can you elaborate the reason on why we need to kill MAY_HAVE_SPARSE_IRQ? What are the benefits after the patch? (As you know we couldn't drop "!SPARSE_IRQ".) > > Kefeng Wang (3): > sh: Cleanup about SPARSE_IRQ > csky: Use SPARSE_IRQ > genirq: Cleanup Kconfig > > arch/csky/Kconfig | 2 +- > arch/sh/Kconfig | 1 - > arch/sh/include/asm/irq.h | 9 ------- > kernel/irq/Kconfig | 50 ++++++++++++++++----------------------- > 4 files changed, 21 insertions(+), 41 deletions(-) > > -- > 2.26.2 >
On 2021/9/28 13:08, Guo Ren wrote: > On Mon, Sep 27, 2021 at 4:11 PM Kefeng Wang <wangkefeng.wang@huawei.com> wrote: >> Most ARCHs support SPARSE_IRQ, and MAY_HAVE_SPARSE_IRQ is useless, and >> only sh and csky select it, but the could use SPARSE_IRQ too, let's >> kill MAY_HAVE_SPARSE_IRQ, also cleanup the kernel/irq/Kconfig a little. > Can you elaborate the reason on why we need to kill MAY_HAVE_SPARSE_IRQ? > What are the benefits after the patch? (As you know we couldn't drop > "!SPARSE_IRQ".) If csky want to keep MAY_HAVE_SPARSE_IRQ, then I won't kill it, or no one use it, then cleanup it.