From: Benjamin Herrenschmidt <benh@kernel.crashing.org>
To: Yinghai Lu <yinghai@kernel.org>
Cc: Ingo Molnar <mingo@elte.hu>, Thomas Gleixner <tglx@linutronix.de>,
"H. Peter Anvin" <hpa@zytor.com>,
Andrew Morton <akpm@linux-foundation.org>,
David Miller <davem@davemloft.net>,
Linus Torvalds <torvalds@linux-foundation.org>,
Johannes Weiner <hannes@cmpxchg.org>,
linux-kernel@vger.kernel.org, linux-arch@vger.kernel.org
Subject: Re: [PATCH 08/39] lmb: Add lmb_reserve_area/lmb_free_area
Date: Tue, 13 Apr 2010 14:15:52 +1000 [thread overview]
Message-ID: <1271132152.13059.63.camel@pasglop> (raw)
In-Reply-To: <1270793048-23796-9-git-send-email-yinghai@kernel.org>
On Thu, 2010-04-08 at 23:03 -0700, Yinghai Lu wrote:
> They will check if the region array is big enough.
>
> __check_and_double_region_array will try to double the region array if that
> array spare slots is not big enough. Old array will be copied to new array.
>
> Arch code should set lmb.default_alloc_limit accordingly, so the new array is in
> accessiable address.
>
> -v2: change get_max_mapped() to lmb.default_alloc_limit according to Michael
> Ellerman and Ben
> change to lmb_reserve_area and lmb_free_area according to Michael Ellerman
> -v3: call check_and_double after reserve/free, so could avoid to use
> find_lmb_area. Suggested by Michael Ellerman
>
> Signed-off-by: Yinghai Lu <yinghai@kernel.org>
So a few things here:
default_alloc_limit: This should be a patch of its own I believe, we
should provide a way for callers to also honor the limit, I'm sure
without that we're going to hit funny problems -especially- if we start
replacing bootmem. (Heh, low/high mem anyone ?)
I would think that the basic lmb_alloc() should be modified to use the
current limit, and maybe add an lmb_alloc_anywhere() as an inline
wrapper to lmb_alloc_base(..., LMB_ALLOC_ANYWHERE); In fact, lmb_alloc()
should become an inline wrapper too.
Also, the way you added the calls to __check_and_double_region_array()
is fishy (what a function name btw !). IE. You added it in 2 or 3
places, missing a whole bunch, which will guarantee some kind of
unexpected behaviour especially when using the _nid variants.
Now, maybe the idea of moving things to -after- the call wasn't that
good. I still don't quite get why we can't do things lazily, especially
if we remove some of the code duplication in there.
In any case, its about time to clarify what is API and what is internal
in LMB and clean up the entry path.
Cheers,
Ben.
> ---
> include/linux/lmb.h | 4 +++
> mm/lmb.c | 66 +++++++++++++++++++++++++++++++++++++++++++++++++++
> 2 files changed, 70 insertions(+), 0 deletions(-)
>
> diff --git a/include/linux/lmb.h b/include/linux/lmb.h
> index 4cf2f3b..598662f 100644
> --- a/include/linux/lmb.h
> +++ b/include/linux/lmb.h
> @@ -33,6 +33,7 @@ struct lmb_region {
> struct lmb {
> unsigned long debug;
