Hi Tom,
Tom Rini trini@konsulko.com wrote on Mon, 8 Apr 2019 07:52:16 -0400:
On Mon, Apr 08, 2019 at 11:43:59AM +0200, Miquel Raynal wrote:
Hello,
Miquel Raynal miquel.raynal@bootlin.com wrote on Sun, 18 Nov 2018 21:13:47 +0100:
Hello,
Miquel Raynal miquel.raynal@bootlin.com wrote on Sun, 18 Nov 2018 21:11:47 +0100:
A Coverity robot reported an integer handling issue (OVERFLOW_BEFORE_WIDEN) in the potentially overflowing expression:
(mtd_div_by_ws(mtd->size, mtd) - mtd_div_by_ws(offs, mtd)) * mtd_oobavail(mtd, ops)While such overflow will certainly never happen due to the numbers handled, it is cleaner to fix this operation anyway.
The problem is that all the maths include 32-bit quantities, while the result is stored in an explicit 64-bit value.
As maxooblen will just be compared with a size_t, let's change the type of the variable to a size_t. This will not fix anything but will clarify a bit the situation. Then, do an explicit cast to fix Coverity warning.
Addresses-Coverity-ID: 184180 ("Integer handling issues") Signed-off-by: Miquel Raynal miquel.raynal@bootlin.com
I forgot to add the link to the corresponding Travis-CI job: https://travis-ci.org/miquelraynal/u-boot/builds/456679982
Gentle ping on this series (patch 1 & 2). They do apply cleanly on current master.
Tom, shall I expect you to take it or is it Jagan's mission?
My current feeling is this goes via Jagan's tree and I wasn't expecting them in for this release. Thanks!
No problem, I know we are late in the merge process, these are just Coverity fixes, they are not urgent, I just am cleaning my branches and I discover many patches which I thought were merged but actually only exist in my local tree, which is sad :)
Thanks, Miquèl