qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Christian Schoenebeck <qemu_oss@crudebyte.com>
To: qemu-devel@nongnu.org
Cc: Will Cohen <wwcohen@gmail.com>,
	Keno Fischer <keno@juliacomputing.com>,
	Greg Kurz <groug@kaod.org>,
	hi@alyssa.is, Michael Roitzsch <reactorcontrol@icloud.com>
Subject: Re: [PATCH v2 04/11] 9p: darwin: Handle struct dirent differences
Date: Wed, 24 Nov 2021 15:58:38 +0100	[thread overview]
Message-ID: <60659730.p4icJFLGV9@silver> (raw)
In-Reply-To: <20211122004913.20052-5-wwcohen@gmail.com>

On Montag, 22. November 2021 01:49:06 CET Will Cohen wrote:
> From: Keno Fischer <keno@juliacomputing.com>
> 
> On darwin d_seekoff exists, but is optional and does not seem to
> be commonly used by file systems. Use `telldir` instead to obtain
> the seek offset.

Are you sure d_seekoff doesn't work on macOS? Because using telldir() instead
is not the same thing. Accessing d_*off is just POD access, whereas telldir()
is a syscall. What you are trying in this patch with telldir() easily gets
hairy.

AFAIK there was d_off in previous versions of macOS, which was then replaced
by d_seekof in macOS 11.1, no?

> Signed-off-by: Keno Fischer <keno@juliacomputing.com>
> [Michael Roitzsch: - Rebase for NixOS]
> Signed-off-by: Michael Roitzsch <reactorcontrol@icloud.com>
> Signed-off-by: Will Cohen <wwcohen@gmail.com>
> ---
>  hw/9pfs/9p-synth.c |  2 ++
>  hw/9pfs/9p.c       | 33 +++++++++++++++++++++++++++++++--
>  hw/9pfs/codir.c    |  4 ++++
>  3 files changed, 37 insertions(+), 2 deletions(-)
> 
> diff --git a/hw/9pfs/9p-synth.c b/hw/9pfs/9p-synth.c
> index 4a4a776d06..09b9c25288 100644
> --- a/hw/9pfs/9p-synth.c
> +++ b/hw/9pfs/9p-synth.c
> @@ -222,7 +222,9 @@ static void synth_direntry(V9fsSynthNode *node,
>  {
>      strcpy(entry->d_name, node->name);
>      entry->d_ino = node->attr->inode;
> +#ifndef CONFIG_DARWIN
>      entry->d_off = off + 1;
> +#endif
>  }

^ That doesn't look like it would work. Compiling sure.

Have you tried running the test cases?
https://wiki.qemu.org/Documentation/9p#Test_Cases

>  static struct dirent *synth_get_dentry(V9fsSynthNode *dir,
> diff --git a/hw/9pfs/9p.c b/hw/9pfs/9p.c
> index f4f0c200c7..c06e8a85a0 100644
> --- a/hw/9pfs/9p.c
> +++ b/hw/9pfs/9p.c
> @@ -2218,6 +2218,25 @@ static int v9fs_xattr_read(V9fsState *s, V9fsPDU
> *pdu, V9fsFidState *fidp, return offset;
>  }
> 
> +/**
> + * Get the seek offset of a dirent. If not available from the structure
> itself, + * obtain it by calling telldir.
> + */
> +static int v9fs_dent_telldir(V9fsPDU *pdu, V9fsFidState *fidp,
> +                             struct dirent *dent)
> +{
> +#ifdef CONFIG_DARWIN
> +    /*
> +     * Darwin has d_seekoff, which appears to function similarly to d_off.
> +     * However, it does not appear to be supported on all file systems,
> +     * so use telldir for correctness.
> +     */
> +    return telldir(fidp->fs.dir.stream);
> +#else
> +    return dent->d_off;
> +#endif

The thing here is, we usually run fs syscalls as coroutines on a worker thread
as they might block for a long time, and in the meantime 9p server's main
thread could handle other tasks. Plus if a fs syscall gets stuck, we can abort
the request, which is not possible if its called directly from main thread.

https://wiki.qemu.org/Documentation/9p#Threads_and_Coroutines

dent->d_off is just POD access, so it is instantanious. But that does not mean
you should wrap that telldir() call now to be a background task, because that
will add other implications. I would rather prefer to clarify first whether
d_*off is really not working on macOS to avoid all the foreseeable trouble.

