Dear Reinhard Meyer, In message <[email protected]> you wrote: > > > dev = (struct eth_device *)malloc(sizeof *dev); > > + memset(dev, 0, sizeof(*dev)); > > > > sprintf (dev->name, "RTL8139#%d", card_number); > > Apparently its quite common NOT to check malloc()'s possible > NULL return value... At least most NET drivers don't seem to... > > Maybe another RFC to avoid duplicating code: > > malloc_cleared_panic() to allocate and clear memory for a > really required structure and put a proper panic message if that > fails. Assuming that continuing u-boot once a driver cannot even > be initialized is futile, that would save even more code in each > driver.
Don't invent the wheel. If you really want to take that route, then copy existing solutions from other projects. Some of them use xmalloc() for this purpose; see for example BusyBox: http://git.busybox.net/busybox/tree/libbb/xfuncs_printf.c lines 44...51 But note that panicing is NOT always the best thing to do. This shouldbe reserved for really unrecoverable cases only. Even if you cannot allocate a struct that is essential for your network driver, then all that is not working is this network driver, so this is NOT a reason to panic U-Boot. If someone cuts the network cable or pulls the plug the end effect is the same, and you don;t want U-Boot to panic because of htat, or do you? Error handling is important, and needs to be done in a sensible way. Best regards, Wolfgang Denk -- DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: [email protected] In the beginning, there was nothing, which exploded. - Terry Pratchett, _Lords and Ladies_ _______________________________________________ U-Boot mailing list [email protected] http://lists.denx.de/mailman/listinfo/u-boot
