From: Juergen Gross <jgross@suse.com>
To: Daniel Kiper <daniel.kiper@oracle.com>
Cc: grub-devel@gnu.org, phcoder@gmail.com, mchang@suse.com,
xen-devel@lists.xen.org
Subject: Re: [PATCH v5 09/11] xen: add capability to load initrd outside of initial mapping
Date: Thu, 25 Feb 2016 16:33:46 +0100 [thread overview]
Message-ID: <56CF1EDA.1000304@suse.com> (raw)
In-Reply-To: <20160225124757.GB3482@olila.local.net-space.pl>
On 25/02/16 13:47, Daniel Kiper wrote:
> On Thu, Feb 25, 2016 at 12:33:35PM +0100, Juergen Gross wrote:
>> Modern pvops linux kernels support an initrd not covered by the initial
>> mapping. This capability is flagged by an elf-note.
>>
>> In case the elf-note is set by the kernel don't place the initrd into
>> the initial mapping. This will allow to load larger initrds and/or
>> support domains with larger memory, as the initial mapping is limited
>> to 2GB and it is containing the p2m list.
>>
>> Signed-off-by: Juergen Gross <jgross@suse.com>
>> ---
>> V5: let call grub_xen_alloc_final() all subfunctions unconditionally
>> and let them decide whether they need to do anything
>> V4: rename grub_xen_alloc_end() to grub_xen_alloc_final()
>> ---
>> grub-core/loader/i386/xen.c | 65 ++++++++++++++++++++++++++++++--------
>> grub-core/loader/i386/xen_fileXX.c | 3 ++
>> include/grub/xen_file.h | 1 +
>> 3 files changed, 56 insertions(+), 13 deletions(-)
>>
>> diff --git a/grub-core/loader/i386/xen.c b/grub-core/loader/i386/xen.c
>> index 2e12763..466f0c0 100644
>> --- a/grub-core/loader/i386/xen.c
>> +++ b/grub-core/loader/i386/xen.c
>> @@ -228,6 +228,9 @@ grub_xen_p2m_alloc (void)
>> grub_size_t p2msize;
>> grub_err_t err;
>>
>> + if (xen_state.virt_mfn_list)
>> + return GRUB_ERR_NONE;
>> +
>> xen_state.state.mfn_list = xen_state.max_addr;
>> xen_state.next_start.mfn_list =
>> xen_state.max_addr + xen_state.xen_inf.virt_base;
>> @@ -250,6 +253,9 @@ grub_xen_special_alloc (void)
>> grub_relocator_chunk_t ch;
>> grub_err_t err;
>>
>> + if (xen_state.virt_start_info)
>> + return GRUB_ERR_NONE;
>> +
>> err = grub_relocator_alloc_chunk_addr (xen_state.relocator, &ch,
>> xen_state.max_addr,
>> sizeof (xen_state.next_start));
>> @@ -281,6 +287,9 @@ grub_xen_pt_alloc (void)
>> grub_uint64_t nr_info_pages;
>> grub_uint64_t nr_pages, nr_pt_pages, nr_need_pages;
>>
>> + if (xen_state.virt_pgtable)
>> + return GRUB_ERR_NONE;
>> +
>> xen_state.next_start.pt_base =
>> xen_state.max_addr + xen_state.xen_inf.virt_base;
>> xen_state.state.paging_start = xen_state.max_addr >> PAGE_SHIFT;
>> @@ -320,6 +329,24 @@ grub_xen_pt_alloc (void)
>> }
>>
>> static grub_err_t
>> +grub_xen_alloc_final (void)
>> +{
>> + grub_err_t err;
>> +
>> + err = grub_xen_p2m_alloc ();
>> + if (err)
>> + return err;
>> + err = grub_xen_special_alloc ();
>> + if (err)
>> + return err;
>> + err = grub_xen_pt_alloc ();
>> + if (err)
>> + return err;
>
> Could you move grub_xen_p2m_alloc() here? This way grub_xen_alloc_final()
> will be real final in patch 11 and you do not need an extra condition
> around grub_xen_p2m_alloc().
No. Page tables must be allocated after p2m unless p2m is outside of
initial kernel mapping (patch 11).
>
>> +
>> + return GRUB_ERR_NONE;
>> +}
>> +
>> +static grub_err_t
>> grub_xen_boot (void)
>> {
>> grub_err_t err;
>> @@ -330,13 +357,7 @@ grub_xen_boot (void)
>> if (grub_xen_n_allocated_shared_pages)
>> return grub_error (GRUB_ERR_BUG, "active grants");
>>
>> - err = grub_xen_p2m_alloc ();
>> - if (err)
>> - return err;
>> - err = grub_xen_special_alloc ();
>> - if (err)
>> - return err;
>> - err = grub_xen_pt_alloc ();
>> + err = grub_xen_alloc_final ();
>
> if (!xen_state.xen_inf.unmapped_initrd)
> {
> err = grub_xen_alloc_final ();
> if (err)
> goto fail;
> }
>
> This way you avoid checks in grub_xen_p2m_alloc(),
> grub_xen_special_alloc() and grub_xen_pt_alloc().
