qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Warner Losh <imp@bsdimp.com>
To: Karim Taha <kariem.taha2.7@gmail.com>
Cc: qemu-devel@nongnu.org, "Stacey Son" <sson@freebsd.org>,
	"Mikaël Urankar" <mikael.urankar@gmail.com>
Subject: Re: [PATCH 17/22] Implement do_obreak function
Date: Sat, 19 Aug 2023 22:40:51 -0600	[thread overview]
Message-ID: <CANCZdfpgdCbnwgHxokDD4jVKHAEoVS+4aY80KwNZdAp7qYxMfA@mail.gmail.com> (raw)
In-Reply-To: <20230819094806.14965-18-kariem.taha2.7@gmail.com>

[-- Attachment #1: Type: text/plain, Size: 5220 bytes --]

On Sat, Aug 19, 2023 at 3:49 AM Karim Taha <kariem.taha2.7@gmail.com> wrote:

> From: Stacey Son <sson@FreeBSD.org>
>
> Co-authored-by: Mikaël Urankar <mikael.urankar@gmail.com>
>
> Signed-off-by: Mikaël Urankar <mikael.urankar@gmail.com>
> Signed-off-by: Stacey Son <sson@FreeBSD.org>
> Signed-off-by: Karim Taha <kariem.taha2.7@gmail.com>
> ---
>  bsd-user/bsd-mem.h            | 85 +++++++++++++++++++++++++++++++++++
>  bsd-user/freebsd/os-syscall.c |  7 +++
>  2 files changed, 92 insertions(+)
>

emacs works with this, iirc, and it's super picky about brk being
pedantically correct.
This also doesn't match current linux-user. Perhaps the caution it uses is
no longer
needed? I think it's another area we should invite the linux-user
maintainers to
comment.