> u64 rmo_size;
> + u64 default_alloc_limit;
> struct lmb_region memory;
> struct lmb_region reserved;
> };
> @@ -83,6 +84,9 @@ lmb_end_pfn(struct lmb_region *type, unsigned long region_nr)
> lmb_size_pages(type, region_nr);
> }
>
> +void lmb_reserve_area(u64 start, u64 end, char *name);
> +void lmb_free_area(u64 start, u64 end);
> +void lmb_add_memory(u64 start, u64 end);
> u64 __lmb_find_area(u64 ei_start, u64 ei_last, u64 start, u64 end,
> u64 size, u64 align);
> u64 lmb_find_area(u64 start, u64 end, u64 size, u64 align);
> diff --git a/mm/lmb.c b/mm/lmb.c
> index 7010212..a514d41 100644
> --- a/mm/lmb.c
> +++ b/mm/lmb.c
> @@ -564,6 +564,72 @@ int lmb_find(struct lmb_property *res)
> return -1;
> }
>
> +static void __init __check_and_double_region_array(struct lmb_region *type,
> + struct lmb_property *static_region)
> +{
> + u64 size, mem;
> + struct lmb_property *new, *old;
> + unsigned long rgnsz = type->nr_regions;
> +
> + /* Do we have enough slots left ? */
> + if ((rgnsz - type->cnt) > 2)
> + return;
> +
> + old = type->region;
> + /* Double the array size */
> + size = sizeof(struct lmb_property) * rgnsz * 2;
> +
> + mem = __lmb_alloc_base(size, sizeof(struct lmb_property), lmb.default_alloc_limit);
> + if (mem == 0)
> + panic("can not find more space for lmb.reserved.region array");
> +
> + new = __va(mem);
> + /* Copy old to new */
> + memcpy(&new[0], &old[0], sizeof(struct lmb_property) * rgnsz);
> + memset(&new[rgnsz], 0, sizeof(struct lmb_property) * rgnsz);
> +
> + memset(&old[0], 0, sizeof(struct lmb_property) * rgnsz);
> + type->region = new;
> + type->nr_regions = rgnsz * 2;
> + printk(KERN_DEBUG "lmb.reserved.region array is doubled to %ld at [%llx - %llx]\n",
> + type->nr_regions, mem, mem + size - 1);
> +
> + /* Free old one ?*/
> + if (old != static_region)
> + lmb_free(__pa(old), sizeof(struct lmb_property) * rgnsz);
> +}
> +
> +void __init lmb_add_memory(u64 start, u64 end)
> +{
> + lmb_add_region(&lmb.memory, start, end - start);
> + __check_and_double_region_array(&lmb.memory, &lmb_memory_region[0]);
> +}
> +
> +void __init lmb_reserve_area(u64 start, u64 end, char *name)
> +{
> + if (start == end)
> + return;
> +
> + if (WARN_ONCE(start > end, "lmb_reserve_area: wrong range [%#llx, %#llx]\n", start, end))
> + return;
> +
> + lmb_add_region(&lmb.reserved, start, end - start);
> + __check_and_double_region_array(&lmb.reserved, &lmb_reserved_region[0]);
> +}
> +
> +void __init lmb_free_area(u64 start, u64 end)
> +{
> + if (start == end)
> + return;
> +
> + if (WARN_ONCE(start > end, "lmb_free_area: wrong range [%#llx, %#llx]\n", start, end))
> + return;
> +
> + /* keep punching hole, could run out of slots too */
> + lmb_free(start, end - start);
> + __check_and_double_region_array(&lmb.reserved, &lmb_reserved_region[0]);
> +}
> +
> u64 __init __weak __lmb_find_area(u64 ei_start, u64 ei_last, u64 start, u64 end,
