On 11/23/23 at 02:43pm, Alexander Gordeev wrote: > On Thu, Nov 23, 2023 at 03:36:52PM +0800, Baoquan He wrote: > > Hi Baoquan, > > > No proof is found to require that S390 architecture has to select > > KEXEC. At least from my testing at below, dropping select of KEXEC won't > > impact anything. > > It does impact the outcome of defconfigs. > Namely, CONFIG_KEXEC is not set with this patch. Right, CONFIG_KEXEC won't be set defaultly with this patch applied. > > > ===testing 1=== > > CONFIG_CRASH_CORE=y > > CONFIG_KEXEC_CORE=y > > CONFIG_CRASH_DUMP=y > > === > > > > ===testing 2=== > > CONFIG_CRASH_CORE=y > > CONFIG_KEXEC_CORE=y > > CONFIG_KEXEC_FILE=y > > CONFIG_CRASH_DUMP=y > > === > > Unfortunately, I do not quite realize what these testings were > and what is the difference between the two. Both these two testings have CONFIG_KEXEC=n, and building all passed. I wound't present their difference, but two cases where no CONFIG_KEXEC is set and no dependency on CONFIG_KEXEC is seen. > > > So drop the select of KEXEC now. > > I suggest dropping this patch. Once the previous two are upstream > we would remove 'select KEXEC' from Kconfig together with defconfig > updates. I see your concern, will drop this one in v2. Thanks for checking these. > > > Signed-off-by: Baoquan He <bhe@xxxxxxxxxx> > > --- > > arch/s390/Kconfig | 1 - > > 1 file changed, 1 deletion(-) > > > > diff --git a/arch/s390/Kconfig b/arch/s390/Kconfig > > index 3bec98d20283..1aec2e692dca 100644 > > --- a/arch/s390/Kconfig > > +++ b/arch/s390/Kconfig > > @@ -217,7 +217,6 @@ config S390 > > select HAVE_VIRT_CPU_ACCOUNTING_IDLE > > select IOMMU_HELPER if PCI > > select IOMMU_SUPPORT if PCI > > - select KEXEC > > select MMU_GATHER_MERGE_VMAS > > select MMU_GATHER_NO_GATHER > > select MMU_GATHER_RCU_TABLE_FREE > > -- > > Thanks! >