Warner



> diff --git a/bsd-user/bsd-mem.h b/bsd-user/bsd-mem.h
> index edbccd3111..6f33148eb7 100644
> --- a/bsd-user/bsd-mem.h
> +++ b/bsd-user/bsd-mem.h
> @@ -167,4 +167,89 @@ static inline abi_long do_bsd_mincore(abi_ulong
> target_addr, abi_ulong len,
>      return ret;
>  }
>
> +#ifdef DO_DEBUG
> +#define DEBUGF_BRK(message, args...) \
> +    do { fprintf(stderr, (message), ## args); } while (0)
> +#else
> +#define DEBUGF_BRK(message, args...)
> +#endif
> +
> +/* do_brk() must return target values and target errnos. */
> +static inline abi_long do_obreak(abi_ulong new_brk)
> +{
> +    abi_long mapped_addr;
> +    int new_alloc_size;
> +
> +    DEBUGF_BRK("do_brk(" TARGET_ABI_FMT_lx ") -> ", new_brk);
> +
> +    if (!new_brk) {
> +        DEBUGF_BRK(TARGET_ABI_FMT_lx " (!new_brk)\n", bsd_target_brk);
> +        return bsd_target_brk;
> +    }
> +    if (new_brk < bsd_target_original_brk) {
> +        DEBUGF_BRK(TARGET_ABI_FMT_lx " (new_brk <
> bsd_target_original_brk)\n",
> +                   bsd_target_brk);
> +        return bsd_target_brk;
> +    }
> +
> +    /*
> +     * If the new brk is less than the highest page reserved to the
> target heap
> +     * allocation, set it and we're almost done...
> +     */
> +    if (new_brk <= brk_page) {
> +        /*
> +         * Heap contents are initialized to zero, as for anonymous mapped
> pages.
> +         */
> +        if (new_brk > bsd_target_brk) {
> +            memset(g2h_untagged(bsd_target_brk), 0, new_brk -
> bsd_target_brk);
> +        }
> +        bsd_target_brk = new_brk;
> +        DEBUGF_BRK(TARGET_ABI_FMT_lx " (new_brk <= brk_page)\n",
> +                   bsd_target_brk);
> +        return bsd_target_brk;
> +    }
> +
> +    /*
> +     * We need to allocate more memory after the brk... Note that we
> don't use
> +     * MAP_FIXED because that will map over the top of any existing
> mapping
> +     * (like the one with the host libc or qemu itself); instead we treat
> +     * "mapped but at wrong address" as a failure and unmap again.
> +     */
> +    new_alloc_size = HOST_PAGE_ALIGN(new_brk - brk_page);
> +    mapped_addr = get_errno(target_mmap(brk_page, new_alloc_size,
> +                                        PROT_READ | PROT_WRITE,
> +                                        MAP_ANON | MAP_PRIVATE, -1, 0));
> +
> +    if (mapped_addr == brk_page) {
> +        /*
> +         * Heap contents are initialized to zero, as for anonymous mapped
> pages.
> +         * Technically the new pages are already initialized to zero
> since they
> +         * *are* anonymous mapped pages, however we have to take care
> with the
> +         * contents that come from the remaining part of the previous
> page: it
> +         * may contains garbage data due to a previous heap usage (grown
> then
> +         * shrunken).
> +         */
> +        memset(g2h_untagged(bsd_target_brk), 0, brk_page -
> bsd_target_brk);
> +
> +        bsd_target_brk = new_brk;
> +        brk_page = HOST_PAGE_ALIGN(bsd_target_brk);
> +        DEBUGF_BRK(TARGET_ABI_FMT_lx " (mapped_addr == brk_page)\n",
> +            bsd_target_brk);
> +        return bsd_target_brk;
> +    } else if (mapped_addr != -1) {
> +        /*
> +         * Mapped but at wrong address, meaning there wasn't actually
> enough
> +         * space for this brk.
> +         */
> +        target_munmap(mapped_addr, new_alloc_size);
> +        mapped_addr = -1;
> +        DEBUGF_BRK(TARGET_ABI_FMT_lx " (mapped_addr != -1)\n",
> bsd_target_brk);
> +    } else {
> +        DEBUGF_BRK(TARGET_ABI_FMT_lx " (otherwise)\n", bsd_target_brk);
> +    }
> +
> +    /* For everything else, return the previous break. */
> +    return bsd_target_brk;
> +}
> +
>  #endif /* BSD_USER_BSD_MEM_H */
> diff --git a/bsd-user/freebsd/os-syscall.c b/bsd-user/freebsd/os-syscall.c
> index 1db0907504..1b9dca9164 100644
> --- a/bsd-user/freebsd/os-syscall.c
> +++ b/bsd-user/freebsd/os-syscall.c
> @@ -543,6 +543,13 @@ static abi_long freebsd_syscall(void *cpu_env, int
> num, abi_long arg1,
>          break;
>  #endif
>
> +        /*
> +         * Misc
> +         */
> +    case TARGET_FREEBSD_NR_break:
> +        ret = do_obreak(arg1);
> +        break;
> +
>          /*
>           * sys{ctl, arch, call}
>           */
> --
> 2.40.0
>
>

[-- Attachment #2: Type: text/html, Size: 6649 bytes --]

  reply	other threads:[~2023-08-20  4:41 UTC|newest]

