qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Paolo Bonzini <pbonzini@redhat.com>
To: "Liu, Jingqi" <jingqi.liu@intel.com>,
	Richard Henderson <rth@twiddle.net>
Cc: "Williams, Dan J" <dan.j.williams@intel.com>,
	"qemu-devel@nongnu.org" <qemu-devel@nongnu.org>
Subject: Re: [PATCH] exec: fetch the alignment of Linux devdax pmem character device nodes
Date: Tue, 7 Apr 2020 10:08:29 +0200	[thread overview]
Message-ID: <64ba0e18-2fa3-f5bd-7711-1d91f1dc74da@redhat.com> (raw)
In-Reply-To: <2c45ec7c-f3ea-2962-47a6-9330f0a2f146@intel.com>

On 07/04/20 09:29, Liu, Jingqi wrote:
> Ping.
> 
> Any comments are appreciated.
> 
> Hi Paolo, Richard,
> 
> Any comments about this ?

I was hoping to get a review from someone else because I have no way to
test it.  But I've now queued the patch, thanks.

Paolo

> 
> Thanks,
> 
> Jingqi
> 
> On 4/1/2020 11:13 AM, Liu, Jingqi wrote:
>> If the backend file is devdax pmem character device, the alignment
>> specified by the option 'align=NUM' in the '-object memory-backend-file'
>> needs to match the alignment requirement of the devdax pmem character
>> device.
>>
>> This patch fetches the devdax pmem file 'align', so that we can compare
>> it with the NUM of 'align=NUM'.
>> The NUM needs to be larger than or equal to the devdax pmem file 'align'.
>>
>> It also fixes the problem that mmap() returns failure in qemu_ram_mmap()
>> when the NUM of 'align=NUM' is less than the devdax pmem file 'align'.
>>
>> Cc: Dan Williams <dan.j.williams@intel.com>
>> Signed-off-by: Jingqi Liu <jingqi.liu@intel.com>
>> ---
>>   exec.c | 46 +++++++++++++++++++++++++++++++++++++++++++++-
>>   1 file changed, 45 insertions(+), 1 deletion(-)
>>
>> diff --git a/exec.c b/exec.c
>> index de9d949902..8221abffec 100644
>> --- a/exec.c
>> +++ b/exec.c
>> @@ -1736,6 +1736,42 @@ static int64_t get_file_size(int fd)
>>       return size;
>>   }
>>   +static int64_t get_file_align(int fd)
>> +{
>> +    int64_t align = -1;
>> +#if defined(__linux__)
>> +    struct stat st;
>> +
>> +    if (fstat(fd, &st) < 0) {
>> +        return -errno;
>> +    }
>> +
>> +    /* Special handling for devdax character devices */
>> +    if (S_ISCHR(st.st_mode)) {
>> +        g_autofree char *subsystem_path = NULL;
>> +        g_autofree char *subsystem = NULL;
>> +
>> +        subsystem_path =
>> g_strdup_printf("/sys/dev/char/%d:%d/subsystem",
>> +                                         major(st.st_rdev),
>> minor(st.st_rdev));
>> +        subsystem = g_file_read_link(subsystem_path, NULL);
>> +
>> +        if (subsystem && g_str_has_suffix(subsystem, "/dax")) {
>> +            g_autofree char *align_path = NULL;
>> +            g_autofree char *align_str = NULL;
>> +
>> +            align_path =
>> g_strdup_printf("/sys/dev/char/%d:%d/device/align",
>> +                                    major(st.st_rdev),
>> minor(st.st_rdev));
>> +
>> +            if (g_file_get_contents(align_path, &align_str, NULL,
>> NULL)) {
>> +                return g_ascii_strtoll(align_str, NULL, 0);
>> +            }
>> +        }
>> +    }
>> +#endif /* defined(__linux__) */
>> +
>> +    return align;
>> +}
>> +
>>   static int file_ram_open(const char *path,
>>                            const char *region_name,
>>                            bool *created,
>> @@ -2275,7 +2311,7 @@ RAMBlock *qemu_ram_alloc_from_fd(ram_addr_t
>> size, MemoryRegion *mr,
>>   {
>>       RAMBlock *new_block;
>>       Error *local_err = NULL;
>> -    int64_t file_size;
>> +    int64_t file_size, file_align;
>>         /* Just support these ram flags by now. */
>>       assert((ram_flags & ~(RAM_SHARED | RAM_PMEM)) == 0);
>> @@ -2311,6 +2347,14 @@ RAMBlock *qemu_ram_alloc_from_fd(ram_addr_t
>> size, MemoryRegion *mr,
>>           return NULL;
>>       }
>>   +    file_align = get_file_align(fd);
>> +    if (file_align > 0 && mr && file_align > mr->align) {
>> +        error_setg(errp, "backing store align 0x%" PRIx64
>> +                   " is larger than 'align' option 0x" RAM_ADDR_FMT,
>> +                   file_align, mr->align);
>> +        return NULL;
>> +    }
>> +
>>       new_block = g_malloc0(sizeof(*new_block));
>>       new_block->mr = mr;
>>       new_block->used_length = size;
> 



  reply	other threads:[~2020-04-07  8:09 UTC|newest]

Thread overview: 20+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-04-01  3:13 [PATCH] exec: fetch the alignment of Linux devdax pmem character device nodes Jingqi Liu
2020-04-07  7:29 ` Liu, Jingqi
2020-04-07  8:08   ` Paolo Bonzini [this message]
2020-04-07  8:16     ` Dan Williams
2020-04-07 11:42       ` Joao Martins
2020-04-07  8:39     ` Liu, Jingqi
2020-04-07 10:59 ` Joao Martins
2020-04-07 14:31   ` Paolo Bonzini
2020-04-07 15:51     ` Joao Martins
2020-04-08  1:16       ` Liu, Jingqi
2020-04-08  9:28         ` Joao Martins
2020-04-07 16:55   ` Dan Williams
2020-04-07 18:28     ` Joao Martins
2020-04-07 18:29       ` Paolo Bonzini
2020-04-08  2:25       ` Liu, Jingqi
2020-04-08  9:42         ` Joao Martins
2020-04-09 14:33           ` Liu, Jingqi
2020-04-09 16:46             ` Dan Williams
2020-04-09 17:02               ` Paolo Bonzini
2020-04-10  1:48               ` Liu, Jingqi

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=64ba0e18-2fa3-f5bd-7711-1d91f1dc74da@redhat.com \
    --to=pbonzini@redhat.com \
    --cc=dan.j.williams@intel.com \
    --cc=jingqi.liu@intel.com \
    --cc=qemu-devel@nongnu.org \
    --cc=rth@twiddle.net \
    /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).