public inbox for [email protected]
 help / color / mirror / Atom feed
From: Ammar Faizi <[email protected]>
To: Alviro Iskandar Setiawan <[email protected]>
Cc: Jens Axboe <[email protected]>,
	Fernanda Ma'rouf <[email protected]>,
	Pavel Begunkov <[email protected]>,
	Hao Xu <[email protected]>,
	io-uring Mailing List <[email protected]>,
	GNU/Weeb Mailing List <[email protected]>
Subject: Re: [PATCH liburing v1 2/9] setup: Handle `get_page_size()` failure (for aarch64 nolibc support)
Date: Wed, 29 Jun 2022 08:00:27 +0700	[thread overview]
Message-ID: <[email protected]> (raw)
In-Reply-To: <CAOG64qOpHNUO3WP6ve98P3zGEAaykpZP_quo6nce-7=H63s8-w@mail.gmail.com>

On 6/29/22 7:50 AM, Alviro Iskandar Setiawan wrote:
> On Wed, Jun 29, 2022 at 7:28 AM Ammar Faizi wrote:
>>          page_size = get_page_size();
>> +       if (page_size < 0)
>> +               return page_size;
>> +
>>          return rings_size(p, entries, cq_entries, page_size);
>>   }
> 
> the current error handling fallback to 4K if fail on sysconf(_SC_PAGESIZE):
> https://github.com/axboe/liburing/blob/68103b731c34a9f83c181cb33eb424f46f3dcb94/src/arch/generic/lib.h#L10-L19
> with this patch, get_page_size() is only possible to return negative
> value on aarch64.

Ah right, this one needs a revision. Either we fallback to 4K, or
return error if we fail.

> i don't understand why the current master branch code fallback to 4K when fail?

Neither do I. Maybe because 4K is widely used page size?

Jens, can you shed some light on this?

   The current upstream does this:

      - For x86/x86-64, it's hard-coded to 4K. So it can't fail.
      - For other archs, if sysconf(_SC_PAGESIZE) fails, we fallback to 4K.

Now we are going to add aarch64, it uses a group of syscalls to get the page
size. So it may fail. What should we do when we fail?

Fallback to 4K? Or return error code from syscall?
       
-- 
Ammar Faizi

  reply	other threads:[~2022-06-29  1:00 UTC|newest]

Thread overview: 19+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-06-29  0:27 [PATCH liburing v1 0/9] aarch64 nolibc support Ammar Faizi
2022-06-29  0:27 ` [PATCH liburing v1 1/9] CHANGELOG: Fixup missing space Ammar Faizi
2022-06-29  0:27 ` [PATCH liburing v1 2/9] setup: Handle `get_page_size()` failure (for aarch64 nolibc support) Ammar Faizi
2022-06-29  0:50   ` Alviro Iskandar Setiawan
2022-06-29  1:00     ` Ammar Faizi [this message]
2022-06-29 15:00       ` Jens Axboe
2022-06-29 15:17         ` Ammar Faizi
2022-06-29  0:27 ` [PATCH liburing v1 3/9] arch: syscall: Add `__sys_open()` syscall Ammar Faizi
2022-06-29  0:27 ` [PATCH liburing v1 4/9] arch: syscall: Add `__sys_read()` syscall Ammar Faizi
2022-06-29  0:27 ` [PATCH liburing v1 5/9] arch/arm64: Rename aarch64 directory to arm64 Ammar Faizi
2022-06-29 14:48   ` Jens Axboe
2022-06-29 14:51     ` Ammar Faizi
2022-06-29  0:27 ` [PATCH liburing v1 6/9] arch: syscall: Remove `__INTERNAL__LIBURING_SYSCALL_H` checks Ammar Faizi
2022-06-29  0:27 ` [PATCH liburing v1 7/9] arch/arm64: Add `get_page_size()` function Ammar Faizi
2022-06-29  0:31   ` Ammar Faizi
2022-06-29 14:49   ` Jens Axboe
2022-06-29 14:52     ` Ammar Faizi
2022-06-29  0:27 ` [PATCH liburing v1 8/9] arch: Enable nolibc support for arm64 Ammar Faizi
2022-06-29  0:27 ` [PATCH liburing v1 9/9] .github: Enable aarch64 nolibc build for GitHub bot Ammar Faizi

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 \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    /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