From: "Chen, Tiejun" <tiejun.chen@intel.com>
To: Jan Beulich <JBeulich@suse.com>
Cc: kevin.tian@intel.com, ian.campbell@citrix.com,
stefano.stabellini@eu.citrix.com, andrew.cooper3@citrix.com,
ian.jackson@eu.citrix.com, xen-devel@lists.xen.org,
yang.z.zhang@intel.com
Subject: Re: [v5][PATCH 09/10] tools:firmware:hvmloader: check to reserve RMRR mappings in e820
Date: Thu, 04 Sep 2014 11:04:57 +0800 [thread overview]
Message-ID: <5407D6D9.40007@intel.com> (raw)
In-Reply-To: <5405A029020000780002FBE2@mail.emea.novell.com>
On 2014/9/2 16:47, Jan Beulich wrote:
>>>> On 26.08.14 at 13:02, <tiejun.chen@intel.com> wrote:
>> +static unsigned int construct_rmrr_e820_maps(unsigned int nr,
s/construct_rmrr_e820_maps/construct_drm_e820_maps
>> + uint32_t nr_map,
>> + struct reserved_device_memory *map,
>> + struct e820entry *e820)
>> +{
>> + unsigned int i = 0, j = 0, m = 0, sum_nr = 0;
>> + uint64_t start, end, rmrr_start, rmrr_end;
I rename this pair of variables in this whole function,
s/rmrr_start/drm_start
s/rmrr_end/drm_end
>> + unsigned int insert = 0, do_insert = 0;
>> +
>> +do_real_construct:
>
> Please indent labels by at least one space (and there are further
> coding style issues to address elsewhere).
Is this necessary?
static struct tmem_object_root * obj_find(struct tmem_pool *pool, struct
oid *oidp)
{
struct rb_node *node;
struct tmem_object_root *obj;
restart_find:
read_lock(&pool->pool_rwlock);
...
read_unlock(&pool->pool_rwlock);
goto restart_find;
>
>> + sum_nr = nr + nr_map;
>
> Afaict neither nr nor nr_map change before you get here the second
> (and last) time. Hence the calculation can be moved up (ideally into
> the initializer of the variable).
+ unsigned int i = 0, j = 0, m = 0, sum_nr = nr + nr_map;
>
>> + for ( i = 0; i < nr_map; i++ )
>> + {
>> + rmrr_start = map[i].start_pfn << PAGE_SHIFT;
>> + rmrr_end = rmrr_start + map[i].nr_pages * PAGE_SIZE;
>> +
>> + for ( j = 0; j < nr; j++ )
>> + {
>> + end = e820[j].addr + e820[j].size;
>> + start = e820[j+1].addr;
>
> This is not valid when j == nr - 1 (last iteration).
>
- for ( j = 0; j < nr; j++ )
+ for ( j = 0; j < nr - 1; j++ )
>> +
>> + /* Between those existing e820 entries. */
>> + if ( (rmrr_start > end) && (rmrr_end < start) )
>> + {
>> + if (do_insert)
- if (do_insert)
+ if ( do_insert )
>> + {
>> + /* Move to free this entry. */
>> + for ( m = sum_nr - 1; m > j; m-- )
>> + {
>> + e820[m].addr = e820[m-1].addr;
>> + e820[m].size = e820[m-1].size;
>> + e820[m].type = e820[m-1].type;
>> + }
>> +
>> + /* Then fill RMRR into that entry. */
>> + e820[j+1].addr = rmrr_start;
>> + e820[j+1].size = rmrr_end - rmrr_start;
>> + e820[j+1].type = E820_RESERVED;
>> + nr++;
>> + }
>> + insert++;
>> + }
>> + /* Already at the end. */
>> + else if ( (rmrr_start > end) && !start )
>> + {
>> + if ( do_insert )
>> + {
>> + e820[nr].addr = rmrr_start;
>> + e820[nr].size = rmrr_end - rmrr_start;
>> + e820[nr].type = E820_RESERVED;
>> + nr++;
>> + }
>> + insert++;
>> + }
>> + }
>> + }
>> +
>> + /* Just return if done. */
>> + if ( do_insert )
>> + return nr;
>> +
>> + /* Fine to construct RMRR mappings into e820. */
>> + if ( insert == nr_map)
- if ( insert == nr_map)
+ if ( insert == nr_map )
>> + {
>> + do_insert = 1;
>> + goto do_real_construct;
>> + }
>> + /* Overlap. */
>> + else
>> + {
>> + printf("RMRR overlap with those existing e820 entries!\n");
>> + printf("So we don't construct RMRR mapping in e820!\n");
s/RMRR/DRM
Thanks
Tiejun
>> + }
>> +
>> + return nr;
>> +}
>> /* Create an E820 table based on memory parameters provided in hvm_info. */
>> int build_e820_table(struct e820entry *e820,
>> unsigned int lowmem_reserved_base,
>> unsigned int bios_image_base)
>> {
>> unsigned int nr = 0;
>> + struct reserved_device_memory *map = NULL;
>> + int rc;
>>
>> if ( !lowmem_reserved_base )
>> lowmem_reserved_base = 0xA0000;
>> @@ -169,6 +247,29 @@ int build_e820_table(struct e820entry *e820,
>> nr++;
>> }
>>
>> + /* We'd better reserve RMRR mapping for each VM to avoid potential
>> + * memory conflict.