> +}
> +
>  static int coroutine_fn v9fs_do_readdir_with_stat(V9fsPDU *pdu,
>                                                    V9fsFidState *fidp,
>                                                    uint32_t max_count)
> @@ -2281,7 +2300,11 @@ static int coroutine_fn
> v9fs_do_readdir_with_stat(V9fsPDU *pdu, count += len;
>          v9fs_stat_free(&v9stat);
>          v9fs_path_free(&path);
> -        saved_dir_pos = dent->d_off;
> +        saved_dir_pos = v9fs_dent_telldir(pdu, fidp, dent);
> +        if (saved_dir_pos < 0) {
> +            err = saved_dir_pos;
> +            break;
> +        }
>      }
> 
>      v9fs_readdir_unlock(&fidp->fs.dir);
> @@ -2420,6 +2443,7 @@ static int coroutine_fn v9fs_do_readdir(V9fsPDU *pdu,
> V9fsFidState *fidp, V9fsString name;
>      int len, err = 0;
>      int32_t count = 0;
> +    off_t off;
>      struct dirent *dent;
>      struct stat *st;
>      struct V9fsDirEnt *entries = NULL;
> @@ -2480,12 +2504,17 @@ static int coroutine_fn v9fs_do_readdir(V9fsPDU
> *pdu, V9fsFidState *fidp, qid.version = 0;
>          }
> 
> +        off = v9fs_dent_telldir(pdu, fidp, dent);
> +        if (off < 0) {
> +            err = off;
> +            break;
> +        }
>          v9fs_string_init(&name);
>          v9fs_string_sprintf(&name, "%s", dent->d_name);
> 
>          /* 11 = 7 + 4 (7 = start offset, 4 = space for storing count) */
>          len = pdu_marshal(pdu, 11 + count, "Qqbs",
> -                          &qid, dent->d_off,
> +                          &qid, off,
>                            dent->d_type, &name);
> 
>          v9fs_string_free(&name);
> diff --git a/hw/9pfs/codir.c b/hw/9pfs/codir.c
> index 032cce04c4..78aca1d98b 100644
> --- a/hw/9pfs/codir.c
> +++ b/hw/9pfs/codir.c
> @@ -167,7 +167,11 @@ static int do_readdir_many(V9fsPDU *pdu, V9fsFidState
> *fidp, }
> 
>          size += len;
> +#ifdef CONFIG_DARWIN
> +        saved_dir_pos = telldir(fidp->fs.dir.stream);
> +#else
>          saved_dir_pos = dent->d_off;
> +#endif
>      }
> 
>      /* restore (last) saved position */




  reply	other threads:[~2021-11-24 14:59 UTC|newest]

Thread overview: 31+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-11-22  0:49 [PATCH v2 00/11] 9p: Add support for darwin Will Cohen
2021-11-22  0:49 ` [PATCH v2 01/11] 9p: linux: Fix a couple Linux assumptions Will Cohen
2021-11-24 12:41   ` Christian Schoenebeck
2021-11-22  0:49 ` [PATCH v2 02/11] 9p: Rename 9p-util -> 9p-util-linux Will Cohen
2021-11-22  0:49 ` [PATCH v2 03/11] 9p: darwin: Handle struct stat(fs) differences Will Cohen
2021-11-24 14:23   ` Christian Schoenebeck
2021-12-01 22:46     ` Will Cohen
2021-12-02 15:35       ` Christian Schoenebeck
2021-11-22  0:49 ` [PATCH v2 04/11] 9p: darwin: Handle struct dirent differences Will Cohen
2021-11-24 14:58   ` Christian Schoenebeck [this message]
2021-11-24 15:45     ` Michael Roitzsch
2021-11-24 19:09       ` Christian Schoenebeck
2022-01-27 21:48         ` Will Cohen
2022-01-28 15:48           ` Christian Schoenebeck
2021-11-22  0:49 ` [PATCH v2 05/11] 9p: darwin: Ignore O_{NOATIME, DIRECT} Will Cohen
2021-11-22  0:49 ` [PATCH v2 06/11] 9p: darwin: Compatibility defn for XATTR_SIZE_MAX Will Cohen
2021-11-24 15:44   ` Christian Schoenebeck
2021-11-22  0:49 ` [PATCH v2 07/11] 9p: darwin: *xattr_nofollow implementations Will Cohen
2021-11-22  0:49 ` [PATCH v2 08/11] 9p: darwin: Compatibility for f/l*xattr Will Cohen
2021-11-24 16:20   ` Christian Schoenebeck
2022-01-27 21:47     ` Will Cohen
2021-11-22  0:49 ` [PATCH v2 09/11] 9p: darwin: Provide fallback impl for utimensat Will Cohen
2021-11-24 17:07   ` Christian Schoenebeck
2021-11-22  0:49 ` [PATCH v2 10/11] 9p: darwin: Implement compatibility for mknodat Will Cohen
2021-11-24 17:20   ` Christian Schoenebeck
2022-01-27 21:47     ` Will Cohen
2022-01-28 15:15       ` Christian Schoenebeck
2022-01-28 18:28         ` Will Cohen
2022-01-31 22:26           ` Will Cohen
2022-02-01 12:44             ` Christian Schoenebeck
2021-11-22  0:49 ` [PATCH v2 11/11] 9p: darwin: meson: Allow VirtFS on Darwin Will Cohen

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=60659730.p4icJFLGV9@silver \
    --to=qemu_oss@crudebyte.com \
    --cc=groug@kaod.org \
    --cc=hi@alyssa.is \
    --cc=keno@juliacomputing.com \
    --cc=qemu-devel@nongnu.org \
    --cc=reactorcontrol@icloud.com \
    --cc=wwcohen@gmail.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).