Received-SPF: pass (google.com: domain of linux-kernel+bounces-17247-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:4601:e00::3 as permitted sender) client-ip=2604:1380:4601:e00::3;
Message-ID: <d6bd59f0-ba79-4e76-9fb7-aa07f86e4043@google.com>
Date: Thu, 4 Jan 2024 16:32:39 -0500
Precedence: bulk
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
Subject: Re: [PATCH bpf-next 2/2] selftests/bpf: add inline assembly helpers
 to access array elements
Content-Language: en-US
To: Andrii Nakryiko <andrii.nakryiko@gmail.com>
Cc: Andrii Nakryiko <andrii@kernel.org>, Alexei Starovoitov <ast@kernel.org>,
 Daniel Borkmann <daniel@iogearbox.net>, Song Liu <song@kernel.org>,
 Yonghong Song <yonghong.song@linux.dev>, mattbobrowski@google.com,
 bpf@vger.kernel.org, linux-kernel@vger.kernel.org
References: <20240103153307.553838-1-brho@google.com>
 <20240103153307.553838-3-brho@google.com>
 <CAEf4BzbKT3LbHQSFwpAfoJuhyGy2NpHk7A6ivkFiutN_jnKHYg@mail.gmail.com>
 <bedf07d1-2cd5-4bc8-9e59-a96479a7ff14@google.com>
 <CAEf4BzauYF4DoQLV6AGfFcq3VgP2yi_Pd6pg2vj2Eb7Rt7j0Pg@mail.gmail.com>
From: Barret Rhoden <brho@google.com>
In-Reply-To: <CAEf4BzauYF4DoQLV6AGfFcq3VgP2yi_Pd6pg2vj2Eb7Rt7j0Pg@mail.gmail.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit

On 1/3/24 16:21, Andrii Nakryiko wrote:
> On Wed, Jan 3, 2024 at 12:06 PM Barret Rhoden<brho@google.com>  wrote:
>> On 1/3/24 14:51, Andrii Nakryiko wrote:
>>> I'm curious how bpf_cmp_likely/bpf_cmp_unlikely (just applied to
>>> bpf-next) compares to this?
>> these work great!
>>
>> e.g.
>>
>>           if (bpf_cmp_likely(idx, <, NR_MAP_ELEMS))
>>                   map_elems[idx] = i;
>>
>> works fine.  since that's essentially the code that bpf_array_elem() was
>> trying to replace, i'd rather just use the new bpf_cmp helpers than have
>> the special array_elem helpers.
> ok, cool, thanks for checking! The less special macros, the better.

sorry - turns out it only worked in testing.  in my actual program, i 
still run into issues.  the comparison is done, which is what bpf_cmp 
enforces.  but the compiler is discarding the comparison.  i have more 
info in the other thread, but figured i'd mention it here too.  =(