No. In case there is no initrd, but the kernel does support an unmapped
initrd, I would omit allocating special pages, p2m and page tables.
>
>> if (err)
>> return err;
>>
>> @@ -610,6 +631,13 @@ grub_cmd_initrd (grub_command_t cmd __attribute__ ((unused)),
>> goto fail;
>> }
>>
>> + if (xen_state.xen_inf.unmapped_initrd)
>> + {
>> + err = grub_xen_alloc_final ();
>> + if (err)
>> + goto fail;
>> + }
>> +
>> if (grub_initrd_init (argc, argv, &initrd_ctx))
>> goto fail;
>>
>> @@ -627,14 +655,24 @@ grub_cmd_initrd (grub_command_t cmd __attribute__ ((unused)),
>> goto fail;
>> }
>>
>> - xen_state.next_start.mod_start =
>> - xen_state.max_addr + xen_state.xen_inf.virt_base;
>> - xen_state.next_start.mod_len = size;
>
> Leave "xen_state.next_start.mod_len = size;" as is here and...
>
>> -
>> - xen_state.max_addr = ALIGN_UP (xen_state.max_addr + size, PAGE_SIZE);
>> + if (xen_state.xen_inf.unmapped_initrd)
>> + {
>> + xen_state.next_start.flags |= SIF_MOD_START_PFN;
>> + xen_state.next_start.mod_start = xen_state.max_addr >> PAGE_SHIFT;
>> + xen_state.next_start.mod_len = size;
>
> ... do not add it here...
>
>> + }
>> + else
>> + {
>> + xen_state.next_start.mod_start =
>> + xen_state.max_addr + xen_state.xen_inf.virt_base;
>> + xen_state.next_start.mod_len = size;
>
> ... and here...
>
>> + }
>
> ... and you do not need curly brackets then.
I still do, as in the "if" part I still have two statements.
>
>> grub_dprintf ("xen", "Initrd, addr=0x%x, size=0x%x\n",
>> - (unsigned) xen_state.next_start.mod_start, (unsigned) size);
>> + (unsigned) (xen_state.max_addr + xen_state.xen_inf.virt_base),
>> + (unsigned) size);
>
> This should be printed only if !xen_state.xen_inf.unmapped_initrd.
Really? Why? We still have an initrd.
> Why do not use xen_state.next_start.mod_start here?
It is the pfn in case of unmapped_initrd.
Juergen
next prev parent reply other threads:[~2016-02-25 15:33 UTC|newest]
Thread overview: 23+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-02-25 11:33 [PATCH v5 00/11] grub-xen: support booting huge pv-domains Juergen Gross
2016-02-25 11:33 ` [PATCH v5 01/11] xen: make xen loader callable multiple times Juergen Gross
2016-02-25 11:33 ` [PATCH v5 02/11] xen: avoid memleaks on error Juergen Gross
2016-02-25 12:14 ` Daniel Kiper
2016-02-25 11:33 ` [PATCH v5 03/11] xen: reduce number of global variables in xen loader Juergen Gross
2016-02-25 11:33 ` [PATCH v5 04/11] xen: add elfnote.h to avoid using numbers instead of constants Juergen Gross
2016-02-25 11:33 ` [PATCH v5 05/11] xen: synchronize xen header Juergen Gross
2016-02-25 11:33 ` [PATCH v5 06/11] xen: factor out p2m list allocation into separate function Juergen Gross
2016-02-25 11:33 ` [PATCH v5 07/11] xen: factor out allocation of special pages " Juergen Gross
2016-02-25 11:33 ` [PATCH v5 08/11] xen: factor out allocation of page tables " Juergen Gross
2016-02-25 11:33 ` [PATCH v5 09/11] xen: add capability to load initrd outside of initial mapping Juergen Gross
2016-02-25 12:47 ` Daniel Kiper
2016-02-25 15:33 ` Juergen Gross [this message]
2016-02-26 14:00 ` Daniel Kiper
2016-02-26 14:28 ` Juergen Gross
2016-02-26 15:41 ` Daniel Kiper
2016-02-29 8:27 ` Juergen Gross
2016-02-29 15:43 ` Daniel Kiper
2016-02-29 15:49 ` Juergen Gross
2016-02-29 16:22 ` Daniel Kiper
2016-02-25 11:33 ` [PATCH v5 10/11] xen: modify page table construction Juergen Gross
2016-02-25 14:45 ` Daniel Kiper
2016-02-25 11:33 ` [PATCH v5 11/11] xen: add capability to load p2m list outside of kernel mapping Juergen Gross
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=56CF1EDA.1000304@suse.com \
--to=jgross@suse.com \
--cc=daniel.kiper@oracle.com \
--cc=grub-devel@gnu.org \
--cc=mchang@suse.com \
--cc=phcoder@gmail.com \
--cc=xen-devel@lists.xen.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).