Thread overview: 79+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-08-19  9:47 [PATCH 00/22] Implement the mmap system call for FreeBSD Karim Taha
2023-08-19  9:47 ` [PATCH 01/22] Implement struct target_ipc_perm Karim Taha
2023-08-19 14:37   ` Richard Henderson
2023-08-20  4:07   ` Warner Losh
2023-08-19  9:47 ` [PATCH 02/22] Implement struct target_shmid_ds Karim Taha
2023-08-19 14:38   ` Richard Henderson
2023-08-20  4:08   ` Warner Losh
2023-08-19  9:47 ` [PATCH 03/22] Declarations for ipc_perm and shmid_ds conversion functions Karim Taha
2023-08-19 14:40   ` Richard Henderson
2023-08-20  4:08   ` Warner Losh
2023-08-19  9:47 ` [PATCH 04/22] Introduce freebsd/os-misc.h to the source tree Karim Taha
2023-08-19 14:40   ` Richard Henderson
2023-08-20  4:09   ` Warner Losh
2023-08-19  9:47 ` [PATCH 05/22] Implement shm_open2(2) system call Karim Taha
2023-08-19 15:10   ` Richard Henderson
2023-08-20  4:16     ` Warner Losh
2023-08-19  9:47 ` [PATCH 06/22] Implement shm_rename(2) " Karim Taha
2023-08-20  4:18   ` Warner Losh
2023-08-20 14:05   ` Richard Henderson
2023-08-19  9:47 ` [PATCH 07/22] Add bsd-mem.c to meson.build Karim Taha
2023-08-20  4:19   ` Warner Losh
2023-08-20 14:06   ` Richard Henderson
2023-08-19  9:47 ` [PATCH 08/22] Implement target_set_brk function in bsd-mem.c instead of os-syscall.c Karim Taha
2023-08-20  4:22   ` Warner Losh
2023-08-20 14:12   ` Richard Henderson
2023-08-19  9:47 ` [PATCH 09/22] Implement ipc_perm conversion between host and target Karim Taha
2023-08-20  4:23   ` Warner Losh
2023-08-20 14:16   ` Richard Henderson
2023-08-19  9:47 ` [PATCH 10/22] Implement shmid_ds " Karim Taha
2023-08-20  4:25   ` Warner Losh
2023-08-20 14:20   ` Richard Henderson
2023-09-03  8:45     ` Kariiem Taha
2023-09-05  1:43       ` Richard Henderson
2023-08-19  9:47 ` [PATCH 11/22] Introduce bsd-mem.h to the source tree Karim Taha
2023-08-20  4:26   ` Warner Losh
2023-08-20 14:21   ` Richard Henderson
2023-08-19  9:47 ` [PATCH 12/22] Implement mmap(2) and munmap(2) Karim Taha
2023-08-20  4:27   ` Warner Losh
2023-08-20 14:25   ` Richard Henderson
2023-08-19  9:47 ` [PATCH 13/22] Implement mprotect(2) Karim Taha
2023-08-20  4:28   ` Warner Losh
2023-08-20 14:25   ` Richard Henderson
2023-08-19  9:47 ` [PATCH 14/22] Implement msync(2) Karim Taha
2023-08-20  4:34   ` Warner Losh
2023-08-20 14:37   ` Richard Henderson
2023-08-19  9:47 ` [PATCH 15/22] Implement mlock(2), munlock(2), mlockall(2), munlockall(2), madvise(2), minherit(2) Karim Taha
2023-08-20  4:37   ` Warner Losh
2023-08-20 14:42     ` Richard Henderson
2023-08-20 14:43   ` Richard Henderson
2023-08-19  9:48 ` [PATCH 16/22] Implement mincore(2) Karim Taha
2023-08-20  4:37   ` Warner Losh
2023-08-20 14:55   ` Richard Henderson
2023-08-19  9:48 ` [PATCH 17/22] Implement do_obreak function Karim Taha
2023-08-20  4:40   ` Warner Losh [this message]
2023-08-20 15:03   ` Richard Henderson
2023-08-19  9:48 ` [PATCH 18/22] Implement shm_open(2) Karim Taha
2023-08-20  4:42   ` Warner Losh
2023-08-20 15:04   ` Richard Henderson
2023-08-20 15:10     ` Richard Henderson
2023-08-19  9:48 ` [PATCH 19/22] Implement shm_unlink(2) and shmget(2) Karim Taha
2023-08-20  4:42   ` Warner Losh
2023-08-20 15:05   ` Richard Henderson
2023-08-20 15:07   ` Richard Henderson
2023-08-19  9:48 ` [PATCH 20/22] Implement shmctl(2) Karim Taha
2023-08-20  4:43   ` Warner Losh
2023-08-20 15:13   ` Richard Henderson
2023-09-09  1:59     ` Karim Taha
2023-09-09 17:51       ` Richard Henderson
2023-08-19  9:48 ` [PATCH 21/22] Implement shmat(2) and shmdt(2) Karim Taha
2023-08-20  4:44   ` Warner Losh
2023-08-20 15:30   ` Richard Henderson
2023-08-22 18:03     ` Warner Losh
2023-08-22 18:11       ` Richard Henderson
2023-08-22 19:54         ` Warner Losh
2023-08-22 21:00           ` Richard Henderson
2023-08-19  9:48 ` [PATCH 22/22] Add stubs for vadvise(), sbrk() and sstk() Karim Taha
2023-08-20  4:45   ` Warner Losh
2023-08-20 15:35   ` Richard Henderson
2023-08-20 20:42     ` Warner Losh

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=CANCZdfpgdCbnwgHxokDD4jVKHAEoVS+4aY80KwNZdAp7qYxMfA@mail.gmail.com \
    --to=imp@bsdimp.com \
    --cc=kariem.taha2.7@gmail.com \
    --cc=mikael.urankar@gmail.com \
    --cc=qemu-devel@nongnu.org \
    --cc=sson@freebsd.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).