> u64 size, u64 align)
> {
next prev parent reply other threads:[~2010-04-13 4:17 UTC|newest]
Thread overview: 90+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-04-09 6:03 [PATCH -v12 00/39] use lmb with x86 Yinghai Lu
2010-04-09 6:03 ` [PATCH 01/39] swiotlb: Use page alignment for early buffer allocation Yinghai Lu
2010-04-09 6:03 ` Yinghai Lu
2010-04-09 6:03 ` [PATCH 02/39] x86: Add sanitize_e820_map() Yinghai Lu
2010-04-09 6:03 ` Yinghai Lu
2010-04-09 6:03 ` [PATCH 03/39] x86: Align e820 ram range to page Yinghai Lu
2010-04-09 6:03 ` [PATCH 04/39] lmb: Move lmb.c to mm/ Yinghai Lu
2010-04-09 6:03 ` Yinghai Lu
2010-04-13 3:52 ` Benjamin Herrenschmidt
2010-04-09 6:03 ` [PATCH 05/39] lmb: Seperate region array from lmb_region struct Yinghai Lu
2010-04-13 3:56 ` Benjamin Herrenschmidt
2010-04-09 6:03 ` [PATCH 06/39] lmb: Seperate __lmb_find_base() from __lmb_alloc_base() Yinghai Lu
2010-04-09 6:03 ` Yinghai Lu
2010-04-13 3:58 ` Benjamin Herrenschmidt
2010-04-13 4:31 ` Yinghai
2010-04-09 6:03 ` [PATCH 07/39] lmb: Add lmb_find_area() Yinghai Lu
2010-04-13 4:05 ` Benjamin Herrenschmidt
2010-04-13 4:29 ` Yinghai
2010-04-13 5:07 ` Benjamin Herrenschmidt
2010-04-13 5:26 ` Yinghai
2010-04-13 5:46 ` H. Peter Anvin
2010-04-13 10:15 ` Benjamin Herrenschmidt
2010-04-09 6:03 ` [PATCH 08/39] lmb: Add lmb_reserve_area/lmb_free_area Yinghai Lu
2010-04-09 6:03 ` Yinghai Lu
2010-04-13 4:15 ` Benjamin Herrenschmidt [this message]
2010-04-09 6:03 ` [PATCH 09/39] bootmem, x86: Add weak version of reserve_bootmem_generic Yinghai Lu
2010-04-09 6:03 ` [PATCH 10/39] lmb: Add lmb_to_bootmem() Yinghai Lu
2010-04-09 6:03 ` Yinghai Lu
2010-04-09 6:03 ` [PATCH 11/39] lmb: Add get_free_all_memory_range() Yinghai Lu
2010-04-09 6:03 ` [PATCH 12/39] lmb: Add lmb_register_active_regions() and lmb_hole_size() Yinghai Lu
2010-04-09 6:03 ` Yinghai Lu
2010-04-09 6:03 ` [PATCH 13/39] lmb: Prepare to include linux/lmb.h in core file Yinghai Lu
2010-04-09 6:03 ` Yinghai Lu
2010-04-09 6:03 ` [PATCH 14/39] lmb: Add find_memory_core_early() Yinghai Lu
2010-04-09 6:03 ` [PATCH 15/39] lmb: Add lmb_find_area_node() Yinghai Lu
2010-04-09 6:03 ` Yinghai Lu
2010-04-13 4:18 ` Benjamin Herrenschmidt
2010-04-13 4:37 ` Yinghai
2010-04-09 6:03 ` [PATCH 16/39] lmb: Add lmb_free_memory_size() Yinghai Lu
2010-04-09 6:03 ` Yinghai Lu
2010-04-09 6:03 ` [PATCH 17/39] lmb: Add lmb_memory_size() Yinghai Lu
2010-04-09 6:03 ` Yinghai Lu
2010-04-09 6:03 ` [PATCH 18/39] lmb: Add lmb_reserve_area_overlap_ok() Yinghai Lu
2010-04-09 6:03 ` Yinghai Lu
2010-04-13 4:21 ` Benjamin Herrenschmidt
2010-04-13 4:44 ` Yinghai
2010-04-13 5:10 ` Benjamin Herrenschmidt
2010-04-13 5:37 ` Yinghai
2010-04-13 5:37 ` Yinghai
2010-04-09 6:03 ` [PATCH 19/39] lmb: Use lmb_debug to control debug message print out Yinghai Lu
2010-04-09 6:03 ` Yinghai Lu
