Re: [PrintLib] Question about ASSERT_UNICODE_BUFFER

[Heyi Guo <heyi.guo@linaro.org>]

Thanks for your confirmation.

Regards,

Gary (Heyi Guo)


在 12/2/2017 1:23 AM, Kinney, Michael D 写道:
>
> Gary,
>
> The correct solution is to align the Unicode string buffer before 
> using those APIs.
>
> The assert for unaligned Unicode string has been in place for over 10 
> years, and we have always
>
> been able to address this type of issue by fixing the calling code.
>
> Mike
>
> *From:*heyi.guo [mailto:heyi.guo@linaro.org]
> *Sent:* Thursday, November 30, 2017 10:58 PM
> *To:* edk2-devel@lists.01.org
> *Cc:* Gao, Liming <liming.gao@intel.com>; Kinney, Michael D 
> <michael.d.kinney@intel.com>
> *Subject:* [PrintLib] Question about ASSERT_UNICODE_BUFFER
>
> Hi folks,
>
> I think platforms supported in EDK2 should enable unaligned data 
> access, so a CHAR16 * pointer still works even if the pointer itself 
> is not aligned by 2. Therefore, can we remove ASSERT_UNICODE_BUFFER 
> statements in PrintLib for Unicode operations? If the platform doesn't 
> support unaligned access, it can generate some kind of exception 
> instead of software assert.
>
> The background is that we are importing a 3rd party driver and it 
> packs a unicode string buffer into a structure, and #pragma pack(1) is 
> used around the definition of this structure. So the alignment of the 
> string buffer depends on the address of the whole structure and the 
> members before it; it happened to be not aligned and we got assert 
> during system post.
>
> Please let me know your comments.
>
> Thanks,
>
> Gary (Heyi Guo)
>