From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Subject: Re: [PATCH v2 4/8] mm/memory_hotplug: Create memory block devices after arch_add_memory() References: <20190507183804.5512-1-david@redhat.com> <20190507183804.5512-5-david@redhat.com> <20190509143151.zexjmwu3ikkmye7i@master> From: David Hildenbrand Message-ID: <28071389-372c-14eb-1209-02464726b4f0@redhat.com> Date: Thu, 9 May 2019 16:58:56 +0200 MIME-Version: 1.0 In-Reply-To: <20190509143151.zexjmwu3ikkmye7i@master> Content-Type: text/plain; charset=utf-8 Content-Language: en-US Content-Transfer-Encoding: 7bit Sender: owner-linux-mm@kvack.org List-Archive: List-Post: To: Wei Yang Cc: linux-mm@kvack.org, linux-kernel@vger.kernel.org, linux-ia64@vger.kernel.org, linuxppc-dev@lists.ozlabs.org, linux-s390@vger.kernel.org, linux-sh@vger.kernel.org, akpm@linux-foundation.org, Dan Williams , Greg Kroah-Hartman , "Rafael J. Wysocki" , "mike.travis@hpe.com" , Ingo Molnar , Andrew Banman , Oscar Salvador , Michal Hocko , Pavel Tatashin , Qian Cai , Arun KS , Mathieu Malaterre List-ID: On 09.05.19 16:31, Wei Yang wrote: > On Tue, May 07, 2019 at 08:38:00PM +0200, David Hildenbrand wrote: >> Only memory to be added to the buddy and to be onlined/offlined by >> user space using memory block devices needs (and should have!) memory >> block devices. >> >> Factor out creation of memory block devices Create all devices after >> arch_add_memory() succeeded. We can later drop the want_memblock parameter, >> because it is now effectively stale. >> >> Only after memory block devices have been added, memory can be onlined >> by user space. This implies, that memory is not visible to user space at >> all before arch_add_memory() succeeded. >> >> Cc: Greg Kroah-Hartman >> Cc: "Rafael J. Wysocki" >> Cc: David Hildenbrand >> Cc: "mike.travis@hpe.com" >> Cc: Andrew Morton >> Cc: Ingo Molnar >> Cc: Andrew Banman >> Cc: Oscar Salvador >> Cc: Michal Hocko >> Cc: Pavel Tatashin >> Cc: Qian Cai >> Cc: Wei Yang >> Cc: Arun KS >> Cc: Mathieu Malaterre >> Signed-off-by: David Hildenbrand >> --- >> drivers/base/memory.c | 70 ++++++++++++++++++++++++++---------------- >> include/linux/memory.h | 2 +- >> mm/memory_hotplug.c | 15 ++++----- >> 3 files changed, 53 insertions(+), 34 deletions(-) >> >> diff --git a/drivers/base/memory.c b/drivers/base/memory.c >> index 6e0cb4fda179..862c202a18ca 100644 >> --- a/drivers/base/memory.c >> +++ b/drivers/base/memory.c >> @@ -701,44 +701,62 @@ static int add_memory_block(int base_section_nr) >> return 0; >> } >> >> +static void unregister_memory(struct memory_block *memory) >> +{ >> + BUG_ON(memory->dev.bus != &memory_subsys); >> + >> + /* drop the ref. we got via find_memory_block() */ >> + put_device(&memory->dev); >> + device_unregister(&memory->dev); >> +} >> + >> /* >> - * need an interface for the VM to add new memory regions, >> - * but without onlining it. >> + * Create memory block devices for the given memory area. Start and size >> + * have to be aligned to memory block granularity. Memory block devices >> + * will be initialized as offline. >> */ >> -int hotplug_memory_register(int nid, struct mem_section *section) >> +int hotplug_memory_register(unsigned long start, unsigned long size) > > One trivial suggestion about the function name. > > For memory_block device, sometimes we use the full name > > find_memory_block > init_memory_block > add_memory_block > > But sometimes we use *nick* name > > hotplug_memory_register > register_memory > unregister_memory > > This is a little bit confusion. > > Can we use one name convention here? We can just go for crate_memory_blocks() and free_memory_blocks(). Or do you have better suggestions? (I would actually even prefer "memory_block_devices", because memory blocks have different meanins) > > [...] > >> /* >> @@ -1106,6 +1100,13 @@ int __ref add_memory_resource(int nid, struct resource *res) >> if (ret < 0) >> goto error; >> >> + /* create memory block devices after memory was added */ >> + ret = hotplug_memory_register(start, size); >> + if (ret) { >> + arch_remove_memory(nid, start, size, NULL); > > Functionally, it works I think. > > But arch_remove_memory() would remove pages from zone. At this point, we just > allocate section/mmap for pages, the zones are empty and pages are not > connected to zone. > > Function zone = page_zone(page); always gets zone #0, since pages->flags is 0 > at this point. This is not exact. > > Would we add some comment to mention this? Or we need to clean up > arch_remove_memory() to take out __remove_zone()? That is precisely what is on my list next (see cover letter).This is already broken when memory that was never onlined is removed again. So I am planning to fix that independently. -- Thanks, David / dhildenb