>> + */
>> + if ( hvm_info->nr_reserved_device_memory_map )
>> + {
>> + if ( !map )
>> + map = mem_alloc(hvm_info->nr_reserved_device_memory_map *
>> sizeof(struct reserved_device_memory), 0);
>> + if ( map )
>> + {
>> + rc = get_reserved_device_memory_map(map,
>> +
>> hvm_info->nr_reserved_device_memory_map);
>> + if ( !rc )
>> + {
>> + nr = construct_rmrr_e820_maps(nr,
>> +
>> hvm_info->nr_reserved_device_memory_map,
>> + map,
>> + e820);
>> + }
>> + }
>> + else
>> + printf("No space to get reserved_device_memory_map!\n");
>> + }
>> +
>> return nr;
>> }
>>
>> --
>> 1.9.1
>
>
>
>
next prev parent reply other threads:[~2014-09-04 3:04 UTC|newest]
Thread overview: 53+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-08-26 11:02 [v5][PATCH 0/10] xen: reserve RMRR to avoid conflicting MMIO/RAM Tiejun Chen
2014-08-26 11:02 ` [v5][PATCH 01/10] xen:vtd:rmrr: export acpi_rmrr_units Tiejun Chen
2014-08-26 11:02 ` [v5][PATCH 02/10] xen:vtd:rmrr: introduce acpi_rmrr_unit_entries Tiejun Chen
2014-08-26 11:02 ` [v5][PATCH 03/10] xen:x86: define a new hypercall to get RMRR mappings Tiejun Chen
2014-08-26 12:02 ` Andrew Cooper
2014-08-26 12:37 ` Jan Beulich
2014-08-27 1:37 ` Chen, Tiejun
2014-08-27 6:51 ` Jan Beulich
2014-08-27 7:21 ` Chen, Tiejun
2014-08-28 2:24 ` Chen, Tiejun
2014-08-28 6:50 ` Jan Beulich
2014-08-28 7:09 ` Chen, Tiejun
2014-08-28 7:19 ` Chen, Tiejun
2014-08-28 7:29 ` Chen, Tiejun
2014-08-28 7:44 ` Jan Beulich
2014-08-29 3:02 ` Chen, Tiejun
2014-08-29 9:18 ` Jan Beulich
2014-09-01 9:44 ` Chen, Tiejun
2014-09-01 10:29 ` Jan Beulich
2014-09-02 9:59 ` Chen, Tiejun
2014-09-02 10:15 ` Jan Beulich
2014-09-02 11:10 ` Chen, Tiejun
2014-09-02 13:15 ` Jan Beulich
2014-09-03 1:45 ` Chen, Tiejun
2014-09-03 8:31 ` Chen, Tiejun
2014-09-03 8:41 ` Jan Beulich
2014-09-03 8:59 ` Chen, Tiejun
2014-09-03 9:01 ` Chen, Tiejun
2014-09-03 9:54 ` Chen, Tiejun
2014-09-03 12:54 ` Jan Beulich
2014-09-04 1:15 ` Chen, Tiejun
2014-09-03 8:35 ` Jan Beulich
2014-08-27 1:15 ` Chen, Tiejun
2014-09-02 8:25 ` Jan Beulich
2014-08-26 11:02 ` [v5][PATCH 04/10] tools:libxc: introduce hypercall for xc_reserved_device_memory_map Tiejun Chen
2014-08-26 11:02 ` [v5][PATCH 05/10] tools:libxc: check if mmio BAR is out of RMRR mappings Tiejun Chen
2014-08-26 11:02 ` [v5][PATCH 06/10] hvm_info_table: introduce nr_reserved_device_memory_map Tiejun Chen
2014-09-02 8:34 ` Jan Beulich
2014-09-04 2:07 ` Chen, Tiejun
2014-09-04 6:32 ` Jan Beulich
2014-09-04 6:55 ` Chen, Tiejun
[not found] ` <54082E3B0200007800030BCB@mail.emea.novell.com>
2014-09-09 6:40 ` Chen, Tiejun
2014-08-26 11:02 ` [v5][PATCH 07/10] xen:x86:: support xc_reserved_device_memory_map in compat case Tiejun Chen
2014-09-02 8:35 ` Jan Beulich
2014-09-04 2:13 ` Chen, Tiejun
2014-08-26 11:02 ` [v5][PATCH 08/10] tools:firmware:hvmloader: introduce hypercall for xc_reserved_device_memory_map Tiejun Chen
2014-09-02 8:37 ` Jan Beulich
2014-08-26 11:02 ` [v5][PATCH 09/10] tools:firmware:hvmloader: check to reserve RMRR mappings in e820 Tiejun Chen
2014-09-02 8:47 ` Jan Beulich
2014-09-04 3:04 ` Chen, Tiejun [this message]
2014-09-04 4:32 ` Chen, Tiejun
2014-09-04 6:36 ` Jan Beulich
2014-08-26 11:03 ` [v5][PATCH 10/10] xen:vtd: make USB RMRR mapping safe Tiejun Chen
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=5407D6D9.40007@intel.com \
--to=tiejun.chen@intel.com \
--cc=JBeulich@suse.com \
--cc=andrew.cooper3@citrix.com \
--cc=ian.campbell@citrix.com \
--cc=ian.jackson@eu.citrix.com \
--cc=kevin.tian@intel.com \
--cc=stefano.stabellini@eu.citrix.com \
--cc=xen-devel@lists.xen.org \
--cc=yang.z.zhang@intel.com \
/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).