From: NeilBrown <neilb@suse.com>
To: Zhilong Liu <zlliu@suse.com>,
"jes.sorensen@gmail.com" <jes.sorensen@gmail.com>
Cc: Linux Raid <linux-raid@vger.kernel.org>
Subject: Re: [mdadm PATCH] mdopen: call "modprobe md_mod" if it might be needed.
Date: Thu, 12 Oct 2017 19:48:26 +1100 [thread overview]
Message-ID: <87zi8wn44l.fsf@notabene.neil.brown.name> (raw)
In-Reply-To: <7631f4ae-bcc5-b36d-bb63-678515109f05@suse.com>
[-- Attachment #1: Type: text/plain, Size: 6829 bytes --]
On Thu, Oct 12 2017, Zhilong Liu wrote:
> On 10/12/2017 08:06 AM, NeilBrown wrote:
>> On Wed, Oct 11 2017, Zhilong Liu wrote:
>>
>>> On 10/11/2017 04:16 AM, NeilBrown wrote:
>>>> On Tue, Oct 10 2017, Zhilong Liu wrote:
>>>>
>>>>> On 09/25/2017 01:52 PM, NeilBrown wrote:
>>>>>> Creating an array by opening a block-device with major number of 9
>>>>>> will transparently load the md module if needed.
>>>>>> Creating an array by opening
>>>>>> /sys/module/md_mod/parameters/new_array
>>>>>> and writing to it won't, it will just fail if md_mod isn't loaded.
>>>>>>
>>>>>> So when opening that file fails with ENOENT, run "modprobe md_mod" and
>>>>>> try again.
>>>>>>
>>>>>> This fixes a bug whereby if you have "CREATE names=yes" in mdadm.conf,
>>>>>> and the md modules isn't loaded, then creating or assembling an
>>>>>> array will not honor the "names=yes" configuration.
>>>>>>
>>>>>> Signed-off-by: NeilBrown <neilb@suse.com>
>>>>>> ---
>>>>>> mdopen.c | 4 ++++
>>>>>> 1 file changed, 4 insertions(+)
>>>>>>
>>>>>> diff --git a/mdopen.c b/mdopen.c
>>>>>> index 3c0052f2db23..dcdc6f23e6c1 100644
>>>>>> --- a/mdopen.c
>>>>>> +++ b/mdopen.c
>>>>>> @@ -312,6 +312,10 @@ int create_mddev(char *dev, char *name, int autof, int trustworthy,
>>>>>> if (block_udev)
>>>>>> udev_block(devnm);
>>>>>> fd = open("/sys/module/md_mod/parameters/new_array", O_WRONLY);
>>>>>> + if (fd < 0 && errno == ENOENT) {
>>>>>> + system("modprobe md_mod");
>>> [nip]
>>>> Hmmm.. that's annoying. I wonder why "system" is marked
>>>> "warn_unused_result".
>>> in /usr/include/stdlib.h:
>>> ...
>>> 712 /* Execute the given line as a shell command.
>>> 713
>>> 714 This function is a cancellation point and therefore not marked with
>>> 715 __THROW. */
>>> 716 extern int system (const char *__command) __wur;
>>> ...
>>>
>>> the "warn_unused_result" is from the __wur parameter, re-compile mdadm
>>> after delete the '__wur',
>>> it works.
>>>
>>>> In this case I really don't care - I'm not convinced an extra error
>>>> message will really help.
>>>> Maybe
>>>> if (system("modprobe md_mod") == 0)
>>>> fd = open("/sys/......", O_WRONLY);
>>> Agree.
>>>> We do what a better error message, then it should be based on 'fd < 0'.
>>>> e.g.
>>>> if (fd < 0 || n != strlen(devnm))
>>>> pr_err("Fail create array using /sys/module/md_mod/parameters/new_array\n");
>>> you mean something like this?
>>>
>>> diff --git a/mdopen.c b/mdopen.c
>>> index dcdc6f2..9de347e 100644
>>> --- a/mdopen.c
>>> +++ b/mdopen.c
>>> @@ -313,14 +313,17 @@ int create_mddev(char *dev, char *name, int autof,
>>> int trustworthy,
>>> udev_block(devnm);
>>> fd = open("/sys/module/md_mod/parameters/new_array",
>>> O_WRONLY);
>>> if (fd < 0 && errno == ENOENT) {
>>> - system("modprobe md_mod");
>>> - fd =
>>> open("/sys/module/md_mod/parameters/new_array", O_WRONLY);
>>> + if (system("modprobe md_mod") == 0)
>>> + fd =
>>> open("/sys/module/md_mod/parameters/new_array",
>>> + O_WRONLY);
>>> }
>>> if (fd >= 0) {
>>> n = write(fd, devnm, strlen(devnm));
>>> close(fd);
>>> }
>>> - if (n < 0) {
>>> + if (fd < 0 || n != strlen(devnm)) {
>>> + pr_err("Fail create array using "
>>> + "/sys/module/md_mod/parameters/new_array\n");
>>> devnm[0] = 0;
>>> udev_unblock();
>>> }
>>>
>> Yes - exactly like that except that I wouldn't wrap the long string.