2010-04-09 6:03 ` [PATCH 20/39] lmb: Add ARCH_DISCARD_LMB to put lmb code to .init Yinghai Lu
2010-04-09 6:03 ` [PATCH 21/39] lmb: Add __free/__reserve/__clear_lmb_reserved_region_array() Yinghai Lu
2010-04-09 6:03 ` Yinghai Lu
2010-04-09 6:03 ` [PATCH 22/39] x86, lmb: Add lmb_find_area_size() Yinghai Lu
2010-04-09 6:03 ` [PATCH 23/39] x86, lmb: Add x86 version of __lmb_find_area() Yinghai Lu
2010-04-09 6:03 ` [PATCH 24/39] x86: Use lmb to replace early_res Yinghai Lu
2010-04-09 6:03 ` [PATCH 25/39] x86: Replace e820_/_early string with lmb_ Yinghai Lu
2010-04-09 6:03 ` Yinghai Lu
2010-04-09 6:03 ` [PATCH 26/39] nobootmem: use lmb.default_alloc_limit in alloc_bootmem path Yinghai Lu
2010-04-13 4:23 ` Benjamin Herrenschmidt
2010-04-13 4:50 ` Yinghai
2010-04-13 5:13 ` Benjamin Herrenschmidt
2010-04-13 5:42 ` Yinghai
2010-04-09 6:03 ` [PATCH 27/39] x86, lmb: turn off ARCH_LMB_FIND_AREA Yinghai Lu
2010-04-09 6:03 ` Yinghai Lu
2010-04-09 6:03 ` [PATCH 28/39] x86: Remove not used early_res code Yinghai Lu
2010-04-09 6:03 ` Yinghai Lu
2010-04-09 6:03 ` [PATCH 29/39] x86, lmb: Use lmb_memory_size()/lmb_free_memory_size() to get correct dma_reserve Yinghai Lu
2010-04-09 6:03 ` Yinghai Lu
2010-04-09 6:03 ` [PATCH 30/39] x86: put 64 bit numa node memmap above 16M Yinghai Lu
2010-04-09 6:03 ` Yinghai Lu
2010-04-09 6:04 ` [PATCH 31/39] x86: Use wake_system_ram_range() instead of e820_any_mapped() in agp path Yinghai Lu
2010-04-09 6:04 ` Yinghai Lu
2010-04-09 6:04 ` [PATCH 32/39] x86: Add get_centaur_ram_top() Yinghai Lu
2010-04-09 6:04 ` Yinghai Lu
2010-04-09 6:04 ` [PATCH 33/39] x86: Change e820_any_mapped() to __init Yinghai Lu
2010-04-09 6:04 ` Yinghai Lu
2010-04-09 6:04 ` [PATCH 34/39] x86: Use walk_system_ream_range() instead of referring e820.map directly for tboot Yinghai Lu
2010-04-09 6:04 ` Yinghai Lu
2010-04-09 6:04 ` [PATCH 35/39] x86: make e820 to be __initdata Yinghai Lu
2010-04-09 6:04 ` Yinghai Lu
2010-04-09 6:04 ` [PATCH 36/39] bootmem: Add nobootmem.c to reduce the #ifdef Yinghai Lu
2010-04-09 6:04 ` Yinghai Lu
2010-04-09 6:04 ` [PATCH 37/39] mm: move contig_page_data define to bootmem.c/nobootmem.c Yinghai Lu
2010-04-09 6:04 ` [PATCH 38/39] lmb: move __alloc_memory_core_early() to nobootmem.c Yinghai Lu
2010-04-09 6:04 ` Yinghai Lu
2010-04-09 6:04 ` [PATCH 39/39] x86: have nobootmem version setup_bootmem_allocator() Yinghai Lu
2010-04-09 6:04 ` Yinghai Lu
2010-04-13 3:41 ` [PATCH -v12 00/39] use lmb with x86 Benjamin Herrenschmidt
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1271132152.13059.63.camel@pasglop \
--to=benh@kernel.crashing.org \
--cc=akpm@linux-foundation.org \
--cc=davem@davemloft.net \
--cc=hannes@cmpxchg.org \
--cc=hpa@zytor.com \
--cc=linux-arch@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@elte.hu \
--cc=tglx@linutronix.de \
--cc=torvalds@linux-foundation.org \
--cc=yinghai@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).