[PATCH v2] erofs-utils: mkfs: Implement 'dsunit' alignment on blobdev

[Friendy.Su at sony.com]

OK, now I got your idea. :)


________________________________________
From: Gao Xiang <hsiangkao at linux.alibaba.com>
Sent: Friday, August 22, 2025 17:58
To: Su, Friendy; linux-erofs at lists.ozlabs.org
Cc: Mo, Yuezhang; Palmer, Daniel (SGC)
Subject: Re: [PATCH v2] erofs-utils: mkfs: Implement 'dsunit' alignment on blobdev

On 2025/8/22 17: 56, Gao Xiang wrote: > > > On 2025/8/22 17: 52, Friendy. Su@ sony. com wrote: >> +       if (cfg. c_chunkbits && dsunit && 1u << (cfg. c_chunkbits - g_sbi. blkszbits) < dsunit) { >> +




On 2025/8/22 17:56, Gao Xiang wrote:
>
>
> On 2025/8/22 17:52, Friendy.Su at sony.com wrote:
>> +       if (cfg.c_chunkbits && dsunit && 1u << (cfg.c_chunkbits - g_sbi.blkszbits) < dsunit) {
>> +               erofs_warn("chunksize %u bytes is smaller than dsunit %u blocks, ignore dsunit !",
>> +                               1u << cfg.c_chunkbits, dsunit);
>> +               dsunit = 0;
>> +       }
>>
>> 'ignore dsunit' means set it to default, default dsunit is 0. Is this correct?
>> > Then sbi->bmgr->dsunit will be set to 'dsunit'.
>
> I think it just ignores `dsunit` since the current behavior also
> ignores `dsunit` for blobchunks.
>
> Let's not introduce extra behavior otherwise it could have three
> different behaviors among different mkfs.erofs versions.
>
> So you could just drop `dsunit = 0` line.

So I tend to just kill `dsunit = 0` here, and change

`if (sbi->bmgr->dsunit > 1) {` into
`if (sbi->bmgr->dsunit >= 1u << (cfg.c_chunkbits - g_sbi.blkszbits)) {`

to match the previous `ignore` behavior for dsunit < chunksize.

Thanks,
Gao Xiang