>> Lines longer than 80 chars are good to avoid, but breaking string
>> literals is worse than having long lines. e.g. it makes searching for
>> the string hard.
>
> Thanks for your detail explanation. I draft it like this, and
> already checked it via to ./linux/scripts/checkpatch.pl.
I wouldn't worry too much about checkpatch.pl.
It is worth looking at what it reports, but if you don't agree or the
maintainer doesn't agree, then feel free to ignore it.
> Is this changing ok to you?
>
> diff --git a/mdopen.c b/mdopen.c
> index dcdc6f2..da8d9d1 100644
> --- a/mdopen.c
> +++ b/mdopen.c
> @@ -26,6 +26,8 @@
> #include "md_p.h"
> #include <ctype.h>
>
> +#define NEW_ARRAY_FILE "/sys/module/md_mod/parameters/new_array"
> +
Splitting this out into a separate string does make sense.
However I would use
static const char new_array_file[] = ....;
rather than #define.
I hadn't noticed that there are two places where we write to new_array.
Maybe that should be split out into a function.
Both should use modprobe if the open fails.
Thanks,
NeilBrown
> void make_parts(char *dev, int cnt)
> {
> /* make 'cnt' partition devices for 'dev'
> @@ -311,16 +313,17 @@ int create_mddev(char *dev, char *name, int autof,
> int trustworthy,
> sprintf(devnm, "md_%s", cname);
> if (block_udev)
> udev_block(devnm);
> - fd = open("/sys/module/md_mod/parameters/new_array", O_WRONLY);
> + fd = open(NEW_ARRAY_FILE, O_WRONLY);
> if (fd < 0 && errno == ENOENT) {
> - system("modprobe md_mod");
> - fd = open("/sys/module/md_mod/parameters/new_array", O_WRONLY);
> + if (system("modprobe md_mod") == 0)
> + fd = open(NEW_ARRAY_FILE, O_WRONLY);
> }
> if (fd >= 0) {
> n = write(fd, devnm, strlen(devnm));
> close(fd);
> }
> - if (n < 0) {
> + if (fd < 0 || n != strlen(devnm)) {
> + pr_err("Fail create array using %s\n", NEW_ARRAY_FILE);
> devnm[0] = 0;
> udev_unblock();
> }
> @@ -331,12 +334,13 @@ int create_mddev(char *dev, char *name, int autof,
> int trustworthy,
> sprintf(devnm, "md%d", num);
> if (block_udev)
> udev_block(devnm);
> - fd = open("/sys/module/md_mod/parameters/new_array", O_WRONLY);
> + fd = open(NEW_ARRAY_FILE, O_WRONLY);
> if (fd >= 0) {
> n = write(fd, devnm, strlen(devnm));
> close(fd);
> }
> - if (n < 0) {
> + if (fd < 0 || n != strlen(devnm)) {
> + pr_err("Fail create array using %s\n", NEW_ARRAY_FILE);
> devnm[0] = 0;
> udev_unblock();
> }
> --
> 2.6.6
>
>
> Thanks,
> -Zhilong
>
>> Thanks,
>> NeilBrown
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 832 bytes --]
next prev parent reply other threads:[~2017-10-12 8:48 UTC|newest]
Thread overview: 15+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-09-25 5:52 [mdadm PATCH] mdopen: call "modprobe md_mod" if it might be needed NeilBrown
2017-09-25 15:26 ` John Stoffel
2017-09-25 23:50 ` NeilBrown
2017-09-26 15:11 ` Jes Sorensen
2017-09-26 19:12 ` Wols Lists
2017-09-26 20:55 ` John Stoffel
2017-09-27 21:30 ` Jes Sorensen
[not found] ` <cba5f77f-d6de-7a6b-35b0-70b7c56eb3f7@suse.com>
2017-10-10 20:16 ` NeilBrown
2017-10-11 7:39 ` Zhilong Liu
2017-10-12 0:06 ` NeilBrown
2017-10-12 3:55 ` Zhilong Liu
2017-10-12 8:48 ` NeilBrown [this message]
2017-10-13 9:16 ` Zhilong Liu
2017-10-15 22:41 ` NeilBrown
2017-10-12 9:55 ` Wols Lists
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=87zi8wn44l.fsf@notabene.neil.brown.name \
--to=neilb@suse.com \
--cc=jes.sorensen@gmail.com \
--cc=linux-raid@vger.kernel.org \
--cc=zlliu@suse.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).